instead of ng_send_data().
The latter could lead to running the IP stack at splimp
instead of splnet, (among other problems) (that MAY be safe
but I wouldn't count on it).
Noticed while preparing a new set of netgraph stuff.
counter register in-CPU.
This is to be used as a fast "timer", where linearity is more important
than time, and multiple lines in the linearity caused by multiple CPUs
in an SMP machine is not a problem.
This adds no code whatsoever to the FreeBSD kernel until it is actually
used, and then as a single-instruction inline routine (except for the
80386 and 80486 where it is some more inline code around nanotime(9).
Reviewed by: bde, kris, jhb
a kevent upon completion of the I/O. Specifically, introduce a new type
of sigevent notification, SIGEV_EVENT. If sigev_notify is SIGEV_EVENT,
then sigev_notify_kqueue names the kqueue that should receive the event
and sigev_value contains the "void *" is copied into the kevent's udata
field.
In contrast to the existing interface, this one: 1) works on
the Alpha 2) avoids the extra copyin() call for the kevent because all
of the information needed is in the sigevent and 3) could be
applied to request a single kevent upon completion of an entire lio_listio().
Reviewed by: jlemon
- move the call to cia_init_sgmap() to after we've determined if we're a pyxis
- convert needed splhigh() in cia_sgmap_invalidate_pyxis() to disable_intr()
Previously, any isa DMA on a pyxis based machine would cause a panic
in cia_sgmap_invalidate_pyxis() because the pyxis workaround was never
setup.
- while i'm at it, convert needed splhigh() in cia_swiz_set_hae_mem to
disable_intr()
in the face of multiple processes doing massive numbers of filesystem
operations. While this patch will work in nearly all situations, there
are still some perverse workloads that can overwhelm the system.
Detecting and handling these perverse workloads will be the subject
of another patch.
Reviewed by: Paul Saab <ps@yahoo-inc.com>
Obtained from: Ethan Solomita <ethan@geocast.com>
could not compress into clusters. This could result in lots of
wasted clusters while recieving small packets from an interface
that uses clusters for all it's packets.
Patch is partially from BSDi (limiting the size of the copy) and
based on a patch for 4.1 by Ian Dowse <iedowse@maths.tcd.ie> and
myself.
Reviewed by: bmilekic
Obtained From: BSDi
Submitted by: iedowse
- 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
Removed most of the hacks that were trying to deal with low-memory
situations prior to now.
The new code is based on the concept that I/O must be able to function in
a low memory situation. All major modules related to I/O (except
networking) have been adjusted to allow allocation out of the system
reserve memory pool. These modules now detect a low memory situation but
rather then block they instead continue to operate, then return resources
to the memory pool instead of cache them or leave them wired.
Code has been added to stall in a low-memory situation prior to a vnode
being locked.
Thus situations where a process blocks in a low-memory condition while
holding a locked vnode have been reduced to near nothing. Not only will
I/O continue to operate, but many prior deadlock conditions simply no
longer exist.
Implement a number of VFS/BIO fixes
(found by Ian): in biodone(), bogus-page replacement code, the loop
was not properly incrementing loop variables prior to a continue
statement. We do not believe this code can be hit anyway but we
aren't taking any chances. We'll turn the whole section into a
panic (as it already is in brelse()) after the release is rolled.
In biodone(), the foff calculation was incorrectly
clamped to the iosize, causing the wrong foff to be calculated
for pages in the case of an I/O error or biodone() called without
initiating I/O. The problem always caused a panic before. Now it
doesn't. The problem is mainly an issue with NFS.
Fixed casts for ~PAGE_MASK. This code worked properly before only
because the calculations use signed arithmatic. Better to properly
extend PAGE_MASK first before inverting it for the 64 bit masking
op.
In brelse(), the bogus_page fixup code was improperly throwing
away the original contents of 'm' when it did the j-loop to
fix the bogus pages. The result was that it would potentially
invalidate parts of the *WRONG* page(!), leading to corruption.
There may still be cases where a background bitmap write is
being duplicated, causing potential corruption. We have identified
a potentially serious bug related to this but the fix is still TBD.
So instead this patch contains a KASSERT to detect the problem
and panic the machine rather then continue to corrupt the filesystem.
The problem does not occur very often.. it is very hard to
reproduce, and it may or may not be the cause of the corruption
people have reported.
Review by: (VFS/BIO: mckusick, Ian Dowse <iedowse@maths.tcd.ie>)
Testing by: (VM/Deadlock) Paul Saab <ps@yahoo-inc.com>
Pre-rfork code assumed inherent locking of a process's file descriptor
array. However, with the advent of rfork() the file descriptor table
could be shared between processes. This patch closes over a dozen
serious race conditions related to one thread manipulating the table
(e.g. closing or dup()ing a descriptor) while another is blocked in
an open(), close(), fcntl(), read(), write(), etc...
PR: kern/11629
Discussed with: Alexander Viro <viro@math.psu.edu>
are in softclock() for a long time. The old code already did an
splx()/slphigh() pair here, I just missed adding in the equivalent mutex
operations on sched_lock earlier.
This makes crash recovery work for stripe sizes that are not multiples of
DEFAULT_REVIVE_BLOCKSIZE (currently 64 kB).
While we're here, fix a few cosmetic nits.
Reviewed by: grog
Sponsored by: Enitel ASA (http://www.enitel.no/)
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
instead of DIAGNOSTIC.
- Remove the p_wchan check as it no longer applies since a process may be
switched out during CURSIG() within msleep() or mawait().
