Aren/NetBSD
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
171,376 Commits 606 Branches 0 Tags 3.1 GiB
1805817468
Commit Graph

6392 Commits

Author SHA1 Message Date
pooka
1805817468 Adjust rump syscalls: remove a few pointless ones and add a few 
missing ones.
 2008-07-01 13:03:08 +00:00
matt
5a4f0c6b2b Change tree op members/typedefs to rbto_compare_* from rb_compare_* 2008-06-30 20:14:09 +00:00
pooka
d496f014ff Don't compile kern_lock for rump any more, it's no longer required. 
Allows us to get rid of the incorrect _RUMPKERNEL ifdefs outside sys/rump.
 2008-06-25 13:16:58 +00:00
ad
40818746b1 Use pool_cache. 2008-06-25 11:05:46 +00:00
ad
d23ae8da8d sigpending1: don't subtract the sigmask (!!). 2008-06-25 11:04:24 +00:00
ad
80892cd092 Use pool_cache. 2008-06-24 18:50:30 +00:00
ad
8c5bcce4b9 execve1: 
- Properly terminate the fake argument list.
- Fix two double-frees.
 2008-06-24 18:04:52 +00:00
ad
9993bc6d33 Nothing uses getsock/getvnode any more. 2008-06-24 11:21:46 +00:00
gmcgarry
8cead24f2a The conversion to c99-style designated initialisers brought in the gcc 
extension of specifying an array range.  Revert to runtime initialization
of limit0.

See here for details: http://gcc.gnu.org/onlinedocs/gcc/Designated-Inits.html
 2008-06-24 10:31:05 +00:00
gmcgarry
a7edd5ccdd Replace gcc-style designated initialisers with c99-style. 2008-06-24 10:27:35 +00:00
gmcgarry
8b957c9d45 ioctl commands are unsigned long. Changes ABI for fsetown() and fgetown() on 64-bit architectures. 2008-06-24 10:26:26 +00:00
gmcgarry
fb6bc48511 ioctl commands are unsigned long. 2008-06-24 10:24:21 +00:00
rmind
59a180ac8f sysctl_proc_stop: fix a lock-leak when kauth returns an error. 
From <kefren>.
 2008-06-23 20:04:36 +00:00
ad
aed68d2d63 getvnode -> fd_getvnode 2008-06-23 11:30:41 +00:00
ad
29957bec2e sys_fcntl: use l_fd, not p_fd. 2008-06-23 11:26:53 +00:00
ad
7756789d3a PR kern/38990: Unmounting a disk fails the first time 
vflush: process vrele_list at least once before looking at the mount.
 2008-06-23 11:23:39 +00:00
ad
bce675d015 When offlining a CPU, ensure that at least one other CPU within the same 
processor set remains online, otherwise the system can deadlock.
 2008-06-22 13:59:06 +00:00
ad
53db9954a3 - Ensure that at least one cpu stays in the default pset, otherwise the 
system will lock up.
- Use cpu_lock, there's no need for a seperate psets_lock.
- Add a sysctl to return the list of psets as a string.
 2008-06-22 12:59:18 +00:00
christos
1d875fc75f Adjust to separate kcpuset_t and cpuset_t. 2008-06-22 00:06:36 +00:00
christos
5e865a2f1b Also enforce that cm->cmsg_len >= CMSG_ALIGN(sizeof cmsghdr), from 
Michael van Elst
 2008-06-20 15:27:50 +00:00
christos
db982eb35b Don't require cm->cmsg_len == control->m_len, just that the cm->cmsg_len 
<= control->m_len, like FreeBSD does. Idea from Taylor R Campbell.
 2008-06-20 15:18:38 +00:00
mlelstv
8b80c3645f reassignbuf() is called by bdirty() from within biodone processing. 
It can have waiters.
 2008-06-18 18:11:25 +00:00
yamt
fff57c5525 merge yamt-pf42 branch. 
(import newer pf from OpenBSD 4.2)

ok'ed by peter@.  requested by core@
 2008-06-18 09:06:25 +00:00
christos
ffabbaf7c6 set mtime/atime properly, not backwards. 2008-06-17 21:02:08 +00:00
mlelstv
76a664c61a Drop !cv_has_waiters assertion. 
bdirty() is called from within biodone() processing before
waiters have been woken up and removed.

N.B. it is also used by smbfs.
 2008-06-17 19:14:14 +00:00
christos
f5be32608a PR/38942: Pedro F. Giffuni: no support for birthtime in utimes(2). 2008-06-17 16:18:01 +00:00
tsutsui
d7d2a525b1 Include <sys/sched.h> before <sys/syscallargs.h> for cpuset_t. 2008-06-17 16:17:21 +00:00
tsutsui
2bc3b0c694 Regen from syscalls.master rev 1.204: 
> Include <sys/sched.h> before <sys/syscallargs.h> for cpuset_t.
 2008-06-17 16:07:57 +00:00
tsutsui
33dfc34006 Include <sys/sched.h> before <sys/syscallargs.h> for cpuset_t. 2008-06-17 16:05:23 +00:00
reinoud
f6a70673ba Mark a buffer busy in getnewbuf() when it came from the pool_cache since 
its not on a free list.

