talexander
/
freebsd_amp_hwpstate
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
freebsd_amp_hwpstate/sys/kern/kern_mutex.c

1685 lines
39 KiB
C
Raw Normal View History

Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
/*-
* Copyright (c) 1998 Berkeley Software Design, Inc. All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* 3. Berkeley Software Design Inc's name may not be used to endorse or
* promote products derived from this software without specific prior
* written permission.
*
* THIS SOFTWARE IS PROVIDED BY BERKELEY SOFTWARE DESIGN INC ``AS IS'' AND
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL BERKELEY SOFTWARE DESIGN INC BE LIABLE
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
* SUCH DAMAGE.
*
* from BSDI $Id: mutex_witness.c,v 1.1.2.20 2000/04/27 03:10:27 cp Exp $
- Make the mutex code almost completely machine independent. This greatly reducues the maintenance load for the mutex code. The only MD portions of the mutex code are in machine/mutex.h now, which include the assembly macros for handling mutexes as well as optionally overriding the mutex micro-operations. For example, we use optimized micro-ops on the x86 platform #ifndef I386_CPU. - Change the behavior of the SMP_DEBUG kernel option. In the new code, mtx_assert() only depends on INVARIANTS, allowing other kernel developers to have working mutex assertiions without having to include all of the mutex debugging code. The SMP_DEBUG kernel option has been renamed to MUTEX_DEBUG and now just controls extra mutex debugging code. - Abolish the ugly mtx_f hack. Instead, we dynamically allocate seperate mtx_debug structures on the fly in mtx_init, except for mutexes that are initiated very early in the boot process. These mutexes are declared using a special MUTEX_DECLARE() macro, and use a new flag MTX_COLD when calling mtx_init. This is still somewhat hackish, but it is less evil than the mtx_f filler struct, and the mtx struct is now the same size with and without mutex debugging code. - Add some micro-micro-operation macros for doing the actual atomic operations on the mutex mtx_lock field to make it easier for other archs to override/optimize mutex ops if needed. These new tiny ops also clean up the code in some places by replacing long atomic operation function calls that spanned 2-3 lines with a short 1-line macro call. - Don't call mi_switch() from mtx_enter_hard() when we block while trying to obtain a sleep mutex. Calling mi_switch() would bogusly release Giant before switching to the next process. Instead, inline most of the code from mi_switch() in the mtx_enter_hard() function. Note that when we finally kill Giant we can back this out and go back to calling mi_switch().
2000-10-20 07:26:37 +00:00
* and BSDI $Id: synch_machdep.c,v 2.3.2.39 2000/04/27 03:10:25 cp Exp $
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
* $FreeBSD$
*/
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
/*
* Machine independent bits of mutex implementation and implementation of
* `witness' structure & related debugging routines.
*/
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
/*
* Main Entry: witness
* Pronunciation: 'wit-n&s
* Function: noun
* Etymology: Middle English witnesse, from Old English witnes knowledge,
* testimony, witness, from 2wit
* Date: before 12th century
* 1 : attestation of a fact or event : TESTIMONY
* 2 : one that gives evidence; specifically : one who testifies in
* a cause or before a judicial tribunal
* 3 : one asked to be present at a transaction so as to be able to
* testify to its having taken place
* 4 : one who has personal knowledge of something
* 5 a : something serving as evidence or proof : SIGN
* b : public affirmation by word or example of usually
* religious faith or conviction <the heroic witness to divine
* life -- Pilot>
* 6 capitalized : a member of the Jehovah's Witnesses
*/
Include the right headers to get the DDB #define and the db_active variable.
2000-11-15 22:08:16 +00:00
#include "opt_ddb.h"
- Use MUTEX_DECLARE() and MTX_COLD for the WITNESS code's internal mutex so it can function before malloc(9) is up and running. - Add two new options WITNESS_DDB and WITNESS_SKIPSPIN. If WITNESS_SKIPSPIN is enabled, then spin mutexes are ignored by the WITNESS code. If WITNESS_DDB is turned on and DDB is compiled into the kernel, then the kernel will drop into DDB when either a lock hierarchy violation occurs or mutexes are held when going to sleep. - Add some new sysctls: debug.witness_ddb is a read-write sysctl that corresponds to WITNESS_DDB. The kernel option merely changes the default value to on at boot. debug.witness_skipspin is a read-only sysctl that one can use to determine if the kernel was compiled with WITNESS_SKIPSPIN. - Wipe out the BSD/OS-specific lock order lists. We get to build our own lists now as we add mutexes to the kernel.
2000-10-27 02:59:30 +00:00
#include "opt_witness.h"
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
#include <sys/param.h>
- Make the mutex code almost completely machine independent. This greatly reducues the maintenance load for the mutex code. The only MD portions of the mutex code are in machine/mutex.h now, which include the assembly macros for handling mutexes as well as optionally overriding the mutex micro-operations. For example, we use optimized micro-ops on the x86 platform #ifndef I386_CPU. - Change the behavior of the SMP_DEBUG kernel option. In the new code, mtx_assert() only depends on INVARIANTS, allowing other kernel developers to have working mutex assertiions without having to include all of the mutex debugging code. The SMP_DEBUG kernel option has been renamed to MUTEX_DEBUG and now just controls extra mutex debugging code. - Abolish the ugly mtx_f hack. Instead, we dynamically allocate seperate mtx_debug structures on the fly in mtx_init, except for mutexes that are initiated very early in the boot process. These mutexes are declared using a special MUTEX_DECLARE() macro, and use a new flag MTX_COLD when calling mtx_init. This is still somewhat hackish, but it is less evil than the mtx_f filler struct, and the mtx struct is now the same size with and without mutex debugging code. - Add some micro-micro-operation macros for doing the actual atomic operations on the mutex mtx_lock field to make it easier for other archs to override/optimize mutex ops if needed. These new tiny ops also clean up the code in some places by replacing long atomic operation function calls that spanned 2-3 lines with a short 1-line macro call. - Don't call mi_switch() from mtx_enter_hard() when we block while trying to obtain a sleep mutex. Calling mi_switch() would bogusly release Giant before switching to the next process. Instead, inline most of the code from mi_switch() in the mtx_enter_hard() function. Note that when we finally kill Giant we can back this out and go back to calling mi_switch().
2000-10-20 07:26:37 +00:00
#include <sys/bus.h>
#include <sys/kernel.h>
#include <sys/malloc.h>
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
#include <sys/proc.h>
- Use MUTEX_DECLARE() and MTX_COLD for the WITNESS code's internal mutex so it can function before malloc(9) is up and running. - Add two new options WITNESS_DDB and WITNESS_SKIPSPIN. If WITNESS_SKIPSPIN is enabled, then spin mutexes are ignored by the WITNESS code. If WITNESS_DDB is turned on and DDB is compiled into the kernel, then the kernel will drop into DDB when either a lock hierarchy violation occurs or mutexes are held when going to sleep. - Add some new sysctls: debug.witness_ddb is a read-write sysctl that corresponds to WITNESS_DDB. The kernel option merely changes the default value to on at boot. debug.witness_skipspin is a read-only sysctl that one can use to determine if the kernel was compiled with WITNESS_SKIPSPIN. - Wipe out the BSD/OS-specific lock order lists. We get to build our own lists now as we add mutexes to the kernel.
2000-10-27 02:59:30 +00:00
#include <sys/sysctl.h>
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
#include <sys/systm.h>
- Make the mutex code almost completely machine independent. This greatly reducues the maintenance load for the mutex code. The only MD portions of the mutex code are in machine/mutex.h now, which include the assembly macros for handling mutexes as well as optionally overriding the mutex micro-operations. For example, we use optimized micro-ops on the x86 platform #ifndef I386_CPU. - Change the behavior of the SMP_DEBUG kernel option. In the new code, mtx_assert() only depends on INVARIANTS, allowing other kernel developers to have working mutex assertiions without having to include all of the mutex debugging code. The SMP_DEBUG kernel option has been renamed to MUTEX_DEBUG and now just controls extra mutex debugging code. - Abolish the ugly mtx_f hack. Instead, we dynamically allocate seperate mtx_debug structures on the fly in mtx_init, except for mutexes that are initiated very early in the boot process. These mutexes are declared using a special MUTEX_DECLARE() macro, and use a new flag MTX_COLD when calling mtx_init. This is still somewhat hackish, but it is less evil than the mtx_f filler struct, and the mtx struct is now the same size with and without mutex debugging code. - Add some micro-micro-operation macros for doing the actual atomic operations on the mutex mtx_lock field to make it easier for other archs to override/optimize mutex ops if needed. These new tiny ops also clean up the code in some places by replacing long atomic operation function calls that spanned 2-3 lines with a short 1-line macro call. - Don't call mi_switch() from mtx_enter_hard() when we block while trying to obtain a sleep mutex. Calling mi_switch() would bogusly release Giant before switching to the next process. Instead, inline most of the code from mi_switch() in the mtx_enter_hard() function. Note that when we finally kill Giant we can back this out and go back to calling mi_switch().
2000-10-20 07:26:37 +00:00
#include <sys/vmmeter.h>
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
#include <sys/ktr.h>
- Make the mutex code almost completely machine independent. This greatly reducues the maintenance load for the mutex code. The only MD portions of the mutex code are in machine/mutex.h now, which include the assembly macros for handling mutexes as well as optionally overriding the mutex micro-operations. For example, we use optimized micro-ops on the x86 platform #ifndef I386_CPU. - Change the behavior of the SMP_DEBUG kernel option. In the new code, mtx_assert() only depends on INVARIANTS, allowing other kernel developers to have working mutex assertiions without having to include all of the mutex debugging code. The SMP_DEBUG kernel option has been renamed to MUTEX_DEBUG and now just controls extra mutex debugging code. - Abolish the ugly mtx_f hack. Instead, we dynamically allocate seperate mtx_debug structures on the fly in mtx_init, except for mutexes that are initiated very early in the boot process. These mutexes are declared using a special MUTEX_DECLARE() macro, and use a new flag MTX_COLD when calling mtx_init. This is still somewhat hackish, but it is less evil than the mtx_f filler struct, and the mtx struct is now the same size with and without mutex debugging code. - Add some micro-micro-operation macros for doing the actual atomic operations on the mutex mtx_lock field to make it easier for other archs to override/optimize mutex ops if needed. These new tiny ops also clean up the code in some places by replacing long atomic operation function calls that spanned 2-3 lines with a short 1-line macro call. - Don't call mi_switch() from mtx_enter_hard() when we block while trying to obtain a sleep mutex. Calling mi_switch() would bogusly release Giant before switching to the next process. Instead, inline most of the code from mi_switch() in the mtx_enter_hard() function. Note that when we finally kill Giant we can back this out and go back to calling mi_switch().
2000-10-20 07:26:37 +00:00
#include <machine/atomic.h>
#include <machine/bus.h>
#include <machine/clock.h>
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
#include <machine/cpu.h>
- Make the mutex code almost completely machine independent. This greatly reducues the maintenance load for the mutex code. The only MD portions of the mutex code are in machine/mutex.h now, which include the assembly macros for handling mutexes as well as optionally overriding the mutex micro-operations. For example, we use optimized micro-ops on the x86 platform #ifndef I386_CPU. - Change the behavior of the SMP_DEBUG kernel option. In the new code, mtx_assert() only depends on INVARIANTS, allowing other kernel developers to have working mutex assertiions without having to include all of the mutex debugging code. The SMP_DEBUG kernel option has been renamed to MUTEX_DEBUG and now just controls extra mutex debugging code. - Abolish the ugly mtx_f hack. Instead, we dynamically allocate seperate mtx_debug structures on the fly in mtx_init, except for mutexes that are initiated very early in the boot process. These mutexes are declared using a special MUTEX_DECLARE() macro, and use a new flag MTX_COLD when calling mtx_init. This is still somewhat hackish, but it is less evil than the mtx_f filler struct, and the mtx struct is now the same size with and without mutex debugging code. - Add some micro-micro-operation macros for doing the actual atomic operations on the mutex mtx_lock field to make it easier for other archs to override/optimize mutex ops if needed. These new tiny ops also clean up the code in some places by replacing long atomic operation function calls that spanned 2-3 lines with a short 1-line macro call. - Don't call mi_switch() from mtx_enter_hard() when we block while trying to obtain a sleep mutex. Calling mi_switch() would bogusly release Giant before switching to the next process. Instead, inline most of the code from mi_switch() in the mtx_enter_hard() function. Note that when we finally kill Giant we can back this out and go back to calling mi_switch().
2000-10-20 07:26:37 +00:00
Include the right headers to get the DDB #define and the db_active variable.
2000-11-15 22:08:16 +00:00
#include <ddb/ddb.h>
- Make the mutex code almost completely machine independent. This greatly reducues the maintenance load for the mutex code. The only MD portions of the mutex code are in machine/mutex.h now, which include the assembly macros for handling mutexes as well as optionally overriding the mutex micro-operations. For example, we use optimized micro-ops on the x86 platform #ifndef I386_CPU. - Change the behavior of the SMP_DEBUG kernel option. In the new code, mtx_assert() only depends on INVARIANTS, allowing other kernel developers to have working mutex assertiions without having to include all of the mutex debugging code. The SMP_DEBUG kernel option has been renamed to MUTEX_DEBUG and now just controls extra mutex debugging code. - Abolish the ugly mtx_f hack. Instead, we dynamically allocate seperate mtx_debug structures on the fly in mtx_init, except for mutexes that are initiated very early in the boot process. These mutexes are declared using a special MUTEX_DECLARE() macro, and use a new flag MTX_COLD when calling mtx_init. This is still somewhat hackish, but it is less evil than the mtx_f filler struct, and the mtx struct is now the same size with and without mutex debugging code. - Add some micro-micro-operation macros for doing the actual atomic operations on the mutex mtx_lock field to make it easier for other archs to override/optimize mutex ops if needed. These new tiny ops also clean up the code in some places by replacing long atomic operation function calls that spanned 2-3 lines with a short 1-line macro call. - Don't call mi_switch() from mtx_enter_hard() when we block while trying to obtain a sleep mutex. Calling mi_switch() would bogusly release Giant before switching to the next process. Instead, inline most of the code from mi_switch() in the mtx_enter_hard() function. Note that when we finally kill Giant we can back this out and go back to calling mi_switch().
2000-10-20 07:26:37 +00:00
#include <vm/vm.h>
#include <vm/vm_extern.h>
#include <sys/mutex.h>
/*
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
* The WITNESS-enabled mutex debug structure.
- Make the mutex code almost completely machine independent. This greatly reducues the maintenance load for the mutex code. The only MD portions of the mutex code are in machine/mutex.h now, which include the assembly macros for handling mutexes as well as optionally overriding the mutex micro-operations. For example, we use optimized micro-ops on the x86 platform #ifndef I386_CPU. - Change the behavior of the SMP_DEBUG kernel option. In the new code, mtx_assert() only depends on INVARIANTS, allowing other kernel developers to have working mutex assertiions without having to include all of the mutex debugging code. The SMP_DEBUG kernel option has been renamed to MUTEX_DEBUG and now just controls extra mutex debugging code. - Abolish the ugly mtx_f hack. Instead, we dynamically allocate seperate mtx_debug structures on the fly in mtx_init, except for mutexes that are initiated very early in the boot process. These mutexes are declared using a special MUTEX_DECLARE() macro, and use a new flag MTX_COLD when calling mtx_init. This is still somewhat hackish, but it is less evil than the mtx_f filler struct, and the mtx struct is now the same size with and without mutex debugging code. - Add some micro-micro-operation macros for doing the actual atomic operations on the mutex mtx_lock field to make it easier for other archs to override/optimize mutex ops if needed. These new tiny ops also clean up the code in some places by replacing long atomic operation function calls that spanned 2-3 lines with a short 1-line macro call. - Don't call mi_switch() from mtx_enter_hard() when we block while trying to obtain a sleep mutex. Calling mi_switch() would bogusly release Giant before switching to the next process. Instead, inline most of the code from mi_switch() in the mtx_enter_hard() function. Note that when we finally kill Giant we can back this out and go back to calling mi_switch().
2000-10-20 07:26:37 +00:00
*/
Split the WITNESS and MUTEX_DEBUG options apart so that WITNESS does not depend on MUTEX_DEBUG. The MUTEX_DEBUG option turns on extra assertions and checks to verify that mutexes themselves are implemented properly. The WITNESS option uses extra checks and diagnostics to verify that other code is using mutexes properly.
2000-12-01 00:10:59 +00:00
#ifdef WITNESS
Move most of sys/mutex.h into kern/kern_mutex.c, thereby making the mutex inline functions non-inlined. Hide parts of the mutex implementation that should not be exposed. Make sure that WITNESS code is not executed during boot until the mutexes are fully initialized by SI_SUB_MUTEX (the original motivation for this commit). Submitted by: peter
2001-01-21 22:34:43 +00:00
struct mtx_debug {
struct witness *mtxd_witness;
LIST_ENTRY(mtx) mtxd_held;
const char *mtxd_file;
int mtxd_line;
};
- Don't use a union and fun tricks to shave one extra pointer off of struct mtx right now as it makes debugging harder. When we are in optimizing mode, we can revisit this. - Fix the KTR trace messages to use %p rather than 0x%p to avoid duplicate 0x's in KTR output. - During witness_fixup, release Giant so that witness doesn't get confused. Also, grab all_mtx while walking the list of mutexes. - Remove w_sleep and w_recurse. Instead, perform checks on mutexes using the mutex's mtx_flags field. - Allow debug.witness_ddb and debug.witness_skipspin to be set from the loader. - Add Giant to the front of existing order_list entries to help ensure Giant is always first. - Add an order entry for the various proc locks. Note that this only helps keep proc in order mostly as the allproc and proctree mutexes are only obtained during a lockmgr operation on the specified mutex.
2001-01-24 10:57:01 +00:00
#define mtx_held mtx_debug->mtxd_held
#define mtx_file mtx_debug->mtxd_file
#define mtx_line mtx_debug->mtxd_line
#define mtx_witness mtx_debug->mtxd_witness
Move most of sys/mutex.h into kern/kern_mutex.c, thereby making the mutex inline functions non-inlined. Hide parts of the mutex implementation that should not be exposed. Make sure that WITNESS code is not executed during boot until the mutexes are fully initialized by SI_SUB_MUTEX (the original motivation for this commit). Submitted by: peter
2001-01-21 22:34:43 +00:00
#endif /* WITNESS */
/*
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
* Internal utility macros.
Move most of sys/mutex.h into kern/kern_mutex.c, thereby making the mutex inline functions non-inlined. Hide parts of the mutex implementation that should not be exposed. Make sure that WITNESS code is not executed during boot until the mutexes are fully initialized by SI_SUB_MUTEX (the original motivation for this commit). Submitted by: peter
2001-01-21 22:34:43 +00:00
*/
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
#define mtx_unowned(m) ((m)->mtx_lock == MTX_UNOWNED)
Move most of sys/mutex.h into kern/kern_mutex.c, thereby making the mutex inline functions non-inlined. Hide parts of the mutex implementation that should not be exposed. Make sure that WITNESS code is not executed during boot until the mutexes are fully initialized by SI_SUB_MUTEX (the original motivation for this commit). Submitted by: peter
2001-01-21 22:34:43 +00:00
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
#define mtx_owner(m) (mtx_unowned((m)) ? NULL \
: (struct proc *)((m)->mtx_lock & MTX_FLAGMASK))
Move most of sys/mutex.h into kern/kern_mutex.c, thereby making the mutex inline functions non-inlined. Hide parts of the mutex implementation that should not be exposed. Make sure that WITNESS code is not executed during boot until the mutexes are fully initialized by SI_SUB_MUTEX (the original motivation for this commit). Submitted by: peter
2001-01-21 22:34:43 +00:00
Implement a unified run queue and adjust priority levels accordingly. - All processes go into the same array of queues, with different scheduling classes using different portions of the array. This allows user processes to have their priorities propogated up into interrupt thread range if need be. - I chose 64 run queues as an arbitrary number that is greater than 32. We used to have 4 separate arrays of 32 queues each, so this may not be optimal. The new run queue code was written with this in mind; changing the number of run queues only requires changing constants in runq.h and adjusting the priority levels. - The new run queue code takes the run queue as a parameter. This is intended to be used to create per-cpu run queues. Implement wrappers for compatibility with the old interface which pass in the global run queue structure. - Group the priority level, user priority, native priority (before propogation) and the scheduling class into a struct priority. - Change any hard coded priority levels that I found to use symbolic constants (TTIPRI and TTOPRI). - Remove the curpriority global variable and use that of curproc. This was used to detect when a process' priority had lowered and it should yield. We now effectively yield on every interrupt. - Activate propogate_priority(). It should now have the desired effect without needing to also propogate the scheduling class. - Temporarily comment out the call to vm_page_zero_idle() in the idle loop. It interfered with propogate_priority() because the idle process needed to do a non-blocking acquire of Giant and then other processes would try to propogate their priority onto it. The idle process should not do anything except idle. vm_page_zero_idle() will return in the form of an idle priority kernel thread which is woken up at apprioriate times by the vm system. - Update struct kinfo_proc to the new priority interface. Deliberately change its size by adjusting the spare fields. It remained the same size, but the layout has changed, so userland processes that use it would parse the data incorrectly. The size constraint should really be changed to an arbitrary version number. Also add a debug.sizeof sysctl node for struct kinfo_proc.
2001-02-12 00:20:08 +00:00
#define SET_PRIO(p, pri) (p)->p_pri.pri_level = (pri)
Move most of sys/mutex.h into kern/kern_mutex.c, thereby making the mutex inline functions non-inlined. Hide parts of the mutex implementation that should not be exposed. Make sure that WITNESS code is not executed during boot until the mutexes are fully initialized by SI_SUB_MUTEX (the original motivation for this commit). Submitted by: peter
2001-01-21 22:34:43 +00:00
/*
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
* Early WITNESS-enabled declarations.
Move most of sys/mutex.h into kern/kern_mutex.c, thereby making the mutex inline functions non-inlined. Hide parts of the mutex implementation that should not be exposed. Make sure that WITNESS code is not executed during boot until the mutexes are fully initialized by SI_SUB_MUTEX (the original motivation for this commit). Submitted by: peter
2001-01-21 22:34:43 +00:00
*/
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
#ifdef WITNESS
Move most of sys/mutex.h into kern/kern_mutex.c, thereby making the mutex inline functions non-inlined. Hide parts of the mutex implementation that should not be exposed. Make sure that WITNESS code is not executed during boot until the mutexes are fully initialized by SI_SUB_MUTEX (the original motivation for this commit). Submitted by: peter
2001-01-21 22:34:43 +00:00
/*
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
* Internal WITNESS routines which must be prototyped early.
*
* XXX: When/if witness code is cleaned up, it would be wise to place all
* witness prototyping early in this file.
*/
static void witness_init(struct mtx *, int flag);
static void witness_destroy(struct mtx *);
static void witness_display(void(*)(const char *fmt, ...));
Move most of sys/mutex.h into kern/kern_mutex.c, thereby making the mutex inline functions non-inlined. Hide parts of the mutex implementation that should not be exposed. Make sure that WITNESS code is not executed during boot until the mutexes are fully initialized by SI_SUB_MUTEX (the original motivation for this commit). Submitted by: peter
2001-01-21 22:34:43 +00:00
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
MALLOC_DEFINE(M_WITNESS, "witness", "witness mtx_debug structure");
Move most of sys/mutex.h into kern/kern_mutex.c, thereby making the mutex inline functions non-inlined. Hide parts of the mutex implementation that should not be exposed. Make sure that WITNESS code is not executed during boot until the mutexes are fully initialized by SI_SUB_MUTEX (the original motivation for this commit). Submitted by: peter
2001-01-21 22:34:43 +00:00
/* All mutexes in system (used for debug/panic) */
- Don't use a union and fun tricks to shave one extra pointer off of struct mtx right now as it makes debugging harder. When we are in optimizing mode, we can revisit this. - Fix the KTR trace messages to use %p rather than 0x%p to avoid duplicate 0x's in KTR output. - During witness_fixup, release Giant so that witness doesn't get confused. Also, grab all_mtx while walking the list of mutexes. - Remove w_sleep and w_recurse. Instead, perform checks on mutexes using the mutex's mtx_flags field. - Allow debug.witness_ddb and debug.witness_skipspin to be set from the loader. - Add Giant to the front of existing order_list entries to help ensure Giant is always first. - Add an order entry for the various proc locks. Note that this only helps keep proc in order mostly as the allproc and proctree mutexes are only obtained during a lockmgr operation on the specified mutex.
2001-01-24 10:57:01 +00:00
static struct mtx_debug all_mtx_debug = { NULL, {NULL, NULL}, NULL, 0 };
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
Remove MUTEX_DECLARE() and MTX_COLD. Instead, postpone full mutex initialization until after malloc() is safe to call, then iterate through all mutexes and complete their initialization. This change is necessary in order to avoid some circular bootstrapping dependencies.
2001-01-21 07:52:20 +00:00
/*
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
* This global is set to 0 once it becomes safe to use the witness code.
Remove MUTEX_DECLARE() and MTX_COLD. Instead, postpone full mutex initialization until after malloc() is safe to call, then iterate through all mutexes and complete their initialization. This change is necessary in order to avoid some circular bootstrapping dependencies.
2001-01-21 07:52:20 +00:00
*/
static int witness_cold = 1;
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
Split the WITNESS and MUTEX_DEBUG options apart so that WITNESS does not depend on MUTEX_DEBUG. The MUTEX_DEBUG option turns on extra assertions and checks to verify that mutexes themselves are implemented properly. The WITNESS option uses extra checks and diagnostics to verify that other code is using mutexes properly.
2000-12-01 00:10:59 +00:00
#else /* WITNESS */
Move most of sys/mutex.h into kern/kern_mutex.c, thereby making the mutex inline functions non-inlined. Hide parts of the mutex implementation that should not be exposed. Make sure that WITNESS code is not executed during boot until the mutexes are fully initialized by SI_SUB_MUTEX (the original motivation for this commit). Submitted by: peter
2001-01-21 22:34:43 +00:00
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
/* XXX XXX XXX
* flag++ is sleazoid way of shuting up warning
Move most of sys/mutex.h into kern/kern_mutex.c, thereby making the mutex inline functions non-inlined. Hide parts of the mutex implementation that should not be exposed. Make sure that WITNESS code is not executed during boot until the mutexes are fully initialized by SI_SUB_MUTEX (the original motivation for this commit). Submitted by: peter
2001-01-21 22:34:43 +00:00
*/
#define witness_init(m, flag) flag++
#define witness_destroy(m)
#define witness_try_enter(m, t, f, l)
Split the WITNESS and MUTEX_DEBUG options apart so that WITNESS does not depend on MUTEX_DEBUG. The MUTEX_DEBUG option turns on extra assertions and checks to verify that mutexes themselves are implemented properly. The WITNESS option uses extra checks and diagnostics to verify that other code is using mutexes properly.
2000-12-01 00:10:59 +00:00
#endif /* WITNESS */
- Make the mutex code almost completely machine independent. This greatly reducues the maintenance load for the mutex code. The only MD portions of the mutex code are in machine/mutex.h now, which include the assembly macros for handling mutexes as well as optionally overriding the mutex micro-operations. For example, we use optimized micro-ops on the x86 platform #ifndef I386_CPU. - Change the behavior of the SMP_DEBUG kernel option. In the new code, mtx_assert() only depends on INVARIANTS, allowing other kernel developers to have working mutex assertiions without having to include all of the mutex debugging code. The SMP_DEBUG kernel option has been renamed to MUTEX_DEBUG and now just controls extra mutex debugging code. - Abolish the ugly mtx_f hack. Instead, we dynamically allocate seperate mtx_debug structures on the fly in mtx_init, except for mutexes that are initiated very early in the boot process. These mutexes are declared using a special MUTEX_DECLARE() macro, and use a new flag MTX_COLD when calling mtx_init. This is still somewhat hackish, but it is less evil than the mtx_f filler struct, and the mtx struct is now the same size with and without mutex debugging code. - Add some micro-micro-operation macros for doing the actual atomic operations on the mutex mtx_lock field to make it easier for other archs to override/optimize mutex ops if needed. These new tiny ops also clean up the code in some places by replacing long atomic operation function calls that spanned 2-3 lines with a short 1-line macro call. - Don't call mi_switch() from mtx_enter_hard() when we block while trying to obtain a sleep mutex. Calling mi_switch() would bogusly release Giant before switching to the next process. Instead, inline most of the code from mi_switch() in the mtx_enter_hard() function. Note that when we finally kill Giant we can back this out and go back to calling mi_switch().
2000-10-20 07:26:37 +00:00
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
/*
* All mutex locks in system are kept on the all_mtx list.
*/
- Don't use a union and fun tricks to shave one extra pointer off of struct mtx right now as it makes debugging harder. When we are in optimizing mode, we can revisit this. - Fix the KTR trace messages to use %p rather than 0x%p to avoid duplicate 0x's in KTR output. - During witness_fixup, release Giant so that witness doesn't get confused. Also, grab all_mtx while walking the list of mutexes. - Remove w_sleep and w_recurse. Instead, perform checks on mutexes using the mutex's mtx_flags field. - Allow debug.witness_ddb and debug.witness_skipspin to be set from the loader. - Add Giant to the front of existing order_list entries to help ensure Giant is always first. - Add an order entry for the various proc locks. Note that this only helps keep proc in order mostly as the allproc and proctree mutexes are only obtained during a lockmgr operation on the specified mutex.
2001-01-24 10:57:01 +00:00
static struct mtx all_mtx = { MTX_UNOWNED, 0, 0, 0, "All mutexes queue head",
TAILQ_HEAD_INITIALIZER(all_mtx.mtx_blocked),
{ NULL, NULL }, &all_mtx, &all_mtx,
#ifdef WITNESS
&all_mtx_debug
#else
NULL
#endif
};
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
/*
* Global variables for book keeping.
*/
- Make the mutex code almost completely machine independent. This greatly reducues the maintenance load for the mutex code. The only MD portions of the mutex code are in machine/mutex.h now, which include the assembly macros for handling mutexes as well as optionally overriding the mutex micro-operations. For example, we use optimized micro-ops on the x86 platform #ifndef I386_CPU. - Change the behavior of the SMP_DEBUG kernel option. In the new code, mtx_assert() only depends on INVARIANTS, allowing other kernel developers to have working mutex assertiions without having to include all of the mutex debugging code. The SMP_DEBUG kernel option has been renamed to MUTEX_DEBUG and now just controls extra mutex debugging code. - Abolish the ugly mtx_f hack. Instead, we dynamically allocate seperate mtx_debug structures on the fly in mtx_init, except for mutexes that are initiated very early in the boot process. These mutexes are declared using a special MUTEX_DECLARE() macro, and use a new flag MTX_COLD when calling mtx_init. This is still somewhat hackish, but it is less evil than the mtx_f filler struct, and the mtx struct is now the same size with and without mutex debugging code. - Add some micro-micro-operation macros for doing the actual atomic operations on the mutex mtx_lock field to make it easier for other archs to override/optimize mutex ops if needed. These new tiny ops also clean up the code in some places by replacing long atomic operation function calls that spanned 2-3 lines with a short 1-line macro call. - Don't call mi_switch() from mtx_enter_hard() when we block while trying to obtain a sleep mutex. Calling mi_switch() would bogusly release Giant before switching to the next process. Instead, inline most of the code from mi_switch() in the mtx_enter_hard() function. Note that when we finally kill Giant we can back this out and go back to calling mi_switch().
2000-10-20 07:26:37 +00:00
static int mtx_cur_cnt;
static int mtx_max_cnt;
- Place back STR string declarations for lock/unlock strings used for KTR_LOCK tracing in order to avoid duplication. - Insert some tracepoints back into the mutex acq/rel code, thus ensuring that we can trace all lock acq/rel's again. - All CURPROC != NULL checks are MPASS()es (under MUTEX_DEBUG) because they signify a serious mutex corruption. - Change up some KASSERT()s to MPASS()es, and vice-versa, depending on the type of problem we're debugging (INVARIANTS is used here to check that the API is being used properly whereas MUTEX_DEBUG is used to ensure that something general isn't happening that will have bad impact on mutex locks). Reminded by: jhb, jake, asmodai
2001-02-11 02:54:16 +00:00
/*
* Couple of strings for KTR_LOCK tracing in order to avoid duplicates.
*/
char STR_mtx_lock_slp[] = "GOT (sleep) %s [%p] r=%d at %s:%d";
char STR_mtx_unlock_slp[] = "REL (sleep) %s [%p] r=%d at %s:%d";
char STR_mtx_lock_spn[] = "GOT (spin) %s [%p] r=%d at %s:%d";
char STR_mtx_unlock_spn[] = "REL (spin) %s [%p] r=%d at %s:%d";
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
/*
* Prototypes for non-exported routines.
*
* NOTE: Prototypes for witness routines are placed at the bottom of the file.
*/
Move most of sys/mutex.h into kern/kern_mutex.c, thereby making the mutex inline functions non-inlined. Hide parts of the mutex implementation that should not be exposed. Make sure that WITNESS code is not executed during boot until the mutexes are fully initialized by SI_SUB_MUTEX (the original motivation for this commit). Submitted by: peter
2001-01-21 22:34:43 +00:00
static void propagate_priority(struct proc *);
- Make the mutex code almost completely machine independent. This greatly reducues the maintenance load for the mutex code. The only MD portions of the mutex code are in machine/mutex.h now, which include the assembly macros for handling mutexes as well as optionally overriding the mutex micro-operations. For example, we use optimized micro-ops on the x86 platform #ifndef I386_CPU. - Change the behavior of the SMP_DEBUG kernel option. In the new code, mtx_assert() only depends on INVARIANTS, allowing other kernel developers to have working mutex assertiions without having to include all of the mutex debugging code. The SMP_DEBUG kernel option has been renamed to MUTEX_DEBUG and now just controls extra mutex debugging code. - Abolish the ugly mtx_f hack. Instead, we dynamically allocate seperate mtx_debug structures on the fly in mtx_init, except for mutexes that are initiated very early in the boot process. These mutexes are declared using a special MUTEX_DECLARE() macro, and use a new flag MTX_COLD when calling mtx_init. This is still somewhat hackish, but it is less evil than the mtx_f filler struct, and the mtx struct is now the same size with and without mutex debugging code. - Add some micro-micro-operation macros for doing the actual atomic operations on the mutex mtx_lock field to make it easier for other archs to override/optimize mutex ops if needed. These new tiny ops also clean up the code in some places by replacing long atomic operation function calls that spanned 2-3 lines with a short 1-line macro call. - Don't call mi_switch() from mtx_enter_hard() when we block while trying to obtain a sleep mutex. Calling mi_switch() would bogusly release Giant before switching to the next process. Instead, inline most of the code from mi_switch() in the mtx_enter_hard() function. Note that when we finally kill Giant we can back this out and go back to calling mi_switch().
2000-10-20 07:26:37 +00:00
static void
propagate_priority(struct proc *p)
{
Implement a unified run queue and adjust priority levels accordingly. - All processes go into the same array of queues, with different scheduling classes using different portions of the array. This allows user processes to have their priorities propogated up into interrupt thread range if need be. - I chose 64 run queues as an arbitrary number that is greater than 32. We used to have 4 separate arrays of 32 queues each, so this may not be optimal. The new run queue code was written with this in mind; changing the number of run queues only requires changing constants in runq.h and adjusting the priority levels. - The new run queue code takes the run queue as a parameter. This is intended to be used to create per-cpu run queues. Implement wrappers for compatibility with the old interface which pass in the global run queue structure. - Group the priority level, user priority, native priority (before propogation) and the scheduling class into a struct priority. - Change any hard coded priority levels that I found to use symbolic constants (TTIPRI and TTOPRI). - Remove the curpriority global variable and use that of curproc. This was used to detect when a process' priority had lowered and it should yield. We now effectively yield on every interrupt. - Activate propogate_priority(). It should now have the desired effect without needing to also propogate the scheduling class. - Temporarily comment out the call to vm_page_zero_idle() in the idle loop. It interfered with propogate_priority() because the idle process needed to do a non-blocking acquire of Giant and then other processes would try to propogate their priority onto it. The idle process should not do anything except idle. vm_page_zero_idle() will return in the form of an idle priority kernel thread which is woken up at apprioriate times by the vm system. - Update struct kinfo_proc to the new priority interface. Deliberately change its size by adjusting the spare fields. It remained the same size, but the layout has changed, so userland processes that use it would parse the data incorrectly. The size constraint should really be changed to an arbitrary version number. Also add a debug.sizeof sysctl node for struct kinfo_proc.
2001-02-12 00:20:08 +00:00
int pri = p->p_pri.pri_level;
- Make the mutex code almost completely machine independent. This greatly reducues the maintenance load for the mutex code. The only MD portions of the mutex code are in machine/mutex.h now, which include the assembly macros for handling mutexes as well as optionally overriding the mutex micro-operations. For example, we use optimized micro-ops on the x86 platform #ifndef I386_CPU. - Change the behavior of the SMP_DEBUG kernel option. In the new code, mtx_assert() only depends on INVARIANTS, allowing other kernel developers to have working mutex assertiions without having to include all of the mutex debugging code. The SMP_DEBUG kernel option has been renamed to MUTEX_DEBUG and now just controls extra mutex debugging code. - Abolish the ugly mtx_f hack. Instead, we dynamically allocate seperate mtx_debug structures on the fly in mtx_init, except for mutexes that are initiated very early in the boot process. These mutexes are declared using a special MUTEX_DECLARE() macro, and use a new flag MTX_COLD when calling mtx_init. This is still somewhat hackish, but it is less evil than the mtx_f filler struct, and the mtx struct is now the same size with and without mutex debugging code. - Add some micro-micro-operation macros for doing the actual atomic operations on the mutex mtx_lock field to make it easier for other archs to override/optimize mutex ops if needed. These new tiny ops also clean up the code in some places by replacing long atomic operation function calls that spanned 2-3 lines with a short 1-line macro call. - Don't call mi_switch() from mtx_enter_hard() when we block while trying to obtain a sleep mutex. Calling mi_switch() would bogusly release Giant before switching to the next process. Instead, inline most of the code from mi_switch() in the mtx_enter_hard() function. Note that when we finally kill Giant we can back this out and go back to calling mi_switch().
2000-10-20 07:26:37 +00:00
struct mtx *m = p->p_blocked;
Fix up priority propagation: - Use a better test for determining when a process is running. - Convert some checks to assertions. - Remove unnecessary tests. - Save the priority before acquiring a mutex rather than in msleep(9).
2000-11-30 00:51:16 +00:00
mtx_assert(&sched_lock, MA_OWNED);
- Make the mutex code almost completely machine independent. This greatly reducues the maintenance load for the mutex code. The only MD portions of the mutex code are in machine/mutex.h now, which include the assembly macros for handling mutexes as well as optionally overriding the mutex micro-operations. For example, we use optimized micro-ops on the x86 platform #ifndef I386_CPU. - Change the behavior of the SMP_DEBUG kernel option. In the new code, mtx_assert() only depends on INVARIANTS, allowing other kernel developers to have working mutex assertiions without having to include all of the mutex debugging code. The SMP_DEBUG kernel option has been renamed to MUTEX_DEBUG and now just controls extra mutex debugging code. - Abolish the ugly mtx_f hack. Instead, we dynamically allocate seperate mtx_debug structures on the fly in mtx_init, except for mutexes that are initiated very early in the boot process. These mutexes are declared using a special MUTEX_DECLARE() macro, and use a new flag MTX_COLD when calling mtx_init. This is still somewhat hackish, but it is less evil than the mtx_f filler struct, and the mtx struct is now the same size with and without mutex debugging code. - Add some micro-micro-operation macros for doing the actual atomic operations on the mutex mtx_lock field to make it easier for other archs to override/optimize mutex ops if needed. These new tiny ops also clean up the code in some places by replacing long atomic operation function calls that spanned 2-3 lines with a short 1-line macro call. - Don't call mi_switch() from mtx_enter_hard() when we block while trying to obtain a sleep mutex. Calling mi_switch() would bogusly release Giant before switching to the next process. Instead, inline most of the code from mi_switch() in the mtx_enter_hard() function. Note that when we finally kill Giant we can back this out and go back to calling mi_switch().
