openbsd
5 years agouse a specific and more relevant status message for failed smarthost
eric [Mon, 19 Aug 2019 15:42:24 +0000 (15:42 +0000)]
use a specific and more relevant status message for failed smarthost
resolution.

ok gilles@

5 years agodocument X509_INFO_new(3) and X509_INFO_free(3)
schwarze [Mon, 19 Aug 2019 13:52:53 +0000 (13:52 +0000)]
document X509_INFO_new(3) and X509_INFO_free(3)

5 years agoadd 6.7 syspatch key
robert [Mon, 19 Aug 2019 13:50:38 +0000 (13:50 +0000)]
add 6.7 syspatch key

5 years agodocument ECDH_compute_key(3) and ECDH_size(3);
schwarze [Mon, 19 Aug 2019 13:08:26 +0000 (13:08 +0000)]
document ECDH_compute_key(3) and ECDH_size(3);
feedback and OK tb@

5 years ago-Pcdrom is nonsensical now, so error out properly on it.
espie [Mon, 19 Aug 2019 12:25:40 +0000 (12:25 +0000)]
-Pcdrom is nonsensical now, so error out properly on it.

Prompted and okay danj@

5 years agosilently add -DDONTLOG
espie [Mon, 19 Aug 2019 12:11:04 +0000 (12:11 +0000)]
silently add -DDONTLOG

5 years agoadd a special flag -DDONTLOG so that regress test won't spam /var/log/messages
espie [Mon, 19 Aug 2019 12:09:07 +0000 (12:09 +0000)]
add a special flag -DDONTLOG so that regress test won't spam /var/log/messages

5 years agodefine CONFIG_X86* in autoconf.h as needed and reduce diff to linux
jsg [Mon, 19 Aug 2019 09:34:13 +0000 (09:34 +0000)]
define CONFIG_X86* in autoconf.h as needed and reduce diff to linux
ok kettenis@

5 years agomatch on r40 ahci
jsg [Mon, 19 Aug 2019 07:27:11 +0000 (07:27 +0000)]
match on r40 ahci

tested by and ok jsing@ ok kettenis@

5 years agoEnable TCP and UDP checksum offloading by default for ix(4).
jan [Mon, 19 Aug 2019 07:07:35 +0000 (07:07 +0000)]
Enable TCP and UDP checksum offloading by default for ix(4).

ok deraadt@

5 years agoAdd missing OHCI clocks for Allwinner R40.
kettenis [Mon, 19 Aug 2019 06:13:44 +0000 (06:13 +0000)]
Add missing OHCI clocks for Allwinner R40.

tested by jsing@

5 years agoEvery "goto bad" in scsi_probedev() deserves a SC_DEBUG().
krw [Sun, 18 Aug 2019 23:58:24 +0000 (23:58 +0000)]
Every "goto bad" in scsi_probedev() deserves a SC_DEBUG().

5 years agoTweak cross references, in particular making sure that
schwarze [Sun, 18 Aug 2019 21:44:10 +0000 (21:44 +0000)]
Tweak cross references, in particular making sure that
all CMS pages are linked to CMS_ContentInfo_new(3) both ways
and that closely related pages reference each other.

5 years agoversion field was reserved but not discussed, this is done now
gilles [Sun, 18 Aug 2019 16:52:02 +0000 (16:52 +0000)]
version field was reserved but not discussed, this is done now

ok millert@ martijn@

5 years agoRename 'link' to 'link0' as it refers to target 0 only.
krw [Sun, 18 Aug 2019 16:21:32 +0000 (16:21 +0000)]
Rename 'link' to 'link0' as it refers to target 0 only.

5 years agoIncrease timeout used when waiting for the Rx FIFO to fill up when in polled
kettenis [Sun, 18 Aug 2019 15:52:45 +0000 (15:52 +0000)]
Increase timeout used when waiting for the Rx FIFO to fill up when in polled
mode.

ok jcs@, mlarkin@

5 years agoIn polled mode, wait on STOP detected bit to be set in the interrupt status
kettenis [Sun, 18 Aug 2019 15:51:18 +0000 (15:51 +0000)]
In polled mode, wait on STOP detected bit to be set in the interrupt status
register like we do for non-polled mode.  This seems to increase the
reliability of i2c transfers on the controller integrated on the Ampare
eMAG processor.

ok jcs@, mlarkin@

5 years agoImplement a few Linux compat ACPI interfaces and enable the ACPI support
kettenis [Sun, 18 Aug 2019 13:11:47 +0000 (13:11 +0000)]
Implement a few Linux compat ACPI interfaces and enable the ACPI support
code in radeon(4) and amdgpu(4).

