1
0
mirror of https://git.FreeBSD.org/src.git synced 2024-12-05 09:14:03 +00:00
freebsd/sys/kern/kern_fork.c

1117 lines
28 KiB
C
Raw Normal View History

/*-
1994-05-24 10:09:53 +00:00
* Copyright (c) 1982, 1986, 1989, 1991, 1993
* The Regents of the University of California. All rights reserved.
* (c) UNIX System Laboratories, Inc.
* All or some portions of this file are derived from material licensed
* to the University of California by American Telephone and Telegraph
* Co. or Unix System Laboratories, Inc. and are reproduced herein with
* the permission of UNIX System Laboratories, Inc.
*
* 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. Neither the name of the University nor the names of its contributors
1994-05-24 10:09:53 +00:00
* may be used to endorse or promote products derived from this software
* without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``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 THE REGENTS OR CONTRIBUTORS 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.
*
* @(#)kern_fork.c 8.6 (Berkeley) 4/8/94
*/
2003-06-11 00:56:59 +00:00
#include <sys/cdefs.h>
__FBSDID("$FreeBSD$");
#include "opt_ktrace.h"
#include "opt_kstack_pages.h"
1994-05-24 10:09:53 +00:00
#include <sys/param.h>
#include <sys/systm.h>
#include <sys/sysproto.h>
#include <sys/eventhandler.h>
#include <sys/fcntl.h>
1994-05-24 10:09:53 +00:00
#include <sys/filedesc.h>
#include <sys/jail.h>
1994-05-24 10:09:53 +00:00
#include <sys/kernel.h>
#include <sys/kthread.h>
#include <sys/sysctl.h>
Rework the witness code to work with sx locks as well as mutexes. - Introduce lock classes and lock objects. Each lock class specifies a name and set of flags (or properties) shared by all locks of a given type. Currently there are three lock classes: spin mutexes, sleep mutexes, and sx locks. A lock object specifies properties of an additional lock along with a lock name and all of the extra stuff needed to make witness work with a given lock. This abstract lock stuff is defined in sys/lock.h. The lockmgr constants, types, and prototypes have been moved to sys/lockmgr.h. For temporary backwards compatability, sys/lock.h includes sys/lockmgr.h. - Replace proc->p_spinlocks with a per-CPU list, PCPU(spinlocks), of spin locks held. By making this per-cpu, we do not have to jump through magic hoops to deal with sched_lock changing ownership during context switches. - Replace proc->p_heldmtx, formerly a list of held sleep mutexes, with proc->p_sleeplocks, which is a list of held sleep locks including sleep mutexes and sx locks. - Add helper macros for logging lock events via the KTR_LOCK KTR logging level so that the log messages are consistent. - Add some new flags that can be passed to mtx_init(): - MTX_NOWITNESS - specifies that this lock should be ignored by witness. This is used for the mutex that blocks a sx lock for example. - MTX_QUIET - this is not new, but you can pass this to mtx_init() now and no events will be logged for this lock, so that one doesn't have to change all the individual mtx_lock/unlock() operations. - All lock objects maintain an initialized flag. Use this flag to export a mtx_initialized() macro that can be safely called from drivers. Also, we on longer walk the all_mtx list if MUTEX_DEBUG is defined as witness performs the corresponding checks using the initialized flag. - The lock order reversal messages have been improved to output slightly more accurate file and line numbers.
2001-03-28 09:03:24 +00:00
#include <sys/lock.h>
1994-05-24 10:09:53 +00:00
#include <sys/malloc.h>
#include <sys/mutex.h>
#include <sys/priv.h>
1994-05-24 10:09:53 +00:00
#include <sys/proc.h>
#include <sys/procdesc.h>
#include <sys/pioctl.h>
#include <sys/ptrace.h>
#include <sys/racct.h>
1994-05-24 10:09:53 +00:00
#include <sys/resourcevar.h>
#include <sys/sched.h>
#include <sys/syscall.h>
#include <sys/vmmeter.h>
1994-05-24 10:09:53 +00:00
#include <sys/vnode.h>
#include <sys/acct.h>
#include <sys/ktr.h>
1994-05-24 10:09:53 +00:00
#include <sys/ktrace.h>
#include <sys/unistd.h>
#include <sys/sdt.h>
#include <sys/sx.h>
#include <sys/sysent.h>
#include <sys/signalvar.h>
1994-05-24 10:09:53 +00:00
#include <security/audit/audit.h>
#include <security/mac/mac_framework.h>
#include <vm/vm.h>
#include <vm/pmap.h>
#include <vm/vm_map.h>
#include <vm/vm_extern.h>
#include <vm/uma.h>
Add an initial NUMA affinity/policy configuration for threads and processes. This is based on work done by jeff@ and jhb@, as well as the numa.diff patch that has been circulating when someone asks for first-touch NUMA on -10 or -11. * Introduce a simple set of VM policy and iterator types. * tie the policy types into the vm_phys path for now, mirroring how the initial first-touch allocation work was enabled. * add syscalls to control changing thread and process defaults. * add a global NUMA VM domain policy. * implement a simple cascade policy order - if a thread policy exists, use it; if a process policy exists, use it; use the default policy. * processes inherit policies from their parent processes, threads inherit policies from their parent threads. * add a simple tool (numactl) to query and modify default thread/process policities. * add documentation for the new syscalls, for numa and for numactl. * re-enable first touch NUMA again by default, as now policies can be set in a variety of methods. This is only relevant for very specific workloads. This doesn't pretend to be a final NUMA solution. The previous defaults in -HEAD (with MAXMEMDOM set) can be achieved by 'sysctl vm.default_policy=rr'. This is only relevant if MAXMEMDOM is set to something other than 1. Ie, if you're using GENERIC or a modified kernel with non-NUMA, then this is a glorified no-op for you. Thank you to Norse Corp for giving me access to rather large (for FreeBSD!) NUMA machines in order to develop and verify this. Thank you to Dell for providing me with dual socket sandybridge and westmere v3 hardware to do NUMA development with. Thank you to Scott Long at Netflix for providing me with access to the two-socket, four-domain haswell v3 hardware. Thank you to Peter Holm for running the stress testing suite against the NUMA branch during various stages of development! Tested: * MIPS (regression testing; non-NUMA) * i386 (regression testing; non-NUMA GENERIC) * amd64 (regression testing; non-NUMA GENERIC) * westmere, 2 socket (thankyou norse!) * sandy bridge, 2 socket (thankyou dell!) * ivy bridge, 2 socket (thankyou norse!) * westmere-EX, 4 socket / 1TB RAM (thankyou norse!) * haswell, 2 socket (thankyou norse!) * haswell v3, 2 socket (thankyou dell) * haswell v3, 2x18 core (thankyou scott long / netflix!) * Peter Holm ran a stress test suite on this work and found one issue, but has not been able to verify it (it doesn't look NUMA related, and he only saw it once over many testing runs.) * I've tested bhyve instances running in fixed NUMA domains and cpusets; all seems to work correctly. Verified: * intel-pcm - pcm-numa.x and pcm-memory.x, whilst selecting different NUMA policies for processes under test. Review: This was reviewed through phabricator (https://reviews.freebsd.org/D2559) as well as privately and via emails to freebsd-arch@. The git history with specific attributes is available at https://github.com/erikarn/freebsd/ in the NUMA branch (https://github.com/erikarn/freebsd/compare/local/adrian_numa_policy). This has been reviewed by a number of people (stas, rpaulo, kib, ngie, wblock) but not achieved a clear consensus. My hope is that with further exposure and testing more functionality can be implemented and evaluated. Notes: * The VM doesn't handle unbalanced domains very well, and if you have an overly unbalanced memory setup whilst under high memory pressure, VM page allocation may fail leading to a kernel panic. This was a problem in the past, but it's much more easily triggered now with these tools. * This work only controls the path through vm_phys; it doesn't yet strongly/predictably affect contigmalloc, KVA placement, UMA, etc. So, driver placement of memory isn't really guaranteed in any way. That's next on my plate. Sponsored by: Norse Corp, Inc.; Dell
2015-07-11 15:21:37 +00:00
#include <vm/vm_domain.h>
#ifdef KDTRACE_HOOKS
#include <sys/dtrace_bsd.h>
dtrace_fork_func_t dtrace_fasttrap_fork;
#endif
SDT_PROVIDER_DECLARE(proc);
SDT_PROBE_DEFINE3(proc, , , create, "struct proc *", "struct proc *", "int");
#ifndef _SYS_SYSPROTO_H_
struct fork_args {
int dummy;
};
#endif
1994-05-24 10:09:53 +00:00
/* ARGSUSED */
int
sys_fork(struct thread *td, struct fork_args *uap)
1994-05-24 10:09:53 +00:00
{
struct fork_req fr;
int error, pid;
bzero(&fr, sizeof(fr));
fr.fr_flags = RFFDG | RFPROC;
fr.fr_pidp = &pid;
error = fork1(td, &fr);
if (error == 0) {
td->td_retval[0] = pid;
td->td_retval[1] = 0;
}
return (error);
1994-05-24 10:09:53 +00:00
}
/* ARGUSED */
int
sys_pdfork(struct thread *td, struct pdfork_args *uap)
{
struct fork_req fr;
int error, fd, pid;
bzero(&fr, sizeof(fr));
fr.fr_flags = RFFDG | RFPROC | RFPROCDESC;
fr.fr_pidp = &pid;
fr.fr_pd_fd = &fd;
fr.fr_pd_flags = uap->flags;
/*
* It is necessary to return fd by reference because 0 is a valid file
* descriptor number, and the child needs to be able to distinguish
* itself from the parent using the return value.
