]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
22 years agoChange the type of `tscan' in vm_object_page_clean() to vm_pindex_t,
iedowse [Sat, 29 Jun 2002 20:04:38 +0000 (20:04 +0000)]
Change the type of `tscan' in vm_object_page_clean() to vm_pindex_t,
as it stores an absolute page index that may not fit in a vm_offset_t.

22 years agoBacked out previous commit. It had nothing to do with KSE and just undid
bde [Sat, 29 Jun 2002 19:53:55 +0000 (19:53 +0000)]
Backed out previous commit.  It had nothing to do with KSE and just undid
some of my style fixes.

22 years agoOops, that was a heck of a lot more debugging left onthan I intended..
julian [Sat, 29 Jun 2002 19:16:10 +0000 (19:16 +0000)]
Oops, that was a heck of a lot more debugging left onthan I intended..
turn it off!

I don't know if people think that these debugging macros are worth keeping
or not but I'll keep them for a short while, while the danger of
queue stuffups in the (rather complicated) run queue code exists.

22 years agoUpdated a comment. Namspace pollution in <sys/select.h> is now moot since
bde [Sat, 29 Jun 2002 19:14:52 +0000 (19:14 +0000)]
Updated a comment.  Namspace pollution in <sys/select.h> is now moot since
it was moved to <sys/selinfo.h>.

Fixed indentation of $FreeBSD$.

22 years agoWhen a netgraph message is read, look for another before returning.
brian [Sat, 29 Jun 2002 18:49:08 +0000 (18:49 +0000)]
When a netgraph message is read, look for another before returning.
This removes a bad latency problem during initial setup where we
end up waiting for too long before reading the connected message
and time the connection out.

Problem figured out by: Andre Albsmeier <andre@albsmeier.net>

22 years agoAdd KSE stubs to MD parts of ia64 code.
julian [Sat, 29 Jun 2002 18:26:18 +0000 (18:26 +0000)]
Add KSE stubs to MD parts of ia64 code.
Dfr will fill these out when we decide to enable KSEs on ia64
(probably not immediatly)

22 years agoReflect new gdb directory structure.
mp [Sat, 29 Jun 2002 18:19:53 +0000 (18:19 +0000)]
Reflect new gdb directory structure.

22 years agoAdd a copy of the sparc64 machine/kse.h to satisfy depencies..
julian [Sat, 29 Jun 2002 18:07:50 +0000 (18:07 +0000)]
Add  a copy of the sparc64 machine/kse.h to satisfy depencies..
dfr will fill in the correct contents at a later time.

22 years agoFix several style bugs:
alfred [Sat, 29 Jun 2002 17:58:44 +0000 (17:58 +0000)]
Fix several style bugs:

close up the continued line after removing the cast made the line.
space before parentheses in indirect function call.

Add an addtional error handler case for the results of callback.

Submitted by: bde

22 years agoUnbreak computation of 'smask' that I broke when removing caddr_t.
alfred [Sat, 29 Jun 2002 17:56:34 +0000 (17:56 +0000)]
Unbreak computation of 'smask' that I broke when removing caddr_t.

Submitted by: bde

22 years agoI don't see any reason to conditionalize this.
obrien [Sat, 29 Jun 2002 17:55:17 +0000 (17:55 +0000)]
I don't see any reason to conditionalize this.

22 years agoKSE test program
julian [Sat, 29 Jun 2002 17:39:07 +0000 (17:39 +0000)]
KSE test program

22 years agoPart 1 of KSE-III
julian [Sat, 29 Jun 2002 17:26:22 +0000 (17:26 +0000)]
Part 1 of KSE-III

The ability to schedule multiple threads per process
(one one cpu) by making ALL system calls optionally asynchronous.
to come: ia64 and power-pc patches, patches for gdb, test program (in tools)

Reviewed by: Almost everyone who counts
(at various times, peter, jhb, matt, alfred, mini, bernd,
and a cast of thousands)

NOTE: this is still Beta code, and contains lots of debugging stuff.
expect slight instability in signals..

22 years agoBmake bits for GDB 5.2.
obrien [Sat, 29 Jun 2002 16:52:15 +0000 (16:52 +0000)]
Bmake bits for GDB 5.2.

22 years agoSwitch over to 3.4p1.
des [Sat, 29 Jun 2002 12:16:50 +0000 (12:16 +0000)]
Switch over to 3.4p1.

22 years agoRegenerate.
des [Sat, 29 Jun 2002 11:58:32 +0000 (11:58 +0000)]
Regenerate.

