openbsd
7 years agoThe EQN_LISTONE box type is pointless.
schwarze [Wed, 5 Jul 2017 15:03:20 +0000 (15:03 +0000)]
The EQN_LISTONE box type is pointless.
Simplify by just using EQN_LIST with expectargs = 1.
Noticed while investigating a bug report from bentley@.
No functional change.

7 years agoEnable NFSCLIENT to let installation over NFS work.
visa [Wed, 5 Jul 2017 14:58:59 +0000 (14:58 +0000)]
Enable NFSCLIENT to let installation over NFS work.

OK kettenis@, deraadt@

7 years agoFix RAMDISK build.
visa [Wed, 5 Jul 2017 14:47:58 +0000 (14:47 +0000)]
Fix RAMDISK build.

OK bluhm@

7 years agomake use of (f)lex -o option to create unique temporary files, so that
espie [Wed, 5 Jul 2017 13:31:40 +0000 (13:31 +0000)]
make use of (f)lex -o option to create unique temporary files, so that
make -j will be happier.

okay millert@

7 years agobased on florian@'s observation and guenther@'s work in kernel makefiles.
espie [Wed, 5 Jul 2017 13:30:01 +0000 (13:30 +0000)]
based on florian@'s observation and guenther@'s work in kernel makefiles.
don't include .d files during obj and cleanup, because those files might
be utterly bogus following an untimely reboot.

This allows cleaning stuff up without needing to manually remove those
files.

okay millert@, kettenis@

7 years agoregen
kettenis [Wed, 5 Jul 2017 12:43:32 +0000 (12:43 +0000)]
regen

7 years agoAdd Intel Braswell Sensor Hub device.
kettenis [Wed, 5 Jul 2017 12:43:10 +0000 (12:43 +0000)]
Add Intel Braswell Sensor Hub device.

7 years agonits about trailing punctuation found with mandoc -Tlint
schwarze [Wed, 5 Jul 2017 12:23:46 +0000 (12:23 +0000)]
nits about trailing punctuation found with mandoc -Tlint

7 years agovoid functions don't return 0
tb [Wed, 5 Jul 2017 11:44:35 +0000 (11:44 +0000)]
void functions don't return 0

From Klemens Nanni

7 years agofix cross references to self; found with mandoc -Tlint
schwarze [Wed, 5 Jul 2017 11:43:09 +0000 (11:43 +0000)]
fix cross references to self; found with mandoc -Tlint

7 years agoConvert pf tagname malloc(9) into pool_get(9) to make it MP safe.
bluhm [Wed, 5 Jul 2017 11:40:17 +0000 (11:40 +0000)]
Convert pf tagname malloc(9) into pool_get(9) to make it MP safe.
While there use TAILQ_FOREACH macro for traversing tags.
OK mpi@

7 years agoThe IP in IP input function strips the outer header and reinserts
bluhm [Wed, 5 Jul 2017 11:34:10 +0000 (11:34 +0000)]
The IP in IP input function strips the outer header and reinserts
the inner IP packet into the internet queue.  The IPv6 local delivery
code has a loop to deal with header chains.  The idea is to use
this loop and avoid the queueing and rescheduling.  The IPsec packet
will be processed in a single flow.
Merge the IP deliver loop from both IP versions into a single
ip_deliver() function that can handle both addresss families.  This
allows to process an IP in IP header like a normal extension header.
If af != AF_UNSPEC, we are already in a deliver loop and have the
kernel look.  Then we can just return the next protocol.  Otherwise
we enqueue.  The dequeue thread has the kernel lock and starts an
IP delivery loop.
OK mpi@

7 years agoDon't stop logging to stderr when running in foreground with -d.
reyk [Wed, 5 Jul 2017 11:11:56 +0000 (11:11 +0000)]
Don't stop logging to stderr when running in foreground with -d.

Pointed out by Kapetanakis Giannis

7 years agoComments are lying.
mpi [Wed, 5 Jul 2017 10:48:41 +0000 (10:48 +0000)]
Comments are lying.

7 years agodocument that we're no longer using 'make depend'
espie [Wed, 5 Jul 2017 10:41:01 +0000 (10:41 +0000)]
document that we're no longer using 'make depend'

7 years agobye bye depends
espie [Wed, 5 Jul 2017 10:22:32 +0000 (10:22 +0000)]
bye bye depends
okay tb@ deraadt@

7 years agoIf we are sending a neighbor solicitation for a link local address
florian [Wed, 5 Jul 2017 09:51:37 +0000 (09:51 +0000)]
If we are sending a neighbor solicitation for a link local address
send it with a link local source address as well.