*/
error = fork1(td, &fr);
if (error == 0) {
td->td_retval[0] = pid;
td->td_retval[1] = 0;
error = copyout(&fd, uap->fdp, sizeof(fd));
}
return (error);
}
1994-05-24 10:09:53 +00:00
/* ARGSUSED */
int
sys_vfork(struct thread *td, struct vfork_args *uap)
1994-05-24 10:09:53 +00:00
{
struct fork_req fr;
int error, pid;
bzero(&fr, sizeof(fr));
fr.fr_flags = RFFDG | RFPROC | RFPPWAIT | RFMEM;
fr.fr_pidp = &pid;
error = fork1(td, &fr);
if (error == 0) {
td->td_retval[0] = pid;
td->td_retval[1] = 0;
}
return (error);
}
1994-05-24 10:09:53 +00:00
int
sys_rfork(struct thread *td, struct rfork_args *uap)
{
struct fork_req fr;
int error, pid;
/* Don't allow kernel-only flags. */
if ((uap->flags & RFKERNELONLY) != 0)
return (EINVAL);
AUDIT_ARG_FFLAGS(uap->flags);
bzero(&fr, sizeof(fr));
fr.fr_flags = uap->flags;
fr.fr_pidp = &pid;
error = fork1(td, &fr);
if (error == 0) {
td->td_retval[0] = pid;
td->td_retval[1] = 0;
}
return (error);
1994-05-24 10:09:53 +00:00
}
int nprocs = 1; /* process 0 */
int lastpid = 0;
SYSCTL_INT(_kern, OID_AUTO, lastpid, CTLFLAG_RD, &lastpid, 0,
"Last used PID");
1994-05-24 10:09:53 +00:00
/*
* Random component to lastpid generation. We mix in a random factor to make
* it a little harder to predict. We sanity check the modulus value to avoid
* doing it in critical paths. Don't let it be too small or we pointlessly
* waste randomness entropy, and don't let it be impossibly large. Using a
* modulus that is too big causes a LOT more process table scans and slows
* down fork processing as the pidchecked caching is defeated.
*/
static int randompid = 0;
static int
sysctl_kern_randompid(SYSCTL_HANDLER_ARGS)
{
int error, pid;
error = sysctl_wire_old_buffer(req, sizeof(int));
if (error != 0)
return(error);
sx_xlock(&allproc_lock);
pid = randompid;
error = sysctl_handle_int(oidp, &pid, 0, req);
if (error == 0 && req->newptr != NULL) {
if (pid < 0 || pid > pid_max - 100) /* out of range */
pid = pid_max - 100;
else if (pid < 2) /* NOP */
pid = 0;
else if (pid < 100) /* Make it reasonable */
pid = 100;
randompid = pid;
}
sx_xunlock(&allproc_lock);
return (error);
}
SYSCTL_PROC(_kern, OID_AUTO, randompid, CTLTYPE_INT|CTLFLAG_RW,
0, 0, sysctl_kern_randompid, "I", "Random PID modulus");
static int
fork_findpid(int flags)
{
struct proc *p;
int trypid;
static int pidchecked = 0;
/*
* Requires allproc_lock in order to iterate over the list
* of processes, and proctree_lock to access p_pgrp.
*/
sx_assert(&allproc_lock, SX_LOCKED);
sx_assert(&proctree_lock, SX_LOCKED);
/*
* Find an unused process ID. We remember a range of unused IDs
* ready to use (from lastpid+1 through pidchecked-1).
*
* If RFHIGHPID is set (used during system boot), do not allocate
* low-numbered pids.
*/
trypid = lastpid + 1;
if (flags & RFHIGHPID) {
if (trypid < 10)
trypid = 10;
} else {
if (randompid)
trypid += arc4random() % randompid;
}
retry:
/*
* If the process ID prototype has wrapped around,
* restart somewhat above 0, as the low-numbered procs
* tend to include daemons that don't exit.
*/
if (trypid >= pid_max) {
trypid = trypid % pid_max;
if (trypid < 100)
trypid += 100;
pidchecked = 0;
}
if (trypid >= pidchecked) {
int doingzomb = 0;
pidchecked = PID_MAX;
/*
* Scan the active and zombie procs to check whether this pid
* is in use. Remember the lowest pid that's greater
* than trypid, so we can avoid checking for a while.
*
* Avoid reuse of the process group id, session id or
* the reaper subtree id. Note that for process group
* and sessions, the amount of reserved pids is
* limited by process limit. For the subtree ids, the
* id is kept reserved only while there is a
* non-reaped process in the subtree, so amount of
* reserved pids is limited by process limit times
* two.
*/
p = LIST_FIRST(&allproc);
again:
for (; p != NULL; p = LIST_NEXT(p, p_list)) {
while (p->p_pid == trypid ||
p->p_reapsubtree == trypid ||
(p->p_pgrp != NULL &&
(p->p_pgrp->pg_id == trypid ||
(p->p_session != NULL &&
p->p_session->s_sid == trypid)))) {
trypid++;
if (trypid >= pidchecked)
goto retry;
}
if (p->p_pid > trypid && pidchecked > p->p_pid)
pidchecked = p->p_pid;
if (p->p_pgrp != NULL) {
if (p->p_pgrp->pg_id > trypid &&
pidchecked > p->p_pgrp->pg_id)
pidchecked = p->p_pgrp->pg_id;
if (p->p_session != NULL &&
p->p_session->s_sid > trypid &&
pidchecked > p->p_session->s_sid)
pidchecked = p->p_session->s_sid;
}
}
if (!doingzomb) {
doingzomb = 1;
p = LIST_FIRST(&zombproc);
goto again;
}
}
/*
* RFHIGHPID does not mess with the lastpid counter during boot.
*/
if (flags & RFHIGHPID)
pidchecked = 0;
else
lastpid = trypid;
return (trypid);
}
static int
fork_norfproc(struct thread *td, int flags)
{
int error;
struct proc *p1;
KASSERT((flags & RFPROC) == 0,
("fork_norfproc called with RFPROC set"));
p1 = td->td_proc;
if (((p1->p_flag & (P_HADTHREADS|P_SYSTEM)) == P_HADTHREADS) &&
(flags & (RFCFDG | RFFDG))) {
PROC_LOCK(p1);
if (thread_single(p1, SINGLE_BOUNDARY)) {
PROC_UNLOCK(p1);
return (ERESTART);
}
PROC_UNLOCK(p1);
}
error = vm_forkproc(td, NULL, NULL, NULL, flags);
if (error)
goto fail;
/*
* Close all file descriptors.