22 years ago<sys/mman.h> requires <sys/types.h>.
des [Sat, 29 Jun 2002 11:57:51 +0000 (11:57 +0000)]
<sys/mman.h> requires <sys/types.h>.

22 years agoResolve conflicts.
des [Sat, 29 Jun 2002 11:48:59 +0000 (11:48 +0000)]
Resolve conflicts.

Sponsored by: DARPA, NAI Labs

22 years agoThis commit was generated by cvs2svn to compensate for changes in r99060,
des [Sat, 29 Jun 2002 11:34:13 +0000 (11:34 +0000)]
This commit was generated by cvs2svn to compensate for changes in r99060,
which included commits to RCS files with non-trunk default branches.

22 years agoVendor import of OpenSSH 3.4p1.
des [Sat, 29 Jun 2002 11:34:13 +0000 (11:34 +0000)]
Vendor import of OpenSSH 3.4p1.

22 years agoCommit config.h so we don't need autoconf to build world.
des [Sat, 29 Jun 2002 11:31:02 +0000 (11:31 +0000)]
Commit config.h so we don't need autoconf to build world.

22 years agocatch up with mextfree callback change when TI_PRIVATE_JUMBOS is defined.
alfred [Sat, 29 Jun 2002 11:26:05 +0000 (11:26 +0000)]
catch up with mextfree callback change when TI_PRIVATE_JUMBOS is defined.

Pointed out by: kdm

22 years agoOpenBSD lifted this code our tree. Preserve the original CVS id.
des [Sat, 29 Jun 2002 11:25:20 +0000 (11:25 +0000)]
OpenBSD lifted this code our tree.  Preserve the original CVS id.

22 years agoUse our __RCSID().
des [Sat, 29 Jun 2002 11:22:20 +0000 (11:22 +0000)]
Use our __RCSID().

22 years agoMake sure the environment variables set by setusercontext() are passed on
des [Sat, 29 Jun 2002 11:21:58 +0000 (11:21 +0000)]
Make sure the environment variables set by setusercontext() are passed on
to the child process.

Reviewed by: ache
Sponsored by: DARPA, NAI Labs

22 years agoCanonicize the host name before looking it up in the host file.
des [Sat, 29 Jun 2002 10:57:53 +0000 (10:57 +0000)]
Canonicize the host name before looking it up in the host file.

Sponsored by: DARPA, NAI Labs

22 years agoApply class-imposed login restrictions.
des [Sat, 29 Jun 2002 10:57:13 +0000 (10:57 +0000)]
Apply class-imposed login restrictions.

Sponsored by: DARPA, NAI Labs

22 years agoPAM support, the FreeBSD way.
des [Sat, 29 Jun 2002 10:56:23 +0000 (10:56 +0000)]
PAM support, the FreeBSD way.

Sponsored by: DARPA, NAI Labs

22 years agoDocument FreeBSD defaults.
des [Sat, 29 Jun 2002 10:55:18 +0000 (10:55 +0000)]
Document FreeBSD defaults.

Sponsored by: DARPA, NAI Labs

22 years agoDocument FreeBSD defaults and paths.
des [Sat, 29 Jun 2002 10:53:57 +0000 (10:53 +0000)]
Document FreeBSD defaults and paths.

Sponsored by: DARPA, NAI Labs

22 years agoRemove duplicate.
des [Sat, 29 Jun 2002 10:52:42 +0000 (10:52 +0000)]
Remove duplicate.

22 years agoApply FreeBSD's configuration defaults.
des [Sat, 29 Jun 2002 10:51:56 +0000 (10:51 +0000)]
Apply FreeBSD's configuration defaults.

Sponsored by: DARPA, NAI Labs

22 years agoAdd the VersionAddendum configuration variable.
des [Sat, 29 Jun 2002 10:49:57 +0000 (10:49 +0000)]
Add the VersionAddendum configuration variable.

Sponsored by: DARPA, NAI Labs

22 years agoSupport OPIE as an alternative to S/Key.
des [Sat, 29 Jun 2002 10:44:37 +0000 (10:44 +0000)]
Support OPIE as an alternative to S/Key.

Sponsored by: DARPA, NAI Labs

22 years agoDocument the upgrade process.
des [Sat, 29 Jun 2002 10:39:14 +0000 (10:39 +0000)]
Document the upgrade process.

22 years agoFiles we don't want to import.
des [Sat, 29 Jun 2002 10:39:02 +0000 (10:39 +0000)]
Files we don't want to import.

