openbsd
8 years agoStop recommending the non-standard and slightly dangerous fgetln(3).
schwarze [Sat, 27 Aug 2016 16:39:42 +0000 (16:39 +0000)]
Stop recommending the non-standard and slightly dangerous fgetln(3).
Recommend POSIX getline(3) instead.

8 years agoWhen a precision is specified for a string format use strnlen()
millert [Sat, 27 Aug 2016 16:10:40 +0000 (16:10 +0000)]
When a precision is specified for a string format use strnlen()
to determine the length instead of doing it manually.  OK schwarze@

8 years agoEnable ALPN regress now that it passes.
jsing [Sat, 27 Aug 2016 16:00:27 +0000 (16:00 +0000)]
Enable ALPN regress now that it passes.

8 years agoBe more strict when parsing TLS extensions.
jsing [Sat, 27 Aug 2016 15:58:06 +0000 (15:58 +0000)]
Be more strict when parsing TLS extensions.

Based on a diff from Kinichiro Inoguchi.

ok beck@

8 years agothe quotes in "unkempt thoughts" were authored by stanislaw lec,
jmc [Sat, 27 Aug 2016 14:29:22 +0000 (14:29 +0000)]
the quotes in "unkempt thoughts" were authored by stanislaw lec,
not stanislaw lem (an easy blunder, i'd say); i've also inserted the "J."
into one of these credits (Stanislaw J. Lec) since all the others have them;

credit to antoni grzymala; diff pockled from netbsd

8 years agoAdd support for the PXN bit in level 1 translation table descriptors and
kettenis [Sat, 27 Aug 2016 14:22:35 +0000 (14:22 +0000)]
Add support for the PXN bit in level 1 translation table descriptors and
enable it on CPUs that support it.  When enabled, this prevents the kernel
from executing userland code.

ok jsg@, tom@

8 years agoAdd support for the watchdog timer found on A31 and later SoCs, which has a
kettenis [Sat, 27 Aug 2016 14:13:14 +0000 (14:13 +0000)]
Add support for the watchdog timer found on A31 and later SoCs, which has a
slightly different register layout.  Also fix the watchdog such that it
actually disables itself when you set the period to 0 and actually gives you
a period that is close to what you asked for.

8 years agoHandle the quirky that the FIFO register sits at a different location on
kettenis [Sat, 27 Aug 2016 12:41:47 +0000 (12:41 +0000)]
Handle the quirky that the FIFO register sits at a different location on
the A31 and later SoCs.

8 years agoimprove revision 1.2: in unusual cases, fgetwc(3) can succeed
schwarze [Sat, 27 Aug 2016 12:08:38 +0000 (12:08 +0000)]
improve revision 1.2: in unusual cases, fgetwc(3) can succeed
even though ferror(3) is already set;
also from Andrey Chernov <ache at freebsd dot org>;
OK millert@

8 years agoAdd support for sun8i-h3, the Allwinner H3. For this SoC, the device tree
kettenis [Sat, 27 Aug 2016 11:39:59 +0000 (11:39 +0000)]
Add support for sun8i-h3, the Allwinner H3.  For this SoC, the device tree
contains a single clock control unit node that handles most clocks.  The
driver handles this through tables that describe the gating, reset signals
and the clock hierarchy.  This description is (deliberately) incomplete.  We
will add clocks on an as-needed basis.  You will need a recent device tree,
that includes a "allwinner,sun8i-h3-ccu" compatible node for things to work.

ok patrick@

8 years agoKill p_instance from proc.c and remove static proc_id unused variables.
rzalamena [Sat, 27 Aug 2016 11:13:16 +0000 (11:13 +0000)]
Kill p_instance from proc.c and remove static proc_id unused variables.

To keep the debug functionality intact and correct we'll use the pid
field in the imsg header to pass the instance number. Remember to always
pass 'ps_instance + 1' otherwise libutil will fill imsg header pid field
with the imsgbuf pid (which is the current process pid).

ok reyk@

8 years agoAdd missing $OpenBSD$ line and remove an unnecessary comment line.
yasuoka [Sat, 27 Aug 2016 09:04:20 +0000 (09:04 +0000)]
Add missing $OpenBSD$ line and remove an unnecessary comment line.

