kettenis [Fri, 5 Aug 2016 20:38:17 +0000 (20:38 +0000)]
We don't want drivers with numbers in their names.
Requested by deraadt@
kettenis [Fri, 5 Aug 2016 19:56:52 +0000 (19:56 +0000)]
Replace inappropriate use of CP15_CNTPCT with CP_DCCIMVAC.
No binary change.
kettenis [Fri, 5 Aug 2016 19:00:25 +0000 (19:00 +0000)]
Switch the sunxi platform over to the new interrupt establish API. This
involves turning sxiahci(4) into a real driver that dynamically attaches.
ehci(4) now also attaches dynamically.
ok jsg@, patrick@
mikeb [Fri, 5 Aug 2016 18:31:21 +0000 (18:31 +0000)]
Switch pending event clearing to an atomic swap operation
Rather than performing an atomic bit clearing for every encountered
event bit set we can adjust the code to perform an atomic swap of a
single row of the events array and decrease the amount of expensive
atomic operations.
Same optimization as for Hyper-V. From FreeBSD.
mikeb [Fri, 5 Aug 2016 18:16:04 +0000 (18:16 +0000)]
Switch pending event clearing to an atomic swap operation
Rather than performing an atomic bit clearing for every encountered
event bit set we can adjust the code to perform an atomic swap of a
single row of the events array and decrease the amount of expensive
atomic operations.
From FreeBSD.
mikeb [Fri, 5 Aug 2016 18:12:48 +0000 (18:12 +0000)]
Fixup incorrect EAGAIN handling
mikeb [Fri, 5 Aug 2016 18:12:20 +0000 (18:12 +0000)]
Read events array offset must be calculated in bytes
mikeb [Fri, 5 Aug 2016 18:05:44 +0000 (18:05 +0000)]
Remove incorrect optimization: ring data portion size is not power of 2
deraadt [Fri, 5 Aug 2016 17:25:51 +0000 (17:25 +0000)]
Do not *printf %s NULL
ok bcook
schwarze [Fri, 5 Aug 2016 17:20:30 +0000 (17:20 +0000)]
fix a typo that prevented names from .Dt from getting priority over
names from .Sh NAME; no dire consequences on OpenBSD since we no
longer have MLINKS for mdoc(7) pages
schwarze [Fri, 5 Aug 2016 17:02:04 +0000 (17:02 +0000)]
* Properly distinguish commands (.Ic) and command modifiers (.Cm).
* Consistently use .Nm for the page name, do not oscillate to .Pa.
deraadt [Fri, 5 Aug 2016 16:46:50 +0000 (16:46 +0000)]
sync
schwarze [Fri, 5 Aug 2016 16:16:06 +0000 (16:16 +0000)]
Obvious minor fixes:
* Add missing .Dv, .Ev, and .Fa macros.
* Delete deprecated .Tn macros.
* Mark up global variable names with .Va, not with .Fa or .Li.
* Mark up config file commands with .Ic, not with .Fa.
* Fix HISTORY, trivial to verify from the CSRG archive CD.
tedu [Fri, 5 Aug 2016 15:42:05 +0000 (15:42 +0000)]
use a larger chunk for getentropy() and save some for next time.
coalesces some syscalls instead of one per random number.
ok deraadt
krw [Fri, 5 Aug 2016 14:02:23 +0000 (14:02 +0000)]
Don't leak the option data of non-DHCPINFORM messages received on
the udp socket.
Found by David Carlier.
ok yasuoka@
tb [Fri, 5 Aug 2016 13:49:19 +0000 (13:49 +0000)]
All four arguments can be omitted by specifying a dash.
While there, remove a superfluous 's'.
discussed with jmc
tb [Fri, 5 Aug 2016 13:43:38 +0000 (13:43 +0000)]
If reps and steps were omitted, while begin and end were specified,
use a default step size of +1 or -1. Fixes the last two failing
regression tests dhhd and dhhd2. Adapted from FreeBSD.
