]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
22 years ago- Make the sio_inited handling more properly MP safe.
John Baldwin [Sat, 20 Oct 2001 03:32:33 +0000 (03:32 +0000)]
- Make the sio_inited handling more properly MP safe.
- If we are booting with a serial console, then pass the MTX_QUIET flag to
  mtx_init of the sio lock as otherwise KTR_VERBOSE simply doesn't work.

22 years agoAssert that a ucred is unshared before we remap its ids.
John Baldwin [Sat, 20 Oct 2001 03:30:34 +0000 (03:30 +0000)]
Assert that a ucred is unshared before we remap its ids.

22 years agoMake the flag field in the ps output one char wider to account for recent
John Baldwin [Sat, 20 Oct 2001 03:22:23 +0000 (03:22 +0000)]
Make the flag field in the ps output one char wider to account for recent
growth in the number of flags used.  Also, if a thread is blocked on a
mutex, print the mutex name in the wait channel column.

22 years agoapparently EV5 2x00s have their t2 CSRs where I'd expect to find
Andrew Gallatin [Sat, 20 Oct 2001 01:27:23 +0000 (01:27 +0000)]
apparently EV5 2x00s have their t2 CSRs where I'd expect to find
them on a  2100A

Thanks to Tyler Willingham <ze_willow@yahoo.com> for letting me
test this on his AS2000

22 years agoThe mtx_init() and sx_init() functions bzero'd locks before handing them
John Baldwin [Sat, 20 Oct 2001 01:22:42 +0000 (01:22 +0000)]
The mtx_init() and sx_init() functions bzero'd locks before handing them
off to witness_init() making the check for double intializating a lock by
testing the LO_INITIALIZED flag moot.  Workaround this by checking the
LO_INITIALIZED flag ourself before we bzero the lock structure.

22 years agoDrop support for x87 emulation. Any CPU one would dare to run 5-CURRENT
David E. O'Brien [Sat, 20 Oct 2001 01:15:54 +0000 (01:15 +0000)]
Drop support for x87 emulation.  Any CPU one would dare to run 5-CURRENT
on would have built-in FP support.

22 years agoTweak the way we determine if an interface needs to have its name translated.
Dag-Erling Smørgrav [Sat, 20 Oct 2001 00:01:26 +0000 (00:01 +0000)]
Tweak the way we determine if an interface needs to have its name translated.
Add some missing break statements in the socket ioctl switch.
Check the return value from copyin() / copyout().
Fix some disorderings and misindentations.
Support a couple more socket ioctls.
Add missing break statements.

22 years agoAdd a sysctl for preventing the sync() in panic() recovery. This can
Peter Wemm [Fri, 19 Oct 2001 23:32:03 +0000 (23:32 +0000)]
Add a sysctl for preventing the sync() in panic() recovery.  This can
be so dangerous it isn't funny.  eg: if you panic inside NFS or softdep,
and then try and sync you run into held locks and cause either deadlocks,
recursive panics or other interesting chaos.  Default is unchanged.

22 years agoFix includes based on recent changes to lock.h, mutex.h and ktr.h.
Mark Peek [Fri, 19 Oct 2001 22:45:46 +0000 (22:45 +0000)]
Fix includes based on recent changes to lock.h, mutex.h and ktr.h.

22 years agoAdd unaligned.c and delete commented out entry for userconfig.c
Doug Rabson [Fri, 19 Oct 2001 22:25:00 +0000 (22:25 +0000)]
Add unaligned.c and delete commented out entry for userconfig.c

22 years agoMake a start at an unaligned trap handler. Only integer loads and stores
Doug Rabson [Fri, 19 Oct 2001 22:23:51 +0000 (22:23 +0000)]
Make a start at an unaligned trap handler. Only integer loads and stores
are handled so far.

22 years agoMake override of nm consistent across all platforms and sort the overrides.
Mark Peek [Fri, 19 Oct 2001 22:19:06 +0000 (22:19 +0000)]
Make override of nm consistent across all platforms and sort the overrides.

Reminded by: obrien

22 years agoMake a start at a regression test for the unaligned trap handler.
Doug Rabson [Fri, 19 Oct 2001 22:10:13 +0000 (22:10 +0000)]
Make a start at a regression test for the unaligned trap handler.