8 years agoPull in <stdio.h> for NULL
guenther [Sat, 27 Aug 2016 07:35:19 +0000 (07:35 +0000)]
Pull in <stdio.h> for NULL

ok deraadt@

8 years agoqaddr_t is (now) dead
guenther [Sat, 27 Aug 2016 05:14:42 +0000 (05:14 +0000)]
qaddr_t is (now) dead

ok deraadt@

8 years agoDeclare lseek() instead of assuming it'll be provided elsewhere
guenther [Sat, 27 Aug 2016 04:57:30 +0000 (04:57 +0000)]
Declare lseek() instead of assuming it'll be provided elsewhere

ok deraadt@

8 years agoMove up the <stdio.h> include so <dev/pci/*.h> have NULL
guenther [Sat, 27 Aug 2016 04:38:48 +0000 (04:38 +0000)]
Move up the <stdio.h> include so <dev/pci/*.h> have NULL

ok deraadt@

8 years agoPull in <stdio.h> for NULL
guenther [Sat, 27 Aug 2016 04:35:18 +0000 (04:35 +0000)]
Pull in <stdio.h> for NULL

ok deraadt@

8 years agoPull in <sys/time.h> for struct timeval
guenther [Sat, 27 Aug 2016 04:34:28 +0000 (04:34 +0000)]
Pull in <sys/time.h> for struct timeval
<sys/*.h> includes go before <*.h> includes

ok deraadt@

8 years agoPull in <sys/time.h> for gettimeofday()
guenther [Sat, 27 Aug 2016 04:32:44 +0000 (04:32 +0000)]
Pull in <sys/time.h> for gettimeofday()
Sort #includes

ok deraadt@

8 years agoALL HAIL ELF!
guenther [Sat, 27 Aug 2016 04:31:22 +0000 (04:31 +0000)]
ALL HAIL ELF!
Since <sys/param.h> isn't needed for __ELF__, pull in <stdio.h> for NULL

ok deraadt@

8 years agoPull in <stdio.h> for NULL
guenther [Sat, 27 Aug 2016 04:29:25 +0000 (04:29 +0000)]
Pull in <stdio.h> for NULL

ok deraadt@

8 years agoPull in <sys/select.h> for fd_set
guenther [Sat, 27 Aug 2016 04:28:28 +0000 (04:28 +0000)]
Pull in <sys/select.h> for fd_set

ok deraadt@

8 years agoPull in <sys/select.h> for fd_set
guenther [Sat, 27 Aug 2016 04:21:08 +0000 (04:21 +0000)]
Pull in <sys/select.h> for fd_set

ok deraadt@

8 years agoPull in <sys/time.h> for gettimeofday()
guenther [Sat, 27 Aug 2016 04:15:52 +0000 (04:15 +0000)]
Pull in <sys/time.h> for gettimeofday()

ok deraadt@

8 years agoPull in <sys/select.h> for howmany() and NBBY, used by various net*/*.h
guenther [Sat, 27 Aug 2016 04:13:43 +0000 (04:13 +0000)]
Pull in <sys/select.h> for howmany() and NBBY, used by various net*/*.h

ok deraadt@

8 years agoPull in <sys/time.h> for struct timespec and timeval
guenther [Sat, 27 Aug 2016 04:07:42 +0000 (04:07 +0000)]
Pull in <sys/time.h> for struct timespec and timeval

ok deraadt@

8 years agoPull in <sys/time.h> for struct timeval
guenther [Sat, 27 Aug 2016 04:05:12 +0000 (04:05 +0000)]
Pull in <sys/time.h> for struct timeval

ok deraadt@

8 years agoPull in <stdlib.h> for NULL
guenther [Sat, 27 Aug 2016 04:04:56 +0000 (04:04 +0000)]
Pull in <stdlib.h> for NULL

ok deraadt@

8 years agoPull in <stdio.h> for NULL
guenther [Sat, 27 Aug 2016 03:55:43 +0000 (03:55 +0000)]
Pull in <stdio.h> for NULL

ok deraadt@

8 years agoPull in <stdio.h> for NULL
guenther [Sat, 27 Aug 2016 03:54:20 +0000 (03:54 +0000)]
Pull in <stdio.h> for NULL

ok deraadt@

8 years agoPull in <stddef.h> for NULL
guenther [Sat, 27 Aug 2016 03:52:25 +0000 (03:52 +0000)]
Pull in <stddef.h> for NULL
(not stdio, because ld.so has a reduced library behind it)