'good direction' deraadt
tb [Fri, 5 Aug 2016 13:33:21 +0000 (13:33 +0000)]
If "%%" appears in the format string, skip it properly.
Part of pjanzen's fixes in r1.8, but was lost in r1.17.
Fixes regression tests wp3, wp4, wp5.
'good direction' deraadt
kettenis [Fri, 5 Aug 2016 13:31:29 +0000 (13:31 +0000)]
Unmask the timer output signal for real.
ok patrick@, jsg@
tb [Fri, 5 Aug 2016 13:24:04 +0000 (13:24 +0000)]
enable dddd regression test
tb [Fri, 5 Aug 2016 13:21:31 +0000 (13:21 +0000)]
Initialize reps, begin, end, steps to their defaults.
Makes jot - - - - behave the same way as jot -, jot - - and jot - - -
instead of spurting out zeroes indefinitely.
Fixes the dddd regression test and matches FreeBSD and NetBSD.
From attila () stalphonsos ! com, thanks!
'good direction' deraadt
visa [Fri, 5 Aug 2016 13:18:27 +0000 (13:18 +0000)]
Drop a pointless link carrier check from the RX path.
jca [Fri, 5 Aug 2016 11:53:23 +0000 (11:53 +0000)]
Using a macro to save 4 chars is mere obfuscation.
jca [Fri, 5 Aug 2016 11:44:10 +0000 (11:44 +0000)]
Kill an unused macro
...instead of leaving an XXX about it.
kettenis [Fri, 5 Aug 2016 11:40:22 +0000 (11:40 +0000)]
imxahci(4) needs ahci.c
jca [Fri, 5 Aug 2016 11:38:00 +0000 (11:38 +0000)]
Add a -u switch to always log route insertions/deletions.
Route updates (and associated warnings) are always prefixed with "RTADD"
or "RTDEL". This is useful for people that previously used the -R
option that got removed.
Tested by Freddy Dissaux.
mikeb [Fri, 5 Aug 2016 11:37:57 +0000 (11:37 +0000)]
Trigger link state update only on media status updates
mikeb [Fri, 5 Aug 2016 11:36:24 +0000 (11:36 +0000)]
Fixup the condition that prevented heartbeat & co. from attaching
jca [Fri, 5 Aug 2016 11:34:51 +0000 (11:34 +0000)]
Use tighter pledge(2) restrictions.
Made possible by the logging overhaul. Tested by Freddy Dissaux.
jca [Fri, 5 Aug 2016 11:32:28 +0000 (11:32 +0000)]
Rework logging.
Use log.c like in many other daemons, with small additions to maintain
a log queue (useful to dump a single log line with several logging
statements). Use log.c to dump the internal state, and remove the -R
option (a similar feature will be implemented differently soon).
The end goal is to avoid touching the filesystem and use much tighter
pledge(2) restrictions.
Tested by Freddy Dissaux.
jca [Fri, 5 Aug 2016 10:34:18 +0000 (10:34 +0000)]
Switch from select(2) to poll(2); ok millert@
No need to check for POLLHUP since those are UDP sockets.
jmc [Fri, 5 Aug 2016 08:12:12 +0000 (08:12 +0000)]
shorten the openssl dsa text;
jmc [Fri, 5 Aug 2016 08:10:51 +0000 (08:10 +0000)]
grammar fix;
martijn [Fri, 5 Aug 2016 07:54:15 +0000 (07:54 +0000)]
Make RES_OPTIONS point directly to resolv.conf(5) instead of going through
resolver(3).
OK jmc@
vgross [Thu, 4 Aug 2016 20:46:24 +0000 (20:46 +0000)]
Commit in6_selectsrc() split again, with missing assignment fixed.
kettenis [Thu, 4 Aug 2016 15:52:52 +0000 (15:52 +0000)]
Dynamically attach ampintc(4) and make it register itself as an interrupt
controller. Switch all i.MX6 devices over to the new FDT-aware interrupt
establish API and enable imxgpc(4).
