dlg [Tue, 27 Oct 2015 13:58:45 +0000 (13:58 +0000)]
fix fallout from chacha changes.
nicm [Tue, 27 Oct 2015 13:23:24 +0000 (13:23 +0000)]
Break the common process set up, event loop and imsg dispatch code
between server and client out into a separate internal API. This will
make it easier to add another process.
renato [Tue, 27 Oct 2015 13:14:44 +0000 (13:14 +0000)]
Remove wrong line.
mpi [Tue, 27 Oct 2015 12:48:14 +0000 (12:48 +0000)]
Check that RTM_GET works without RTAX_GATEWAY argument for multipath
route entries.
benno [Tue, 27 Oct 2015 12:27:54 +0000 (12:27 +0000)]
change cipher-server-preference to be on by default. It can be disabled with
no cipher-server-preference
this makes more clients select ciphers with pfs.
requested and ok by reyk@
mpi [Tue, 27 Oct 2015 12:06:37 +0000 (12:06 +0000)]
Rewrite in_ouraddr() to not use ``rt_ifa'' since it is not obvious that
the lifetime of an ``ifa'' is tied to a route entry, so it might no
longer be valid after calling rtfree(9).
While here put a KERNEL_LOCK() around the per-ifp address list iteration.
ok bluhm@
mikeb [Tue, 27 Oct 2015 12:00:25 +0000 (12:00 +0000)]
Use verbose defines instead of hardcoded values for clarity when
initializing hash objects. No binary or functional change.
jca [Tue, 27 Oct 2015 11:47:17 +0000 (11:47 +0000)]
Consistency in error messages.
No functional change.
mikeb [Tue, 27 Oct 2015 11:13:06 +0000 (11:13 +0000)]
Sync chacha_ivsetup to the version in ssh so that we could
specify custom counter value when setting up Chacha context.
ok reyk djm
mpi [Tue, 27 Oct 2015 10:54:52 +0000 (10:54 +0000)]
Move code around for clarity, no functional change.
ok bluhm@
mpi [Tue, 27 Oct 2015 10:52:17 +0000 (10:52 +0000)]
Use rt_ifidx rather than rt_ifp.
ok bluhm@
nicm [Tue, 27 Oct 2015 09:28:31 +0000 (09:28 +0000)]
Count brackets in #{?...} so that nested conditional formats work, from
Daniel De Graaf.
nicm [Tue, 27 Oct 2015 09:18:06 +0000 (09:18 +0000)]
The format callback may not always succeed, so we need to check for
NULL. From Patrick Palka.
nicm [Tue, 27 Oct 2015 09:15:21 +0000 (09:15 +0000)]
Use copy-mode -et= in WheelUpPane binding, from Patrick Palka.
djm [Tue, 27 Oct 2015 08:54:52 +0000 (08:54 +0000)]
fix execv arguments in a way less likely to cause grief for -portable;
ok dtucker@
doug [Tue, 27 Oct 2015 07:58:00 +0000 (07:58 +0000)]
Fix nsd and unbound's error checking with SSL_CTX_set_options().
LibreSSL, BoringSSL and OpenSSL have all been redefining SSL_OP_* flags to
be 0 when support is removed. This breaks the error checking that nsd and
unbound were doing. It should check against the flag value itself to allow
for 0.
input deraadt@, jsing@
tweak + ok miod@, guenther@
jmc [Tue, 27 Oct 2015 07:33:42 +0000 (07:33 +0000)]
tweak previous;
jmc [Tue, 27 Oct 2015 07:26:43 +0000 (07:26 +0000)]
tweak previous;
guenther [Tue, 27 Oct 2015 05:54:49 +0000 (05:54 +0000)]
Wrap the remaining math functions in libc: __fpclassify*(), __flt_rounds(),
and ldexp().
ok millert@
yasuoka [Tue, 27 Oct 2015 04:48:06 +0000 (04:48 +0000)]
Free the received radius packet when it is duplicated.
diff from Yuuichi Someya
guenther [Tue, 27 Oct 2015 04:41:18 +0000 (04:41 +0000)]
The fd of the passwd database is marked close-on-exec now
Incorrect text pointed out by tedu@
ok deraadt@ millert@ tedu@
yasuoka [Tue, 27 Oct 2015 04:30:44 +0000 (04:30 +0000)]
Initialize module_radius_req before use it.
diff from Yuuichi Someya
yasuoka [Tue, 27 Oct 2015 04:27:01 +0000 (04:27 +0000)]
Set O_NONBLOCK for UDP sockets not to block on recv(). Actually
block had happened if an error of the socket is handled by send().
diff from Yuuichi Someya.
