calls it rather than obtaining and releasing it a lot in proc_compare.
- Collect all of the data gathering and stick it just after the
proc_compare loop. This way, we only have to grab sched_lock once now
when handling SIGINFO. All the printf's are done after the values are
calculated.
Submitted mostly by: bde
The counters are incremented when a thread goes to sleep and decremented
either when a thread is woken up by another thread or when the sleep
times out. There existed a race where the sleep count could be decremented
twice resulting in an eventual underflow.
Move the decrementing of the "counters" to the thread initiating the sleep
and thus remedy the problem.
SIGXCPU signal, and killing of processes that exceed their allowed run
time until they can play nice with sched_lock. Right now they are just
potentital panics waiting to happen. The printf() has bitten several
people.
not hold sched_lock while calling ttyprintf(). If we are on a serial
console, then ttyprintf() will end up getting the sio lock, resulting in
a lock order violation.
Noticed by: des
This is an odd one. This patch appears to fix a panic related to background
bitmap writes (for FFS), though neither Kirk, Ian, or I can figure out how
B_CLUSTEROK could possibly be set on a bitmap block to cause the clustering
code to improperly cluster with a buffer undergoing a background write.
In anycase, the clustering code is very fragile and this patch helps with
that, as well as possibly fixing a bug Andre was having.
Suggested by: Ian Dowse <iedowse@maths.tcd.ie>
Testing by: Andre Albsmeier <andre.albsmeier@mchp.siemens.de>
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
allocation routines are being called safely. Since we drop our relevant
mbuf mutex and acquire Giant before we call kmem_malloc(), we have
to make sure that this does not pave the way for a fatal lock order
reversal. Check that either Giant is already held (in which case it's safe
to grab it again and recurse on it) or, if Giant is not held, that no
other locks are held before we try to acquire Giant.
Similarily, add a KASSERT valid in the WITNESS case in m_reclaim() to
nail callers who end up in m_reclaim() and hold a lock.
Pointed out by: jhb
while we are copying it to the kinfo_proc structure.
- Test against p_stat to see if we are blocked on a mutex.
- Terminate ki_mtxname with a null char rather than ki_wmesg.
m_reclaim() and re-acquire it when m_reclaim() returns. This means that
we now call the drain routines without holding the mutex lock and
recursing into it. This was done for mainly two reasons:
(i) Avoid the long recursion; long recursions are typically bad and this
is the case here because we block all other code from freeing mbufs
if they need to. Doing that is kind of counter-productive, since we're
really hoping that someone will free.
(ii) More importantly, avoid a potential lock order reversal. Right now,
not all the locks have been added to our networking code; but
without this change, we're introducing the possibility for deadlock.
Consider for example ip_drain(). We will likely eventually introduce
a lock for ipq there, and so ip_freef() will be called with ipq lock
held. But, ip_freef() calls m_freem() which in turn acquires the
mmbfree lock. Since we were previously calling ip_drain() with mmbfree
held, our lock order would be: mmbfree->ipq->mmbfree. Some other code
may very well lock ipq first and then call ip_freef(). This would
result in the regular lock order, ipq->mmbfree. Clearly, we have
deadlock if one thread acquires the ipq lock and sits waiting for
mmbfree while another thread calling m_reclaim() acquires mmbfree
and sits waiting for the ipq lock.
Also, make sure to add a comment above m_reclaim()'s definition briefly
explaining this. Also document this above the call to m_reclaim() in
m_mballoc_wait().
Suggested and reviewed by: alfred
the pipe, then we were corrupting the pipe_zone free list by calling
pipeclose on rpipe twice. NULL out rpipe to avoid this.
Reviewed by: dillon
Reviewed by: iedowse
- Provide TUNABLE_INT() hooks for ktr_cpumask, ktr_mask, and ktr_verbose
so that they can be set from the loader by their respective sysctl names.
For example, to turn on KTR_INTR and KTR_PROC in ktr_mask, one could
stick 'debug.ktr.mask="0x1200"' in /boot/loader.conf.
o Use 8 space hard tabs
o Eliminate trailing white space (while I'm here, just in a couple of places)
o wrap mostly at 80 columns (printf literal strings being the notable
exception)
o use return (foo) consistantly
o use 0 vs NULL more consistantly
o use queue(3) xxx_FOREACH macros where appropriate (some places used it
before, others didn't).
o use BSD line continuation parameters
Pendants will likely notice minor style(9) violations, but for the
most part the file now looks much much closer to style(9) and is
mostly self-consistant.
Approved in principle by: dfr
Reviewed by: md5 (no changes to the .o)
to the SYSCTL_ADD_FOO() macros is a constant that should be turned into
a string via the pre-processor. Instead, require it to be an explicit
string so that names can be generated on the fly.
- Make some of the char * arguments to sysctl_add_oid() const to quiet
warnings.
file.
While there fix the layout of function headers (noticable in long headers)
Fix up some style nits. It's Perl and should be written in that style.
out: label in psignal() did not grab sched_lock before trying to release
it. Also, the previous version had several cases where it grabbed
sched_lock before jumping to out: unneccessarily, so rework this a bit.
The runfast: and out: labels must be called with sched_lock released, and
the run: label must be called with it held. Appropriate mtx_assert()'s
have been added that should catch any bugs that may still be in this
code.
Noticed by: bde
attaching to running processes, it completely breaks normal debugging.
A better fix is in the works, but cannot be properly tested until
the problem with gdb hanging the system in -current is solved.
process. This fixes a problem when attaching to a process in gdb
and the process staying in the STOP'd state after quiting gdb.
This whole process seems a bit suspect, but this seems to work.
Reviewed by: peter
in 4.2-REL which I ripped out in -stable and -current when implementing the
low-memory handling solution. However, maxlaunder turns out to be the saving
grace in certain very heavily loaded systems (e.g. newsreader box). The new
algorithm limits the number of pages laundered in the first pageout daemon
pass. If that is not sufficient then suceessive will be run without any
limit.
Write I/O is now pipelined using two sysctls, vfs.lorunningspace and
vfs.hirunningspace. This prevents excessive buffered writes in the
disk queues which cause long (multi-second) delays for reads. It leads
to more stable (less jerky) and generally faster I/O streaming to disk
by allowing required read ops (e.g. for indirect blocks and such) to occur
without interrupting the write stream, amoung other things.
NOTE: eventually, filesystem write I/O pipelining needs to be done on a
per-device basis. At the moment it is globalized.