2000-10-20 07:26:37 +00:00
for (;;) {
struct proc *p1;
p = mtx_owner(m);
if (p == NULL) {
/*
* This really isn't quite right. Really
* ought to bump priority of process that
* next acquires the mutex.
*/
MPASS(m->mtx_lock == MTX_CONTESTED);
return;
}
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
- Make the mutex code almost completely machine independent. This greatly reducues the maintenance load for the mutex code. The only MD portions of the mutex code are in machine/mutex.h now, which include the assembly macros for handling mutexes as well as optionally overriding the mutex micro-operations. For example, we use optimized micro-ops on the x86 platform #ifndef I386_CPU. - Change the behavior of the SMP_DEBUG kernel option. In the new code, mtx_assert() only depends on INVARIANTS, allowing other kernel developers to have working mutex assertiions without having to include all of the mutex debugging code. The SMP_DEBUG kernel option has been renamed to MUTEX_DEBUG and now just controls extra mutex debugging code. - Abolish the ugly mtx_f hack. Instead, we dynamically allocate seperate mtx_debug structures on the fly in mtx_init, except for mutexes that are initiated very early in the boot process. These mutexes are declared using a special MUTEX_DECLARE() macro, and use a new flag MTX_COLD when calling mtx_init. This is still somewhat hackish, but it is less evil than the mtx_f filler struct, and the mtx struct is now the same size with and without mutex debugging code. - Add some micro-micro-operation macros for doing the actual atomic operations on the mutex mtx_lock field to make it easier for other archs to override/optimize mutex ops if needed. These new tiny ops also clean up the code in some places by replacing long atomic operation function calls that spanned 2-3 lines with a short 1-line macro call. - Don't call mi_switch() from mtx_enter_hard() when we block while trying to obtain a sleep mutex. Calling mi_switch() would bogusly release Giant before switching to the next process. Instead, inline most of the code from mi_switch() in the mtx_enter_hard() function. Note that when we finally kill Giant we can back this out and go back to calling mi_switch().
2000-10-20 07:26:37 +00:00
MPASS(p->p_magic == P_MAGIC);
Fix up priority propagation: - Use a better test for determining when a process is running. - Convert some checks to assertions. - Remove unnecessary tests. - Save the priority before acquiring a mutex rather than in msleep(9).
2000-11-30 00:51:16 +00:00
KASSERT(p->p_stat != SSLEEP, ("sleeping process owns a mutex"));
Implement a unified run queue and adjust priority levels accordingly. - All processes go into the same array of queues, with different scheduling classes using different portions of the array. This allows user processes to have their priorities propogated up into interrupt thread range if need be. - I chose 64 run queues as an arbitrary number that is greater than 32. We used to have 4 separate arrays of 32 queues each, so this may not be optimal. The new run queue code was written with this in mind; changing the number of run queues only requires changing constants in runq.h and adjusting the priority levels. - The new run queue code takes the run queue as a parameter. This is intended to be used to create per-cpu run queues. Implement wrappers for compatibility with the old interface which pass in the global run queue structure. - Group the priority level, user priority, native priority (before propogation) and the scheduling class into a struct priority. - Change any hard coded priority levels that I found to use symbolic constants (TTIPRI and TTOPRI). - Remove the curpriority global variable and use that of curproc. This was used to detect when a process' priority had lowered and it should yield. We now effectively yield on every interrupt. - Activate propogate_priority(). It should now have the desired effect without needing to also propogate the scheduling class. - Temporarily comment out the call to vm_page_zero_idle() in the idle loop. It interfered with propogate_priority() because the idle process needed to do a non-blocking acquire of Giant and then other processes would try to propogate their priority onto it. The idle process should not do anything except idle. vm_page_zero_idle() will return in the form of an idle priority kernel thread which is woken up at apprioriate times by the vm system. - Update struct kinfo_proc to the new priority interface. Deliberately change its size by adjusting the spare fields. It remained the same size, but the layout has changed, so userland processes that use it would parse the data incorrectly. The size constraint should really be changed to an arbitrary version number. Also add a debug.sizeof sysctl node for struct kinfo_proc.
2001-02-12 00:20:08 +00:00
if (p->p_pri.pri_level <= pri)
- Make the mutex code almost completely machine independent. This greatly reducues the maintenance load for the mutex code. The only MD portions of the mutex code are in machine/mutex.h now, which include the assembly macros for handling mutexes as well as optionally overriding the mutex micro-operations. For example, we use optimized micro-ops on the x86 platform #ifndef I386_CPU. - Change the behavior of the SMP_DEBUG kernel option. In the new code, mtx_assert() only depends on INVARIANTS, allowing other kernel developers to have working mutex assertiions without having to include all of the mutex debugging code. The SMP_DEBUG kernel option has been renamed to MUTEX_DEBUG and now just controls extra mutex debugging code. - Abolish the ugly mtx_f hack. Instead, we dynamically allocate seperate mtx_debug structures on the fly in mtx_init, except for mutexes that are initiated very early in the boot process. These mutexes are declared using a special MUTEX_DECLARE() macro, and use a new flag MTX_COLD when calling mtx_init. This is still somewhat hackish, but it is less evil than the mtx_f filler struct, and the mtx struct is now the same size with and without mutex debugging code. - Add some micro-micro-operation macros for doing the actual atomic operations on the mutex mtx_lock field to make it easier for other archs to override/optimize mutex ops if needed. These new tiny ops also clean up the code in some places by replacing long atomic operation function calls that spanned 2-3 lines with a short 1-line macro call. - Don't call mi_switch() from mtx_enter_hard() when we block while trying to obtain a sleep mutex. Calling mi_switch() would bogusly release Giant before switching to the next process. Instead, inline most of the code from mi_switch() in the mtx_enter_hard() function. Note that when we finally kill Giant we can back this out and go back to calling mi_switch().
2000-10-20 07:26:37 +00:00
return;
Fix up priority propagation: - Use a better test for determining when a process is running. - Convert some checks to assertions. - Remove unnecessary tests. - Save the priority before acquiring a mutex rather than in msleep(9).
2000-11-30 00:51:16 +00:00
/*
* Bump this process' priority.
*/
SET_PRIO(p, pri);
- Make the mutex code almost completely machine independent. This greatly reducues the maintenance load for the mutex code. The only MD portions of the mutex code are in machine/mutex.h now, which include the assembly macros for handling mutexes as well as optionally overriding the mutex micro-operations. For example, we use optimized micro-ops on the x86 platform #ifndef I386_CPU. - Change the behavior of the SMP_DEBUG kernel option. In the new code, mtx_assert() only depends on INVARIANTS, allowing other kernel developers to have working mutex assertiions without having to include all of the mutex debugging code. The SMP_DEBUG kernel option has been renamed to MUTEX_DEBUG and now just controls extra mutex debugging code. - Abolish the ugly mtx_f hack. Instead, we dynamically allocate seperate mtx_debug structures on the fly in mtx_init, except for mutexes that are initiated very early in the boot process. These mutexes are declared using a special MUTEX_DECLARE() macro, and use a new flag MTX_COLD when calling mtx_init. This is still somewhat hackish, but it is less evil than the mtx_f filler struct, and the mtx struct is now the same size with and without mutex debugging code. - Add some micro-micro-operation macros for doing the actual atomic operations on the mutex mtx_lock field to make it easier for other archs to override/optimize mutex ops if needed. These new tiny ops also clean up the code in some places by replacing long atomic operation function calls that spanned 2-3 lines with a short 1-line macro call. - Don't call mi_switch() from mtx_enter_hard() when we block while trying to obtain a sleep mutex. Calling mi_switch() would bogusly release Giant before switching to the next process. Instead, inline most of the code from mi_switch() in the mtx_enter_hard() function. Note that when we finally kill Giant we can back this out and go back to calling mi_switch().
2000-10-20 07:26:37 +00:00
/*
* If lock holder is actually running, just bump priority.
*/
- Use the NOCPU constant. - Move the ithread spin locks before sched lock and clk in preparation for future commits to the ithread code.
2001-02-22 02:12:54 +00:00
if (p->p_oncpu != NOCPU) {
- Make the mutex code almost completely machine independent. This greatly reducues the maintenance load for the mutex code. The only MD portions of the mutex code are in machine/mutex.h now, which include the assembly macros for handling mutexes as well as optionally overriding the mutex micro-operations. For example, we use optimized micro-ops on the x86 platform #ifndef I386_CPU. - Change the behavior of the SMP_DEBUG kernel option. In the new code, mtx_assert() only depends on INVARIANTS, allowing other kernel developers to have working mutex assertiions without having to include all of the mutex debugging code. The SMP_DEBUG kernel option has been renamed to MUTEX_DEBUG and now just controls extra mutex debugging code. - Abolish the ugly mtx_f hack. Instead, we dynamically allocate seperate mtx_debug structures on the fly in mtx_init, except for mutexes that are initiated very early in the boot process. These mutexes are declared using a special MUTEX_DECLARE() macro, and use a new flag MTX_COLD when calling mtx_init. This is still somewhat hackish, but it is less evil than the mtx_f filler struct, and the mtx struct is now the same size with and without mutex debugging code. - Add some micro-micro-operation macros for doing the actual atomic operations on the mutex mtx_lock field to make it easier for other archs to override/optimize mutex ops if needed. These new tiny ops also clean up the code in some places by replacing long atomic operation function calls that spanned 2-3 lines with a short 1-line macro call. - Don't call mi_switch() from mtx_enter_hard() when we block while trying to obtain a sleep mutex. Calling mi_switch() would bogusly release Giant before switching to the next process. Instead, inline most of the code from mi_switch() in the mtx_enter_hard() function. Note that when we finally kill Giant we can back this out and go back to calling mi_switch().
2000-10-20 07:26:37 +00:00
MPASS(p->p_stat == SRUN || p->p_stat == SZOMB);
return;
}
Implement a unified run queue and adjust priority levels accordingly. - All processes go into the same array of queues, with different scheduling classes using different portions of the array. This allows user processes to have their priorities propogated up into interrupt thread range if need be. - I chose 64 run queues as an arbitrary number that is greater than 32. We used to have 4 separate arrays of 32 queues each, so this may not be optimal. The new run queue code was written with this in mind; changing the number of run queues only requires changing constants in runq.h and adjusting the priority levels. - The new run queue code takes the run queue as a parameter. This is intended to be used to create per-cpu run queues. Implement wrappers for compatibility with the old interface which pass in the global run queue structure. - Group the priority level, user priority, native priority (before propogation) and the scheduling class into a struct priority. - Change any hard coded priority levels that I found to use symbolic constants (TTIPRI and TTOPRI). - Remove the curpriority global variable and use that of curproc. This was used to detect when a process' priority had lowered and it should yield. We now effectively yield on every interrupt. - Activate propogate_priority(). It should now have the desired effect without needing to also propogate the scheduling class. - Temporarily comment out the call to vm_page_zero_idle() in the idle loop. It interfered with propogate_priority() because the idle process needed to do a non-blocking acquire of Giant and then other processes would try to propogate their priority onto it. The idle process should not do anything except idle. vm_page_zero_idle() will return in the form of an idle priority kernel thread which is woken up at apprioriate times by the vm system. - Update struct kinfo_proc to the new priority interface. Deliberately change its size by adjusting the spare fields. It remained the same size, but the layout has changed, so userland processes that use it would parse the data incorrectly. The size constraint should really be changed to an arbitrary version number. Also add a debug.sizeof sysctl node for struct kinfo_proc.
2001-02-12 00:20:08 +00:00
- Add an extra check in priority_propagation() for UP systems to ensure we don't end up back at ourselves which would indicate deadlock. - Add the proc lock to the witness dup_list as we may hold more than one process lock at a time. - Don't assert a mutex is owned in _mtx_unlock_sleep() as that is too late. We do the checks in the macros instead.
2001-03-07 02:45:15 +00:00
#ifndef SMP
/*
* For UP, we check to see if p is curproc (this shouldn't
* ever happen however as it would mean we are in a deadlock.)
*/
KASSERT(p != curproc, ("Deadlock detected"));
#endif
- Make the mutex code almost completely machine independent. This greatly reducues the maintenance load for the mutex code. The only MD portions of the mutex code are in machine/mutex.h now, which include the assembly macros for handling mutexes as well as optionally overriding the mutex micro-operations. For example, we use optimized micro-ops on the x86 platform #ifndef I386_CPU. - Change the behavior of the SMP_DEBUG kernel option. In the new code, mtx_assert() only depends on INVARIANTS, allowing other kernel developers to have working mutex assertiions without having to include all of the mutex debugging code. The SMP_DEBUG kernel option has been renamed to MUTEX_DEBUG and now just controls extra mutex debugging code. - Abolish the ugly mtx_f hack. Instead, we dynamically allocate seperate mtx_debug structures on the fly in mtx_init, except for mutexes that are initiated very early in the boot process. These mutexes are declared using a special MUTEX_DECLARE() macro, and use a new flag MTX_COLD when calling mtx_init. This is still somewhat hackish, but it is less evil than the mtx_f filler struct, and the mtx struct is now the same size with and without mutex debugging code. - Add some micro-micro-operation macros for doing the actual atomic operations on the mutex mtx_lock field to make it easier for other archs to override/optimize mutex ops if needed. These new tiny ops also clean up the code in some places by replacing long atomic operation function calls that spanned 2-3 lines with a short 1-line macro call. - Don't call mi_switch() from mtx_enter_hard() when we block while trying to obtain a sleep mutex. Calling mi_switch() would bogusly release Giant before switching to the next process. Instead, inline most of the code from mi_switch() in the mtx_enter_hard() function. Note that when we finally kill Giant we can back this out and go back to calling mi_switch().
2000-10-20 07:26:37 +00:00
/*
* If on run queue move to new run queue, and
* quit.
*/
if (p->p_stat == SRUN) {
MPASS(p->p_blocked == NULL);
remrunqueue(p);
setrunqueue(p);
return;
}
/*
Fix up priority propagation: - Use a better test for determining when a process is running. - Convert some checks to assertions. - Remove unnecessary tests. - Save the priority before acquiring a mutex rather than in msleep(9).
2000-11-30 00:51:16 +00:00
* If we aren't blocked on a mutex, we should be.
- Make the mutex code almost completely machine independent. This greatly reducues the maintenance load for the mutex code. The only MD portions of the mutex code are in machine/mutex.h now, which include the assembly macros for handling mutexes as well as optionally overriding the mutex micro-operations. For example, we use optimized micro-ops on the x86 platform #ifndef I386_CPU. - Change the behavior of the SMP_DEBUG kernel option. In the new code, mtx_assert() only depends on INVARIANTS, allowing other kernel developers to have working mutex assertiions without having to include all of the mutex debugging code. The SMP_DEBUG kernel option has been renamed to MUTEX_DEBUG and now just controls extra mutex debugging code. - Abolish the ugly mtx_f hack. Instead, we dynamically allocate seperate mtx_debug structures on the fly in mtx_init, except for mutexes that are initiated very early in the boot process. These mutexes are declared using a special MUTEX_DECLARE() macro, and use a new flag MTX_COLD when calling mtx_init. This is still somewhat hackish, but it is less evil than the mtx_f filler struct, and the mtx struct is now the same size with and without mutex debugging code. - Add some micro-micro-operation macros for doing the actual atomic operations on the mutex mtx_lock field to make it easier for other archs to override/optimize mutex ops if needed. These new tiny ops also clean up the code in some places by replacing long atomic operation function calls that spanned 2-3 lines with a short 1-line macro call. - Don't call mi_switch() from mtx_enter_hard() when we block while trying to obtain a sleep mutex. Calling mi_switch() would bogusly release Giant before switching to the next process. Instead, inline most of the code from mi_switch() in the mtx_enter_hard() function. Note that when we finally kill Giant we can back this out and go back to calling mi_switch().
2000-10-20 07:26:37 +00:00
*/
Fix up priority propagation: - Use a better test for determining when a process is running. - Convert some checks to assertions. - Remove unnecessary tests. - Save the priority before acquiring a mutex rather than in msleep(9).
2000-11-30 00:51:16 +00:00
KASSERT(p->p_stat == SMTX, (
"process %d(%s):%d holds %s but isn't blocked on a mutex\n",
p->p_pid, p->p_comm, p->p_stat,
m->mtx_description));
- Make the mutex code almost completely machine independent. This greatly reducues the maintenance load for the mutex code. The only MD portions of the mutex code are in machine/mutex.h now, which include the assembly macros for handling mutexes as well as optionally overriding the mutex micro-operations. For example, we use optimized micro-ops on the x86 platform #ifndef I386_CPU. - Change the behavior of the SMP_DEBUG kernel option. In the new code, mtx_assert() only depends on INVARIANTS, allowing other kernel developers to have working mutex assertiions without having to include all of the mutex debugging code. The SMP_DEBUG kernel option has been renamed to MUTEX_DEBUG and now just controls extra mutex debugging code. - Abolish the ugly mtx_f hack. Instead, we dynamically allocate seperate mtx_debug structures on the fly in mtx_init, except for mutexes that are initiated very early in the boot process. These mutexes are declared using a special MUTEX_DECLARE() macro, and use a new flag MTX_COLD when calling mtx_init. This is still somewhat hackish, but it is less evil than the mtx_f filler struct, and the mtx struct is now the same size with and without mutex debugging code. - Add some micro-micro-operation macros for doing the actual atomic operations on the mutex mtx_lock field to make it easier for other archs to override/optimize mutex ops if needed. These new tiny ops also clean up the code in some places by replacing long atomic operation function calls that spanned 2-3 lines with a short 1-line macro call. - Don't call mi_switch() from mtx_enter_hard() when we block while trying to obtain a sleep mutex. Calling mi_switch() would bogusly release Giant before switching to the next process. Instead, inline most of the code from mi_switch() in the mtx_enter_hard() function. Note that when we finally kill Giant we can back this out and go back to calling mi_switch().
2000-10-20 07:26:37 +00:00
/*
* Pick up the mutex that p is blocked on.
*/
m = p->p_blocked;
MPASS(m != NULL);
/*
* Check if the proc needs to be moved up on
* the blocked chain
*/
Fix up priority propagation: - Use a better test for determining when a process is running. - Convert some checks to assertions. - Remove unnecessary tests. - Save the priority before acquiring a mutex rather than in msleep(9).
2000-11-30 00:51:16 +00:00
if (p == TAILQ_FIRST(&m->mtx_blocked)) {
continue;
}
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
Implement a unified run queue and adjust priority levels accordingly. - All processes go into the same array of queues, with different scheduling classes using different portions of the array. This allows user processes to have their priorities propogated up into interrupt thread range if need be. - I chose 64 run queues as an arbitrary number that is greater than 32. We used to have 4 separate arrays of 32 queues each, so this may not be optimal. The new run queue code was written with this in mind; changing the number of run queues only requires changing constants in runq.h and adjusting the priority levels. - The new run queue code takes the run queue as a parameter. This is intended to be used to create per-cpu run queues. Implement wrappers for compatibility with the old interface which pass in the global run queue structure. - Group the priority level, user priority, native priority (before propogation) and the scheduling class into a struct priority. - Change any hard coded priority levels that I found to use symbolic constants (TTIPRI and TTOPRI). - Remove the curpriority global variable and use that of curproc. This was used to detect when a process' priority had lowered and it should yield. We now effectively yield on every interrupt. - Activate propogate_priority(). It should now have the desired effect without needing to also propogate the scheduling class. - Temporarily comment out the call to vm_page_zero_idle() in the idle loop. It interfered with propogate_priority() because the idle process needed to do a non-blocking acquire of Giant and then other processes would try to propogate their priority onto it. The idle process should not do anything except idle. vm_page_zero_idle() will return in the form of an idle priority kernel thread which is woken up at apprioriate times by the vm system. - Update struct kinfo_proc to the new priority interface. Deliberately change its size by adjusting the spare fields. It remained the same size, but the layout has changed, so userland processes that use it would parse the data incorrectly. The size constraint should really be changed to an arbitrary version number. Also add a debug.sizeof sysctl node for struct kinfo_proc.
2001-02-12 00:20:08 +00:00
p1 = TAILQ_PREV(p, procqueue, p_procq);
if (p1->p_pri.pri_level <= pri) {
- Make the mutex code almost completely machine independent. This greatly reducues the maintenance load for the mutex code. The only MD portions of the mutex code are in machine/mutex.h now, which include the assembly macros for handling mutexes as well as optionally overriding the mutex micro-operations. For example, we use optimized micro-ops on the x86 platform #ifndef I386_CPU. - Change the behavior of the SMP_DEBUG kernel option. In the new code, mtx_assert() only depends on INVARIANTS, allowing other kernel developers to have working mutex assertiions without having to include all of the mutex debugging code. The SMP_DEBUG kernel option has been renamed to MUTEX_DEBUG and now just controls extra mutex debugging code. - Abolish the ugly mtx_f hack. Instead, we dynamically allocate seperate mtx_debug structures on the fly in mtx_init, except for mutexes that are initiated very early in the boot process. These mutexes are declared using a special MUTEX_DECLARE() macro, and use a new flag MTX_COLD when calling mtx_init. This is still somewhat hackish, but it is less evil than the mtx_f filler struct, and the mtx struct is now the same size with and without mutex debugging code. - Add some micro-micro-operation macros for doing the actual atomic operations on the mutex mtx_lock field to make it easier for other archs to override/optimize mutex ops if needed. These new tiny ops also clean up the code in some places by replacing long atomic operation function calls that spanned 2-3 lines with a short 1-line macro call. - Don't call mi_switch() from mtx_enter_hard() when we block while trying to obtain a sleep mutex. Calling mi_switch() would bogusly release Giant before switching to the next process. Instead, inline most of the code from mi_switch() in the mtx_enter_hard() function. Note that when we finally kill Giant we can back this out and go back to calling mi_switch().
2000-10-20 07:26:37 +00:00
continue;
}
/*
Fix up priority propagation: - Use a better test for determining when a process is running. - Convert some checks to assertions. - Remove unnecessary tests. - Save the priority before acquiring a mutex rather than in msleep(9).
2000-11-30 00:51:16 +00:00
* Remove proc from blocked chain and determine where
* it should be moved up to. Since we know that p1 has
* a lower priority than p, we know that at least one
* process in the chain has a lower priority and that
* p1 will thus not be NULL after the loop.
- Make the mutex code almost completely machine independent. This greatly reducues the maintenance load for the mutex code. The only MD portions of the mutex code are in machine/mutex.h now, which include the assembly macros for handling mutexes as well as optionally overriding the mutex micro-operations. For example, we use optimized micro-ops on the x86 platform #ifndef I386_CPU. - Change the behavior of the SMP_DEBUG kernel option. In the new code, mtx_assert() only depends on INVARIANTS, allowing other kernel developers to have working mutex assertiions without having to include all of the mutex debugging code. The SMP_DEBUG kernel option has been renamed to MUTEX_DEBUG and now just controls extra mutex debugging code. - Abolish the ugly mtx_f hack. Instead, we dynamically allocate seperate mtx_debug structures on the fly in mtx_init, except for mutexes that are initiated very early in the boot process. These mutexes are declared using a special MUTEX_DECLARE() macro, and use a new flag MTX_COLD when calling mtx_init. This is still somewhat hackish, but it is less evil than the mtx_f filler struct, and the mtx struct is now the same size with and without mutex debugging code. - Add some micro-micro-operation macros for doing the actual atomic operations on the mutex mtx_lock field to make it easier for other archs to override/optimize mutex ops if needed. These new tiny ops also clean up the code in some places by replacing long atomic operation function calls that spanned 2-3 lines with a short 1-line macro call. - Don't call mi_switch() from mtx_enter_hard() when we block while trying to obtain a sleep mutex. Calling mi_switch() would bogusly release Giant before switching to the next process. Instead, inline most of the code from mi_switch() in the mtx_enter_hard() function. Note that when we finally kill Giant we can back this out and go back to calling mi_switch().
2000-10-20 07:26:37 +00:00
*/
TAILQ_REMOVE(&m->mtx_blocked, p, p_procq);
TAILQ_FOREACH(p1, &m->mtx_blocked, p_procq) {
MPASS(p1->p_magic == P_MAGIC);
Implement a unified run queue and adjust priority levels accordingly. - All processes go into the same array of queues, with different scheduling classes using different portions of the array. This allows user processes to have their priorities propogated up into interrupt thread range if need be. - I chose 64 run queues as an arbitrary number that is greater than 32. We used to have 4 separate arrays of 32 queues each, so this may not be optimal. The new run queue code was written with this in mind; changing the number of run queues only requires changing constants in runq.h and adjusting the priority levels. - The new run queue code takes the run queue as a parameter. This is intended to be used to create per-cpu run queues. Implement wrappers for compatibility with the old interface which pass in the global run queue structure. - Group the priority level, user priority, native priority (before propogation) and the scheduling class into a struct priority. - Change any hard coded priority levels that I found to use symbolic constants (TTIPRI and TTOPRI). - Remove the curpriority global variable and use that of curproc. This was used to detect when a process' priority had lowered and it should yield. We now effectively yield on every interrupt. - Activate propogate_priority(). It should now have the desired effect without needing to also propogate the scheduling class. - Temporarily comment out the call to vm_page_zero_idle() in the idle loop. It interfered with propogate_priority() because the idle process needed to do a non-blocking acquire of Giant and then other processes would try to propogate their priority onto it. The idle process should not do anything except idle. vm_page_zero_idle() will return in the form of an idle priority kernel thread which is woken up at apprioriate times by the vm system. - Update struct kinfo_proc to the new priority interface. Deliberately change its size by adjusting the spare fields. It remained the same size, but the layout has changed, so userland processes that use it would parse the data incorrectly. The size constraint should really be changed to an arbitrary version number. Also add a debug.sizeof sysctl node for struct kinfo_proc.
2001-02-12 00:20:08 +00:00
if (p1->p_pri.pri_level > pri)
- Make the mutex code almost completely machine independent. This greatly reducues the maintenance load for the mutex code. The only MD portions of the mutex code are in machine/mutex.h now, which include the assembly macros for handling mutexes as well as optionally overriding the mutex micro-operations. For example, we use optimized micro-ops on the x86 platform #ifndef I386_CPU. - Change the behavior of the SMP_DEBUG kernel option. In the new code, mtx_assert() only depends on INVARIANTS, allowing other kernel developers to have working mutex assertiions without having to include all of the mutex debugging code. The SMP_DEBUG kernel option has been renamed to MUTEX_DEBUG and now just controls extra mutex debugging code. - Abolish the ugly mtx_f hack. Instead, we dynamically allocate seperate mtx_debug structures on the fly in mtx_init, except for mutexes that are initiated very early in the boot process. These mutexes are declared using a special MUTEX_DECLARE() macro, and use a new flag MTX_COLD when calling mtx_init. This is still somewhat hackish, but it is less evil than the mtx_f filler struct, and the mtx struct is now the same size with and without mutex debugging code. - Add some micro-micro-operation macros for doing the actual atomic operations on the mutex mtx_lock field to make it easier for other archs to override/optimize mutex ops if needed. These new tiny ops also clean up the code in some places by replacing long atomic operation function calls that spanned 2-3 lines with a short 1-line macro call. - Don't call mi_switch() from mtx_enter_hard() when we block while trying to obtain a sleep mutex. Calling mi_switch() would bogusly release Giant before switching to the next process. Instead, inline most of the code from mi_switch() in the mtx_enter_hard() function. Note that when we finally kill Giant we can back this out and go back to calling mi_switch().
2000-10-20 07:26:37 +00:00
break;
}
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
Fix up priority propagation: - Use a better test for determining when a process is running. - Convert some checks to assertions. - Remove unnecessary tests. - Save the priority before acquiring a mutex rather than in msleep(9).
2000-11-30 00:51:16 +00:00
MPASS(p1 != NULL);
TAILQ_INSERT_BEFORE(p1, p, p_procq);
- Make the mutex code almost completely machine independent. This greatly reducues the maintenance load for the mutex code. The only MD portions of the mutex code are in machine/mutex.h now, which include the assembly macros for handling mutexes as well as optionally overriding the mutex micro-operations. For example, we use optimized micro-ops on the x86 platform #ifndef I386_CPU. - Change the behavior of the SMP_DEBUG kernel option. In the new code, mtx_assert() only depends on INVARIANTS, allowing other kernel developers to have working mutex assertiions without having to include all of the mutex debugging code. The SMP_DEBUG kernel option has been renamed to MUTEX_DEBUG and now just controls extra mutex debugging code. - Abolish the ugly mtx_f hack. Instead, we dynamically allocate seperate mtx_debug structures on the fly in mtx_init, except for mutexes that are initiated very early in the boot process. These mutexes are declared using a special MUTEX_DECLARE() macro, and use a new flag MTX_COLD when calling mtx_init. This is still somewhat hackish, but it is less evil than the mtx_f filler struct, and the mtx struct is now the same size with and without mutex debugging code. - Add some micro-micro-operation macros for doing the actual atomic operations on the mutex mtx_lock field to make it easier for other archs to override/optimize mutex ops if needed. These new tiny ops also clean up the code in some places by replacing long atomic operation function calls that spanned 2-3 lines with a short 1-line macro call. - Don't call mi_switch() from mtx_enter_hard() when we block while trying to obtain a sleep mutex. Calling mi_switch() would bogusly release Giant before switching to the next process. Instead, inline most of the code from mi_switch() in the mtx_enter_hard() function. Note that when we finally kill Giant we can back this out and go back to calling mi_switch().
2000-10-20 07:26:37 +00:00
CTR4(KTR_LOCK,
- Don't use a union and fun tricks to shave one extra pointer off of struct mtx right now as it makes debugging harder. When we are in optimizing mode, we can revisit this. - Fix the KTR trace messages to use %p rather than 0x%p to avoid duplicate 0x's in KTR output. - During witness_fixup, release Giant so that witness doesn't get confused. Also, grab all_mtx while walking the list of mutexes. - Remove w_sleep and w_recurse. Instead, perform checks on mutexes using the mutex's mtx_flags field. - Allow debug.witness_ddb and debug.witness_skipspin to be set from the loader. - Add Giant to the front of existing order_list entries to help ensure Giant is always first. - Add an order entry for the various proc locks. Note that this only helps keep proc in order mostly as the allproc and proctree mutexes are only obtained during a lockmgr operation on the specified mutex.
2001-01-24 10:57:01 +00:00
"propagate_priority: p %p moved before %p on [%p] %s",
- Make the mutex code almost completely machine independent. This greatly reducues the maintenance load for the mutex code. The only MD portions of the mutex code are in machine/mutex.h now, which include the assembly macros for handling mutexes as well as optionally overriding the mutex micro-operations. For example, we use optimized micro-ops on the x86 platform #ifndef I386_CPU. - Change the behavior of the SMP_DEBUG kernel option. In the new code, mtx_assert() only depends on INVARIANTS, allowing other kernel developers to have working mutex assertiions without having to include all of the mutex debugging code. The SMP_DEBUG kernel option has been renamed to MUTEX_DEBUG and now just controls extra mutex debugging code. - Abolish the ugly mtx_f hack. Instead, we dynamically allocate seperate mtx_debug structures on the fly in mtx_init, except for mutexes that are initiated very early in the boot process. These mutexes are declared using a special MUTEX_DECLARE() macro, and use a new flag MTX_COLD when calling mtx_init. This is still somewhat hackish, but it is less evil than the mtx_f filler struct, and the mtx struct is now the same size with and without mutex debugging code. - Add some micro-micro-operation macros for doing the actual atomic operations on the mutex mtx_lock field to make it easier for other archs to override/optimize mutex ops if needed. These new tiny ops also clean up the code in some places by replacing long atomic operation function calls that spanned 2-3 lines with a short 1-line macro call. - Don't call mi_switch() from mtx_enter_hard() when we block while trying to obtain a sleep mutex. Calling mi_switch() would bogusly release Giant before switching to the next process. Instead, inline most of the code from mi_switch() in the mtx_enter_hard() function. Note that when we finally kill Giant we can back this out and go back to calling mi_switch().
2000-10-20 07:26:37 +00:00
p, p1, m, m->mtx_description);
}
}
Move most of sys/mutex.h into kern/kern_mutex.c, thereby making the mutex inline functions non-inlined. Hide parts of the mutex implementation that should not be exposed. Make sure that WITNESS code is not executed during boot until the mutexes are fully initialized by SI_SUB_MUTEX (the original motivation for this commit). Submitted by: peter
2001-01-21 22:34:43 +00:00
/*
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
* The important part of mtx_trylock{,_flags}()
* Tries to acquire lock `m.' We do NOT handle recursion here; we assume that
* if we're called, it's because we know we don't already own this lock.
Move most of sys/mutex.h into kern/kern_mutex.c, thereby making the mutex inline functions non-inlined. Hide parts of the mutex implementation that should not be exposed. Make sure that WITNESS code is not executed during boot until the mutexes are fully initialized by SI_SUB_MUTEX (the original motivation for this commit). Submitted by: peter
2001-01-21 22:34:43 +00:00
*/
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
int
_mtx_trylock(struct mtx *m, int opts, const char *file, int line)
Move most of sys/mutex.h into kern/kern_mutex.c, thereby making the mutex inline functions non-inlined. Hide parts of the mutex implementation that should not be exposed. Make sure that WITNESS code is not executed during boot until the mutexes are fully initialized by SI_SUB_MUTEX (the original motivation for this commit). Submitted by: peter
2001-01-21 22:34:43 +00:00
{
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
int rval;
Move most of sys/mutex.h into kern/kern_mutex.c, thereby making the mutex inline functions non-inlined. Hide parts of the mutex implementation that should not be exposed. Make sure that WITNESS code is not executed during boot until the mutexes are fully initialized by SI_SUB_MUTEX (the original motivation for this commit). Submitted by: peter
2001-01-21 22:34:43 +00:00
Change all instances of `CURPROC' and `CURTHD' to `curproc,' in order to stay consistent. Requested by: bde
2001-02-12 03:15:43 +00:00
MPASS(curproc != NULL);
Move most of sys/mutex.h into kern/kern_mutex.c, thereby making the mutex inline functions non-inlined. Hide parts of the mutex implementation that should not be exposed. Make sure that WITNESS code is not executed during boot until the mutexes are fully initialized by SI_SUB_MUTEX (the original motivation for this commit). Submitted by: peter
2001-01-21 22:34:43 +00:00
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
/*
* _mtx_trylock does not accept MTX_NOSWITCH option.
*/
- Place back STR string declarations for lock/unlock strings used for KTR_LOCK tracing in order to avoid duplication. - Insert some tracepoints back into the mutex acq/rel code, thus ensuring that we can trace all lock acq/rel's again. - All CURPROC != NULL checks are MPASS()es (under MUTEX_DEBUG) because they signify a serious mutex corruption. - Change up some KASSERT()s to MPASS()es, and vice-versa, depending on the type of problem we're debugging (INVARIANTS is used here to check that the API is being used properly whereas MUTEX_DEBUG is used to ensure that something general isn't happening that will have bad impact on mutex locks). Reminded by: jhb, jake, asmodai
2001-02-11 02:54:16 +00:00
KASSERT((opts & MTX_NOSWITCH) == 0,
("mtx_trylock() called with invalid option flag(s) %d", opts));
Move most of sys/mutex.h into kern/kern_mutex.c, thereby making the mutex inline functions non-inlined. Hide parts of the mutex implementation that should not be exposed. Make sure that WITNESS code is not executed during boot until the mutexes are fully initialized by SI_SUB_MUTEX (the original motivation for this commit). Submitted by: peter
2001-01-21 22:34:43 +00:00
Change all instances of `CURPROC' and `CURTHD' to `curproc,' in order to stay consistent. Requested by: bde
2001-02-12 03:15:43 +00:00
rval = _obtain_lock(m, curproc);
Move most of sys/mutex.h into kern/kern_mutex.c, thereby making the mutex inline functions non-inlined. Hide parts of the mutex implementation that should not be exposed. Make sure that WITNESS code is not executed during boot until the mutexes are fully initialized by SI_SUB_MUTEX (the original motivation for this commit). Submitted by: peter
2001-01-21 22:34:43 +00:00
#ifdef WITNESS
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
if (rval && m->mtx_witness != NULL) {
/*
* We do not handle recursion in _mtx_trylock; see the
* note at the top of the routine.
*/
- Place back STR string declarations for lock/unlock strings used for KTR_LOCK tracing in order to avoid duplication. - Insert some tracepoints back into the mutex acq/rel code, thus ensuring that we can trace all lock acq/rel's again. - All CURPROC != NULL checks are MPASS()es (under MUTEX_DEBUG) because they signify a serious mutex corruption. - Change up some KASSERT()s to MPASS()es, and vice-versa, depending on the type of problem we're debugging (INVARIANTS is used here to check that the API is being used properly whereas MUTEX_DEBUG is used to ensure that something general isn't happening that will have bad impact on mutex locks). Reminded by: jhb, jake, asmodai
2001-02-11 02:54:16 +00:00
KASSERT(!mtx_recursed(m),
("mtx_trylock() called on a recursed mutex"));
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
witness_try_enter(m, (opts | m->mtx_flags), file, line);
Move most of sys/mutex.h into kern/kern_mutex.c, thereby making the mutex inline functions non-inlined. Hide parts of the mutex implementation that should not be exposed. Make sure that WITNESS code is not executed during boot until the mutexes are fully initialized by SI_SUB_MUTEX (the original motivation for this commit). Submitted by: peter
2001-01-21 22:34:43 +00:00
}
#endif /* WITNESS */
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
if ((opts & MTX_QUIET) == 0)
- Axe RETIP() as it was very i386 specific and unwieldy. Instead, use the passed in filename and line number in the KTR tracepoint message. - Even though it is #if 0'd code, change the code to detect that a process is an interrupt thread to check p->p_ithd against NULL rather than checking non-existant process flags from BSD/OS. - Use '%p' to print pointers in KTR log messages instead of assuming sizeof(int) == sizeof(void *). - Don't set p_mtxname to NULL when releasing a mutex. It doesn't hurt to leave it set (we don't clear w_mesg for example) and at least at one time in the past, there used to be race conditions in the kernel that would result in setting this to NULL causing the kernel to dereference NULL. - Make the _mtx_assert() function be compiled in if INVARIANTS_SUPPORT is defined rather than if INVARIANTS is defined so that a KLD compiled with INVARIANTS that uses mtx_assert() can be used with a kernel that just has INVARIANT_SUPPORT compiled in.
