summaryrefslogtreecommitdiff
path: root/sys/miscfs
AgeCommit message (Collapse)Author
2005-05-31Protect the run queues with SCHED_LOCK, not just spl (ot nothing at all inArtur Grabowski
one case fixed here). miod@ "appears to be harmless" markus@ ok
2005-05-29sched work by niklas and art backed out; causes panicsTheo de Raadt
2005-05-26bye bye, rest in attic ad infinitum, amenPedro Martelletto
2005-05-25This patch is mortly art's work and was done *a year* ago. Art wants to thankNiklas Hallqvist
everyone for the prompt review and ok of this work ;-) Yeah, that includes me too, or maybe especially me. I am sorry. Change the sched_lock to a mutex. This fixes, among other things, the infamous "telnet localhost &" problem. The real bug in that case was that the sched_lock which is by design a non-recursive lock, was recursively acquired, and not enough releases made us hold the lock in the idle loop, blocking scheduling on the other processors. Some of the other processors would hold the biglock though, which made it impossible for cpu 0 to enter the kernel... A nice deadlock. Let me just say debugging this for days just to realize that it was all fixed in an old diff noone ever ok'd was somewhat of an anti-climax. This diff also changes splsched to be correct for all our architectures.
2005-05-24better handling of disappearing devices, in spec_close(), don't try toPedro Martelletto
double-lock the vnode if we're coming from vclean()
2005-04-21careful strlcpy and snprintf return handling; ok pedro beckTheo de Raadt
2005-04-16Remove regs and fpregs pseudo-files from procfs.Mark Kettenis
ok deraadt@, miod@
2005-04-01calling pfind() and checking that the process doesn't exist is notPedro Martelletto
enough to allow us to call vgone() from procfs_inactive(). to avoid a deadlock, check for VXLOCK as well, in case we were called from vclean(). problem report from Sho Fujita, okay tedu@.
2004-11-29Don't use seltrue() in poll ops since it requires v_rdev dereferencingAlexander Yurchenko
which is not possible here. Problem found and fixed by form@. ok millert@ fgsch@ pedro@
2004-11-18Better and simpler check for negative offset; from Christer ObergTodd C. Miller
2004-10-26typo, from Joris VinkPedro Martelletto
2004-10-19use queue.h provided macro when traversing the process list, no changesPedro Martelletto
in generated object file. from Joris Vink.
2004-09-01Make sure we don't call kernfs_xread() with a negative offset.Todd C. Miller
Problem found by Christer Oberg. OK henning@, deraadt@
2004-07-22remove p arg from fdplockTed Unangst
2004-07-11explicitly initialize a variable that is used to control the vnodePedro Martelletto
flushing loop, otherwise we could hard-lock the machine when unmounting an union filesystem ok tedu@
2004-07-10make sure we pass vop_whiteout() a locked vnode, ok tedu@Pedro Martelletto
2004-06-24This moves access to wall and uptime variables in MI code,Thorsten Lockert
encapsulating all such access into wall-defined functions that makes sure locking is done as needed. It also cleans up some uses of wall time vs. uptime some places, but there is sure to be more of these needed as well, particularily in MD code. Also, many current calls to microtime() should probably be changed to getmicrotime(), or to the {,get}microuptime() versions. ok art@ deraadt@ aaron@ matthieu@ beck@ sturm@ millert@ others "Oh, that is not your problem!" from miod@
2004-06-12fix ncpu vs ncpus glitch, spotted by Sven DehmlowTheo de Raadt
2004-06-06portal_connect needs splsoftnet.Ted Unangst
using CMSG_ALIGN was wrong, userland fires in data not so aligned. if fd_getfile returns NULL, don't try to close the fd, since it's not there.