This helps upstream routers with their own source address
selection.

A reoccurring scenario is:
- gateway on fe80::1%if
- the gateway does not have an IP in the same prefix as our global address

When we want to talk to the outside world we first need to resolve the
gateway. We copy the source address from our outgoing packet to the
neighbor solicitation packet (a global address) and ask for layer2
information of a link local address.

The upstream router now needs to do source address selection of it's
own. Since we are coming from a global address and there is no address
from the same prefix the router uses another global address lying
around.

We then drop this with "ND packet from non-neighbor".

Reported over the years by a few people, most recently by Marc Peters
on bugs@ who confirmed that this fixes the problem.

OK stsp@, mpi@

7 years agoSome documentation improvements:
mpi [Wed, 5 Jul 2017 09:40:16 +0000 (09:40 +0000)]
Some documentation improvements:

- Fix TLS s/server/client/

- Use 'remote loghost' consistently, even if it's not clear to which
  endpoint this correspond.

- Replace 'forwarding' by 'sending' to remove the ambiguity about the
  inserted hostname.

- Do not use the word 'server' with 'socket' to avoid confusion with
  a TLS server.

- Prefer 'senders' than 'clients' when it comes to spoofing, to reduce
  one usage of the word 'client.

ok jmc@, bluhm@

7 years agoremove useless (void)printf casts, diff from Klemens Nanni, massaged
florian [Wed, 5 Jul 2017 07:15:40 +0000 (07:15 +0000)]
remove useless (void)printf casts, diff from Klemens Nanni, massaged
by me.

7 years agoAdd tests for file completion in ksh emacs mode, currently failing.
anton [Wed, 5 Jul 2017 06:31:59 +0000 (06:31 +0000)]
Add tests for file completion in ksh emacs mode, currently failing.

While here, pass the v option to hexdump in order to output all data.

7 years agoSwitch to build user if run as root. Prompt output $ or # affects test.
bluhm [Tue, 4 Jul 2017 23:54:52 +0000 (23:54 +0000)]
Switch to build user if run as root.  Prompt output $ or # affects test.

7 years agocross reference to self; found with mandoc(1)
schwarze [Tue, 4 Jul 2017 23:38:49 +0000 (23:38 +0000)]
cross reference to self; found with mandoc(1)

7 years agoRevert back previous, pledge cannot be enabled on the privsep'd proc yet, at
mestre [Tue, 4 Jul 2017 23:13:09 +0000 (23:13 +0000)]
Revert back previous, pledge cannot be enabled on the privsep'd proc yet, at
least not as is

Reported by tim@, OK deraadt@ to backout the pledge for now

7 years agoFix file regress from stdin if obj directory exists.
bluhm [Tue, 4 Jul 2017 23:12:01 +0000 (23:12 +0000)]
Fix file regress from stdin if obj directory exists.

7 years agoFix the only remaining mandoc(1) ERROR in the base system.
schwarze [Tue, 4 Jul 2017 23:11:26 +0000 (23:11 +0000)]
Fix the only remaining mandoc(1) ERROR in the base system.
(Can't resist the temptation to commit 'cause it allows
such a cute commit message.)

7 years agoproper escaping to prevent information loss; found with mandoc -Tlint
schwarze [Tue, 4 Jul 2017 23:07:40 +0000 (23:07 +0000)]
proper escaping to prevent information loss; found with mandoc -Tlint

7 years agoFix handling of \} on roff request lines.
schwarze [Tue, 4 Jul 2017 22:49:59 +0000 (22:49 +0000)]
Fix handling of \} on roff request lines.
Cures bogus error messages in pages generated with pod2man(1).

7 years agoIgnore entries with specific PCI subvendor/subdevice such that we don't
kettenis [Tue, 4 Jul 2017 22:27:23 +0000 (22:27 +0000)]
Ignore entries with specific PCI subvendor/subdevice such that we don't
inadvertedly match the Intel HD Graphics P4000 as a "Quanta transcode"
device.

Thanks to Joe Gidi for figuring out that I inadvertedly brought this back.
Hopefully fixing it this way prevents it from happening again.