2001-02-24 19:36:13 +00:00
CTR5(KTR_LOCK, "TRY_LOCK %s [%p] result=%d at %s:%d",
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
m->mtx_description, m, rval, file, line);
Move most of sys/mutex.h into kern/kern_mutex.c, thereby making the mutex inline functions non-inlined. Hide parts of the mutex implementation that should not be exposed. Make sure that WITNESS code is not executed during boot until the mutexes are fully initialized by SI_SUB_MUTEX (the original motivation for this commit). Submitted by: peter
2001-01-21 22:34:43 +00:00
return rval;
}
/*
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
* _mtx_lock_sleep: the tougher part of acquiring an MTX_DEF lock.
*
* We call this if the lock is either contested (i.e. we need to go to
* sleep waiting for it), or if we need to recurse on it.
Move most of sys/mutex.h into kern/kern_mutex.c, thereby making the mutex inline functions non-inlined. Hide parts of the mutex implementation that should not be exposed. Make sure that WITNESS code is not executed during boot until the mutexes are fully initialized by SI_SUB_MUTEX (the original motivation for this commit). Submitted by: peter
2001-01-21 22:34:43 +00:00
*/
void
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
_mtx_lock_sleep(struct mtx *m, int opts, const char *file, int line)
Move most of sys/mutex.h into kern/kern_mutex.c, thereby making the mutex inline functions non-inlined. Hide parts of the mutex implementation that should not be exposed. Make sure that WITNESS code is not executed during boot until the mutexes are fully initialized by SI_SUB_MUTEX (the original motivation for this commit). Submitted by: peter
2001-01-21 22:34:43 +00:00
{
Change all instances of `CURPROC' and `CURTHD' to `curproc,' in order to stay consistent. Requested by: bde
2001-02-12 03:15:43 +00:00
struct proc *p = curproc;
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
if ((m->mtx_lock & MTX_FLAGMASK) == (uintptr_t)p) {
m->mtx_recurse++;
atomic_set_ptr(&m->mtx_lock, MTX_RECURSED);
if ((opts & MTX_QUIET) == 0)
- Place back STR string declarations for lock/unlock strings used for KTR_LOCK tracing in order to avoid duplication. - Insert some tracepoints back into the mutex acq/rel code, thus ensuring that we can trace all lock acq/rel's again. - All CURPROC != NULL checks are MPASS()es (under MUTEX_DEBUG) because they signify a serious mutex corruption. - Change up some KASSERT()s to MPASS()es, and vice-versa, depending on the type of problem we're debugging (INVARIANTS is used here to check that the API is being used properly whereas MUTEX_DEBUG is used to ensure that something general isn't happening that will have bad impact on mutex locks). Reminded by: jhb, jake, asmodai
2001-02-11 02:54:16 +00:00
CTR1(KTR_LOCK, "_mtx_lock_sleep: %p recursing", m);
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
return;
Move most of sys/mutex.h into kern/kern_mutex.c, thereby making the mutex inline functions non-inlined. Hide parts of the mutex implementation that should not be exposed. Make sure that WITNESS code is not executed during boot until the mutexes are fully initialized by SI_SUB_MUTEX (the original motivation for this commit). Submitted by: peter
2001-01-21 22:34:43 +00:00
}
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
if ((opts & MTX_QUIET) == 0)
- Axe RETIP() as it was very i386 specific and unwieldy. Instead, use the passed in filename and line number in the KTR tracepoint message. - Even though it is #if 0'd code, change the code to detect that a process is an interrupt thread to check p->p_ithd against NULL rather than checking non-existant process flags from BSD/OS. - Use '%p' to print pointers in KTR log messages instead of assuming sizeof(int) == sizeof(void *). - Don't set p_mtxname to NULL when releasing a mutex. It doesn't hurt to leave it set (we don't clear w_mesg for example) and at least at one time in the past, there used to be race conditions in the kernel that would result in setting this to NULL causing the kernel to dereference NULL. - Make the _mtx_assert() function be compiled in if INVARIANTS_SUPPORT is defined rather than if INVARIANTS is defined so that a KLD compiled with INVARIANTS that uses mtx_assert() can be used with a kernel that just has INVARIANT_SUPPORT compiled in.
2001-02-24 19:36:13 +00:00
CTR4(KTR_LOCK,
"_mtx_lock_sleep: %s contested (lock=%p) at %s:%d",
m->mtx_description, (void *)m->mtx_lock, file, line);
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
while (!_obtain_lock(m, p)) {
uintptr_t v;
struct proc *p1;
- Make the mutex code almost completely machine independent. This greatly reducues the maintenance load for the mutex code. The only MD portions of the mutex code are in machine/mutex.h now, which include the assembly macros for handling mutexes as well as optionally overriding the mutex micro-operations. For example, we use optimized micro-ops on the x86 platform #ifndef I386_CPU. - Change the behavior of the SMP_DEBUG kernel option. In the new code, mtx_assert() only depends on INVARIANTS, allowing other kernel developers to have working mutex assertiions without having to include all of the mutex debugging code. The SMP_DEBUG kernel option has been renamed to MUTEX_DEBUG and now just controls extra mutex debugging code. - Abolish the ugly mtx_f hack. Instead, we dynamically allocate seperate mtx_debug structures on the fly in mtx_init, except for mutexes that are initiated very early in the boot process. These mutexes are declared using a special MUTEX_DECLARE() macro, and use a new flag MTX_COLD when calling mtx_init. This is still somewhat hackish, but it is less evil than the mtx_f filler struct, and the mtx struct is now the same size with and without mutex debugging code. - Add some micro-micro-operation macros for doing the actual atomic operations on the mutex mtx_lock field to make it easier for other archs to override/optimize mutex ops if needed. These new tiny ops also clean up the code in some places by replacing long atomic operation function calls that spanned 2-3 lines with a short 1-line macro call. - Don't call mi_switch() from mtx_enter_hard() when we block while trying to obtain a sleep mutex. Calling mi_switch() would bogusly release Giant before switching to the next process. Instead, inline most of the code from mi_switch() in the mtx_enter_hard() function. Note that when we finally kill Giant we can back this out and go back to calling mi_switch().
2000-10-20 07:26:37 +00:00
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
mtx_lock_spin(&sched_lock);
/*
* Check if the lock has been released while spinning for
* the sched_lock.
*/
if ((v = m->mtx_lock) == MTX_UNOWNED) {
mtx_unlock_spin(&sched_lock);
continue;
- Make the mutex code almost completely machine independent. This greatly reducues the maintenance load for the mutex code. The only MD portions of the mutex code are in machine/mutex.h now, which include the assembly macros for handling mutexes as well as optionally overriding the mutex micro-operations. For example, we use optimized micro-ops on the x86 platform #ifndef I386_CPU. - Change the behavior of the SMP_DEBUG kernel option. In the new code, mtx_assert() only depends on INVARIANTS, allowing other kernel developers to have working mutex assertiions without having to include all of the mutex debugging code. The SMP_DEBUG kernel option has been renamed to MUTEX_DEBUG and now just controls extra mutex debugging code. - Abolish the ugly mtx_f hack. Instead, we dynamically allocate seperate mtx_debug structures on the fly in mtx_init, except for mutexes that are initiated very early in the boot process. These mutexes are declared using a special MUTEX_DECLARE() macro, and use a new flag MTX_COLD when calling mtx_init. This is still somewhat hackish, but it is less evil than the mtx_f filler struct, and the mtx struct is now the same size with and without mutex debugging code. - Add some micro-micro-operation macros for doing the actual atomic operations on the mutex mtx_lock field to make it easier for other archs to override/optimize mutex ops if needed. These new tiny ops also clean up the code in some places by replacing long atomic operation function calls that spanned 2-3 lines with a short 1-line macro call. - Don't call mi_switch() from mtx_enter_hard() when we block while trying to obtain a sleep mutex. Calling mi_switch() would bogusly release Giant before switching to the next process. Instead, inline most of the code from mi_switch() in the mtx_enter_hard() function. Note that when we finally kill Giant we can back this out and go back to calling mi_switch().
2000-10-20 07:26:37 +00:00
}
Fix up priority propagation: - Use a better test for determining when a process is running. - Convert some checks to assertions. - Remove unnecessary tests. - Save the priority before acquiring a mutex rather than in msleep(9).
2000-11-30 00:51:16 +00:00
/*
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
* The mutex was marked contested on release. This means that
* there are processes blocked on it.
Fix up priority propagation: - Use a better test for determining when a process is running. - Convert some checks to assertions. - Remove unnecessary tests. - Save the priority before acquiring a mutex rather than in msleep(9).
2000-11-30 00:51:16 +00:00
*/
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
if (v == MTX_CONTESTED) {
p1 = TAILQ_FIRST(&m->mtx_blocked);
- Place back STR string declarations for lock/unlock strings used for KTR_LOCK tracing in order to avoid duplication. - Insert some tracepoints back into the mutex acq/rel code, thus ensuring that we can trace all lock acq/rel's again. - All CURPROC != NULL checks are MPASS()es (under MUTEX_DEBUG) because they signify a serious mutex corruption. - Change up some KASSERT()s to MPASS()es, and vice-versa, depending on the type of problem we're debugging (INVARIANTS is used here to check that the API is being used properly whereas MUTEX_DEBUG is used to ensure that something general isn't happening that will have bad impact on mutex locks). Reminded by: jhb, jake, asmodai
2001-02-11 02:54:16 +00:00
MPASS(p1 != NULL);
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
m->mtx_lock = (uintptr_t)p | MTX_CONTESTED;
Implement a unified run queue and adjust priority levels accordingly. - All processes go into the same array of queues, with different scheduling classes using different portions of the array. This allows user processes to have their priorities propogated up into interrupt thread range if need be. - I chose 64 run queues as an arbitrary number that is greater than 32. We used to have 4 separate arrays of 32 queues each, so this may not be optimal. The new run queue code was written with this in mind; changing the number of run queues only requires changing constants in runq.h and adjusting the priority levels. - The new run queue code takes the run queue as a parameter. This is intended to be used to create per-cpu run queues. Implement wrappers for compatibility with the old interface which pass in the global run queue structure. - Group the priority level, user priority, native priority (before propogation) and the scheduling class into a struct priority. - Change any hard coded priority levels that I found to use symbolic constants (TTIPRI and TTOPRI). - Remove the curpriority global variable and use that of curproc. This was used to detect when a process' priority had lowered and it should yield. We now effectively yield on every interrupt. - Activate propogate_priority(). It should now have the desired effect without needing to also propogate the scheduling class. - Temporarily comment out the call to vm_page_zero_idle() in the idle loop. It interfered with propogate_priority() because the idle process needed to do a non-blocking acquire of Giant and then other processes would try to propogate their priority onto it. The idle process should not do anything except idle. vm_page_zero_idle() will return in the form of an idle priority kernel thread which is woken up at apprioriate times by the vm system. - Update struct kinfo_proc to the new priority interface. Deliberately change its size by adjusting the spare fields. It remained the same size, but the layout has changed, so userland processes that use it would parse the data incorrectly. The size constraint should really be changed to an arbitrary version number. Also add a debug.sizeof sysctl node for struct kinfo_proc.
2001-02-12 00:20:08 +00:00
if (p1->p_pri.pri_level < p->p_pri.pri_level)
SET_PRIO(p, p1->p_pri.pri_level);
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
mtx_unlock_spin(&sched_lock);
return;
}
- Make the mutex code almost completely machine independent. This greatly reducues the maintenance load for the mutex code. The only MD portions of the mutex code are in machine/mutex.h now, which include the assembly macros for handling mutexes as well as optionally overriding the mutex micro-operations. For example, we use optimized micro-ops on the x86 platform #ifndef I386_CPU. - Change the behavior of the SMP_DEBUG kernel option. In the new code, mtx_assert() only depends on INVARIANTS, allowing other kernel developers to have working mutex assertiions without having to include all of the mutex debugging code. The SMP_DEBUG kernel option has been renamed to MUTEX_DEBUG and now just controls extra mutex debugging code. - Abolish the ugly mtx_f hack. Instead, we dynamically allocate seperate mtx_debug structures on the fly in mtx_init, except for mutexes that are initiated very early in the boot process. These mutexes are declared using a special MUTEX_DECLARE() macro, and use a new flag MTX_COLD when calling mtx_init. This is still somewhat hackish, but it is less evil than the mtx_f filler struct, and the mtx struct is now the same size with and without mutex debugging code. - Add some micro-micro-operation macros for doing the actual atomic operations on the mutex mtx_lock field to make it easier for other archs to override/optimize mutex ops if needed. These new tiny ops also clean up the code in some places by replacing long atomic operation function calls that spanned 2-3 lines with a short 1-line macro call. - Don't call mi_switch() from mtx_enter_hard() when we block while trying to obtain a sleep mutex. Calling mi_switch() would bogusly release Giant before switching to the next process. Instead, inline most of the code from mi_switch() in the mtx_enter_hard() function. Note that when we finally kill Giant we can back this out and go back to calling mi_switch().
2000-10-20 07:26:37 +00:00
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
/*
* If the mutex isn't already contested and a failure occurs
* setting the contested bit, the mutex was either released
* or the state of the MTX_RECURSED bit changed.
*/
if ((v & MTX_CONTESTED) == 0 &&
!atomic_cmpset_ptr(&m->mtx_lock, (void *)v,
(void *)(v | MTX_CONTESTED))) {
mtx_unlock_spin(&sched_lock);
continue;
}
- Make the mutex code almost completely machine independent. This greatly reducues the maintenance load for the mutex code. The only MD portions of the mutex code are in machine/mutex.h now, which include the assembly macros for handling mutexes as well as optionally overriding the mutex micro-operations. For example, we use optimized micro-ops on the x86 platform #ifndef I386_CPU. - Change the behavior of the SMP_DEBUG kernel option. In the new code, mtx_assert() only depends on INVARIANTS, allowing other kernel developers to have working mutex assertiions without having to include all of the mutex debugging code. The SMP_DEBUG kernel option has been renamed to MUTEX_DEBUG and now just controls extra mutex debugging code. - Abolish the ugly mtx_f hack. Instead, we dynamically allocate seperate mtx_debug structures on the fly in mtx_init, except for mutexes that are initiated very early in the boot process. These mutexes are declared using a special MUTEX_DECLARE() macro, and use a new flag MTX_COLD when calling mtx_init. This is still somewhat hackish, but it is less evil than the mtx_f filler struct, and the mtx struct is now the same size with and without mutex debugging code. - Add some micro-micro-operation macros for doing the actual atomic operations on the mutex mtx_lock field to make it easier for other archs to override/optimize mutex ops if needed. These new tiny ops also clean up the code in some places by replacing long atomic operation function calls that spanned 2-3 lines with a short 1-line macro call. - Don't call mi_switch() from mtx_enter_hard() when we block while trying to obtain a sleep mutex. Calling mi_switch() would bogusly release Giant before switching to the next process. Instead, inline most of the code from mi_switch() in the mtx_enter_hard() function. Note that when we finally kill Giant we can back this out and go back to calling mi_switch().
2000-10-20 07:26:37 +00:00
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
/*
* We deffinately must sleep for this lock.
*/
mtx_assert(m, MA_NOTOWNED);
- Make the mutex code almost completely machine independent. This greatly reducues the maintenance load for the mutex code. The only MD portions of the mutex code are in machine/mutex.h now, which include the assembly macros for handling mutexes as well as optionally overriding the mutex micro-operations. For example, we use optimized micro-ops on the x86 platform #ifndef I386_CPU. - Change the behavior of the SMP_DEBUG kernel option. In the new code, mtx_assert() only depends on INVARIANTS, allowing other kernel developers to have working mutex assertiions without having to include all of the mutex debugging code. The SMP_DEBUG kernel option has been renamed to MUTEX_DEBUG and now just controls extra mutex debugging code. - Abolish the ugly mtx_f hack. Instead, we dynamically allocate seperate mtx_debug structures on the fly in mtx_init, except for mutexes that are initiated very early in the boot process. These mutexes are declared using a special MUTEX_DECLARE() macro, and use a new flag MTX_COLD when calling mtx_init. This is still somewhat hackish, but it is less evil than the mtx_f filler struct, and the mtx struct is now the same size with and without mutex debugging code. - Add some micro-micro-operation macros for doing the actual atomic operations on the mutex mtx_lock field to make it easier for other archs to override/optimize mutex ops if needed. These new tiny ops also clean up the code in some places by replacing long atomic operation function calls that spanned 2-3 lines with a short 1-line macro call. - Don't call mi_switch() from mtx_enter_hard() when we block while trying to obtain a sleep mutex. Calling mi_switch() would bogusly release Giant before switching to the next process. Instead, inline most of the code from mi_switch() in the mtx_enter_hard() function. Note that when we finally kill Giant we can back this out and go back to calling mi_switch().
2000-10-20 07:26:37 +00:00
#ifdef notyet
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
/*
* If we're borrowing an interrupted thread's VM context, we
* must clean up before going to sleep.
*/
- Axe RETIP() as it was very i386 specific and unwieldy. Instead, use the passed in filename and line number in the KTR tracepoint message. - Even though it is #if 0'd code, change the code to detect that a process is an interrupt thread to check p->p_ithd against NULL rather than checking non-existant process flags from BSD/OS. - Use '%p' to print pointers in KTR log messages instead of assuming sizeof(int) == sizeof(void *). - Don't set p_mtxname to NULL when releasing a mutex. It doesn't hurt to leave it set (we don't clear w_mesg for example) and at least at one time in the past, there used to be race conditions in the kernel that would result in setting this to NULL causing the kernel to dereference NULL. - Make the _mtx_assert() function be compiled in if INVARIANTS_SUPPORT is defined rather than if INVARIANTS is defined so that a KLD compiled with INVARIANTS that uses mtx_assert() can be used with a kernel that just has INVARIANT_SUPPORT compiled in.
2001-02-24 19:36:13 +00:00
if (p->p_ithd != NULL) {
struct ithd *it = p->p_ithd;
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
if (it->it_interrupted) {
if ((opts & MTX_QUIET) == 0)
CTR2(KTR_LOCK,
- Axe RETIP() as it was very i386 specific and unwieldy. Instead, use the passed in filename and line number in the KTR tracepoint message. - Even though it is #if 0'd code, change the code to detect that a process is an interrupt thread to check p->p_ithd against NULL rather than checking non-existant process flags from BSD/OS. - Use '%p' to print pointers in KTR log messages instead of assuming sizeof(int) == sizeof(void *). - Don't set p_mtxname to NULL when releasing a mutex. It doesn't hurt to leave it set (we don't clear w_mesg for example) and at least at one time in the past, there used to be race conditions in the kernel that would result in setting this to NULL causing the kernel to dereference NULL. - Make the _mtx_assert() function be compiled in if INVARIANTS_SUPPORT is defined rather than if INVARIANTS is defined so that a KLD compiled with INVARIANTS that uses mtx_assert() can be used with a kernel that just has INVARIANT_SUPPORT compiled in.
2001-02-24 19:36:13 +00:00
"_mtx_lock_sleep: %p interrupted %p",
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
it, it->it_interrupted);
intr_thd_fixup(it);
- Make the mutex code almost completely machine independent. This greatly reducues the maintenance load for the mutex code. The only MD portions of the mutex code are in machine/mutex.h now, which include the assembly macros for handling mutexes as well as optionally overriding the mutex micro-operations. For example, we use optimized micro-ops on the x86 platform #ifndef I386_CPU. - Change the behavior of the SMP_DEBUG kernel option. In the new code, mtx_assert() only depends on INVARIANTS, allowing other kernel developers to have working mutex assertiions without having to include all of the mutex debugging code. The SMP_DEBUG kernel option has been renamed to MUTEX_DEBUG and now just controls extra mutex debugging code. - Abolish the ugly mtx_f hack. Instead, we dynamically allocate seperate mtx_debug structures on the fly in mtx_init, except for mutexes that are initiated very early in the boot process. These mutexes are declared using a special MUTEX_DECLARE() macro, and use a new flag MTX_COLD when calling mtx_init. This is still somewhat hackish, but it is less evil than the mtx_f filler struct, and the mtx struct is now the same size with and without mutex debugging code. - Add some micro-micro-operation macros for doing the actual atomic operations on the mutex mtx_lock field to make it easier for other archs to override/optimize mutex ops if needed. These new tiny ops also clean up the code in some places by replacing long atomic operation function calls that spanned 2-3 lines with a short 1-line macro call. - Don't call mi_switch() from mtx_enter_hard() when we block while trying to obtain a sleep mutex. Calling mi_switch() would bogusly release Giant before switching to the next process. Instead, inline most of the code from mi_switch() in the mtx_enter_hard() function. Note that when we finally kill Giant we can back this out and go back to calling mi_switch().
2000-10-20 07:26:37 +00:00
}
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
}
- Make the mutex code almost completely machine independent. This greatly reducues the maintenance load for the mutex code. The only MD portions of the mutex code are in machine/mutex.h now, which include the assembly macros for handling mutexes as well as optionally overriding the mutex micro-operations. For example, we use optimized micro-ops on the x86 platform #ifndef I386_CPU. - Change the behavior of the SMP_DEBUG kernel option. In the new code, mtx_assert() only depends on INVARIANTS, allowing other kernel developers to have working mutex assertiions without having to include all of the mutex debugging code. The SMP_DEBUG kernel option has been renamed to MUTEX_DEBUG and now just controls extra mutex debugging code. - Abolish the ugly mtx_f hack. Instead, we dynamically allocate seperate mtx_debug structures on the fly in mtx_init, except for mutexes that are initiated very early in the boot process. These mutexes are declared using a special MUTEX_DECLARE() macro, and use a new flag MTX_COLD when calling mtx_init. This is still somewhat hackish, but it is less evil than the mtx_f filler struct, and the mtx struct is now the same size with and without mutex debugging code. - Add some micro-micro-operation macros for doing the actual atomic operations on the mutex mtx_lock field to make it easier for other archs to override/optimize mutex ops if needed. These new tiny ops also clean up the code in some places by replacing long atomic operation function calls that spanned 2-3 lines with a short 1-line macro call. - Don't call mi_switch() from mtx_enter_hard() when we block while trying to obtain a sleep mutex. Calling mi_switch() would bogusly release Giant before switching to the next process. Instead, inline most of the code from mi_switch() in the mtx_enter_hard() function. Note that when we finally kill Giant we can back this out and go back to calling mi_switch().
2000-10-20 07:26:37 +00:00
#endif
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
/*
* Put us on the list of threads blocked on this mutex.
*/
if (TAILQ_EMPTY(&m->mtx_blocked)) {
p1 = (struct proc *)(m->mtx_lock & MTX_FLAGMASK);
LIST_INSERT_HEAD(&p1->p_contested, m, mtx_contested);
TAILQ_INSERT_TAIL(&m->mtx_blocked, p, p_procq);
} else {
TAILQ_FOREACH(p1, &m->mtx_blocked, p_procq)
Implement a unified run queue and adjust priority levels accordingly. - All processes go into the same array of queues, with different scheduling classes using different portions of the array. This allows user processes to have their priorities propogated up into interrupt thread range if need be. - I chose 64 run queues as an arbitrary number that is greater than 32. We used to have 4 separate arrays of 32 queues each, so this may not be optimal. The new run queue code was written with this in mind; changing the number of run queues only requires changing constants in runq.h and adjusting the priority levels. - The new run queue code takes the run queue as a parameter. This is intended to be used to create per-cpu run queues. Implement wrappers for compatibility with the old interface which pass in the global run queue structure. - Group the priority level, user priority, native priority (before propogation) and the scheduling class into a struct priority. - Change any hard coded priority levels that I found to use symbolic constants (TTIPRI and TTOPRI). - Remove the curpriority global variable and use that of curproc. This was used to detect when a process' priority had lowered and it should yield. We now effectively yield on every interrupt. - Activate propogate_priority(). It should now have the desired effect without needing to also propogate the scheduling class. - Temporarily comment out the call to vm_page_zero_idle() in the idle loop. It interfered with propogate_priority() because the idle process needed to do a non-blocking acquire of Giant and then other processes would try to propogate their priority onto it. The idle process should not do anything except idle. vm_page_zero_idle() will return in the form of an idle priority kernel thread which is woken up at apprioriate times by the vm system. - Update struct kinfo_proc to the new priority interface. Deliberately change its size by adjusting the spare fields. It remained the same size, but the layout has changed, so userland processes that use it would parse the data incorrectly. The size constraint should really be changed to an arbitrary version number. Also add a debug.sizeof sysctl node for struct kinfo_proc.
2001-02-12 00:20:08 +00:00
if (p1->p_pri.pri_level > p->p_pri.pri_level)
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
break;
if (p1)
TAILQ_INSERT_BEFORE(p1, p, p_procq);
else
- Make the mutex code almost completely machine independent. This greatly reducues the maintenance load for the mutex code. The only MD portions of the mutex code are in machine/mutex.h now, which include the assembly macros for handling mutexes as well as optionally overriding the mutex micro-operations. For example, we use optimized micro-ops on the x86 platform #ifndef I386_CPU. - Change the behavior of the SMP_DEBUG kernel option. In the new code, mtx_assert() only depends on INVARIANTS, allowing other kernel developers to have working mutex assertiions without having to include all of the mutex debugging code. The SMP_DEBUG kernel option has been renamed to MUTEX_DEBUG and now just controls extra mutex debugging code. - Abolish the ugly mtx_f hack. Instead, we dynamically allocate seperate mtx_debug structures on the fly in mtx_init, except for mutexes that are initiated very early in the boot process. These mutexes are declared using a special MUTEX_DECLARE() macro, and use a new flag MTX_COLD when calling mtx_init. This is still somewhat hackish, but it is less evil than the mtx_f filler struct, and the mtx struct is now the same size with and without mutex debugging code. - Add some micro-micro-operation macros for doing the actual atomic operations on the mutex mtx_lock field to make it easier for other archs to override/optimize mutex ops if needed. These new tiny ops also clean up the code in some places by replacing long atomic operation function calls that spanned 2-3 lines with a short 1-line macro call. - Don't call mi_switch() from mtx_enter_hard() when we block while trying to obtain a sleep mutex. Calling mi_switch() would bogusly release Giant before switching to the next process. Instead, inline most of the code from mi_switch() in the mtx_enter_hard() function. Note that when we finally kill Giant we can back this out and go back to calling mi_switch().
2000-10-20 07:26:37 +00:00
TAILQ_INSERT_TAIL(&m->mtx_blocked, p, p_procq);
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
}
- Make the mutex code almost completely machine independent. This greatly reducues the maintenance load for the mutex code. The only MD portions of the mutex code are in machine/mutex.h now, which include the assembly macros for handling mutexes as well as optionally overriding the mutex micro-operations. For example, we use optimized micro-ops on the x86 platform #ifndef I386_CPU. - Change the behavior of the SMP_DEBUG kernel option. In the new code, mtx_assert() only depends on INVARIANTS, allowing other kernel developers to have working mutex assertiions without having to include all of the mutex debugging code. The SMP_DEBUG kernel option has been renamed to MUTEX_DEBUG and now just controls extra mutex debugging code. - Abolish the ugly mtx_f hack. Instead, we dynamically allocate seperate mtx_debug structures on the fly in mtx_init, except for mutexes that are initiated very early in the boot process. These mutexes are declared using a special MUTEX_DECLARE() macro, and use a new flag MTX_COLD when calling mtx_init. This is still somewhat hackish, but it is less evil than the mtx_f filler struct, and the mtx struct is now the same size with and without mutex debugging code. - Add some micro-micro-operation macros for doing the actual atomic operations on the mutex mtx_lock field to make it easier for other archs to override/optimize mutex ops if needed. These new tiny ops also clean up the code in some places by replacing long atomic operation function calls that spanned 2-3 lines with a short 1-line macro call. - Don't call mi_switch() from mtx_enter_hard() when we block while trying to obtain a sleep mutex. Calling mi_switch() would bogusly release Giant before switching to the next process. Instead, inline most of the code from mi_switch() in the mtx_enter_hard() function. Note that when we finally kill Giant we can back this out and go back to calling mi_switch().
2000-10-20 07:26:37 +00:00
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
/*
* Save who we're blocked on.
*/
p->p_blocked = m;
p->p_mtxname = m->mtx_description;
p->p_stat = SMTX;
propagate_priority(p);
if ((opts & MTX_QUIET) == 0)
CTR3(KTR_LOCK,
"_mtx_lock_sleep: p %p blocked on [%p] %s", p, m,
m->mtx_description);
mi_switch();
if ((opts & MTX_QUIET) == 0)
CTR3(KTR_LOCK,
"_mtx_lock_sleep: p %p free from blocked on [%p] %s",
p, m, m->mtx_description);
mtx_unlock_spin(&sched_lock);
}
return;
}
/*
* _mtx_lock_spin: the tougher part of acquiring an MTX_SPIN lock.
*
* This is only called if we need to actually spin for the lock. Recursion
* is handled inline.
*/
void
_mtx_lock_spin(struct mtx *m, int opts, u_int mtx_intr, const char *file,
int line)
{
int i = 0;
if ((opts & MTX_QUIET) == 0)
- Place back STR string declarations for lock/unlock strings used for KTR_LOCK tracing in order to avoid duplication. - Insert some tracepoints back into the mutex acq/rel code, thus ensuring that we can trace all lock acq/rel's again. - All CURPROC != NULL checks are MPASS()es (under MUTEX_DEBUG) because they signify a serious mutex corruption. - Change up some KASSERT()s to MPASS()es, and vice-versa, depending on the type of problem we're debugging (INVARIANTS is used here to check that the API is being used properly whereas MUTEX_DEBUG is used to ensure that something general isn't happening that will have bad impact on mutex locks). Reminded by: jhb, jake, asmodai
2001-02-11 02:54:16 +00:00
CTR1(KTR_LOCK, "_mtx_lock_spin: %p spinning", m);
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
for (;;) {
Change all instances of `CURPROC' and `CURTHD' to `curproc,' in order to stay consistent. Requested by: bde
2001-02-12 03:15:43 +00:00
if (_obtain_lock(m, curproc))
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
break;
while (m->mtx_lock != MTX_UNOWNED) {
if (i++ < 1000000)
continue;
if (i++ < 6000000)
DELAY(1);
- Make the mutex code almost completely machine independent. This greatly reducues the maintenance load for the mutex code. The only MD portions of the mutex code are in machine/mutex.h now, which include the assembly macros for handling mutexes as well as optionally overriding the mutex micro-operations. For example, we use optimized micro-ops on the x86 platform #ifndef I386_CPU. - Change the behavior of the SMP_DEBUG kernel option. In the new code, mtx_assert() only depends on INVARIANTS, allowing other kernel developers to have working mutex assertiions without having to include all of the mutex debugging code. The SMP_DEBUG kernel option has been renamed to MUTEX_DEBUG and now just controls extra mutex debugging code. - Abolish the ugly mtx_f hack. Instead, we dynamically allocate seperate mtx_debug structures on the fly in mtx_init, except for mutexes that are initiated very early in the boot process. These mutexes are declared using a special MUTEX_DECLARE() macro, and use a new flag MTX_COLD when calling mtx_init. This is still somewhat hackish, but it is less evil than the mtx_f filler struct, and the mtx struct is now the same size with and without mutex debugging code. - Add some micro-micro-operation macros for doing the actual atomic operations on the mutex mtx_lock field to make it easier for other archs to override/optimize mutex ops if needed. These new tiny ops also clean up the code in some places by replacing long atomic operation function calls that spanned 2-3 lines with a short 1-line macro call. - Don't call mi_switch() from mtx_enter_hard() when we block while trying to obtain a sleep mutex. Calling mi_switch() would bogusly release Giant before switching to the next process. Instead, inline most of the code from mi_switch() in the mtx_enter_hard() function. Note that when we finally kill Giant we can back this out and go back to calling mi_switch().
2000-10-20 07:26:37 +00:00
#ifdef DDB
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
else if (!db_active)
- Make the mutex code almost completely machine independent. This greatly reducues the maintenance load for the mutex code. The only MD portions of the mutex code are in machine/mutex.h now, which include the assembly macros for handling mutexes as well as optionally overriding the mutex micro-operations. For example, we use optimized micro-ops on the x86 platform #ifndef I386_CPU. - Change the behavior of the SMP_DEBUG kernel option. In the new code, mtx_assert() only depends on INVARIANTS, allowing other kernel developers to have working mutex assertiions without having to include all of the mutex debugging code. The SMP_DEBUG kernel option has been renamed to MUTEX_DEBUG and now just controls extra mutex debugging code. - Abolish the ugly mtx_f hack. Instead, we dynamically allocate seperate mtx_debug structures on the fly in mtx_init, except for mutexes that are initiated very early in the boot process. These mutexes are declared using a special MUTEX_DECLARE() macro, and use a new flag MTX_COLD when calling mtx_init. This is still somewhat hackish, but it is less evil than the mtx_f filler struct, and the mtx struct is now the same size with and without mutex debugging code. - Add some micro-micro-operation macros for doing the actual atomic operations on the mutex mtx_lock field to make it easier for other archs to override/optimize mutex ops if needed. These new tiny ops also clean up the code in some places by replacing long atomic operation function calls that spanned 2-3 lines with a short 1-line macro call. - Don't call mi_switch() from mtx_enter_hard() when we block while trying to obtain a sleep mutex. Calling mi_switch() would bogusly release Giant before switching to the next process. Instead, inline most of the code from mi_switch() in the mtx_enter_hard() function. Note that when we finally kill Giant we can back this out and go back to calling mi_switch().
2000-10-20 07:26:37 +00:00
#else
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
else
- Make the mutex code almost completely machine independent. This greatly reducues the maintenance load for the mutex code. The only MD portions of the mutex code are in machine/mutex.h now, which include the assembly macros for handling mutexes as well as optionally overriding the mutex micro-operations. For example, we use optimized micro-ops on the x86 platform #ifndef I386_CPU. - Change the behavior of the SMP_DEBUG kernel option. In the new code, mtx_assert() only depends on INVARIANTS, allowing other kernel developers to have working mutex assertiions without having to include all of the mutex debugging code. The SMP_DEBUG kernel option has been renamed to MUTEX_DEBUG and now just controls extra mutex debugging code. - Abolish the ugly mtx_f hack. Instead, we dynamically allocate seperate mtx_debug structures on the fly in mtx_init, except for mutexes that are initiated very early in the boot process. These mutexes are declared using a special MUTEX_DECLARE() macro, and use a new flag MTX_COLD when calling mtx_init. This is still somewhat hackish, but it is less evil than the mtx_f filler struct, and the mtx struct is now the same size with and without mutex debugging code. - Add some micro-micro-operation macros for doing the actual atomic operations on the mutex mtx_lock field to make it easier for other archs to override/optimize mutex ops if needed. These new tiny ops also clean up the code in some places by replacing long atomic operation function calls that spanned 2-3 lines with a short 1-line macro call. - Don't call mi_switch() from mtx_enter_hard() when we block while trying to obtain a sleep mutex. Calling mi_switch() would bogusly release Giant before switching to the next process. Instead, inline most of the code from mi_switch() in the mtx_enter_hard() function. Note that when we finally kill Giant we can back this out and go back to calling mi_switch().
2000-10-20 07:26:37 +00:00
#endif
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
panic("spin lock %s held by %p for > 5 seconds",
m->mtx_description, (void *)m->mtx_lock);
- Make the mutex code almost completely machine independent. This greatly reducues the maintenance load for the mutex code. The only MD portions of the mutex code are in machine/mutex.h now, which include the assembly macros for handling mutexes as well as optionally overriding the mutex micro-operations. For example, we use optimized micro-ops on the x86 platform #ifndef I386_CPU. - Change the behavior of the SMP_DEBUG kernel option. In the new code, mtx_assert() only depends on INVARIANTS, allowing other kernel developers to have working mutex assertiions without having to include all of the mutex debugging code. The SMP_DEBUG kernel option has been renamed to MUTEX_DEBUG and now just controls extra mutex debugging code. - Abolish the ugly mtx_f hack. Instead, we dynamically allocate seperate mtx_debug structures on the fly in mtx_init, except for mutexes that are initiated very early in the boot process. These mutexes are declared using a special MUTEX_DECLARE() macro, and use a new flag MTX_COLD when calling mtx_init. This is still somewhat hackish, but it is less evil than the mtx_f filler struct, and the mtx struct is now the same size with and without mutex debugging code. - Add some micro-micro-operation macros for doing the actual atomic operations on the mutex mtx_lock field to make it easier for other archs to override/optimize mutex ops if needed. These new tiny ops also clean up the code in some places by replacing long atomic operation function calls that spanned 2-3 lines with a short 1-line macro call. - Don't call mi_switch() from mtx_enter_hard() when we block while trying to obtain a sleep mutex. Calling mi_switch() would bogusly release Giant before switching to the next process. Instead, inline most of the code from mi_switch() in the mtx_enter_hard() function. Note that when we finally kill Giant we can back this out and go back to calling mi_switch().
2000-10-20 07:26:37 +00:00
}
}
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
m->mtx_saveintr = mtx_intr;
if ((opts & MTX_QUIET) == 0)
CTR1(KTR_LOCK, "_mtx_lock_spin: %p spin done", m);
return;
- Make the mutex code almost completely machine independent. This greatly reducues the maintenance load for the mutex code. The only MD portions of the mutex code are in machine/mutex.h now, which include the assembly macros for handling mutexes as well as optionally overriding the mutex micro-operations. For example, we use optimized micro-ops on the x86 platform #ifndef I386_CPU. - Change the behavior of the SMP_DEBUG kernel option. In the new code, mtx_assert() only depends on INVARIANTS, allowing other kernel developers to have working mutex assertiions without having to include all of the mutex debugging code. The SMP_DEBUG kernel option has been renamed to MUTEX_DEBUG and now just controls extra mutex debugging code. - Abolish the ugly mtx_f hack. Instead, we dynamically allocate seperate mtx_debug structures on the fly in mtx_init, except for mutexes that are initiated very early in the boot process. These mutexes are declared using a special MUTEX_DECLARE() macro, and use a new flag MTX_COLD when calling mtx_init. This is still somewhat hackish, but it is less evil than the mtx_f filler struct, and the mtx struct is now the same size with and without mutex debugging code. - Add some micro-micro-operation macros for doing the actual atomic operations on the mutex mtx_lock field to make it easier for other archs to override/optimize mutex ops if needed. These new tiny ops also clean up the code in some places by replacing long atomic operation function calls that spanned 2-3 lines with a short 1-line macro call. - Don't call mi_switch() from mtx_enter_hard() when we block while trying to obtain a sleep mutex. Calling mi_switch() would bogusly release Giant before switching to the next process. Instead, inline most of the code from mi_switch() in the mtx_enter_hard() function. Note that when we finally kill Giant we can back this out and go back to calling mi_switch().
2000-10-20 07:26:37 +00:00
}
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
/*
* _mtx_unlock_sleep: the tougher part of releasing an MTX_DEF lock.
*
* We are only called here if the lock is recursed or contested (i.e. we
* need to wake up a blocked thread).
