VM pages into mbufs as it can -- up to the free send socket buffer space.
The outer loop then drops the whole mbuf chain into the send socket buffer,
calls tcp_output() on it and then waits until 50% of the socket buffer are
free again to repeat the cycle. This way tcp_output() gets the full amount
of data to work with and can issue up to 64K sends for TSO to chop up in
the network adapter without using any CPU cycles. Thus it gets very efficient
especially with the readahead the VM and I/O system do.
The previous sendfile(2) code simply looped over the file, turned each 4K
page into an mbuf and sent it off. This had the effect that TSO could only
generate 2 packets per send instead of up to 44 at its maximum of 64K.
Add experimental SF_MNOWAIT flag to sendfile(2) to return ENOMEM instead of
sleeping on mbuf allocation failures.
Benchmarking shows significant improvements (95% confidence):
45% less cpu (or 1.81 times better) with new sendfile vs. old sendfile (non-TSO)
83% less cpu (or 5.7 times better) with new sendfile vs. old sendfile (TSO)
(Sender AMD Opteron 852 (2.6GHz) with em(4) PCI-X-133 interface and receiver
DELL Poweredge SC1425 P-IV Xeon 3.2GHz with em(4) LOM connected back to back
at 1000Base-TX full duplex.)
Sponsored by: TCP/IP Optimization Fundraise 2005
MFC after: 3 month
a lock to prevent interspersed strings written from different CPUs
at the same time.
To avoid putting a buffer on the stack or having to malloc one,
space is incorporated in the per-cpu structure. The buffer
size if 128 bytes; chosen because it's the next power of 2 size
up from 80 characters.
String writes to the console are buffered up the end of the line
or until the buffer fills. Then the buffer is flushed to all
console devices.
Existing low level console output via cnputc() is unaffected by
this change. ithread calls to log() are also unaffected to avoid
blocking those threads.
A minor change to the behaviour in a panic situation is that
console output will still be buffered, but won't be written to
a tty as before. This should prevent interspersed panic output
as a number of CPUs panic before we end up single threaded
running ddb.
Reviewed by: scottl, jhb
MFC after: 2 weeks
- Add FS_GJOURNAL flag which enables gjournal support on a file system.
- Add cg_unrefs field to the cylinder group structure which holds
number of unreferenced (orphaned) inodes in the given cylinder group.
- Add fs_unrefs field to the super block structure which holds
total number of unreferenced (orphaned) inodes.
- When file or a directory is orphaned (last reference is removed, but
object is still open), increase fs_unrefs and cg_unrefs fields,
which is a hint for fsck in which cylinder groups looks for such
(orphaned) objects.
- When file is last closed, decrease {fs,cg}_unrefs fields.
- Add VV_DELETED vnode flag which points at orphaned objects.
Sponsored by: home.pl
Call vfs_setdirty_locked_object() from vfs_busy_pages() instead of
vfs_setdirty(), thereby eliminating a second acquisition and release
of the same vm object lock.
queues lock to BIO_READ operations. Recent changes to the implementation
of the per-page flags have eliminated the need for the page queues lock
in the other cases.
to twice unlock the vnode. Check that ni_vp and ni_dvp are different before
doing second unlock.
Reviewed by: rwatson
Approved by: pjd (mentor)
MFC after: 1 week
counters of allocs/frees/use for each malloc type to calculating InUse,
MemUse, and Requests as displayed by the userspace vmstat -m. This is
more useful when debugging malloc(9)-related memory leaks, where the
count of allocs/frees may not usefully reflect that current memory
allocation (i.e., when highly variable size allocations occur with the
same malloc type, such as with contigmalloc).
MFC after: 3 days
Limitations observed by: scottl
to and from struct timespec, to replace the crummy conversion
function which have been copy&pasted into three different
filesystems already.
Apart from general crummyness as indicated by code like:
for (year = 1970;; year++) {
inc = year & 0x03 ? 365 : 366;
if (days < inc)
break;
days -= inc;
}
They also contain specialized crummyness which tries to compensate
for the general crummyness by caching recent conversion results,
with no regard for locking or consistency.
These replacement functions are smaller, O(1) and handle the Y2.1K
leap-year correctly.
Ideally, these functions should live in a module of their own,
which the three offending filesystems would depend on, but the
size is 877 bytes of code (on i386), so that would be false
economy.
begun with a repo-copy of mac.h to mac_framework.h. sys/mac.h now
contains the userspace and user<->kernel API and definitions, with all
in-kernel interfaces moved to mac_framework.h, which is now included
across most of the kernel instead.
This change is the first step in a larger cleanup and sweep of MAC
Framework interfaces in the kernel, and will not be MFC'd.
Obtained from: TrustedBSD Project
Sponsored by: SPARTA
processes. It was originally added back when support for Linux threads
(and thus shared sigacts objects) was added, but no one knows why. My
guess is that at some point during the Linux threads patches, the sigacts
object was torn down during exit1(), so this check was added to prevent
a panic for that race. However, the stuff that was actually committed to
the tree doesn't teardown sigacts until wait() making the above race moot.
