Commit Graph

83812 Commits

Author SHA1 Message Date
Kirk McKusick
0f5f789c0d The buffer daemon cannot skip over buffers owned by locked inodes as
they may be the only viable ones to flush. Thus it will now wait for
an inode lock if the other alternatives will result in rollbacks (and
immediate redirtying of the buffer). If only buffers with rollbacks
are available, one will be flushed, but then the buffer daemon will
wait briefly before proceeding. Failing to wait briefly effectively
deadlocks a uniprocessor since every other process writing to that
filesystem will wait for the buffer daemon to clean up which takes
close enough to forever to feel like a deadlock.

Reported by:	Archie Cobbs <archie@dellroad.org>
Sponsored by:   DARPA & NAI Labs.
Approved by:	re
2002-12-14 01:35:30 +00:00
Tim J. Robbins
2e3684f843 When job control is disabled, never show the job id when reporting the
status of a background process that has terminated because of a signal.
2002-12-14 01:08:04 +00:00
Archie Cobbs
f3059f3906 Fix two bugs:
(a) Save control message return address only if NGM_MPPC_CONFIG_DECOMP
    (b) Properly count the number of required re-key operations
	when we loose synchronization and have to resync

MFC after:	3 days
2002-12-14 00:56:07 +00:00
Bill Fenner
3dc331b9f4 Add missing "of". 2002-12-14 00:20:43 +00:00
Tim J. Robbins
56c1b4ab77 libmchain no longer exports m_fixhdr(); remove it from EXPORT_SYMS. 2002-12-14 00:01:51 +00:00
Tim J. Robbins
578dcf0c31 Keep trying to flush the vnode list for the mount while some are still
busy and we are making progress towards making them not busy. This is
needed because smbfs vnodes reference their parent directory but may
appear after their parent in the mount's vnode list; one pass over the
list is not sufficient in this case.

This stops attempts to unmount idle smbfs mounts failing with EBUSY.
2002-12-13 23:44:00 +00:00
Warner Losh
96ac3c4b16 o Don't consider LOOPBACK devices as configured... 2002-12-13 23:36:31 +00:00
Scott Long
1f380ae506 Move the amd(4) driver to it's own directory in preparation for it growing
an sbus front-end.
2002-12-13 22:59:18 +00:00
Alfred Perlstein
bc9e75d7ca Backout removal SCARGS, the code freeze is only "selectively" over. 2002-12-13 22:41:47 +00:00
Alfred Perlstein
0bbe7292e1 Remove SCARGS.
Reviewed by: md5
2002-12-13 22:27:25 +00:00
Bill Fenner
83d000df26 Fix misspelling.
PR:		docs/44930
Submitted by:	Michael McGoldrick <mmcgoldrick@linuxdriven.net>
2002-12-13 22:26:10 +00:00
Bill Fenner
43ac5a2340 Add an implementation of the POSIX.1 sockatmark(3). 2002-12-13 22:22:55 +00:00
Poul-Henning Kamp
cc0163a3c4 Add a couple of KASSERTS, just in case. 2002-12-13 22:04:45 +00:00
Maxime Henrion
84c64d8895 Correct a harmless problem when creating the DMA tag
used to map mbufs.  The maximum size should be MCLBYTES
and not SIS_TX_LIST_SZ.  This is probably a typo in
the original commit.

Tested by:	cognet
2002-12-13 21:34:35 +00:00
Poul-Henning Kamp
821a4d01ea Don't interpret the hotspots relative to all slices on a slicer, but
relative to the parent device.
2002-12-13 21:31:13 +00:00
Poul-Henning Kamp
ef7b370b00 Fix spelling in comment. 2002-12-13 21:10:59 +00:00
Warner Losh
9cf092a263 MFp4: Make ifs consistant with rest of file. 2002-12-13 19:47:07 +00:00
Eric Moore
c664b80dbc Problem: Hang while bootup on Perc2/SC.
Fix: Error handling wasn't implemented
in newly implemented polling code for quartz
controllers.

Approved by: re@
2002-12-13 17:27:44 +00:00
Ruslan Ermilov
c8d40b7d34 mdoc(7) police: sort xrefs in SEE ALSO. 2002-12-13 16:53:51 +00:00
Tom Rhodes
97d0a1b862 Remove the incorrect information about dump(8) and snapshots. All the
dump(8) issues are assumed to be fixed.

PR:		46224
Submitted By:	Chris Knight <chris@aims.com.au>
Approved By:	re (murray)
2002-12-13 16:35:23 +00:00
Ruslan Ermilov
480565d5e4 mdoc(7) police: markup overhaul. 2002-12-13 15:27:27 +00:00
Ruslan Ermilov
f393f49865 mdoc(7) police: markup overhaul. 2002-12-13 10:38:42 +00:00
Tim J. Robbins
e3e797439e Fix build with SMB_VNODE_DEBUG defined; use td_proc->p_pid instead of
the nonexistent td_pid.
2002-12-13 10:25:08 +00:00
Tim J. Robbins
11de0c59e9 Store a reference to the parent directory's vnode in struct smbnode,
not to the parent's smbnode, which may be freed during the lifetime
of the child if the mount is forcibly unmounted. umount -f should now
work properly (ie. not panic) on smbfs mounts.
2002-12-13 10:15:01 +00:00
Ruslan Ermilov
c40d6a177d mdoc(7) police: markup overhaul. 2002-12-13 10:04:12 +00:00
Tim J. Robbins
9d0fffd3ca Drop filedesc lock and acquire Giant around calls to malloc() and free().
These call uma_large_malloc() and uma_large_free() which require Giant.
Fixes panic when descriptor table is larger than KMEM_ZMAX bytes
noticed by kkenn.