- Remove an extra sanity check only needed during the early SMPng work.
with Julian and Archie.
Implement a new ``sizedstring'' parse type for dealing with field pairs
consisting of a uint16_t followed by a data field of that size, and use
this to deal with the data_len and data fields.
Written by: Archie with some input by me
Agreed in principle by: julian
untimeout() not being called with Giant in those functions. For now,
use the sched_lock to protect the callout wheel in softclock() and in
the various timeout and callout functions.
Noticed by: tegge
16-bit mode. Technically, pcn_probe() is destructive because once the
chip goes into 32-bit mode, the only way to get it out again is a
hardware reset. And once the device is in 32-bit mode, the lnc driver
won't be able to talk to it. So if pcn_probe() is called before the
lnc probe routine, and pcn_probe() rejects the chip as one it doesn't
support, the lnc driver will be SOL.
I don't like this. I think it's a design flaw that you can't switch
the chip out of 32-bit mode once it's selected. The only 'right'
solution is for the pcn driver to support all of the PCI devices
in 32-bit mode, however I don't have samples of all the PCnet series
cards for testing.
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
linux_rt_sendsig() and restore the same signal mask linux does
in rt_sigreturn(). This gets us saving/restoring all 64-bits of the
linux sigset_t in rt signals.
Reviewed by: marcel
idea to be holding the sched_lock while we are calling it. As such,
release sched_lock before calling CURSIG() in msleep() and mawait() and
reacquire it after CURSIG() returns.
Submitted by: witness
to our native connect(). This is required to deal with the differences
in the way linux handles connects on non-blocking sockets.
This gets the private beta of the Compaq Linux/alpha JDK working
on FreeBSD/alpha
Approved by: marcel
tsleep(). Namely, mawait() takes an extra argument which is a mutex
to drop when going to sleep. Just as with msleep(), if the priority
argument includes the PDROP flag, then the mutex will be dropped and will
not be reacquired when the process wakes up.
- Add in a backwards compatible macro await() that passes in NULL as the
mutex argument to mawait().
except that it uses the MTX_NOSWITCH flag while it releases Giant via
mtx_exit().
- Add a mtx_recursed() primitive. This primitive should only be used on
a mutex owned by the current process. It will return non-zero if the
mutex is recursively owned, or zero otherwise.
- Add two new flags MA_RECURSED and MA_NOTRECURSED that can be used in
conjuction with MA_OWNED to control the assertion checked by mtx_assert().
- Fix some of the KTR tracepoint strings to use %p when displaying the lock
field of a mutex, which is a uintptr_t.
macros which provide the same functionality and are a bit more
efficient, convert use of CIRCLEQ's in netgraph PPP code to TAILQ's.
Reviewed by: Archie Cobbs <archie@dellroad.org>
sched_lock. This is needed for kernel threads that are created before
interrupts are enabled. kthreads created by kld's that are created at
SI_SUB_KLD such as the random kthread.
Tested by: phk
macros which provide the same functionality and are a bit more
efficient, convert use of CIRCLEQ's in resource manager to TAILQ's.
Approved by: Garrett Wollman <wollman@khavrinen.lcs.mit.edu>
is to first write the deleted directory entry to disk, second write
the zero'ed inode to disk, and finally to release the freed blocks
and the inode back to the cylinder-group map. As this ordering
requires two disk writes to occur which are normally spaced about
30 seconds apart (except when memory is under duress), it takes
about a minute from the time that a file is deleted until its inode
and data blocks show up in the cylinder-group map for reallocation.
If a file has had only a brief lifetime (less than 30 seconds from
creation to deletion), neither its inode nor its directory entry
may have been written to disk. If its directory entry has not been
written to disk, then we need not wait for that directory block to
be written as the on-disk directory block does not reference the
inode. Similarly, if the allocated inode has never been written to
disk, we do not have to wait for it to be written back either as
its on-disk representation is still zero'ed out. Thus, in the case
of a short lived file, we can simply release the blocks and inode
to the cylinder-group map immediately. As the inode and its blocks
are released immediately, they are immediately available for other
uses. If they are not released for a minute, then other inodes and
blocks must be allocated for short lived files, cluttering up the
vnode and buffer caches. The previous code was a bit too aggressive
in trying to release the blocks and inode back to the cylinder-group
map resulting in their being made available when in fact the inode
on disk had not yet been zero'ed. This patch takes a more conservative
approach to doing the release which avoids doing the release prematurely.
of the data structures to include new members that weren't defined in the
manual I have.
I opted to use Doug Ambrisko's WEP patches since David Cornejo's patches
did not include the necessary changes to ancontrol(8) to actually enable
and use WEP.
NOTE: I don't currently have access to an Aironet card, so I can't test
any of this. Everything compiles and close scrutiny doesn't reveal any
obvious problems, but Murphy's Law applies. This means I will probably
leave these changes in -current for a bit longer than usual until I'm
sure they work right.
linux_sigset_t by updating the linux_sigframe struct so as to include
linux's "extramask" field. This field contains the upper 32-bits of
the sigset. extramask sits behind a linux_fpstate struct, which I've
defined primarily for padding purposes.
While we're here, define LINUX_NSIG in terms of LINUX_NBPW (32) and
LINUX_NSIG_WORDS (2).
This fixes problems where threaded apps would accumulate a large
number of zombies. This was happening because the exit signal resides
in the upper 32-bits of the sigset and was never getting unmasked by
the manager thread after the first child exited.
PR: misc/18530 (may be related, originator not yet contacted)
Reviewed by: marcel