Re-allowing signals here lets one interrupt a NFS request during process
teardown (such as closing descriptors) on an interruptible mount.
Requested by: kib (long time ago)
MFC after: 1 week
vnode' v_rdev and increment the dev threadcount , as well as clear it
(in devfs_reclaim) under the dev_lock().
Reviewed by: tegge
Approved by: pjd (mentor)
- Count (scheduling of) software interrupts (SWIs) as SWIs, not as
hardware interrupts.
- Don't count (scheduling of) delayed SWIs as interrupts at all, since
in the delayed case it is expected that there are many more scheduling
calls than handling calls. Perhaps all interrupts should be counted
only when they are handled, but it is only counts of delayed SWIs that
shouldn never be combined with the other counts.
subr_trap.c:
- Count (handling of) Asynchronous System Traps (ASTs) as traps, not as
software interrupts.
Before these changes, the counter for SWIs only counted ASTs, and SWIs
weren't counted separately, but a subcounter for ASTs alone is less
needed than for most other exception sources.
4.4BSD-Lite uses the counters for similar things (actually matching
their names) on its main arches (hp300, ..., !i386) where more of the
exceptions are in hardware.
Implement the linux_io_* syscalls (AIO). They are only enabled if the native
AIO code is available (either compiled in to the kernel or as a module) at
the time the functions are used. If the AIO stuff is not available there
will be a ENOSYS.
From the submitter:
---snip---
DESIGN NOTES:
1. Linux permits a process to own multiple AIO queues (distinguished by
"context"), but FreeBSD creates only one single AIO queue per process.
My code maintains a request queue (STAILQ of queue(3)) per "context",
and throws all AIO requests of all contexts owned by a process into
the single FreeBSD per-process AIO queue.
When the process calls io_destroy(2), io_getevents(2), io_submit(2) and
io_cancel(2), my code can pick out requests owned by the specified context
from the single FreeBSD per-process AIO queue according to the per-context
request queues maintained by my code.
2. The request queue maintained by my code stores contrast information between
Linux IO control blocks (struct linux_iocb) and FreeBSD IO control blocks
(struct aiocb). FreeBSD IO control block actually exists in userland memory
space, required by FreeBSD native aio_XXXXXX(2).
3. It is quite troubling that the function io_getevents() of libaio-0.3.105
needs to use Linux-specific "struct aio_ring", which is a partial mirror
of context in user space. I would rather take the address of context in
kernel as the context ID, but the io_getevents() of libaio forces me to
take the address of the "ring" in user space as the context ID.
To my surprise, one comment line in the file "io_getevents.c" of
libaio-0.3.105 reads:
Ben will hate me for this
REFERENCE:
1. Linux kernel source code: http://www.kernel.org/pub/linux/kernel/v2.6/
(include/linux/aio_abi.h, fs/aio.c)
2. Linux manual pages: http://www.kernel.org/pub/linux/docs/manpages/
(io_setup(2), io_destroy(2), io_getevents(2), io_submit(2), io_cancel(2))
3. Linux Scalability Effort: http://lse.sourceforge.net/io/aio.html
The design notes: http://lse.sourceforge.net/io/aionotes.txt
4. The package libaio, both source and binary:
http://rpmfind.net/linux/rpm2html/search.php?query=libaio
Simple transparent interface to Linux AIO system calls.
5. Libaio-oracle: http://oss.oracle.com/projects/libaio-oracle/
POSIX AIO implementation based on Linux AIO system calls (depending on
libaio).
---snip---
Submitted by: Li, Xiao <intron@intron.ac>
if backward copatibility options are present) from attempting
to free memory that wasn't allocated. This is an old bug, and
previously it would attempt to free a null pointer. I noticed
this bug when working on the previous revision, but forgot to
fix it.
Security: local DoS
Reported by: Peter Holm
MFC after: 3 days
method is defined, to avoid memory being modified after free.
Temporarily increase refcount in destroy_devl() to avoid a double free
if dev_rel() is called while waiting for thread count to reach zero.
removals, including failures, into the callwheel.
XXX: Most of the CTR() macros are called with callout_lock spin mutex
held, thus won't be logged into file, if KTR_ALQ is used. Moving the
CTR() macros out from the spinlocked code would require copying of all
arguments. I'm too lazy to do this.
calls are not used by libthr in RELENG_6 and HEAD, it is only used by
the libthr in RELENG-5, the _umtx_op system call can do more incremental
dirty works than these two system calls without having to introduce new
system calls or throw away old system calls when things are going on.
unmount when mp structure is reused while waiting for coveredvp lock.
Introduce struct mount generation count, increment it on each reuse and
compare the generations before and after obtaining the coveredvp lock.
Reviewed by: tegge, pjd
Approved by: pjd (mentor)
MFC after: 2 weeks
Split subr_clock.c in two parts (by repo-copy):
subr_clock.c contains generic RTC and calendaric stuff. etc.
subr_rtc.c contains the newbus'ified RTC interface.
Centralize the machdep.{adjkerntz,disable_rtc_set,wall_cmos_clock}
sysctls and associated variables into subr_clock.c. They are
not machine dependent and we have generic code that relies on being
present so they are not even optional.
