openbsd
10 years agoMake sure state changes are properly serialized.
mpi [Tue, 19 Aug 2014 12:24:04 +0000 (12:24 +0000)]
Make sure state changes are properly serialized.

When pms(4) is attached to a touchpad it generally presents two different
wsmouse(4) devices: one for the touchpad itself and one for the clitpad
and/or some interleaved packets.  But since both devices are writing to
the same pckbc slot, a race can occur if they try to change the state at
the same time.

So prevent two process opening the two /dev/wsmouse* node at the same time
to corrupt the magic sequences needed to enable/disable the touchpad.

ok schadchin@

10 years agoin myx_start, replace
dlg [Tue, 19 Aug 2014 11:13:16 +0000 (11:13 +0000)]
in myx_start, replace

while (space) {
IFQ_POLL;
myx_dequeue(free descr);
IFQ_DEQUEUE;
etc;
}

with

while (space && myx_dequeue(free descr)) {
IFQ_DEQUEUE;
etc;
}

10 years agold.so(1): Missing stack var initialization
uebayasi [Tue, 19 Aug 2014 10:01:50 +0000 (10:01 +0000)]
ld.so(1): Missing stack var initialization

Explicitly zero-clear stack array to read dynamic segments.  This code has
been working because stack is mapped as anon and zero'ed right after execve(2).

Found by work-in-progress Valgrind port for OpenBSD/amd64.

OK millert@ guenther@

10 years agopppx(4): Correct pppx_dev_lookup() and pppx_dev2pxd() local declarations
uebayasi [Tue, 19 Aug 2014 08:04:51 +0000 (08:04 +0000)]
pppx(4): Correct pppx_dev_lookup() and pppx_dev2pxd() local declarations

Those two functions take one dev_t argument, not int.  Match declarations
with reality.  No functional changes.

10 years agolibroff depends on libmandoc and so libmandoc should be included first.
daniel [Tue, 19 Aug 2014 05:18:16 +0000 (05:18 +0000)]
libroff depends on libmandoc and so libmandoc should be included first.

this change makes the code slightly more portable by removing a gcc
extension (incomplete enum types) which is not part of ISO C.

ok schwarze@

10 years agoadd -4 and -6 flags for saying if you want to force ipv4 or ipv6
dlg [Tue, 19 Aug 2014 03:28:53 +0000 (03:28 +0000)]
add -4 and -6 flags for saying if you want to force ipv4 or ipv6
respectively.

tweaks from jmc@
ok djm@

10 years agoThe _exit codes in syslogd privsep.c were the wrong way around.
bluhm [Tue, 19 Aug 2014 00:53:01 +0000 (00:53 +0000)]
The _exit codes in syslogd privsep.c were the wrong way around.
Fatal errors should exit with 1, and regular shutdown should result
in exit with 0.
OK deraadt@

10 years agoReplace manually written function names with __func__.
bluhm [Tue, 19 Aug 2014 00:28:48 +0000 (00:28 +0000)]
Replace manually written function names with __func__.
OK sthen@

10 years agoFix trailing white spaces.
bluhm [Tue, 19 Aug 2014 00:24:00 +0000 (00:24 +0000)]
Fix trailing white spaces.
OK sthen@

10 years agoMake the perl modules consistent for the multiple regression tests.
bluhm [Mon, 18 Aug 2014 22:58:19 +0000 (22:58 +0000)]
Make the perl modules consistent for the multiple regression tests.
This includes coding style, better error messages and variable
naming.

10 years agoFix read access to uninitialized memory found by kristaps@ with valgrind,
schwarze [Mon, 18 Aug 2014 22:21:52 +0000 (22:21 +0000)]
Fix read access to uninitialized memory found by kristaps@ with valgrind,
patch from kristaps@.
Theoretically, this could lead to a buffer overrun and segfault,
but only for very long output lines (about 1000 charecters) of
exactly the right length or if by ill chance, the complete unused,
uninitialized tail of the line output buffer (at least 1000 characters
total length) would contain all blank characters.  Anyway, the
uninitialized data read wasn't used for anything.

10 years agoAdd a description of what the relayd and sosplice regression tests
bluhm [Mon, 18 Aug 2014 22:05:08 +0000 (22:05 +0000)]
Add a description of what the relayd and sosplice regression tests
do.  Also explain the optional environment variables.

