openbsd
7 years agoRemove double assignments
mikeb [Fri, 28 Apr 2017 14:15:45 +0000 (14:15 +0000)]
Remove double assignments

7 years agoStart deleting redundant features that are already covered
schwarze [Fri, 28 Apr 2017 14:11:01 +0000 (14:11 +0000)]
Start deleting redundant features that are already covered
by mandoc(1) -Tlint, as suggested by the upstream author
and maintainer Thomas Klausner <wiz @ NetBSD>.
The ultimate goal is to get rid of mdoclint and integrate
all its functionality into mandoc, but that will still take
some time, so simplify mdoclint for the time being.
jmc@ "go ahead", OK wiz@, also committed upstream

This first patch removes the AUTHORS check that mandoc does, too.

7 years agoDocument futex(2) with a lot of inputs from schwarze@
mpi [Fri, 28 Apr 2017 14:02:57 +0000 (14:02 +0000)]
Document futex(2) with a lot of inputs from schwarze@

7 years agoDisplay futex(2) operations and arguments.
mpi [Fri, 28 Apr 2017 13:53:05 +0000 (13:53 +0000)]
Display futex(2) operations and arguments.

7 years agoregen
mpi [Fri, 28 Apr 2017 13:51:14 +0000 (13:51 +0000)]
regen

7 years agoAdd futex(2) syscall based on a sane subset of its Linux equivalent.
mpi [Fri, 28 Apr 2017 13:50:54 +0000 (13:50 +0000)]
Add futex(2) syscall based on a sane subset of its Linux equivalent.

The syscall is marked NOLOCK and only FUTEX_WAIT grabs the KERNEL_LOCK()
because of PCATCH and the signal nightmare.

Serialization of threads is currently done with a global & exclusive
rwlock.

Note that the current implementation still use copyin(9) which is not
guaranteed to be atomic.  Committing now such that remaining issues can
be addressed in-tree.

With inputs from guenther@, kettenis@ and visa@.

ok deraadt@, visa@

7 years agoadd City of Derry Airport
benno [Fri, 28 Apr 2017 13:50:02 +0000 (13:50 +0000)]
add City of Derry Airport
ok "i have been there"@

7 years agoDefault for xterm-keys was wrong, stop documenting it.
nicm [Fri, 28 Apr 2017 13:39:59 +0000 (13:39 +0000)]
Default for xterm-keys was wrong, stop documenting it.

7 years agoInstall a few more x86 intrinsics header files.
kettenis [Fri, 28 Apr 2017 11:28:20 +0000 (11:28 +0000)]
Install a few more x86 intrinsics header files.

spotted by espie@

7 years agoadd some comments. no functional change
mlarkin [Fri, 28 Apr 2017 10:09:37 +0000 (10:09 +0000)]
add some comments. no functional change

7 years agorename i8253 "counter" to "channel", a better name for what we are
mlarkin [Fri, 28 Apr 2017 08:14:48 +0000 (08:14 +0000)]
rename i8253 "counter" to "channel", a better name for what we are
emulating

7 years agovmm: don't use invvpid if we didn't detect vpid capability during
mlarkin [Fri, 28 Apr 2017 07:44:36 +0000 (07:44 +0000)]
vmm: don't use invvpid if we didn't detect vpid capability during
vcpu setup

7 years agosort;
jmc [Fri, 28 Apr 2017 06:15:03 +0000 (06:15 +0000)]
sort;

7 years agoMerge missing bits from Colin Watson's patch in bz#2658 which make integrity
dtucker [Fri, 28 Apr 2017 04:16:27 +0000 (04:16 +0000)]
Merge missing bits from Colin Watson's patch in bz#2658 which make integrity
tests more robust against timeouts.  ok djm@

7 years agoPull back some shell portability fixes from -portable to make future syncs
dtucker [Fri, 28 Apr 2017 04:00:14 +0000 (04:00 +0000)]
Pull back some shell portability fixes from -portable to make future syncs
easier.  Note that this also changes the number of bytes skipped on "Bad
Packet" errors from 2 to 3 because the worst case is changing the high byte
of the length field in which case skipping 3 bytes is needed.  ok djm@

