summaryrefslogtreecommitdiff
path: root/sys/kern
AgeCommit message (Collapse)Author
2001-11-15Make sure that stuff on the syncer worklist has VBIOONSYNCLIST setArtur Grabowski
and stuff that isn't on the worklist doesn't have it set.
2001-11-15New function vn_marktext - mark a vnode as executing a text image.Artur Grabowski
Use where VTEXT was set in vnode flags before. Doesn't do anything else (yet).
2001-11-14Better exit codes. ok mickey,millert,miod,deraadtHakan Olsson
2001-11-14in poll, if fd is < 0, clear revents; bjy@mogua.orgTheo de Raadt
2001-11-14Compat layer for 1k Vax aout, from NetBSD.Hugh Graham
2001-11-12Remove unnecessary check for NULL vnode in reassignbuf.Artur Grabowski
2001-11-12Bring in more changes from NetBSD. Mostly pagedaemon improvements.Artur Grabowski
2001-11-11Let ltsleep take a const wmesg.Artur Grabowski
2001-11-10Move maxdmap and maxsmap to kern_resource.cArtur Grabowski
2001-11-10Merge in some parts of the ubc work that has been done in NetBSD that are notArtur Grabowski
UBC, but prerequsites for it. - Create a daemon that processes async I/O (swap and paging in the future) requests that need processing in process context and that were processed in the pagedaemon before. - Convert some ugly ifdef DIAGNOSTIC code to less intrusive KASSERTs. - misc other cleanups.
2001-11-09Use bufpool, not malloc.Artur Grabowski
2001-11-09Create bufpool - a pool of struct bufs.Artur Grabowski
2001-11-07new flag to lockmgr. LK_RECURSEFAIL - even if the lock can recurse fail.Artur Grabowski
2001-11-07Add an alignment argument to uvm_map that specifies an alignment hintArtur Grabowski
for the virtual address.
2001-11-06Replace inclusion of <vm/foo.h> with the correct <uvm/bar.h> when necessary.Miod Vallat
(Look ma, I might have broken the tree)
2001-11-06Let fork1, uvm_fork, and cpu_fork take a function/argument pair as argument,Artur Grabowski
instead of doing fork1, cpu_set_kpc. This lets us retire cpu_set_kpc and avoid a multiprocessor race. This commit breaks vax because it doesn't look like any other arch, someone working on vax might want to look at this and try to adapt the code to be more like the rest of the world. Idea and uvm parts from NetBSD.
2001-11-06More sync to NetBSD.Artur Grabowski
- Use malloc/free instead of MALLOC/FREE for variable sized allocations. - Move the memory inheritance code to sys/mman.h and rename from VM_* to MAP_* - various cleanups and simplifications.
2001-11-05let finishdup take the proc as arguemnt instead of the filedesc.Artur Grabowski
2001-11-02Move the accounting disk space watcher into a kernel thread.Artur Grabowski
Since we do filesystem operations, we need process context. This should fix accounting on nfs.
2001-10-28we try to test things before commit, artTheo de Raadt
2001-10-28Sprinkle some assertions in the buffer code.Artur Grabowski
Add a new DEBUG function "buf_print" that prints the contents of struct buf.
2001-10-27Every file must have a license blob, even if the blob is larger than the file.Artur Grabowski
2001-10-26 - every new fd created by falloc() is marked as larval and should not be usedArtur Grabowski
any anyone. Every caller of falloc matures the fd when it's usable. - Since every lookup in the fd table must now check this flag and all of them do the same thing, move all the necessary checks into a function - fd_getfile.
2001-10-26 - split fdalloc into two functions. fdalloc now returns ENOSPC if itArtur Grabowski
would have to grow the table and fdexpand that grows the table. Since fdexpand can sleep we have to restart all operations if we have to call it. - dup2 changed so that finishdup frees the destination descriptor. - unp_externalize is a completly rewritten version from NetBSD. Changes mostly inspired by NetBSD.
2001-10-14Disable fd sharing for this release. Sorry.Artur Grabowski
deraadt@ ok.
2001-10-11Fix flawed logic when deciding if we should sleep whenArtur Grabowski
we are below the low watermark or if we should try to use up all buffers.