7 years ago< and > keys to scroll preview list left and right in tree mode.
nicm [Tue, 4 Jul 2017 22:21:31 +0000 (22:21 +0000)]
< and > keys to scroll preview list left and right in tree mode.

7 years agoFix building the libedit regress programs.
bluhm [Tue, 4 Jul 2017 21:56:05 +0000 (21:56 +0000)]
Fix building the libedit regress programs.

7 years agodelete tokens that have never been used.
benno [Tue, 4 Jul 2017 21:13:03 +0000 (21:13 +0000)]
delete tokens that have never been used.
From Rob Pierce
ok yacc and gcc

7 years agorename fetch_state() to fetch_ifstate(). The word "state" is a bit
benno [Tue, 4 Jul 2017 21:09:52 +0000 (21:09 +0000)]
rename fetch_state() to fetch_ifstate(). The word "state" is a bit
overloaded in this daemon.
From Rob Pierce

7 years agoFix some variable alignment whitespace.
benno [Tue, 4 Jul 2017 21:04:14 +0000 (21:04 +0000)]
Fix some variable alignment whitespace.
From Rob Pierce

7 years agoIt turns out association of tbl spans with layout rows is simpler than
schwarze [Tue, 4 Jul 2017 20:59:17 +0000 (20:59 +0000)]
It turns out association of tbl spans with layout rows is simpler than
i thought.  Fixing a bug in curs_addch(3) and minus 25 lines of code.

7 years agoif configured, set the ttl (IPV6_UNICAST_HOPS) for ipv6 tcp checks.
benno [Tue, 4 Jul 2017 20:27:09 +0000 (20:27 +0000)]
if configured, set the ttl (IPV6_UNICAST_HOPS) for ipv6 tcp checks.
From Kapetanakis Giannis, thanks.
ok florian@

7 years agoDetect upfront whether we have a particular set installed and if not, do *not*
ajacoutot [Tue, 4 Jul 2017 20:25:53 +0000 (20:25 +0000)]
Detect upfront whether we have a particular set installed and if not, do *not*
list a matching syspatch as available. This will allow to skip the syspatches
for the X sets for example if we don't have them installed. If we do install
these sets a posteriori, then syspatch *will* list the X syspatches if there are
any.

req. by many
ok robert@

7 years agomake relayd not crash in relay_udp_server() when using a dns relay.
benno [Tue, 4 Jul 2017 19:59:51 +0000 (19:59 +0000)]
make relayd not crash in relay_udp_server() when using a dns relay.
needs revisiting. From Rivo Nurges, thanks.
ok florian@

7 years agoSwitch reorder_kernel() from sending emails to logging via syslog
rpe [Tue, 4 Jul 2017 19:02:11 +0000 (19:02 +0000)]
Switch reorder_kernel() from sending emails to logging via syslog
and to write the logfile inside the kernel compile dir.

- turn the whole reorder_kernel function into a subshell {} -> ()
- create kernel compile dir early on
- redirect all stdout/stderr to a logfile inside this dir
- setup ERR trap handler that
  - disables the EXIT trap handler
  - syslogs the error and hints to the logfile
  - additionally sends this message to the console
- setup EXIT trap handler that syslogs success
- wipe only the content instead of the whole kernel compile dir
- reestablish stdout redirection to the log after the wipe
- remove -q option of sha256 to log check result
- run reorder_kernel() in the background

OK deraadt@ tb@

7 years agoDo not add isakmpd obj directory to .PATH. We must compile some
bluhm [Tue, 4 Jul 2017 18:02:21 +0000 (18:02 +0000)]
Do not add isakmpd obj directory to .PATH.  We must compile some
fake objects from regress ourselves.

7 years agosome of this code was written in an era when spaces cost extra.
tedu [Tue, 4 Jul 2017 17:29:51 +0000 (17:29 +0000)]
some of this code was written in an era when spaces cost extra.
add a little breathing room.