2004-06-01use M_MISCFSMNT to allocate nullfs memory, ok tedu@Pedro Martelletto
2004-05-28fix for: !(flags && FORCECLOSE)Mike Pechkin
hackers@ ok
2004-05-26typosPedro Martelletto
2004-05-20remove more caddr_tTed Unangst
2004-05-20Properly flush instruction cache for ptrace(PT_WRTIE_{DI}, ...) on powerpcMark Kettenis
and m68k. ok drahn@, millert@
2004-05-18useless caddr_t casts removal, same sha1sPedro Martelletto
2004-05-18fix memory leak, ok tedu@Pedro Martelletto
2004-05-14Fix a bug that occurs when a FIFO is opened for writing withTodd C. Miller
O_NONBLOCK set and there are no readers. Before returning ENXIO fifo_open calls VOP_CLOSE (and hence fifo_close). However, since fi_writers has not yet been incremented, when fifo_close decements fi_writers it is one too few. This could cause qmail processes to spin, consuming all the CPU. Noticed by avsm@ and henning@, test case provided by claudio@, Ok pedro@
2004-05-14use pool for namei pathbuf. testing ok millert@ tdeval@Ted Unangst
2004-05-05make sure uio_offset is a safe value, with suggestions from millert@Ted Unangst
ok deraadt@ millert@ problem noticed by deprotect.com
2004-04-25useless castsTed Unangst
2004-04-25if you don't want the parent, don't use WANTPARENT. from pedro martellettoTed Unangst
2004-04-25only use um_cred for lookups, fixes pr 745.Ted Unangst
from pedro martelletto
2004-04-25fix typos/spelling in comments, from pedro martellettoJolan Luff
2004-04-23use CMSG macros. netbsd via pedro marteletto.Ted Unangst
2004-03-03allow force umount here tooTed Unangst
2004-03-03repair status by not repeating fields. fixes pr2101 from jim razmusTed Unangst
2004-03-03support forced unmounts. pr2394 from peter wernerTed Unangst
2004-03-03save room for nul, so we can unmount later. pr2327 from peter wernerTed Unangst
2004-03-02prevent memory leak on revoke by freeing resources in a real reclaimTed Unangst
function. also check fifoinfo is valid before deref. ok millert@
2004-02-24FIFO fixes adapted from FreeBSD:Todd C. Miller
o use different wchan string for reading and writing o make O_RDWR not block forever o remove some useless casts o reorganize the normal, blocking code path (ie: O_NONBLOCK not set) o fix select/poll semantics wrt EOF. With these changes we pass the FIFO regress. OK tedu@
2004-02-21fiddle with filedesc lock to prevent recursion. crash and fix confirmedTed Unangst
by marc balmer
2004-01-28Don't rely on vp->v_usecount to tell when we can dispose of our resourcesTodd C. Miller
in fifo_close(). If the FIFO is accessed via a layed fs (e.g. nullfs), v_usecount will always be 1. Instead, just check fip->if_readers and fip->fi_writers. If either one is non-zero we know the FIFO is in use and we should not free up its resources. Adapted from FreeBSD, NetBSD has an equivalent change (but they keep a counter instead).
2003-12-09missing ar; form Alf Schlichting <leopold-bloom@foni.net>Michael Shalayeff
2003-11-29typos from Michael Coulter;Jason McIntyre
2003-11-15set boottime's time to boottime, handy for find. requested by markus.Ted Unangst
don't bother with microtime. time should be more than enough precision. from netbsd.
2003-11-08typos from Jonathon Gray;Jason McIntyre
2003-10-24style(9)ed panics, from tom cosgrove.Ted Unangst
2003-09-23Replace select backends with poll backends. selscan() and pollscan()Todd C. Miller
now call the poll backend. With this change we implement greater poll(2) functionality instead of emulating it via the select backend. Adapted from NetBSD and including some changes from FreeBSD. Tested by many, deraadt@ OK
2003-08-15change arguments to suser. suser now takes the process, and a flagsTed Unangst
argument. old cred only calls user suser_ucred. this will allow future work to more flexibly implement the idea of a root process. looks like something i saw in freebsd, but a little different. use of suser_ucred vs suser in file system code should be looked at again, for the moment semantics remain unchanged. review and input from art@ testing and further review miod@
2003-08-14constify vfsops; tedu@ okMichael Shalayeff
generated by cgit v1.2.3 (git 2.25.1) at 2024-12-15 17:56:25 +0000