yasuoka [Tue, 27 Oct 2015 04:18:36 +0000 (04:18 +0000)]
Fix radiusd module to set O_NONBLOCK properly.
diff from Yuuichi Someya.
renato [Tue, 27 Oct 2015 03:33:56 +0000 (03:33 +0000)]
Add a man page for eigrpd.conf.
renato [Tue, 27 Oct 2015 03:30:23 +0000 (03:30 +0000)]
Minor tweaks.
renato [Tue, 27 Oct 2015 03:27:35 +0000 (03:27 +0000)]
Whenever a summary route is activated, install a respective blackhole
route in the FIB.
renato [Tue, 27 Oct 2015 03:25:55 +0000 (03:25 +0000)]
Introduce a new flag to identify "self" neighbors attached to local interfaces.
renato [Tue, 27 Oct 2015 03:23:30 +0000 (03:23 +0000)]
Print a missing "metric" before the actual metric when printing a
redistribute line.
djm [Tue, 27 Oct 2015 01:44:45 +0000 (01:44 +0000)]
log certificate serial in verbose() messages to match the main
auth success/fail message; ok dtucker@
djm [Tue, 27 Oct 2015 00:49:53 +0000 (00:49 +0000)]
avoid de-const warning & shrink; ok dtucker@
nicm [Mon, 26 Oct 2015 23:16:18 +0000 (23:16 +0000)]
If a mouse event has no key binding, pass it through to the pane it
happened in, not the active pane like normal key presses. Fixes problems
seen by Enrico Ghirardi.
nicm [Mon, 26 Oct 2015 23:06:18 +0000 (23:06 +0000)]
Default bindings for mouse wheel on status line to change window (like
we had before), from Patrick Palka.
jca [Mon, 26 Oct 2015 22:24:44 +0000 (22:24 +0000)]
Tidy up error() levels: merge COMPILE and COMPILE2, and kill ERROR (unused)
ok tobias@
jca [Mon, 26 Oct 2015 22:22:56 +0000 (22:22 +0000)]
Use error(FATAL) instead of what was previously err(1,...)
ok tobias@
naddy [Mon, 26 Oct 2015 22:03:06 +0000 (22:03 +0000)]
miscellaneous cruft:
* remove lint
* unifdef S_IFLNK, S_ISFIFO, S_ISLNK, S_ISSOCK, O_APPEND
* remove 0, NOTUSED, notdef -- untouched since original import
* remove IIASA
nicm [Mon, 26 Oct 2015 22:03:04 +0000 (22:03 +0000)]
Handle unknown keys more gracefully, return a string instead of NULL.
naddy [Mon, 26 Oct 2015 21:57:42 +0000 (21:57 +0000)]
remove the profiling and debugging ifdefs; ok zhuk@
naddy [Mon, 26 Oct 2015 20:44:00 +0000 (20:44 +0000)]
document the -p and -l file enquiries and explicitly do not document -s
with/ok schwarze@
rpe [Mon, 26 Oct 2015 19:24:04 +0000 (19:24 +0000)]
The hostname variable is not used since r1.99. Remove it and use
stripcom() output directly with the hostname command.
OK deraadt@ krw@
semarie [Mon, 26 Oct 2015 18:52:19 +0000 (18:52 +0000)]
update regress suite for pledge
jasper [Mon, 26 Oct 2015 18:09:32 +0000 (18:09 +0000)]
add tests for whitespace surrounding properties
jasper [Mon, 26 Oct 2015 18:08:44 +0000 (18:08 +0000)]
Don't reject properties with no whitespace after the colon; it's perfectly
valid to have a line like 'Requires:requiree'.
Confirmed with fd.o and this fixes using the libczmq pc file
jmc [Mon, 26 Oct 2015 18:05:25 +0000 (18:05 +0000)]
tweak previous;
jmc [Mon, 26 Oct 2015 17:56:20 +0000 (17:56 +0000)]
no more TMPDIR;
zhuk [Mon, 26 Oct 2015 17:55:32 +0000 (17:55 +0000)]
Put repquota in line with other utilities: you either wants list everyting,
effectively using -a, or you specify filesystems one-by-one. Doing both
resulted in a funny error messages, so just don't allow that silly behaviour.
okay millert@
deraadt [Mon, 26 Oct 2015 17:52:19 +0000 (17:52 +0000)]
Add ppoll() to "stdio"
mikeb [Mon, 26 Oct 2015 17:22:50 +0000 (17:22 +0000)]
Use axf's hashsize as a block size in the authenticated encryption routine.