22 years agoAdd an inline to call eieio.
benno [Sat, 29 Jun 2002 10:00:07 +0000 (10:00 +0000)]
Add an inline to call eieio.

("Enforce In-order Execution of I/O".  I am not making this up.)

22 years agoWe don't need to clear RI in the MSR when entering a critical section.
benno [Sat, 29 Jun 2002 09:55:57 +0000 (09:55 +0000)]
We don't need to clear RI in the MSR when entering a critical section.

22 years agoAdd in_cksum.c
benno [Sat, 29 Jun 2002 09:50:20 +0000 (09:50 +0000)]
Add in_cksum.c

22 years agoin_cksum et al.
benno [Sat, 29 Jun 2002 09:49:26 +0000 (09:49 +0000)]
in_cksum et al.

Submitted by: Peter Grehan <peterg@ptree32.com.au>

22 years agoImplement vtophys()
benno [Sat, 29 Jun 2002 09:47:06 +0000 (09:47 +0000)]
Implement vtophys()

22 years agoAdd pmap_mapdev and pmap_unmapdev.
benno [Sat, 29 Jun 2002 09:45:59 +0000 (09:45 +0000)]
Add pmap_mapdev and pmap_unmapdev.

22 years ago- Initialise battable to cover I/O spaces.
benno [Sat, 29 Jun 2002 09:43:59 +0000 (09:43 +0000)]
- Initialise battable to cover I/O spaces.
- Statically size the bpvo entries to avoid conflicts between bpvo allocation
  and the vm allocator.
- Shift pmap_init2 code into pmap_init.
- Add UMA_ZONE_VM flag to uma_zcreate.

Submitted by: Peter Grehan <peterg@ptree32.com.au>

22 years agoTo quote Peter:
benno [Sat, 29 Jun 2002 09:39:44 +0000 (09:39 +0000)]
To quote Peter:

The case in cpu_switch() where there isn't a higher priority thread
(choosethread() == curthread) uses r4 as the PCB context pointer. However, the
use of r4 after the label L2 is incorrect, since it was probably trashed by
the call to choosethread, and in any case was set up to curthread at the start
of the routine.

This condition will occur when an interrupt thread schedules a netisr, which
is a lower priority thread.

Another (probably unnecessary) difference is that I was paranoid about
register trashing, so I decided to save r2 and r13 as well.

Submitted by: Peter Grehan <peterg@ptree32.com.au>

22 years agomempcy/bcopy handles overlapping copies so make ovbcopy call it.
benno [Sat, 29 Jun 2002 09:34:54 +0000 (09:34 +0000)]
mempcy/bcopy handles overlapping copies so make ovbcopy call it.

22 years agoAdd BOOTP_NFSROOT support code.
benno [Sat, 29 Jun 2002 09:33:18 +0000 (09:33 +0000)]
Add BOOTP_NFSROOT support code.

22 years ago- Use tmpstk exclusively in the init path.
benno [Sat, 29 Jun 2002 09:31:27 +0000 (09:31 +0000)]
- Use tmpstk exclusively in the init path.
- Remove redundant code.

Submitted by: Peter Grehan <peterg@ptree32.com.au>

22 years agoMany fixes to low-level trap and interrupt handling:
benno [Sat, 29 Jun 2002 09:28:21 +0000 (09:28 +0000)]
Many fixes to low-level trap and interrupt handling:

- Tidy up clock code.  Don't repeatedly call hardclock().
- Remove intrnames, decrnest and intrcnt from locore.s
- Coalesce all trap handling into a single stub that then calls a dispatch
  function.

Submitted by: Peter Grehan <peterg@ptree32.com.au>

22 years agoTo avoid anybody else getting caught out, these two files are really
peter [Sat, 29 Jun 2002 09:21:14 +0000 (09:21 +0000)]
To avoid anybody else getting caught out, these two files are really
in /sys/boot/common and are actually used there.

22 years agoConvert this from mostly inline assembler to mostly C.
benno [Sat, 29 Jun 2002 09:16:30 +0000 (09:16 +0000)]
Convert this from mostly inline assembler to mostly C.

Submitted by: Peter Grehan <peterg@ptree32.com.au>

22 years agoUpdate from NetBSD 1.3 -> 1.6. Most notable, rev 1.6:
peter [Sat, 29 Jun 2002 09:00:47 +0000 (09:00 +0000)]
Update from NetBSD 1.3 -> 1.6.  Most notable, rev 1.6:
"Make in_cksum work on little endian machines"

This would explain a few things. :-)

22 years agoDon't use the static thread.. it is going away.
julian [Sat, 29 Jun 2002 07:47:20 +0000 (07:47 +0000)]
Don't use the static thread.. it is going away.