7 years agocalculate multiples of 4 smarter; from Klemens Nanni
florian [Tue, 4 Jul 2017 15:55:22 +0000 (15:55 +0000)]
calculate multiples of 4 smarter; from Klemens Nanni

7 years agoIt does not make sense to use ioctl(FIONREAD) with TLS as libtls
bluhm [Tue, 4 Jul 2017 15:52:26 +0000 (15:52 +0000)]
It does not make sense to use ioctl(FIONREAD) with TLS as libtls
has already read the buffer from kernel to user land.  I have blindly
copied this code from libevent for syslogd(8) TLS, remove it together
with the bug.  It caused hangs in ldapd(8).
report, analysis, testing, OK Seiya Kawashima, Robert Klein, gsoares@

7 years agoreplace perror(3) with err(3); from Klemens Nanni
florian [Tue, 4 Jul 2017 15:43:38 +0000 (15:43 +0000)]
replace perror(3) with err(3); from Klemens Nanni

7 years ago1. mild deprecation notice
schwarze [Tue, 4 Jul 2017 15:33:03 +0000 (15:33 +0000)]
1. mild deprecation notice
2. point to getline (suggested by nicm@)
3. cross reference fgetc(3) rather than putc(3)
4. add missing error handling to the example code

OK nicm@

7 years agozap .o, put both .c at deraadt@'s request
espie [Tue, 4 Jul 2017 14:54:15 +0000 (14:54 +0000)]
zap .o, put both .c at deraadt@'s request

7 years agoMessages of the -Wbase level now print STYLE:. Since this
schwarze [Tue, 4 Jul 2017 14:53:23 +0000 (14:53 +0000)]
Messages of the -Wbase level now print STYLE:.  Since this
causes horrible churn anyway, profit of the opportunity to stop
excessive testing, such that this is hopefully the last instance
of such churn.  Consistently use OpenBSD RCS tags, blank .Os,
blank fourth .TH argument, and Mdocdate like everywhere else.
Use -Ios=OpenBSD for platform-independent predictable output.

7 years agoPrinting "BASE:" in messages about violations of base system conventions
schwarze [Tue, 4 Jul 2017 14:40:30 +0000 (14:40 +0000)]
Printing "BASE:" in messages about violations of base system conventions
is confusing, simply print "STYLE:", which is intuitive and does not
sound excessively alarming; suggested by jmc@, OK tedu@ jmc@.

7 years agoEnable the use of futex(2) in librthread on mips64.
visa [Tue, 4 Jul 2017 14:33:43 +0000 (14:33 +0000)]
Enable the use of futex(2) in librthread on mips64.

OK mpi@, deraadt@

7 years agoRevert previous, it exposed two edge cases still requiring the
mpi [Tue, 4 Jul 2017 14:10:15 +0000 (14:10 +0000)]
Revert previous, it exposed two edge cases still requiring the
KERNEL_LOCK().

- radix_node_head are still allocated with malloc(9) and pf_table can
  free(9) some.

- pfsync(4) might send some traffic which can exercise IPsec code

Found by bluhm@ and Hrvoje Popovski

7 years agoShould have committed this two days ago, last part that ran during depend
espie [Tue, 4 Jul 2017 13:10:13 +0000 (13:10 +0000)]
Should have committed this two days ago, last part that ran during depend
can run normally

7 years agoAlways hold the socket lock when calling sblock().
mpi [Tue, 4 Jul 2017 12:58:32 +0000 (12:58 +0000)]
Always hold the socket lock when calling sblock().

Implicitely protects `so_state' with the socket lock in sosend().

ok visa@, bluhm@

7 years agoAssert that the socket lock is held when `so_state' is modified.
mpi [Tue, 4 Jul 2017 12:52:48 +0000 (12:52 +0000)]
Assert that the socket lock is held when `so_state' is modified.

ok bluhm@, visa@

7 years agoAssert that the socket lock is held when `so_qlen' is modified.
mpi [Tue, 4 Jul 2017 12:51:18 +0000 (12:51 +0000)]
Assert that the socket lock is held when `so_qlen' is modified.

ok bluhm@, visa@

7 years agoconsistently use the evtimer wrappers around the connection timeout.
florian [Tue, 4 Jul 2017 12:48:36 +0000 (12:48 +0000)]
consistently use the evtimer wrappers around the connection timeout.

this is instead of setting the timeout up with event_set and
event_add, but removing it with evtimer_del. this uses evtimer_set
and evtimer_add.

prompted by dlg's commit to identd
OK dlg

7 years agoInclude stdarg.h explicitly instead of relying on event.h to bring it in.
florian [Tue, 4 Jul 2017 12:47:51 +0000 (12:47 +0000)]
Include stdarg.h explicitly instead of relying on event.h to bring it in.
prompted by dlg's commit to identd
OK dlg

7 years agocomplete "nodepend" in source
espie [Tue, 4 Jul 2017 12:43:07 +0000 (12:43 +0000)]
complete "nodepend" in source

- mix-up between .ALLSRC and .IMPSRC, compile the right gc.c file
- add explicit targets for pre-generated stuff

7 years agoHandle 0 size of preview box in caller.
nicm [Tue, 4 Jul 2017 12:26:14 +0000 (12:26 +0000)]
Handle 0 size of preview box in caller.