*/
if (flags & RFCFDG) {
struct filedesc *fdtmp;
fdtmp = fdinit(td->td_proc->p_fd, false);
Merge Capsicum overhaul: - Capability is no longer separate descriptor type. Now every descriptor has set of its own capability rights. - The cap_new(2) system call is left, but it is no longer documented and should not be used in new code. - The new syscall cap_rights_limit(2) should be used instead of cap_new(2), which limits capability rights of the given descriptor without creating a new one. - The cap_getrights(2) syscall is renamed to cap_rights_get(2). - If CAP_IOCTL capability right is present we can further reduce allowed ioctls list with the new cap_ioctls_limit(2) syscall. List of allowed ioctls can be retrived with cap_ioctls_get(2) syscall. - If CAP_FCNTL capability right is present we can further reduce fcntls that can be used with the new cap_fcntls_limit(2) syscall and retrive them with cap_fcntls_get(2). - To support ioctl and fcntl white-listing the filedesc structure was heavly modified. - The audit subsystem, kdump and procstat tools were updated to recognize new syscalls. - Capability rights were revised and eventhough I tried hard to provide backward API and ABI compatibility there are some incompatible changes that are described in detail below: CAP_CREATE old behaviour: - Allow for openat(2)+O_CREAT. - Allow for linkat(2). - Allow for symlinkat(2). CAP_CREATE new behaviour: - Allow for openat(2)+O_CREAT. Added CAP_LINKAT: - Allow for linkat(2). ABI: Reuses CAP_RMDIR bit. - Allow to be target for renameat(2). Added CAP_SYMLINKAT: - Allow for symlinkat(2). Removed CAP_DELETE. Old behaviour: - Allow for unlinkat(2) when removing non-directory object. - Allow to be source for renameat(2). Removed CAP_RMDIR. Old behaviour: - Allow for unlinkat(2) when removing directory. Added CAP_RENAMEAT: - Required for source directory for the renameat(2) syscall. Added CAP_UNLINKAT (effectively it replaces CAP_DELETE and CAP_RMDIR): - Allow for unlinkat(2) on any object. - Required if target of renameat(2) exists and will be removed by this call. Removed CAP_MAPEXEC. CAP_MMAP old behaviour: - Allow for mmap(2) with any combination of PROT_NONE, PROT_READ and PROT_WRITE. CAP_MMAP new behaviour: - Allow for mmap(2)+PROT_NONE. Added CAP_MMAP_R: - Allow for mmap(PROT_READ). Added CAP_MMAP_W: - Allow for mmap(PROT_WRITE). Added CAP_MMAP_X: - Allow for mmap(PROT_EXEC). Added CAP_MMAP_RW: - Allow for mmap(PROT_READ | PROT_WRITE). Added CAP_MMAP_RX: - Allow for mmap(PROT_READ | PROT_EXEC). Added CAP_MMAP_WX: - Allow for mmap(PROT_WRITE | PROT_EXEC). Added CAP_MMAP_RWX: - Allow for mmap(PROT_READ | PROT_WRITE | PROT_EXEC). Renamed CAP_MKDIR to CAP_MKDIRAT. Renamed CAP_MKFIFO to CAP_MKFIFOAT. Renamed CAP_MKNODE to CAP_MKNODEAT. CAP_READ old behaviour: - Allow pread(2). - Disallow read(2), readv(2) (if there is no CAP_SEEK). CAP_READ new behaviour: - Allow read(2), readv(2). - Disallow pread(2) (CAP_SEEK was also required). CAP_WRITE old behaviour: - Allow pwrite(2). - Disallow write(2), writev(2) (if there is no CAP_SEEK). CAP_WRITE new behaviour: - Allow write(2), writev(2). - Disallow pwrite(2) (CAP_SEEK was also required). Added convinient defines: #define CAP_PREAD (CAP_SEEK | CAP_READ) #define CAP_PWRITE (CAP_SEEK | CAP_WRITE) #define CAP_MMAP_R (CAP_MMAP | CAP_SEEK | CAP_READ) #define CAP_MMAP_W (CAP_MMAP | CAP_SEEK | CAP_WRITE) #define CAP_MMAP_X (CAP_MMAP | CAP_SEEK | 0x0000000000000008ULL) #define CAP_MMAP_RW (CAP_MMAP_R | CAP_MMAP_W) #define CAP_MMAP_RX (CAP_MMAP_R | CAP_MMAP_X) #define CAP_MMAP_WX (CAP_MMAP_W | CAP_MMAP_X) #define CAP_MMAP_RWX (CAP_MMAP_R | CAP_MMAP_W | CAP_MMAP_X) #define CAP_RECV CAP_READ #define CAP_SEND CAP_WRITE #define CAP_SOCK_CLIENT \ (CAP_CONNECT | CAP_GETPEERNAME | CAP_GETSOCKNAME | CAP_GETSOCKOPT | \ CAP_PEELOFF | CAP_RECV | CAP_SEND | CAP_SETSOCKOPT | CAP_SHUTDOWN) #define CAP_SOCK_SERVER \ (CAP_ACCEPT | CAP_BIND | CAP_GETPEERNAME | CAP_GETSOCKNAME | \ CAP_GETSOCKOPT | CAP_LISTEN | CAP_PEELOFF | CAP_RECV | CAP_SEND | \ CAP_SETSOCKOPT | CAP_SHUTDOWN) Added defines for backward API compatibility: #define CAP_MAPEXEC CAP_MMAP_X #define CAP_DELETE CAP_UNLINKAT #define CAP_MKDIR CAP_MKDIRAT #define CAP_RMDIR CAP_UNLINKAT #define CAP_MKFIFO CAP_MKFIFOAT #define CAP_MKNOD CAP_MKNODAT #define CAP_SOCK_ALL (CAP_SOCK_CLIENT | CAP_SOCK_SERVER) Sponsored by: The FreeBSD Foundation Reviewed by: Christoph Mallon <christoph.mallon@gmx.de> Many aspects discussed with: rwatson, benl, jonathan ABI compatibility discussed with: kib
2013-03-02 00:53:12 +00:00
fdescfree(td);
p1->p_fd = fdtmp;
}
/*
* Unshare file descriptors (from parent).
*/
if (flags & RFFDG)
fdunshare(td);
fail:
if (((p1->p_flag & (P_HADTHREADS|P_SYSTEM)) == P_HADTHREADS) &&
(flags & (RFCFDG | RFFDG))) {
PROC_LOCK(p1);
thread_single_end(p1, SINGLE_BOUNDARY);
PROC_UNLOCK(p1);
}
return (error);
}
static void
do_fork(struct thread *td, struct fork_req *fr, struct proc *p2, struct thread *td2,
struct vmspace *vm2, struct file *fp_procdesc)
1994-05-24 10:09:53 +00:00
{
struct proc *p1, *pptr;
int trypid;
struct filedesc *fd;
struct filedesc_to_leader *fdtol;
struct sigacts *newsigacts;
sx_assert(&proctree_lock, SX_SLOCKED);
sx_assert(&allproc_lock, SX_XLOCKED);
1994-05-24 10:09:53 +00:00
p1 = td->td_proc;
trypid = fork_findpid(fr->fr_flags);
1994-05-24 10:09:53 +00:00
sx_sunlock(&proctree_lock);
1994-05-24 10:09:53 +00:00
p2->p_state = PRS_NEW; /* protect against others */
p2->p_pid = trypid;
AUDIT_ARG_PID(p2->p_pid);
LIST_INSERT_HEAD(&allproc, p2, p_list);
allproc_gen++;
LIST_INSERT_HEAD(PIDHASH(p2->p_pid), p2, p_hash);
tidhash_add(td2);
Close race conditions between fork() and [sg]etpriority()'s PRIO_USER case, possibly also other places that deferences p_ucred. In the past, we insert a new process into the allproc list right after PID allocation, and release the allproc_lock sx. Because most content in new proc's structure is not yet initialized, this could lead to undefined result if we do not handle PRS_NEW with care. The problem with PRS_NEW state is that it does not provide fine grained information about how much initialization is done for a new process. By defination, after PRIO_USER setpriority(), all processes that belongs to given user should have their nice value set to the specified value. Therefore, if p_{start,end}copy section was done for a PRS_NEW process, we can not safely ignore it because p_nice is in this area. On the other hand, we should be careful on PRS_NEW processes because we do not allow non-root users to lower their nice values, and without a successful copy of the copy section, we can get stale values that is inherted from the uninitialized area of the process structure. This commit tries to close the race condition by grabbing proc mutex *before* we release allproc_lock xlock, and do copy as well as zero immediately after the allproc_lock xunlock. This guarantees that the new process would have its p_copy and p_zero sections, as well as user credential informaion initialized. In getpriority() case, instead of grabbing PROC_LOCK for a PRS_NEW process, we just skip the process in question, because it does not affect the final result of the call, as the p_nice value would be copied from its parent, and we will see it during allproc traverse. Other potential solutions are still under evaluation. Discussed with: davidxu, jhb, rwatson PR: kern/108071 MFC after: 2 weeks
2007-02-26 03:38:09 +00:00
PROC_LOCK(p2);
PROC_LOCK(p1);
sx_xunlock(&allproc_lock);
Close race conditions between fork() and [sg]etpriority()'s PRIO_USER case, possibly also other places that deferences p_ucred. In the past, we insert a new process into the allproc list right after PID allocation, and release the allproc_lock sx. Because most content in new proc's structure is not yet initialized, this could lead to undefined result if we do not handle PRS_NEW with care. The problem with PRS_NEW state is that it does not provide fine grained information about how much initialization is done for a new process. By defination, after PRIO_USER setpriority(), all processes that belongs to given user should have their nice value set to the specified value. Therefore, if p_{start,end}copy section was done for a PRS_NEW process, we can not safely ignore it because p_nice is in this area. On the other hand, we should be careful on PRS_NEW processes because we do not allow non-root users to lower their nice values, and without a successful copy of the copy section, we can get stale values that is inherted from the uninitialized area of the process structure. This commit tries to close the race condition by grabbing proc mutex *before* we release allproc_lock xlock, and do copy as well as zero immediately after the allproc_lock xunlock. This guarantees that the new process would have its p_copy and p_zero sections, as well as user credential informaion initialized. In getpriority() case, instead of grabbing PROC_LOCK for a PRS_NEW process, we just skip the process in question, because it does not affect the final result of the call, as the p_nice value would be copied from its parent, and we will see it during allproc traverse. Other potential solutions are still under evaluation. Discussed with: davidxu, jhb, rwatson PR: kern/108071 MFC after: 2 weeks
2007-02-26 03:38:09 +00:00
bcopy(&p1->p_startcopy, &p2->p_startcopy,
__rangeof(struct proc, p_startcopy, p_endcopy));
pargs_hold(p2->p_args);