This may break other platforms that use ampintc(4) as their interrupt
controller. We will fix this as soon as possible by switching them over
to the new interrupt establish API as well.
This also removes the hack in fec(4) to avoid the gpio-based interrupt
workaround. This commit is an essential step on the road to fix that
issue properly.
ok patrick@, jsg@
kettenis [Thu, 4 Aug 2016 14:31:50 +0000 (14:31 +0000)]
Add imxgpc(4), a driver for the i.MX6 General Power Controller (GPC) which
acts as a "transparent" interrupt controller. This driver just hands
interrupt handlers to its parent (the Cortex-A9 GIC). It doesn't implement
any powermanagement features yet.
ok patrick@
kettenis [Thu, 4 Aug 2016 14:14:05 +0000 (14:14 +0000)]
After fixing up some of the AR8035 PHY registers, issue a reset. This seems
to fix autonegotiation at 100BaseTX. Drop the bit of code that powers up
the PHY as the reset should already take care of that.
ok patrick@
visa [Thu, 4 Aug 2016 13:22:20 +0000 (13:22 +0000)]
Mention jumbo frames.
visa [Thu, 4 Aug 2016 13:10:31 +0000 (13:10 +0000)]
Add support for jumbo frames.
kettenis [Thu, 4 Aug 2016 12:17:36 +0000 (12:17 +0000)]
Add support for pre-registering interrupts. This allows device drivers to
establish interrupts before their interrupt controller attaches, solving
dependency problems in various device trees.
Also add support for handing interrupt handlers over to parent interrupt
controllers.
ok jsg@ patrick@ (on an earlier diff)
schwarze [Thu, 4 Aug 2016 09:33:53 +0000 (09:33 +0000)]
Fix an assertion failure that happened when trying to add a page
with makewhatis -d to a completely empty database.
Reported by Mark Patruck <mark at wrapped dot cx>, thanks!
schwarze [Thu, 4 Aug 2016 09:18:53 +0000 (09:18 +0000)]
adjust database file name and version number to what was committed
jmc [Thu, 4 Aug 2016 05:39:28 +0000 (05:39 +0000)]
tpm tweaks;
krw [Wed, 3 Aug 2016 23:37:25 +0000 (23:37 +0000)]
Apply a warning fix from lukem@NetBSD circa 1997 (more '()'), and
one from he@NetBSD circa 1998 (more '{}').
Makes gcc happier.
krw [Wed, 3 Aug 2016 23:22:48 +0000 (23:22 +0000)]
Apply fix from David Binderman via NetBSD. "Use all 3 of b, p, n
to determine if route changed".
Spotted by the Echelon team with AppChecker static analyzer.
NetBSD fix pointed out by millert@
krw [Wed, 3 Aug 2016 23:13:54 +0000 (23:13 +0000)]
Apply fix from mrouted 3.8 via NetBSD. Our mbone support may
be slightly dated.
Spotted by the Echelon team with AppChecker static analyzer.
jcs [Wed, 3 Aug 2016 22:50:47 +0000 (22:50 +0000)]
move acpimcfg to keep list sorted
jcs [Wed, 3 Aug 2016 22:47:43 +0000 (22:47 +0000)]
reference dwiic here too
jcs [Wed, 3 Aug 2016 22:46:44 +0000 (22:46 +0000)]
add a man page for tpm(4)
vgross [Wed, 3 Aug 2016 20:45:36 +0000 (20:45 +0000)]
Fix typo, spotted by Holger Mikolon <holger@mikolon.com>
jcs [Wed, 3 Aug 2016 17:33:50 +0000 (17:33 +0000)]
not having a century is a known layout by now
jcs [Wed, 3 Aug 2016 17:29:18 +0000 (17:29 +0000)]
don't complain about invalid checksum if the bios isn't writing it
properly, the user isn't going to do anything about it
jcs [Wed, 3 Aug 2016 17:23:38 +0000 (17:23 +0000)]
add a minimal TPM 1.2 driver just to issue a "save state" command
before suspending, in order to fix suspend/resume on some newer
machines.