7 years agoinclude key fingerprint in "Offering public key" debug message
djm [Fri, 28 Apr 2017 03:24:53 +0000 (03:24 +0000)]
include key fingerprint in "Offering public key" debug message

7 years agoAvoid relying on implementation-specific behavior when detecting
millert [Fri, 28 Apr 2017 03:21:12 +0000 (03:21 +0000)]
Avoid relying on implementation-specific behavior when detecting
whether the timestamp or file size overflowed.  If time_t and
off_t are not either 32-bit or 64-bit scp will exit with an error.
OK djm@

7 years agoAdd SyslogFacility option to ssh(1) matching the equivalent option in
dtucker [Fri, 28 Apr 2017 03:20:27 +0000 (03:20 +0000)]
Add SyslogFacility option to ssh(1) matching the equivalent option in
sshd(8).  bz#2705, patch from erahn at arista.com, ok djm@

7 years agontohl() returns uint32_t so it cannot be < 0. Since we're storing
millert [Thu, 27 Apr 2017 23:54:08 +0000 (23:54 +0000)]
ntohl() returns uint32_t so it cannot be < 0.  Since we're storing
the result in an int check for > INT_MAX instead.  OK bluhm@

7 years agoRemove "len < 0" check; len is socklen_t (uint32_t) so can't be
millert [Thu, 27 Apr 2017 23:52:35 +0000 (23:52 +0000)]
Remove "len < 0" check; len is socklen_t (uint32_t) so can't be
negative.  Quiets a warning from clang.  OK bluhm@

7 years agoBring over the changes to mainbus(4) and simplebus(4) from arm64.
kettenis [Thu, 27 Apr 2017 22:41:46 +0000 (22:41 +0000)]
Bring over the changes to mainbus(4) and simplebus(4) from arm64.

7 years agoMerge trivial patches from pkgsrc to get rid of gratuitious differences;
schwarze [Thu, 27 Apr 2017 22:27:43 +0000 (22:27 +0000)]
Merge trivial patches from pkgsrc to get rid of gratuitious differences;
no functional change on OpenBSD: rev.s 1.50, 1.52, 1.62, 1.63
Of course, keep intentional differences.

7 years agoDetect links to self by parsing .Nm in SYNOPSIS.
schwarze [Thu, 27 Apr 2017 22:07:06 +0000 (22:07 +0000)]
Detect links to self by parsing .Nm in SYNOPSIS.
Functionality suggested and patch OK'ed by jmc@.
Patch also accepted upstream in pkgsrc by wiz@.

7 years agoEnforce that sysctl kern.somaxconn and sominconn can only be set
bluhm [Thu, 27 Apr 2017 21:39:27 +0000 (21:39 +0000)]
Enforce that sysctl kern.somaxconn and sominconn can only be set
to valid values.  The so_qlimit is type short.
report Dillon Jay Pena; OK deraadt@

7 years agoChange len arg of getidle() from size_t to int since it is used as
millert [Thu, 27 Apr 2017 21:28:00 +0000 (21:28 +0000)]
Change len arg of getidle() from size_t to int since it is used as
a length parameter to a "%.*s" printf format.  Quiets a compiler
warning (both gcc and clang).

7 years agochallenge is an array, not a pointer, therefore cannot be NULL.
millert [Thu, 27 Apr 2017 20:55:52 +0000 (20:55 +0000)]
challenge is an array, not a pointer, therefore cannot be NULL.
Quiets a clang warning.

7 years agotls_free(3) and tls_config_free(3) accept NULL;
schwarze [Thu, 27 Apr 2017 20:41:08 +0000 (20:41 +0000)]
tls_free(3) and tls_config_free(3) accept NULL;
patch from Matthew Martin <phy1729 at gmail dot com>,
slightly tweaked by me

7 years agosyspatch needs to create shared libs with the same object sequence instead
robert [Thu, 27 Apr 2017 17:41:47 +0000 (17:41 +0000)]
syspatch needs to create shared libs with the same object sequence instead
of randomizing the order so let's read the libs on the system and link the
same way in case a syspatch is being built

7 years agotrack size of pss allocation, so that length can be passed to free(9)
deraadt [Thu, 27 Apr 2017 16:34:18 +0000 (16:34 +0000)]
track size of pss allocation, so that length can be passed to free(9)
ok mlarkin

7 years agoclang warns about some of the strlcpy arguments here, which aren't the
tedu [Thu, 27 Apr 2017 16:09:32 +0000 (16:09 +0000)]
clang warns about some of the strlcpy arguments here, which aren't the
typical idiom because there's invisible size dependencies. rewrite some
of it to use memcpy, which makes clear the lengths are the same.