*/
- Make the mutex code almost completely machine independent. This greatly reducues the maintenance load for the mutex code. The only MD portions of the mutex code are in machine/mutex.h now, which include the assembly macros for handling mutexes as well as optionally overriding the mutex micro-operations. For example, we use optimized micro-ops on the x86 platform #ifndef I386_CPU. - Change the behavior of the SMP_DEBUG kernel option. In the new code, mtx_assert() only depends on INVARIANTS, allowing other kernel developers to have working mutex assertiions without having to include all of the mutex debugging code. The SMP_DEBUG kernel option has been renamed to MUTEX_DEBUG and now just controls extra mutex debugging code. - Abolish the ugly mtx_f hack. Instead, we dynamically allocate seperate mtx_debug structures on the fly in mtx_init, except for mutexes that are initiated very early in the boot process. These mutexes are declared using a special MUTEX_DECLARE() macro, and use a new flag MTX_COLD when calling mtx_init. This is still somewhat hackish, but it is less evil than the mtx_f filler struct, and the mtx struct is now the same size with and without mutex debugging code. - Add some micro-micro-operation macros for doing the actual atomic operations on the mutex mtx_lock field to make it easier for other archs to override/optimize mutex ops if needed. These new tiny ops also clean up the code in some places by replacing long atomic operation function calls that spanned 2-3 lines with a short 1-line macro call. - Don't call mi_switch() from mtx_enter_hard() when we block while trying to obtain a sleep mutex. Calling mi_switch() would bogusly release Giant before switching to the next process. Instead, inline most of the code from mi_switch() in the mtx_enter_hard() function. Note that when we finally kill Giant we can back this out and go back to calling mi_switch().
2000-10-20 07:26:37 +00:00
void
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
_mtx_unlock_sleep(struct mtx *m, int opts, const char *file, int line)
- Make the mutex code almost completely machine independent. This greatly reducues the maintenance load for the mutex code. The only MD portions of the mutex code are in machine/mutex.h now, which include the assembly macros for handling mutexes as well as optionally overriding the mutex micro-operations. For example, we use optimized micro-ops on the x86 platform #ifndef I386_CPU. - Change the behavior of the SMP_DEBUG kernel option. In the new code, mtx_assert() only depends on INVARIANTS, allowing other kernel developers to have working mutex assertiions without having to include all of the mutex debugging code. The SMP_DEBUG kernel option has been renamed to MUTEX_DEBUG and now just controls extra mutex debugging code. - Abolish the ugly mtx_f hack. Instead, we dynamically allocate seperate mtx_debug structures on the fly in mtx_init, except for mutexes that are initiated very early in the boot process. These mutexes are declared using a special MUTEX_DECLARE() macro, and use a new flag MTX_COLD when calling mtx_init. This is still somewhat hackish, but it is less evil than the mtx_f filler struct, and the mtx struct is now the same size with and without mutex debugging code. - Add some micro-micro-operation macros for doing the actual atomic operations on the mutex mtx_lock field to make it easier for other archs to override/optimize mutex ops if needed. These new tiny ops also clean up the code in some places by replacing long atomic operation function calls that spanned 2-3 lines with a short 1-line macro call. - Don't call mi_switch() from mtx_enter_hard() when we block while trying to obtain a sleep mutex. Calling mi_switch() would bogusly release Giant before switching to the next process. Instead, inline most of the code from mi_switch() in the mtx_enter_hard() function. Note that when we finally kill Giant we can back this out and go back to calling mi_switch().
2000-10-20 07:26:37 +00:00
{
struct proc *p, *p1;
struct mtx *m1;
int pri;
Change all instances of `CURPROC' and `CURTHD' to `curproc,' in order to stay consistent. Requested by: bde
2001-02-12 03:15:43 +00:00
p = curproc;
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
if (mtx_recursed(m)) {
if (--(m->mtx_recurse) == 0)
atomic_clear_ptr(&m->mtx_lock, MTX_RECURSED);
if ((opts & MTX_QUIET) == 0)
CTR1(KTR_LOCK, "_mtx_unlock_sleep: %p unrecurse", m);
return;
}
mtx_lock_spin(&sched_lock);
if ((opts & MTX_QUIET) == 0)
CTR1(KTR_LOCK, "_mtx_unlock_sleep: %p contested", m);
p1 = TAILQ_FIRST(&m->mtx_blocked);
MPASS(p->p_magic == P_MAGIC);
MPASS(p1->p_magic == P_MAGIC);
TAILQ_REMOVE(&m->mtx_blocked, p1, p_procq);
if (TAILQ_EMPTY(&m->mtx_blocked)) {
LIST_REMOVE(m, mtx_contested);
_release_lock_quick(m);
if ((opts & MTX_QUIET) == 0)
CTR1(KTR_LOCK, "_mtx_unlock_sleep: %p not held", m);
} else
atomic_store_rel_ptr(&m->mtx_lock, (void *)MTX_CONTESTED);
Implement a unified run queue and adjust priority levels accordingly. - All processes go into the same array of queues, with different scheduling classes using different portions of the array. This allows user processes to have their priorities propogated up into interrupt thread range if need be. - I chose 64 run queues as an arbitrary number that is greater than 32. We used to have 4 separate arrays of 32 queues each, so this may not be optimal. The new run queue code was written with this in mind; changing the number of run queues only requires changing constants in runq.h and adjusting the priority levels. - The new run queue code takes the run queue as a parameter. This is intended to be used to create per-cpu run queues. Implement wrappers for compatibility with the old interface which pass in the global run queue structure. - Group the priority level, user priority, native priority (before propogation) and the scheduling class into a struct priority. - Change any hard coded priority levels that I found to use symbolic constants (TTIPRI and TTOPRI). - Remove the curpriority global variable and use that of curproc. This was used to detect when a process' priority had lowered and it should yield. We now effectively yield on every interrupt. - Activate propogate_priority(). It should now have the desired effect without needing to also propogate the scheduling class. - Temporarily comment out the call to vm_page_zero_idle() in the idle loop. It interfered with propogate_priority() because the idle process needed to do a non-blocking acquire of Giant and then other processes would try to propogate their priority onto it. The idle process should not do anything except idle. vm_page_zero_idle() will return in the form of an idle priority kernel thread which is woken up at apprioriate times by the vm system. - Update struct kinfo_proc to the new priority interface. Deliberately change its size by adjusting the spare fields. It remained the same size, but the layout has changed, so userland processes that use it would parse the data incorrectly. The size constraint should really be changed to an arbitrary version number. Also add a debug.sizeof sysctl node for struct kinfo_proc.
2001-02-12 00:20:08 +00:00
pri = PRI_MAX;
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
LIST_FOREACH(m1, &p->p_contested, mtx_contested) {
Implement a unified run queue and adjust priority levels accordingly. - All processes go into the same array of queues, with different scheduling classes using different portions of the array. This allows user processes to have their priorities propogated up into interrupt thread range if need be. - I chose 64 run queues as an arbitrary number that is greater than 32. We used to have 4 separate arrays of 32 queues each, so this may not be optimal. The new run queue code was written with this in mind; changing the number of run queues only requires changing constants in runq.h and adjusting the priority levels. - The new run queue code takes the run queue as a parameter. This is intended to be used to create per-cpu run queues. Implement wrappers for compatibility with the old interface which pass in the global run queue structure. - Group the priority level, user priority, native priority (before propogation) and the scheduling class into a struct priority. - Change any hard coded priority levels that I found to use symbolic constants (TTIPRI and TTOPRI). - Remove the curpriority global variable and use that of curproc. This was used to detect when a process' priority had lowered and it should yield. We now effectively yield on every interrupt. - Activate propogate_priority(). It should now have the desired effect without needing to also propogate the scheduling class. - Temporarily comment out the call to vm_page_zero_idle() in the idle loop. It interfered with propogate_priority() because the idle process needed to do a non-blocking acquire of Giant and then other processes would try to propogate their priority onto it. The idle process should not do anything except idle. vm_page_zero_idle() will return in the form of an idle priority kernel thread which is woken up at apprioriate times by the vm system. - Update struct kinfo_proc to the new priority interface. Deliberately change its size by adjusting the spare fields. It remained the same size, but the layout has changed, so userland processes that use it would parse the data incorrectly. The size constraint should really be changed to an arbitrary version number. Also add a debug.sizeof sysctl node for struct kinfo_proc.
2001-02-12 00:20:08 +00:00
int cp = TAILQ_FIRST(&m1->mtx_blocked)->p_pri.pri_level;
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
if (cp < pri)
pri = cp;
}
Implement a unified run queue and adjust priority levels accordingly. - All processes go into the same array of queues, with different scheduling classes using different portions of the array. This allows user processes to have their priorities propogated up into interrupt thread range if need be. - I chose 64 run queues as an arbitrary number that is greater than 32. We used to have 4 separate arrays of 32 queues each, so this may not be optimal. The new run queue code was written with this in mind; changing the number of run queues only requires changing constants in runq.h and adjusting the priority levels. - The new run queue code takes the run queue as a parameter. This is intended to be used to create per-cpu run queues. Implement wrappers for compatibility with the old interface which pass in the global run queue structure. - Group the priority level, user priority, native priority (before propogation) and the scheduling class into a struct priority. - Change any hard coded priority levels that I found to use symbolic constants (TTIPRI and TTOPRI). - Remove the curpriority global variable and use that of curproc. This was used to detect when a process' priority had lowered and it should yield. We now effectively yield on every interrupt. - Activate propogate_priority(). It should now have the desired effect without needing to also propogate the scheduling class. - Temporarily comment out the call to vm_page_zero_idle() in the idle loop. It interfered with propogate_priority() because the idle process needed to do a non-blocking acquire of Giant and then other processes would try to propogate their priority onto it. The idle process should not do anything except idle. vm_page_zero_idle() will return in the form of an idle priority kernel thread which is woken up at apprioriate times by the vm system. - Update struct kinfo_proc to the new priority interface. Deliberately change its size by adjusting the spare fields. It remained the same size, but the layout has changed, so userland processes that use it would parse the data incorrectly. The size constraint should really be changed to an arbitrary version number. Also add a debug.sizeof sysctl node for struct kinfo_proc.
2001-02-12 00:20:08 +00:00
if (pri > p->p_pri.pri_native)
pri = p->p_pri.pri_native;
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
SET_PRIO(p, pri);
if ((opts & MTX_QUIET) == 0)
CTR2(KTR_LOCK, "_mtx_unlock_sleep: %p contested setrunqueue %p",
m, p1);
p1->p_blocked = NULL;
p1->p_stat = SRUN;
setrunqueue(p1);
Implement a unified run queue and adjust priority levels accordingly. - All processes go into the same array of queues, with different scheduling classes using different portions of the array. This allows user processes to have their priorities propogated up into interrupt thread range if need be. - I chose 64 run queues as an arbitrary number that is greater than 32. We used to have 4 separate arrays of 32 queues each, so this may not be optimal. The new run queue code was written with this in mind; changing the number of run queues only requires changing constants in runq.h and adjusting the priority levels. - The new run queue code takes the run queue as a parameter. This is intended to be used to create per-cpu run queues. Implement wrappers for compatibility with the old interface which pass in the global run queue structure. - Group the priority level, user priority, native priority (before propogation) and the scheduling class into a struct priority. - Change any hard coded priority levels that I found to use symbolic constants (TTIPRI and TTOPRI). - Remove the curpriority global variable and use that of curproc. This was used to detect when a process' priority had lowered and it should yield. We now effectively yield on every interrupt. - Activate propogate_priority(). It should now have the desired effect without needing to also propogate the scheduling class. - Temporarily comment out the call to vm_page_zero_idle() in the idle loop. It interfered with propogate_priority() because the idle process needed to do a non-blocking acquire of Giant and then other processes would try to propogate their priority onto it. The idle process should not do anything except idle. vm_page_zero_idle() will return in the form of an idle priority kernel thread which is woken up at apprioriate times by the vm system. - Update struct kinfo_proc to the new priority interface. Deliberately change its size by adjusting the spare fields. It remained the same size, but the layout has changed, so userland processes that use it would parse the data incorrectly. The size constraint should really be changed to an arbitrary version number. Also add a debug.sizeof sysctl node for struct kinfo_proc.
2001-02-12 00:20:08 +00:00
if ((opts & MTX_NOSWITCH) == 0 && p1->p_pri.pri_level < pri) {
- Make the mutex code almost completely machine independent. This greatly reducues the maintenance load for the mutex code. The only MD portions of the mutex code are in machine/mutex.h now, which include the assembly macros for handling mutexes as well as optionally overriding the mutex micro-operations. For example, we use optimized micro-ops on the x86 platform #ifndef I386_CPU. - Change the behavior of the SMP_DEBUG kernel option. In the new code, mtx_assert() only depends on INVARIANTS, allowing other kernel developers to have working mutex assertiions without having to include all of the mutex debugging code. The SMP_DEBUG kernel option has been renamed to MUTEX_DEBUG and now just controls extra mutex debugging code. - Abolish the ugly mtx_f hack. Instead, we dynamically allocate seperate mtx_debug structures on the fly in mtx_init, except for mutexes that are initiated very early in the boot process. These mutexes are declared using a special MUTEX_DECLARE() macro, and use a new flag MTX_COLD when calling mtx_init. This is still somewhat hackish, but it is less evil than the mtx_f filler struct, and the mtx struct is now the same size with and without mutex debugging code. - Add some micro-micro-operation macros for doing the actual atomic operations on the mutex mtx_lock field to make it easier for other archs to override/optimize mutex ops if needed. These new tiny ops also clean up the code in some places by replacing long atomic operation function calls that spanned 2-3 lines with a short 1-line macro call. - Don't call mi_switch() from mtx_enter_hard() when we block while trying to obtain a sleep mutex. Calling mi_switch() would bogusly release Giant before switching to the next process. Instead, inline most of the code from mi_switch() in the mtx_enter_hard() function. Note that when we finally kill Giant we can back this out and go back to calling mi_switch().
2000-10-20 07:26:37 +00:00
#ifdef notyet
- Axe RETIP() as it was very i386 specific and unwieldy. Instead, use the passed in filename and line number in the KTR tracepoint message. - Even though it is #if 0'd code, change the code to detect that a process is an interrupt thread to check p->p_ithd against NULL rather than checking non-existant process flags from BSD/OS. - Use '%p' to print pointers in KTR log messages instead of assuming sizeof(int) == sizeof(void *). - Don't set p_mtxname to NULL when releasing a mutex. It doesn't hurt to leave it set (we don't clear w_mesg for example) and at least at one time in the past, there used to be race conditions in the kernel that would result in setting this to NULL causing the kernel to dereference NULL. - Make the _mtx_assert() function be compiled in if INVARIANTS_SUPPORT is defined rather than if INVARIANTS is defined so that a KLD compiled with INVARIANTS that uses mtx_assert() can be used with a kernel that just has INVARIANT_SUPPORT compiled in.
2001-02-24 19:36:13 +00:00
if (p->p_ithd != NULL) {
struct ithd *it = p->p_ithd;
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
if (it->it_interrupted) {
if ((opts & MTX_QUIET) == 0)
CTR2(KTR_LOCK,
- Axe RETIP() as it was very i386 specific and unwieldy. Instead, use the passed in filename and line number in the KTR tracepoint message. - Even though it is #if 0'd code, change the code to detect that a process is an interrupt thread to check p->p_ithd against NULL rather than checking non-existant process flags from BSD/OS. - Use '%p' to print pointers in KTR log messages instead of assuming sizeof(int) == sizeof(void *). - Don't set p_mtxname to NULL when releasing a mutex. It doesn't hurt to leave it set (we don't clear w_mesg for example) and at least at one time in the past, there used to be race conditions in the kernel that would result in setting this to NULL causing the kernel to dereference NULL. - Make the _mtx_assert() function be compiled in if INVARIANTS_SUPPORT is defined rather than if INVARIANTS is defined so that a KLD compiled with INVARIANTS that uses mtx_assert() can be used with a kernel that just has INVARIANT_SUPPORT compiled in.
2001-02-24 19:36:13 +00:00
"_mtx_unlock_sleep: %p interrupted %p",
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
it, it->it_interrupted);
intr_thd_fixup(it);
- Make the mutex code almost completely machine independent. This greatly reducues the maintenance load for the mutex code. The only MD portions of the mutex code are in machine/mutex.h now, which include the assembly macros for handling mutexes as well as optionally overriding the mutex micro-operations. For example, we use optimized micro-ops on the x86 platform #ifndef I386_CPU. - Change the behavior of the SMP_DEBUG kernel option. In the new code, mtx_assert() only depends on INVARIANTS, allowing other kernel developers to have working mutex assertiions without having to include all of the mutex debugging code. The SMP_DEBUG kernel option has been renamed to MUTEX_DEBUG and now just controls extra mutex debugging code. - Abolish the ugly mtx_f hack. Instead, we dynamically allocate seperate mtx_debug structures on the fly in mtx_init, except for mutexes that are initiated very early in the boot process. These mutexes are declared using a special MUTEX_DECLARE() macro, and use a new flag MTX_COLD when calling mtx_init. This is still somewhat hackish, but it is less evil than the mtx_f filler struct, and the mtx struct is now the same size with and without mutex debugging code. - Add some micro-micro-operation macros for doing the actual atomic operations on the mutex mtx_lock field to make it easier for other archs to override/optimize mutex ops if needed. These new tiny ops also clean up the code in some places by replacing long atomic operation function calls that spanned 2-3 lines with a short 1-line macro call. - Don't call mi_switch() from mtx_enter_hard() when we block while trying to obtain a sleep mutex. Calling mi_switch() would bogusly release Giant before switching to the next process. Instead, inline most of the code from mi_switch() in the mtx_enter_hard() function. Note that when we finally kill Giant we can back this out and go back to calling mi_switch().
2000-10-20 07:26:37 +00:00
}
}
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
#endif
setrunqueue(p);
if ((opts & MTX_QUIET) == 0)
CTR2(KTR_LOCK,
"_mtx_unlock_sleep: %p switching out lock=%p", m,
(void *)m->mtx_lock);
mi_switch();
if ((opts & MTX_QUIET) == 0)
CTR2(KTR_LOCK, "_mtx_unlock_sleep: %p resuming lock=%p",
m, (void *)m->mtx_lock);
- Make the mutex code almost completely machine independent. This greatly reducues the maintenance load for the mutex code. The only MD portions of the mutex code are in machine/mutex.h now, which include the assembly macros for handling mutexes as well as optionally overriding the mutex micro-operations. For example, we use optimized micro-ops on the x86 platform #ifndef I386_CPU. - Change the behavior of the SMP_DEBUG kernel option. In the new code, mtx_assert() only depends on INVARIANTS, allowing other kernel developers to have working mutex assertiions without having to include all of the mutex debugging code. The SMP_DEBUG kernel option has been renamed to MUTEX_DEBUG and now just controls extra mutex debugging code. - Abolish the ugly mtx_f hack. Instead, we dynamically allocate seperate mtx_debug structures on the fly in mtx_init, except for mutexes that are initiated very early in the boot process. These mutexes are declared using a special MUTEX_DECLARE() macro, and use a new flag MTX_COLD when calling mtx_init. This is still somewhat hackish, but it is less evil than the mtx_f filler struct, and the mtx struct is now the same size with and without mutex debugging code. - Add some micro-micro-operation macros for doing the actual atomic operations on the mutex mtx_lock field to make it easier for other archs to override/optimize mutex ops if needed. These new tiny ops also clean up the code in some places by replacing long atomic operation function calls that spanned 2-3 lines with a short 1-line macro call. - Don't call mi_switch() from mtx_enter_hard() when we block while trying to obtain a sleep mutex. Calling mi_switch() would bogusly release Giant before switching to the next process. Instead, inline most of the code from mi_switch() in the mtx_enter_hard() function. Note that when we finally kill Giant we can back this out and go back to calling mi_switch().
2000-10-20 07:26:37 +00:00
}
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
mtx_unlock_spin(&sched_lock);
return;
- Make the mutex code almost completely machine independent. This greatly reducues the maintenance load for the mutex code. The only MD portions of the mutex code are in machine/mutex.h now, which include the assembly macros for handling mutexes as well as optionally overriding the mutex micro-operations. For example, we use optimized micro-ops on the x86 platform #ifndef I386_CPU. - Change the behavior of the SMP_DEBUG kernel option. In the new code, mtx_assert() only depends on INVARIANTS, allowing other kernel developers to have working mutex assertiions without having to include all of the mutex debugging code. The SMP_DEBUG kernel option has been renamed to MUTEX_DEBUG and now just controls extra mutex debugging code. - Abolish the ugly mtx_f hack. Instead, we dynamically allocate seperate mtx_debug structures on the fly in mtx_init, except for mutexes that are initiated very early in the boot process. These mutexes are declared using a special MUTEX_DECLARE() macro, and use a new flag MTX_COLD when calling mtx_init. This is still somewhat hackish, but it is less evil than the mtx_f filler struct, and the mtx struct is now the same size with and without mutex debugging code. - Add some micro-micro-operation macros for doing the actual atomic operations on the mutex mtx_lock field to make it easier for other archs to override/optimize mutex ops if needed. These new tiny ops also clean up the code in some places by replacing long atomic operation function calls that spanned 2-3 lines with a short 1-line macro call. - Don't call mi_switch() from mtx_enter_hard() when we block while trying to obtain a sleep mutex. Calling mi_switch() would bogusly release Giant before switching to the next process. Instead, inline most of the code from mi_switch() in the mtx_enter_hard() function. Note that when we finally kill Giant we can back this out and go back to calling mi_switch().
2000-10-20 07:26:37 +00:00
}
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
/*
* All the unlocking of MTX_SPIN locks is done inline.
* See the _rel_spin_lock() macro for the details.
*/
/*
- Axe RETIP() as it was very i386 specific and unwieldy. Instead, use the passed in filename and line number in the KTR tracepoint message. - Even though it is #if 0'd code, change the code to detect that a process is an interrupt thread to check p->p_ithd against NULL rather than checking non-existant process flags from BSD/OS. - Use '%p' to print pointers in KTR log messages instead of assuming sizeof(int) == sizeof(void *). - Don't set p_mtxname to NULL when releasing a mutex. It doesn't hurt to leave it set (we don't clear w_mesg for example) and at least at one time in the past, there used to be race conditions in the kernel that would result in setting this to NULL causing the kernel to dereference NULL. - Make the _mtx_assert() function be compiled in if INVARIANTS_SUPPORT is defined rather than if INVARIANTS is defined so that a KLD compiled with INVARIANTS that uses mtx_assert() can be used with a kernel that just has INVARIANT_SUPPORT compiled in.
2001-02-24 19:36:13 +00:00
* The backing function for the INVARIANTS-enabled mtx_assert()
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
*/
Grrr, s/INVARIANTS_SUPPORT/INVARIANT_SUPPORT/.
2001-02-24 21:29:32 +00:00
#ifdef INVARIANT_SUPPORT
Move most of sys/mutex.h into kern/kern_mutex.c, thereby making the mutex inline functions non-inlined. Hide parts of the mutex implementation that should not be exposed. Make sure that WITNESS code is not executed during boot until the mutexes are fully initialized by SI_SUB_MUTEX (the original motivation for this commit). Submitted by: peter
2001-01-21 22:34:43 +00:00
void
Print correct file name and line number in mtx_assert(). Noticed by: jake
2001-01-22 05:56:55 +00:00
_mtx_assert(struct mtx *m, int what, const char *file, int line)
Move most of sys/mutex.h into kern/kern_mutex.c, thereby making the mutex inline functions non-inlined. Hide parts of the mutex implementation that should not be exposed. Make sure that WITNESS code is not executed during boot until the mutexes are fully initialized by SI_SUB_MUTEX (the original motivation for this commit). Submitted by: peter
2001-01-21 22:34:43 +00:00
{
Remove brackets around variables in a function that used to be a macro.
2001-02-25 16:18:13 +00:00
switch (what) {
Move most of sys/mutex.h into kern/kern_mutex.c, thereby making the mutex inline functions non-inlined. Hide parts of the mutex implementation that should not be exposed. Make sure that WITNESS code is not executed during boot until the mutexes are fully initialized by SI_SUB_MUTEX (the original motivation for this commit). Submitted by: peter
2001-01-21 22:34:43 +00:00
case MA_OWNED:
case MA_OWNED | MA_RECURSED:
case MA_OWNED | MA_NOTRECURSED:
Remove brackets around variables in a function that used to be a macro.
2001-02-25 16:18:13 +00:00
if (!mtx_owned(m))
Move most of sys/mutex.h into kern/kern_mutex.c, thereby making the mutex inline functions non-inlined. Hide parts of the mutex implementation that should not be exposed. Make sure that WITNESS code is not executed during boot until the mutexes are fully initialized by SI_SUB_MUTEX (the original motivation for this commit). Submitted by: peter
2001-01-21 22:34:43 +00:00
panic("mutex %s not owned at %s:%d",
Remove brackets around variables in a function that used to be a macro.
2001-02-25 16:18:13 +00:00
m->mtx_description, file, line);
if (mtx_recursed(m)) {
if ((what & MA_NOTRECURSED) != 0)
Move most of sys/mutex.h into kern/kern_mutex.c, thereby making the mutex inline functions non-inlined. Hide parts of the mutex implementation that should not be exposed. Make sure that WITNESS code is not executed during boot until the mutexes are fully initialized by SI_SUB_MUTEX (the original motivation for this commit). Submitted by: peter
2001-01-21 22:34:43 +00:00
panic("mutex %s recursed at %s:%d",
Remove brackets around variables in a function that used to be a macro.
2001-02-25 16:18:13 +00:00
m->mtx_description, file, line);
} else if ((what & MA_RECURSED) != 0) {
Move most of sys/mutex.h into kern/kern_mutex.c, thereby making the mutex inline functions non-inlined. Hide parts of the mutex implementation that should not be exposed. Make sure that WITNESS code is not executed during boot until the mutexes are fully initialized by SI_SUB_MUTEX (the original motivation for this commit). Submitted by: peter
2001-01-21 22:34:43 +00:00
panic("mutex %s unrecursed at %s:%d",
Remove brackets around variables in a function that used to be a macro.
2001-02-25 16:18:13 +00:00
m->mtx_description, file, line);
Move most of sys/mutex.h into kern/kern_mutex.c, thereby making the mutex inline functions non-inlined. Hide parts of the mutex implementation that should not be exposed. Make sure that WITNESS code is not executed during boot until the mutexes are fully initialized by SI_SUB_MUTEX (the original motivation for this commit). Submitted by: peter
2001-01-21 22:34:43 +00:00
}
break;
case MA_NOTOWNED:
Remove brackets around variables in a function that used to be a macro.
2001-02-25 16:18:13 +00:00
if (mtx_owned(m))
Move most of sys/mutex.h into kern/kern_mutex.c, thereby making the mutex inline functions non-inlined. Hide parts of the mutex implementation that should not be exposed. Make sure that WITNESS code is not executed during boot until the mutexes are fully initialized by SI_SUB_MUTEX (the original motivation for this commit). Submitted by: peter
2001-01-21 22:34:43 +00:00
panic("mutex %s owned at %s:%d",
Remove brackets around variables in a function that used to be a macro.
2001-02-25 16:18:13 +00:00
m->mtx_description, file, line);
Move most of sys/mutex.h into kern/kern_mutex.c, thereby making the mutex inline functions non-inlined. Hide parts of the mutex implementation that should not be exposed. Make sure that WITNESS code is not executed during boot until the mutexes are fully initialized by SI_SUB_MUTEX (the original motivation for this commit). Submitted by: peter
2001-01-21 22:34:43 +00:00
break;
default:
Print correct file name and line number in mtx_assert(). Noticed by: jake
2001-01-22 05:56:55 +00:00
panic("unknown mtx_assert at %s:%d", file, line);
Move most of sys/mutex.h into kern/kern_mutex.c, thereby making the mutex inline functions non-inlined. Hide parts of the mutex implementation that should not be exposed. Make sure that WITNESS code is not executed during boot until the mutexes are fully initialized by SI_SUB_MUTEX (the original motivation for this commit). Submitted by: peter
2001-01-21 22:34:43 +00:00
}
}
#endif
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
/*
* The MUTEX_DEBUG-enabled mtx_validate()
*/
- Make the mutex code almost completely machine independent. This greatly reducues the maintenance load for the mutex code. The only MD portions of the mutex code are in machine/mutex.h now, which include the assembly macros for handling mutexes as well as optionally overriding the mutex micro-operations. For example, we use optimized micro-ops on the x86 platform #ifndef I386_CPU. - Change the behavior of the SMP_DEBUG kernel option. In the new code, mtx_assert() only depends on INVARIANTS, allowing other kernel developers to have working mutex assertiions without having to include all of the mutex debugging code. The SMP_DEBUG kernel option has been renamed to MUTEX_DEBUG and now just controls extra mutex debugging code. - Abolish the ugly mtx_f hack. Instead, we dynamically allocate seperate mtx_debug structures on the fly in mtx_init, except for mutexes that are initiated very early in the boot process. These mutexes are declared using a special MUTEX_DECLARE() macro, and use a new flag MTX_COLD when calling mtx_init. This is still somewhat hackish, but it is less evil than the mtx_f filler struct, and the mtx struct is now the same size with and without mutex debugging code. - Add some micro-micro-operation macros for doing the actual atomic operations on the mutex mtx_lock field to make it easier for other archs to override/optimize mutex ops if needed. These new tiny ops also clean up the code in some places by replacing long atomic operation function calls that spanned 2-3 lines with a short 1-line macro call. - Don't call mi_switch() from mtx_enter_hard() when we block while trying to obtain a sleep mutex. Calling mi_switch() would bogusly release Giant before switching to the next process. Instead, inline most of the code from mi_switch() in the mtx_enter_hard() function. Note that when we finally kill Giant we can back this out and go back to calling mi_switch().
2000-10-20 07:26:37 +00:00
#define MV_DESTROY 0 /* validate before destory */
#define MV_INIT 1 /* validate before init */
#ifdef MUTEX_DEBUG
int mtx_validate __P((struct mtx *, int));
int
mtx_validate(struct mtx *m, int when)
{
struct mtx *mp;
int i;
int retval = 0;
Remove MUTEX_DECLARE() and MTX_COLD. Instead, postpone full mutex initialization until after malloc() is safe to call, then iterate through all mutexes and complete their initialization. This change is necessary in order to avoid some circular bootstrapping dependencies.
2001-01-21 07:52:20 +00:00
#ifdef WITNESS
if (witness_cold)
return 0;
#endif
- Make the mutex code almost completely machine independent. This greatly reducues the maintenance load for the mutex code. The only MD portions of the mutex code are in machine/mutex.h now, which include the assembly macros for handling mutexes as well as optionally overriding the mutex micro-operations. For example, we use optimized micro-ops on the x86 platform #ifndef I386_CPU. - Change the behavior of the SMP_DEBUG kernel option. In the new code, mtx_assert() only depends on INVARIANTS, allowing other kernel developers to have working mutex assertiions without having to include all of the mutex debugging code. The SMP_DEBUG kernel option has been renamed to MUTEX_DEBUG and now just controls extra mutex debugging code. - Abolish the ugly mtx_f hack. Instead, we dynamically allocate seperate mtx_debug structures on the fly in mtx_init, except for mutexes that are initiated very early in the boot process. These mutexes are declared using a special MUTEX_DECLARE() macro, and use a new flag MTX_COLD when calling mtx_init. This is still somewhat hackish, but it is less evil than the mtx_f filler struct, and the mtx struct is now the same size with and without mutex debugging code. - Add some micro-micro-operation macros for doing the actual atomic operations on the mutex mtx_lock field to make it easier for other archs to override/optimize mutex ops if needed. These new tiny ops also clean up the code in some places by replacing long atomic operation function calls that spanned 2-3 lines with a short 1-line macro call. - Don't call mi_switch() from mtx_enter_hard() when we block while trying to obtain a sleep mutex. Calling mi_switch() would bogusly release Giant before switching to the next process. Instead, inline most of the code from mi_switch() in the mtx_enter_hard() function. Note that when we finally kill Giant we can back this out and go back to calling mi_switch().
2000-10-20 07:26:37 +00:00
if (m == &all_mtx || cold)
return 0;
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
mtx_lock(&all_mtx);
- Make the mutex code almost completely machine independent. This greatly reducues the maintenance load for the mutex code. The only MD portions of the mutex code are in machine/mutex.h now, which include the assembly macros for handling mutexes as well as optionally overriding the mutex micro-operations. For example, we use optimized micro-ops on the x86 platform #ifndef I386_CPU. - Change the behavior of the SMP_DEBUG kernel option. In the new code, mtx_assert() only depends on INVARIANTS, allowing other kernel developers to have working mutex assertiions without having to include all of the mutex debugging code. The SMP_DEBUG kernel option has been renamed to MUTEX_DEBUG and now just controls extra mutex debugging code. - Abolish the ugly mtx_f hack. Instead, we dynamically allocate seperate mtx_debug structures on the fly in mtx_init, except for mutexes that are initiated very early in the boot process. These mutexes are declared using a special MUTEX_DECLARE() macro, and use a new flag MTX_COLD when calling mtx_init. This is still somewhat hackish, but it is less evil than the mtx_f filler struct, and the mtx struct is now the same size with and without mutex debugging code. - Add some micro-micro-operation macros for doing the actual atomic operations on the mutex mtx_lock field to make it easier for other archs to override/optimize mutex ops if needed. These new tiny ops also clean up the code in some places by replacing long atomic operation function calls that spanned 2-3 lines with a short 1-line macro call. - Don't call mi_switch() from mtx_enter_hard() when we block while trying to obtain a sleep mutex. Calling mi_switch() would bogusly release Giant before switching to the next process. Instead, inline most of the code from mi_switch() in the mtx_enter_hard() function. Note that when we finally kill Giant we can back this out and go back to calling mi_switch().
2000-10-20 07:26:37 +00:00
/*
* XXX - When kernacc() is fixed on the alpha to handle K0_SEG memory properly
* we can re-enable the kernacc() checks.
*/
#ifndef __alpha__
MPASS(kernacc((caddr_t)all_mtx.mtx_next, sizeof(uintptr_t),
VM_PROT_READ) == 1);
#endif
MPASS(all_mtx.mtx_next->mtx_prev == &all_mtx);
for (i = 0, mp = all_mtx.mtx_next; mp != &all_mtx; mp = mp->mtx_next) {
#ifndef __alpha__
if (kernacc((caddr_t)mp->mtx_next, sizeof(uintptr_t),
VM_PROT_READ) != 1) {
panic("mtx_validate: mp=%p mp->mtx_next=%p",
mp, mp->mtx_next);
}
#endif
i++;
if (i > mtx_cur_cnt) {
panic("mtx_validate: too many in chain, known=%d\n",
mtx_cur_cnt);
}
}
MPASS(i == mtx_cur_cnt);
switch (when) {
case MV_DESTROY:
for (mp = all_mtx.mtx_next; mp != &all_mtx; mp = mp->mtx_next)
if (mp == m)
break;
MPASS(mp == m);
break;
case MV_INIT:
for (mp = all_mtx.mtx_next; mp != &all_mtx; mp = mp->mtx_next)
if (mp == m) {
/*
* Not good. This mutex already exists.
*/
printf("re-initing existing mutex %s\n",
m->mtx_description);
MPASS(m->mtx_lock == MTX_UNOWNED);
retval = 1;
}
}
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
mtx_unlock(&all_mtx);
- Make the mutex code almost completely machine independent. This greatly reducues the maintenance load for the mutex code. The only MD portions of the mutex code are in machine/mutex.h now, which include the assembly macros for handling mutexes as well as optionally overriding the mutex micro-operations. For example, we use optimized micro-ops on the x86 platform #ifndef I386_CPU. - Change the behavior of the SMP_DEBUG kernel option. In the new code, mtx_assert() only depends on INVARIANTS, allowing other kernel developers to have working mutex assertiions without having to include all of the mutex debugging code. The SMP_DEBUG kernel option has been renamed to MUTEX_DEBUG and now just controls extra mutex debugging code. - Abolish the ugly mtx_f hack. Instead, we dynamically allocate seperate mtx_debug structures on the fly in mtx_init, except for mutexes that are initiated very early in the boot process. These mutexes are declared using a special MUTEX_DECLARE() macro, and use a new flag MTX_COLD when calling mtx_init. This is still somewhat hackish, but it is less evil than the mtx_f filler struct, and the mtx struct is now the same size with and without mutex debugging code. - Add some micro-micro-operation macros for doing the actual atomic operations on the mutex mtx_lock field to make it easier for other archs to override/optimize mutex ops if needed. These new tiny ops also clean up the code in some places by replacing long atomic operation function calls that spanned 2-3 lines with a short 1-line macro call. - Don't call mi_switch() from mtx_enter_hard() when we block while trying to obtain a sleep mutex. Calling mi_switch() would bogusly release Giant before switching to the next process. Instead, inline most of the code from mi_switch() in the mtx_enter_hard() function. Note that when we finally kill Giant we can back this out and go back to calling mi_switch().
2000-10-20 07:26:37 +00:00
return (retval);
}
#endif
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
/*
* Mutex initialization routine; initialize lock `m' of type contained in
* `opts' with options contained in `opts' and description `description.'
* Place on "all_mtx" queue.
*/
- Make the mutex code almost completely machine independent. This greatly reducues the maintenance load for the mutex code. The only MD portions of the mutex code are in machine/mutex.h now, which include the assembly macros for handling mutexes as well as optionally overriding the mutex micro-operations. For example, we use optimized micro-ops on the x86 platform #ifndef I386_CPU. - Change the behavior of the SMP_DEBUG kernel option. In the new code, mtx_assert() only depends on INVARIANTS, allowing other kernel developers to have working mutex assertiions without having to include all of the mutex debugging code. The SMP_DEBUG kernel option has been renamed to MUTEX_DEBUG and now just controls extra mutex debugging code. - Abolish the ugly mtx_f hack. Instead, we dynamically allocate seperate mtx_debug structures on the fly in mtx_init, except for mutexes that are initiated very early in the boot process. These mutexes are declared using a special MUTEX_DECLARE() macro, and use a new flag MTX_COLD when calling mtx_init. This is still somewhat hackish, but it is less evil than the mtx_f filler struct, and the mtx struct is now the same size with and without mutex debugging code. - Add some micro-micro-operation macros for doing the actual atomic operations on the mutex mtx_lock field to make it easier for other archs to override/optimize mutex ops if needed. These new tiny ops also clean up the code in some places by replacing long atomic operation function calls that spanned 2-3 lines with a short 1-line macro call. - Don't call mi_switch() from mtx_enter_hard() when we block while trying to obtain a sleep mutex. Calling mi_switch() would bogusly release Giant before switching to the next process. Instead, inline most of the code from mi_switch() in the mtx_enter_hard() function. Note that when we finally kill Giant we can back this out and go back to calling mi_switch().
2000-10-20 07:26:37 +00:00
void
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
mtx_init(struct mtx *m, const char *description, int opts)
- Make the mutex code almost completely machine independent. This greatly reducues the maintenance load for the mutex code. The only MD portions of the mutex code are in machine/mutex.h now, which include the assembly macros for handling mutexes as well as optionally overriding the mutex micro-operations. For example, we use optimized micro-ops on the x86 platform #ifndef I386_CPU. - Change the behavior of the SMP_DEBUG kernel option. In the new code, mtx_assert() only depends on INVARIANTS, allowing other kernel developers to have working mutex assertiions without having to include all of the mutex debugging code. The SMP_DEBUG kernel option has been renamed to MUTEX_DEBUG and now just controls extra mutex debugging code. - Abolish the ugly mtx_f hack. Instead, we dynamically allocate seperate mtx_debug structures on the fly in mtx_init, except for mutexes that are initiated very early in the boot process. These mutexes are declared using a special MUTEX_DECLARE() macro, and use a new flag MTX_COLD when calling mtx_init. This is still somewhat hackish, but it is less evil than the mtx_f filler struct, and the mtx struct is now the same size with and without mutex debugging code. - Add some micro-micro-operation macros for doing the actual atomic operations on the mutex mtx_lock field to make it easier for other archs to override/optimize mutex ops if needed. These new tiny ops also clean up the code in some places by replacing long atomic operation function calls that spanned 2-3 lines with a short 1-line macro call. - Don't call mi_switch() from mtx_enter_hard() when we block while trying to obtain a sleep mutex. Calling mi_switch() would bogusly release Giant before switching to the next process. Instead, inline most of the code from mi_switch() in the mtx_enter_hard() function. Note that when we finally kill Giant we can back this out and go back to calling mi_switch().