10 years agoMake out of band data tests more reliable. As multiple tcp urgent
bluhm [Mon, 18 Aug 2014 21:53:40 +0000 (21:53 +0000)]
Make out of band data tests more reliable.  As multiple tcp urgent
bytes cannot be handled correctly, allow another variation.

10 years agoMake the forking tests more reliable. During copy the read(2)
bluhm [Mon, 18 Aug 2014 21:51:45 +0000 (21:51 +0000)]
Make the forking tests more reliable.  During copy the read(2)
system call might fail with EAGAIN when the other forked process
has already read the data.

10 years agoMake the multiple UDP packet test more reliable. The receive buffer
bluhm [Mon, 18 Aug 2014 21:49:17 +0000 (21:49 +0000)]
Make the multiple UDP packet test more reliable.  The receive buffer
limit also counts the mbufs in it, so packets could be dropped.
Increase both send and receive socket buffer size.

10 years agoAdd test cases for multiple small HTTP requests. Many PUT requests
bluhm [Mon, 18 Aug 2014 21:43:45 +0000 (21:43 +0000)]
Add test cases for multiple small HTTP requests.  Many PUT requests
with ony byte content length triggered relayd to hang.  This has
been fixed already.

10 years agoVariables to count seen packets should be unsigned.
tobias [Mon, 18 Aug 2014 20:58:56 +0000 (20:58 +0000)]
Variables to count seen packets should be unsigned.

ok krw@

10 years agosync
deraadt [Mon, 18 Aug 2014 19:41:18 +0000 (19:41 +0000)]
sync

10 years agoreplace more ROTATE macros with plain-old C code.
bcook [Mon, 18 Aug 2014 19:15:34 +0000 (19:15 +0000)]
replace more ROTATE macros with plain-old C code.

Let the compiler optimize these. Even older versions of gcc generate
equal or better quality code than the inline asm.

ok miod@

10 years agoremove return value from HOST_c2l/l2c macros
bcook [Mon, 18 Aug 2014 19:11:48 +0000 (19:11 +0000)]
remove return value from HOST_c2l/l2c macros

These macros and asm inlines simulate a function returning a value, but
nothing ever uses this return value. Remove the pseudo-returns and
(void) casts discarding the unused values.

This, maybe unsurprisingly, speeds things up a bit. It also removes the
GCC 4.9 warnings about unused values.

ok miod@ deraadt@

10 years agoExplicitely need <sys/pool.h> now.
miod [Mon, 18 Aug 2014 17:56:45 +0000 (17:56 +0000)]
Explicitely need <sys/pool.h> now.

10 years agoSigh, ignoring instruction fetch bus errors for the kernel code should not
miod [Mon, 18 Aug 2014 17:23:06 +0000 (17:23 +0000)]
Sigh, ignoring instruction fetch bus errors for the kernel code should not
depend upon the address being at the beginning of a cache line, for we may
arrive in the middle of a line thanks to a branch. Noticed the hard way...

10 years agoWhen the first child of the node being validated gets deleted during
schwarze [Mon, 18 Aug 2014 16:26:13 +0000 (16:26 +0000)]
When the first child of the node being validated gets deleted during
validation, man_node_unlink() switches to MAN_NEXT_CHILD.  After
that, we have to switch back to MAN_NEXT_SIBLING after completing
validation, or subsequent parsing would add content into an already
closed node, clobbering potentially existing children, causing
information loss and a memory leak.  Bug found by kristaps@ with
valgrind in groff(7) on Mac OS X.

Note that the switch back must be conditional, for if the node being
validated itself gets deleted, we must *not* go to MAN_NEXT_SIBLING,
which would not only yield wrong results in general but also crash
in malformed manuals having an empty paragraph before the first .SH,
for example OpenBSD c++filt(1).

While here, add the missing <sys/types.h> as required before mandoc.h.

10 years agopool debug back on
tedu [Mon, 18 Aug 2014 14:35:37 +0000 (14:35 +0000)]
pool debug back on

10 years agoa small tweak to improve malloc in multithreaded programs. we don't need
tedu [Mon, 18 Aug 2014 14:34:58 +0000 (14:34 +0000)]
a small tweak to improve malloc in multithreaded programs. we don't need
to hold the malloc lock across mmap syscalls in all cases. dropping it
allows another thread to access the existing chunk cache if necessary.
could be improved to be a bit more aggressive, but i've been testing this
simple diff for some time now with good results.