8 years agoPull in <sys/time.h> for struct timespec
guenther [Sat, 27 Aug 2016 03:48:00 +0000 (03:48 +0000)]
Pull in <sys/time.h> for struct timespec

ok deraadt@

8 years agoPull in <stdio.h> for NULL
guenther [Sat, 27 Aug 2016 03:34:45 +0000 (03:34 +0000)]
Pull in <stdio.h> for NULL

ok deraadt@

8 years agosome updates for recent changes.
tedu [Sat, 27 Aug 2016 03:31:47 +0000 (03:31 +0000)]
some updates for recent changes.
not all tests passing yet since i don't really understand what's going on.

8 years agoPull in <sys/select.h> for fd_set
guenther [Sat, 27 Aug 2016 02:11:27 +0000 (02:11 +0000)]
Pull in <sys/select.h> for fd_set
Pull in <time.h> for time()

ok deraadt@

8 years agoPull in <sys/select.h> for fd_set
guenther [Sat, 27 Aug 2016 02:06:40 +0000 (02:06 +0000)]
Pull in <sys/select.h> for fd_set

ok deraadt@

8 years agoPull in <sys/time.h> for struct timespec, gettimeofday(), clock_gettime(),
guenther [Sat, 27 Aug 2016 02:02:44 +0000 (02:02 +0000)]
Pull in <sys/time.h> for struct timespec, gettimeofday(), clock_gettime(),
and setitimer()

ok deraadt@

8 years agoPull in <time.h> for time() and perhaps other functions
guenther [Sat, 27 Aug 2016 02:00:10 +0000 (02:00 +0000)]
Pull in <time.h> for time() and perhaps other functions

ok deraadt@

8 years agoPull in <sys/time.h> for struct timespec
guenther [Sat, 27 Aug 2016 01:57:27 +0000 (01:57 +0000)]
Pull in <sys/time.h> for struct timespec

ok deraadt@

8 years agoPull in <sys/time.h> for struct timespec
guenther [Sat, 27 Aug 2016 01:56:07 +0000 (01:56 +0000)]
Pull in <sys/time.h> for struct timespec
<sys/fcntl.h> shouldn't be used by userspace; this already uses <fcntl.h>

ok deraadt@

8 years agoPull in <sys/time.h> for struct timespec
guenther [Sat, 27 Aug 2016 01:55:30 +0000 (01:55 +0000)]
Pull in <sys/time.h> for struct timespec

ok deraadt@

8 years agoPull in <sys/time.h> for struct timespec, timeval, or clockrate
guenther [Sat, 27 Aug 2016 01:50:07 +0000 (01:50 +0000)]
Pull in <sys/time.h> for struct timespec, timeval, or clockrate

ok deraadt@

8 years agoPull in <sys/time.h> for gettimeofday()
guenther [Sat, 27 Aug 2016 01:42:37 +0000 (01:42 +0000)]
Pull in <sys/time.h> for gettimeofday()

ok deraadt@

8 years agoUse %zu for size_t arguments
guenther [Sat, 27 Aug 2016 01:30:39 +0000 (01:30 +0000)]
Use %zu for size_t arguments