No change for GCM, however upcoming changes will rely on this.
nicm [Mon, 26 Oct 2015 17:17:06 +0000 (17:17 +0000)]
Some extra logging of where keys are actually going.
deraadt [Mon, 26 Oct 2015 16:57:13 +0000 (16:57 +0000)]
Rather than setpassent(1) for the lifetime of the program, use two smaller
windows of setpassent(1) + entpwent()
ok tedu millert
sunil [Mon, 26 Oct 2015 16:38:06 +0000 (16:38 +0000)]
Fix makemap(8) for values containing a '#'.
Ok millert@ gilles@
krw [Mon, 26 Oct 2015 16:32:33 +0000 (16:32 +0000)]
Give dhclient(8) the ability to use option 119, a.k.a. "Domain
Search" if supplied by the server.
Requested by a few. Original diff from Ray Lai via tech@.
Tested & ok claudio@
naddy [Mon, 26 Oct 2015 16:31:09 +0000 (16:31 +0000)]
Remove the #define Strfoo() s_strfoo() redirection and move everything
directly to Strfoo(). Purely mechanical. ok deraadt@
naddy [Mon, 26 Oct 2015 16:27:04 +0000 (16:27 +0000)]
Clean up cruft made visible by the NLS unifdefing:
* remove setlocale() calls
* remove write-only variable AsciiOnly
* remove now unused string constants STRLANG, STRLC_CTYPE
* remove hardcoded support for ISO8859-1
ok millert@
millert [Mon, 26 Oct 2015 15:50:06 +0000 (15:50 +0000)]
Remove TMPDIR support, it is not really useful in crontab.
OK deraadt@
mpi [Mon, 26 Oct 2015 15:49:13 +0000 (15:49 +0000)]
Use rt_ifidx rather than rt_ifp.
ok bluhm@
zhuk [Mon, 26 Oct 2015 15:29:17 +0000 (15:29 +0000)]
Put off some "notyet" dust from login.c.
okay millert@ and deraadt@
millert [Mon, 26 Oct 2015 15:16:30 +0000 (15:16 +0000)]
Use SOCK_NONBLOCK and SOCK_CLOEXEC instead of fcntl() calls.
OK guenther@
krw [Mon, 26 Oct 2015 15:08:26 +0000 (15:08 +0000)]
Add GPT editing. Based on GSOC 2014 work by Markus Muller.
tedu [Mon, 26 Oct 2015 15:04:51 +0000 (15:04 +0000)]
don't keep the passwd database open; callers do not expect that they will
become responsible for calling endpwent.
ok deraadt
naddy [Mon, 26 Oct 2015 15:01:15 +0000 (15:01 +0000)]
unifdef BUILTIN (unused), NLS (always set), SHORT_STRINGS (always set)
and remove a few unused defines; no binary changes; ok tedu@ zhuk@
mmcc [Mon, 26 Oct 2015 14:48:54 +0000 (14:48 +0000)]
(char *)0 -> NULL
ok tedu@
krw [Mon, 26 Oct 2015 14:37:50 +0000 (14:37 +0000)]
Flesh out GPT info and tweak other verbiage to be less MBR specific.
millert [Mon, 26 Oct 2015 14:27:41 +0000 (14:27 +0000)]
Remove some useless defines and the set_cron_uid() function.
OK jung@
bluhm [Mon, 26 Oct 2015 14:12:13 +0000 (14:12 +0000)]
Let SLIST_REMOVE invalidate the pointer in more cases.
OK jsing@
mmcc [Mon, 26 Oct 2015 14:08:47 +0000 (14:08 +0000)]
Rename err() to error() to prevent confusion with the stdlib function.
Discussed with jca@.
tedu [Mon, 26 Oct 2015 13:56:18 +0000 (13:56 +0000)]
remove the silly username cache here and use libc user_from_uid.
libc isn't (yet) any better, but it can improve with time.
deraadt [Mon, 26 Oct 2015 13:12:10 +0000 (13:12 +0000)]
remove TMPDIR support
deraadt [Mon, 26 Oct 2015 13:11:42 +0000 (13:11 +0000)]
sync
deraadt [Mon, 26 Oct 2015 13:07:13 +0000 (13:07 +0000)]
rip6query goes to the attic. Hopefully route6d (rip6d?) will get some
love instead, prod prod.
ok claudio jca mpi
tedu [Mon, 26 Oct 2015 12:52:27 +0000 (12:52 +0000)]
remove TMPDIR support. ok millert
tedu [Mon, 26 Oct 2015 12:44:22 +0000 (12:44 +0000)]
print uptime too. ok deraadt
tedu [Mon, 26 Oct 2015 12:24:48 +0000 (12:24 +0000)]
one second amnesty for timeouts so we don't spin with short timeouts
tedu [Mon, 26 Oct 2015 12:23:40 +0000 (12:23 +0000)]
set listening sockets to nonblocking, then handle as many incoming requests
as we can in the loop instead of keventing per request.