2000-10-20 07:26:37 +00:00
{
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
if ((opts & MTX_QUIET) == 0)
CTR2(KTR_LOCK, "mtx_init %p (%s)", m, description);
- Make the mutex code almost completely machine independent. This greatly reducues the maintenance load for the mutex code. The only MD portions of the mutex code are in machine/mutex.h now, which include the assembly macros for handling mutexes as well as optionally overriding the mutex micro-operations. For example, we use optimized micro-ops on the x86 platform #ifndef I386_CPU. - Change the behavior of the SMP_DEBUG kernel option. In the new code, mtx_assert() only depends on INVARIANTS, allowing other kernel developers to have working mutex assertiions without having to include all of the mutex debugging code. The SMP_DEBUG kernel option has been renamed to MUTEX_DEBUG and now just controls extra mutex debugging code. - Abolish the ugly mtx_f hack. Instead, we dynamically allocate seperate mtx_debug structures on the fly in mtx_init, except for mutexes that are initiated very early in the boot process. These mutexes are declared using a special MUTEX_DECLARE() macro, and use a new flag MTX_COLD when calling mtx_init. This is still somewhat hackish, but it is less evil than the mtx_f filler struct, and the mtx struct is now the same size with and without mutex debugging code. - Add some micro-micro-operation macros for doing the actual atomic operations on the mutex mtx_lock field to make it easier for other archs to override/optimize mutex ops if needed. These new tiny ops also clean up the code in some places by replacing long atomic operation function calls that spanned 2-3 lines with a short 1-line macro call. - Don't call mi_switch() from mtx_enter_hard() when we block while trying to obtain a sleep mutex. Calling mi_switch() would bogusly release Giant before switching to the next process. Instead, inline most of the code from mi_switch() in the mtx_enter_hard() function. Note that when we finally kill Giant we can back this out and go back to calling mi_switch().
2000-10-20 07:26:37 +00:00
#ifdef MUTEX_DEBUG
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
/* Diagnostic and error correction */
if (mtx_validate(m, MV_INIT))
- Make the mutex code almost completely machine independent. This greatly reducues the maintenance load for the mutex code. The only MD portions of the mutex code are in machine/mutex.h now, which include the assembly macros for handling mutexes as well as optionally overriding the mutex micro-operations. For example, we use optimized micro-ops on the x86 platform #ifndef I386_CPU. - Change the behavior of the SMP_DEBUG kernel option. In the new code, mtx_assert() only depends on INVARIANTS, allowing other kernel developers to have working mutex assertiions without having to include all of the mutex debugging code. The SMP_DEBUG kernel option has been renamed to MUTEX_DEBUG and now just controls extra mutex debugging code. - Abolish the ugly mtx_f hack. Instead, we dynamically allocate seperate mtx_debug structures on the fly in mtx_init, except for mutexes that are initiated very early in the boot process. These mutexes are declared using a special MUTEX_DECLARE() macro, and use a new flag MTX_COLD when calling mtx_init. This is still somewhat hackish, but it is less evil than the mtx_f filler struct, and the mtx struct is now the same size with and without mutex debugging code. - Add some micro-micro-operation macros for doing the actual atomic operations on the mutex mtx_lock field to make it easier for other archs to override/optimize mutex ops if needed. These new tiny ops also clean up the code in some places by replacing long atomic operation function calls that spanned 2-3 lines with a short 1-line macro call. - Don't call mi_switch() from mtx_enter_hard() when we block while trying to obtain a sleep mutex. Calling mi_switch() would bogusly release Giant before switching to the next process. Instead, inline most of the code from mi_switch() in the mtx_enter_hard() function. Note that when we finally kill Giant we can back this out and go back to calling mi_switch().
2000-10-20 07:26:37 +00:00
return;
Split the WITNESS and MUTEX_DEBUG options apart so that WITNESS does not depend on MUTEX_DEBUG. The MUTEX_DEBUG option turns on extra assertions and checks to verify that mutexes themselves are implemented properly. The WITNESS option uses extra checks and diagnostics to verify that other code is using mutexes properly.
2000-12-01 00:10:59 +00:00
#endif
- Make the mutex code almost completely machine independent. This greatly reducues the maintenance load for the mutex code. The only MD portions of the mutex code are in machine/mutex.h now, which include the assembly macros for handling mutexes as well as optionally overriding the mutex micro-operations. For example, we use optimized micro-ops on the x86 platform #ifndef I386_CPU. - Change the behavior of the SMP_DEBUG kernel option. In the new code, mtx_assert() only depends on INVARIANTS, allowing other kernel developers to have working mutex assertiions without having to include all of the mutex debugging code. The SMP_DEBUG kernel option has been renamed to MUTEX_DEBUG and now just controls extra mutex debugging code. - Abolish the ugly mtx_f hack. Instead, we dynamically allocate seperate mtx_debug structures on the fly in mtx_init, except for mutexes that are initiated very early in the boot process. These mutexes are declared using a special MUTEX_DECLARE() macro, and use a new flag MTX_COLD when calling mtx_init. This is still somewhat hackish, but it is less evil than the mtx_f filler struct, and the mtx struct is now the same size with and without mutex debugging code. - Add some micro-micro-operation macros for doing the actual atomic operations on the mutex mtx_lock field to make it easier for other archs to override/optimize mutex ops if needed. These new tiny ops also clean up the code in some places by replacing long atomic operation function calls that spanned 2-3 lines with a short 1-line macro call. - Don't call mi_switch() from mtx_enter_hard() when we block while trying to obtain a sleep mutex. Calling mi_switch() would bogusly release Giant before switching to the next process. Instead, inline most of the code from mi_switch() in the mtx_enter_hard() function. Note that when we finally kill Giant we can back this out and go back to calling mi_switch().
2000-10-20 07:26:37 +00:00
bzero((void *)m, sizeof *m);
TAILQ_INIT(&m->mtx_blocked);
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
Split the WITNESS and MUTEX_DEBUG options apart so that WITNESS does not depend on MUTEX_DEBUG. The MUTEX_DEBUG option turns on extra assertions and checks to verify that mutexes themselves are implemented properly. The WITNESS option uses extra checks and diagnostics to verify that other code is using mutexes properly.
2000-12-01 00:10:59 +00:00
#ifdef WITNESS
Remove MUTEX_DECLARE() and MTX_COLD. Instead, postpone full mutex initialization until after malloc() is safe to call, then iterate through all mutexes and complete their initialization. This change is necessary in order to avoid some circular bootstrapping dependencies.
2001-01-21 07:52:20 +00:00
if (!witness_cold) {
- Don't use a union and fun tricks to shave one extra pointer off of struct mtx right now as it makes debugging harder. When we are in optimizing mode, we can revisit this. - Fix the KTR trace messages to use %p rather than 0x%p to avoid duplicate 0x's in KTR output. - During witness_fixup, release Giant so that witness doesn't get confused. Also, grab all_mtx while walking the list of mutexes. - Remove w_sleep and w_recurse. Instead, perform checks on mutexes using the mutex's mtx_flags field. - Allow debug.witness_ddb and debug.witness_skipspin to be set from the loader. - Add Giant to the front of existing order_list entries to help ensure Giant is always first. - Add an order entry for the various proc locks. Note that this only helps keep proc in order mostly as the allproc and proctree mutexes are only obtained during a lockmgr operation on the specified mutex.
2001-01-24 10:57:01 +00:00
m->mtx_debug = malloc(sizeof(struct mtx_debug),
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
M_WITNESS, M_NOWAIT | M_ZERO);
- Don't use a union and fun tricks to shave one extra pointer off of struct mtx right now as it makes debugging harder. When we are in optimizing mode, we can revisit this. - Fix the KTR trace messages to use %p rather than 0x%p to avoid duplicate 0x's in KTR output. - During witness_fixup, release Giant so that witness doesn't get confused. Also, grab all_mtx while walking the list of mutexes. - Remove w_sleep and w_recurse. Instead, perform checks on mutexes using the mutex's mtx_flags field. - Allow debug.witness_ddb and debug.witness_skipspin to be set from the loader. - Add Giant to the front of existing order_list entries to help ensure Giant is always first. - Add an order entry for the various proc locks. Note that this only helps keep proc in order mostly as the allproc and proctree mutexes are only obtained during a lockmgr operation on the specified mutex.
2001-01-24 10:57:01 +00:00
MPASS(m->mtx_debug != NULL);
Remove MUTEX_DECLARE() and MTX_COLD. Instead, postpone full mutex initialization until after malloc() is safe to call, then iterate through all mutexes and complete their initialization. This change is necessary in order to avoid some circular bootstrapping dependencies.
2001-01-21 07:52:20 +00:00
}
#endif
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
m->mtx_description = description;
m->mtx_flags = opts;
- Make the mutex code almost completely machine independent. This greatly reducues the maintenance load for the mutex code. The only MD portions of the mutex code are in machine/mutex.h now, which include the assembly macros for handling mutexes as well as optionally overriding the mutex micro-operations. For example, we use optimized micro-ops on the x86 platform #ifndef I386_CPU. - Change the behavior of the SMP_DEBUG kernel option. In the new code, mtx_assert() only depends on INVARIANTS, allowing other kernel developers to have working mutex assertiions without having to include all of the mutex debugging code. The SMP_DEBUG kernel option has been renamed to MUTEX_DEBUG and now just controls extra mutex debugging code. - Abolish the ugly mtx_f hack. Instead, we dynamically allocate seperate mtx_debug structures on the fly in mtx_init, except for mutexes that are initiated very early in the boot process. These mutexes are declared using a special MUTEX_DECLARE() macro, and use a new flag MTX_COLD when calling mtx_init. This is still somewhat hackish, but it is less evil than the mtx_f filler struct, and the mtx struct is now the same size with and without mutex debugging code. - Add some micro-micro-operation macros for doing the actual atomic operations on the mutex mtx_lock field to make it easier for other archs to override/optimize mutex ops if needed. These new tiny ops also clean up the code in some places by replacing long atomic operation function calls that spanned 2-3 lines with a short 1-line macro call. - Don't call mi_switch() from mtx_enter_hard() when we block while trying to obtain a sleep mutex. Calling mi_switch() would bogusly release Giant before switching to the next process. Instead, inline most of the code from mi_switch() in the mtx_enter_hard() function. Note that when we finally kill Giant we can back this out and go back to calling mi_switch().
2000-10-20 07:26:37 +00:00
m->mtx_lock = MTX_UNOWNED;
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
- Make the mutex code almost completely machine independent. This greatly reducues the maintenance load for the mutex code. The only MD portions of the mutex code are in machine/mutex.h now, which include the assembly macros for handling mutexes as well as optionally overriding the mutex micro-operations. For example, we use optimized micro-ops on the x86 platform #ifndef I386_CPU. - Change the behavior of the SMP_DEBUG kernel option. In the new code, mtx_assert() only depends on INVARIANTS, allowing other kernel developers to have working mutex assertiions without having to include all of the mutex debugging code. The SMP_DEBUG kernel option has been renamed to MUTEX_DEBUG and now just controls extra mutex debugging code. - Abolish the ugly mtx_f hack. Instead, we dynamically allocate seperate mtx_debug structures on the fly in mtx_init, except for mutexes that are initiated very early in the boot process. These mutexes are declared using a special MUTEX_DECLARE() macro, and use a new flag MTX_COLD when calling mtx_init. This is still somewhat hackish, but it is less evil than the mtx_f filler struct, and the mtx struct is now the same size with and without mutex debugging code. - Add some micro-micro-operation macros for doing the actual atomic operations on the mutex mtx_lock field to make it easier for other archs to override/optimize mutex ops if needed. These new tiny ops also clean up the code in some places by replacing long atomic operation function calls that spanned 2-3 lines with a short 1-line macro call. - Don't call mi_switch() from mtx_enter_hard() when we block while trying to obtain a sleep mutex. Calling mi_switch() would bogusly release Giant before switching to the next process. Instead, inline most of the code from mi_switch() in the mtx_enter_hard() function. Note that when we finally kill Giant we can back this out and go back to calling mi_switch().
2000-10-20 07:26:37 +00:00
/* Put on all mutex queue */
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
mtx_lock(&all_mtx);
- Make the mutex code almost completely machine independent. This greatly reducues the maintenance load for the mutex code. The only MD portions of the mutex code are in machine/mutex.h now, which include the assembly macros for handling mutexes as well as optionally overriding the mutex micro-operations. For example, we use optimized micro-ops on the x86 platform #ifndef I386_CPU. - Change the behavior of the SMP_DEBUG kernel option. In the new code, mtx_assert() only depends on INVARIANTS, allowing other kernel developers to have working mutex assertiions without having to include all of the mutex debugging code. The SMP_DEBUG kernel option has been renamed to MUTEX_DEBUG and now just controls extra mutex debugging code. - Abolish the ugly mtx_f hack. Instead, we dynamically allocate seperate mtx_debug structures on the fly in mtx_init, except for mutexes that are initiated very early in the boot process. These mutexes are declared using a special MUTEX_DECLARE() macro, and use a new flag MTX_COLD when calling mtx_init. This is still somewhat hackish, but it is less evil than the mtx_f filler struct, and the mtx struct is now the same size with and without mutex debugging code. - Add some micro-micro-operation macros for doing the actual atomic operations on the mutex mtx_lock field to make it easier for other archs to override/optimize mutex ops if needed. These new tiny ops also clean up the code in some places by replacing long atomic operation function calls that spanned 2-3 lines with a short 1-line macro call. - Don't call mi_switch() from mtx_enter_hard() when we block while trying to obtain a sleep mutex. Calling mi_switch() would bogusly release Giant before switching to the next process. Instead, inline most of the code from mi_switch() in the mtx_enter_hard() function. Note that when we finally kill Giant we can back this out and go back to calling mi_switch().
2000-10-20 07:26:37 +00:00
m->mtx_next = &all_mtx;
m->mtx_prev = all_mtx.mtx_prev;
m->mtx_prev->mtx_next = m;
all_mtx.mtx_prev = m;
if (++mtx_cur_cnt > mtx_max_cnt)
mtx_max_cnt = mtx_cur_cnt;
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
mtx_unlock(&all_mtx);
Remove MUTEX_DECLARE() and MTX_COLD. Instead, postpone full mutex initialization until after malloc() is safe to call, then iterate through all mutexes and complete their initialization. This change is necessary in order to avoid some circular bootstrapping dependencies.
2001-01-21 07:52:20 +00:00
#ifdef WITNESS
if (!witness_cold)
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
witness_init(m, opts);
Remove MUTEX_DECLARE() and MTX_COLD. Instead, postpone full mutex initialization until after malloc() is safe to call, then iterate through all mutexes and complete their initialization. This change is necessary in order to avoid some circular bootstrapping dependencies.
2001-01-21 07:52:20 +00:00
#endif
- Make the mutex code almost completely machine independent. This greatly reducues the maintenance load for the mutex code. The only MD portions of the mutex code are in machine/mutex.h now, which include the assembly macros for handling mutexes as well as optionally overriding the mutex micro-operations. For example, we use optimized micro-ops on the x86 platform #ifndef I386_CPU. - Change the behavior of the SMP_DEBUG kernel option. In the new code, mtx_assert() only depends on INVARIANTS, allowing other kernel developers to have working mutex assertiions without having to include all of the mutex debugging code. The SMP_DEBUG kernel option has been renamed to MUTEX_DEBUG and now just controls extra mutex debugging code. - Abolish the ugly mtx_f hack. Instead, we dynamically allocate seperate mtx_debug structures on the fly in mtx_init, except for mutexes that are initiated very early in the boot process. These mutexes are declared using a special MUTEX_DECLARE() macro, and use a new flag MTX_COLD when calling mtx_init. This is still somewhat hackish, but it is less evil than the mtx_f filler struct, and the mtx struct is now the same size with and without mutex debugging code. - Add some micro-micro-operation macros for doing the actual atomic operations on the mutex mtx_lock field to make it easier for other archs to override/optimize mutex ops if needed. These new tiny ops also clean up the code in some places by replacing long atomic operation function calls that spanned 2-3 lines with a short 1-line macro call. - Don't call mi_switch() from mtx_enter_hard() when we block while trying to obtain a sleep mutex. Calling mi_switch() would bogusly release Giant before switching to the next process. Instead, inline most of the code from mi_switch() in the mtx_enter_hard() function. Note that when we finally kill Giant we can back this out and go back to calling mi_switch().
2000-10-20 07:26:37 +00:00
}
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
/*
* Remove lock `m' from all_mtx queue.
*/
- Make the mutex code almost completely machine independent. This greatly reducues the maintenance load for the mutex code. The only MD portions of the mutex code are in machine/mutex.h now, which include the assembly macros for handling mutexes as well as optionally overriding the mutex micro-operations. For example, we use optimized micro-ops on the x86 platform #ifndef I386_CPU. - Change the behavior of the SMP_DEBUG kernel option. In the new code, mtx_assert() only depends on INVARIANTS, allowing other kernel developers to have working mutex assertiions without having to include all of the mutex debugging code. The SMP_DEBUG kernel option has been renamed to MUTEX_DEBUG and now just controls extra mutex debugging code. - Abolish the ugly mtx_f hack. Instead, we dynamically allocate seperate mtx_debug structures on the fly in mtx_init, except for mutexes that are initiated very early in the boot process. These mutexes are declared using a special MUTEX_DECLARE() macro, and use a new flag MTX_COLD when calling mtx_init. This is still somewhat hackish, but it is less evil than the mtx_f filler struct, and the mtx struct is now the same size with and without mutex debugging code. - Add some micro-micro-operation macros for doing the actual atomic operations on the mutex mtx_lock field to make it easier for other archs to override/optimize mutex ops if needed. These new tiny ops also clean up the code in some places by replacing long atomic operation function calls that spanned 2-3 lines with a short 1-line macro call. - Don't call mi_switch() from mtx_enter_hard() when we block while trying to obtain a sleep mutex. Calling mi_switch() would bogusly release Giant before switching to the next process. Instead, inline most of the code from mi_switch() in the mtx_enter_hard() function. Note that when we finally kill Giant we can back this out and go back to calling mi_switch().
2000-10-20 07:26:37 +00:00
void
mtx_destroy(struct mtx *m)
{
Remove MUTEX_DECLARE() and MTX_COLD. Instead, postpone full mutex initialization until after malloc() is safe to call, then iterate through all mutexes and complete their initialization. This change is necessary in order to avoid some circular bootstrapping dependencies.
2001-01-21 07:52:20 +00:00
#ifdef WITNESS
KASSERT(!witness_cold, ("%s: Cannot destroy while still cold\n",
__FUNCTION__));
#endif
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
- Don't use a union and fun tricks to shave one extra pointer off of struct mtx right now as it makes debugging harder. When we are in optimizing mode, we can revisit this. - Fix the KTR trace messages to use %p rather than 0x%p to avoid duplicate 0x's in KTR output. - During witness_fixup, release Giant so that witness doesn't get confused. Also, grab all_mtx while walking the list of mutexes. - Remove w_sleep and w_recurse. Instead, perform checks on mutexes using the mutex's mtx_flags field. - Allow debug.witness_ddb and debug.witness_skipspin to be set from the loader. - Add Giant to the front of existing order_list entries to help ensure Giant is always first. - Add an order entry for the various proc locks. Note that this only helps keep proc in order mostly as the allproc and proctree mutexes are only obtained during a lockmgr operation on the specified mutex.
2001-01-24 10:57:01 +00:00
CTR2(KTR_LOCK, "mtx_destroy %p (%s)", m, m->mtx_description);
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
- Make the mutex code almost completely machine independent. This greatly reducues the maintenance load for the mutex code. The only MD portions of the mutex code are in machine/mutex.h now, which include the assembly macros for handling mutexes as well as optionally overriding the mutex micro-operations. For example, we use optimized micro-ops on the x86 platform #ifndef I386_CPU. - Change the behavior of the SMP_DEBUG kernel option. In the new code, mtx_assert() only depends on INVARIANTS, allowing other kernel developers to have working mutex assertiions without having to include all of the mutex debugging code. The SMP_DEBUG kernel option has been renamed to MUTEX_DEBUG and now just controls extra mutex debugging code. - Abolish the ugly mtx_f hack. Instead, we dynamically allocate seperate mtx_debug structures on the fly in mtx_init, except for mutexes that are initiated very early in the boot process. These mutexes are declared using a special MUTEX_DECLARE() macro, and use a new flag MTX_COLD when calling mtx_init. This is still somewhat hackish, but it is less evil than the mtx_f filler struct, and the mtx struct is now the same size with and without mutex debugging code. - Add some micro-micro-operation macros for doing the actual atomic operations on the mutex mtx_lock field to make it easier for other archs to override/optimize mutex ops if needed. These new tiny ops also clean up the code in some places by replacing long atomic operation function calls that spanned 2-3 lines with a short 1-line macro call. - Don't call mi_switch() from mtx_enter_hard() when we block while trying to obtain a sleep mutex. Calling mi_switch() would bogusly release Giant before switching to the next process. Instead, inline most of the code from mi_switch() in the mtx_enter_hard() function. Note that when we finally kill Giant we can back this out and go back to calling mi_switch().
2000-10-20 07:26:37 +00:00
#ifdef MUTEX_DEBUG
if (m->mtx_next == NULL)
panic("mtx_destroy: %p (%s) already destroyed",
m, m->mtx_description);
if (!mtx_owned(m)) {
MPASS(m->mtx_lock == MTX_UNOWNED);
} else {
Implement MTX_RECURSE flag for mtx_init(). All calls to mtx_init() for mutexes that recurse must now include the MTX_RECURSE bit in the flag argument variable. This change is in preparation for an upcoming (further) mutex API cleanup. The witness code will call panic() if a lock is found to recurse but the MTX_RECURSE bit was not set during the lock's initialization. The old MTX_RECURSE "state" bit (in mtx_lock) has been renamed to MTX_RECURSED, which is more appropriate given its meaning. The following locks have been made "recursive," thus far: eventhandler, Giant, callout, sched_lock, possibly some others declared in the architecture-specific code, all of the network card driver locks in pci/, as well as some other locks in dev/ stuff that I've found to be recursive. Reviewed by: jhb
2001-01-19 01:59:14 +00:00
MPASS((m->mtx_lock & (MTX_RECURSED|MTX_CONTESTED)) == 0);
- Make the mutex code almost completely machine independent. This greatly reducues the maintenance load for the mutex code. The only MD portions of the mutex code are in machine/mutex.h now, which include the assembly macros for handling mutexes as well as optionally overriding the mutex micro-operations. For example, we use optimized micro-ops on the x86 platform #ifndef I386_CPU. - Change the behavior of the SMP_DEBUG kernel option. In the new code, mtx_assert() only depends on INVARIANTS, allowing other kernel developers to have working mutex assertiions without having to include all of the mutex debugging code. The SMP_DEBUG kernel option has been renamed to MUTEX_DEBUG and now just controls extra mutex debugging code. - Abolish the ugly mtx_f hack. Instead, we dynamically allocate seperate mtx_debug structures on the fly in mtx_init, except for mutexes that are initiated very early in the boot process. These mutexes are declared using a special MUTEX_DECLARE() macro, and use a new flag MTX_COLD when calling mtx_init. This is still somewhat hackish, but it is less evil than the mtx_f filler struct, and the mtx struct is now the same size with and without mutex debugging code. - Add some micro-micro-operation macros for doing the actual atomic operations on the mutex mtx_lock field to make it easier for other archs to override/optimize mutex ops if needed. These new tiny ops also clean up the code in some places by replacing long atomic operation function calls that spanned 2-3 lines with a short 1-line macro call. - Don't call mi_switch() from mtx_enter_hard() when we block while trying to obtain a sleep mutex. Calling mi_switch() would bogusly release Giant before switching to the next process. Instead, inline most of the code from mi_switch() in the mtx_enter_hard() function. Note that when we finally kill Giant we can back this out and go back to calling mi_switch().
2000-10-20 07:26:37 +00:00
}
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
/* diagnostic */
mtx_validate(m, MV_DESTROY);
- Make the mutex code almost completely machine independent. This greatly reducues the maintenance load for the mutex code. The only MD portions of the mutex code are in machine/mutex.h now, which include the assembly macros for handling mutexes as well as optionally overriding the mutex micro-operations. For example, we use optimized micro-ops on the x86 platform #ifndef I386_CPU. - Change the behavior of the SMP_DEBUG kernel option. In the new code, mtx_assert() only depends on INVARIANTS, allowing other kernel developers to have working mutex assertiions without having to include all of the mutex debugging code. The SMP_DEBUG kernel option has been renamed to MUTEX_DEBUG and now just controls extra mutex debugging code. - Abolish the ugly mtx_f hack. Instead, we dynamically allocate seperate mtx_debug structures on the fly in mtx_init, except for mutexes that are initiated very early in the boot process. These mutexes are declared using a special MUTEX_DECLARE() macro, and use a new flag MTX_COLD when calling mtx_init. This is still somewhat hackish, but it is less evil than the mtx_f filler struct, and the mtx struct is now the same size with and without mutex debugging code. - Add some micro-micro-operation macros for doing the actual atomic operations on the mutex mtx_lock field to make it easier for other archs to override/optimize mutex ops if needed. These new tiny ops also clean up the code in some places by replacing long atomic operation function calls that spanned 2-3 lines with a short 1-line macro call. - Don't call mi_switch() from mtx_enter_hard() when we block while trying to obtain a sleep mutex. Calling mi_switch() would bogusly release Giant before switching to the next process. Instead, inline most of the code from mi_switch() in the mtx_enter_hard() function. Note that when we finally kill Giant we can back this out and go back to calling mi_switch().
2000-10-20 07:26:37 +00:00
#endif
#ifdef WITNESS
if (m->mtx_witness)
witness_destroy(m);
#endif /* WITNESS */
/* Remove from the all mutex queue */
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
mtx_lock(&all_mtx);
- Make the mutex code almost completely machine independent. This greatly reducues the maintenance load for the mutex code. The only MD portions of the mutex code are in machine/mutex.h now, which include the assembly macros for handling mutexes as well as optionally overriding the mutex micro-operations. For example, we use optimized micro-ops on the x86 platform #ifndef I386_CPU. - Change the behavior of the SMP_DEBUG kernel option. In the new code, mtx_assert() only depends on INVARIANTS, allowing other kernel developers to have working mutex assertiions without having to include all of the mutex debugging code. The SMP_DEBUG kernel option has been renamed to MUTEX_DEBUG and now just controls extra mutex debugging code. - Abolish the ugly mtx_f hack. Instead, we dynamically allocate seperate mtx_debug structures on the fly in mtx_init, except for mutexes that are initiated very early in the boot process. These mutexes are declared using a special MUTEX_DECLARE() macro, and use a new flag MTX_COLD when calling mtx_init. This is still somewhat hackish, but it is less evil than the mtx_f filler struct, and the mtx struct is now the same size with and without mutex debugging code. - Add some micro-micro-operation macros for doing the actual atomic operations on the mutex mtx_lock field to make it easier for other archs to override/optimize mutex ops if needed. These new tiny ops also clean up the code in some places by replacing long atomic operation function calls that spanned 2-3 lines with a short 1-line macro call. - Don't call mi_switch() from mtx_enter_hard() when we block while trying to obtain a sleep mutex. Calling mi_switch() would bogusly release Giant before switching to the next process. Instead, inline most of the code from mi_switch() in the mtx_enter_hard() function. Note that when we finally kill Giant we can back this out and go back to calling mi_switch().
2000-10-20 07:26:37 +00:00
m->mtx_next->mtx_prev = m->mtx_prev;
m->mtx_prev->mtx_next = m->mtx_next;
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
- Make the mutex code almost completely machine independent. This greatly reducues the maintenance load for the mutex code. The only MD portions of the mutex code are in machine/mutex.h now, which include the assembly macros for handling mutexes as well as optionally overriding the mutex micro-operations. For example, we use optimized micro-ops on the x86 platform #ifndef I386_CPU. - Change the behavior of the SMP_DEBUG kernel option. In the new code, mtx_assert() only depends on INVARIANTS, allowing other kernel developers to have working mutex assertiions without having to include all of the mutex debugging code. The SMP_DEBUG kernel option has been renamed to MUTEX_DEBUG and now just controls extra mutex debugging code. - Abolish the ugly mtx_f hack. Instead, we dynamically allocate seperate mtx_debug structures on the fly in mtx_init, except for mutexes that are initiated very early in the boot process. These mutexes are declared using a special MUTEX_DECLARE() macro, and use a new flag MTX_COLD when calling mtx_init. This is still somewhat hackish, but it is less evil than the mtx_f filler struct, and the mtx struct is now the same size with and without mutex debugging code. - Add some micro-micro-operation macros for doing the actual atomic operations on the mutex mtx_lock field to make it easier for other archs to override/optimize mutex ops if needed. These new tiny ops also clean up the code in some places by replacing long atomic operation function calls that spanned 2-3 lines with a short 1-line macro call. - Don't call mi_switch() from mtx_enter_hard() when we block while trying to obtain a sleep mutex. Calling mi_switch() would bogusly release Giant before switching to the next process. Instead, inline most of the code from mi_switch() in the mtx_enter_hard() function. Note that when we finally kill Giant we can back this out and go back to calling mi_switch().
2000-10-20 07:26:37 +00:00
#ifdef MUTEX_DEBUG
m->mtx_next = m->mtx_prev = NULL;
Split the WITNESS and MUTEX_DEBUG options apart so that WITNESS does not depend on MUTEX_DEBUG. The MUTEX_DEBUG option turns on extra assertions and checks to verify that mutexes themselves are implemented properly. The WITNESS option uses extra checks and diagnostics to verify that other code is using mutexes properly.
2000-12-01 00:10:59 +00:00
#endif
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
Split the WITNESS and MUTEX_DEBUG options apart so that WITNESS does not depend on MUTEX_DEBUG. The MUTEX_DEBUG option turns on extra assertions and checks to verify that mutexes themselves are implemented properly. The WITNESS option uses extra checks and diagnostics to verify that other code is using mutexes properly.
2000-12-01 00:10:59 +00:00
#ifdef WITNESS
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
free(m->mtx_debug, M_WITNESS);
- Don't use a union and fun tricks to shave one extra pointer off of struct mtx right now as it makes debugging harder. When we are in optimizing mode, we can revisit this. - Fix the KTR trace messages to use %p rather than 0x%p to avoid duplicate 0x's in KTR output. - During witness_fixup, release Giant so that witness doesn't get confused. Also, grab all_mtx while walking the list of mutexes. - Remove w_sleep and w_recurse. Instead, perform checks on mutexes using the mutex's mtx_flags field. - Allow debug.witness_ddb and debug.witness_skipspin to be set from the loader. - Add Giant to the front of existing order_list entries to help ensure Giant is always first. - Add an order entry for the various proc locks. Note that this only helps keep proc in order mostly as the allproc and proctree mutexes are only obtained during a lockmgr operation on the specified mutex.
2001-01-24 10:57:01 +00:00
m->mtx_debug = NULL;
- Make the mutex code almost completely machine independent. This greatly reducues the maintenance load for the mutex code. The only MD portions of the mutex code are in machine/mutex.h now, which include the assembly macros for handling mutexes as well as optionally overriding the mutex micro-operations. For example, we use optimized micro-ops on the x86 platform #ifndef I386_CPU. - Change the behavior of the SMP_DEBUG kernel option. In the new code, mtx_assert() only depends on INVARIANTS, allowing other kernel developers to have working mutex assertiions without having to include all of the mutex debugging code. The SMP_DEBUG kernel option has been renamed to MUTEX_DEBUG and now just controls extra mutex debugging code. - Abolish the ugly mtx_f hack. Instead, we dynamically allocate seperate mtx_debug structures on the fly in mtx_init, except for mutexes that are initiated very early in the boot process. These mutexes are declared using a special MUTEX_DECLARE() macro, and use a new flag MTX_COLD when calling mtx_init. This is still somewhat hackish, but it is less evil than the mtx_f filler struct, and the mtx struct is now the same size with and without mutex debugging code. - Add some micro-micro-operation macros for doing the actual atomic operations on the mutex mtx_lock field to make it easier for other archs to override/optimize mutex ops if needed. These new tiny ops also clean up the code in some places by replacing long atomic operation function calls that spanned 2-3 lines with a short 1-line macro call. - Don't call mi_switch() from mtx_enter_hard() when we block while trying to obtain a sleep mutex. Calling mi_switch() would bogusly release Giant before switching to the next process. Instead, inline most of the code from mi_switch() in the mtx_enter_hard() function. Note that when we finally kill Giant we can back this out and go back to calling mi_switch().
2000-10-20 07:26:37 +00:00
#endif
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
- Make the mutex code almost completely machine independent. This greatly reducues the maintenance load for the mutex code. The only MD portions of the mutex code are in machine/mutex.h now, which include the assembly macros for handling mutexes as well as optionally overriding the mutex micro-operations. For example, we use optimized micro-ops on the x86 platform #ifndef I386_CPU. - Change the behavior of the SMP_DEBUG kernel option. In the new code, mtx_assert() only depends on INVARIANTS, allowing other kernel developers to have working mutex assertiions without having to include all of the mutex debugging code. The SMP_DEBUG kernel option has been renamed to MUTEX_DEBUG and now just controls extra mutex debugging code. - Abolish the ugly mtx_f hack. Instead, we dynamically allocate seperate mtx_debug structures on the fly in mtx_init, except for mutexes that are initiated very early in the boot process. These mutexes are declared using a special MUTEX_DECLARE() macro, and use a new flag MTX_COLD when calling mtx_init. This is still somewhat hackish, but it is less evil than the mtx_f filler struct, and the mtx struct is now the same size with and without mutex debugging code. - Add some micro-micro-operation macros for doing the actual atomic operations on the mutex mtx_lock field to make it easier for other archs to override/optimize mutex ops if needed. These new tiny ops also clean up the code in some places by replacing long atomic operation function calls that spanned 2-3 lines with a short 1-line macro call. - Don't call mi_switch() from mtx_enter_hard() when we block while trying to obtain a sleep mutex. Calling mi_switch() would bogusly release Giant before switching to the next process. Instead, inline most of the code from mi_switch() in the mtx_enter_hard() function. Note that when we finally kill Giant we can back this out and go back to calling mi_switch().
2000-10-20 07:26:37 +00:00
mtx_cur_cnt--;
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
mtx_unlock(&all_mtx);
- Make the mutex code almost completely machine independent. This greatly reducues the maintenance load for the mutex code. The only MD portions of the mutex code are in machine/mutex.h now, which include the assembly macros for handling mutexes as well as optionally overriding the mutex micro-operations. For example, we use optimized micro-ops on the x86 platform #ifndef I386_CPU. - Change the behavior of the SMP_DEBUG kernel option. In the new code, mtx_assert() only depends on INVARIANTS, allowing other kernel developers to have working mutex assertiions without having to include all of the mutex debugging code. The SMP_DEBUG kernel option has been renamed to MUTEX_DEBUG and now just controls extra mutex debugging code. - Abolish the ugly mtx_f hack. Instead, we dynamically allocate seperate mtx_debug structures on the fly in mtx_init, except for mutexes that are initiated very early in the boot process. These mutexes are declared using a special MUTEX_DECLARE() macro, and use a new flag MTX_COLD when calling mtx_init. This is still somewhat hackish, but it is less evil than the mtx_f filler struct, and the mtx struct is now the same size with and without mutex debugging code. - Add some micro-micro-operation macros for doing the actual atomic operations on the mutex mtx_lock field to make it easier for other archs to override/optimize mutex ops if needed. These new tiny ops also clean up the code in some places by replacing long atomic operation function calls that spanned 2-3 lines with a short 1-line macro call. - Don't call mi_switch() from mtx_enter_hard() when we block while trying to obtain a sleep mutex. Calling mi_switch() would bogusly release Giant before switching to the next process. Instead, inline most of the code from mi_switch() in the mtx_enter_hard() function. Note that when we finally kill Giant we can back this out and go back to calling mi_switch().
2000-10-20 07:26:37 +00:00
}
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
- Don't use a union and fun tricks to shave one extra pointer off of struct mtx right now as it makes debugging harder. When we are in optimizing mode, we can revisit this. - Fix the KTR trace messages to use %p rather than 0x%p to avoid duplicate 0x's in KTR output. - During witness_fixup, release Giant so that witness doesn't get confused. Also, grab all_mtx while walking the list of mutexes. - Remove w_sleep and w_recurse. Instead, perform checks on mutexes using the mutex's mtx_flags field. - Allow debug.witness_ddb and debug.witness_skipspin to be set from the loader. - Add Giant to the front of existing order_list entries to help ensure Giant is always first. - Add an order entry for the various proc locks. Note that this only helps keep proc in order mostly as the allproc and proctree mutexes are only obtained during a lockmgr operation on the specified mutex.
2001-01-24 10:57:01 +00:00
/*
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
* The WITNESS-enabled diagnostic code.
- Don't use a union and fun tricks to shave one extra pointer off of struct mtx right now as it makes debugging harder. When we are in optimizing mode, we can revisit this. - Fix the KTR trace messages to use %p rather than 0x%p to avoid duplicate 0x's in KTR output. - During witness_fixup, release Giant so that witness doesn't get confused. Also, grab all_mtx while walking the list of mutexes. - Remove w_sleep and w_recurse. Instead, perform checks on mutexes using the mutex's mtx_flags field. - Allow debug.witness_ddb and debug.witness_skipspin to be set from the loader. - Add Giant to the front of existing order_list entries to help ensure Giant is always first. - Add an order entry for the various proc locks. Note that this only helps keep proc in order mostly as the allproc and proctree mutexes are only obtained during a lockmgr operation on the specified mutex.
2001-01-24 10:57:01 +00:00
*/
#ifdef WITNESS
Remove MUTEX_DECLARE() and MTX_COLD. Instead, postpone full mutex initialization until after malloc() is safe to call, then iterate through all mutexes and complete their initialization. This change is necessary in order to avoid some circular bootstrapping dependencies.
2001-01-21 07:52:20 +00:00
static void
witness_fixup(void *dummy __unused)
{
struct mtx *mp;
- Don't use a union and fun tricks to shave one extra pointer off of struct mtx right now as it makes debugging harder. When we are in optimizing mode, we can revisit this. - Fix the KTR trace messages to use %p rather than 0x%p to avoid duplicate 0x's in KTR output. - During witness_fixup, release Giant so that witness doesn't get confused. Also, grab all_mtx while walking the list of mutexes. - Remove w_sleep and w_recurse. Instead, perform checks on mutexes using the mutex's mtx_flags field. - Allow debug.witness_ddb and debug.witness_skipspin to be set from the loader. - Add Giant to the front of existing order_list entries to help ensure Giant is always first. - Add an order entry for the various proc locks. Note that this only helps keep proc in order mostly as the allproc and proctree mutexes are only obtained during a lockmgr operation on the specified mutex.
2001-01-24 10:57:01 +00:00
/*
* We have to release Giant before initializing its witness
* structure so that WITNESS doesn't get confused.