Add an initial NUMA affinity/policy configuration for threads and processes. This is based on work done by jeff@ and jhb@, as well as the numa.diff patch that has been circulating when someone asks for first-touch NUMA on -10 or -11. * Introduce a simple set of VM policy and iterator types. * tie the policy types into the vm_phys path for now, mirroring how the initial first-touch allocation work was enabled. * add syscalls to control changing thread and process defaults. * add a global NUMA VM domain policy. * implement a simple cascade policy order - if a thread policy exists, use it; if a process policy exists, use it; use the default policy. * processes inherit policies from their parent processes, threads inherit policies from their parent threads. * add a simple tool (numactl) to query and modify default thread/process policities. * add documentation for the new syscalls, for numa and for numactl. * re-enable first touch NUMA again by default, as now policies can be set in a variety of methods. This is only relevant for very specific workloads. This doesn't pretend to be a final NUMA solution. The previous defaults in -HEAD (with MAXMEMDOM set) can be achieved by 'sysctl vm.default_policy=rr'. This is only relevant if MAXMEMDOM is set to something other than 1. Ie, if you're using GENERIC or a modified kernel with non-NUMA, then this is a glorified no-op for you. Thank you to Norse Corp for giving me access to rather large (for FreeBSD!) NUMA machines in order to develop and verify this. Thank you to Dell for providing me with dual socket sandybridge and westmere v3 hardware to do NUMA development with. Thank you to Scott Long at Netflix for providing me with access to the two-socket, four-domain haswell v3 hardware. Thank you to Peter Holm for running the stress testing suite against the NUMA branch during various stages of development! Tested: * MIPS (regression testing; non-NUMA) * i386 (regression testing; non-NUMA GENERIC) * amd64 (regression testing; non-NUMA GENERIC) * westmere, 2 socket (thankyou norse!) * sandy bridge, 2 socket (thankyou dell!) * ivy bridge, 2 socket (thankyou norse!) * westmere-EX, 4 socket / 1TB RAM (thankyou norse!) * haswell, 2 socket (thankyou norse!) * haswell v3, 2 socket (thankyou dell) * haswell v3, 2x18 core (thankyou scott long / netflix!) * Peter Holm ran a stress test suite on this work and found one issue, but has not been able to verify it (it doesn't look NUMA related, and he only saw it once over many testing runs.) * I've tested bhyve instances running in fixed NUMA domains and cpusets; all seems to work correctly. Verified: * intel-pcm - pcm-numa.x and pcm-memory.x, whilst selecting different NUMA policies for processes under test. Review: This was reviewed through phabricator (https://reviews.freebsd.org/D2559) as well as privately and via emails to freebsd-arch@. The git history with specific attributes is available at https://github.com/erikarn/freebsd/ in the NUMA branch (https://github.com/erikarn/freebsd/compare/local/adrian_numa_policy). This has been reviewed by a number of people (stas, rpaulo, kib, ngie, wblock) but not achieved a clear consensus. My hope is that with further exposure and testing more functionality can be implemented and evaluated. Notes: * The VM doesn't handle unbalanced domains very well, and if you have an overly unbalanced memory setup whilst under high memory pressure, VM page allocation may fail leading to a kernel panic. This was a problem in the past, but it's much more easily triggered now with these tools. * This work only controls the path through vm_phys; it doesn't yet strongly/predictably affect contigmalloc, KVA placement, UMA, etc. So, driver placement of memory isn't really guaranteed in any way. That's next on my plate. Sponsored by: Norse Corp, Inc.; Dell
2015-07-11 15:21:37 +00:00
Close race conditions between fork() and [sg]etpriority()'s PRIO_USER case, possibly also other places that deferences p_ucred. In the past, we insert a new process into the allproc list right after PID allocation, and release the allproc_lock sx. Because most content in new proc's structure is not yet initialized, this could lead to undefined result if we do not handle PRS_NEW with care. The problem with PRS_NEW state is that it does not provide fine grained information about how much initialization is done for a new process. By defination, after PRIO_USER setpriority(), all processes that belongs to given user should have their nice value set to the specified value. Therefore, if p_{start,end}copy section was done for a PRS_NEW process, we can not safely ignore it because p_nice is in this area. On the other hand, we should be careful on PRS_NEW processes because we do not allow non-root users to lower their nice values, and without a successful copy of the copy section, we can get stale values that is inherted from the uninitialized area of the process structure. This commit tries to close the race condition by grabbing proc mutex *before* we release allproc_lock xlock, and do copy as well as zero immediately after the allproc_lock xunlock. This guarantees that the new process would have its p_copy and p_zero sections, as well as user credential informaion initialized. In getpriority() case, instead of grabbing PROC_LOCK for a PRS_NEW process, we just skip the process in question, because it does not affect the final result of the call, as the p_nice value would be copied from its parent, and we will see it during allproc traverse. Other potential solutions are still under evaluation. Discussed with: davidxu, jhb, rwatson PR: kern/108071 MFC after: 2 weeks
2007-02-26 03:38:09 +00:00
PROC_UNLOCK(p1);
bzero(&p2->p_startzero,
__rangeof(struct proc, p_startzero, p_endzero));
/* Tell the prison that we exist. */
prison_proc_hold(p2->p_ucred->cr_prison);
MFp4: Bring in updated jail support from bz_jail branch. This enhances the current jail implementation to permit multiple addresses per jail. In addtion to IPv4, IPv6 is supported as well. Due to updated checks it is even possible to have jails without an IP address at all, which basically gives one a chroot with restricted process view, no networking,.. SCTP support was updated and supports IPv6 in jails as well. Cpuset support permits jails to be bound to specific processor sets after creation. Jails can have an unrestricted (no duplicate protection, etc.) name in addition to the hostname. The jail name cannot be changed from within a jail and is considered to be used for management purposes or as audit-token in the future. DDB 'show jails' command was added to aid debugging. Proper compat support permits 32bit jail binaries to be used on 64bit systems to manage jails. Also backward compatibility was preserved where possible: for jail v1 syscalls, as well as with user space management utilities. Both jail as well as prison version were updated for the new features. A gap was intentionally left as the intermediate versions had been used by various patches floating around the last years. Bump __FreeBSD_version for the afore mentioned and in kernel changes. Special thanks to: - Pawel Jakub Dawidek (pjd) for his multi-IPv4 patches and Olivier Houchard (cognet) for initial single-IPv6 patches. - Jeff Roberson (jeff) and Randall Stewart (rrs) for their help, ideas and review on cpuset and SCTP support. - Robert Watson (rwatson) for lots and lots of help, discussions, suggestions and review of most of the patch at various stages. - John Baldwin (jhb) for his help. - Simon L. Nielsen (simon) as early adopter testing changes on cluster machines as well as all the testers and people who provided feedback the last months on freebsd-jail and other channels. - My employer, CK Software GmbH, for the support so I could work on this. Reviewed by: (see above) MFC after: 3 months (this is just so that I get the mail) X-MFC Before: 7.2-RELEASE if possible
2008-11-29 14:32:14 +00:00
Close race conditions between fork() and [sg]etpriority()'s PRIO_USER case, possibly also other places that deferences p_ucred. In the past, we insert a new process into the allproc list right after PID allocation, and release the allproc_lock sx. Because most content in new proc's structure is not yet initialized, this could lead to undefined result if we do not handle PRS_NEW with care. The problem with PRS_NEW state is that it does not provide fine grained information about how much initialization is done for a new process. By defination, after PRIO_USER setpriority(), all processes that belongs to given user should have their nice value set to the specified value. Therefore, if p_{start,end}copy section was done for a PRS_NEW process, we can not safely ignore it because p_nice is in this area. On the other hand, we should be careful on PRS_NEW processes because we do not allow non-root users to lower their nice values, and without a successful copy of the copy section, we can get stale values that is inherted from the uninitialized area of the process structure. This commit tries to close the race condition by grabbing proc mutex *before* we release allproc_lock xlock, and do copy as well as zero immediately after the allproc_lock xunlock. This guarantees that the new process would have its p_copy and p_zero sections, as well as user credential informaion initialized. In getpriority() case, instead of grabbing PROC_LOCK for a PRS_NEW process, we just skip the process in question, because it does not affect the final result of the call, as the p_nice value would be copied from its parent, and we will see it during allproc traverse. Other potential solutions are still under evaluation. Discussed with: davidxu, jhb, rwatson PR: kern/108071 MFC after: 2 weeks
2007-02-26 03:38:09 +00:00
PROC_UNLOCK(p2);
/*
* Malloc things while we don't hold any locks.
*/
if (fr->fr_flags & RFSIGSHARE)
newsigacts = NULL;
else
newsigacts = sigacts_alloc();
/*
* Copy filedesc.
*/
if (fr->fr_flags & RFCFDG) {
fd = fdinit(p1->p_fd, false);
fdtol = NULL;
} else if (fr->fr_flags & RFFDG) {
fd = fdcopy(p1->p_fd);
fdtol = NULL;
} else {
fd = fdshare(p1->p_fd);
if (p1->p_fdtol == NULL)
p1->p_fdtol = filedesc_to_leader_alloc(NULL, NULL,
p1->p_leader);
if ((fr->fr_flags & RFTHREAD) != 0) {
/*
* Shared file descriptor table, and shared
* process leaders.