8 years agoPull in <time.h> for one or more of gmtime, strftime, strptime, time,
guenther [Sat, 27 Aug 2016 01:26:22 +0000 (01:26 +0000)]
Pull in <time.h> for one or more of gmtime, strftime, strptime, time,
timegm, and tzset

ok deraadt@

8 years agoImplement bus dma support for loading raw mappings so that we can use
patrick [Fri, 26 Aug 2016 21:50:42 +0000 (21:50 +0000)]
Implement bus dma support for loading raw mappings so that we can use
xhci(4) on ARM.  The only way the load raw operation can get to know
about the coherent flag is via the segments.  Store it there when the
memory is initially mapped.  Also store the virtual address which we
need to know when we have to flush the caches on a non-coherent mapping.

ok kettenis@

8 years agoAdd support for the Allwinner H3 (sun8i-h3).
kettenis [Fri, 26 Aug 2016 21:23:41 +0000 (21:23 +0000)]
Add support for the Allwinner H3 (sun8i-h3).

8 years agoSince the caller already checks and handles COHERENT we don't need
patrick [Fri, 26 Aug 2016 21:14:58 +0000 (21:14 +0000)]
Since the caller already checks and handles COHERENT we don't need
to explicitly check for the flag.

ok kettenis@

8 years agogrrr warning
espie [Fri, 26 Aug 2016 18:19:21 +0000 (18:19 +0000)]
grrr warning

8 years agoRemove cpu_cpwait() calls; they are no-ops on armv7.
kettenis [Fri, 26 Aug 2016 16:02:33 +0000 (16:02 +0000)]
Remove cpu_cpwait() calls; they are no-ops on armv7.

ok tom@

8 years agoRepeated occurances of the idiom buf[5][BUFSIZ] -- ridiculous. Give each
deraadt [Fri, 26 Aug 2016 15:04:15 +0000 (15:04 +0000)]
Repeated occurances of the idiom buf[5][BUFSIZ] -- ridiculous.  Give each
buf a special name, recognize that most are PATH_MAX, and remove a few that
are not needed at all.
ok jsing beck

8 years agoMake regress Makefile consistent with other tests, use bsd.regress.mk.
bluhm [Fri, 26 Aug 2016 13:10:30 +0000 (13:10 +0000)]
Make regress Makefile consistent with other tests, use bsd.regress.mk.
OK otto@

8 years agoKill the ps_ninstances from proc.c.
rzalamena [Fri, 26 Aug 2016 12:24:21 +0000 (12:24 +0000)]
Kill the ps_ninstances from proc.c.

We got the same information in ps_instances[proc] (more accurate) and
we avoid allocating unnecessary memory for pipe storage.

ok reyk@

8 years agoRemove the code that switches around MMU domains on armv7. MMU domains are
kettenis [Fri, 26 Aug 2016 11:59:04 +0000 (11:59 +0000)]
Remove the code that switches around MMU domains on armv7.  MMU domains are
basically a relic from the past.  Using them doesn't make a lot of sense
the way our pmaps work.  Support for MMU domains isn't present in
long-descriptor translation table format, so it is clearly on its way out.

Based on a diff from Artituri Alm.

ok patrick@

8 years agoReplace the static env variables with a single global variable.
rzalamena [Fri, 26 Aug 2016 10:46:39 +0000 (10:46 +0000)]
Replace the static env variables with a single global variable.

ok reyk@

8 years agoPull in <time.h> for time()
guenther [Fri, 26 Aug 2016 09:10:11 +0000 (09:10 +0000)]
Pull in <time.h> for time()

ok deraadt@

8 years agoPull in <time.h> for struct tm, used in rcs.h
guenther [Fri, 26 Aug 2016 09:02:54 +0000 (09:02 +0000)]
Pull in <time.h> for struct tm, used in rcs.h

ok deraadt@

8 years agoPull in <sys/select.h> for fd_set
guenther [Fri, 26 Aug 2016 08:52:19 +0000 (08:52 +0000)]
Pull in <sys/select.h> for fd_set

ok deraadt@

8 years agoPull in <time.h> for ctime_r, gmtime, etc
guenther [Fri, 26 Aug 2016 08:50:31 +0000 (08:50 +0000)]
Pull in <time.h> for ctime_r, gmtime, etc

ok deraadt@

8 years agoPull in <sys/time.h> for gettimeofday()
guenther [Fri, 26 Aug 2016 08:44:04 +0000 (08:44 +0000)]
Pull in <sys/time.h> for gettimeofday()

ok deraadt@

8 years agoRemove unused prototype for sxiccmu_enablemodule().
mglocker [Fri, 26 Aug 2016 08:30:24 +0000 (08:30 +0000)]
Remove unused prototype for sxiccmu_enablemodule().