22 years agoBring a few multi-architecture arch= attributes into the hardware
Bruce A. Mah [Fri, 19 Oct 2001 20:56:31 +0000 (20:56 +0000)]
Bring a few multi-architecture arch= attributes into the hardware
list, based primarily on dmesg logs posted to various mailing lists.

22 years agoTranslate various userland traps into SIGBUS (instead of just panicing).
Doug Rabson [Fri, 19 Oct 2001 20:46:12 +0000 (20:46 +0000)]
Translate various userland traps into SIGBUS (instead of just panicing).

22 years agoSupport a comma-separated list of architectures for the arch=
Bruce A. Mah [Fri, 19 Oct 2001 20:41:09 +0000 (20:41 +0000)]
Support a comma-separated list of architectures for the arch=
attribute.  This change is necessary in order for the release
documentation to scale to support more than two architectures.

Based on work by: hrs

22 years agoMake the logic more explicit.
David E. O'Brien [Fri, 19 Oct 2001 20:20:16 +0000 (20:20 +0000)]
Make the logic more explicit.

22 years agoFix the typedef of va_list.
Ruslan Ermilov [Fri, 19 Oct 2001 20:07:46 +0000 (20:07 +0000)]
Fix the typedef of va_list.

22 years agoNew release note: xl(4) checksum offloading.
Bruce A. Mah [Fri, 19 Oct 2001 20:07:07 +0000 (20:07 +0000)]
New release note:  xl(4) checksum offloading.

Reviewed by: alc

22 years agoTry two on the preprocessing logic.
Ruslan Ermilov [Fri, 19 Oct 2001 20:06:23 +0000 (20:06 +0000)]
Try two on the preprocessing logic.

Reviewed by: obrien

22 years agoNew release note: route(8) add -ifp/-ifa.
Bruce A. Mah [Fri, 19 Oct 2001 20:02:22 +0000 (20:02 +0000)]
New release note:  route(8) add -ifp/-ifa.

Submitted by: ru

22 years agoTry two on the preprocessing logic.
David E. O'Brien [Fri, 19 Oct 2001 20:01:01 +0000 (20:01 +0000)]
Try two on the preprocessing logic.

Reviewed by: ru

22 years ago- Move the definition of LOCK_DEBUG back to sys/lock.h from sys/_lock.h.
John Baldwin [Fri, 19 Oct 2001 19:28:18 +0000 (19:28 +0000)]
- Move the definition of LOCK_DEBUG back to sys/lock.h from sys/_lock.h.
- Change LOCK_DEBUG so that it is always on if KTR is compiled in
  regardless of the state of KTR_COMPILE.  This means that we no longer
  need to include sys/ktr.h before sys/lock.h to ensure a valid setting
  for LOCK_DEBUG.
- Change the use of LOCK_DEBUG so that it is now always defined and its
  value is used instead of merely its definition.  That is, instead of
  #ifdef LOCK_DEBUG, code should now use #if LOCK_DEBUG > 0.
- Use this latest to #error out in sys/mutex.h if sys/lock.h isn't
  included before sys/mutex.h to ensure that the proper versions of the
  mutex operations are used.
- As a result of (2) sys/mutex.h no longer includes sys/ktr.h in the
  KERNEL case.

Requested by: bde (1)

22 years agoRemove unneeded sys/mutex.h includes.
John Baldwin [Fri, 19 Oct 2001 19:23:32 +0000 (19:23 +0000)]
Remove unneeded sys/mutex.h includes.

22 years agoMove sys/mutex.h include after sys/lock.h to ensure we use the proper
John Baldwin [Fri, 19 Oct 2001 19:22:04 +0000 (19:22 +0000)]
Move sys/mutex.h include after sys/lock.h to ensure we use the proper
versions of the mutex operations.

22 years agoBlah, fix braino where ru had to remind me of proper preprocessor syntax.
David E. O'Brien [Fri, 19 Oct 2001 19:17:11 +0000 (19:17 +0000)]
Blah, fix braino where ru had to remind me of proper preprocessor syntax.
Bad fingers, no cookie.

22 years agoProperly handle backslash newline within an identifier or keyword.
Jens Schweikhardt [Fri, 19 Oct 2001 19:10:36 +0000 (19:10 +0000)]
Properly handle backslash newline within an identifier or keyword.