*/
fdtol = p1->p_fdtol;
Replace custom file descriptor array sleep lock constructed using a mutex and flags with an sxlock. This leads to a significant and measurable performance improvement as a result of access to shared locking for frequent lookup operations, reduced general overhead, and reduced overhead in the event of contention. All of these are imported for threaded applications where simultaneous access to a shared file descriptor array occurs frequently. Kris has reported 2x-4x transaction rate improvements on 8-core MySQL benchmarks; smaller improvements can be expected for many workloads as a result of reduced overhead. - Generally eliminate the distinction between "fast" and regular acquisisition of the filedesc lock; the plan is that they will now all be fast. Change all locking instances to either shared or exclusive locks. - Correct a bug (pointed out by kib) in fdfree() where previously msleep() was called without the mutex held; sx_sleep() is now always called with the sxlock held exclusively. - Universally hold the struct file lock over changes to struct file, rather than the filedesc lock or no lock. Always update the f_ops field last. A further memory barrier is required here in the future (discussed with jhb). - Improve locking and reference management in linux_at(), which fails to properly acquire vnode references before using vnode pointers. Annotate improper use of vn_fullpath(), which will be replaced at a future date. In fcntl(), we conservatively acquire an exclusive lock, even though in some cases a shared lock may be sufficient, which should be revisited. The dropping of the filedesc lock in fdgrowtable() is no longer required as the sxlock can be held over the sleep operation; we should consider removing that (pointed out by attilio). Tested by: kris Discussed with: jhb, kris, attilio, jeff
2007-04-04 09:11:34 +00:00
FILEDESC_XLOCK(p1->p_fd);
fdtol->fdl_refcount++;
Replace custom file descriptor array sleep lock constructed using a mutex and flags with an sxlock. This leads to a significant and measurable performance improvement as a result of access to shared locking for frequent lookup operations, reduced general overhead, and reduced overhead in the event of contention. All of these are imported for threaded applications where simultaneous access to a shared file descriptor array occurs frequently. Kris has reported 2x-4x transaction rate improvements on 8-core MySQL benchmarks; smaller improvements can be expected for many workloads as a result of reduced overhead. - Generally eliminate the distinction between "fast" and regular acquisisition of the filedesc lock; the plan is that they will now all be fast. Change all locking instances to either shared or exclusive locks. - Correct a bug (pointed out by kib) in fdfree() where previously msleep() was called without the mutex held; sx_sleep() is now always called with the sxlock held exclusively. - Universally hold the struct file lock over changes to struct file, rather than the filedesc lock or no lock. Always update the f_ops field last. A further memory barrier is required here in the future (discussed with jhb). - Improve locking and reference management in linux_at(), which fails to properly acquire vnode references before using vnode pointers. Annotate improper use of vn_fullpath(), which will be replaced at a future date. In fcntl(), we conservatively acquire an exclusive lock, even though in some cases a shared lock may be sufficient, which should be revisited. The dropping of the filedesc lock in fdgrowtable() is no longer required as the sxlock can be held over the sleep operation; we should consider removing that (pointed out by attilio). Tested by: kris Discussed with: jhb, kris, attilio, jeff
2007-04-04 09:11:34 +00:00
FILEDESC_XUNLOCK(p1->p_fd);
} else {
/*
* Shared file descriptor table, and different
* process leaders.
*/
fdtol = filedesc_to_leader_alloc(p1->p_fdtol,
p1->p_fd, p2);
}
}
1994-05-24 10:09:53 +00:00
/*
* Make a proc table entry for the new process.
* Start by zeroing the section of proc that is zero-initialized,
* then copy the section that is copied directly from the parent.
*/
PROC_LOCK(p2);
PROC_LOCK(p1);
bzero(&td2->td_startzero,
__rangeof(struct thread, td_startzero, td_endzero));
bcopy(&td->td_startcopy, &td2->td_startcopy,
__rangeof(struct thread, td_startcopy, td_endcopy));
1994-05-24 10:09:53 +00:00
bcopy(&p2->p_comm, &td2->td_name, sizeof(td2->td_name));
td2->td_sigstk = td->td_sigstk;
td2->td_flags = TDF_INMEM;
td2->td_lend_user_pri = PRI_MAX;
Change the curvnet variable from a global const struct vnet *, previously always pointing to the default vnet context, to a dynamically changing thread-local one. The currvnet context should be set on entry to networking code via CURVNET_SET() macros, and reverted to previous state via CURVNET_RESTORE(). Recursions on curvnet are permitted, though strongly discuouraged. This change should have no functional impact on nooptions VIMAGE kernel builds, where CURVNET_* macros expand to whitespace. The curthread->td_vnet (aka curvnet) variable's purpose is to be an indicator of the vnet context in which the current network-related operation takes place, in case we cannot deduce the current vnet context from any other source, such as by looking at mbuf's m->m_pkthdr.rcvif->if_vnet, sockets's so->so_vnet etc. Moreover, so far curvnet has turned out to be an invaluable consistency checking aid: it helps to catch cases when sockets, ifnets or any other vnet-aware structures may have leaked from one vnet to another. The exact placement of the CURVNET_SET() / CURVNET_RESTORE() macros was a result of an empirical iterative process, whith an aim to reduce recursions on CURVNET_SET() to a minimum, while still reducing the scope of CURVNET_SET() to networking only operations - the alternative would be calling CURVNET_SET() on each system call entry. In general, curvnet has to be set in three typicall cases: when processing socket-related requests from userspace or from within the kernel; when processing inbound traffic flowing from device drivers to upper layers of the networking stack, and when executing timer-driven networking functions. This change also introduces a DDB subcommand to show the list of all vnet instances. Approved by: julian (mentor)
2009-05-05 10:56:12 +00:00
#ifdef VIMAGE
td2->td_vnet = NULL;
td2->td_vnet_lpush = NULL;
#endif
/*
* Allow the scheduler to initialize the child.
*/
thread_lock(td);
sched_fork(td, td2);
thread_unlock(td);
1994-05-24 10:09:53 +00:00
/*
* Duplicate sub-structures as needed.
* Increase reference counts on shared objects.
*/
p2->p_flag = P_INMEM;
p2->p_flag2 = p1->p_flag2 & (P2_NOTRACE | P2_NOTRACE_EXEC | P2_TRAPCAP);
p2->p_swtick = ticks;
if (p1->p_flag & P_PROFIL)
startprofclock(p2);
Add an initial NUMA affinity/policy configuration for threads and processes. This is based on work done by jeff@ and jhb@, as well as the numa.diff patch that has been circulating when someone asks for first-touch NUMA on -10 or -11. * Introduce a simple set of VM policy and iterator types. * tie the policy types into the vm_phys path for now, mirroring how the initial first-touch allocation work was enabled. * add syscalls to control changing thread and process defaults. * add a global NUMA VM domain policy. * implement a simple cascade policy order - if a thread policy exists, use it; if a process policy exists, use it; use the default policy. * processes inherit policies from their parent processes, threads inherit policies from their parent threads. * add a simple tool (numactl) to query and modify default thread/process policities. * add documentation for the new syscalls, for numa and for numactl. * re-enable first touch NUMA again by default, as now policies can be set in a variety of methods. This is only relevant for very specific workloads. This doesn't pretend to be a final NUMA solution. The previous defaults in -HEAD (with MAXMEMDOM set) can be achieved by 'sysctl vm.default_policy=rr'. This is only relevant if MAXMEMDOM is set to something other than 1. Ie, if you're using GENERIC or a modified kernel with non-NUMA, then this is a glorified no-op for you. Thank you to Norse Corp for giving me access to rather large (for FreeBSD!) NUMA machines in order to develop and verify this. Thank you to Dell for providing me with dual socket sandybridge and westmere v3 hardware to do NUMA development with. Thank you to Scott Long at Netflix for providing me with access to the two-socket, four-domain haswell v3 hardware. Thank you to Peter Holm for running the stress testing suite against the NUMA branch during various stages of development! Tested: * MIPS (regression testing; non-NUMA) * i386 (regression testing; non-NUMA GENERIC) * amd64 (regression testing; non-NUMA GENERIC) * westmere, 2 socket (thankyou norse!) * sandy bridge, 2 socket (thankyou dell!) * ivy bridge, 2 socket (thankyou norse!) * westmere-EX, 4 socket / 1TB RAM (thankyou norse!) * haswell, 2 socket (thankyou norse!) * haswell v3, 2 socket (thankyou dell) * haswell v3, 2x18 core (thankyou scott long / netflix!) * Peter Holm ran a stress test suite on this work and found one issue, but has not been able to verify it (it doesn't look NUMA related, and he only saw it once over many testing runs.) * I've tested bhyve instances running in fixed NUMA domains and cpusets; all seems to work correctly. Verified: * intel-pcm - pcm-numa.x and pcm-memory.x, whilst selecting different NUMA policies for processes under test. Review: This was reviewed through phabricator (https://reviews.freebsd.org/D2559) as well as privately and via emails to freebsd-arch@. The git history with specific attributes is available at https://github.com/erikarn/freebsd/ in the NUMA branch (https://github.com/erikarn/freebsd/compare/local/adrian_numa_policy). This has been reviewed by a number of people (stas, rpaulo, kib, ngie, wblock) but not achieved a clear consensus. My hope is that with further exposure and testing more functionality can be implemented and evaluated. Notes: * The VM doesn't handle unbalanced domains very well, and if you have an overly unbalanced memory setup whilst under high memory pressure, VM page allocation may fail leading to a kernel panic. This was a problem in the past, but it's much more easily triggered now with these tools. * This work only controls the path through vm_phys; it doesn't yet strongly/predictably affect contigmalloc, KVA placement, UMA, etc. So, driver placement of memory isn't really guaranteed in any way. That's next on my plate. Sponsored by: Norse Corp, Inc.; Dell
2015-07-11 15:21:37 +00:00
/*
* Whilst the proc lock is held, copy the VM domain data out
* using the VM domain method.