ok kettenis

8 years agoPull in <time.h> for clock_gettime()
guenther [Fri, 26 Aug 2016 08:25:02 +0000 (08:25 +0000)]
Pull in <time.h> for clock_gettime()

ok deraadt@

8 years agoThe *_HEAD_INITIALIZER() macros are documented as taking the struct,
guenther [Fri, 26 Aug 2016 07:12:30 +0000 (07:12 +0000)]
The *_HEAD_INITIALIZER() macros are documented as taking the struct,
not a pointer to it

8 years agoPull in <sys/time.h> for clock_ymdhms bits
guenther [Fri, 26 Aug 2016 07:09:56 +0000 (07:09 +0000)]
Pull in <sys/time.h> for clock_ymdhms bits

ok deraadt@

8 years agotrim down some NBBY references. 8 bits ought to be enough for anyone.
tedu [Fri, 26 Aug 2016 06:32:10 +0000 (06:32 +0000)]
trim down some NBBY references. 8 bits ought to be enough for anyone.

8 years agoPull in <sys/select.h> for fd_set
guenther [Fri, 26 Aug 2016 06:18:01 +0000 (06:18 +0000)]
Pull in <sys/select.h> for fd_set

ok deraadt@

8 years agoAdd <time.h> for time(); sort <*.h> includes
guenther [Fri, 26 Aug 2016 06:06:58 +0000 (06:06 +0000)]
Add <time.h> for time(); sort <*.h> includes

ok deraadt@

8 years agoAdd <time.h> for time() and ctime(); sort <net*/*.h>
guenther [Fri, 26 Aug 2016 06:04:10 +0000 (06:04 +0000)]
Add <time.h> for time() and ctime(); sort <net*/*.h>

ok deraadt@

8 years agoPull in <sys/time.h> for gettimeofday()
guenther [Fri, 26 Aug 2016 05:50:29 +0000 (05:50 +0000)]
Pull in <sys/time.h> for gettimeofday()

8 years agoNeed <time.h> for time() and others. Pulled in by coincidence previously
guenther [Fri, 26 Aug 2016 05:06:14 +0000 (05:06 +0000)]
Need <time.h> for time() and others.  Pulled in by coincidence previously

noted by tb@

8 years ago-E NONE has *never* worked; it was always -E none. It's a dubious option
guenther [Fri, 26 Aug 2016 04:40:27 +0000 (04:40 +0000)]
-E NONE has *never* worked; it was always -E none.  It's a dubious option
value that merits a warning in the manpage and using 2 billion will get
you practically the same effect, so delete the -E none support

8 years agooptions.h is only used by options.c; merge it into the .c file
guenther [Fri, 26 Aug 2016 04:33:11 +0000 (04:33 +0000)]
options.h is only used by options.c; merge it into the .c file
reduce and sort #includes

8 years agoreduce and sort #includes
guenther [Fri, 26 Aug 2016 04:31:35 +0000 (04:31 +0000)]
reduce and sort #includes

8 years agotables.h is only used by tables.c; merge it into the .c file
guenther [Fri, 26 Aug 2016 04:23:44 +0000 (04:23 +0000)]
tables.h is only used by tables.c; merge it into the .c file
<sys/time.h> is unnecessary; sort the #includes

8 years ago<sys/time.h>, <errno.h>, and <stdlib.h> are unnecessary; sort #includes
guenther [Fri, 26 Aug 2016 04:22:13 +0000 (04:22 +0000)]
<sys/time.h>, <errno.h>, and <stdlib.h> are unnecessary; sort #includes

8 years agosel_subs.h is only used by sel_subs.c; merge it into the .c file
guenther [Fri, 26 Aug 2016 04:20:38 +0000 (04:20 +0000)]
sel_subs.h is only used by sel_subs.c; merge it into the .c file
<sys/time.h> and <unistd.h> are unnecessary, but <time.h> is; sort #includes