Reviewed by:	jhb
2002-12-13 09:59:40 +00:00
Bill Fenner
ed334c502f Renumber divert.
Forgotten by:	fenner
Approved by:	re (murray)
2002-12-13 09:18:09 +00:00
Ruslan Ermilov
f17c7afd96 mdoc(7) police: SEE ALSO references aren't real sentences. 2002-12-13 09:06:44 +00:00
Ruslan Ermilov
6971333705 mdoc(7) police: sort xrefs in SEE ALSO. 2002-12-13 09:05:50 +00:00
Murray Stokely
e46b1dd3c0 HEAD is 5.0-CURRENT again. 2002-12-13 08:48:34 +00:00
Murray Stokely
865fa6a43a HEAD reverts to 5.0-CURRENT
RELENG_5_0 is 5.0-RC
5.0-RELEASE will be built off of the RELENG_5_0 branch
2002-12-13 08:47:15 +00:00
Murray Stokely
8e414f48c1 Move the creation of the DRIVERS.TXT file to a more appropriate area
of this file, so it will be created for other platforms that use a
third floppy disk, not just i386.

Approved by:	re
2002-12-13 06:54:33 +00:00
David E. O'Brien
9dc9a9b380 Update for the just-done-import. 2002-12-13 05:03:15 +00:00
David E. O'Brien
cf709e0a4f This commit was generated by cvs2svn to compensate for changes in r107806,
which included commits to RCS files with non-trunk default branches.
2002-12-13 04:59:48 +00:00
David E. O'Brien
813da98d0e Vendor import of bwk's 29-Nov-2002 release.
Most significant update is the inclusion of our port's locale patches.
2002-12-13 04:59:48 +00:00
Jake Burkholder
f2de139378 Don't run kldxref when installing modules on sparc64. kldxref and the whole
module dependency system rely on linker behaviour that is machine dependent
and not part of the elf spec, and only work by accident on other platforms.

Approved by:	re
2002-12-13 00:32:29 +00:00
Greg Lehey
0538c676cc Create raw plexes and subdisks, not the old volume-related ones. This
crept in in the transition to devfs, and caused a number of "device
not configured" errors.

Approved by:	re (rwatson)
2002-12-13 00:27:32 +00:00
Murray Stokely
67cf4c38dc Add more information about the third floppy disk, drivers.flp.
Approved by:	re (jhb)
2002-12-12 23:11:43 +00:00
Chris Costello
4e9f9c6d30 Use the correct compartment notation in sample labels.
Sponsored by:	DARPA, Network Associates Labs
Approved by:	re (blanket)
2002-12-12 22:47:08 +00:00
Chris Costello
b4036ebd30 Change an "i.e." to an "e.g." This will be changed when a better
solution is worked out to clarify more specific compartment-related
cases.

Submitted by:	dcs
Sponsored by:	DARPA, Network Associates Labs
Approved by:	re (blanket)
2002-12-12 22:27:00 +00:00
Scott Long
a9f3d2c756 sable the CAM interface on some more cards. It seemed to be confusing and
upsetting the firmware there.

Thanks to imp@freebsd.org for suffering through testing with this.

Approved by:	re
2002-12-12 22:23:06 +00:00
Tom Rhodes
6883c2e5ae Fix the HISTORY to match reality. They were never MFC'ed to 4.X.
Submitted by:	R. Imura <imura@ryu16.org>
Approved by:	re (murray)
2002-12-12 22:22:51 +00:00
Chris Costello
12cfff2842 Add and fix cross-references.
Prompted by:	rwatson, dcs
Sponsored by:	DARPA, Network Associates Labs
2002-12-12 22:17:44 +00:00
Brian Feldman
77ae814e61 Account for fts(3)'s FTS_SLNONE case for symbolic links which have
a nonexistant target, in addition to the FTS_SL previously, so e.g.
setfmac -h sebsd/system_u:object_r:malloc_conf_t /etc/malloc.conf
succeeds.

Approved by:	re
2002-12-12 20:32:45 +00:00
Tom Rhodes
0509e374ac Update the pcm(4) manual page to explain devfs, fix a grammar error and
cross-ref devfs(5).

Approved by:	re (jhb)
2002-12-12 19:50:09 +00:00
Bernd Walter
765187cac7 The uftdi driver runs with multiple instances.
Remove a comment claiming the opposite.

Approved by: re (rwatson)
             gallatin (Mentor)
MFC after: 1 week
2002-12-12 19:03:43 +00:00
Yaroslav Tykhiy
3aa2a1d0e4 Clear the "device open" flag in daopen() before returning a error,
so the device won't stay marked as open whereas it isn't.

Approved by:	re, njl
MFC after:	1 week
2002-12-12 18:09:35 +00:00
Mark Murray
340a8a91c3 Fix a buffer overrun in /dev/random which, due to the nature of the
kernel memory allocator, is harmless. This could be a problem for
other systems, though. I've modified Darren's patch a little.

Original patch by: Darren Schack, Isilon Systems, Inc <darrens@isilon.com>
Also analysed by:  SGI, and in particular Divy Le Ray of SGI
OK'ed by:          re(rwatson)
2002-12-12 17:38:45 +00:00
Ruslan Ermilov
8d5d039f80 Uniformly refer to a file system as "file system".
Approved by:	re
2002-12-12 17:26:04 +00:00
Scott Long
e66cdcd7e0 Incorporate some suggestions from several others about talking about
device hints.

Approved by:	re
2002-12-12 15:40:11 +00:00