PR: bin/6015
Submitted by: myself (schweikh)
Patch by: Alexey V.Neyman <alex.neyman@auriga.ru>
Tested by: indenting my chess problem solver and running its test suite
MFC after: 3 weeks

22 years agoFix this so it compiles cleanly for alpha. Tried to do some minimal testing.
Matt Jacob [Fri, 19 Oct 2001 18:29:57 +0000 (18:29 +0000)]
Fix this so it compiles cleanly for alpha. Tried to do some minimal testing.

Reviewed by: freebsd-net

22 years agoArgh! I updated the version number in the MODULE_DEPEND() thingamagook but
Dag-Erling Smørgrav [Fri, 19 Oct 2001 18:23:51 +0000 (18:23 +0000)]
Argh!  I updated the version number in the MODULE_DEPEND() thingamagook but
not in the actual MODULE_VERSION().  Pass me the pointy hat.

22 years agoNew release notes: libstand/loader(8) bzip2 support, isdnphone(8) -k,
Bruce A. Mah [Fri, 19 Oct 2001 16:27:05 +0000 (16:27 +0000)]
New release notes:  libstand/loader(8) bzip2 support, isdnphone(8) -k,
arp(8) fddi/atm output.

MFC noted:  arp(8) interface name in output.

22 years agoCleanup of the stdarg code.
Mark Peek [Fri, 19 Oct 2001 16:15:46 +0000 (16:15 +0000)]
Cleanup of the stdarg code.

Submitted by: ru

22 years agoInclude the right i4b files for NetBSD
Brian Somers [Fri, 19 Oct 2001 15:56:56 +0000 (15:56 +0000)]
Include the right i4b files for NetBSD

22 years agoFix the le driver - it has been broken for a while and these fixes have
Jonathan Lemon [Fri, 19 Oct 2001 15:47:41 +0000 (15:47 +0000)]
Fix the le driver - it has been broken for a while and these fixes have
been sitting neglected in my tree.

22 years agoNew release and hardware note: gx(4).
Bruce A. Mah [Fri, 19 Oct 2001 15:01:54 +0000 (15:01 +0000)]
New release and hardware note:  gx(4).

While I'm here, fix a minor markup glitch in bge(4) release note.

22 years agomdoc(7) police: join OS version with the corresponding macro.
Ruslan Ermilov [Fri, 19 Oct 2001 14:44:13 +0000 (14:44 +0000)]
mdoc(7) police: join OS version with the corresponding macro.

22 years agoMake an ATAPI CD device appear as /dev/acdN and the individual
Søren Schmidt [Fri, 19 Oct 2001 13:34:48 +0000 (13:34 +0000)]
Make an ATAPI CD device appear as /dev/acdN and the individual
tracks as /dev/acdNtY.

This solves the problems with having to deal with howmany and which
devices are open.

For hysterical reasons ONLY, make a link to both the acdNa & acdNc devices.

22 years agoJust use RSYSCALL.
Ruslan Ermilov [Fri, 19 Oct 2001 13:32:24 +0000 (13:32 +0000)]
Just use RSYSCALL.

22 years agosignanosleep(2) hasn't existed since 1998.
Ruslan Ermilov [Fri, 19 Oct 2001 13:01:57 +0000 (13:01 +0000)]
signanosleep(2) hasn't existed since 1998.

22 years agoRespect system ${CFLAGS} when building Perl modules.
Ruslan Ermilov [Fri, 19 Oct 2001 12:58:22 +0000 (12:58 +0000)]
Respect system ${CFLAGS} when building Perl modules.

bsd.obj.mk -> bsd.prog.mk in modules makefiles, as the
latter automatically includes ../Makefile.inc and adds
-I${DESTDIR}/usr/include to ${CFLAGS} needed for "make
world" which is built with -nostdinc.

Reviewed by: MAINTAINER timeout

22 years agoFixed style bugs in previous commit:
Bruce Evans [Fri, 19 Oct 2001 12:45:59 +0000 (12:45 +0000)]
Fixed style bugs in previous commit:
- don't comment out the old version; just modify it.
- don't 2 sets of 3 style bugs that weren't present in the old version.

22 years agoMFi386: sys/i386/isa/npx.c revisions from 1.114 to 1.116
Yoshihiro Takahashi [Fri, 19 Oct 2001 12:30:16 +0000 (12:30 +0000)]
MFi386: sys/i386/isa/npx.c revisions from 1.114 to 1.116

22 years agoFixed most style bugs in previous commit.
Bruce Evans [Fri, 19 Oct 2001 12:27:26 +0000 (12:27 +0000)]
Fixed most style bugs in previous commit.