tedu [Mon, 26 Oct 2015 12:17:03 +0000 (12:17 +0000)]
dns check needs to be done on the kernel address after copyin
deraadt [Mon, 26 Oct 2015 12:13:45 +0000 (12:13 +0000)]
Allow NET_RT_IFLIST in pledge "dns" as well
claudio [Mon, 26 Oct 2015 11:46:25 +0000 (11:46 +0000)]
Kernel got fixed we no longer need to check if rtm_pid is 0 to filter out
route messages that are not from us.
deraadt [Mon, 26 Oct 2015 11:17:52 +0000 (11:17 +0000)]
If the system call is entirely unpermitted, code will be 0, and there is
no pledge to recommend.
deraadt [Mon, 26 Oct 2015 11:11:45 +0000 (11:11 +0000)]
add setreuid/setregid to "id"
deraadt [Mon, 26 Oct 2015 11:10:53 +0000 (11:10 +0000)]
stop saving a pidfile; ok jung zhuk
jung [Mon, 26 Oct 2015 11:03:29 +0000 (11:03 +0000)]
fix PATH_INFO for / requests
diff from Denis Fondras
ok reyk
bluhm [Mon, 26 Oct 2015 10:43:42 +0000 (10:43 +0000)]
Nothing in base uses native language support anymore. Remove NLS
support from the makefiles.
OK naddy@
tedu [Mon, 26 Oct 2015 10:34:18 +0000 (10:34 +0000)]
remove libutil from dpadd too
tedu [Mon, 26 Oct 2015 10:32:56 +0000 (10:32 +0000)]
remove libutil from dpadd too
deraadt [Mon, 26 Oct 2015 10:12:46 +0000 (10:12 +0000)]
last consumer in base of pidfile()...
jung [Mon, 26 Oct 2015 10:08:14 +0000 (10:08 +0000)]
remove util.h include and -lutil after pidfile removal
ok deraadt
deraadt [Mon, 26 Oct 2015 09:58:18 +0000 (09:58 +0000)]
Stop writing a pidfile
ok benno jung
deraadt [Mon, 26 Oct 2015 09:57:55 +0000 (09:57 +0000)]
stop using pidfile()
ok benno jung jca
jung [Mon, 26 Oct 2015 09:56:01 +0000 (09:56 +0000)]
remove unneeded pidfile() from smtpd
ok gilles
jsg [Mon, 26 Oct 2015 09:34:50 +0000 (09:34 +0000)]
Sync "ifconfig create" interface list with ifconfig -C output.
Adds pair, svlan, tap and vxlan.
jsg [Mon, 26 Oct 2015 09:22:03 +0000 (09:22 +0000)]
avoid a potential double free
ok gilles@
vgross [Mon, 26 Oct 2015 08:06:51 +0000 (08:06 +0000)]
Add autoport to netinet6 regress tests
vgross [Mon, 26 Oct 2015 08:04:10 +0000 (08:04 +0000)]
Add regress test on local route insertions
semarie [Mon, 26 Oct 2015 07:44:43 +0000 (07:44 +0000)]
change some pledge_fail() error/code
- for PLEDGE_FATTR: the code is more accurate to be PLEDGE_FATTR, has
p->p_pledgenote could have multiple bits inside, and it is the lake of
PLEDGE_FATTR that make the check fail.
- for PLEDGE_RECVFD and PLEDGE_SENDFD, change the error to be EINVAL: we fail,
not because of lake of PLEDGE_RECVFD / PLEDGE_SENDFD permission, but because
the value passed is invalid in pledged program (for example trying to send a
VDIR).
"go ahead" deraadt@
semarie [Mon, 26 Oct 2015 07:24:20 +0000 (07:24 +0000)]
make pledge_check(), used for syscall check with pledge, returns an error and
provide the required pledge request for pledge_fail().
ok deraadt@
yasuoka [Mon, 26 Oct 2015 06:44:40 +0000 (06:44 +0000)]
Using realm for bsdauth is misleading. It is currently unusable.
semarie [Mon, 26 Oct 2015 05:28:37 +0000 (05:28 +0000)]
temporary unhook regress/sys/kern/pledge from regress
dtucker [Mon, 26 Oct 2015 02:50:58 +0000 (02:50 +0000)]
Fix typo certopt->certopts in shell variable. This would cause the test to
hang at a host key prompt if you have an A or CNAME for "proxy" in your local
domain.
mmcc [Mon, 26 Oct 2015 02:33:07 +0000 (02:33 +0000)]
Cast isdigit()'s argument to unsigned char.
ok guenther@
schwarze [Mon, 26 Oct 2015 00:49:34 +0000 (00:49 +0000)]
adjust macro usage to the usual conventions