more cleanup and TPM 2.0 support coming later.
earlier version ok pirofti
ok deraadt
mikeb [Wed, 3 Aug 2016 17:14:41 +0000 (17:14 +0000)]
Use atomic operations to manipulate event masking bits
krw [Wed, 3 Aug 2016 16:32:08 +0000 (16:32 +0000)]
strnames[i][0] == 'k' && strnames[i][0] == 'f' should be
strnames[i][0] == 'k' && strnames[i][1] == 'f'.
Spotted by the Echelon team with AppChecker static analyzer.
millert@ naddy@ and ncurses 6.0 all agree.
krw [Wed, 3 Aug 2016 16:27:25 +0000 (16:27 +0000)]
A couple of "a->blah == a->blah" -> "a->blah == b->blah".
Spotted by the Echelon team with AppChecker static analyzer.
ok sashan@
mikeb [Wed, 3 Aug 2016 15:08:06 +0000 (15:08 +0000)]
Remove the periodic timer and do rescheduling during Rx completion
This change adds a check into the Rx ring completion routine that
schedules an interrupt task to be executed immediately after if
consumer index has already advanced itself. The benefit of doing
this compared to an additional loop after replenishing the ring
(as done in FreeBSD for example) is that first of all this goes
through the loop in the taskqueue thread with a yeild check to
prevent CPU hogging and second is that it triggers Tx completion
as well since interrupt handler runs both.
mikeb [Wed, 3 Aug 2016 14:55:57 +0000 (14:55 +0000)]
Use an atomic operation to clear pending event bits
Pending event bits are located in a shared memory and are potentially
accessed by multiple CPUs running dom0 and the guest VM. It appears
that a failure to synchronize changes to this shared memory leads to
race conditions resulting in the guest missing out on notifications.
krw [Wed, 3 Aug 2016 13:44:49 +0000 (13:44 +0000)]
Fold umass_atapi_attach() and umass_scsi_setup() into umass_scsi_attach() to
shorten and simplify code.
No intentional functional change.
kettenis [Wed, 3 Aug 2016 11:52:43 +0000 (11:52 +0000)]
Simplify the way we handle TLB flushes. Since ARMv7 effectively has a
unified TLB there is not much point in optimizing TLB flushing for pages
that have never been executable. The only difference is a flush of the
branch predictor and even that isn't necessary anymore on all but the oldest
Cortex cores.
ok patrick@
nicm [Wed, 3 Aug 2016 09:08:40 +0000 (09:08 +0000)]
We only replace the first %%, not multiple (use %2 for second).
nicm [Wed, 3 Aug 2016 09:07:02 +0000 (09:07 +0000)]
Fix minimum size when pane status line is enabled, reported by Y Petremann.
kettenis [Wed, 3 Aug 2016 07:05:05 +0000 (07:05 +0000)]
The ARMv7 architecture deprecates the separate Instruction and Data TLB
maintanenance instruction and recommends to only use the instructions that
operate on the unified TLB. Those instructions will flush both TLBs on
implementations that still have separate Instruction and Data TLBs. Switch
the TLB maintenance primitives over to use those. This allows us to reduce
the number of primitives on armv7. We still keep separate "ID" and "D"
variants as we still have to flush the branch predictor when changing the
mappings of pages that contain instructions.
ok jsg@
jmc [Wed, 3 Aug 2016 06:43:21 +0000 (06:43 +0000)]
shorten the openssl dhparam text;
guenther helped rewrite the -dsaparam parts.