*/
vm_domain_policy_init(&p2->p_vm_dom_policy);
vm_domain_policy_localcopy(&p2->p_vm_dom_policy,
&p1->p_vm_dom_policy);
if (fr->fr_flags & RFSIGSHARE) {
p2->p_sigacts = sigacts_hold(p1->p_sigacts);
} else {
sigacts_copy(newsigacts, p1->p_sigacts);
p2->p_sigacts = newsigacts;
}
if (fr->fr_flags & RFTSIGZMB)
p2->p_sigparent = RFTSIGNUM(fr->fr_flags);
else if (fr->fr_flags & RFLINUXTHPN)
p2->p_sigparent = SIGUSR1;
else
p2->p_sigparent = SIGCHLD;
1994-05-24 10:09:53 +00:00
p2->p_textvp = p1->p_textvp;
p2->p_fd = fd;
p2->p_fdtol = fdtol;
if (p1->p_flag2 & P2_INHERIT_PROTECTED) {
p2->p_flag |= P_PROTECTED;
p2->p_flag2 |= P2_INHERIT_PROTECTED;
}
1994-05-24 10:09:53 +00:00
/*
* p_limit is copy-on-write. Bump its refcount.
1994-05-24 10:09:53 +00:00
*/
lim_fork(p1, p2);
thread_cow_get_proc(td2, p2);
pstats_fork(p1->p_stats, p2->p_stats);
PROC_UNLOCK(p1);
PROC_UNLOCK(p2);
1994-05-24 10:09:53 +00:00
/* Bump references to the text vnode (for procfs). */
if (p2->p_textvp)
vrefact(p2->p_textvp);
/*
* Set up linkage for kernel based threading.
*/
if ((fr->fr_flags & RFTHREAD) != 0) {
mtx_lock(&ppeers_lock);
p2->p_peers = p1->p_peers;
p1->p_peers = p2;
p2->p_leader = p1->p_leader;
mtx_unlock(&ppeers_lock);
PROC_LOCK(p1->p_leader);
if ((p1->p_leader->p_flag & P_WEXIT) != 0) {
PROC_UNLOCK(p1->p_leader);
/*
* The task leader is exiting, so process p1 is
* going to be killed shortly. Since p1 obviously
* isn't dead yet, we know that the leader is either
* sending SIGKILL's to all the processes in this
* task or is sleeping waiting for all the peers to
* exit. We let p1 complete the fork, but we need
* to go ahead and kill the new process p2 since
* the task leader may not get a chance to send
* SIGKILL to it. We leave it on the list so that
* the task leader will wait for this new process
* to commit suicide.
*/
PROC_LOCK(p2);
kern_psignal(p2, SIGKILL);
PROC_UNLOCK(p2);
} else
PROC_UNLOCK(p1->p_leader);
} else {
p2->p_peers = NULL;
p2->p_leader = p2;
}
sx_xlock(&proctree_lock);
PGRP_LOCK(p1->p_pgrp);
PROC_LOCK(p2);
PROC_LOCK(p1);
/*
* Preserve some more flags in subprocess. P_PROFIL has already
* been preserved.
*/
p2->p_flag |= p1->p_flag & P_SUGID;
td2->td_pflags |= (td->td_pflags & TDP_ALTSTACK) | TDP_FORKING;
SESS_LOCK(p1->p_session);
1994-05-24 10:09:53 +00:00
if (p1->p_session->s_ttyvp != NULL && p1->p_flag & P_CONTROLT)
p2->p_flag |= P_CONTROLT;
SESS_UNLOCK(p1->p_session);
if (fr->fr_flags & RFPPWAIT)
1994-05-24 10:09:53 +00:00
p2->p_flag |= P_PPWAIT;
p2->p_pgrp = p1->p_pgrp;
LIST_INSERT_AFTER(p1, p2, p_pglist);
2002-04-02 17:12:06 +00:00
PGRP_UNLOCK(p1->p_pgrp);
LIST_INIT(&p2->p_children);
LIST_INIT(&p2->p_orphans);
callout_init_mtx(&p2->p_itcallout, &p2->p_mtx, 0);
/*
* If PF_FORK is set, the child process inherits the
* procfs ioctl flags from its parent.
*/
if (p1->p_pfsflags & PF_FORK) {
p2->p_stops = p1->p_stops;
p2->p_pfsflags = p1->p_pfsflags;
}
1994-05-24 10:09:53 +00:00
/*
* This begins the section where we must prevent the parent
* from being swapped.
1994-05-24 10:09:53 +00:00
*/
_PHOLD(p1);
PROC_UNLOCK(p1);
/*
* Attach the new process to its parent.
*
* If RFNOWAIT is set, the newly created process becomes a child
* of init. This effectively disassociates the child from the
* parent.
*/
if ((fr->fr_flags & RFNOWAIT) != 0) {
pptr = p1->p_reaper;
p2->p_reaper = pptr;
} else {
p2->p_reaper = (p1->p_treeflag & P_TREE_REAPER) != 0 ?
p1 : p1->p_reaper;
pptr = p1;
}
p2->p_pptr = pptr;
LIST_INSERT_HEAD(&pptr->p_children, p2, p_sibling);
LIST_INIT(&p2->p_reaplist);
LIST_INSERT_HEAD(&p2->p_reaper->p_reaplist, p2, p_reapsibling);
if (p2->p_reaper == p1)
p2->p_reapsubtree = p2->p_pid;
sx_xunlock(&proctree_lock);
/* Inform accounting that we have forked. */
p2->p_acflag = AFORK;
PROC_UNLOCK(p2);
#ifdef KTRACE
ktrprocfork(p1, p2);
#endif
/*
The biggie: Get rid of the UPAGES from the top of the per-process address space. (!) Have each process use the kernel stack and pcb in the kvm space. Since the stacks are at a different address, we cannot copy the stack at fork() and allow the child to return up through the function call tree to return to user mode - create a new execution context and have the new process begin executing from cpu_switch() and go to user mode directly. In theory this should speed up fork a bit. Context switch the tss_esp0 pointer in the common tss. This is a lot simpler since than swithching the gdt[GPROC0_SEL].sd.sd_base pointer to each process's tss since the esp0 pointer is a 32 bit pointer, and the sd_base setting is split into three different bit sections at non-aligned boundaries and requires a lot of twiddling to reset. The 8K of memory at the top of the process space is now empty, and unmapped (and unmappable, it's higher than VM_MAXUSER_ADDRESS). Simplity the pmap code to manage process contexts, we no longer have to double map the UPAGES, this simplifies and should measuably speed up fork(). The following parts came from John Dyson: Set PG_G on the UPAGES that are now in kernel context, and invalidate them when swapping them out. Move the upages object (upobj) from the vmspace to the proc structure. Now that the UPAGES (pcb and kernel stack) are out of user space, make rfork(..RFMEM..) do what was intended by sharing the vmspace entirely via reference counting rather than simply inheriting the mappings.
1997-04-07 07:16:06 +00:00
* Finish creating the child process. It will return via a different
* execution path later. (ie: directly into user mode)
*/
vm_forkproc(td, p2, td2, vm2, fr->fr_flags);
1994-05-24 10:09:53 +00:00
if (fr->fr_flags == (RFFDG | RFPROC)) {
PCPU_INC(cnt.v_forks);
PCPU_ADD(cnt.v_forkpages, p2->p_vmspace->vm_dsize +
p2->p_vmspace->vm_ssize);
} else if (fr->fr_flags == (RFFDG | RFPROC | RFPPWAIT | RFMEM)) {
PCPU_INC(cnt.v_vforks);
PCPU_ADD(cnt.v_vforkpages, p2->p_vmspace->vm_dsize +
p2->p_vmspace->vm_ssize);
} else if (p1 == &proc0) {
PCPU_INC(cnt.v_kthreads);
PCPU_ADD(cnt.v_kthreadpages, p2->p_vmspace->vm_dsize +
p2->p_vmspace->vm_ssize);
} else {
PCPU_INC(cnt.v_rforks);
PCPU_ADD(cnt.v_rforkpages, p2->p_vmspace->vm_dsize +
p2->p_vmspace->vm_ssize);
}
/*
* Associate the process descriptor with the process before anything
* can happen that might cause that process to need the descriptor.
* However, don't do this until after fork(2) can no longer fail.
*/
if (fr->fr_flags & RFPROCDESC)
procdesc_new(p2, fr->fr_pd_flags);
/*
* Both processes are set up, now check if any loadable modules want
* to adjust anything.
*/
EVENTHANDLER_INVOKE(process_fork, p1, p2, fr->fr_flags);
/*
* Set the child start time and mark the process as being complete.