10 years agothis uses pools, but relied on mbuf.h to provide them.
dlg [Mon, 18 Aug 2014 13:29:13 +0000 (13:29 +0000)]
this uses pools, but relied on mbuf.h to provide them.

found by benoit@

10 years agokristaps@ found this with valgrind, merge his patch from bsd.lv:
schwarze [Mon, 18 Aug 2014 13:25:54 +0000 (13:25 +0000)]
kristaps@ found this with valgrind, merge his patch from bsd.lv:
Fix a corner case where \H<nil> (where <nil> is the \0 character) would
cause mandoc_escape() to read past the end of an allocated string.
Found when a script scanning of all Mac OSX manuals accidentally also
scanned binary (gzip'd) files, discussed with schwarze@ on tech@mdocml.

10 years agoSync proc.c with httpd. httpd needs SIGUSR1 but snmpd will ignore it
reyk [Mon, 18 Aug 2014 13:13:42 +0000 (13:13 +0000)]
Sync proc.c with httpd.  httpd needs SIGUSR1 but snmpd will ignore it
now instead of terminating the process.

ok blambert@

10 years agoSync proc.c with httpd. httpd needs SIGUSR1 but relayd will ignore it
reyk [Mon, 18 Aug 2014 12:59:00 +0000 (12:59 +0000)]
Sync proc.c with httpd.  httpd needs SIGUSR1 but relayd will ignore it
now instead of terminating the process.

ok florian@

10 years agoSync proc.c with httpd. httpd needs SIGUSR1 but iked will ignore it
reyk [Mon, 18 Aug 2014 09:43:02 +0000 (09:43 +0000)]
Sync proc.c with httpd.  httpd needs SIGUSR1 but iked will ignore it
now instead of terminating the process.

ok mikeb@

10 years agodont rely on mbuf.h to provide pool.h.
dlg [Mon, 18 Aug 2014 05:11:03 +0000 (05:11 +0000)]
dont rely on mbuf.h to provide pool.h.

ok miod@, who has offerred to help with any MD fallout
ok guenther@

10 years agoAdd mention of execvpe() setting errno on failure.
guenther [Mon, 18 Aug 2014 04:48:22 +0000 (04:48 +0000)]
Add mention of execvpe() setting errno on failure.

from Justus Baumgartner (meddev50 (at) gmail.com)

10 years agoits a bit silly to include mbuf.h to get to malloc.h. just include
dlg [Mon, 18 Aug 2014 04:23:11 +0000 (04:23 +0000)]
its a bit silly to include mbuf.h to get to malloc.h. just include
malloc.h instead.

10 years agointroduce mbuf list and queue apis. both manage fifo lists of mbufs
dlg [Mon, 18 Aug 2014 04:06:16 +0000 (04:06 +0000)]
introduce mbuf list and queue apis. both manage fifo lists of mbufs
and a count of the mbufs.

struct mbuf_list and the ml_foo() apis can be used to build lists of
mbufs where you dont need locking (eg, on the stack).

struct mbuf_queue and mq_foo() wrap mbuf_lists with a mutex, and
limits the number of mbufs that can be queued. they can be useful
for moving mbufs between contexts/subsystems.

with help from jmc@ for the manpage bits
mpi@ is keen

10 years agoAdd fancy printing of ktrace()'s ops argument
guenther [Mon, 18 Aug 2014 03:29:53 +0000 (03:29 +0000)]
Add fancy printing of ktrace()'s ops argument
mquery() has the exact same argument layout as mmap(), so share the case
Fix a couple brace placement glitches

10 years agoexternal page headers use an RB tree to find the page header
dlg [Mon, 18 Aug 2014 01:28:44 +0000 (01:28 +0000)]
external page headers use an RB tree to find the page header
containing an item when its returned to the pool. this means you
need to do an inexact comparison between an items address and the
page address, cos a pool page can contain many items.

previously this used RB_FIND with a compare function that would do math
on every node comparison to see if one node (the key) was within the other
node (the tree element).

this cuts it over to using RB_NFIND to find the closest tree node
instead of the exact tree node. the node compares turns into simple
< and > operations, which inline very nicely with the RB_NFIND. the
constraint (an item must be within a page) is then checked only
once after the NFIND call.

feedback from matthew@ and tedu@

10 years agoUse %#o instead of %#x for mode_t
guenther [Sun, 17 Aug 2014 22:43:07 +0000 (22:43 +0000)]
Use %#o instead of %#x for mode_t