Also change buf_init() to not automatically mark buffers `busy' since this
only makes sense for bufcache buffers.

Mark all buf_init'd buffers 'busy' on the places where they ought to be
flagged as such to not confuse the buffer cache.

Fixes PR 38923.
 2008-06-17 14:53:10 +00:00
ad
4810cbe2bf sleepq_block: add a comment. 2008-06-17 09:11:25 +00:00
ad
cad3a145a3 PR kern/38761: new (?) race in buffer cache code 
Back out the workaround from cv_has_waiters(), which is not longer needed.
Removal was missed earlier.
 2008-06-16 12:03:01 +00:00
ad
a9c4cefebd PR kern/38927: processes getting stuck in uvm_map (cv_timedwait), hanging 
machine

Assume that a vnode (and associated data structures) costs 2kB in the
worst imaginable case. Don't allow sysctl to set desiredvnodes to a
value that would use more than 75% of KVA or 75% of physical memory.
 2008-06-16 11:26:28 +00:00
ad
45850c3df9 PPWAIT need only be locked by proc_lock, so move it to proc::p_lflag. 2008-06-16 10:15:57 +00:00
ad
672f91757e PR kern/38761: new (?) race in buffer cache code 
- Back out the previous workaround now that the sleep queue code has
  been changed to never let the queue become empty if there are valid
  waiters.
- Use sleepq_hashlock() to improve clarity.
- Sprinkle some assertions.
 2008-06-16 10:03:47 +00:00
ad
461a5fb24b PR kern/38761: new (?) race in buffer cache code 
sleepq_changepri, sleepq_lendpri: don't let an active sleep queue head become
empty. The condvar code inspects the queue head without holding the sleep
queue lock and needs to see a non-empty queue if there are waiters.
 2008-06-16 10:02:15 +00:00
ad
5adf7333fd - PPWAIT is need only be locked by proc_lock, so move it to proc::p_lflag. 
- Remove a few needless lock acquires from exec/fork/exit.
- Sprinkle branch hints.

No functional change.
 2008-06-16 09:51:14 +00:00
ad
b0ac1133c6 Sprinkle some assertions. 2008-06-16 09:48:13 +00:00
ad
5bfa865311 Sprinkle more assertions. 2008-06-16 09:47:55 +00:00
ad
b7f5063255 lwp_lock_retry: return a pointer to the lock acquired. No functional change. 2008-06-16 09:45:20 +00:00
rmind
481ae1556f - Add general cpuset macros. 
- Use kcpuset name for kernel-only functions.
- Use cpuid_t to specify CPU ID.
- Unify all cpuset users.

API is expected to be stable now.
 2008-06-16 01:41:20 +00:00
christos
09e715922e regen 2008-06-16 01:00:21 +00:00
uebayasi
9d4479f7e9 Typo. 2008-06-16 00:54:53 +00:00
christos
20d8fbfbe4 typo 2008-06-16 00:54:10 +00:00
christos
88c08e3b98 regen 2008-06-16 00:32:23 +00:00
christos
5c45905c42 adjust prototypes for the affinity calls. 2008-06-16 00:31:17 +00:00
rmind
a6092d3c4a gencpuset: avoid returning of unitialized error. 
sched_setaffinity: instead of E2BIG, return EINVAL in a case of invalid
CPU, which is more consistent.
 2008-06-15 23:29:09 +00:00
christos
f30b5785d5 Don't expose struct cpuset, share the l_affinity flag and only allocate it 
if we need to. This is not a compatible change, but the syscalls are new
enough and they don't need to be versioned. Approved by rmind.
 2008-06-15 20:32:57 +00:00
chris
44d939d150 Fix for biowait hangs, and possibly other condvar hangs. Also should fix 
PR kern/38761.

The condvar must access the sleepq with the sleepq lock held, doing so
is causing inconsistent sleepq state to be read.

This is because some accesses to the sleepq don't come via the cv code,
but are call directly into sleepq_changepri and sleepq_lendpri, which take
the sleepq lock, and removes then re-inserts lwps into the sleepq.

Running a build.sh with -j8 now completes on my quad-core, also tested by
Simon@ on a 8-core server and matt@ on a quad-core.

I believe there is room to be more efficient with this, as we now take the
sleepq lock for all cv_broadcast and cv_signal calls. I'll look into this
and post a diff to tech-kern.
 2008-06-15 09:56:18 +00:00
drochner
a748803de9 tighten type checking: use device_t instead of void* at some places 2008-06-11 15:56:11 +00:00