djm [Wed, 3 Aug 2016 05:41:57 +0000 (05:41 +0000)]
small refactor of cipher.c: make ciphercontext opaque to callers
feedback and ok markus@
dtucker [Wed, 3 Aug 2016 04:23:55 +0000 (04:23 +0000)]
Fix bug introduced in rev 1.467 which causes "buffer_get_bignum_ret:
incomplete message" errors when built with WITH_SSH1 and run such that no
Protocol 1 ephemeral host key is generated (eg "Protocol 2", no SSH1 host
key supplied). Reported by rainer.laatsch at t-online.de, ok deraadt@
martijn [Tue, 2 Aug 2016 20:17:06 +0000 (20:17 +0000)]
Fix the begin of word bug in s-command, similar to sed.
nit fix and OK schwarze@
jca [Tue, 2 Aug 2016 17:00:09 +0000 (17:00 +0000)]
Move to libevent; ok florian@
jca [Tue, 2 Aug 2016 16:17:54 +0000 (16:17 +0000)]
Ensure that "lim" is initialized and reset at each loop run.
Avoids possible invalid memory accesses. ok florian@
jca [Tue, 2 Aug 2016 16:09:55 +0000 (16:09 +0000)]
Ignore the -bindir option, meaningless on OpenBSD.
Lack of -bindir handling is not a problem right now because -bindir is
passed to cc(1) which ignores it, and because read(2) on /usr/local/bin
returns 0 right now, so ld(1) copes with it. But there are plans to
make read(2) fail with EISDIR.
Input from and bulk builds thanks to ajacoutot@, ok ajacoutot@ espie@
jca [Tue, 2 Aug 2016 16:05:32 +0000 (16:05 +0000)]
Allow specifying an alternate socket path.
This allows one to run multiple ripd instances, for example to serve
multiple rdomains. Diff from Nima GHOTBI, ok claudio@ florian@ benno@
jcs [Tue, 2 Aug 2016 16:00:10 +0000 (16:00 +0000)]
capitalize a question like all the others
ok krw
jsing [Tue, 2 Aug 2016 07:47:11 +0000 (07:47 +0000)]
Revert previous since it adds new symbols.
Requested by deraadt@
benno [Mon, 1 Aug 2016 21:25:53 +0000 (21:25 +0000)]
whitespace and 80 columns, no binary change
benno [Mon, 1 Aug 2016 21:15:30 +0000 (21:15 +0000)]
sync http.h with relayd
ok reyk@
benno [Mon, 1 Aug 2016 21:14:45 +0000 (21:14 +0000)]
rfc 7230 3.3.3 says: response with a status code of 1xx
(Informational) or 204 (No Content) MUST not have a Content-Length.
Add a check for that.
ok reyk@
kettenis [Mon, 1 Aug 2016 21:08:20 +0000 (21:08 +0000)]
Looks like patrick@ committed an older version of the diff. Fix this, and
rename a few variables and functions to be a bit more consistent with other
armv7 code.
schwarze [Mon, 1 Aug 2016 20:46:33 +0000 (20:46 +0000)]
Repair makewhatis -t, sorry for breaking it in the previous commit.
Committing this quickly because it hurts package builders.
Regression reported by naddy@.
martijn [Mon, 1 Aug 2016 20:30:25 +0000 (20:30 +0000)]
Allow a bel character in wall. This was missed when migrating away from
vis(3). This brings the allowed characters on par with write(1).
Noticed by consus <at> gmx <dot> com
OK millert@
bentley [Mon, 1 Aug 2016 18:27:35 +0000 (18:27 +0000)]
Remove vi's "directory" option and TMPDIR support.
ok jung@
jmc [Mon, 1 Aug 2016 18:20:19 +0000 (18:20 +0000)]
the default smtpd setup does not require newaliases be run;
noted by consus
ok gilles
jsing [Mon, 1 Aug 2016 17:40:23 +0000 (17:40 +0000)]
Bump TLS_API for addition of ALPN support.
jsing [Mon, 1 Aug 2016 17:32:19 +0000 (17:32 +0000)]
Add ALPN support to libtls.
ok beck@ doug@
jca [Mon, 1 Aug 2016 17:04:18 +0000 (17:04 +0000)]
zap unused var
stefan [Mon, 1 Aug 2016 16:40:09 +0000 (16:40 +0000)]
Make sure variables are initialized before using them
In elf.c, do not increment `s' before it is initialized. At the time
of the increment, `s' is otherwise unused anyway.