*/
PROC_LOCK(p2);
PROC_LOCK(p1);
microuptime(&p2->p_stats->p_start);
PROC_SLOCK(p2);
p2->p_state = PRS_NORMAL;
PROC_SUNLOCK(p2);
#ifdef KDTRACE_HOOKS
/*
* Tell the DTrace fasttrap provider about the new process so that any
* tracepoints inherited from the parent can be removed. We have to do
* this only after p_state is PRS_NORMAL since the fasttrap module will
* use pfind() later on.
*/
if ((fr->fr_flags & RFMEM) == 0 && dtrace_fasttrap_fork)
dtrace_fasttrap_fork(p1, p2);
#endif
/*
* Hold the process so that it cannot exit after we make it runnable,
* but before we wait for the debugger.
*/
_PHOLD(p2);
if (p1->p_ptevents & PTRACE_FORK) {
/*
* Arrange for debugger to receive the fork event.
*
* We can report PL_FLAG_FORKED regardless of
* P_FOLLOWFORK settings, but it does not make a sense
* for runaway child.
*/
td->td_dbgflags |= TDB_FORK;
td->td_dbg_forked = p2->p_pid;
td2->td_dbgflags |= TDB_STOPATFORK;
}
if (fr->fr_flags & RFPPWAIT) {
td->td_pflags |= TDP_RFPPWAIT;
td->td_rfppwait_p = p2;
td->td_dbgflags |= TDB_VFORK;
}
PROC_UNLOCK(p2);
1994-05-24 10:09:53 +00:00
/*
* Now can be swapped.
*/
_PRELE(p1);
PROC_UNLOCK(p1);
1994-05-24 10:09:53 +00:00
/*
* Tell any interested parties about the new process.
*/
When filt_proc() removes event from the knlist due to the process exiting (NOTE_EXIT->knlist_remove_inevent()), two things happen: - knote kn_knlist pointer is reset - INFLUX knote is removed from the process knlist. And, there are two consequences: - KN_LIST_UNLOCK() on such knote is nop - there is nothing which would block exit1() from processing past the knlist_destroy() (and knlist_destroy() resets knlist lock pointers). Both consequences result either in leaked process lock, or dereferencing NULL function pointers for locking. Handle this by stopping embedding the process knlist into struct proc. Instead, the knlist is allocated together with struct proc, but marked as autodestroy on the zombie reap, by knlist_detach() function. The knlist is freed when last kevent is removed from the list, in particular, at the zombie reap time if the list is empty. As result, the knlist_remove_inevent() is no longer needed and removed. Other changes: In filt_procattach(), clear NOTE_EXEC and NOTE_FORK desired events from kn_sfflags for knote registered by kernel to only get NOTE_CHILD notifications. The flags leak resulted in excessive NOTE_EXEC/NOTE_FORK reports. Fix immediate note activation in filt_procattach(). Condition should be either the immediate CHILD_NOTE activation, or immediate NOTE_EXIT report for the exiting process. In knote_fork(), do not perform racy check for KN_INFLUX before kq lock is taken. Besides being racy, it did not accounted for notes just added by scan (KN_SCAN). Some minor and incomplete style fixes. Analyzed and tested by: Eric Badger <eric@badgerio.us> Reviewed by: jhb Sponsored by: The FreeBSD Foundation MFC after: 2 weeks Approved by: re (gjb) Differential revision: https://reviews.freebsd.org/D6859
2016-06-27 21:52:17 +00:00
knote_fork(p1->p_klist, p2->p_pid);
SDT_PROBE3(proc, , , create, p2, p1, fr->fr_flags);
if (fr->fr_flags & RFPROCDESC) {
procdesc_finit(p2->p_procdesc, fp_procdesc);
fdrop(fp_procdesc, td);
}
if ((fr->fr_flags & RFSTOPPED) == 0) {
/*
* If RFSTOPPED not requested, make child runnable and
* add to run queue.
*/
thread_lock(td2);
TD_SET_CAN_RUN(td2);
sched_add(td2, SRQ_BORING);
thread_unlock(td2);
if (fr->fr_pidp != NULL)
*fr->fr_pidp = p2->p_pid;
} else {
*fr->fr_procp = p2;
}
PROC_LOCK(p2);
/*
* Wait until debugger is attached to child.
*/
while (td2->td_proc == p2 && (td2->td_dbgflags & TDB_STOPATFORK) != 0)
cv_wait(&p2->p_dbgwait, &p2->p_mtx);
_PRELE(p2);
racct_proc_fork_done(p2);
PROC_UNLOCK(p2);
}
int
fork1(struct thread *td, struct fork_req *fr)
{
struct proc *p1, *newproc;
struct thread *td2;
struct vmspace *vm2;
struct file *fp_procdesc;
vm_ooffset_t mem_charged;
int error, nprocs_new, ok;
static int curfail;
static struct timeval lastfail;
int flags, pages;
flags = fr->fr_flags;
pages = fr->fr_pages;
if ((flags & RFSTOPPED) != 0)
MPASS(fr->fr_procp != NULL && fr->fr_pidp == NULL);
else
MPASS(fr->fr_procp == NULL);
/* Check for the undefined or unimplemented flags. */
if ((flags & ~(RFFLAGS | RFTSIGFLAGS(RFTSIGMASK))) != 0)
return (EINVAL);
/* Signal value requires RFTSIGZMB. */
if ((flags & RFTSIGFLAGS(RFTSIGMASK)) != 0 && (flags & RFTSIGZMB) == 0)
return (EINVAL);
/* Can't copy and clear. */
if ((flags & (RFFDG|RFCFDG)) == (RFFDG|RFCFDG))
return (EINVAL);
/* Check the validity of the signal number. */
if ((flags & RFTSIGZMB) != 0 && (u_int)RFTSIGNUM(flags) > _SIG_MAXSIG)
return (EINVAL);
if ((flags & RFPROCDESC) != 0) {
/* Can't not create a process yet get a process descriptor. */
if ((flags & RFPROC) == 0)
return (EINVAL);
/* Must provide a place to put a procdesc if creating one. */
if (fr->fr_pd_fd == NULL)
return (EINVAL);
/* Check if we are using supported flags. */
if ((fr->fr_pd_flags & ~PD_ALLOWED_AT_FORK) != 0)
return (EINVAL);
}
p1 = td->td_proc;
/*
* Here we don't create a new process, but we divorce
* certain parts of a process from itself.
*/
if ((flags & RFPROC) == 0) {
if (fr->fr_procp != NULL)
*fr->fr_procp = NULL;
else if (fr->fr_pidp != NULL)
*fr->fr_pidp = 0;
return (fork_norfproc(td, flags));
}
fp_procdesc = NULL;
newproc = NULL;
vm2 = NULL;
/*
* Increment the nprocs resource before allocations occur.
* Although process entries are dynamically created, we still
* keep a global limit on the maximum number we will
* create. There are hard-limits as to the number of processes
* that can run, established by the KVA and memory usage for
* the process data.
*
* Don't allow a nonprivileged user to use the last ten
* processes; don't let root exceed the limit.
*/
nprocs_new = atomic_fetchadd_int(&nprocs, 1) + 1;
if ((nprocs_new >= maxproc - 10 && priv_check_cred(td->td_ucred,
PRIV_MAXPROC, 0) != 0) || nprocs_new >= maxproc) {
error = EAGAIN;
sx_xlock(&allproc_lock);
if (ppsratecheck(&lastfail, &curfail, 1)) {
printf("maxproc limit exceeded by uid %u (pid %d); "
"see tuning(7) and login.conf(5)\n",
td->td_ucred->cr_ruid, p1->p_pid);
}
sx_xunlock(&allproc_lock);
goto fail2;
}
/*
* If required, create a process descriptor in the parent first; we
* will abandon it if something goes wrong. We don't finit() until
* later.
*/
if (flags & RFPROCDESC) {
error = procdesc_falloc(td, &fp_procdesc, fr->fr_pd_fd,
fr->fr_pd_flags, fr->fr_pd_fcaps);
if (error != 0)
goto fail2;
}
mem_charged = 0;
if (pages == 0)
pages = kstack_pages;
/* Allocate new proc. */
newproc = uma_zalloc(proc_zone, M_WAITOK);
td2 = FIRST_THREAD_IN_PROC(newproc);
if (td2 == NULL) {
td2 = thread_alloc(pages);
if (td2 == NULL) {
error = ENOMEM;
goto fail2;
}
proc_linkup(newproc, td2);
} else {
if (td2->td_kstack == 0 || td2->td_kstack_pages != pages) {
if (td2->td_kstack != 0)
vm_thread_dispose(td2);
if (!thread_alloc_stack(td2, pages)) {
error = ENOMEM;
goto fail2;
}
}
}
if ((flags & RFMEM) == 0) {
vm2 = vmspace_fork(p1->p_vmspace, &mem_charged);
if (vm2 == NULL) {
error = ENOMEM;
goto fail2;
}
if (!swap_reserve(mem_charged)) {
/*
* The swap reservation failed. The accounting
* from the entries of the copied vm2 will be
* subtracted in vmspace_free(), so force the
* reservation there.
*/
swap_reserve_force(mem_charged);
error = ENOMEM;
goto fail2;
}
} else
vm2 = NULL;
/*
* XXX: This is ugly; when we copy resource usage, we need to bump
* per-cred resource counters.