22 years agoFixed style bugs from revisions 8.1 (CSRG), 1.14, 1.19 and 1.22.
Ruslan Ermilov [Fri, 19 Oct 2001 12:24:54 +0000 (12:24 +0000)]
Fixed style bugs from revisions 8.1 (CSRG), 1.14, 1.19 and 1.22.

22 years agoChanged IO_NPXSIZE to 8
Yoshihiro Takahashi [Fri, 19 Oct 2001 12:24:45 +0000 (12:24 +0000)]
Changed IO_NPXSIZE to 8

22 years agoFixed missing DPADD and misplaced $FreeBSD$ in previous commit.
Bruce Evans [Fri, 19 Oct 2001 12:15:56 +0000 (12:15 +0000)]
Fixed missing DPADD and misplaced $FreeBSD$ in previous commit.

22 years agoMFi386: sys/i386/isa/clock.c revision 1.177
Yoshihiro Takahashi [Fri, 19 Oct 2001 11:52:49 +0000 (11:52 +0000)]
MFi386: sys/i386/isa/clock.c revision 1.177

22 years agoMFi386: sys/i386/isa/pcaudio.c revision 1.63
Yoshihiro Takahashi [Fri, 19 Oct 2001 11:50:00 +0000 (11:50 +0000)]
MFi386: sys/i386/isa/pcaudio.c revision 1.63

22 years agoMerged from sys/isa/sio.c revisions from 1.344 to 1.347
Yoshihiro Takahashi [Fri, 19 Oct 2001 11:48:47 +0000 (11:48 +0000)]
Merged from sys/isa/sio.c revisions from 1.344 to 1.347

22 years agoMFi386: sys/i386/conf/GENERIC revision 1.320
Yoshihiro Takahashi [Fri, 19 Oct 2001 11:46:13 +0000 (11:46 +0000)]
MFi386: sys/i386/conf/GENERIC revision 1.320

22 years agoMFi386: options.i386 revision 1.158
Yoshihiro Takahashi [Fri, 19 Oct 2001 11:45:05 +0000 (11:45 +0000)]
MFi386: options.i386 revision 1.158

22 years agoChange rmskey() to rmopie() - we don't have skey anymore
Andrey A. Chernov [Fri, 19 Oct 2001 10:45:14 +0000 (10:45 +0000)]
Change rmskey() to rmopie() - we don't have skey anymore

22 years agoAdd NO_WERROR so the build won't die because of discarded qualifiers
Dag-Erling Smørgrav [Fri, 19 Oct 2001 10:08:38 +0000 (10:08 +0000)]
Add NO_WERROR so the build won't die because of discarded qualifiers

22 years agoBack out part of previous commit which was gcc-centric
Dag-Erling Smørgrav [Fri, 19 Oct 2001 10:08:05 +0000 (10:08 +0000)]
Back out part of previous commit which was gcc-centric

22 years agoRework pmap so that it separates the PTE structure from the pv_entry
Doug Rabson [Fri, 19 Oct 2001 09:47:02 +0000 (09:47 +0000)]
Rework pmap so that it separates the PTE structure from the pv_entry
structure. This makes it possible to pre-allocate PTEs for the kernel,
which is necessary for a reliable implementation of pmap_kenter(). This
also avoids wasting space (about 48 bytes per page) for kernel mappings
and user mappings of memory-mapped devices.

This also fixes a bug with the previous version where the implementation
required the pv_entry structure to be physically contiguous but did not
enforce this (the structure size was not a power of two). This meant
that the pv_entry free list was quickly corrupted as soon as the system
was even mildly loaded.

22 years agoBuild the linux module on Alpha again. Note that the linux module
Marcel Moolenaar [Fri, 19 Oct 2001 08:33:56 +0000 (08:33 +0000)]
Build the linux module on Alpha again. Note that the linux module
is not added to the MI list of modules. The current state of
affairs is such that the linux module is the exception, not the
common case.

22 years agoAfter removing the bogus differences, any real incompatibilities
Marcel Moolenaar [Fri, 19 Oct 2001 08:20:15 +0000 (08:20 +0000)]
After removing the bogus differences, any real incompatibilities
have been merged into /sys/compat/linux/linux_ioctl.h. This file
is now unused.