22 years agoremove cruft
julian [Sat, 29 Jun 2002 07:16:24 +0000 (07:16 +0000)]
remove cruft

22 years agoAdd files that are new for KSE.
julian [Sat, 29 Jun 2002 07:04:59 +0000 (07:04 +0000)]
Add files that are new for KSE.

22 years agoFix a deletion during traversal tailq bug.
jake [Sat, 29 Jun 2002 06:37:24 +0000 (06:37 +0000)]
Fix a deletion during traversal tailq bug.

22 years agoAvoid truncating filenames with snprintf().
tjr [Sat, 29 Jun 2002 05:31:26 +0000 (05:31 +0000)]
Avoid truncating filenames with snprintf().

22 years agoAvoid truncating the pathname to UNIX Domain Sockets with snprintf(),
tjr [Sat, 29 Jun 2002 04:52:33 +0000 (04:52 +0000)]
Avoid truncating the pathname to UNIX Domain Sockets with snprintf(),
giving a more sensible warning when the (relatively meagre) sun_path
limit is exceeded.

22 years agoRename the db command lockedvnodes to lockedvnods so that it fits on the
obrien [Sat, 29 Jun 2002 04:45:09 +0000 (04:45 +0000)]
Rename the db command lockedvnodes to lockedvnods so that it fits on the
help screen and one doens't think we have a lockedvnodesmap command.

22 years agoDo NOT enable ATA_STATIC_ID -- cmd646 controller will be !ata2!,
obrien [Sat, 29 Jun 2002 04:42:03 +0000 (04:42 +0000)]
Do NOT enable ATA_STATIC_ID -- cmd646 controller will be !ata2!,
and you will not mount an ATA /:
mountroot> ufs:/ad0a
Mounting root from ufs:/ad0a
setrootbyname failed
ffs_mountroot: can't find rootvp

22 years agoCheck that snprintf() does not truncate the destination pathname when
tjr [Sat, 29 Jun 2002 04:10:40 +0000 (04:10 +0000)]
Check that snprintf() does not truncate the destination pathname when
linking into a directory.

22 years agoUse __FBSDID.
obrien [Sat, 29 Jun 2002 03:23:51 +0000 (03:23 +0000)]
Use __FBSDID.

22 years agoAdd frexp(3).
obrien [Sat, 29 Jun 2002 03:23:18 +0000 (03:23 +0000)]
Add frexp(3).

Obtained from: OpenBSD

22 years agoTurn back on GDB.
obrien [Sat, 29 Jun 2002 03:19:35 +0000 (03:19 +0000)]
Turn back on GDB.
gdbserver does not build, but I've gone about as far with GDB as I'm going
to right now.

22 years agoBest guess at configurations for our other arches. (totally untested)
obrien [Sat, 29 Jun 2002 03:18:28 +0000 (03:18 +0000)]
Best guess at configurations for our other arches.  (totally untested)

22 years agoAllow to compile and at least startup on sparc64.
obrien [Sat, 29 Jun 2002 03:16:40 +0000 (03:16 +0000)]
Allow to compile and at least startup on sparc64.

22 years agoRemove a couple of __P() stragglers.
peter [Sat, 29 Jun 2002 02:32:34 +0000 (02:32 +0000)]
Remove a couple of __P() stragglers.

22 years agomore caddr_t removal.
alfred [Sat, 29 Jun 2002 02:00:02 +0000 (02:00 +0000)]
more caddr_t removal.

22 years agoAdd another Intel chipset (i82562).
silby [Sat, 29 Jun 2002 01:56:14 +0000 (01:56 +0000)]
Add another Intel chipset (i82562).

PR: 39974
Submitted by: Morten Aaboe Jensen <morten@codemonkey.dk>
MFC after: 1 day

22 years agoMore caddr_t removal, make fo_ioctl take a void * instead of a caddr_t.
alfred [Sat, 29 Jun 2002 01:50:25 +0000 (01:50 +0000)]
More caddr_t removal, make fo_ioctl take a void * instead of a caddr_t.

22 years agocatch up with mextadd callback taking a void argument instead of a caddr_t.
alfred [Sat, 29 Jun 2002 01:49:22 +0000 (01:49 +0000)]
catch up with mextadd callback taking a void argument instead of a caddr_t.