7 years agoBackout previous due to a bug discovered by zhuk@ that requires some tinkering
anton [Tue, 4 Jul 2017 11:46:15 +0000 (11:46 +0000)]
Backout previous due to a bug discovered by zhuk@ that requires some tinkering
and is not an easy fix for now.

7 years agoAdd GRACEFUL_SHUTDOWN example to bgpd.conf
job [Tue, 4 Jul 2017 09:33:36 +0000 (09:33 +0000)]
Add GRACEFUL_SHUTDOWN example to bgpd.conf

OK phessler@

7 years agoExport our atomic definitions to userland instead of gcc builtins.
mpi [Tue, 4 Jul 2017 09:00:12 +0000 (09:00 +0000)]
Export our atomic definitions to userland instead of gcc builtins.

ok kettenis@, dlg@

7 years agogenerate the parser directly instead of fucking around
espie [Tue, 4 Jul 2017 08:39:57 +0000 (08:39 +0000)]
generate the parser directly instead of fucking around

7 years agoAdd support for pattern substitution to variables in ksh using a common syntax
anton [Tue, 4 Jul 2017 07:29:32 +0000 (07:29 +0000)]
Add support for pattern substitution to variables in ksh using a common syntax
borrowed from ksh93.

Survived a ports build performed by naddy@ and encouraged by many.

7 years agoconsistently use the evtimer wrappers around the connection timeout.
dlg [Tue, 4 Jul 2017 01:09:42 +0000 (01:09 +0000)]
consistently use the evtimer wrappers around the connection timeout.

this is instead of setting the timeout up with event_set and
event_add, but removing it with evtimer_del. this uses evtimer_set
and evtimer_add.

7 years agoinclude stdarg.h explicitly instead of relying on event.h to bring it in.
dlg [Tue, 4 Jul 2017 01:07:32 +0000 (01:07 +0000)]
include stdarg.h explicitly instead of relying on event.h to bring it in.

7 years agolet BUILDFIRST just build first, using the same trick guenther@ did in
espie [Tue, 4 Jul 2017 00:59:11 +0000 (00:59 +0000)]
let BUILDFIRST just build first, using the same trick guenther@ did in
kernel Makefiles (thank you!), also add lex and yacc generated files to
that list.

Let those lex/yacc variables be defined for bsd.lib.mk as well.

Complete the SRCS -> OBJS hints with the library dependency

This should fix most dependency problems.

Commit now, so that people can check if something else is still amiss.
krw@ agrees

7 years agoproperly adjust the ttl of replies instead of freezing them in time
tedu [Tue, 4 Jul 2017 00:30:45 +0000 (00:30 +0000)]
properly adjust the ttl of replies instead of freezing them in time

7 years agoblacklist MSI on intel m.2 optane parts. it just doesnt work there.
dlg [Mon, 3 Jul 2017 23:27:53 +0000 (23:27 +0000)]
blacklist MSI on intel m.2 optane parts. it just doesnt work there.

discovered by jmatthew@
tweaks and ok jsg@ kettenis@

7 years agoDo not grab the socket lock in doaccept() twice. Pass NOTE_SUBMIT
bluhm [Mon, 3 Jul 2017 23:05:21 +0000 (23:05 +0000)]
Do not grab the socket lock in doaccept() twice.  Pass NOTE_SUBMIT
to KNOTE() as we are already holding the lock.  Fixes "panic:
rw_enter: netlock locking against myself" reported by Gregor Best
and reproduced with src/regress/lib/libtls/gotls.
OK millert@

7 years agoChange session and window preview so that the current window or pane is
nicm [Mon, 3 Jul 2017 22:48:02 +0000 (22:48 +0000)]
Change session and window preview so that the current window or pane is
always shown.