*/
proc_set_cred_init(newproc, crhold(td->td_ucred));
/*
* Initialize resource accounting for the child process.
*/
error = racct_proc_fork(p1, newproc);
if (error != 0) {
error = EAGAIN;
goto fail1;
}
#ifdef MAC
mac_proc_init(newproc);
#endif
When filt_proc() removes event from the knlist due to the process exiting (NOTE_EXIT->knlist_remove_inevent()), two things happen: - knote kn_knlist pointer is reset - INFLUX knote is removed from the process knlist. And, there are two consequences: - KN_LIST_UNLOCK() on such knote is nop - there is nothing which would block exit1() from processing past the knlist_destroy() (and knlist_destroy() resets knlist lock pointers). Both consequences result either in leaked process lock, or dereferencing NULL function pointers for locking. Handle this by stopping embedding the process knlist into struct proc. Instead, the knlist is allocated together with struct proc, but marked as autodestroy on the zombie reap, by knlist_detach() function. The knlist is freed when last kevent is removed from the list, in particular, at the zombie reap time if the list is empty. As result, the knlist_remove_inevent() is no longer needed and removed. Other changes: In filt_procattach(), clear NOTE_EXEC and NOTE_FORK desired events from kn_sfflags for knote registered by kernel to only get NOTE_CHILD notifications. The flags leak resulted in excessive NOTE_EXEC/NOTE_FORK reports. Fix immediate note activation in filt_procattach(). Condition should be either the immediate CHILD_NOTE activation, or immediate NOTE_EXIT report for the exiting process. In knote_fork(), do not perform racy check for KN_INFLUX before kq lock is taken. Besides being racy, it did not accounted for notes just added by scan (KN_SCAN). Some minor and incomplete style fixes. Analyzed and tested by: Eric Badger <eric@badgerio.us> Reviewed by: jhb Sponsored by: The FreeBSD Foundation MFC after: 2 weeks Approved by: re (gjb) Differential revision: https://reviews.freebsd.org/D6859
2016-06-27 21:52:17 +00:00
newproc->p_klist = knlist_alloc(&newproc->p_mtx);
STAILQ_INIT(&newproc->p_ktr);
/* We have to lock the process tree while we look for a pid. */
sx_slock(&proctree_lock);
sx_xlock(&allproc_lock);
/*
* Increment the count of procs running with this uid. Don't allow
* a nonprivileged user to exceed their current limit.
*
* XXXRW: Can we avoid privilege here if it's not needed?
*/
error = priv_check_cred(td->td_ucred, PRIV_PROC_LIMIT, 0);
if (error == 0)
ok = chgproccnt(td->td_ucred->cr_ruidinfo, 1, 0);
else {
ok = chgproccnt(td->td_ucred->cr_ruidinfo, 1,
lim_cur(td, RLIMIT_NPROC));
}
if (ok) {
do_fork(td, fr, newproc, td2, vm2, fp_procdesc);
return (0);
}
error = EAGAIN;
sx_sunlock(&proctree_lock);
sx_xunlock(&allproc_lock);
#ifdef MAC
mac_proc_destroy(newproc);
#endif
racct_proc_exit(newproc);
fail1:
crfree(newproc->p_ucred);
newproc->p_ucred = NULL;
fail2:
if (vm2 != NULL)
vmspace_free(vm2);
uma_zfree(proc_zone, newproc);
2013-02-17 11:47:01 +00:00
if ((flags & RFPROCDESC) != 0 && fp_procdesc != NULL) {
fdclose(td, fp_procdesc, *fr->fr_pd_fd);
fdrop(fp_procdesc, td);
}
atomic_add_int(&nprocs, -1);
pause("fork", hz / 2);
return (error);
1994-05-24 10:09:53 +00:00
}
/*
* Handle the return of a child process from fork1(). This function
* is called from the MD fork_trampoline() entry point.
*/
void
fork_exit(void (*callout)(void *, struct trapframe *), void *arg,
struct trapframe *frame)
{
struct proc *p;
struct thread *td;
struct thread *dtd;
td = curthread;
p = td->td_proc;
KASSERT(p->p_state == PRS_NORMAL, ("executing process is still new"));
CTR4(KTR_PROC, "fork_exit: new thread %p (td_sched %p, pid %d, %s)",
td, td_get_sched(td), p->p_pid, td->td_name);
sched_fork_exit(td);
/*
* Processes normally resume in mi_switch() after being
* cpu_switch()'ed to, but when children start up they arrive here
* instead, so we must do much the same things as mi_switch() would.
*/
if ((dtd = PCPU_GET(deadthread))) {
PCPU_SET(deadthread, NULL);
thread_stash(dtd);
}
thread_unlock(td);
/*
* cpu_fork_kthread_handler intercepts this function call to
* have this call a non-return function to stay in kernel mode.
* initproc has its own fork handler, but it does return.
*/
KASSERT(callout != NULL, ("NULL callout in fork_exit"));
callout(arg, frame);
/*
* Check if a kernel thread misbehaved and returned from its main
* function.
*/
if (p->p_flag & P_KPROC) {
printf("Kernel thread \"%s\" (pid %d) exited prematurely.\n",
td->td_name, p->p_pid);
kthread_exit();
}
mtx_assert(&Giant, MA_NOTOWNED);
if (p->p_sysent->sv_schedtail != NULL)
(p->p_sysent->sv_schedtail)(td);
td->td_pflags &= ~TDP_FORKING;
}
/*
* Simplified back end of syscall(), used when returning from fork()
* directly into user mode. This function is passed in to fork_exit()
* as the first parameter and is called when returning to a new
* userland process.
*/
void
fork_return(struct thread *td, struct trapframe *frame)
{
struct proc *p, *dbg;
p = td->td_proc;
if (td->td_dbgflags & TDB_STOPATFORK) {
sx_xlock(&proctree_lock);
PROC_LOCK(p);
if (p->p_pptr->p_ptevents & PTRACE_FORK) {
/*
* If debugger still wants auto-attach for the
* parent's children, do it now.
*/
dbg = p->p_pptr->p_pptr;
proc_set_traced(p, true);
CTR2(KTR_PTRACE,
"fork_return: attaching to new child pid %d: oppid %d",
p->p_pid, p->p_oppid);
proc_reparent(p, dbg);
sx_xunlock(&proctree_lock);
When a debugger attaches to the process, SIGSTOP is sent to the target. Due to a way issignal() selects the next signal to deliver and report, if the simultaneous or already pending another signal exists, that signal might be reported by the next waitpid(2) call. This causes minor annoyance for debuggers, which must be prepared to take any signal as the first event, then filter SIGSTOP later. More importantly, for tools like gcore(1), which attach and then detach without processing events, SIGSTOP might leak to be delivered after PT_DETACH. This results in the process being unintentionally stopped after detach, which is fatal for automatic tools. The solution is to force SIGSTOP to be the first signal reported after the attach. Attach code is modified to set P2_PTRACE_FSTP to indicate that the attaching ritual was not yet finished, and issignal() prefers SIGSTOP in that condition. Also, the thread which handles P2_PTRACE_FSTP is made to guarantee to own p_xthread during the first waitpid(2). All that ensures that SIGSTOP is consumed first. Additionally, if P2_PTRACE_FSTP is still set on detach, which means that waitpid(2) was not called at all, SIGSTOP is removed from the queue, ensuring that the process is resumed on detach. In issignal(), when acting on STOPing signals, remove the signal from queue before suspending. Otherwise parallel attach could result in ptracestop() acting on that STOP as if it was the STOP signal from the attach. Then SIGSTOP from attach leaks again. As a minor refactoring, some bits of the common attach code is moved to new helper proc_set_traced(). Reported by: markj Reviewed by: jhb, markj Tested by: pho Sponsored by: The FreeBSD Foundation MFC after: 2 weeks Differential revision: https://reviews.freebsd.org/D7256
2016-07-28 08:41:13 +00:00
td->td_dbgflags |= TDB_CHILD | TDB_SCX | TDB_FSTP;
ptracestop(td, SIGSTOP);
td->td_dbgflags &= ~(TDB_CHILD | TDB_SCX);
} else {
/*
* ... otherwise clear the request.
*/
sx_xunlock(&proctree_lock);
td->td_dbgflags &= ~TDB_STOPATFORK;
cv_broadcast(&p->p_dbgwait);
}
PROC_UNLOCK(p);
} else if (p->p_flag & P_TRACED || td->td_dbgflags & TDB_BORN) {
/*
* This is the start of a new thread in a traced
* process. Report a system call exit event.
*/
PROC_LOCK(p);
td->td_dbgflags |= TDB_SCX;
_STOPEVENT(p, S_SCX, td->td_dbg_sc_code);
if ((p->p_ptevents & PTRACE_SCX) != 0 ||
(td->td_dbgflags & TDB_BORN) != 0)
ptracestop(td, SIGTRAP);
td->td_dbgflags &= ~(TDB_SCX | TDB_BORN);
PROC_UNLOCK(p);
}
userret(td, frame);
#ifdef KTRACE
if (KTRPOINT(td, KTR_SYSRET))
ktrsysret(SYS_fork, 0, 0);
#endif
}