22 years agoThis commit was generated by cvs2svn to compensate for changes in r99005,
obrien [Sat, 29 Jun 2002 01:43:09 +0000 (01:43 +0000)]
This commit was generated by cvs2svn to compensate for changes in r99005,
which included commits to RCS files with non-trunk default branches.

22 years agoVendor import of GDB 5.2.
obrien [Sat, 29 Jun 2002 01:43:09 +0000 (01:43 +0000)]
Vendor import of GDB 5.2.
These bits are taken from the FSF gdb_5_2-branch anoncvs repo
on 27-June-2002 12:01:00 EDT.

22 years agocatch up with ext_free prototype change.
alfred [Sat, 29 Jun 2002 01:36:59 +0000 (01:36 +0000)]
catch up with ext_free prototype change.

22 years agoBmake bits for GDB 5.2.
obrien [Sat, 29 Jun 2002 01:16:01 +0000 (01:16 +0000)]
Bmake bits for GDB 5.2.

22 years ago- kernel core debugging for i386 (Mark Peek)
obrien [Sat, 29 Jun 2002 00:42:52 +0000 (00:42 +0000)]
- kernel core debugging for i386 (Mark Peek)
- thread debugging for i386 (Mark Peek) and alpha (Doug Rabson)

Submitted by: mp,dfr

22 years agoNote that this is a modified version of GDB.
obrien [Sat, 29 Jun 2002 00:41:02 +0000 (00:41 +0000)]
Note that this is a modified version of GDB.

22 years agoUse the stock source.
obrien [Sat, 29 Jun 2002 00:38:44 +0000 (00:38 +0000)]
Use the stock source.

22 years agoBest that I can tell, we've never used these files.
obrien [Sat, 29 Jun 2002 00:36:19 +0000 (00:36 +0000)]
Best that I can tell, we've never used these files.

22 years agoMore caddr_t removal.
alfred [Sat, 29 Jun 2002 00:29:12 +0000 (00:29 +0000)]
More caddr_t removal.
Change struct knote's kn_hook from caddr_t to void *.

22 years agonuke more instances of caddr_t
alfred [Sat, 29 Jun 2002 00:02:01 +0000 (00:02 +0000)]
nuke more instances of caddr_t

22 years agom_extadd takes a void (*freef)(void *, void *) now, not a
alfred [Sat, 29 Jun 2002 00:01:46 +0000 (00:01 +0000)]
m_extadd takes a void (*freef)(void *, void *) now, not a
void (*freef)(caddr_t, void *).

22 years agoremove or replace caddr_t with void.
alfred [Fri, 28 Jun 2002 23:48:23 +0000 (23:48 +0000)]
remove or replace caddr_t with void.
make the mbuf external free function take a void * rather than caddr_t.

22 years agonuke caddr_t.
alfred [Fri, 28 Jun 2002 23:17:36 +0000 (23:17 +0000)]
nuke caddr_t.

22 years agochange struct socket -> so_pcb from caddr_t to void *.
alfred [Fri, 28 Jun 2002 23:17:08 +0000 (23:17 +0000)]
change struct socket -> so_pcb from caddr_t to void *.

22 years agoRemove unneeded casts to caddr_t.
alfred [Fri, 28 Jun 2002 23:02:38 +0000 (23:02 +0000)]
Remove unneeded casts to caddr_t.

22 years agochange f_data field in struct file from caddr_t to void *.
alfred [Fri, 28 Jun 2002 23:00:32 +0000 (23:00 +0000)]
change f_data field in struct file from caddr_t to void *.

22 years agoSimplify TERM handling since now libutil not overwrites existen TERM for "term"
ache [Fri, 28 Jun 2002 22:56:31 +0000 (22:56 +0000)]
Simplify TERM handling since now libutil not overwrites existen TERM for "term"

22 years agodocument that the pipe fo_stat routine doesn't need locks because it's
alfred [Fri, 28 Jun 2002 22:35:12 +0000 (22:35 +0000)]
document that the pipe fo_stat routine doesn't need locks because it's
a read operation.