10 years agoDisplay symbolicly the mode argument of mkdir, mkfifo, mknod, and umask
guenther [Sun, 17 Aug 2014 22:25:53 +0000 (22:25 +0000)]
Display symbolicly the mode argument of mkdir, mkfifo, mknod, and umask

10 years agotypo, sorry
schwarze [Sun, 17 Aug 2014 22:08:53 +0000 (22:08 +0000)]
typo, sorry

10 years agoWhile all current callers pass valid data to ascii_hspan() only,
schwarze [Sun, 17 Aug 2014 22:06:49 +0000 (22:06 +0000)]
While all current callers pass valid data to ascii_hspan() only,
it's safer to assume incoming enum data might be invalid
and catch it instead of happily returning an unitialized int.
No functional change right now.

10 years agotypo; Steven Honeyman <stevenhoneyman at gmail dot com>
schwarze [Sun, 17 Aug 2014 21:02:06 +0000 (21:02 +0000)]
typo; Steven Honeyman <stevenhoneyman at gmail dot com>

10 years agoSync regress/usr.sbin/relayd with recent http.h changes
doug [Sun, 17 Aug 2014 21:01:24 +0000 (21:01 +0000)]
Sync regress/usr.sbin/relayd with recent http.h changes

10 years agodon;t mark up {};
jmc [Sun, 17 Aug 2014 18:46:29 +0000 (18:46 +0000)]
don;t mark up {};

10 years agoSegfault fixes from kristaps@.
schwarze [Sun, 17 Aug 2014 18:42:07 +0000 (18:42 +0000)]
Segfault fixes from kristaps@.
Note that .It and .Nm blocks without bodies cannot be generated by
valid mdoc(7) syntax but are a adequate representations of invalid
mdoc(7) constructs like Bl -hang It Bo Sh and Sh SYNOPSIS Nm Bo Sh.

10 years agoExecute /etc/netstart using sh(1) instead of sourcing it.
ajacoutot [Sun, 17 Aug 2014 14:43:34 +0000 (14:43 +0000)]
Execute /etc/netstart using sh(1) instead of sourcing it.
Committing early to make sure we have time to fix any side-effect.

ok deraadt@

10 years agoWhen running interactively, display the output of the "check" action
ajacoutot [Sun, 17 Aug 2014 14:41:03 +0000 (14:41 +0000)]
When running interactively, display the output of the "check" action
(ok or failed) like we do with all other actions.

ok jung@ rpe@

10 years agoFix regression tests for ld.so
doug [Sun, 17 Aug 2014 12:30:28 +0000 (12:30 +0000)]
Fix regression tests for ld.so

It now runs correctly with 'make depend regress'

10 years agoOn kernels compiled with R10000 support, ignore (by simply returning)
miod [Sun, 17 Aug 2014 11:11:34 +0000 (11:11 +0000)]
On kernels compiled with R10000 support, ignore (by simply returning)
`bus error upon instruction fetch' exceptions where the faulting address is
in the kernel, and at the very beginning of an I$ cache line.
(I've experienced these on an R16000 Fuel since several months already)

10 years agoi broke the userland shim used for the extent regress test when i
dlg [Sun, 17 Aug 2014 09:48:55 +0000 (09:48 +0000)]
i broke the userland shim used for the extent regress test when i
made it so struct pool was only visible to _KERNEL. tedu broke it
too when he added the size argument to the kernel free
functions.

this fixes both issues. the main change is to provide a local version of
struct pool with just the bit (pr_size) needed for extent to run.

if extents take advantage of more malloc/pool features (eg, {M,PR}_ZERO
then this will need to be updated again.

found by and based on a diff from Theo Buehler
ok mpi@

10 years agosync with hackathons.html
jsg [Sun, 17 Aug 2014 09:00:44 +0000 (09:00 +0000)]
sync with hackathons.html

10 years agoupdate the built-ins list:
jmc [Sun, 17 Aug 2014 07:15:41 +0000 (07:15 +0000)]
update the built-ins list:
- "times" is both promoted to posix and special (ooh!)
- "pwd" is promoted to posix, but not special (aah!)

while here, jiggle the text somewhat to clarify that "non-POSIX"
actually meant when posix mode is off, not mandated by posix

joint work with guenther

10 years agorepair operation of kern.arandom, which will only allow a buffer of
deraadt [Sat, 16 Aug 2014 21:39:16 +0000 (21:39 +0000)]
repair operation of kern.arandom, which will only allow a buffer of
512 bytes.  As a result, it stopped working...
ok miod