22 years agoFix Alpha related brokenness. We used to have a MD linux_ioctl.h
Marcel Moolenaar [Fri, 19 Oct 2001 08:18:31 +0000 (08:18 +0000)]
Fix Alpha related brokenness. We used to have a MD linux_ioctl.h
that appeared to be very different from the MI version. These
differences were mostly bogus and caused by copying octal
definitions and write them as hexadecimal values without doing
any base conversion (ie 010 was copied to 0x10). After filtering
out these differences, any remaining (real) incompatibilities
have been merged into the MI header file to make them more visible.

While here, fix the termios <-> termio conversion WRT to the c_cc
field for Alpha. The termios values do not match the termio values
and thus prevents us from copying.

By eliminating the Alpha MD copy of linux_ioctl.h we also fixed
the recent build breakage caused by putting new bits in the MI
header and not in the MD header.

22 years agos/kernal/kernel
Alfred Perlstein [Fri, 19 Oct 2001 07:52:11 +0000 (07:52 +0000)]
s/kernal/kernel

22 years agoHandle the lack of nfs server or client support in the kernel by
Doug Barton [Fri, 19 Oct 2001 06:50:52 +0000 (06:50 +0000)]
Handle the lack of nfs server or client support in the kernel by
kldload'ing the appropriate modules before enabling the service.

22 years agoLinux compiles broken on alpha again. Remove.
Matt Jacob [Fri, 19 Oct 2001 05:04:02 +0000 (05:04 +0000)]
Linux compiles broken on alpha again. Remove.

22 years agoAdd manpage for gx driver.
Jonathan Lemon [Fri, 19 Oct 2001 02:49:41 +0000 (02:49 +0000)]
Add manpage for gx driver.

22 years agoAdd entry for the PRO/1000.
Jonathan Lemon [Fri, 19 Oct 2001 02:28:12 +0000 (02:28 +0000)]
Add entry for the PRO/1000.

22 years agoHook the gx module up to the build.
Jonathan Lemon [Fri, 19 Oct 2001 02:21:16 +0000 (02:21 +0000)]
Hook the gx module up to the build.

22 years agoAdd a driver for the Intel PRO/1000 Gigabit Ethernet boards.
Jonathan Lemon [Fri, 19 Oct 2001 02:17:32 +0000 (02:17 +0000)]
Add a driver for the Intel PRO/1000 Gigabit Ethernet boards.

22 years ago#if 0 out some code that depends on other uncommitted patches.
Dag-Erling Smørgrav [Fri, 19 Oct 2001 01:52:13 +0000 (01:52 +0000)]
#if 0 out some code that depends on other uncommitted patches.

22 years agoAdapt to pseudofs changes (dynamic initialization, not static).
Dag-Erling Smørgrav [Fri, 19 Oct 2001 01:45:03 +0000 (01:45 +0000)]
Adapt to pseudofs changes (dynamic initialization, not static).
Use the new linux_ifname() function from the linuxulator rather than roll
our own interface name translation.

22 years agoSwitch to dynamic rather than static initialization.
Dag-Erling Smørgrav [Fri, 19 Oct 2001 01:43:06 +0000 (01:43 +0000)]
Switch to dynamic rather than static initialization.
This makes it possible (in theory) for nodes to be added and / or removed
from pseudofs filesystems at runtime.

22 years agoAdd support for the "device private" ioctls soon to be used by the an driver.
Dag-Erling Smørgrav [Fri, 19 Oct 2001 01:38:10 +0000 (01:38 +0000)]
Add support for the "device private" ioctls soon to be used by the an driver.
Also slightly change the name translation policy - only rename interfaces
that have the IFF_BROADCAST flag set.  This is not perfect, but is closer to
how Linux names network interfaces.

22 years agoAdd two "device private" ioctls needed by the Aironet (an) driver.
Dag-Erling Smørgrav [Fri, 19 Oct 2001 01:20:21 +0000 (01:20 +0000)]
Add two "device private" ioctls needed by the Aironet (an) driver.
Move some ioctls that were out of order.

22 years agoWhitespace fix.
Dag-Erling Smørgrav [Fri, 19 Oct 2001 00:46:17 +0000 (00:46 +0000)]
Whitespace fix.