8 years agopat_rep.h is only used by pat_rep.c; merge it into the .c file
guenther [Fri, 26 Aug 2016 04:19:28 +0000 (04:19 +0000)]
pat_rep.h is only used by pat_rep.c; merge it into the .c file
<sys/time.h>, <errno.h> and <unistd.h> are unnecessary; sort #includes

8 years agoftree.h is only used by ftree.c; merge it into the .c file
guenther [Fri, 26 Aug 2016 04:17:48 +0000 (04:17 +0000)]
ftree.h is only used by ftree.c; merge it into the .c file
<sys/time.h> is unnecessary; sort #includes

8 years agoDon't need <sys/time.h> or "options.h" here
guenther [Fri, 26 Aug 2016 04:11:16 +0000 (04:11 +0000)]
Don't need <sys/time.h> or "options.h" here

8 years agocache.h is only used by cache.c; merge it into the .c file
guenther [Fri, 26 Aug 2016 04:08:18 +0000 (04:08 +0000)]
cache.h is only used by cache.c; merge it into the .c file
<unistd.h> and <sys/time.h> are unneeded here

ok by general acclaim

8 years agocorrect INTR_MSG_MASK to include INTR_MSG_1
jsg [Fri, 26 Aug 2016 00:45:50 +0000 (00:45 +0000)]
correct INTR_MSG_MASK to include INTR_MSG_1
ok dlg@

8 years agoadd a sIgnore opcode that silently ignores options and use it to
djm [Thu, 25 Aug 2016 23:57:54 +0000 (23:57 +0000)]
add a sIgnore opcode that silently ignores options and use it to
suppress noisy deprecation warnings for the Protocol directive.

req henning, ok markus

8 years agoremove superfluous NOTREACHED comment
djm [Thu, 25 Aug 2016 23:56:51 +0000 (23:56 +0000)]
remove superfluous NOTREACHED comment

8 years agoSet relayd socket buffer size to reasonable value to make test pass
bluhm [Thu, 25 Aug 2016 23:37:03 +0000 (23:37 +0000)]
Set relayd socket buffer size to reasonable value to make test pass
in different environments.

8 years agoMake relayd test slow-consumer more reliable. Set SO_SNDTIMEO
bluhm [Thu, 25 Aug 2016 22:56:13 +0000 (22:56 +0000)]
Make relayd test slow-consumer more reliable.  Set SO_SNDTIMEO
sockopt correctly with Perl pack on i386.  Make it possible to grep
in each others client and server logfile.  Client does not simply
sleep but waits for short write at server.

8 years agoAfter read errors, fgetln(3) sometimes succeeded (returning non-NULL)
schwarze [Thu, 25 Aug 2016 19:21:33 +0000 (19:21 +0000)]
After read errors, fgetln(3) sometimes succeeded (returning non-NULL)
and failed (setting errno and ferror(3)) both at the same time.
That's a bad idea in general, and here in particular since
returning partial lines was neither reliable (sometimes, you
got NULL anyway) nor predictable (almost always, the line would
be truncated long before the actual read error).
Instead, on read failure, fail properly and always return NULL.
Issue found in a discussion with Andrey Chernov <ache at freebsd dot org>
who finally agreed to move FreeBSD into the same direction.
The fix is joint work with and OK by millert@.

8 years agoFix sosplice tests on 32 bit systems by putting the correct number
bluhm [Thu, 25 Aug 2016 17:13:59 +0000 (17:13 +0000)]
Fix sosplice tests on 32 bit systems by putting the correct number
of bytes into the time_t Perl pack template.  Also fix error messages.

8 years ago-F before -f (as -I before -i currently is...);
jmc [Thu, 25 Aug 2016 16:27:16 +0000 (16:27 +0000)]
-F before -f (as -I before -i currently is...);

8 years agoRemove a check to expire received prefix.
mpi [Thu, 25 Aug 2016 16:12:16 +0000 (16:12 +0000)]
Remove a check to expire received prefix.