Requested by: rwatson

22 years agoIn namei(), we use a NULL thread for uio_td when doing a VOP_READLINK().
jhb [Fri, 28 Jun 2002 21:53:08 +0000 (21:53 +0000)]
In namei(), we use a NULL thread for uio_td when doing a VOP_READLINK().
nfs_readlink() calls nfs_bioread() which passes in uio_td as the thread
argument to nfs_getcacheblk().  In nfs_getcacheblk() we dereference the
thread pointer to get a process pointer to pass to nfs_sigintr().  This
obviously results in a panic. :)

Rather than change nfs_getcacheblk() to check if the thread pointer is
NULL when calling nfs_sigintr() like other callers do, change
nfs_sigintr() to take a thread as the last argument instead of a
process so none of the callers have to care if the thread is NULL or not.

22 years agoAdd two new submodes to the AES encryption method.
phk [Fri, 28 Jun 2002 21:25:15 +0000 (21:25 +0000)]
Add two new submodes to the AES encryption method.

This method is now suitable for encrypting swap spaces.

Sponsored by: DARPA & NAI Labs.

22 years agoAdd a module for src/tools/regression (`regression').
jmallett [Fri, 28 Jun 2002 21:13:06 +0000 (21:13 +0000)]
Add a module for src/tools/regression (`regression').

Add a MAINTAINERS line for the regression module, specifically referring to
src/tools/regression/usr.bin, right now, but applicable to other things, to
make clear that I am willing to help write new tests.  The framework is all
modularised now, so it is easy to write new tests, etc., and since I'd like
to see tests for more and more things as bugs get fixed, it seems to be the
right thing to do to stand up and offer to help people write tests.

22 years agoImprove the VOP locking asserts
jeff [Fri, 28 Jun 2002 20:58:14 +0000 (20:58 +0000)]
Improve the VOP locking asserts

 - Add vfs_badlock_print to control whether or not we print lock violations
 - Add vfs_badlock_panic to control whether we panic on lock violations

Both default to on to mimic the original behavior if DEBUG_VFS_LOCKS is on.

22 years agoIn vn_mkdir(), use vrele() instead of vput() on the parent directory
iedowse [Fri, 28 Jun 2002 20:06:47 +0000 (20:06 +0000)]
In vn_mkdir(), use vrele() instead of vput() on the parent directory
vnode in the case that the target exists and is the same vnode as
the parent (i.e. "mkdir ."). The namei() call does not leave the
vnode locked in this case even though you might expect it to.

This bug was mostly harmless in practice because unlocking an already
unlocked vnode currently does not trigger any panics or warnings.

Reviewed by: jeff

22 years agoOne possible code path for syncache_respond() is:
jlemon [Fri, 28 Jun 2002 19:12:38 +0000 (19:12 +0000)]
One possible code path for syncache_respond() is:

 syncache_respond(A), ip_output(), ip_input(), tcp_input(), syncache_badack(B)

Which winds up deleting a different entry from the syncache.  Handle
this by not utilizing the next entry in the timer chain until after
syncache_respond() completes.  The case of A == B should not be possible.

Problem found by: Don Bowman <don@sandvine.com>

22 years agoClean up vn_rdwr locking.
jeff [Fri, 28 Jun 2002 17:51:11 +0000 (17:51 +0000)]
Clean up vn_rdwr locking.

 - Do shared locks on read.
 - Only do vn_{start,finished}_write when writing.

22 years agoFix a case where a vnode got explicitly unlocked after the pointer to it
green [Fri, 28 Jun 2002 16:17:47 +0000 (16:17 +0000)]
Fix a case where a vnode got explicitly unlocked after the pointer to it
got set to NULL.

Revision 1.355: in the box

22 years agoInclude 'sshd' to the lists of forbidden users.
maxim [Fri, 28 Jun 2002 15:46:29 +0000 (15:46 +0000)]
Include 'sshd' to the lists of forbidden users.

Reviewed by: cvs-committers

22 years agoAdd additional field 'overwrite' to login_vars. It mainly needed to handle
ache [Fri, 28 Jun 2002 14:45:30 +0000 (14:45 +0000)]
Add additional field 'overwrite' to login_vars. It mainly needed to handle
"term" according to manpage, i.e. not overwrite it, if already present in
environment.

22 years agoMake sigpending and sigsuspend account for signals that are pending on
deischen [Fri, 28 Jun 2002 13:28:41 +0000 (13:28 +0000)]
Make sigpending and sigsuspend account for signals that are pending on
the process as well as pending on the current thread.

Reported by: Andrew MacIntyre <andymac@bullseye.apana.org.au>

22 years agoAdd a wrapper for pselect() in order to make it a cancellation point.
deischen [Fri, 28 Jun 2002 13:26:02 +0000 (13:26 +0000)]
Add a wrapper for pselect() in order to make it a cancellation point.

Prompted by: wollman