10 years agoCover execvpe() in the description of what the envp arg means.
guenther [Sat, 16 Aug 2014 21:09:36 +0000 (21:09 +0000)]
Cover execvpe() in the description of what the envp arg means.

from Justus Baumgartner (meddev50 (at) gmail.com)
ok jmc@

10 years agoIf a stray .It follows .El, we are no longer in the list,
schwarze [Sat, 16 Aug 2014 19:47:17 +0000 (19:47 +0000)]
If a stray .It follows .El, we are no longer in the list,
even though the list is still the last processed macro.
This fixes a regression introduced in mdoc_macro.c rev. 1.95:
Ulrich Spoerlein <uqs at FreeBSD> reports that various of their
kernel manuals trigger assertions.

10 years agoreplace sprintf/strdup with asprintf in engine test
bcook [Sat, 16 Aug 2014 18:47:40 +0000 (18:47 +0000)]
replace sprintf/strdup with asprintf in engine test

10 years agoonly build the getrandom path if SYS_getrandom is defined.
bcook [Sat, 16 Aug 2014 18:42:41 +0000 (18:42 +0000)]
only build the getrandom path if SYS_getrandom is defined.

like the sysctl path

10 years agoFixed typo in "distinct".
tobias [Sat, 16 Aug 2014 17:33:40 +0000 (17:33 +0000)]
Fixed typo in "distinct".

ok jmc@

10 years agouse C rather than C++ array initialization syntax
bcook [Sat, 16 Aug 2014 17:30:12 +0000 (17:30 +0000)]
use C rather than C++ array initialization syntax

this causes errors with stricter C compilers

10 years agogetrandom(2) support for getentropy_linux
bcook [Sat, 16 Aug 2014 17:21:56 +0000 (17:21 +0000)]
getrandom(2) support for getentropy_linux

This enables support for the new getrandom(2) syscall in Linux 3.17.

If the call exists and fails, return a failure in getentropy(2) emulation as
well. This adds a EINTR check in case the urandom pool is not initialized.

Tested on Fedora Rawhide with 3.17rc0 and Ubuntu 14.04

ok deraadt@

10 years agoreduce cutoff for "hours only" start times to reduce window of ambiguity.
tedu [Sat, 16 Aug 2014 16:38:38 +0000 (16:38 +0000)]
reduce cutoff for "hours only" start times to reduce window of ambiguity.
ok deraadt

10 years agoI found a number of interactive events which can cause signals, and go
deraadt [Sat, 16 Aug 2014 07:49:27 +0000 (07:49 +0000)]
I found a number of interactive events which can cause signals, and go
down paths not previously marked as signal handled unsafe.  Try to clean
up a few of them especially regarding errno, mark others as unsafe, and
repair a few by avoiding stdio.  Glanced at by misc people in Slovenia,
but considered too risky before release..

10 years agosync
deraadt [Sat, 16 Aug 2014 07:42:37 +0000 (07:42 +0000)]
sync

10 years agouse void * instead of char *. ok jsing
tedu [Fri, 15 Aug 2014 16:55:32 +0000 (16:55 +0000)]
use void * instead of char *. ok jsing

10 years agoUpdate to tzdata2014f from ftp.iana.org.
millert [Fri, 15 Aug 2014 16:48:58 +0000 (16:48 +0000)]
Update to tzdata2014f from ftp.iana.org.

10 years agoPublic key operations are no longer supported
mikeb [Fri, 15 Aug 2014 15:46:24 +0000 (15:46 +0000)]
Public key operations are no longer supported

10 years agoRemove support for public key operations
mikeb [Fri, 15 Aug 2014 15:43:27 +0000 (15:43 +0000)]
Remove support for public key operations

10 years agoRemove support for public key operations
mikeb [Fri, 15 Aug 2014 15:37:51 +0000 (15:37 +0000)]
Remove support for public key operations

10 years agoConvert 3DES regress test from /dev/crypto to using kernel source code
mikeb [Fri, 15 Aug 2014 15:13:38 +0000 (15:13 +0000)]
Convert 3DES regress test from /dev/crypto to using kernel source code
directly with a simplified CBC implementation.

10 years agoConvert AES-CTR regress test from /dev/crypto to using kernel source
mikeb [Fri, 15 Aug 2014 14:39:04 +0000 (14:39 +0000)]
Convert AES-CTR regress test from /dev/crypto to using kernel source
code directly.  This test is converted the same way jsing@ has recently
converted an XTS test by pulling in xform.c code.