7 years agoRemove the "volatile" attribute from bpf_error(). The "volatile"
millert [Thu, 27 Apr 2017 15:46:14 +0000 (15:46 +0000)]
Remove the "volatile" attribute from bpf_error().  The "volatile"
attribute is an obsolete way of saying "noreturn".  clang doesn't
recognize it and the function is already marked __dead so it is
superfluous anyway.

7 years agoRemove useless check for ks_env_table being NULL. Since ks_env_table
millert [Thu, 27 Apr 2017 15:35:36 +0000 (15:35 +0000)]
Remove useless check for ks_env_table being NULL.  Since ks_env_table
is an array embedded in struct keynote_session it can never be NULL.
Quiets a clang warning.

7 years agoremove a static array unused since rev 1.306 spotted by clang
jsg [Thu, 27 Apr 2017 13:40:05 +0000 (13:40 +0000)]
remove a static array unused since rev 1.306 spotted by clang
ok djm@

7 years agoDon't send multiple error messages in response to a single command
mikeb [Thu, 27 Apr 2017 13:30:54 +0000 (13:30 +0000)]
Don't send multiple error messages in response to a single command

While a few commands (such as HELP and NOOP) are allowed in between
USER and PASS, most of the others are not.  However if such command
is issued, ftpd should abort the command evaluation early and reply
with "530 Please login with USER and PASS."

From form@, with input from deraadt@ and OK millert@

7 years agoAvoid potential signed int overflow when parsing the file size.
millert [Thu, 27 Apr 2017 11:53:12 +0000 (11:53 +0000)]
Avoid potential signed int overflow when parsing the file size.
Use strtoul() instead of parsing manually.  OK djm@

7 years agoUnifdef KADB.
mpi [Thu, 27 Apr 2017 11:48:08 +0000 (11:48 +0000)]
Unifdef KADB.

ok deraadt@

7 years agoUse (32-bit) word-sized access in the a4x bus space routine even if only
kettenis [Thu, 27 Apr 2017 10:57:05 +0000 (10:57 +0000)]
Use (32-bit) word-sized access in the a4x bus space routine even if only
a byte or a half-word is needed.  Certain implementations of the Synopsis
Designware copy-and-paste logic blocks don't respond to transactions that
are smaller than a word.

Fixes the serial console on boards with a Rockchip RK3288.

7 years agoAdd code to identify the CPUs on arm64 systems. The primary CPU is attached
kettenis [Thu, 27 Apr 2017 10:23:19 +0000 (10:23 +0000)]
Add code to identify the CPUs on arm64 systems.  The primary CPU is attached
and identified early on.  For the secondary CPUs this happens late, such that
the drivers we need to spin up CPUs, such as psci(4), will be available.

This also fixes some code in simplebus(4) where the return value of
OF_getprop() was not properly checked.

Heavily based on an earlier diff from drahn@.

ok drahn@, jsg@

7 years agorename a struct that was denoted as "VMX only" to make it more clear
mlarkin [Thu, 27 Apr 2017 07:15:35 +0000 (07:15 +0000)]
rename a struct that was denoted as "VMX only" to make it more clear
that it can be used in SVM and VMX.

no functional change

7 years agouse a more descriptive value from the VEI_DIR_xxx enum instead of a
mlarkin [Thu, 27 Apr 2017 06:49:05 +0000 (06:49 +0000)]
use a more descriptive value from the VEI_DIR_xxx enum instead of a
hardcoded number. no functional change

7 years agospelling;
jmc [Thu, 27 Apr 2017 06:22:11 +0000 (06:22 +0000)]
spelling;

7 years agovmm(4): proper save/restore of FPU context during entry/exit.
mlarkin [Thu, 27 Apr 2017 06:16:39 +0000 (06:16 +0000)]
vmm(4): proper save/restore of FPU context during entry/exit.