*/
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
mtx_unlock(&Giant);
- Don't use a union and fun tricks to shave one extra pointer off of struct mtx right now as it makes debugging harder. When we are in optimizing mode, we can revisit this. - Fix the KTR trace messages to use %p rather than 0x%p to avoid duplicate 0x's in KTR output. - During witness_fixup, release Giant so that witness doesn't get confused. Also, grab all_mtx while walking the list of mutexes. - Remove w_sleep and w_recurse. Instead, perform checks on mutexes using the mutex's mtx_flags field. - Allow debug.witness_ddb and debug.witness_skipspin to be set from the loader. - Add Giant to the front of existing order_list entries to help ensure Giant is always first. - Add an order entry for the various proc locks. Note that this only helps keep proc in order mostly as the allproc and proctree mutexes are only obtained during a lockmgr operation on the specified mutex.
2001-01-24 10:57:01 +00:00
mtx_assert(&Giant, MA_NOTOWNED);
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
mtx_lock(&all_mtx);
Remove MUTEX_DECLARE() and MTX_COLD. Instead, postpone full mutex initialization until after malloc() is safe to call, then iterate through all mutexes and complete their initialization. This change is necessary in order to avoid some circular bootstrapping dependencies.
2001-01-21 07:52:20 +00:00
/* Iterate through all mutexes and finish up mutex initialization. */
for (mp = all_mtx.mtx_next; mp != &all_mtx; mp = mp->mtx_next) {
- Don't use a union and fun tricks to shave one extra pointer off of struct mtx right now as it makes debugging harder. When we are in optimizing mode, we can revisit this. - Fix the KTR trace messages to use %p rather than 0x%p to avoid duplicate 0x's in KTR output. - During witness_fixup, release Giant so that witness doesn't get confused. Also, grab all_mtx while walking the list of mutexes. - Remove w_sleep and w_recurse. Instead, perform checks on mutexes using the mutex's mtx_flags field. - Allow debug.witness_ddb and debug.witness_skipspin to be set from the loader. - Add Giant to the front of existing order_list entries to help ensure Giant is always first. - Add an order entry for the various proc locks. Note that this only helps keep proc in order mostly as the allproc and proctree mutexes are only obtained during a lockmgr operation on the specified mutex.
2001-01-24 10:57:01 +00:00
mp->mtx_debug = malloc(sizeof(struct mtx_debug),
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
M_WITNESS, M_NOWAIT | M_ZERO);
- Don't use a union and fun tricks to shave one extra pointer off of struct mtx right now as it makes debugging harder. When we are in optimizing mode, we can revisit this. - Fix the KTR trace messages to use %p rather than 0x%p to avoid duplicate 0x's in KTR output. - During witness_fixup, release Giant so that witness doesn't get confused. Also, grab all_mtx while walking the list of mutexes. - Remove w_sleep and w_recurse. Instead, perform checks on mutexes using the mutex's mtx_flags field. - Allow debug.witness_ddb and debug.witness_skipspin to be set from the loader. - Add Giant to the front of existing order_list entries to help ensure Giant is always first. - Add an order entry for the various proc locks. Note that this only helps keep proc in order mostly as the allproc and proctree mutexes are only obtained during a lockmgr operation on the specified mutex.
2001-01-24 10:57:01 +00:00
MPASS(mp->mtx_debug != NULL);
Remove MUTEX_DECLARE() and MTX_COLD. Instead, postpone full mutex initialization until after malloc() is safe to call, then iterate through all mutexes and complete their initialization. This change is necessary in order to avoid some circular bootstrapping dependencies.
2001-01-21 07:52:20 +00:00
witness_init(mp, mp->mtx_flags);
}
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
mtx_unlock(&all_mtx);
Remove MUTEX_DECLARE() and MTX_COLD. Instead, postpone full mutex initialization until after malloc() is safe to call, then iterate through all mutexes and complete their initialization. This change is necessary in order to avoid some circular bootstrapping dependencies.
2001-01-21 07:52:20 +00:00
/* Mark the witness code as being ready for use. */
atomic_store_rel_int(&witness_cold, 0);
- Don't use a union and fun tricks to shave one extra pointer off of struct mtx right now as it makes debugging harder. When we are in optimizing mode, we can revisit this. - Fix the KTR trace messages to use %p rather than 0x%p to avoid duplicate 0x's in KTR output. - During witness_fixup, release Giant so that witness doesn't get confused. Also, grab all_mtx while walking the list of mutexes. - Remove w_sleep and w_recurse. Instead, perform checks on mutexes using the mutex's mtx_flags field. - Allow debug.witness_ddb and debug.witness_skipspin to be set from the loader. - Add Giant to the front of existing order_list entries to help ensure Giant is always first. - Add an order entry for the various proc locks. Note that this only helps keep proc in order mostly as the allproc and proctree mutexes are only obtained during a lockmgr operation on the specified mutex.
2001-01-24 10:57:01 +00:00
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
mtx_lock(&Giant);
Remove MUTEX_DECLARE() and MTX_COLD. Instead, postpone full mutex initialization until after malloc() is safe to call, then iterate through all mutexes and complete their initialization. This change is necessary in order to avoid some circular bootstrapping dependencies.
2001-01-21 07:52:20 +00:00
}
SYSINIT(wtnsfxup, SI_SUB_MUTEX, SI_ORDER_FIRST, witness_fixup, NULL)
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
#define WITNESS_COUNT 200
#define WITNESS_NCHILDREN 2
Actually enable the witness code if the WITNESS kernel option is enabled.
2000-10-20 21:58:11 +00:00
int witness_watch = 1;
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
Remove the mtx_t, witness_t, and witness_blessed_t types. Instead, just use struct mtx, struct witness, and struct witness_blessed. Requested by: bde
2000-09-14 20:15:16 +00:00
struct witness {
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
struct witness *w_next;
Propogate the 'const'ness of mutex descriptions to the witness code to quiet warnings.
2000-10-20 22:45:01 +00:00
const char *w_description;
Rename mtx_enter(), mtx_try_enter(), and mtx_exit() and wrap them with cpp macros that expand to pass filename and line number information. This is necessary since we're using inline functions instead of macros now. Add const to the filename pointers passed througout the mtx and witness code.
2000-09-08 21:48:06 +00:00
const char *w_file;
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
int w_line;
struct witness *w_morechildren;
u_char w_childcnt;
u_char w_Giant_squawked:1;
u_char w_other_squawked:1;
u_char w_same_squawked:1;
Implement MTX_RECURSE flag for mtx_init(). All calls to mtx_init() for mutexes that recurse must now include the MTX_RECURSE bit in the flag argument variable. This change is in preparation for an upcoming (further) mutex API cleanup. The witness code will call panic() if a lock is found to recurse but the MTX_RECURSE bit was not set during the lock's initialization. The old MTX_RECURSE "state" bit (in mtx_lock) has been renamed to MTX_RECURSED, which is more appropriate given its meaning. The following locks have been made "recursive," thus far: eventhandler, Giant, callout, sched_lock, possibly some others declared in the architecture-specific code, all of the network card driver locks in pci/, as well as some other locks in dev/ stuff that I've found to be recursive. Reviewed by: jhb
2001-01-19 01:59:14 +00:00
u_char w_spin:1; /* MTX_SPIN type mutex. */
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
u_int w_level;
struct witness *w_children[WITNESS_NCHILDREN];
Remove the mtx_t, witness_t, and witness_blessed_t types. Instead, just use struct mtx, struct witness, and struct witness_blessed. Requested by: bde
2000-09-14 20:15:16 +00:00
};
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
Remove the mtx_t, witness_t, and witness_blessed_t types. Instead, just use struct mtx, struct witness, and struct witness_blessed. Requested by: bde
2000-09-14 20:15:16 +00:00
struct witness_blessed {
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
char *b_lock1;
char *b_lock2;
Remove the mtx_t, witness_t, and witness_blessed_t types. Instead, just use struct mtx, struct witness, and struct witness_blessed. Requested by: bde
2000-09-14 20:15:16 +00:00
};
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
- Use MUTEX_DECLARE() and MTX_COLD for the WITNESS code's internal mutex so it can function before malloc(9) is up and running. - Add two new options WITNESS_DDB and WITNESS_SKIPSPIN. If WITNESS_SKIPSPIN is enabled, then spin mutexes are ignored by the WITNESS code. If WITNESS_DDB is turned on and DDB is compiled into the kernel, then the kernel will drop into DDB when either a lock hierarchy violation occurs or mutexes are held when going to sleep. - Add some new sysctls: debug.witness_ddb is a read-write sysctl that corresponds to WITNESS_DDB. The kernel option merely changes the default value to on at boot. debug.witness_skipspin is a read-only sysctl that one can use to determine if the kernel was compiled with WITNESS_SKIPSPIN. - Wipe out the BSD/OS-specific lock order lists. We get to build our own lists now as we add mutexes to the kernel.
2000-10-27 02:59:30 +00:00
#ifdef DDB
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
/*
- Use MUTEX_DECLARE() and MTX_COLD for the WITNESS code's internal mutex so it can function before malloc(9) is up and running. - Add two new options WITNESS_DDB and WITNESS_SKIPSPIN. If WITNESS_SKIPSPIN is enabled, then spin mutexes are ignored by the WITNESS code. If WITNESS_DDB is turned on and DDB is compiled into the kernel, then the kernel will drop into DDB when either a lock hierarchy violation occurs or mutexes are held when going to sleep. - Add some new sysctls: debug.witness_ddb is a read-write sysctl that corresponds to WITNESS_DDB. The kernel option merely changes the default value to on at boot. debug.witness_skipspin is a read-only sysctl that one can use to determine if the kernel was compiled with WITNESS_SKIPSPIN. - Wipe out the BSD/OS-specific lock order lists. We get to build our own lists now as we add mutexes to the kernel.
2000-10-27 02:59:30 +00:00
* When DDB is enabled and witness_ddb is set to 1, it will cause the system to
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
* drop into kdebug() when:
* - a lock heirarchy violation occurs
* - locks are held when going to sleep.
*/
- Don't use a union and fun tricks to shave one extra pointer off of struct mtx right now as it makes debugging harder. When we are in optimizing mode, we can revisit this. - Fix the KTR trace messages to use %p rather than 0x%p to avoid duplicate 0x's in KTR output. - During witness_fixup, release Giant so that witness doesn't get confused. Also, grab all_mtx while walking the list of mutexes. - Remove w_sleep and w_recurse. Instead, perform checks on mutexes using the mutex's mtx_flags field. - Allow debug.witness_ddb and debug.witness_skipspin to be set from the loader. - Add Giant to the front of existing order_list entries to help ensure Giant is always first. - Add an order entry for the various proc locks. Note that this only helps keep proc in order mostly as the allproc and proctree mutexes are only obtained during a lockmgr operation on the specified mutex.
2001-01-24 10:57:01 +00:00
int witness_ddb;
- Use MUTEX_DECLARE() and MTX_COLD for the WITNESS code's internal mutex so it can function before malloc(9) is up and running. - Add two new options WITNESS_DDB and WITNESS_SKIPSPIN. If WITNESS_SKIPSPIN is enabled, then spin mutexes are ignored by the WITNESS code. If WITNESS_DDB is turned on and DDB is compiled into the kernel, then the kernel will drop into DDB when either a lock hierarchy violation occurs or mutexes are held when going to sleep. - Add some new sysctls: debug.witness_ddb is a read-write sysctl that corresponds to WITNESS_DDB. The kernel option merely changes the default value to on at boot. debug.witness_skipspin is a read-only sysctl that one can use to determine if the kernel was compiled with WITNESS_SKIPSPIN. - Wipe out the BSD/OS-specific lock order lists. We get to build our own lists now as we add mutexes to the kernel.
2000-10-27 02:59:30 +00:00
#ifdef WITNESS_DDB
- Don't use a union and fun tricks to shave one extra pointer off of struct mtx right now as it makes debugging harder. When we are in optimizing mode, we can revisit this. - Fix the KTR trace messages to use %p rather than 0x%p to avoid duplicate 0x's in KTR output. - During witness_fixup, release Giant so that witness doesn't get confused. Also, grab all_mtx while walking the list of mutexes. - Remove w_sleep and w_recurse. Instead, perform checks on mutexes using the mutex's mtx_flags field. - Allow debug.witness_ddb and debug.witness_skipspin to be set from the loader. - Add Giant to the front of existing order_list entries to help ensure Giant is always first. - Add an order entry for the various proc locks. Note that this only helps keep proc in order mostly as the allproc and proctree mutexes are only obtained during a lockmgr operation on the specified mutex.
2001-01-24 10:57:01 +00:00
TUNABLE_INT_DECL("debug.witness_ddb", 1, witness_ddb);
- Use MUTEX_DECLARE() and MTX_COLD for the WITNESS code's internal mutex so it can function before malloc(9) is up and running. - Add two new options WITNESS_DDB and WITNESS_SKIPSPIN. If WITNESS_SKIPSPIN is enabled, then spin mutexes are ignored by the WITNESS code. If WITNESS_DDB is turned on and DDB is compiled into the kernel, then the kernel will drop into DDB when either a lock hierarchy violation occurs or mutexes are held when going to sleep. - Add some new sysctls: debug.witness_ddb is a read-write sysctl that corresponds to WITNESS_DDB. The kernel option merely changes the default value to on at boot. debug.witness_skipspin is a read-only sysctl that one can use to determine if the kernel was compiled with WITNESS_SKIPSPIN. - Wipe out the BSD/OS-specific lock order lists. We get to build our own lists now as we add mutexes to the kernel.
2000-10-27 02:59:30 +00:00
#else
- Don't use a union and fun tricks to shave one extra pointer off of struct mtx right now as it makes debugging harder. When we are in optimizing mode, we can revisit this. - Fix the KTR trace messages to use %p rather than 0x%p to avoid duplicate 0x's in KTR output. - During witness_fixup, release Giant so that witness doesn't get confused. Also, grab all_mtx while walking the list of mutexes. - Remove w_sleep and w_recurse. Instead, perform checks on mutexes using the mutex's mtx_flags field. - Allow debug.witness_ddb and debug.witness_skipspin to be set from the loader. - Add Giant to the front of existing order_list entries to help ensure Giant is always first. - Add an order entry for the various proc locks. Note that this only helps keep proc in order mostly as the allproc and proctree mutexes are only obtained during a lockmgr operation on the specified mutex.
2001-01-24 10:57:01 +00:00
TUNABLE_INT_DECL("debug.witness_ddb", 0, witness_ddb);
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
#endif
- Use MUTEX_DECLARE() and MTX_COLD for the WITNESS code's internal mutex so it can function before malloc(9) is up and running. - Add two new options WITNESS_DDB and WITNESS_SKIPSPIN. If WITNESS_SKIPSPIN is enabled, then spin mutexes are ignored by the WITNESS code. If WITNESS_DDB is turned on and DDB is compiled into the kernel, then the kernel will drop into DDB when either a lock hierarchy violation occurs or mutexes are held when going to sleep. - Add some new sysctls: debug.witness_ddb is a read-write sysctl that corresponds to WITNESS_DDB. The kernel option merely changes the default value to on at boot. debug.witness_skipspin is a read-only sysctl that one can use to determine if the kernel was compiled with WITNESS_SKIPSPIN. - Wipe out the BSD/OS-specific lock order lists. We get to build our own lists now as we add mutexes to the kernel.
2000-10-27 02:59:30 +00:00
SYSCTL_INT(_debug, OID_AUTO, witness_ddb, CTLFLAG_RW, &witness_ddb, 0, "");
#endif /* DDB */
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
- Don't use a union and fun tricks to shave one extra pointer off of struct mtx right now as it makes debugging harder. When we are in optimizing mode, we can revisit this. - Fix the KTR trace messages to use %p rather than 0x%p to avoid duplicate 0x's in KTR output. - During witness_fixup, release Giant so that witness doesn't get confused. Also, grab all_mtx while walking the list of mutexes. - Remove w_sleep and w_recurse. Instead, perform checks on mutexes using the mutex's mtx_flags field. - Allow debug.witness_ddb and debug.witness_skipspin to be set from the loader. - Add Giant to the front of existing order_list entries to help ensure Giant is always first. - Add an order entry for the various proc locks. Note that this only helps keep proc in order mostly as the allproc and proctree mutexes are only obtained during a lockmgr operation on the specified mutex.
2001-01-24 10:57:01 +00:00
int witness_skipspin;
- Use MUTEX_DECLARE() and MTX_COLD for the WITNESS code's internal mutex so it can function before malloc(9) is up and running. - Add two new options WITNESS_DDB and WITNESS_SKIPSPIN. If WITNESS_SKIPSPIN is enabled, then spin mutexes are ignored by the WITNESS code. If WITNESS_DDB is turned on and DDB is compiled into the kernel, then the kernel will drop into DDB when either a lock hierarchy violation occurs or mutexes are held when going to sleep. - Add some new sysctls: debug.witness_ddb is a read-write sysctl that corresponds to WITNESS_DDB. The kernel option merely changes the default value to on at boot. debug.witness_skipspin is a read-only sysctl that one can use to determine if the kernel was compiled with WITNESS_SKIPSPIN. - Wipe out the BSD/OS-specific lock order lists. We get to build our own lists now as we add mutexes to the kernel.
2000-10-27 02:59:30 +00:00
#ifdef WITNESS_SKIPSPIN
- Don't use a union and fun tricks to shave one extra pointer off of struct mtx right now as it makes debugging harder. When we are in optimizing mode, we can revisit this. - Fix the KTR trace messages to use %p rather than 0x%p to avoid duplicate 0x's in KTR output. - During witness_fixup, release Giant so that witness doesn't get confused. Also, grab all_mtx while walking the list of mutexes. - Remove w_sleep and w_recurse. Instead, perform checks on mutexes using the mutex's mtx_flags field. - Allow debug.witness_ddb and debug.witness_skipspin to be set from the loader. - Add Giant to the front of existing order_list entries to help ensure Giant is always first. - Add an order entry for the various proc locks. Note that this only helps keep proc in order mostly as the allproc and proctree mutexes are only obtained during a lockmgr operation on the specified mutex.
2001-01-24 10:57:01 +00:00
TUNABLE_INT_DECL("debug.witness_skipspin", 1, witness_skipspin);
- Use MUTEX_DECLARE() and MTX_COLD for the WITNESS code's internal mutex so it can function before malloc(9) is up and running. - Add two new options WITNESS_DDB and WITNESS_SKIPSPIN. If WITNESS_SKIPSPIN is enabled, then spin mutexes are ignored by the WITNESS code. If WITNESS_DDB is turned on and DDB is compiled into the kernel, then the kernel will drop into DDB when either a lock hierarchy violation occurs or mutexes are held when going to sleep. - Add some new sysctls: debug.witness_ddb is a read-write sysctl that corresponds to WITNESS_DDB. The kernel option merely changes the default value to on at boot. debug.witness_skipspin is a read-only sysctl that one can use to determine if the kernel was compiled with WITNESS_SKIPSPIN. - Wipe out the BSD/OS-specific lock order lists. We get to build our own lists now as we add mutexes to the kernel.
2000-10-27 02:59:30 +00:00
#else
- Don't use a union and fun tricks to shave one extra pointer off of struct mtx right now as it makes debugging harder. When we are in optimizing mode, we can revisit this. - Fix the KTR trace messages to use %p rather than 0x%p to avoid duplicate 0x's in KTR output. - During witness_fixup, release Giant so that witness doesn't get confused. Also, grab all_mtx while walking the list of mutexes. - Remove w_sleep and w_recurse. Instead, perform checks on mutexes using the mutex's mtx_flags field. - Allow debug.witness_ddb and debug.witness_skipspin to be set from the loader. - Add Giant to the front of existing order_list entries to help ensure Giant is always first. - Add an order entry for the various proc locks. Note that this only helps keep proc in order mostly as the allproc and proctree mutexes are only obtained during a lockmgr operation on the specified mutex.
2001-01-24 10:57:01 +00:00
TUNABLE_INT_DECL("debug.witness_skipspin", 0, witness_skipspin);
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
#endif
- Use MUTEX_DECLARE() and MTX_COLD for the WITNESS code's internal mutex so it can function before malloc(9) is up and running. - Add two new options WITNESS_DDB and WITNESS_SKIPSPIN. If WITNESS_SKIPSPIN is enabled, then spin mutexes are ignored by the WITNESS code. If WITNESS_DDB is turned on and DDB is compiled into the kernel, then the kernel will drop into DDB when either a lock hierarchy violation occurs or mutexes are held when going to sleep. - Add some new sysctls: debug.witness_ddb is a read-write sysctl that corresponds to WITNESS_DDB. The kernel option merely changes the default value to on at boot. debug.witness_skipspin is a read-only sysctl that one can use to determine if the kernel was compiled with WITNESS_SKIPSPIN. - Wipe out the BSD/OS-specific lock order lists. We get to build our own lists now as we add mutexes to the kernel.
2000-10-27 02:59:30 +00:00
SYSCTL_INT(_debug, OID_AUTO, witness_skipspin, CTLFLAG_RD, &witness_skipspin, 0,
"");
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
/*
* Witness-enabled globals
*/
Remove MUTEX_DECLARE() and MTX_COLD. Instead, postpone full mutex initialization until after malloc() is safe to call, then iterate through all mutexes and complete their initialization. This change is necessary in order to avoid some circular bootstrapping dependencies.
2001-01-21 07:52:20 +00:00
static struct mtx w_mtx;
Remove the mtx_t, witness_t, and witness_blessed_t types. Instead, just use struct mtx, struct witness, and struct witness_blessed. Requested by: bde
2000-09-14 20:15:16 +00:00
static struct witness *w_free;
static struct witness *w_all;
static int w_inited;
static int witness_dead; /* fatal error, probably no memory */
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
Remove the mtx_t, witness_t, and witness_blessed_t types. Instead, just use struct mtx, struct witness, and struct witness_blessed. Requested by: bde
2000-09-14 20:15:16 +00:00
static struct witness w_data[WITNESS_COUNT];
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
/*
* Internal witness routine prototypes
*/
static struct witness *enroll(const char *description, int flag);
static int itismychild(struct witness *parent, struct witness *child);
static void removechild(struct witness *parent, struct witness *child);
static int isitmychild(struct witness *parent, struct witness *child);
static int isitmydescendant(struct witness *parent, struct witness *child);
static int dup_ok(struct witness *);
static int blessed(struct witness *, struct witness *);
static void
witness_displaydescendants(void(*)(const char *fmt, ...), struct witness *);
static void witness_leveldescendents(struct witness *parent, int level);
static void witness_levelall(void);
static struct witness * witness_get(void);
static void witness_free(struct witness *m);
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
static char *ignore_list[] = {
"witness lock",
NULL
};
static char *spin_order_list[] = {
- Change the 'witness_list' ddb command to 'show mutexes'. Note that this will only display sleep mutexes held by the current process. - Clean up some nits in the witness_display() function and add a ddb command 'show witness' that dumps the hierarchy and order lists to the console. - Use queue(3) macros where appropriate. - Resort the spin lock order list so that "com" is before "sched_lock". Also, add appropriate #ifdef's around SMP and i386-specific mutexes. - Add two new mutexes used to protect the ithread lists and tables to the order list. Requested by: bde (1)
2001-02-09 15:19:41 +00:00
#if defined(__i386__) && defined (SMP)
"com",
#endif
- Use MUTEX_DECLARE() and MTX_COLD for the WITNESS code's internal mutex so it can function before malloc(9) is up and running. - Add two new options WITNESS_DDB and WITNESS_SKIPSPIN. If WITNESS_SKIPSPIN is enabled, then spin mutexes are ignored by the WITNESS code. If WITNESS_DDB is turned on and DDB is compiled into the kernel, then the kernel will drop into DDB when either a lock hierarchy violation occurs or mutexes are held when going to sleep. - Add some new sysctls: debug.witness_ddb is a read-write sysctl that corresponds to WITNESS_DDB. The kernel option merely changes the default value to on at boot. debug.witness_skipspin is a read-only sysctl that one can use to determine if the kernel was compiled with WITNESS_SKIPSPIN. - Wipe out the BSD/OS-specific lock order lists. We get to build our own lists now as we add mutexes to the kernel.
2000-10-27 02:59:30 +00:00
"sio",
- Change the 'witness_list' ddb command to 'show mutexes'. Note that this will only display sleep mutexes held by the current process. - Clean up some nits in the witness_display() function and add a ddb command 'show witness' that dumps the hierarchy and order lists to the console. - Use queue(3) macros where appropriate. - Resort the spin lock order list so that "com" is before "sched_lock". Also, add appropriate #ifdef's around SMP and i386-specific mutexes. - Add two new mutexes used to protect the ithread lists and tables to the order list. Requested by: bde (1)
2001-02-09 15:19:41 +00:00
#ifdef __i386__
"cy",
#endif
Shuffle netgraph mutexes a bit and hold a reference on a node from the function that is calling the destructor.
2001-02-28 18:49:09 +00:00
"ng_node",
"ng_worklist",
- Use the NOCPU constant. - Move the ithread spin locks before sched lock and clk in preparation for future commits to the ithread code.
2001-02-22 02:12:54 +00:00
"ithread table lock",
"ithread list lock",
The sched_lock mutex goes after the sio mutex in the locking order since a software interrupt can be scheduled in the sio interrupt handler while the sio mutex is held.
2000-11-29 18:38:14 +00:00
"sched lock",
Don't release and acquire Giant in mi_switch(). Instead, release and acquire Giant as needed in functions that call mi_switch(). The releases need to be done outside of the sched_lock to avoid potential deadlocks from trying to acquire Giant while interrupts are disabled. Submitted by: witness
2000-11-16 02:16:44 +00:00
#ifdef __i386__
"clk",
#endif
- Protect the callout wheel with a separate spin mutex, callout_lock. - Use the mutex in hardclock to ensure no races between it and softclock. - Make softclock be INTR_MPSAFE and provide a flag, CALLOUT_MPSAFE, which specifies that a callout handler does not need giant. There is still no way to set this flag when regstering a callout. Reviewed by: -smp@, jlemon
2000-11-19 06:02:32 +00:00
"callout",
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
/*
* leaf locks
*/
- Change the 'witness_list' ddb command to 'show mutexes'. Note that this will only display sleep mutexes held by the current process. - Clean up some nits in the witness_display() function and add a ddb command 'show witness' that dumps the hierarchy and order lists to the console. - Use queue(3) macros where appropriate. - Resort the spin lock order list so that "com" is before "sched_lock". Also, add appropriate #ifdef's around SMP and i386-specific mutexes. - Add two new mutexes used to protect the ithread lists and tables to the order list. Requested by: bde (1)
2001-02-09 15:19:41 +00:00
#ifdef SMP
Convert all simplelocks to mutexes and remove the simplelock implementations.
2001-01-24 12:35:55 +00:00
#ifdef __i386__
"ap boot",
"imen",
#endif
"smp rendezvous",
- Change the 'witness_list' ddb command to 'show mutexes'. Note that this will only display sleep mutexes held by the current process. - Clean up some nits in the witness_display() function and add a ddb command 'show witness' that dumps the hierarchy and order lists to the console. - Use queue(3) macros where appropriate. - Resort the spin lock order list so that "com" is before "sched_lock". Also, add appropriate #ifdef's around SMP and i386-specific mutexes. - Add two new mutexes used to protect the ithread lists and tables to the order list. Requested by: bde (1)
2001-02-09 15:19:41 +00:00
#endif
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
NULL
};
static char *order_list[] = {
Unify the two sleep lock order lists to enforce the process lock -> uidinfo lock locking order.
2001-02-09 20:52:02 +00:00
"Giant", "proctree", "allproc", "process lock", "uidinfo hash",
"uidinfo struct", NULL,
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
NULL
};
static char *dup_list[] = {
- Add an extra check in priority_propagation() for UP systems to ensure we don't end up back at ourselves which would indicate deadlock. - Add the proc lock to the witness dup_list as we may hold more than one process lock at a time. - Don't assert a mutex is owned in _mtx_unlock_sleep() as that is too late. We do the checks in the macros instead.
2001-03-07 02:45:15 +00:00
"process lock",
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
NULL
};
static char *sleep_list[] = {
- Split the run queue and sleep queue linkage, so that a process may block on a mutex while on the sleep queue without corrupting it. - Move dropping of Giant to after the acquire of sched_lock. Tested by: John Hay <jhay@icomtek.csir.co.za> jhb
2000-11-17 18:09:18 +00:00
"Giant",
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
NULL
};
/*
* Pairs of locks which have been blessed
* Don't complain about order problems with blessed locks
*/
Remove the mtx_t, witness_t, and witness_blessed_t types. Instead, just use struct mtx, struct witness, and struct witness_blessed. Requested by: bde
2000-09-14 20:15:16 +00:00
static struct witness_blessed blessed_list[] = {
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
};
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
static int blessed_count =
sizeof(blessed_list) / sizeof(struct witness_blessed);
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
Move most of sys/mutex.h into kern/kern_mutex.c, thereby making the mutex inline functions non-inlined. Hide parts of the mutex implementation that should not be exposed. Make sure that WITNESS code is not executed during boot until the mutexes are fully initialized by SI_SUB_MUTEX (the original motivation for this commit). Submitted by: peter
2001-01-21 22:34:43 +00:00
static void
Remove the mtx_t, witness_t, and witness_blessed_t types. Instead, just use struct mtx, struct witness, and struct witness_blessed. Requested by: bde
2000-09-14 20:15:16 +00:00
witness_init(struct mtx *m, int flag)
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
{
m->mtx_witness = enroll(m->mtx_description, flag);
}
Move most of sys/mutex.h into kern/kern_mutex.c, thereby making the mutex inline functions non-inlined. Hide parts of the mutex implementation that should not be exposed. Make sure that WITNESS code is not executed during boot until the mutexes are fully initialized by SI_SUB_MUTEX (the original motivation for this commit). Submitted by: peter
2001-01-21 22:34:43 +00:00
static void
Remove the mtx_t, witness_t, and witness_blessed_t types. Instead, just use struct mtx, struct witness, and struct witness_blessed. Requested by: bde
2000-09-14 20:15:16 +00:00
witness_destroy(struct mtx *m)
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
{
Remove the mtx_t, witness_t, and witness_blessed_t types. Instead, just use struct mtx, struct witness, and struct witness_blessed. Requested by: bde
2000-09-14 20:15:16 +00:00
struct mtx *m1;
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
struct proc *p;
Change all instances of `CURPROC' and `CURTHD' to `curproc,' in order to stay consistent. Requested by: bde
2001-02-12 03:15:43 +00:00
p = curproc;
- Change the 'witness_list' ddb command to 'show mutexes'. Note that this will only display sleep mutexes held by the current process. - Clean up some nits in the witness_display() function and add a ddb command 'show witness' that dumps the hierarchy and order lists to the console. - Use queue(3) macros where appropriate. - Resort the spin lock order list so that "com" is before "sched_lock". Also, add appropriate #ifdef's around SMP and i386-specific mutexes. - Add two new mutexes used to protect the ithread lists and tables to the order list. Requested by: bde (1)
2001-02-09 15:19:41 +00:00
LIST_FOREACH(m1, &p->p_heldmtx, mtx_held) {
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
if (m1 == m) {
LIST_REMOVE(m, mtx_held);
break;
}
}
return;
}
Move most of sys/mutex.h into kern/kern_mutex.c, thereby making the mutex inline functions non-inlined. Hide parts of the mutex implementation that should not be exposed. Make sure that WITNESS code is not executed during boot until the mutexes are fully initialized by SI_SUB_MUTEX (the original motivation for this commit). Submitted by: peter
2001-01-21 22:34:43 +00:00
static void
witness_display(void(*prnt)(const char *fmt, ...))
{
struct witness *w, *w1;
- Change the 'witness_list' ddb command to 'show mutexes'. Note that this will only display sleep mutexes held by the current process. - Clean up some nits in the witness_display() function and add a ddb command 'show witness' that dumps the hierarchy and order lists to the console. - Use queue(3) macros where appropriate. - Resort the spin lock order list so that "com" is before "sched_lock". Also, add appropriate #ifdef's around SMP and i386-specific mutexes. - Add two new mutexes used to protect the ithread lists and tables to the order list. Requested by: bde (1)
2001-02-09 15:19:41 +00:00
int level, found;
Move most of sys/mutex.h into kern/kern_mutex.c, thereby making the mutex inline functions non-inlined. Hide parts of the mutex implementation that should not be exposed. Make sure that WITNESS code is not executed during boot until the mutexes are fully initialized by SI_SUB_MUTEX (the original motivation for this commit). Submitted by: peter
2001-01-21 22:34:43 +00:00
KASSERT(!witness_cold, ("%s: witness_cold\n", __FUNCTION__));
witness_levelall();
- Change the 'witness_list' ddb command to 'show mutexes'. Note that this will only display sleep mutexes held by the current process. - Clean up some nits in the witness_display() function and add a ddb command 'show witness' that dumps the hierarchy and order lists to the console. - Use queue(3) macros where appropriate. - Resort the spin lock order list so that "com" is before "sched_lock". Also, add appropriate #ifdef's around SMP and i386-specific mutexes. - Add two new mutexes used to protect the ithread lists and tables to the order list. Requested by: bde (1)
2001-02-09 15:19:41 +00:00
/*
* First, handle sleep mutexes which have been acquired at least
* once.
*/
prnt("Sleep mutexes:\n");
Move most of sys/mutex.h into kern/kern_mutex.c, thereby making the mutex inline functions non-inlined. Hide parts of the mutex implementation that should not be exposed. Make sure that WITNESS code is not executed during boot until the mutexes are fully initialized by SI_SUB_MUTEX (the original motivation for this commit). Submitted by: peter
2001-01-21 22:34:43 +00:00
for (w = w_all; w; w = w->w_next) {
- Change the 'witness_list' ddb command to 'show mutexes'. Note that this will only display sleep mutexes held by the current process. - Clean up some nits in the witness_display() function and add a ddb command 'show witness' that dumps the hierarchy and order lists to the console. - Use queue(3) macros where appropriate. - Resort the spin lock order list so that "com" is before "sched_lock". Also, add appropriate #ifdef's around SMP and i386-specific mutexes. - Add two new mutexes used to protect the ithread lists and tables to the order list. Requested by: bde (1)
2001-02-09 15:19:41 +00:00
if (w->w_file == NULL || w->w_spin)
Move most of sys/mutex.h into kern/kern_mutex.c, thereby making the mutex inline functions non-inlined. Hide parts of the mutex implementation that should not be exposed. Make sure that WITNESS code is not executed during boot until the mutexes are fully initialized by SI_SUB_MUTEX (the original motivation for this commit). Submitted by: peter
2001-01-21 22:34:43 +00:00
continue;
for (w1 = w_all; w1; w1 = w1->w_next) {
if (isitmychild(w1, w))
break;
}
if (w1 != NULL)
continue;
/*
* This lock has no anscestors, display its descendants.
*/
witness_displaydescendants(prnt, w);
}
- Change the 'witness_list' ddb command to 'show mutexes'. Note that this will only display sleep mutexes held by the current process. - Clean up some nits in the witness_display() function and add a ddb command 'show witness' that dumps the hierarchy and order lists to the console. - Use queue(3) macros where appropriate. - Resort the spin lock order list so that "com" is before "sched_lock". Also, add appropriate #ifdef's around SMP and i386-specific mutexes. - Add two new mutexes used to protect the ithread lists and tables to the order list. Requested by: bde (1)
2001-02-09 15:19:41 +00:00
/*
* Now do spin mutexes which have been acquired at least once.
*/
prnt("\nSpin mutexes:\n");
level = 0;
while (level < sizeof(spin_order_list) / sizeof(char *)) {
found = 0;
for (w = w_all; w; w = w->w_next) {
if (w->w_file == NULL || !w->w_spin)
continue;
if (w->w_level == 1 << level) {
witness_displaydescendants(prnt, w);
level++;
found = 1;
}
}
if (found == 0)
level++;
}
/*
* Finally, any mutexes which have not been acquired yet.
*/
prnt("\nMutexes which were never acquired:\n");
Move most of sys/mutex.h into kern/kern_mutex.c, thereby making the mutex inline functions non-inlined. Hide parts of the mutex implementation that should not be exposed. Make sure that WITNESS code is not executed during boot until the mutexes are fully initialized by SI_SUB_MUTEX (the original motivation for this commit). Submitted by: peter
2001-01-21 22:34:43 +00:00
for (w = w_all; w; w = w->w_next) {
if (w->w_file != NULL)
continue;
prnt("%s\n", w->w_description);
}
}
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
void
Remove the mtx_t, witness_t, and witness_blessed_t types. Instead, just use struct mtx, struct witness, and struct witness_blessed. Requested by: bde
2000-09-14 20:15:16 +00:00
witness_enter(struct mtx *m, int flags, const char *file, int line)
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
{
Remove the mtx_t, witness_t, and witness_blessed_t types. Instead, just use struct mtx, struct witness, and struct witness_blessed. Requested by: bde
2000-09-14 20:15:16 +00:00
struct witness *w, *w1;
struct mtx *m1;
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
struct proc *p;
int i;
- Use MUTEX_DECLARE() and MTX_COLD for the WITNESS code's internal mutex so it can function before malloc(9) is up and running. - Add two new options WITNESS_DDB and WITNESS_SKIPSPIN. If WITNESS_SKIPSPIN is enabled, then spin mutexes are ignored by the WITNESS code. If WITNESS_DDB is turned on and DDB is compiled into the kernel, then the kernel will drop into DDB when either a lock hierarchy violation occurs or mutexes are held when going to sleep. - Add some new sysctls: debug.witness_ddb is a read-write sysctl that corresponds to WITNESS_DDB. The kernel option merely changes the default value to on at boot. debug.witness_skipspin is a read-only sysctl that one can use to determine if the kernel was compiled with WITNESS_SKIPSPIN. - Wipe out the BSD/OS-specific lock order lists. We get to build our own lists now as we add mutexes to the kernel.
2000-10-27 02:59:30 +00:00
#ifdef DDB
int go_into_ddb = 0;
#endif /* DDB */
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
Move most of sys/mutex.h into kern/kern_mutex.c, thereby making the mutex inline functions non-inlined. Hide parts of the mutex implementation that should not be exposed. Make sure that WITNESS code is not executed during boot until the mutexes are fully initialized by SI_SUB_MUTEX (the original motivation for this commit). Submitted by: peter
2001-01-21 22:34:43 +00:00
if (witness_cold || m->mtx_witness == NULL || panicstr)
- Add a new flag MTX_QUIET that can be passed to the various mtx_* functions. If this flag is set, then no KTR log messages are issued. This is useful for blocking excessive logging, such as with the internal mutex used by the witness code. - Use MTX_QUIET on all of the mtx_enter/exit operations on the internal mutex used by the witness code. - If we are in a panic, don't do witness checks in witness_enter(), witness_exit(), and witness_try_enter(), just return.