This check does not make sense since November 2000 when IPv6 autoconf
address deletion has been made independent from prefix lifetimes.

Fix a case when a route was added to the table but the corresponding
address was not, leaving v6 unusable.

Found the hardway by and ok sthen@

8 years agowhen using -o, we may restart a match in the middle of the line.
tedu [Thu, 25 Aug 2016 15:11:05 +0000 (15:11 +0000)]
when using -o, we may restart a match in the middle of the line.
set NOTBOL so that anchored patterns don't match.
from a patch by DaniĆ«l de Kok in freebsd bug 201650
ok martijn

8 years agomake the full filesystem check optional, by popular demand.
espie [Thu, 25 Aug 2016 14:58:43 +0000 (14:58 +0000)]
make the full filesystem check optional, by popular demand.
doubling -q removes the plist thorough check.

8 years agoAmend after clean-up of fortran.port.mk
dcoppa [Thu, 25 Aug 2016 14:57:35 +0000 (14:57 +0000)]
Amend after clean-up of fortran.port.mk

8 years agoshorten the pkeyutl text;
jmc [Thu, 25 Aug 2016 14:37:28 +0000 (14:37 +0000)]
shorten the pkeyutl text;
help/ok guenther

8 years agoSpliced TCP sockets become faster when the output part is running
bluhm [Thu, 25 Aug 2016 14:13:19 +0000 (14:13 +0000)]
Spliced TCP sockets become faster when the output part is running
as its own task thread.  This is inspired by userland copy where a
process also has to go through the scheduler.  This gives the socket
buffer a chance to be filled up and tcp_output() is called less
often and with bigger chunks.
When two kernel tasks share all the workload, the current scheduler
implementation will hang userland processes on single cpu machines.
As a workaround put a yield() into the splicing thread after each
task execution.  This reduces the number of calls of tcp_output()
even more.
OK tedu@ mpi@

8 years agoCompletely revert the M_WAIT change on the cluster allocation and
bluhm [Thu, 25 Aug 2016 13:59:16 +0000 (13:59 +0000)]
Completely revert the M_WAIT change on the cluster allocation and
bring back the behaviour of rev 1.72.  Although allocating small
mbufs when allocating an mbuf cluster fails seems suboptimal, this
should not be changed as a side effect when introducing m_getuio().
OK claudio@

8 years agoSimplify ip6_getpmtu() to use a 'struct rtentry *' instead of two
mpi [Thu, 25 Aug 2016 12:30:16 +0000 (12:30 +0000)]
Simplify ip6_getpmtu() to use a 'struct rtentry *' instead of two
'struct route_in6 *'.

This is another little step towards deprecating 'struct route{,_in6}'

Inputs from and ok bluhm@

8 years agogoda@ contributed most of the OpenFlow 1.3.5 parts in ofp.h (for the
reyk [Thu, 25 Aug 2016 11:00:44 +0000 (11:00 +0000)]
goda@ contributed most of the OpenFlow 1.3.5 parts in ofp.h (for the
"still-unreleased" switch(4) driver), so it is only fair to add his
copyright.  This header file will eventually move to net/ofp.h.

8 years agoFix pfsync(4)'s carp examples: as of ip_carp.c r1.245, carpdev must be
sthen [Thu, 25 Aug 2016 10:51:37 +0000 (10:51 +0000)]
Fix pfsync(4)'s carp examples: as of ip_carp.c r1.245, carpdev must be
specified.  From Bryan Stenson.

8 years agoFix rtadvd.conf(5) example for pinfoflags, it accepts numbers not strings.
sthen [Thu, 25 Aug 2016 09:41:13 +0000 (09:41 +0000)]
Fix rtadvd.conf(5) example for pinfoflags, it accepts numbers not strings.
ok phessler

8 years agoDo not crash if display-message used without a client, issue reported by
nicm [Thu, 25 Aug 2016 09:33:14 +0000 (09:33 +0000)]
Do not crash if display-message used without a client, issue reported by
Serge Aleynikov, fix from Thomas Adam.