tested by reyk, dcoppa, and a few others.

ok kettenis@ on the fpu bits
ok deraadt@ on the vmm bits

7 years agoRestore if_iwm.c r1.173 (mira retry change) which got incarcerated by
stsp [Thu, 27 Apr 2017 06:14:23 +0000 (06:14 +0000)]
Restore if_iwm.c r1.173 (mira retry change) which got incarcerated by
the backout police even though it wasn't responsible for anything.
pointed out by tb@

7 years agoiwm(4) monitor mode support was backed out for now.
stsp [Thu, 27 Apr 2017 05:49:32 +0000 (05:49 +0000)]
iwm(4) monitor mode support was backed out for now.

7 years agoRevert if_iwm.c to the state of r1.170 because ETOOMANYDRAGONSATONCE.
stsp [Thu, 27 Apr 2017 05:46:51 +0000 (05:46 +0000)]
Revert if_iwm.c to the state of r1.170 because ETOOMANYDRAGONSATONCE.
Performance regression reported by deraadt@ and uvm_fault reported by fcambus@.
I will be poking further at this out of tree.

7 years agoadd support for -msave-args in gcc on amd64
dlg [Thu, 27 Apr 2017 00:02:03 +0000 (00:02 +0000)]
add support for -msave-args in gcc on amd64

i got sick of not having arguments in ddb stack traces on amd64,
which is because amd64 passes arguments in registers, and it's
impossible to figure out where they go without dwarf info, and when
you have dwarf info it is complicated.

solaris has a simple solution for this. they tweaked their compilers
to accept an -msave-args option which makes functions store their
arguments on the stack, while maintaining compatability with the
System V AMD64 ABI. tools (eg, ddb) can then look at the stack to
get access to function arguments in traces.

this ports their changes to gcc 3 to our gcc.

ok deraadt@

7 years agosilence clang warnings: add an extra pair of parentheses and stop
naddy [Wed, 26 Apr 2017 21:25:43 +0000 (21:25 +0000)]
silence clang warnings: add an extra pair of parentheses and stop
passing empty format strings to printf-family functions; ok millert@

7 years agoIn -C mode, process all checksums that match the specified file(s),
millert [Wed, 26 Apr 2017 21:10:42 +0000 (21:10 +0000)]
In -C mode, process all checksums that match the specified file(s),
not just the first one that matches.  OK deraadt@

7 years agoremove some repetitive text chunks; ok krw
jmc [Wed, 26 Apr 2017 21:07:21 +0000 (21:07 +0000)]
remove some repetitive text chunks; ok krw

7 years agoPerform H-FSC root queue allocation in the kernel
mikeb [Wed, 26 Apr 2017 15:50:59 +0000 (15:50 +0000)]
Perform H-FSC root queue allocation in the kernel

Since only leaf queues can have packets assigned to them,
H-FSC requires the user specified root queue to have a
parent.  To simplify userland tools and the configuration
interface, the kernel can be leveraged to set it up.

ok henning

7 years agoInstall arm64 manpages: eeprom(8), MAKEDEV(8). ok phessler kettenis
sthen [Wed, 26 Apr 2017 14:53:11 +0000 (14:53 +0000)]
Install arm64 manpages: eeprom(8), MAKEDEV(8).  ok phessler kettenis

7 years agoRemove extraneous ", NULL" in the assignment of msgstr which was
millert [Wed, 26 Apr 2017 13:14:28 +0000 (13:14 +0000)]
Remove extraneous ", NULL" in the assignment of msgstr which was
leftover from when msg_cat() was removed.  From Anton Lindqvist

7 years agocope with IP address changes. before, we were trying to resend the msg
henning [Wed, 26 Apr 2017 10:42:38 +0000 (10:42 +0000)]
cope with IP address changes. before, we were trying to resend the msg
with the no-longer-available address over and over and over, requiring
iked to be restarted eventually. instead, on EADDRNOTAVAIL, schedule
SA deletion so a new one is set up shortly thereafter. ok reyk mikeb

7 years agoConvert some hardcoded function names in printfs into %s / __func__.
mlarkin [Wed, 26 Apr 2017 09:53:28 +0000 (09:53 +0000)]
Convert some hardcoded function names in printfs into %s / __func__.