22 years agoPrint statistics for AF_IPX.
Matthew N. Dodd [Fri, 19 Oct 2001 00:40:51 +0000 (00:40 +0000)]
Print statistics for AF_IPX.
Note that the IPX code doesn't update these correctly yet, but should.

22 years agoPrint [fddi] or [atm] tags for ARP addresses on interfaces of that type.
Matthew N. Dodd [Fri, 19 Oct 2001 00:33:26 +0000 (00:33 +0000)]
Print [fddi] or [atm] tags for ARP addresses on interfaces of that type.

22 years agoUse normal CWARNFLAGS and defines when compiling on FreeBSD.
Mark Peek [Fri, 19 Oct 2001 00:32:00 +0000 (00:32 +0000)]
Use normal CWARNFLAGS and defines when compiling on FreeBSD.

22 years agoTurn on the libedit support in bc. Turn off libreadline.
Matthew N. Dodd [Fri, 19 Oct 2001 00:24:39 +0000 (00:24 +0000)]
Turn on the libedit support in bc.  Turn off libreadline.

22 years agoAdd libedit support to tftp.
Matthew N. Dodd [Fri, 19 Oct 2001 00:23:07 +0000 (00:23 +0000)]
Add libedit support to tftp.

22 years agoAdd library exposed by KDE's use if this module.
Mark Murray [Thu, 18 Oct 2001 20:05:20 +0000 (20:05 +0000)]
Add library exposed by KDE's use if this module.

22 years agoAdd verbage for MODULES_OVERRIDE.
Warner Losh [Thu, 18 Oct 2001 19:44:13 +0000 (19:44 +0000)]
Add verbage for MODULES_OVERRIDE.

22 years agoUpdate to note that rpc.statd and rpc.lockd are now needed for
Alfred Perlstein [Thu, 18 Oct 2001 19:37:57 +0000 (19:37 +0000)]
Update to note that rpc.statd and rpc.lockd are now needed for
client side NFS mounts.

Stumbled upon by: rwatson

22 years agoAdd support for the gcc-2.95 stdarg implementation.
Mark Peek [Thu, 18 Oct 2001 19:11:12 +0000 (19:11 +0000)]
Add support for the gcc-2.95 stdarg implementation.

22 years agoProtect against deranged fabric nameservers that spit out 10000 identical
Matt Jacob [Thu, 18 Oct 2001 17:26:52 +0000 (17:26 +0000)]
Protect against deranged fabric nameservers that spit out 10000 identical
port numbers.

MFC after: 1 day

22 years agoBack previous revision out until it has been discussed on -arch and
Sheldon Hearn [Thu, 18 Oct 2001 16:53:20 +0000 (16:53 +0000)]
Back previous revision out until it has been discussed on -arch and
motivated.  Currently, it is under dispute.

22 years agoBack previous revision out until it has been discussed on -arch and
Sheldon Hearn [Thu, 18 Oct 2001 16:41:58 +0000 (16:41 +0000)]
Back previous revision out until it has been discussed on -arch and
motivated.  Currently, it is under dispute.

22 years agoShift the code which packs and unpacks instruction bundles out of DDB
Doug Rabson [Thu, 18 Oct 2001 16:20:04 +0000 (16:20 +0000)]
Shift the code which packs and unpacks instruction bundles out of DDB
since it is useful for various emulations duties (e.g. unaligned trap
handling).

22 years agoMy attempts at minimizing the number of #def's got me in trouble.
David E. O'Brien [Thu, 18 Oct 2001 16:07:21 +0000 (16:07 +0000)]
My attempts at minimizing the number of #def's got me in trouble.

22 years agoNote MFC of syslogd -b.
Dima Dorfman [Thu, 18 Oct 2001 15:23:55 +0000 (15:23 +0000)]
Note MFC of syslogd -b.

22 years agoAdd experimental support for sending keypad facility messages.
Hellmuth Michaelis [Thu, 18 Oct 2001 11:58:49 +0000 (11:58 +0000)]
Add experimental support for sending keypad facility messages.
MFC after: 2 months

22 years agoFix diskless clients by removing the code for calculating the minimum
Ollivier Robert [Thu, 18 Oct 2001 09:48:28 +0000 (09:48 +0000)]
Fix diskless clients by removing the code for calculating the minimum
value for cpg. The change was bogus.