2001-10-11Count pages not buffers. This fixes deadlock condition which mainlyGrigoriy Orlov
occurs on the fs with large block size. We can have situation where numcleanbufs < locleanbufs and numdirtybufs < hidirtybufs. So, buffer flushing daemon never wakeups and other processes asleep forever waiting for a clean buffers. We count pages only for the dirty buffers which are on freelist(BQ_DIRTY). niklas@ found this. Rename flasher to cleaner. Suggested by costa@. Discussed with niklas@, costa@, millert@, art@. Ok deraadt@.
2001-10-07Clean up fdrelease.Artur Grabowski
2001-10-07gc UF_MAPPED.Artur Grabowski
2001-10-07Setting the VTIME to a value larger than 214 lead to an integer overflowArtur Grabowski
in the calculation of the sleep timeout which could lead to a panic. Instead of doing the very error-prone, and in more than one way incorrect, sleep time calculation, allocate an external timeout and let it wake us. Reported by: Raan Young <raan@graand-visions.com>
2001-10-02Bounds check index into routing table. Thanks to Ken Ashcraft of StanfordConstantine Sapuntzakis
for finding this bug.
2001-10-01Avoid panic when unconfiguring raid device which was autoconfigured andGrigoriy Orlov
so has fdp->fd_cdir = NULL. From Tertou <TDeval@PrimeOBJ.COM>
2001-09-29Avoid panic at time of raid autoconfiguration. Raid kernel thread startsGrigoriy Orlov
when root filesystem isn't configured yet and some kernel variables can't be initialized. From Thierry Deval <TDeval@PrimeOBJ.COM>
2001-09-29Move signal initialization before dostartuphooks to avoid panic atGrigoriy Orlov
start of raid autoconfiguration thread. Required for upcoming RAIDFrame code update. From Thierry Deval <TDeval@PrimeOBJ.COM>
2001-09-28Random cleanup (I don't like Ms in my tree).Artur Grabowski
2001-09-28QBITS is always defined, no point in having all those cofusing ifdefs.Artur Grabowski
2001-09-28Redundant prototypes.Artur Grabowski
2001-09-28sysctl() support for getting the SYSV *info structs and the associatedTodd C. Miller
SYSV structs. To be used by ipcs(1). Based on work by simonb@netbsd.org
2001-09-26At sonewconn() time, copy so_siguid & so_sigeuid to the newly created socket.Theo de Raadt
Stops a nasty little program supplied by gustavo@core-sdi.com
2001-09-21execsigs() could modify the wrong sigacts; based on David.Arana@eng.sun.com ↵Theo de Raadt
saying he could not understand how this worked
2001-09-20cosmetic.Artur Grabowski
2001-09-20occured->occurredMike Pechkin
idea from deraadt@ via NetBSD millert@ ok
2001-09-20Get rid of B_VFLUSH. Tested by costa and me.Grigoriy Orlov
2001-09-19Get rid of B_VFLUSH. Not relevant after the end of the AGE queue.Constantine Sapuntzakis
2001-09-19merge vm/vm_kern.h into uvm/uvm_extern.h; art@ okMichael Shalayeff
2001-09-19No need for this complicated (and bug-prone) method for waking up the flusher.Artur Grabowski
2001-09-19Unify elf32 and elf64 code with macros. This forced a renamingArtur Grabowski
of a few functions from elf32.
2001-09-17Don't treated return length of recvfrom() as a size_t... it's a socklen_t.Jason Wright
2001-09-17 The first implementation of the buffer flushing daemon. It solves ourGrigoriy Orlov
problem when syncer can't do its work because of vnode locks (PR1983). This also solves our problem where bigger number of buffers results in a much worse perfomance. In my configuration (i386, 128mb, BUFCACHEPERCENT=35) this speedup tar -xzf ports.tar.gz in 2-4 times. In configuration with low number of buffers and softupdates this may slowdown some operations up to 15%. The major difference with current buffer cache is that new implementation uses separate queues for dirty and clean buffers. I.e. BQ_LRU and BQ_AGE replaced by BQ_CLEAN and BQ_DIRTY. This simplifies things a lot and doesn't affect perfomance in a bad manner. Thanks to art and costa for pointing on errors. Tested by brad, millert, naddy, art, jj, camield art, millert ok.
2001-09-17Allocate the profiling buffer from kernel_map, not with malloc.Artur Grabowski
generated by cgit v1.2.3 (git 2.25.1) at 2025-01-01 13:29:47 +0000