Some of these pointed out by Michael W. Bombardieri, I went through
the file and fixed the rest.

No functional change.

7 years agoGet rid of convoluted macros in iwm's receive code path.
stsp [Wed, 26 Apr 2017 09:19:56 +0000 (09:19 +0000)]
Get rid of convoluted macros in iwm's receive code path.

Since we now sync the entire Rx DMA buffer the SYNC_RESP_STRUCT() macro which
was used to sync just parts of the buffer can be replaced by assignments.
The SYNC_RESP_PTR() macro was already unused so just remove it.
The ADVANCE_RXQ() macro was used just once so expand it in place.

7 years agoRemove the size argument of iwm_rx_addbuf() and use a constant internally.
stsp [Wed, 26 Apr 2017 08:19:09 +0000 (08:19 +0000)]
Remove the size argument of iwm_rx_addbuf() and use a constant internally.
All callers were passing the same constant IWM_RBUF_SIZE, which is derived
from the smallest RX buffer the hardware supports (4k). Smaller sizes don't
make sense anyway and larger sizes are not on our horizon for now.

7 years agoConvert some hardcoded function names in printfs into %s / __func__.
mlarkin [Wed, 26 Apr 2017 08:02:14 +0000 (08:02 +0000)]
Convert some hardcoded function names in printfs into %s / __func__.

Some of these pointed out by Michael W. Bombardieri, I went through
the file and fixed the rest.

No functional change.

7 years agoTell MiRA about the actual number of retries instead of pretending that
stsp [Wed, 26 Apr 2017 07:53:17 +0000 (07:53 +0000)]
Tell MiRA about the actual number of retries instead of pretending that
a frame which was retried N times was retried just once.
This punishes retry-heavy data rates even more and should reduce latency.
It does not seem to have a noticable effect on net throughput in my testing.
If this change causes throughput problems for anyone, let me know.
ok tb@ mlarkin@

7 years agoFix a typo relating to savefpu alignment
mlarkin [Wed, 26 Apr 2017 07:05:24 +0000 (07:05 +0000)]
Fix a typo relating to savefpu alignment

ok kettenis

7 years agotweak previous;
jmc [Wed, 26 Apr 2017 05:55:27 +0000 (05:55 +0000)]
tweak previous;

7 years agoWe are men of action. Lies do not become us.
beck [Wed, 26 Apr 2017 02:43:34 +0000 (02:43 +0000)]
We are men of action. Lies do not become us.

7 years agoHave libtool ignore 'static-libtool-libs'. Previously it was unsupported and
sthen [Tue, 25 Apr 2017 21:13:15 +0000 (21:13 +0000)]
Have libtool ignore 'static-libtool-libs'. Previously it was unsupported and
passed straight through to the compiler, which gcc doesn't mind, but clang is
unhappy about.  Suggestion/ok aja@

7 years agoUnbreak multiple interface configurations like trunk or carp. Ensure
rpe [Tue, 25 Apr 2017 19:24:05 +0000 (19:24 +0000)]
Unbreak multiple interface configurations like trunk or carp. Ensure
that the noglob option is disabled at the end of parse_hn_line()
and ifstart().

7 years agoUnbreak netstart for multiple inteface configurations like trunk
rpe [Tue, 25 Apr 2017 19:17:33 +0000 (19:17 +0000)]
Unbreak netstart for multiple inteface configurations like trunk
or carp. Ensure that the noglob option is disabled at the end of
parse_hn_line() and ifstart().

Reported by Christer Solskogen and Stefan Wollny, thanks!

7 years agoMake full width panes try to play more nicely with terminal copy and
nicm [Tue, 25 Apr 2017 18:30:29 +0000 (18:30 +0000)]
Make full width panes try to play more nicely with terminal copy and
paste by avoiding explicit line wraps if we think the terminal will wrap
anyway.

7 years agoWhen we write out the grid including escape sequences, an SGR 0 needs to
nicm [Tue, 25 Apr 2017 18:20:51 +0000 (18:20 +0000)]
When we write out the grid including escape sequences, an SGR 0 needs to
cause the colours to be written again. Also treat colours separately
from attributes so that RGB colours will work.