Submitted by: bde
MFC after: 2 days

22 years agoFix AUTH callback negotiations.
Brian Somers [Thu, 18 Oct 2001 08:48:04 +0000 (08:48 +0000)]
Fix AUTH callback negotiations.

MFC after: 1 week

22 years agoWhen the peer fails to specify an MRU and a 1500 byte MRU is not
Brian Somers [Thu, 18 Oct 2001 08:47:56 +0000 (08:47 +0000)]
When the peer fails to specify an MRU and a 1500 byte MRU is not
allowed either because of the transport or configuration, send a
MRU NAK only once, then allow the negotiations to proceed.

rfc1661 says that 1500 should always be allowed and rfc2516 says
that 1492 is the maximum for PPPoE.  This changes ppp so that it
only weakly suggests 1492, then goes with the default (leaving
the problem in the hands of the peer WRT how they set their MTU).

MFC after: 1 week

22 years agoTons of type, style and warning fixes that have been rotting in my tree for
Dag-Erling Smørgrav [Thu, 18 Oct 2001 08:29:26 +0000 (08:29 +0000)]
Tons of type, style and warning fixes that have been rotting in my tree for
ages - some of which wouldn't be necessary if gcc wasn't broken or TPTB were
willing to do something (-fno-builtin) about it.

22 years agokeypad facility information element: add keypad facility information decoding.
Hellmuth Michaelis [Thu, 18 Oct 2001 08:03:23 +0000 (08:03 +0000)]
keypad facility information element: add keypad facility information decoding.
MFC after: 1 month

22 years agoFix typos in previous commit:
Marcel Moolenaar [Thu, 18 Oct 2001 05:21:52 +0000 (05:21 +0000)]
Fix typos in previous commit:

o  s/sys_narg/sy_narg/
o  s/SYS_MPSAFE/SYF_MPSAFE/

22 years agoAdd support for "__gnuc_va_list". Some overly "smart" libraries assume
David E. O'Brien [Thu, 18 Oct 2001 00:27:39 +0000 (00:27 +0000)]
Add support for "__gnuc_va_list".  Some overly "smart" libraries assume
the existence of the __gnuc_va_list type[*] because our compiler is GCC.

[*] __gnuc_va_list is defined in the GCC ginclude/stdarg.h replacement
headerwhich we don't use.

22 years agoSync comments with i386.
David E. O'Brien [Thu, 18 Oct 2001 00:00:31 +0000 (00:00 +0000)]
Sync comments with i386.

22 years agoMinor comment tweaking.
David E. O'Brien [Wed, 17 Oct 2001 23:59:29 +0000 (23:59 +0000)]
Minor comment tweaking.

22 years ago- Small cleanups to the Giant handling in trap().
John Baldwin [Wed, 17 Oct 2001 22:04:45 +0000 (22:04 +0000)]
- Small cleanups to the Giant handling in trap().
- Only release Giant in trap() if we locked it, otherwise we could release
  Giant in a kernel trap if we didn't get it for a page fault and the
  previous frame had grabbed the lock.
- Only get Giant for !MP safe syscalls.

22 years agoSmall cleanups to the handling of Giant in trap().
John Baldwin [Wed, 17 Oct 2001 21:55:56 +0000 (21:55 +0000)]
Small cleanups to the handling of Giant in trap().

22 years agoUse the new SIOCGIFINDEX ioctl to efficiently map a name to an index.
Jonathan Lemon [Wed, 17 Oct 2001 20:08:15 +0000 (20:08 +0000)]
Use the new SIOCGIFINDEX ioctl to efficiently map a name to an index.
If the syscall fails, fall back on the old method as a compatability
measure.

22 years agoAdd a SIOCGIFINDEX ioctl, which returns the index of a named interface.
Jonathan Lemon [Wed, 17 Oct 2001 19:40:44 +0000 (19:40 +0000)]
Add a SIOCGIFINDEX ioctl, which returns the index of a named interface.
This will be used to more efficiently support if_nametoindex(3).

22 years agoCleanup ifunit(), so it uses the dev_named() function to map an interface
Jonathan Lemon [Wed, 17 Oct 2001 18:58:14 +0000 (18:58 +0000)]
Cleanup ifunit(), so it uses the dev_named() function to map an interface
name into a device.