2000-12-13 21:53:42 +00:00
return;
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
w = m->mtx_witness;
Change all instances of `CURPROC' and `CURTHD' to `curproc,' in order to stay consistent. Requested by: bde
2001-02-12 03:15:43 +00:00
p = curproc;
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
if (flags & MTX_SPIN) {
- Don't use a union and fun tricks to shave one extra pointer off of struct mtx right now as it makes debugging harder. When we are in optimizing mode, we can revisit this. - Fix the KTR trace messages to use %p rather than 0x%p to avoid duplicate 0x's in KTR output. - During witness_fixup, release Giant so that witness doesn't get confused. Also, grab all_mtx while walking the list of mutexes. - Remove w_sleep and w_recurse. Instead, perform checks on mutexes using the mutex's mtx_flags field. - Allow debug.witness_ddb and debug.witness_skipspin to be set from the loader. - Add Giant to the front of existing order_list entries to help ensure Giant is always first. - Add an order entry for the various proc locks. Note that this only helps keep proc in order mostly as the allproc and proctree mutexes are only obtained during a lockmgr operation on the specified mutex.
2001-01-24 10:57:01 +00:00
if ((m->mtx_flags & MTX_SPIN) == 0)
Style cleanups. No functional changes.
2000-09-09 23:18:48 +00:00
panic("mutex_enter: MTX_SPIN on MTX_DEF mutex %s @"
" %s:%d", m->mtx_description, file, line);
Implement MTX_RECURSE flag for mtx_init(). All calls to mtx_init() for mutexes that recurse must now include the MTX_RECURSE bit in the flag argument variable. This change is in preparation for an upcoming (further) mutex API cleanup. The witness code will call panic() if a lock is found to recurse but the MTX_RECURSE bit was not set during the lock's initialization. The old MTX_RECURSE "state" bit (in mtx_lock) has been renamed to MTX_RECURSED, which is more appropriate given its meaning. The following locks have been made "recursive," thus far: eventhandler, Giant, callout, sched_lock, possibly some others declared in the architecture-specific code, all of the network card driver locks in pci/, as well as some other locks in dev/ stuff that I've found to be recursive. Reviewed by: jhb
2001-01-19 01:59:14 +00:00
if (mtx_recursed(m)) {
- Don't use a union and fun tricks to shave one extra pointer off of struct mtx right now as it makes debugging harder. When we are in optimizing mode, we can revisit this. - Fix the KTR trace messages to use %p rather than 0x%p to avoid duplicate 0x's in KTR output. - During witness_fixup, release Giant so that witness doesn't get confused. Also, grab all_mtx while walking the list of mutexes. - Remove w_sleep and w_recurse. Instead, perform checks on mutexes using the mutex's mtx_flags field. - Allow debug.witness_ddb and debug.witness_skipspin to be set from the loader. - Add Giant to the front of existing order_list entries to help ensure Giant is always first. - Add an order entry for the various proc locks. Note that this only helps keep proc in order mostly as the allproc and proctree mutexes are only obtained during a lockmgr operation on the specified mutex.
2001-01-24 10:57:01 +00:00
if ((m->mtx_flags & MTX_RECURSE) == 0)
Implement MTX_RECURSE flag for mtx_init(). All calls to mtx_init() for mutexes that recurse must now include the MTX_RECURSE bit in the flag argument variable. This change is in preparation for an upcoming (further) mutex API cleanup. The witness code will call panic() if a lock is found to recurse but the MTX_RECURSE bit was not set during the lock's initialization. The old MTX_RECURSE "state" bit (in mtx_lock) has been renamed to MTX_RECURSED, which is more appropriate given its meaning. The following locks have been made "recursive," thus far: eventhandler, Giant, callout, sched_lock, possibly some others declared in the architecture-specific code, all of the network card driver locks in pci/, as well as some other locks in dev/ stuff that I've found to be recursive. Reviewed by: jhb
2001-01-19 01:59:14 +00:00
panic("mutex_enter: recursion on non-recursive"
" mutex %s @ %s:%d", m->mtx_description,
file, line);
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
return;
Implement MTX_RECURSE flag for mtx_init(). All calls to mtx_init() for mutexes that recurse must now include the MTX_RECURSE bit in the flag argument variable. This change is in preparation for an upcoming (further) mutex API cleanup. The witness code will call panic() if a lock is found to recurse but the MTX_RECURSE bit was not set during the lock's initialization. The old MTX_RECURSE "state" bit (in mtx_lock) has been renamed to MTX_RECURSED, which is more appropriate given its meaning. The following locks have been made "recursive," thus far: eventhandler, Giant, callout, sched_lock, possibly some others declared in the architecture-specific code, all of the network card driver locks in pci/, as well as some other locks in dev/ stuff that I've found to be recursive. Reviewed by: jhb
2001-01-19 01:59:14 +00:00
}
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
mtx_lock_spin_flags(&w_mtx, MTX_QUIET);
Use PCPU_GET, PCPU_PTR and PCPU_SET to access all per-cpu variables other then curproc.
2001-01-10 04:43:51 +00:00
i = PCPU_GET(witness_spin_check);
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
if (i != 0 && w->w_level < i) {
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
mtx_unlock_spin_flags(&w_mtx, MTX_QUIET);
Style cleanups. No functional changes.
2000-09-09 23:18:48 +00:00
panic("mutex_enter(%s:%x, MTX_SPIN) out of order @"
" %s:%d already holding %s:%x",
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
m->mtx_description, w->w_level, file, line,
spin_order_list[ffs(i)-1], i);
}
PCPU_SET(witness_spin_check, i | w->w_level);
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
mtx_unlock_spin_flags(&w_mtx, MTX_QUIET);
Fix mtx_legal2block. The only time that it is bad to block on a mutex is if we hold a spin mutex, since we can trivially get into deadlocks if we start switching out of processes that hold spinlocks. Checking to see if interrupts were disabled was a sort of cheap way of doing this since most of the time interrupts were only disabled when holding a spin lock. At least on the i386. To fix this properly, use a per-process counter p_spinlocks that counts the number of spin locks currently held, and instead of checking to see if interrupts are disabled in the witness code, check to see if we hold any spin locks. Since child processes always start up with the sched lock magically held in fork_exit(), we initialize p_spinlocks to 1 for child processes. Note that proc0 doesn't go through fork_exit(), so it starts with no spin locks held. Consulting from: cp
2001-03-09 07:24:17 +00:00
p->p_spinlocks++;
MPASS(p->p_spinlocks > 0);
Save the line number and filename of the last mtx_enter operation for spin locks. We already do this for sleep locks.
2000-11-29 18:37:01 +00:00
w->w_file = file;
w->w_line = line;
m->mtx_line = line;
m->mtx_file = file;
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
return;
}
- Don't use a union and fun tricks to shave one extra pointer off of struct mtx right now as it makes debugging harder. When we are in optimizing mode, we can revisit this. - Fix the KTR trace messages to use %p rather than 0x%p to avoid duplicate 0x's in KTR output. - During witness_fixup, release Giant so that witness doesn't get confused. Also, grab all_mtx while walking the list of mutexes. - Remove w_sleep and w_recurse. Instead, perform checks on mutexes using the mutex's mtx_flags field. - Allow debug.witness_ddb and debug.witness_skipspin to be set from the loader. - Add Giant to the front of existing order_list entries to help ensure Giant is always first. - Add an order entry for the various proc locks. Note that this only helps keep proc in order mostly as the allproc and proctree mutexes are only obtained during a lockmgr operation on the specified mutex.
2001-01-24 10:57:01 +00:00
if ((m->mtx_flags & MTX_SPIN) != 0)
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
panic("mutex_enter: MTX_DEF on MTX_SPIN mutex %s @ %s:%d",
m->mtx_description, file, line);
Implement MTX_RECURSE flag for mtx_init(). All calls to mtx_init() for mutexes that recurse must now include the MTX_RECURSE bit in the flag argument variable. This change is in preparation for an upcoming (further) mutex API cleanup. The witness code will call panic() if a lock is found to recurse but the MTX_RECURSE bit was not set during the lock's initialization. The old MTX_RECURSE "state" bit (in mtx_lock) has been renamed to MTX_RECURSED, which is more appropriate given its meaning. The following locks have been made "recursive," thus far: eventhandler, Giant, callout, sched_lock, possibly some others declared in the architecture-specific code, all of the network card driver locks in pci/, as well as some other locks in dev/ stuff that I've found to be recursive. Reviewed by: jhb
2001-01-19 01:59:14 +00:00
if (mtx_recursed(m)) {
- Don't use a union and fun tricks to shave one extra pointer off of struct mtx right now as it makes debugging harder. When we are in optimizing mode, we can revisit this. - Fix the KTR trace messages to use %p rather than 0x%p to avoid duplicate 0x's in KTR output. - During witness_fixup, release Giant so that witness doesn't get confused. Also, grab all_mtx while walking the list of mutexes. - Remove w_sleep and w_recurse. Instead, perform checks on mutexes using the mutex's mtx_flags field. - Allow debug.witness_ddb and debug.witness_skipspin to be set from the loader. - Add Giant to the front of existing order_list entries to help ensure Giant is always first. - Add an order entry for the various proc locks. Note that this only helps keep proc in order mostly as the allproc and proctree mutexes are only obtained during a lockmgr operation on the specified mutex.
2001-01-24 10:57:01 +00:00
if ((m->mtx_flags & MTX_RECURSE) == 0)
Implement MTX_RECURSE flag for mtx_init(). All calls to mtx_init() for mutexes that recurse must now include the MTX_RECURSE bit in the flag argument variable. This change is in preparation for an upcoming (further) mutex API cleanup. The witness code will call panic() if a lock is found to recurse but the MTX_RECURSE bit was not set during the lock's initialization. The old MTX_RECURSE "state" bit (in mtx_lock) has been renamed to MTX_RECURSED, which is more appropriate given its meaning. The following locks have been made "recursive," thus far: eventhandler, Giant, callout, sched_lock, possibly some others declared in the architecture-specific code, all of the network card driver locks in pci/, as well as some other locks in dev/ stuff that I've found to be recursive. Reviewed by: jhb
2001-01-19 01:59:14 +00:00
panic("mutex_enter: recursion on non-recursive"
" mutex %s @ %s:%d", m->mtx_description,
file, line);
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
return;
Implement MTX_RECURSE flag for mtx_init(). All calls to mtx_init() for mutexes that recurse must now include the MTX_RECURSE bit in the flag argument variable. This change is in preparation for an upcoming (further) mutex API cleanup. The witness code will call panic() if a lock is found to recurse but the MTX_RECURSE bit was not set during the lock's initialization. The old MTX_RECURSE "state" bit (in mtx_lock) has been renamed to MTX_RECURSED, which is more appropriate given its meaning. The following locks have been made "recursive," thus far: eventhandler, Giant, callout, sched_lock, possibly some others declared in the architecture-specific code, all of the network card driver locks in pci/, as well as some other locks in dev/ stuff that I've found to be recursive. Reviewed by: jhb
2001-01-19 01:59:14 +00:00
}
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
if (witness_dead)
goto out;
- Add a new flag MTX_QUIET that can be passed to the various mtx_* functions. If this flag is set, then no KTR log messages are issued. This is useful for blocking excessive logging, such as with the internal mutex used by the witness code. - Use MTX_QUIET on all of the mtx_enter/exit operations on the internal mutex used by the witness code. - If we are in a panic, don't do witness checks in witness_enter(), witness_exit(), and witness_try_enter(), just return.
2000-12-13 21:53:42 +00:00
if (cold)
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
goto out;
Fix mtx_legal2block. The only time that it is bad to block on a mutex is if we hold a spin mutex, since we can trivially get into deadlocks if we start switching out of processes that hold spinlocks. Checking to see if interrupts were disabled was a sort of cheap way of doing this since most of the time interrupts were only disabled when holding a spin lock. At least on the i386. To fix this properly, use a per-process counter p_spinlocks that counts the number of spin locks currently held, and instead of checking to see if interrupts are disabled in the witness code, check to see if we hold any spin locks. Since child processes always start up with the sched lock magically held in fork_exit(), we initialize p_spinlocks to 1 for child processes. Note that proc0 doesn't go through fork_exit(), so it starts with no spin locks held. Consulting from: cp
2001-03-09 07:24:17 +00:00
if (p->p_spinlocks != 0)
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
panic("blockable mtx_lock() of %s when not legal @ %s:%d",
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
m->mtx_description, file, line);
/*
* Is this the first mutex acquired
*/
if ((m1 = LIST_FIRST(&p->p_heldmtx)) == NULL)
goto out;
if ((w1 = m1->mtx_witness) == w) {
if (w->w_same_squawked || dup_ok(w))
goto out;
w->w_same_squawked = 1;
printf("acquring duplicate lock of same type: \"%s\"\n",
m->mtx_description);
printf(" 1st @ %s:%d\n", w->w_file, w->w_line);
printf(" 2nd @ %s:%d\n", file, line);
- Use MUTEX_DECLARE() and MTX_COLD for the WITNESS code's internal mutex so it can function before malloc(9) is up and running. - Add two new options WITNESS_DDB and WITNESS_SKIPSPIN. If WITNESS_SKIPSPIN is enabled, then spin mutexes are ignored by the WITNESS code. If WITNESS_DDB is turned on and DDB is compiled into the kernel, then the kernel will drop into DDB when either a lock hierarchy violation occurs or mutexes are held when going to sleep. - Add some new sysctls: debug.witness_ddb is a read-write sysctl that corresponds to WITNESS_DDB. The kernel option merely changes the default value to on at boot. debug.witness_skipspin is a read-only sysctl that one can use to determine if the kernel was compiled with WITNESS_SKIPSPIN. - Wipe out the BSD/OS-specific lock order lists. We get to build our own lists now as we add mutexes to the kernel.
2000-10-27 02:59:30 +00:00
#ifdef DDB
go_into_ddb = 1;
#endif /* DDB */
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
goto out;
}
MPASS(!mtx_owned(&w_mtx));
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
mtx_lock_spin_flags(&w_mtx, MTX_QUIET);
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
/*
* If we have a known higher number just say ok
*/
if (witness_watch > 1 && w->w_level > w1->w_level) {
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
mtx_unlock_spin_flags(&w_mtx, MTX_QUIET);
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
goto out;
}
if (isitmydescendant(m1->mtx_witness, w)) {
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
mtx_unlock_spin_flags(&w_mtx, MTX_QUIET);
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
goto out;
}
for (i = 0; m1 != NULL; m1 = LIST_NEXT(m1, mtx_held), i++) {
- Make the mutex code almost completely machine independent. This greatly reducues the maintenance load for the mutex code. The only MD portions of the mutex code are in machine/mutex.h now, which include the assembly macros for handling mutexes as well as optionally overriding the mutex micro-operations. For example, we use optimized micro-ops on the x86 platform #ifndef I386_CPU. - Change the behavior of the SMP_DEBUG kernel option. In the new code, mtx_assert() only depends on INVARIANTS, allowing other kernel developers to have working mutex assertiions without having to include all of the mutex debugging code. The SMP_DEBUG kernel option has been renamed to MUTEX_DEBUG and now just controls extra mutex debugging code. - Abolish the ugly mtx_f hack. Instead, we dynamically allocate seperate mtx_debug structures on the fly in mtx_init, except for mutexes that are initiated very early in the boot process. These mutexes are declared using a special MUTEX_DECLARE() macro, and use a new flag MTX_COLD when calling mtx_init. This is still somewhat hackish, but it is less evil than the mtx_f filler struct, and the mtx struct is now the same size with and without mutex debugging code. - Add some micro-micro-operation macros for doing the actual atomic operations on the mutex mtx_lock field to make it easier for other archs to override/optimize mutex ops if needed. These new tiny ops also clean up the code in some places by replacing long atomic operation function calls that spanned 2-3 lines with a short 1-line macro call. - Don't call mi_switch() from mtx_enter_hard() when we block while trying to obtain a sleep mutex. Calling mi_switch() would bogusly release Giant before switching to the next process. Instead, inline most of the code from mi_switch() in the mtx_enter_hard() function. Note that when we finally kill Giant we can back this out and go back to calling mi_switch().
2000-10-20 07:26:37 +00:00
MPASS(i < 200);
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
w1 = m1->mtx_witness;
if (isitmydescendant(w, w1)) {
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
mtx_unlock_spin_flags(&w_mtx, MTX_QUIET);
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
if (blessed(w, w1))
goto out;
if (m1 == &Giant) {
if (w1->w_Giant_squawked)
goto out;
else
w1->w_Giant_squawked = 1;
} else {
if (w1->w_other_squawked)
goto out;
else
w1->w_other_squawked = 1;
}
printf("lock order reversal\n");
printf(" 1st %s last acquired @ %s:%d\n",
w->w_description, w->w_file, w->w_line);
printf(" 2nd %p %s @ %s:%d\n",
m1, w1->w_description, w1->w_file, w1->w_line);
printf(" 3rd %p %s @ %s:%d\n",
m, w->w_description, file, line);
- Use MUTEX_DECLARE() and MTX_COLD for the WITNESS code's internal mutex so it can function before malloc(9) is up and running. - Add two new options WITNESS_DDB and WITNESS_SKIPSPIN. If WITNESS_SKIPSPIN is enabled, then spin mutexes are ignored by the WITNESS code. If WITNESS_DDB is turned on and DDB is compiled into the kernel, then the kernel will drop into DDB when either a lock hierarchy violation occurs or mutexes are held when going to sleep. - Add some new sysctls: debug.witness_ddb is a read-write sysctl that corresponds to WITNESS_DDB. The kernel option merely changes the default value to on at boot. debug.witness_skipspin is a read-only sysctl that one can use to determine if the kernel was compiled with WITNESS_SKIPSPIN. - Wipe out the BSD/OS-specific lock order lists. We get to build our own lists now as we add mutexes to the kernel.
2000-10-27 02:59:30 +00:00
#ifdef DDB
go_into_ddb = 1;
#endif /* DDB */
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
goto out;
}
}
m1 = LIST_FIRST(&p->p_heldmtx);
if (!itismychild(m1->mtx_witness, w))
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
mtx_unlock_spin_flags(&w_mtx, MTX_QUIET);
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
out:
- Use MUTEX_DECLARE() and MTX_COLD for the WITNESS code's internal mutex so it can function before malloc(9) is up and running. - Add two new options WITNESS_DDB and WITNESS_SKIPSPIN. If WITNESS_SKIPSPIN is enabled, then spin mutexes are ignored by the WITNESS code. If WITNESS_DDB is turned on and DDB is compiled into the kernel, then the kernel will drop into DDB when either a lock hierarchy violation occurs or mutexes are held when going to sleep. - Add some new sysctls: debug.witness_ddb is a read-write sysctl that corresponds to WITNESS_DDB. The kernel option merely changes the default value to on at boot. debug.witness_skipspin is a read-only sysctl that one can use to determine if the kernel was compiled with WITNESS_SKIPSPIN. - Wipe out the BSD/OS-specific lock order lists. We get to build our own lists now as we add mutexes to the kernel.
2000-10-27 02:59:30 +00:00
#ifdef DDB
if (witness_ddb && go_into_ddb)
Debugger("witness_enter");
#endif /* DDB */
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
w->w_file = file;
w->w_line = line;
m->mtx_line = line;
m->mtx_file = file;
/*
Minor whitespace nit in a comment.
2000-11-10 21:21:20 +00:00
* If this pays off it likely means that a mutex being witnessed
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
* is acquired in hardclock. Put it in the ignore list. It is
* likely not the mutex this assert fails on.
*/
- Make the mutex code almost completely machine independent. This greatly reducues the maintenance load for the mutex code. The only MD portions of the mutex code are in machine/mutex.h now, which include the assembly macros for handling mutexes as well as optionally overriding the mutex micro-operations. For example, we use optimized micro-ops on the x86 platform #ifndef I386_CPU. - Change the behavior of the SMP_DEBUG kernel option. In the new code, mtx_assert() only depends on INVARIANTS, allowing other kernel developers to have working mutex assertiions without having to include all of the mutex debugging code. The SMP_DEBUG kernel option has been renamed to MUTEX_DEBUG and now just controls extra mutex debugging code. - Abolish the ugly mtx_f hack. Instead, we dynamically allocate seperate mtx_debug structures on the fly in mtx_init, except for mutexes that are initiated very early in the boot process. These mutexes are declared using a special MUTEX_DECLARE() macro, and use a new flag MTX_COLD when calling mtx_init. This is still somewhat hackish, but it is less evil than the mtx_f filler struct, and the mtx struct is now the same size with and without mutex debugging code. - Add some micro-micro-operation macros for doing the actual atomic operations on the mutex mtx_lock field to make it easier for other archs to override/optimize mutex ops if needed. These new tiny ops also clean up the code in some places by replacing long atomic operation function calls that spanned 2-3 lines with a short 1-line macro call. - Don't call mi_switch() from mtx_enter_hard() when we block while trying to obtain a sleep mutex. Calling mi_switch() would bogusly release Giant before switching to the next process. Instead, inline most of the code from mi_switch() in the mtx_enter_hard() function. Note that when we finally kill Giant we can back this out and go back to calling mi_switch().
2000-10-20 07:26:37 +00:00
MPASS(m->mtx_held.le_prev == NULL);
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
LIST_INSERT_HEAD(&p->p_heldmtx, (struct mtx*)m, mtx_held);
}
void
Remove the mtx_t, witness_t, and witness_blessed_t types. Instead, just use struct mtx, struct witness, and struct witness_blessed. Requested by: bde
2000-09-14 20:15:16 +00:00
witness_try_enter(struct mtx *m, int flags, const char *file, int line)
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
{
struct proc *p;
Remove the mtx_t, witness_t, and witness_blessed_t types. Instead, just use struct mtx, struct witness, and struct witness_blessed. Requested by: bde
2000-09-14 20:15:16 +00:00
struct witness *w = m->mtx_witness;
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
Remove MUTEX_DECLARE() and MTX_COLD. Instead, postpone full mutex initialization until after malloc() is safe to call, then iterate through all mutexes and complete their initialization. This change is necessary in order to avoid some circular bootstrapping dependencies.
2001-01-21 07:52:20 +00:00
if (witness_cold)
return;
- Add a new flag MTX_QUIET that can be passed to the various mtx_* functions. If this flag is set, then no KTR log messages are issued. This is useful for blocking excessive logging, such as with the internal mutex used by the witness code. - Use MTX_QUIET on all of the mtx_enter/exit operations on the internal mutex used by the witness code. - If we are in a panic, don't do witness checks in witness_enter(), witness_exit(), and witness_try_enter(), just return.
2000-12-13 21:53:42 +00:00
if (panicstr)
return;
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
if (flags & MTX_SPIN) {
- Don't use a union and fun tricks to shave one extra pointer off of struct mtx right now as it makes debugging harder. When we are in optimizing mode, we can revisit this. - Fix the KTR trace messages to use %p rather than 0x%p to avoid duplicate 0x's in KTR output. - During witness_fixup, release Giant so that witness doesn't get confused. Also, grab all_mtx while walking the list of mutexes. - Remove w_sleep and w_recurse. Instead, perform checks on mutexes using the mutex's mtx_flags field. - Allow debug.witness_ddb and debug.witness_skipspin to be set from the loader. - Add Giant to the front of existing order_list entries to help ensure Giant is always first. - Add an order entry for the various proc locks. Note that this only helps keep proc in order mostly as the allproc and proctree mutexes are only obtained during a lockmgr operation on the specified mutex.
2001-01-24 10:57:01 +00:00
if ((m->mtx_flags & MTX_SPIN) == 0)
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
panic("mutex_try_enter: "
"MTX_SPIN on MTX_DEF mutex %s @ %s:%d",
m->mtx_description, file, line);
Implement MTX_RECURSE flag for mtx_init(). All calls to mtx_init() for mutexes that recurse must now include the MTX_RECURSE bit in the flag argument variable. This change is in preparation for an upcoming (further) mutex API cleanup. The witness code will call panic() if a lock is found to recurse but the MTX_RECURSE bit was not set during the lock's initialization. The old MTX_RECURSE "state" bit (in mtx_lock) has been renamed to MTX_RECURSED, which is more appropriate given its meaning. The following locks have been made "recursive," thus far: eventhandler, Giant, callout, sched_lock, possibly some others declared in the architecture-specific code, all of the network card driver locks in pci/, as well as some other locks in dev/ stuff that I've found to be recursive. Reviewed by: jhb
2001-01-19 01:59:14 +00:00
if (mtx_recursed(m)) {
- Don't use a union and fun tricks to shave one extra pointer off of struct mtx right now as it makes debugging harder. When we are in optimizing mode, we can revisit this. - Fix the KTR trace messages to use %p rather than 0x%p to avoid duplicate 0x's in KTR output. - During witness_fixup, release Giant so that witness doesn't get confused. Also, grab all_mtx while walking the list of mutexes. - Remove w_sleep and w_recurse. Instead, perform checks on mutexes using the mutex's mtx_flags field. - Allow debug.witness_ddb and debug.witness_skipspin to be set from the loader. - Add Giant to the front of existing order_list entries to help ensure Giant is always first. - Add an order entry for the various proc locks. Note that this only helps keep proc in order mostly as the allproc and proctree mutexes are only obtained during a lockmgr operation on the specified mutex.
2001-01-24 10:57:01 +00:00
if ((m->mtx_flags & MTX_RECURSE) == 0)
Implement MTX_RECURSE flag for mtx_init(). All calls to mtx_init() for mutexes that recurse must now include the MTX_RECURSE bit in the flag argument variable. This change is in preparation for an upcoming (further) mutex API cleanup. The witness code will call panic() if a lock is found to recurse but the MTX_RECURSE bit was not set during the lock's initialization. The old MTX_RECURSE "state" bit (in mtx_lock) has been renamed to MTX_RECURSED, which is more appropriate given its meaning. The following locks have been made "recursive," thus far: eventhandler, Giant, callout, sched_lock, possibly some others declared in the architecture-specific code, all of the network card driver locks in pci/, as well as some other locks in dev/ stuff that I've found to be recursive. Reviewed by: jhb
2001-01-19 01:59:14 +00:00
panic("mutex_try_enter: recursion on"
" non-recursive mutex %s @ %s:%d",
m->mtx_description, file, line);
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
return;
Implement MTX_RECURSE flag for mtx_init(). All calls to mtx_init() for mutexes that recurse must now include the MTX_RECURSE bit in the flag argument variable. This change is in preparation for an upcoming (further) mutex API cleanup. The witness code will call panic() if a lock is found to recurse but the MTX_RECURSE bit was not set during the lock's initialization. The old MTX_RECURSE "state" bit (in mtx_lock) has been renamed to MTX_RECURSED, which is more appropriate given its meaning. The following locks have been made "recursive," thus far: eventhandler, Giant, callout, sched_lock, possibly some others declared in the architecture-specific code, all of the network card driver locks in pci/, as well as some other locks in dev/ stuff that I've found to be recursive. Reviewed by: jhb
2001-01-19 01:59:14 +00:00
}
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
mtx_lock_spin_flags(&w_mtx, MTX_QUIET);
Use PCPU_GET, PCPU_PTR and PCPU_SET to access all per-cpu variables other then curproc.
2001-01-10 04:43:51 +00:00
PCPU_SET(witness_spin_check,
PCPU_GET(witness_spin_check) | w->w_level);
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
mtx_unlock_spin_flags(&w_mtx, MTX_QUIET);
Save the line number and filename of the last mtx_enter operation for spin locks. We already do this for sleep locks.
2000-11-29 18:37:01 +00:00
w->w_file = file;
w->w_line = line;
m->mtx_line = line;
m->mtx_file = file;
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
return;
}
- Don't use a union and fun tricks to shave one extra pointer off of struct mtx right now as it makes debugging harder. When we are in optimizing mode, we can revisit this. - Fix the KTR trace messages to use %p rather than 0x%p to avoid duplicate 0x's in KTR output. - During witness_fixup, release Giant so that witness doesn't get confused. Also, grab all_mtx while walking the list of mutexes. - Remove w_sleep and w_recurse. Instead, perform checks on mutexes using the mutex's mtx_flags field. - Allow debug.witness_ddb and debug.witness_skipspin to be set from the loader. - Add Giant to the front of existing order_list entries to help ensure Giant is always first. - Add an order entry for the various proc locks. Note that this only helps keep proc in order mostly as the allproc and proctree mutexes are only obtained during a lockmgr operation on the specified mutex.
2001-01-24 10:57:01 +00:00
if ((m->mtx_flags & MTX_SPIN) != 0)
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
panic("mutex_try_enter: MTX_DEF on MTX_SPIN mutex %s @ %s:%d",
m->mtx_description, file, line);
Implement MTX_RECURSE flag for mtx_init(). All calls to mtx_init() for mutexes that recurse must now include the MTX_RECURSE bit in the flag argument variable. This change is in preparation for an upcoming (further) mutex API cleanup. The witness code will call panic() if a lock is found to recurse but the MTX_RECURSE bit was not set during the lock's initialization. The old MTX_RECURSE "state" bit (in mtx_lock) has been renamed to MTX_RECURSED, which is more appropriate given its meaning. The following locks have been made "recursive," thus far: eventhandler, Giant, callout, sched_lock, possibly some others declared in the architecture-specific code, all of the network card driver locks in pci/, as well as some other locks in dev/ stuff that I've found to be recursive. Reviewed by: jhb
2001-01-19 01:59:14 +00:00
if (mtx_recursed(m)) {
- Don't use a union and fun tricks to shave one extra pointer off of struct mtx right now as it makes debugging harder. When we are in optimizing mode, we can revisit this. - Fix the KTR trace messages to use %p rather than 0x%p to avoid duplicate 0x's in KTR output. - During witness_fixup, release Giant so that witness doesn't get confused. Also, grab all_mtx while walking the list of mutexes. - Remove w_sleep and w_recurse. Instead, perform checks on mutexes using the mutex's mtx_flags field. - Allow debug.witness_ddb and debug.witness_skipspin to be set from the loader. - Add Giant to the front of existing order_list entries to help ensure Giant is always first. - Add an order entry for the various proc locks. Note that this only helps keep proc in order mostly as the allproc and proctree mutexes are only obtained during a lockmgr operation on the specified mutex.
2001-01-24 10:57:01 +00:00
if ((m->mtx_flags & MTX_RECURSE) == 0)
Implement MTX_RECURSE flag for mtx_init(). All calls to mtx_init() for mutexes that recurse must now include the MTX_RECURSE bit in the flag argument variable. This change is in preparation for an upcoming (further) mutex API cleanup. The witness code will call panic() if a lock is found to recurse but the MTX_RECURSE bit was not set during the lock's initialization. The old MTX_RECURSE "state" bit (in mtx_lock) has been renamed to MTX_RECURSED, which is more appropriate given its meaning. The following locks have been made "recursive," thus far: eventhandler, Giant, callout, sched_lock, possibly some others declared in the architecture-specific code, all of the network card driver locks in pci/, as well as some other locks in dev/ stuff that I've found to be recursive. Reviewed by: jhb
2001-01-19 01:59:14 +00:00
panic("mutex_try_enter: recursion on non-recursive"
" mutex %s @ %s:%d", m->mtx_description, file,
line);
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
return;
Implement MTX_RECURSE flag for mtx_init(). All calls to mtx_init() for mutexes that recurse must now include the MTX_RECURSE bit in the flag argument variable. This change is in preparation for an upcoming (further) mutex API cleanup. The witness code will call panic() if a lock is found to recurse but the MTX_RECURSE bit was not set during the lock's initialization. The old MTX_RECURSE "state" bit (in mtx_lock) has been renamed to MTX_RECURSED, which is more appropriate given its meaning. The following locks have been made "recursive," thus far: eventhandler, Giant, callout, sched_lock, possibly some others declared in the architecture-specific code, all of the network card driver locks in pci/, as well as some other locks in dev/ stuff that I've found to be recursive. Reviewed by: jhb
2001-01-19 01:59:14 +00:00
}
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
w->w_file = file;
w->w_line = line;
m->mtx_line = line;
m->mtx_file = file;
Change all instances of `CURPROC' and `CURTHD' to `curproc,' in order to stay consistent. Requested by: bde
2001-02-12 03:15:43 +00:00
p = curproc;
- Make the mutex code almost completely machine independent. This greatly reducues the maintenance load for the mutex code. The only MD portions of the mutex code are in machine/mutex.h now, which include the assembly macros for handling mutexes as well as optionally overriding the mutex micro-operations. For example, we use optimized micro-ops on the x86 platform #ifndef I386_CPU. - Change the behavior of the SMP_DEBUG kernel option. In the new code, mtx_assert() only depends on INVARIANTS, allowing other kernel developers to have working mutex assertiions without having to include all of the mutex debugging code. The SMP_DEBUG kernel option has been renamed to MUTEX_DEBUG and now just controls extra mutex debugging code. - Abolish the ugly mtx_f hack. Instead, we dynamically allocate seperate mtx_debug structures on the fly in mtx_init, except for mutexes that are initiated very early in the boot process. These mutexes are declared using a special MUTEX_DECLARE() macro, and use a new flag MTX_COLD when calling mtx_init. This is still somewhat hackish, but it is less evil than the mtx_f filler struct, and the mtx struct is now the same size with and without mutex debugging code. - Add some micro-micro-operation macros for doing the actual atomic operations on the mutex mtx_lock field to make it easier for other archs to override/optimize mutex ops if needed. These new tiny ops also clean up the code in some places by replacing long atomic operation function calls that spanned 2-3 lines with a short 1-line macro call. - Don't call mi_switch() from mtx_enter_hard() when we block while trying to obtain a sleep mutex. Calling mi_switch() would bogusly release Giant before switching to the next process. Instead, inline most of the code from mi_switch() in the mtx_enter_hard() function. Note that when we finally kill Giant we can back this out and go back to calling mi_switch().
2000-10-20 07:26:37 +00:00
MPASS(m->mtx_held.le_prev == NULL);
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
LIST_INSERT_HEAD(&p->p_heldmtx, (struct mtx*)m, mtx_held);
}
void
Move most of sys/mutex.h into kern/kern_mutex.c, thereby making the mutex inline functions non-inlined. Hide parts of the mutex implementation that should not be exposed. Make sure that WITNESS code is not executed during boot until the mutexes are fully initialized by SI_SUB_MUTEX (the original motivation for this commit). Submitted by: peter
2001-01-21 22:34:43 +00:00
witness_exit(struct mtx *m, int flags, const char *file, int line)
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
{
Move most of sys/mutex.h into kern/kern_mutex.c, thereby making the mutex inline functions non-inlined. Hide parts of the mutex implementation that should not be exposed. Make sure that WITNESS code is not executed during boot until the mutexes are fully initialized by SI_SUB_MUTEX (the original motivation for this commit). Submitted by: peter
2001-01-21 22:34:43 +00:00
struct witness *w;
Fix mtx_legal2block. The only time that it is bad to block on a mutex is if we hold a spin mutex, since we can trivially get into deadlocks if we start switching out of processes that hold spinlocks. Checking to see if interrupts were disabled was a sort of cheap way of doing this since most of the time interrupts were only disabled when holding a spin lock. At least on the i386. To fix this properly, use a per-process counter p_spinlocks that counts the number of spin locks currently held, and instead of checking to see if interrupts are disabled in the witness code, check to see if we hold any spin locks. Since child processes always start up with the sched lock magically held in fork_exit(), we initialize p_spinlocks to 1 for child processes. Note that proc0 doesn't go through fork_exit(), so it starts with no spin locks held. Consulting from: cp
2001-03-09 07:24:17 +00:00
struct proc *p;
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
Move most of sys/mutex.h into kern/kern_mutex.c, thereby making the mutex inline functions non-inlined. Hide parts of the mutex implementation that should not be exposed. Make sure that WITNESS code is not executed during boot until the mutexes are fully initialized by SI_SUB_MUTEX (the original motivation for this commit). Submitted by: peter
2001-01-21 22:34:43 +00:00
if (witness_cold || m->mtx_witness == NULL || panicstr)
return;
w = m->mtx_witness;
Fix mtx_legal2block. The only time that it is bad to block on a mutex is if we hold a spin mutex, since we can trivially get into deadlocks if we start switching out of processes that hold spinlocks. Checking to see if interrupts were disabled was a sort of cheap way of doing this since most of the time interrupts were only disabled when holding a spin lock. At least on the i386. To fix this properly, use a per-process counter p_spinlocks that counts the number of spin locks currently held, and instead of checking to see if interrupts are disabled in the witness code, check to see if we hold any spin locks. Since child processes always start up with the sched lock magically held in fork_exit(), we initialize p_spinlocks to 1 for child processes. Note that proc0 doesn't go through fork_exit(), so it starts with no spin locks held. Consulting from: cp
2001-03-09 07:24:17 +00:00
p = curproc;
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
Move most of sys/mutex.h into kern/kern_mutex.c, thereby making the mutex inline functions non-inlined. Hide parts of the mutex implementation that should not be exposed. Make sure that WITNESS code is not executed during boot until the mutexes are fully initialized by SI_SUB_MUTEX (the original motivation for this commit). Submitted by: peter
2001-01-21 22:34:43 +00:00
if (flags & MTX_SPIN) {
- Don't use a union and fun tricks to shave one extra pointer off of struct mtx right now as it makes debugging harder. When we are in optimizing mode, we can revisit this. - Fix the KTR trace messages to use %p rather than 0x%p to avoid duplicate 0x's in KTR output. - During witness_fixup, release Giant so that witness doesn't get confused. Also, grab all_mtx while walking the list of mutexes. - Remove w_sleep and w_recurse. Instead, perform checks on mutexes using the mutex's mtx_flags field. - Allow debug.witness_ddb and debug.witness_skipspin to be set from the loader. - Add Giant to the front of existing order_list entries to help ensure Giant is always first. - Add an order entry for the various proc locks. Note that this only helps keep proc in order mostly as the allproc and proctree mutexes are only obtained during a lockmgr operation on the specified mutex.
2001-01-24 10:57:01 +00:00
if ((m->mtx_flags & MTX_SPIN) == 0)
Move most of sys/mutex.h into kern/kern_mutex.c, thereby making the mutex inline functions non-inlined. Hide parts of the mutex implementation that should not be exposed. Make sure that WITNESS code is not executed during boot until the mutexes are fully initialized by SI_SUB_MUTEX (the original motivation for this commit). Submitted by: peter
2001-01-21 22:34:43 +00:00
panic("mutex_exit: MTX_SPIN on MTX_DEF mutex %s @"
" %s:%d", m->mtx_description, file, line);
if (mtx_recursed(m)) {
- Don't use a union and fun tricks to shave one extra pointer off of struct mtx right now as it makes debugging harder. When we are in optimizing mode, we can revisit this. - Fix the KTR trace messages to use %p rather than 0x%p to avoid duplicate 0x's in KTR output. - During witness_fixup, release Giant so that witness doesn't get confused. Also, grab all_mtx while walking the list of mutexes. - Remove w_sleep and w_recurse. Instead, perform checks on mutexes using the mutex's mtx_flags field. - Allow debug.witness_ddb and debug.witness_skipspin to be set from the loader. - Add Giant to the front of existing order_list entries to help ensure Giant is always first. - Add an order entry for the various proc locks. Note that this only helps keep proc in order mostly as the allproc and proctree mutexes are only obtained during a lockmgr operation on the specified mutex.