7 years agono need to generate y.tab.h if nothing uses it, set YFLAGS to nothing
espie [Mon, 3 Jul 2017 22:21:47 +0000 (22:21 +0000)]
no need to generate y.tab.h if nothing uses it, set YFLAGS to nothing
instead of CLEANFILES += y.tab.h

okay millert@

7 years agosizeof(struct sockaddr_in) != sizeof(struct in_addr).
krw [Mon, 3 Jul 2017 22:06:11 +0000 (22:06 +0000)]
sizeof(struct sockaddr_in) != sizeof(struct in_addr).

Fix construction of the unicast UDP packets being sent
out via sendmsg(). Accidentally broken in r1.32 (5.6)
when types were juggled.

ok bluhm@

7 years agoAdd more checks and debug info to identify the sporadic test failures.
bluhm [Mon, 3 Jul 2017 19:35:06 +0000 (19:35 +0000)]
Add more checks and debug info to identify the sporadic test failures.

7 years agoBoth PF_ROUTE and PF_KEY need to start to take care of their own PCB list
claudio [Mon, 3 Jul 2017 19:23:47 +0000 (19:23 +0000)]
Both PF_ROUTE and PF_KEY need to start to take care of their own PCB list
to simplify the locking of those lists. For now move the LIST_ENTRY from
rawcb into routecb and in the pfkey case a new keycb. Do some cleanup and
renaming in the pfkey code but more to come.
OK bluhm@ mpi@

7 years agos/log_warn/log_warnx/ where errno is not set
florian [Mon, 3 Jul 2017 19:02:04 +0000 (19:02 +0000)]
s/log_warn/log_warnx/ where errno is not set

7 years agoRemove obvious clear_config() comments and misleading state_change() comments.
benno [Mon, 3 Jul 2017 18:45:34 +0000 (18:45 +0000)]
Remove obvious clear_config() comments and misleading state_change() comments.
Also relocate do_action() calls for the init block from change_state() to
occur with the corresponding do_action() calls for the body block.
From Rob Pierce
ok stsp@ and i

7 years agoGenerate source and header files for regress on demand.
bluhm [Mon, 3 Jul 2017 18:19:55 +0000 (18:19 +0000)]
Generate source and header files for regress on demand.
OK espie@

7 years agoreport trailing delimiters after macros where they are usually a mistake;
schwarze [Mon, 3 Jul 2017 17:33:01 +0000 (17:33 +0000)]
report trailing delimiters after macros where they are usually a mistake;
the idea came up in a discussion with Thomas Klausner <wiz at NetBSD>

7 years agoUse copyin32() instead of copyin() in RDHWR emulation.
visa [Mon, 3 Jul 2017 16:37:07 +0000 (16:37 +0000)]
Use copyin32() instead of copyin() in RDHWR emulation.
This reduces emulation time about 5%.

7 years agodon't bother caching invalid or very short lived responses
tedu [Mon, 3 Jul 2017 16:36:48 +0000 (16:36 +0000)]
don't bother caching invalid or very short lived responses

7 years agoLink all lib regress tests to the build.
bluhm [Mon, 3 Jul 2017 16:24:51 +0000 (16:24 +0000)]
Link all lib regress tests to the build.

7 years agoallows working without depends, as discovered and discussed with drahn@
espie [Mon, 3 Jul 2017 15:17:11 +0000 (15:17 +0000)]
allows working without depends, as discovered and discussed with drahn@
okay drahn@

7 years agore-add dependencies from SRCS to corresponding OBJS.
espie [Mon, 3 Jul 2017 15:11:02 +0000 (15:11 +0000)]
re-add dependencies from SRCS to corresponding OBJS.
the old mkdep run did walk the SRCS list and write those as a side-effect.

Without this, bsd.prog.mk .SUFFIXES order will have OBJS depend on the
wrong src file.

Causes of the alpha/stand/bootxx breakage.