7 years agoTest listening on multiple syslog over TLS sockets.
bluhm [Tue, 25 Apr 2017 17:49:11 +0000 (17:49 +0000)]
Test listening on multiple syslog over TLS sockets.

7 years agoAllow syslogd(8) to listen on multiple addresses for incomming TLS
bluhm [Tue, 25 Apr 2017 17:45:50 +0000 (17:45 +0000)]
Allow syslogd(8) to listen on multiple addresses for incomming TLS
connections.  This expands the feature from UDP and TCP to syslog
over TLS.
input jmc@; OK millert@

7 years agoUse strtonum instead of strto{,u}ll for simpler and better overflow
tb [Tue, 25 Apr 2017 17:33:16 +0000 (17:33 +0000)]
Use strtonum instead of strto{,u}ll for simpler and better overflow
checking, make somaxconn and sominconn unsigned.

Issue reported by orge on freenode, thanks!
Input, patient explanations and ok deraadt, millert.

7 years agoGenerate randomized MAC addresses earlier to keep them across reboots.
reyk [Tue, 25 Apr 2017 16:38:23 +0000 (16:38 +0000)]
Generate randomized MAC addresses earlier to keep them across reboots.

OK deraadt@

7 years agoDo not update TERM into config file parsing has finished.
nicm [Tue, 25 Apr 2017 15:35:10 +0000 (15:35 +0000)]
Do not update TERM into config file parsing has finished.

7 years agoBlock the initial client if there is one until the configuration file
nicm [Tue, 25 Apr 2017 14:46:23 +0000 (14:46 +0000)]
Block the initial client if there is one until the configuration file
has finished loading.

7 years agosome ip6 setup notes; help/ok bluhm
jmc [Tue, 25 Apr 2017 13:52:08 +0000 (13:52 +0000)]
some ip6 setup notes; help/ok bluhm

7 years agoif-shell doesn't need to queue its error message into a callback, and in
nicm [Tue, 25 Apr 2017 11:49:35 +0000 (11:49 +0000)]
if-shell doesn't need to queue its error message into a callback, and in
fact it can't do so because the item it was working with will have been
freed. Reported by Daniel Hahler.

7 years agospacing
reyk [Tue, 25 Apr 2017 06:44:35 +0000 (06:44 +0000)]
spacing

7 years agoContinue parser unification:
schwarze [Mon, 24 Apr 2017 23:06:09 +0000 (23:06 +0000)]
Continue parser unification:
* Make enum rofft an internal interface as enum roff_tok in "roff.h".
* Represent mdoc and man macros in enum roff_tok.
* Make TOKEN_NONE a proper enum value and use it throughout.
* Put the prologue macros first in the macro tables.
* Unify mdoc_macroname[] and man_macroname[] into roff_name[].

7 years ago'rtsol' has been removed from hostname.if(5) recently.
rpe [Mon, 24 Apr 2017 20:50:42 +0000 (20:50 +0000)]
'rtsol' has been removed from hostname.if(5) recently.

OK jmc@

7 years agoReplace 'rtsol' with 'autoconf' to reflect the change in the installer.
rpe [Mon, 24 Apr 2017 20:44:24 +0000 (20:44 +0000)]
Replace 'rtsol' with 'autoconf' to reflect the change in the installer.

7 years agoPrint bgp administrative shutdown messages (draft-ietf-idr-shutdown)
benno [Mon, 24 Apr 2017 20:35:35 +0000 (20:35 +0000)]
Print bgp administrative shutdown messages (draft-ietf-idr-shutdown)
From Job Snijders <job -AT instituut -DOT- net>, thanks!

ok canacar@, deraadt@

7 years agoIntroduce a new function parse_hn_line() that replaces the existing
rpe [Mon, 24 Apr 2017 20:31:48 +0000 (20:31 +0000)]
Introduce a new function parse_hn_line() that replaces the existing
hostname.if(5) parsing code in ifstart().
Add a -n option to netstart to only print the interface configuration
commands instead of executing them.
Add a HN_DIR variable, that points to the directory of the hostname.if
files (default /etc) that allows for future regression tests.