2001-01-24 10:57:01 +00:00
if ((m->mtx_flags & MTX_RECURSE) == 0)
Move most of sys/mutex.h into kern/kern_mutex.c, thereby making the mutex inline functions non-inlined. Hide parts of the mutex implementation that should not be exposed. Make sure that WITNESS code is not executed during boot until the mutexes are fully initialized by SI_SUB_MUTEX (the original motivation for this commit). Submitted by: peter
2001-01-21 22:34:43 +00:00
panic("mutex_exit: recursion on non-recursive"
" mutex %s @ %s:%d", m->mtx_description,
file, line);
return;
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
}
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
mtx_lock_spin_flags(&w_mtx, MTX_QUIET);
Move most of sys/mutex.h into kern/kern_mutex.c, thereby making the mutex inline functions non-inlined. Hide parts of the mutex implementation that should not be exposed. Make sure that WITNESS code is not executed during boot until the mutexes are fully initialized by SI_SUB_MUTEX (the original motivation for this commit). Submitted by: peter
2001-01-21 22:34:43 +00:00
PCPU_SET(witness_spin_check,
PCPU_GET(witness_spin_check) & ~w->w_level);
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
mtx_unlock_spin_flags(&w_mtx, MTX_QUIET);
Fix mtx_legal2block. The only time that it is bad to block on a mutex is if we hold a spin mutex, since we can trivially get into deadlocks if we start switching out of processes that hold spinlocks. Checking to see if interrupts were disabled was a sort of cheap way of doing this since most of the time interrupts were only disabled when holding a spin lock. At least on the i386. To fix this properly, use a per-process counter p_spinlocks that counts the number of spin locks currently held, and instead of checking to see if interrupts are disabled in the witness code, check to see if we hold any spin locks. Since child processes always start up with the sched lock magically held in fork_exit(), we initialize p_spinlocks to 1 for child processes. Note that proc0 doesn't go through fork_exit(), so it starts with no spin locks held. Consulting from: cp
2001-03-09 07:24:17 +00:00
MPASS(p->p_spinlocks > 0);
p->p_spinlocks--;
Move most of sys/mutex.h into kern/kern_mutex.c, thereby making the mutex inline functions non-inlined. Hide parts of the mutex implementation that should not be exposed. Make sure that WITNESS code is not executed during boot until the mutexes are fully initialized by SI_SUB_MUTEX (the original motivation for this commit). Submitted by: peter
2001-01-21 22:34:43 +00:00
return;
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
}
- Don't use a union and fun tricks to shave one extra pointer off of struct mtx right now as it makes debugging harder. When we are in optimizing mode, we can revisit this. - Fix the KTR trace messages to use %p rather than 0x%p to avoid duplicate 0x's in KTR output. - During witness_fixup, release Giant so that witness doesn't get confused. Also, grab all_mtx while walking the list of mutexes. - Remove w_sleep and w_recurse. Instead, perform checks on mutexes using the mutex's mtx_flags field. - Allow debug.witness_ddb and debug.witness_skipspin to be set from the loader. - Add Giant to the front of existing order_list entries to help ensure Giant is always first. - Add an order entry for the various proc locks. Note that this only helps keep proc in order mostly as the allproc and proctree mutexes are only obtained during a lockmgr operation on the specified mutex.
2001-01-24 10:57:01 +00:00
if ((m->mtx_flags & MTX_SPIN) != 0)
Move most of sys/mutex.h into kern/kern_mutex.c, thereby making the mutex inline functions non-inlined. Hide parts of the mutex implementation that should not be exposed. Make sure that WITNESS code is not executed during boot until the mutexes are fully initialized by SI_SUB_MUTEX (the original motivation for this commit). Submitted by: peter
2001-01-21 22:34:43 +00:00
panic("mutex_exit: MTX_DEF on MTX_SPIN mutex %s @ %s:%d",
m->mtx_description, file, line);
if (mtx_recursed(m)) {
- Don't use a union and fun tricks to shave one extra pointer off of struct mtx right now as it makes debugging harder. When we are in optimizing mode, we can revisit this. - Fix the KTR trace messages to use %p rather than 0x%p to avoid duplicate 0x's in KTR output. - During witness_fixup, release Giant so that witness doesn't get confused. Also, grab all_mtx while walking the list of mutexes. - Remove w_sleep and w_recurse. Instead, perform checks on mutexes using the mutex's mtx_flags field. - Allow debug.witness_ddb and debug.witness_skipspin to be set from the loader. - Add Giant to the front of existing order_list entries to help ensure Giant is always first. - Add an order entry for the various proc locks. Note that this only helps keep proc in order mostly as the allproc and proctree mutexes are only obtained during a lockmgr operation on the specified mutex.
2001-01-24 10:57:01 +00:00
if ((m->mtx_flags & MTX_RECURSE) == 0)
Move most of sys/mutex.h into kern/kern_mutex.c, thereby making the mutex inline functions non-inlined. Hide parts of the mutex implementation that should not be exposed. Make sure that WITNESS code is not executed during boot until the mutexes are fully initialized by SI_SUB_MUTEX (the original motivation for this commit). Submitted by: peter
2001-01-21 22:34:43 +00:00
panic("mutex_exit: recursion on non-recursive"
" mutex %s @ %s:%d", m->mtx_description,
file, line);
return;
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
}
Move most of sys/mutex.h into kern/kern_mutex.c, thereby making the mutex inline functions non-inlined. Hide parts of the mutex implementation that should not be exposed. Make sure that WITNESS code is not executed during boot until the mutexes are fully initialized by SI_SUB_MUTEX (the original motivation for this commit). Submitted by: peter
2001-01-21 22:34:43 +00:00
Fix mtx_legal2block. The only time that it is bad to block on a mutex is if we hold a spin mutex, since we can trivially get into deadlocks if we start switching out of processes that hold spinlocks. Checking to see if interrupts were disabled was a sort of cheap way of doing this since most of the time interrupts were only disabled when holding a spin lock. At least on the i386. To fix this properly, use a per-process counter p_spinlocks that counts the number of spin locks currently held, and instead of checking to see if interrupts are disabled in the witness code, check to see if we hold any spin locks. Since child processes always start up with the sched lock magically held in fork_exit(), we initialize p_spinlocks to 1 for child processes. Note that proc0 doesn't go through fork_exit(), so it starts with no spin locks held. Consulting from: cp
2001-03-09 07:24:17 +00:00
if ((flags & MTX_NOSWITCH) == 0 && p->p_spinlocks != 0 && !cold)
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
panic("switchable mtx_unlock() of %s when not legal @ %s:%d",
Move most of sys/mutex.h into kern/kern_mutex.c, thereby making the mutex inline functions non-inlined. Hide parts of the mutex implementation that should not be exposed. Make sure that WITNESS code is not executed during boot until the mutexes are fully initialized by SI_SUB_MUTEX (the original motivation for this commit). Submitted by: peter
2001-01-21 22:34:43 +00:00
m->mtx_description, file, line);
LIST_REMOVE(m, mtx_held);
m->mtx_held.le_prev = NULL;
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
}
int
Remove the mtx_t, witness_t, and witness_blessed_t types. Instead, just use struct mtx, struct witness, and struct witness_blessed. Requested by: bde
2000-09-14 20:15:16 +00:00
witness_sleep(int check_only, struct mtx *mtx, const char *file, int line)
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
{
Remove the mtx_t, witness_t, and witness_blessed_t types. Instead, just use struct mtx, struct witness, and struct witness_blessed. Requested by: bde
2000-09-14 20:15:16 +00:00
struct mtx *m;
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
struct proc *p;
char **sleep;
int n = 0;
Remove MUTEX_DECLARE() and MTX_COLD. Instead, postpone full mutex initialization until after malloc() is safe to call, then iterate through all mutexes and complete their initialization. This change is necessary in order to avoid some circular bootstrapping dependencies.
2001-01-21 07:52:20 +00:00
KASSERT(!witness_cold, ("%s: witness_cold\n", __FUNCTION__));
Change all instances of `CURPROC' and `CURTHD' to `curproc,' in order to stay consistent. Requested by: bde
2001-02-12 03:15:43 +00:00
p = curproc;
- Change the 'witness_list' ddb command to 'show mutexes'. Note that this will only display sleep mutexes held by the current process. - Clean up some nits in the witness_display() function and add a ddb command 'show witness' that dumps the hierarchy and order lists to the console. - Use queue(3) macros where appropriate. - Resort the spin lock order list so that "com" is before "sched_lock". Also, add appropriate #ifdef's around SMP and i386-specific mutexes. - Add two new mutexes used to protect the ithread lists and tables to the order list. Requested by: bde (1)
2001-02-09 15:19:41 +00:00
LIST_FOREACH(m, &p->p_heldmtx, mtx_held) {
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
if (m == mtx)
continue;
for (sleep = sleep_list; *sleep!= NULL; sleep++)
if (strcmp(m->mtx_description, *sleep) == 0)
goto next;
- Change the 'witness_list' ddb command to 'show mutexes'. Note that this will only display sleep mutexes held by the current process. - Clean up some nits in the witness_display() function and add a ddb command 'show witness' that dumps the hierarchy and order lists to the console. - Use queue(3) macros where appropriate. - Resort the spin lock order list so that "com" is before "sched_lock". Also, add appropriate #ifdef's around SMP and i386-specific mutexes. - Add two new mutexes used to protect the ithread lists and tables to the order list. Requested by: bde (1)
2001-02-09 15:19:41 +00:00
if (n == 0)
printf("Whee!\n");
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
printf("%s:%d: %s with \"%s\" locked from %s:%d\n",
file, line, check_only ? "could sleep" : "sleeping",
m->mtx_description,
m->mtx_witness->w_file, m->mtx_witness->w_line);
n++;
next:
}
- Use MUTEX_DECLARE() and MTX_COLD for the WITNESS code's internal mutex so it can function before malloc(9) is up and running. - Add two new options WITNESS_DDB and WITNESS_SKIPSPIN. If WITNESS_SKIPSPIN is enabled, then spin mutexes are ignored by the WITNESS code. If WITNESS_DDB is turned on and DDB is compiled into the kernel, then the kernel will drop into DDB when either a lock hierarchy violation occurs or mutexes are held when going to sleep. - Add some new sysctls: debug.witness_ddb is a read-write sysctl that corresponds to WITNESS_DDB. The kernel option merely changes the default value to on at boot. debug.witness_skipspin is a read-only sysctl that one can use to determine if the kernel was compiled with WITNESS_SKIPSPIN. - Wipe out the BSD/OS-specific lock order lists. We get to build our own lists now as we add mutexes to the kernel.
2000-10-27 02:59:30 +00:00
#ifdef DDB
if (witness_ddb && n)
Debugger("witness_sleep");
#endif /* DDB */
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
return (n);
}
Remove the mtx_t, witness_t, and witness_blessed_t types. Instead, just use struct mtx, struct witness, and struct witness_blessed. Requested by: bde
2000-09-14 20:15:16 +00:00
static struct witness *
Propogate the 'const'ness of mutex descriptions to the witness code to quiet warnings.
2000-10-20 22:45:01 +00:00
enroll(const char *description, int flag)
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
{
int i;
Remove the mtx_t, witness_t, and witness_blessed_t types. Instead, just use struct mtx, struct witness, and struct witness_blessed. Requested by: bde
2000-09-14 20:15:16 +00:00
struct witness *w, *w1;
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
char **ignore;
char **order;
if (!witness_watch)
return (NULL);
for (ignore = ignore_list; *ignore != NULL; ignore++)
if (strcmp(description, *ignore) == 0)
return (NULL);
if (w_inited == 0) {
Remove MUTEX_DECLARE() and MTX_COLD. Instead, postpone full mutex initialization until after malloc() is safe to call, then iterate through all mutexes and complete their initialization. This change is necessary in order to avoid some circular bootstrapping dependencies.
2001-01-21 07:52:20 +00:00
mtx_init(&w_mtx, "witness lock", MTX_SPIN);
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
for (i = 0; i < WITNESS_COUNT; i++) {
w = &w_data[i];
witness_free(w);
}
w_inited = 1;
for (order = order_list; *order != NULL; order++) {
w = enroll(*order, MTX_DEF);
w->w_file = "order list";
for (order++; *order != NULL; order++) {
w1 = enroll(*order, MTX_DEF);
w1->w_file = "order list";
itismychild(w, w1);
w = w1;
}
}
}
if ((flag & MTX_SPIN) && witness_skipspin)
return (NULL);
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
mtx_lock_spin_flags(&w_mtx, MTX_QUIET);
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
for (w = w_all; w; w = w->w_next) {
if (strcmp(description, w->w_description) == 0) {
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
mtx_unlock_spin_flags(&w_mtx, MTX_QUIET);
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
return (w);
}
}
if ((w = witness_get()) == NULL)
return (NULL);
w->w_next = w_all;
w_all = w;
w->w_description = description;
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
mtx_unlock_spin_flags(&w_mtx, MTX_QUIET);
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
if (flag & MTX_SPIN) {
w->w_spin = 1;
i = 1;
for (order = spin_order_list; *order != NULL; order++) {
if (strcmp(description, *order) == 0)
break;
i <<= 1;
}
if (*order == NULL)
panic("spin lock %s not in order list", description);
w->w_level = i;
- Don't use a union and fun tricks to shave one extra pointer off of struct mtx right now as it makes debugging harder. When we are in optimizing mode, we can revisit this. - Fix the KTR trace messages to use %p rather than 0x%p to avoid duplicate 0x's in KTR output. - During witness_fixup, release Giant so that witness doesn't get confused. Also, grab all_mtx while walking the list of mutexes. - Remove w_sleep and w_recurse. Instead, perform checks on mutexes using the mutex's mtx_flags field. - Allow debug.witness_ddb and debug.witness_skipspin to be set from the loader. - Add Giant to the front of existing order_list entries to help ensure Giant is always first. - Add an order entry for the various proc locks. Note that this only helps keep proc in order mostly as the allproc and proctree mutexes are only obtained during a lockmgr operation on the specified mutex.
2001-01-24 10:57:01 +00:00
}
Implement MTX_RECURSE flag for mtx_init(). All calls to mtx_init() for mutexes that recurse must now include the MTX_RECURSE bit in the flag argument variable. This change is in preparation for an upcoming (further) mutex API cleanup. The witness code will call panic() if a lock is found to recurse but the MTX_RECURSE bit was not set during the lock's initialization. The old MTX_RECURSE "state" bit (in mtx_lock) has been renamed to MTX_RECURSED, which is more appropriate given its meaning. The following locks have been made "recursive," thus far: eventhandler, Giant, callout, sched_lock, possibly some others declared in the architecture-specific code, all of the network card driver locks in pci/, as well as some other locks in dev/ stuff that I've found to be recursive. Reviewed by: jhb
2001-01-19 01:59:14 +00:00
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
return (w);
}
static int
Remove the mtx_t, witness_t, and witness_blessed_t types. Instead, just use struct mtx, struct witness, and struct witness_blessed. Requested by: bde
2000-09-14 20:15:16 +00:00
itismychild(struct witness *parent, struct witness *child)
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
{
static int recursed;
/*
* Insert "child" after "parent"
*/
while (parent->w_morechildren)
parent = parent->w_morechildren;
if (parent->w_childcnt == WITNESS_NCHILDREN) {
if ((parent->w_morechildren = witness_get()) == NULL)
return (1);
parent = parent->w_morechildren;
}
- Make the mutex code almost completely machine independent. This greatly reducues the maintenance load for the mutex code. The only MD portions of the mutex code are in machine/mutex.h now, which include the assembly macros for handling mutexes as well as optionally overriding the mutex micro-operations. For example, we use optimized micro-ops on the x86 platform #ifndef I386_CPU. - Change the behavior of the SMP_DEBUG kernel option. In the new code, mtx_assert() only depends on INVARIANTS, allowing other kernel developers to have working mutex assertiions without having to include all of the mutex debugging code. The SMP_DEBUG kernel option has been renamed to MUTEX_DEBUG and now just controls extra mutex debugging code. - Abolish the ugly mtx_f hack. Instead, we dynamically allocate seperate mtx_debug structures on the fly in mtx_init, except for mutexes that are initiated very early in the boot process. These mutexes are declared using a special MUTEX_DECLARE() macro, and use a new flag MTX_COLD when calling mtx_init. This is still somewhat hackish, but it is less evil than the mtx_f filler struct, and the mtx struct is now the same size with and without mutex debugging code. - Add some micro-micro-operation macros for doing the actual atomic operations on the mutex mtx_lock field to make it easier for other archs to override/optimize mutex ops if needed. These new tiny ops also clean up the code in some places by replacing long atomic operation function calls that spanned 2-3 lines with a short 1-line macro call. - Don't call mi_switch() from mtx_enter_hard() when we block while trying to obtain a sleep mutex. Calling mi_switch() would bogusly release Giant before switching to the next process. Instead, inline most of the code from mi_switch() in the mtx_enter_hard() function. Note that when we finally kill Giant we can back this out and go back to calling mi_switch().
2000-10-20 07:26:37 +00:00
MPASS(child != NULL);
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
parent->w_children[parent->w_childcnt++] = child;
/*
* now prune whole tree
*/
if (recursed)
return (0);
recursed = 1;
for (child = w_all; child != NULL; child = child->w_next) {
for (parent = w_all; parent != NULL;
parent = parent->w_next) {
if (!isitmychild(parent, child))
continue;
removechild(parent, child);
if (isitmydescendant(parent, child))
continue;
itismychild(parent, child);
}
}
recursed = 0;
witness_levelall();
return (0);
}
static void
Remove the mtx_t, witness_t, and witness_blessed_t types. Instead, just use struct mtx, struct witness, and struct witness_blessed. Requested by: bde
2000-09-14 20:15:16 +00:00
removechild(struct witness *parent, struct witness *child)
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
{
Remove the mtx_t, witness_t, and witness_blessed_t types. Instead, just use struct mtx, struct witness, and struct witness_blessed. Requested by: bde
2000-09-14 20:15:16 +00:00
struct witness *w, *w1;
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
int i;
for (w = parent; w != NULL; w = w->w_morechildren)
for (i = 0; i < w->w_childcnt; i++)
if (w->w_children[i] == child)
goto found;
return;
found:
for (w1 = w; w1->w_morechildren != NULL; w1 = w1->w_morechildren)
continue;
w->w_children[i] = w1->w_children[--w1->w_childcnt];
- Make the mutex code almost completely machine independent. This greatly reducues the maintenance load for the mutex code. The only MD portions of the mutex code are in machine/mutex.h now, which include the assembly macros for handling mutexes as well as optionally overriding the mutex micro-operations. For example, we use optimized micro-ops on the x86 platform #ifndef I386_CPU. - Change the behavior of the SMP_DEBUG kernel option. In the new code, mtx_assert() only depends on INVARIANTS, allowing other kernel developers to have working mutex assertiions without having to include all of the mutex debugging code. The SMP_DEBUG kernel option has been renamed to MUTEX_DEBUG and now just controls extra mutex debugging code. - Abolish the ugly mtx_f hack. Instead, we dynamically allocate seperate mtx_debug structures on the fly in mtx_init, except for mutexes that are initiated very early in the boot process. These mutexes are declared using a special MUTEX_DECLARE() macro, and use a new flag MTX_COLD when calling mtx_init. This is still somewhat hackish, but it is less evil than the mtx_f filler struct, and the mtx struct is now the same size with and without mutex debugging code. - Add some micro-micro-operation macros for doing the actual atomic operations on the mutex mtx_lock field to make it easier for other archs to override/optimize mutex ops if needed. These new tiny ops also clean up the code in some places by replacing long atomic operation function calls that spanned 2-3 lines with a short 1-line macro call. - Don't call mi_switch() from mtx_enter_hard() when we block while trying to obtain a sleep mutex. Calling mi_switch() would bogusly release Giant before switching to the next process. Instead, inline most of the code from mi_switch() in the mtx_enter_hard() function. Note that when we finally kill Giant we can back this out and go back to calling mi_switch().
2000-10-20 07:26:37 +00:00
MPASS(w->w_children[i] != NULL);
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
if (w1->w_childcnt != 0)
return;
if (w1 == parent)
return;
for (w = parent; w->w_morechildren != w1; w = w->w_morechildren)
continue;
w->w_morechildren = 0;
witness_free(w1);
}
static int
Remove the mtx_t, witness_t, and witness_blessed_t types. Instead, just use struct mtx, struct witness, and struct witness_blessed. Requested by: bde
2000-09-14 20:15:16 +00:00
isitmychild(struct witness *parent, struct witness *child)
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
{
Remove the mtx_t, witness_t, and witness_blessed_t types. Instead, just use struct mtx, struct witness, and struct witness_blessed. Requested by: bde
2000-09-14 20:15:16 +00:00
struct witness *w;
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
int i;
for (w = parent; w != NULL; w = w->w_morechildren) {
for (i = 0; i < w->w_childcnt; i++) {
if (w->w_children[i] == child)
return (1);
}
}
return (0);
}
static int
Remove the mtx_t, witness_t, and witness_blessed_t types. Instead, just use struct mtx, struct witness, and struct witness_blessed. Requested by: bde
2000-09-14 20:15:16 +00:00
isitmydescendant(struct witness *parent, struct witness *child)
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
{
Remove the mtx_t, witness_t, and witness_blessed_t types. Instead, just use struct mtx, struct witness, and struct witness_blessed. Requested by: bde
2000-09-14 20:15:16 +00:00
struct witness *w;
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
int i;
int j;
for (j = 0, w = parent; w != NULL; w = w->w_morechildren, j++) {
- Make the mutex code almost completely machine independent. This greatly reducues the maintenance load for the mutex code. The only MD portions of the mutex code are in machine/mutex.h now, which include the assembly macros for handling mutexes as well as optionally overriding the mutex micro-operations. For example, we use optimized micro-ops on the x86 platform #ifndef I386_CPU. - Change the behavior of the SMP_DEBUG kernel option. In the new code, mtx_assert() only depends on INVARIANTS, allowing other kernel developers to have working mutex assertiions without having to include all of the mutex debugging code. The SMP_DEBUG kernel option has been renamed to MUTEX_DEBUG and now just controls extra mutex debugging code. - Abolish the ugly mtx_f hack. Instead, we dynamically allocate seperate mtx_debug structures on the fly in mtx_init, except for mutexes that are initiated very early in the boot process. These mutexes are declared using a special MUTEX_DECLARE() macro, and use a new flag MTX_COLD when calling mtx_init. This is still somewhat hackish, but it is less evil than the mtx_f filler struct, and the mtx struct is now the same size with and without mutex debugging code. - Add some micro-micro-operation macros for doing the actual atomic operations on the mutex mtx_lock field to make it easier for other archs to override/optimize mutex ops if needed. These new tiny ops also clean up the code in some places by replacing long atomic operation function calls that spanned 2-3 lines with a short 1-line macro call. - Don't call mi_switch() from mtx_enter_hard() when we block while trying to obtain a sleep mutex. Calling mi_switch() would bogusly release Giant before switching to the next process. Instead, inline most of the code from mi_switch() in the mtx_enter_hard() function. Note that when we finally kill Giant we can back this out and go back to calling mi_switch().
2000-10-20 07:26:37 +00:00
MPASS(j < 1000);
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
for (i = 0; i < w->w_childcnt; i++) {
if (w->w_children[i] == child)
return (1);
}
for (i = 0; i < w->w_childcnt; i++) {
if (isitmydescendant(w->w_children[i], child))
return (1);
}
}
return (0);
}
void
witness_levelall (void)
{
Remove the mtx_t, witness_t, and witness_blessed_t types. Instead, just use struct mtx, struct witness, and struct witness_blessed. Requested by: bde
2000-09-14 20:15:16 +00:00
struct witness *w, *w1;
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
for (w = w_all; w; w = w->w_next)
Implement MTX_RECURSE flag for mtx_init(). All calls to mtx_init() for mutexes that recurse must now include the MTX_RECURSE bit in the flag argument variable. This change is in preparation for an upcoming (further) mutex API cleanup. The witness code will call panic() if a lock is found to recurse but the MTX_RECURSE bit was not set during the lock's initialization. The old MTX_RECURSE "state" bit (in mtx_lock) has been renamed to MTX_RECURSED, which is more appropriate given its meaning. The following locks have been made "recursive," thus far: eventhandler, Giant, callout, sched_lock, possibly some others declared in the architecture-specific code, all of the network card driver locks in pci/, as well as some other locks in dev/ stuff that I've found to be recursive. Reviewed by: jhb
2001-01-19 01:59:14 +00:00
if (!(w->w_spin))
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
w->w_level = 0;
for (w = w_all; w; w = w->w_next) {
if (w->w_spin)
continue;
for (w1 = w_all; w1; w1 = w1->w_next) {
if (isitmychild(w1, w))
break;
}
if (w1 != NULL)
continue;
witness_leveldescendents(w, 0);
}
}
static void
Remove the mtx_t, witness_t, and witness_blessed_t types. Instead, just use struct mtx, struct witness, and struct witness_blessed. Requested by: bde
2000-09-14 20:15:16 +00:00
witness_leveldescendents(struct witness *parent, int level)
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
{
int i;
Remove the mtx_t, witness_t, and witness_blessed_t types. Instead, just use struct mtx, struct witness, and struct witness_blessed. Requested by: bde
2000-09-14 20:15:16 +00:00
struct witness *w;
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
if (parent->w_level < level)
parent->w_level = level;
level++;
for (w = parent; w != NULL; w = w->w_morechildren)
for (i = 0; i < w->w_childcnt; i++)
witness_leveldescendents(w->w_children[i], level);
}
static void
Remove the mtx_t, witness_t, and witness_blessed_t types. Instead, just use struct mtx, struct witness, and struct witness_blessed. Requested by: bde
2000-09-14 20:15:16 +00:00
witness_displaydescendants(void(*prnt)(const char *fmt, ...),
struct witness *parent)
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
{
Remove the mtx_t, witness_t, and witness_blessed_t types. Instead, just use struct mtx, struct witness, and struct witness_blessed. Requested by: bde
2000-09-14 20:15:16 +00:00
struct witness *w;
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
int i;
- Change the 'witness_list' ddb command to 'show mutexes'. Note that this will only display sleep mutexes held by the current process. - Clean up some nits in the witness_display() function and add a ddb command 'show witness' that dumps the hierarchy and order lists to the console. - Use queue(3) macros where appropriate. - Resort the spin lock order list so that "com" is before "sched_lock". Also, add appropriate #ifdef's around SMP and i386-specific mutexes. - Add two new mutexes used to protect the ithread lists and tables to the order list. Requested by: bde (1)
2001-02-09 15:19:41 +00:00
int level;
level = parent->w_spin ? ffs(parent->w_level) : parent->w_level;
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
prnt("%d", level);
if (level < 10)
prnt(" ");
for (i = 0; i < level; i++)
prnt(" ");
prnt("%s", parent->w_description);
- Change the 'witness_list' ddb command to 'show mutexes'. Note that this will only display sleep mutexes held by the current process. - Clean up some nits in the witness_display() function and add a ddb command 'show witness' that dumps the hierarchy and order lists to the console. - Use queue(3) macros where appropriate. - Resort the spin lock order list so that "com" is before "sched_lock". Also, add appropriate #ifdef's around SMP and i386-specific mutexes. - Add two new mutexes used to protect the ithread lists and tables to the order list. Requested by: bde (1)
2001-02-09 15:19:41 +00:00
if (parent->w_file != NULL)
prnt(" -- last acquired @ %s:%d\n", parent->w_file,
parent->w_line);
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
for (w = parent; w != NULL; w = w->w_morechildren)
for (i = 0; i < w->w_childcnt; i++)
witness_displaydescendants(prnt, w->w_children[i]);
}
static int
Remove the mtx_t, witness_t, and witness_blessed_t types. Instead, just use struct mtx, struct witness, and struct witness_blessed. Requested by: bde
2000-09-14 20:15:16 +00:00
dup_ok(struct witness *w)
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
{
char **dup;
for (dup = dup_list; *dup!= NULL; dup++)
if (strcmp(w->w_description, *dup) == 0)
return (1);
return (0);
}
static int
Remove the mtx_t, witness_t, and witness_blessed_t types. Instead, just use struct mtx, struct witness, and struct witness_blessed. Requested by: bde
2000-09-14 20:15:16 +00:00
blessed(struct witness *w1, struct witness *w2)
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
{
int i;
Remove the mtx_t, witness_t, and witness_blessed_t types. Instead, just use struct mtx, struct witness, and struct witness_blessed. Requested by: bde
2000-09-14 20:15:16 +00:00
struct witness_blessed *b;
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
for (i = 0; i < blessed_count; i++) {
b = &blessed_list[i];
if (strcmp(w1->w_description, b->b_lock1) == 0) {
if (strcmp(w2->w_description, b->b_lock2) == 0)
return (1);
continue;
}
if (strcmp(w1->w_description, b->b_lock2) == 0)
if (strcmp(w2->w_description, b->b_lock1) == 0)
return (1);
}
return (0);
}
Remove the mtx_t, witness_t, and witness_blessed_t types. Instead, just use struct mtx, struct witness, and struct witness_blessed. Requested by: bde
2000-09-14 20:15:16 +00:00
static struct witness *
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
witness_get()
{
Remove the mtx_t, witness_t, and witness_blessed_t types. Instead, just use struct mtx, struct witness, and struct witness_blessed. Requested by: bde
2000-09-14 20:15:16 +00:00
struct witness *w;
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
if ((w = w_free) == NULL) {
witness_dead = 1;
Change and clean the mutex lock interface. mtx_enter(lock, type) becomes: mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks) mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized) similarily, for releasing a lock, we now have: mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN. We change the caller interface for the two different types of locks because the semantics are entirely different for each case, and this makes it explicitly clear and, at the same time, it rids us of the extra `type' argument. The enter->lock and exit->unlock change has been made with the idea that we're "locking data" and not "entering locked code" in mind. Further, remove all additional "flags" previously passed to the lock acquire/release routines with the exception of two: MTX_QUIET and MTX_NOSWITCH The functionality of these flags is preserved and they can be passed to the lock/unlock routines by calling the corresponding wrappers: mtx_{lock, unlock}_flags(lock, flag(s)) and mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN locks, respectively. Re-inline some lock acq/rel code; in the sleep lock case, we only inline the _obtain_lock()s in order to ensure that the inlined code fits into a cache line. In the spin lock case, we inline recursion and actually only perform a function call if we need to spin. This change has been made with the idea that we generally tend to avoid spin locks and that also the spin locks that we do have and are heavily used (i.e. sched_lock) do recurse, and therefore in an effort to reduce function call overhead for some architectures (such as alpha), we inline recursion for this case. Create a new malloc type for the witness code and retire from using the M_DEV type. The new type is called M_WITNESS and is only declared if WITNESS is enabled. Begin cleaning up some machdep/mutex.h code - specifically updated the "optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently need those. Finally, caught up to the interface changes in all sys code. Contributors: jake, jhb, jasone (in no particular order)
2001-02-09 06:11:45 +00:00
mtx_unlock_spin_flags(&w_mtx, MTX_QUIET);
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
printf("witness exhausted\n");
return (NULL);
}
w_free = w->w_next;
Remove the mtx_t, witness_t, and witness_blessed_t types. Instead, just use struct mtx, struct witness, and struct witness_blessed. Requested by: bde
2000-09-14 20:15:16 +00:00
bzero(w, sizeof(*w));
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
return (w);
}
static void
Remove the mtx_t, witness_t, and witness_blessed_t types. Instead, just use struct mtx, struct witness, and struct witness_blessed. Requested by: bde
2000-09-14 20:15:16 +00:00
witness_free(struct witness *w)
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
{
w->w_next = w_free;
w_free = w;
}
- Add code to detect if a system call returns with locks other than Giant held and panic if so (conditional on witness). - Change witness_list to return the number of locks held so this is easier. - Add kern/syscalls.c to the kernel build if witness is defined so that the panic message can contain the name of the offending system call. - Add assertions that Giant and sched_lock are not held when returning from a system call, which were missing for alpha and ia64.
2000-12-12 01:14:32 +00:00
int
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
witness_list(struct proc *p)
{
Remove the mtx_t, witness_t, and witness_blessed_t types. Instead, just use struct mtx, struct witness, and struct witness_blessed. Requested by: bde
2000-09-14 20:15:16 +00:00
struct mtx *m;
- Add code to detect if a system call returns with locks other than Giant held and panic if so (conditional on witness). - Change witness_list to return the number of locks held so this is easier. - Add kern/syscalls.c to the kernel build if witness is defined so that the panic message can contain the name of the offending system call. - Add assertions that Giant and sched_lock are not held when returning from a system call, which were missing for alpha and ia64.
2000-12-12 01:14:32 +00:00
int nheld;
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
Remove MUTEX_DECLARE() and MTX_COLD. Instead, postpone full mutex initialization until after malloc() is safe to call, then iterate through all mutexes and complete their initialization. This change is necessary in order to avoid some circular bootstrapping dependencies.
2001-01-21 07:52:20 +00:00
KASSERT(!witness_cold, ("%s: witness_cold\n", __FUNCTION__));
- Add code to detect if a system call returns with locks other than Giant held and panic if so (conditional on witness). - Change witness_list to return the number of locks held so this is easier. - Add kern/syscalls.c to the kernel build if witness is defined so that the panic message can contain the name of the offending system call. - Add assertions that Giant and sched_lock are not held when returning from a system call, which were missing for alpha and ia64.
2000-12-12 01:14:32 +00:00
nheld = 0;
- Change the 'witness_list' ddb command to 'show mutexes'. Note that this will only display sleep mutexes held by the current process. - Clean up some nits in the witness_display() function and add a ddb command 'show witness' that dumps the hierarchy and order lists to the console. - Use queue(3) macros where appropriate. - Resort the spin lock order list so that "com" is before "sched_lock". Also, add appropriate #ifdef's around SMP and i386-specific mutexes. - Add two new mutexes used to protect the ithread lists and tables to the order list. Requested by: bde (1)
2001-02-09 15:19:41 +00:00
LIST_FOREACH(m, &p->p_heldmtx, mtx_held) {
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
printf("\t\"%s\" (%p) locked at %s:%d\n",
m->mtx_description, m,
m->mtx_witness->w_file, m->mtx_witness->w_line);
- Add code to detect if a system call returns with locks other than Giant held and panic if so (conditional on witness). - Change witness_list to return the number of locks held so this is easier. - Add kern/syscalls.c to the kernel build if witness is defined so that the panic message can contain the name of the offending system call. - Add assertions that Giant and sched_lock are not held when returning from a system call, which were missing for alpha and ia64.
2000-12-12 01:14:32 +00:00
nheld++;
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
}
- Add code to detect if a system call returns with locks other than Giant held and panic if so (conditional on witness). - Change witness_list to return the number of locks held so this is easier. - Add kern/syscalls.c to the kernel build if witness is defined so that the panic message can contain the name of the offending system call. - Add assertions that Giant and sched_lock are not held when returning from a system call, which were missing for alpha and ia64.
2000-12-12 01:14:32 +00:00
return (nheld);
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
}
Add a new ddb command 'witness_list' that lists the mutexes held by curproc. Requested by: peter
2001-01-27 07:51:34 +00:00
#ifdef DDB
- Change the 'witness_list' ddb command to 'show mutexes'. Note that this will only display sleep mutexes held by the current process. - Clean up some nits in the witness_display() function and add a ddb command 'show witness' that dumps the hierarchy and order lists to the console. - Use queue(3) macros where appropriate. - Resort the spin lock order list so that "com" is before "sched_lock". Also, add appropriate #ifdef's around SMP and i386-specific mutexes. - Add two new mutexes used to protect the ithread lists and tables to the order list. Requested by: bde (1)
2001-02-09 15:19:41 +00:00
DB_SHOW_COMMAND(mutexes, db_witness_list)
Add a new ddb command 'witness_list' that lists the mutexes held by curproc. Requested by: peter
2001-01-27 07:51:34 +00:00
{
Change all instances of `CURPROC' and `CURTHD' to `curproc,' in order to stay consistent. Requested by: bde
2001-02-12 03:15:43 +00:00
witness_list(curproc);
Add a new ddb command 'witness_list' that lists the mutexes held by curproc. Requested by: peter
2001-01-27 07:51:34 +00:00
}
- Change the 'witness_list' ddb command to 'show mutexes'. Note that this will only display sleep mutexes held by the current process. - Clean up some nits in the witness_display() function and add a ddb command 'show witness' that dumps the hierarchy and order lists to the console. - Use queue(3) macros where appropriate. - Resort the spin lock order list so that "com" is before "sched_lock". Also, add appropriate #ifdef's around SMP and i386-specific mutexes. - Add two new mutexes used to protect the ithread lists and tables to the order list. Requested by: bde (1)
2001-02-09 15:19:41 +00:00
DB_SHOW_COMMAND(witness, db_witness_display)
{
witness_display(db_printf);
}
Add a new ddb command 'witness_list' that lists the mutexes held by curproc. Requested by: peter
2001-01-27 07:51:34 +00:00
#endif
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
void
Remove the mtx_t, witness_t, and witness_blessed_t types. Instead, just use struct mtx, struct witness, and struct witness_blessed. Requested by: bde
2000-09-14 20:15:16 +00:00
witness_save(struct mtx *m, const char **filep, int *linep)
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
{
Remove MUTEX_DECLARE() and MTX_COLD. Instead, postpone full mutex initialization until after malloc() is safe to call, then iterate through all mutexes and complete their initialization. This change is necessary in order to avoid some circular bootstrapping dependencies.
2001-01-21 07:52:20 +00:00
KASSERT(!witness_cold, ("%s: witness_cold\n", __FUNCTION__));
Move most of sys/mutex.h into kern/kern_mutex.c, thereby making the mutex inline functions non-inlined. Hide parts of the mutex implementation that should not be exposed. Make sure that WITNESS code is not executed during boot until the mutexes are fully initialized by SI_SUB_MUTEX (the original motivation for this commit). Submitted by: peter
2001-01-21 22:34:43 +00:00
if (m->mtx_witness == NULL)
return;
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
*filep = m->mtx_witness->w_file;
*linep = m->mtx_witness->w_line;
}
void
Remove the mtx_t, witness_t, and witness_blessed_t types. Instead, just use struct mtx, struct witness, and struct witness_blessed. Requested by: bde
2000-09-14 20:15:16 +00:00
witness_restore(struct mtx *m, const char *file, int line)
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
{
Remove MUTEX_DECLARE() and MTX_COLD. Instead, postpone full mutex initialization until after malloc() is safe to call, then iterate through all mutexes and complete their initialization. This change is necessary in order to avoid some circular bootstrapping dependencies.
2001-01-21 07:52:20 +00:00
KASSERT(!witness_cold, ("%s: witness_cold\n", __FUNCTION__));
Move most of sys/mutex.h into kern/kern_mutex.c, thereby making the mutex inline functions non-inlined. Hide parts of the mutex implementation that should not be exposed. Make sure that WITNESS code is not executed during boot until the mutexes are fully initialized by SI_SUB_MUTEX (the original motivation for this commit). Submitted by: peter
2001-01-21 22:34:43 +00:00
if (m->mtx_witness == NULL)
return;
Major update to the way synchronization is done in the kernel. Highlights include: * Mutual exclusion is used instead of spl*(). See mutex(9). (Note: The alpha port is still in transition and currently uses both.) * Per-CPU idle processes. * Interrupts are run in their own separate kernel threads and can be preempted (i386 only). Partially contributed by: BSDi (BSD/OS) Submissions by (at least): cp, dfr, dillon, grog, jake, jhb, sheldonh
2000-09-07 01:33:02 +00:00
m->mtx_witness->w_file = file;
m->mtx_witness->w_line = line;
}
Split the WITNESS and MUTEX_DEBUG options apart so that WITNESS does not depend on MUTEX_DEBUG. The MUTEX_DEBUG option turns on extra assertions and checks to verify that mutexes themselves are implemented properly. The WITNESS option uses extra checks and diagnostics to verify that other code is using mutexes properly.
2000-12-01 00:10:59 +00:00
#endif /* WITNESS */