7 years agomarkup fixes;
jmc [Mon, 3 Jul 2017 14:30:11 +0000 (14:30 +0000)]
markup fixes;

7 years agowarn about time machines; suggested by Thomas Klausner <wiz @ NetBSD>
schwarze [Mon, 3 Jul 2017 13:40:00 +0000 (13:40 +0000)]
warn about time machines; suggested by Thomas Klausner <wiz @ NetBSD>

7 years agomerge error, revealed by clang; ok kettenis@
naddy [Mon, 3 Jul 2017 13:26:04 +0000 (13:26 +0000)]
merge error, revealed by clang; ok kettenis@

7 years agoChange previous to not wait for both process exit and pty close -
nicm [Mon, 3 Jul 2017 12:38:50 +0000 (12:38 +0000)]
Change previous to not wait for both process exit and pty close -
instead if there is a pipe-pane active, do not exit until all data is
read (including any libevent hasn't seen yet). Fixes problem reported by
Theo Buehler and still seems to solve the original issue.

7 years agoRemove mdoclint(1) from the tree.
schwarze [Mon, 3 Jul 2017 11:21:23 +0000 (11:21 +0000)]
Remove mdoclint(1) from the tree.
All functionality is now covered by mandoc -Tlint.
OK jmc@ wiz@

7 years agocheck that a cachehit hasn't expired before using it.
tedu [Mon, 3 Jul 2017 09:39:48 +0000 (09:39 +0000)]
check that a cachehit hasn't expired before using it.

7 years agoReplace slot time durations with macros.
kevlo [Mon, 3 Jul 2017 09:21:09 +0000 (09:21 +0000)]
Replace slot time durations with macros.

ok stsp@

7 years agoProtect `so_state', `so_snd' and `so_rcv' with the socket lock in
mpi [Mon, 3 Jul 2017 08:31:39 +0000 (08:31 +0000)]
Protect `so_state', `so_snd' and `so_rcv' with the socket lock in
kqueue filters.

ok millert@, bluhm@, visa@

7 years agoProtect `so_state', `so_error' and `so_qlen' with the socket lock in
mpi [Mon, 3 Jul 2017 08:29:24 +0000 (08:29 +0000)]
Protect `so_state', `so_error' and `so_qlen' with the socket lock in
kqueue filters.

ok millert@, bluhm@, visa@

7 years agoNote arm64 and armv7.
visa [Mon, 3 Jul 2017 08:27:59 +0000 (08:27 +0000)]
Note arm64 and armv7.

7 years agoNot all sockets require the NET_LOCK(). So use solock() instead in
mpi [Mon, 3 Jul 2017 08:27:07 +0000 (08:27 +0000)]
Not all sockets require the NET_LOCK().  So use solock() instead in
fifo_close().

ok millert@, visa@

7 years agoUpdate attachment on octeon.
visa [Mon, 3 Jul 2017 08:23:36 +0000 (08:23 +0000)]
Update attachment on octeon.

7 years agoAttach com(4) using fdt on octeon.
visa [Mon, 3 Jul 2017 08:17:20 +0000 (08:17 +0000)]
Attach com(4) using fdt on octeon.

The relevant part of uartbus(4) is made part of the com(4) glue
to avoid extra maneuvers in the code.

7 years agoDo not close panes until process has exited and any outstanding data
nicm [Mon, 3 Jul 2017 08:16:03 +0000 (08:16 +0000)]
Do not close panes until process has exited and any outstanding data
has been written to the pipe-pane event if there is one. GitHub issue 991.

7 years agoDrop the KERNEL_LOCK() in pf_purge_thread().
mpi [Mon, 3 Jul 2017 08:11:21 +0000 (08:11 +0000)]
Drop the KERNEL_LOCK() in pf_purge_thread().

The NET_LOCK() is currently what guarantees that accesses to PF data
structures are serialized.

While here use rwsleep(9) instead of calling NET_LOCK()/NET_UNLOCK()
for every iteration.

Tested by Hrvoje Popovski, ok sashan@, visa@

7 years agoTry C.UTF-8 which is also a commonly useful locale on some platforms,
nicm [Mon, 3 Jul 2017 08:08:30 +0000 (08:08 +0000)]
Try C.UTF-8 which is also a commonly useful locale on some platforms,
from Romain Francoise.

7 years agoAdd the first UART into the fdt blob.
visa [Mon, 3 Jul 2017 07:58:42 +0000 (07:58 +0000)]
Add the first UART into the fdt blob.

7 years agoexplicitly pass -MD -MP in DFLAGS, so that even if CFLAGS are forced
espie [Mon, 3 Jul 2017 07:54:18 +0000 (07:54 +0000)]
explicitly pass -MD -MP in DFLAGS, so that even if CFLAGS are forced
on the command line, things will still work