In elflink.c, initialize sec_contents and l_sec_contents to make
sure that the free(sec_contents) and free(l_sec_contents) are called
on valid pointers.
ok jca@ guenther@
stefan [Mon, 1 Aug 2016 16:32:10 +0000 (16:32 +0000)]
Do not pass VM name directly to setproctitle format string.
Use a secure idiom instead.
ok mlarkin@
tedu [Mon, 1 Aug 2016 15:58:22 +0000 (15:58 +0000)]
bring the light of ansi to a few more files
mikeb [Mon, 1 Aug 2016 14:37:39 +0000 (14:37 +0000)]
Don't forget to destroy the taskqueue on interrupt disestablish
patrick [Mon, 1 Aug 2016 14:17:00 +0000 (14:17 +0000)]
Implement an FDT-aware interrupt establish API. This means the drivers
don't need to know where to attach to. Instead the API will take care
of finding the correct interrupt establish for a given device node and
will call it with the correct data.
Adapted from the OFW GPIO framework.
ok kettenis@
mikeb [Mon, 1 Aug 2016 13:48:33 +0000 (13:48 +0000)]
Mark shared producer and consumer indices volatile
schwarze [Mon, 1 Aug 2016 12:24:37 +0000 (12:24 +0000)]
document the new file format
krw [Mon, 1 Aug 2016 10:56:31 +0000 (10:56 +0000)]
Set SDEV_UMASS on all umass devices, even when they use ATAPI or
UFI command sets. Probe all the LUNs a device reports instead of
assuming ATAPI and UFI devices always have 1 LUN.
Motivated by the device David Vasek reported as claiming to be
ATAPI but wanting to use >1 LUN. Committing quickly to smoke out
the inevitable device that works in the exact opposite way. All
hail USB.
ok kettenis@
schwarze [Mon, 1 Aug 2016 10:32:39 +0000 (10:32 +0000)]
Remove the dependency on SQLite without loss of functionality.
Drop the obsolete names_check() now that we deleted MLINKS.
Run "doas makewhatis" after compiling and installing this.
Earlier version tested by jmc@ and jturner@;
"commit it all" deraadt@ "commit and dodge" krw@
jmc [Mon, 1 Aug 2016 07:23:29 +0000 (07:23 +0000)]
shorten the openssl dgst text; in particular, do not try to
list all the available digests;
jmc [Mon, 1 Aug 2016 07:22:00 +0000 (07:22 +0000)]
update currency exchange rates;
espie [Mon, 1 Aug 2016 00:48:22 +0000 (00:48 +0000)]
document default for DISTDIR
schwarze [Sun, 31 Jul 2016 23:35:26 +0000 (23:35 +0000)]
Some base system pages, for example perl(1), contain non-ASCII
characters in their source code, so switch on charset autodetection
in the same way as in man(1) itself.
Issue reported by Pavan Maddamsetti at gmail dot com on bugs@.
kettenis [Sun, 31 Jul 2016 22:27:07 +0000 (22:27 +0000)]
Remove devmap stuff which is unused on armv7.
ok patrick@
kettenis [Sun, 31 Jul 2016 22:04:44 +0000 (22:04 +0000)]
According to te armv7 ARM TLB entries that caused a Permission fault might
be held in the TLB. On top of that valid page table entries might be
speculatively loaded into the TLB. As a result we need to flush TLB entries
even when the page in question has not been referenced.
Fixes pmap_fault_fixup messages on Cortex-A53, and presumably also on
Cortex-A7.
ok patrick@, guenther@