10 years agoConvert AES regress test from /dev/crypto to using kernel source code
mikeb [Fri, 15 Aug 2014 14:36:19 +0000 (14:36 +0000)]
Convert AES regress test from /dev/crypto to using kernel source code
directly.  This test case uses ECB vectors, therefore no chaining is
required.

10 years agolofn(4) and nofn(4) are gone
mikeb [Fri, 15 Aug 2014 14:16:01 +0000 (14:16 +0000)]
lofn(4) and nofn(4) are gone

10 years agoWith deprecation of crypto(4) interface lofn(4) and nofn(4) become
mikeb [Fri, 15 Aug 2014 14:12:07 +0000 (14:12 +0000)]
With deprecation of crypto(4) interface lofn(4) and nofn(4) become
obsolete.  No objections from the usual suspects.

10 years agounlink lofn(4) and nofn(4)
mikeb [Fri, 15 Aug 2014 14:07:39 +0000 (14:07 +0000)]
unlink lofn(4) and nofn(4)

10 years agolofn(4) and nofn(4) are going away
mikeb [Fri, 15 Aug 2014 14:06:47 +0000 (14:06 +0000)]
lofn(4) and nofn(4) are going away

10 years agoAlign install.md files with installer style.
rpe [Fri, 15 Aug 2014 09:45:54 +0000 (09:45 +0000)]
Align install.md files with installer style.

- { foo ; bar ; } -> { foo; bar; }
- if foo ; then -> if foo; then

OK halex@

10 years agoshorten share/doc/{html,mg} to share/doc, since nothing currently
jmc [Fri, 15 Aug 2014 06:44:29 +0000 (06:44 +0000)]
shorten share/doc/{html,mg} to share/doc, since nothing currently
gets installed in html (i think!);

10 years agoadmission of hawaii happened in 1959, not 1984;
jmc [Fri, 15 Aug 2014 06:05:18 +0000 (06:05 +0000)]
admission of hawaii happened in 1959, not 1984;

from the freebsd bugs database, bug 192651:
Reported: 2014-08-14 13:17 UTC by Richard

10 years agoXPG requires insque() and remque() to work with linear lists and not just
guenther [Fri, 15 Aug 2014 04:14:36 +0000 (04:14 +0000)]
XPG requires insque() and remque() to work with linear lists and not just
circular lists.  Amazingly, they managed to extend the requirements to no
longer match the behavior of the VAX instructions they were modeled after,
so the trivial VAX ASM versions have to go.  Nice job breaking it, X/Open!

Based on a diff from enh (at) google.com
ok miod@

10 years agoUse O_CLOEXEC wherever we open a file and then call fcntl(F_SETFD, FD_CLOEXEC)
guenther [Fri, 15 Aug 2014 03:51:40 +0000 (03:51 +0000)]
Use O_CLOEXEC wherever we open a file and then call fcntl(F_SETFD, FD_CLOEXEC)
on it, simplifying error checking, reducing system calls, and improving
thread-safety for libraries.

ok miod@

10 years agoCreate a function which loads sgd in the mfi_iop_ops struct so that skinny
yasuoka [Fri, 15 Aug 2014 02:27:02 +0000 (02:27 +0000)]
Create a function which loads sgd in the mfi_iop_ops struct so that skinny
adapters can use "IEEE sgl".

tested dlg yasuoka
ok dlg jsg

10 years agoSome compilers apparently worry that abort() might return
schwarze [Thu, 14 Aug 2014 22:32:28 +0000 (22:32 +0000)]
Some compilers apparently worry that abort() might return
and then throw a "may be used uninitialized" warning, so
sprinkle some /* NOTREACHED */.  No functional change.
Noticed by Thomas Klausner <wiz at NetBSD dot org>.