- add new parse_hn_line() function
- change ifstart()
  - rename $if to $_if
  - don't ifconfig or ifconfig create if -n option is used
  - replace hostname.if(5) parsing code with new parse_hn_line()
  - just print configuration commands if -n option is used
- autoconf now happens in ifstart(), remove ifv6autoconf()
- introduce HN_DIR variable for the hostname.if file location
- add handling of the -n option to only print config commands
- ensure -n is only used if interfaces are specified as parameters

Discussed with and positive feedback from many
'commit' deraadt@
OK sthen@

7 years agoIntroduce a new function parse_hn_line() that replaces the existing
rpe [Mon, 24 Apr 2017 20:27:59 +0000 (20:27 +0000)]
Introduce a new function parse_hn_line() that replaces the existing
hostname.if(5) parsing code in ifstart().

- change v6_config()
  - replace rtsol with autoconf in "IPv6 addres for <if>" question
  - write 'inet6 autoconf" to hostname.if file instead of rtsol
- add parse_hn_line() function with installer specific changes
  - skip shell command and bridge config lines
  - skip inet6 config (incl. dest lines) if there's no inet6 support
  - skip dhcp config if there's no dhclient
- change ifstart() to use parse_hn_line()

Discussed with and positive feedback from many
'commit' deraadt@
conditional OK sthen@ (untested)

7 years agoAdd support for Cortex-A12. Even though ARM rebranded these as Cortex-A17
kettenis [Mon, 24 Apr 2017 18:15:16 +0000 (18:15 +0000)]
Add support for Cortex-A12.  Even though ARM rebranded these as Cortex-A17
they have a different ID from "real" Cortex-A17 cores.

ok phessler@, patrick@

7 years agoa little more consistent text for previous; ok krw
jmc [Mon, 24 Apr 2017 18:05:52 +0000 (18:05 +0000)]
a little more consistent text for previous; ok krw

7 years agoAdd 'echo-client-id' statment, so that RFC 6842 behaviour can be
krw [Mon, 24 Apr 2017 14:58:36 +0000 (14:58 +0000)]
Add 'echo-client-id' statment, so that RFC 6842 behaviour can be
turned off for those clients and networks that find it impossible to
move past RFC 2131. Modelled on the same statement in recent ISC
versions, though we default to 'on' (a.k.a. RFC 6842) rather that
'off' (a.k.a. RFC 2131).

Problems reported by Bastien Durel (Xerox Phaser 6022 printer) and
Bryan Vyhmeister (Hon Hai Precision router) via misc@. Thanks!

7 years agoProcess fdt before iobus devices. Needed later when interrupt and
visa [Mon, 24 Apr 2017 14:10:19 +0000 (14:10 +0000)]
Process fdt before iobus devices. Needed later when interrupt and
MDIO controller drivers are attached using fdt.

7 years agoupdate for the llvm libs; help/ok kettenis
jmc [Mon, 24 Apr 2017 13:58:39 +0000 (13:58 +0000)]
update for the llvm libs; help/ok kettenis

7 years agoAdd an fdt blob for octeon systems whose firmware does not provide
visa [Mon, 24 Apr 2017 13:54:28 +0000 (13:54 +0000)]
Add an fdt blob for octeon systems whose firmware does not provide
a device-tree. It will be needed later when more device drivers
are attached using fdt.

OK kettenis@

7 years agoSince rev 1.1 ix has attempted to require a 64 bit BAR, the test for
jsg [Mon, 24 Apr 2017 13:20:46 +0000 (13:20 +0000)]
Since rev 1.1 ix has attempted to require a 64 bit BAR, the test for
this was wrong and was corrected at the end of last year in rev 1.140.
Before then a 64 bit BAR was not enforced as the test was wrong.

It turns out there exist 82598 parts which have a 32 bit BAR so change
the test to only require a memory BAR and not a 64 bit memory BAR.

Problem reported by Robert Blacquiere. ok mikeb@

7 years agoDocument support for 3168 and 8265 devices in iwm(4).
stsp [Mon, 24 Apr 2017 09:53:43 +0000 (09:53 +0000)]
Document support for 3168 and 8265 devices in iwm(4).

7 years agoDocument monitor mode support in iwm(4).
stsp [Mon, 24 Apr 2017 09:50:04 +0000 (09:50 +0000)]
Document monitor mode support in iwm(4).