10 years agoClarify that .Em and .Sy are physical, not semantic markup,
schwarze [Thu, 14 Aug 2014 20:55:41 +0000 (20:55 +0000)]
Clarify that .Em and .Sy are physical, not semantic markup,
explain appropriate usage, and provide some examples.
ok jmc@

10 years agoregen
millert [Thu, 14 Aug 2014 20:29:31 +0000 (20:29 +0000)]
regen

10 years agoAdd USB hubs found on Intel Z97 chipset boards (one attached to
millert [Thu, 14 Aug 2014 20:29:08 +0000 (20:29 +0000)]
Add USB hubs found on Intel Z97 chipset boards (one attached to
each EHCI root hub).  OK deraadt@ jsg@

10 years agodouble word fix-o;
jmc [Thu, 14 Aug 2014 20:06:41 +0000 (20:06 +0000)]
double word fix-o;

10 years agoBackport from binutils-2.17:
guenther [Thu, 14 Aug 2014 19:43:45 +0000 (19:43 +0000)]
Backport from binutils-2.17:
2005-08-17  Alan Modra  <amodra@bigpond.net.au>
* elfxx-mips.c (_bfd_mips_elf_hide_symbol): Cope with being called
without any got section.

ok miod@

10 years agoremove two duplicate command entries;
schwarze [Thu, 14 Aug 2014 19:08:21 +0000 (19:08 +0000)]
remove two duplicate command entries;
reported by Denis Fondras <openbsd at ledeuns dot net>

10 years agofixed overrid(d)en typo
tobias [Thu, 14 Aug 2014 17:55:28 +0000 (17:55 +0000)]
fixed overrid(d)en typo

millert@ and jmc@ agree that "overriden" is wrong

10 years agoUnused variable, and more fixes in the attach failure path.
miod [Thu, 14 Aug 2014 17:49:50 +0000 (17:49 +0000)]
Unused variable, and more fixes in the attach failure path.

10 years agoBring back 1.173 (reverting 1.174) - mips64 systems are still unhappy when
miod [Thu, 14 Aug 2014 17:21:38 +0000 (17:21 +0000)]
Bring back 1.173 (reverting 1.174) - mips64 systems are still unhappy when
the hint returned is over VM_MAXUSER_ADDRESS, apparently; better be safe for
now while this is investigated further.

10 years agoNo need for <netinet/in_systm.h> nor <netinet/tcpip.h>.
mpi [Thu, 14 Aug 2014 12:55:50 +0000 (12:55 +0000)]
No need for <netinet/in_systm.h> nor <netinet/tcpip.h>.

10 years agoNo use for <netinet/in_systm.h> nor <netinet/tcpip.h>.
mpi [Thu, 14 Aug 2014 12:44:44 +0000 (12:44 +0000)]
No use for <netinet/in_systm.h> nor <netinet/tcpip.h>.

10 years agoKill bandwidth monitoring meters.
mpi [Thu, 14 Aug 2014 12:34:04 +0000 (12:34 +0000)]
Kill bandwidth monitoring meters.

10 years agoAdd bounce matching for [] and {} like mg already does with () and like
bcallah [Thu, 14 Aug 2014 12:22:58 +0000 (12:22 +0000)]
Add bounce matching for [] and {} like mg already does with () and like
Emacs does.
ok florian@

10 years agoCompare number of allocated clusters with a low watermark, not
mikeb [Thu, 14 Aug 2014 11:38:14 +0000 (11:38 +0000)]
Compare number of allocated clusters with a low watermark, not
a magic number 4 since sometimes we can't fit a single packet
(jumbo frame) into 4 clusters.

OK dlg

10 years agoSync with httpd, including the following change from doug@:
reyk [Thu, 14 Aug 2014 10:30:52 +0000 (10:30 +0000)]
Sync with httpd, including the following change from doug@:
"Sync with RFC 7230-7235 phrases and IANA registered status codes.
ok reyk@"

10 years agoImplement rxrinfo ioctl for cluster usage statistics
mikeb [Thu, 14 Aug 2014 09:52:03 +0000 (09:52 +0000)]
Implement rxrinfo ioctl for cluster usage statistics

10 years agoSync with RFC 7230-7235 phrases and IANA registered status codes.
doug [Thu, 14 Aug 2014 09:12:26 +0000 (09:12 +0000)]
Sync with RFC 7230-7235 phrases and IANA registered status codes.

ok reyk@

10 years agoNo need for raw_cb.h
mpi [Thu, 14 Aug 2014 09:01:47 +0000 (09:01 +0000)]
No need for raw_cb.h

10 years agoKill MRT_{ADD,DEL}_BW_UPCALL interfaces and the bandwidth monitoring
mpi [Thu, 14 Aug 2014 08:22:38 +0000 (08:22 +0000)]
Kill MRT_{ADD,DEL}_BW_UPCALL interfaces and the bandwidth monitoring
code that comes with them.

ok mikeb@, henning@