7 years agoImplement monitor mode support for iwm(4).
stsp [Mon, 24 Apr 2017 09:48:42 +0000 (09:48 +0000)]
Implement monitor mode support for iwm(4).

The magic commands to capture raw 802.11 frames on e.g. channel 1 are:
ifconfig iwm0 mediaopt monitor chan 1
tcpdump -n -i iwm0 -y IEEE802_11_RADIO -s 4096 -w /tmp/iwm.pcap
ifconfig iwm0 -mediaopt monitor -chan

7 years agoImplement support for multiple packets per receive DMA buffer in iwm(4).
stsp [Mon, 24 Apr 2017 09:31:31 +0000 (09:31 +0000)]
Implement support for multiple packets per receive DMA buffer in iwm(4).

We can now clear SINGLE_FRAME_MSK in the receive configuration register.
Linux has not set this bit in years, and Dragonfly stopped setting it
in commit b5eb43f0280bbcfd26af51cf5a4b8e8ff3590b67.
This is a prerequisite for monitor mode since the firmware seems to ignore
the SINGLE_FRAME_MSK bit while running in monitor mode and sends us garbage.

Tested on 7260, 7265, and 8260 devices.

7 years agoStop dumping registers on "normal" segmentation faults. Do print the contents
kettenis [Mon, 24 Apr 2017 08:09:13 +0000 (08:09 +0000)]
Stop dumping registers on "normal" segmentation faults.  Do print the contents
of the esr_el1 register for unhandled userland exceptions (and continue to
dump the registers in that case).

Handle breakpoint traps and make sure we call refreshcreds() on userland traps.

ok jsg@

7 years agoDon't overflow the allocated DHCP response buffer.
reyk [Mon, 24 Apr 2017 07:14:27 +0000 (07:14 +0000)]
Don't overflow the allocated DHCP response buffer.

When copying the DHCP/BOOTP response into the allocated buffer, use
the calculated response length instead of the maximum size of a DHCP
packet.  The response length was verified correctly but the very last
step missed to use it.

Found by Caspar Schutijser and malloc.conf(5) "S"

OK deraadt@

7 years agoFix configuration of ASN1_DN IDs.
reyk [Mon, 24 Apr 2017 07:07:25 +0000 (07:07 +0000)]
Fix configuration of ASN1_DN IDs.

Public key authentication uses public key files that are stored in the
/etc/iked/pubkeys/ directory where the IKE IDs are encoded as filenames.
This does not simply work with ASN1_DNs where the IDs include slashes
and other special characters. Instead of breaking and failing when an
ASN1_DN is configured, simply skip the public key lookup but allow
to use it with certificates or PSKs.

Reported and fix tested by Igor V. Gubenko - Thanks.

7 years agoFix typos. From Michael W. Bombardieri <mb at ii.net>. Thanks!
ratchov [Mon, 24 Apr 2017 06:47:41 +0000 (06:47 +0000)]
Fix typos. From Michael W. Bombardieri <mb at ii.net>. Thanks!

7 years agoPrint usage on stderr as do other programs. From Michael
ratchov [Mon, 24 Apr 2017 06:45:56 +0000 (06:45 +0000)]
Print usage on stderr as do other programs. From Michael
Bombardieri <mb at ii.net>, thanks.

7 years agoHandle unequal numbers of Tx and Rx streams in MiRA.
stsp [Sun, 23 Apr 2017 21:04:55 +0000 (21:04 +0000)]
Handle unequal numbers of Tx and Rx streams in MiRA.
Problem reported by Colton Lewis on misc@
ok tb@

7 years agoOnly use ED for clear screen if at the bottom, same as earlier fix to
nicm [Sun, 23 Apr 2017 18:13:24 +0000 (18:13 +0000)]
Only use ED for clear screen if at the bottom, same as earlier fix to
clear end of screen.

7 years agoSome of the LOG_NOTICE messages from PF were seen in normal operations
sthen [Sun, 23 Apr 2017 11:37:11 +0000 (11:37 +0000)]
Some of the LOG_NOTICE messages from PF were seen in normal operations
with certain rulesets and excessively noisy; move them to LOG_INFO (which was
previously unused).  ok benno@