ok jsg@

5 years agominor cleanup:
schwarze [Sun, 18 Aug 2019 13:02:37 +0000 (13:02 +0000)]
minor cleanup:
* avoid jumping back and forth between use cases
* delete duplicate information
* and minor wording improvements

5 years agominor cleanup:
schwarze [Sun, 18 Aug 2019 12:06:51 +0000 (12:06 +0000)]
minor cleanup:
* add the missing STANDARDS sections
* mark up ASN.1 type names
* GOST does not need an ENGINE in LibreSSL, so don't use it as an example
* and minor wording improvements and typo fixes

5 years agominor cleanup:
schwarze [Sun, 18 Aug 2019 10:54:57 +0000 (10:54 +0000)]
minor cleanup:
* mark up ASN.1 type and field names
* move the RFC reference to STANDARDS
* and minor wording improvements

5 years agosome cleanup:
schwarze [Sun, 18 Aug 2019 10:19:17 +0000 (10:19 +0000)]
some cleanup:
* do not jump back and forth among functions
* show data type - NID correspondance in a table
* make the difference between content type and embedded content clearer
* add the missing STANDARDS section
* mark up ASN.1 type names
* remove some text that says nothing
* and minor wording improvements

5 years agoWhen activating or detaching a target don't search the scsi_link SLIST
krw [Sun, 18 Aug 2019 02:43:52 +0000 (02:43 +0000)]
When activating or detaching a target don't search the scsi_link SLIST
for each target:lun. Just travese the SLIST once taking care of relevant
scsi_link's as they are encountered.

ok jmatthew@

5 years agosc_buswidth field in struct scsi_link is redundant. Just use
krw [Sun, 18 Aug 2019 00:58:54 +0000 (00:58 +0000)]
sc_buswidth field in struct scsi_link is redundant. Just use
adapter_link->adapter_buswidth, which supplied the value for
sc_buswidth and is never changed.

5 years agoNuke some unused variables, tweak some declarations and
krw [Sat, 17 Aug 2019 15:31:41 +0000 (15:31 +0000)]
Nuke some unused variables, tweak some declarations and
variable names into a consistant idiom.

5 years agosome cleanup:
schwarze [Sat, 17 Aug 2019 15:23:26 +0000 (15:23 +0000)]
some cleanup:
* add the missing STANDARDS section
* more precision below RETURN VALUES
* simplify some overly verbose text
* mark up ASN.1 type names
* and minor wording improvements and typo fixes

5 years agono need to escape "-", generally speaking; instances remain in EXAMPLES;
jmc [Sat, 17 Aug 2019 14:43:21 +0000 (14:43 +0000)]
no need to escape "-", generally speaking; instances remain in EXAMPLES;
(see ingo's notes in mandoc_char.7, if explanation neccessary)

ok schwarze gilles

5 years agominor cleanup:
schwarze [Sat, 17 Aug 2019 14:41:01 +0000 (14:41 +0000)]
minor cleanup:
* add the missing STANDARDS section
* mark up ASN.1 type names
* avoid some repetitions
* make some lists more palatable in -column form
* and minor wording improvements and typo fixes

5 years agosignal handlers should not call exit() due to possibility of reentering
deraadt [Sat, 17 Aug 2019 14:25:06 +0000 (14:25 +0000)]
signal handlers should not call exit() due to possibility of reentering
libc (stdio etc), instead do the unlink tasks then call _exit() instead
ok millert

5 years agodrm/i915: Fix wrong escape clock divisor init for GLK
jsg [Sat, 17 Aug 2019 08:27:43 +0000 (08:27 +0000)]
drm/i915: Fix wrong escape clock divisor init for GLK

From Stanislav Lisovskiy
edc388566a1d25f88e631bd8170462d4f86c5a24 in linux 4.19.y/4.19.67
73a0ff0b30af79bf0303d557eb82f1d1945bb6ee in mainline linux

5 years agodrm: silence variable 'conn' set but not used
jsg [Sat, 17 Aug 2019 08:25:28 +0000 (08:25 +0000)]
drm: silence variable 'conn' set but not used

From Qian Cai
991c4756be69417f720734ceb32a22071e0aa0af in linux 4.19.y/4.19.67
bbb6fc43f131f77fcb7ae8081f6d7c51396a2120 in mainline linux

5 years agodrm/amd/display: Increase size of audios array
jsg [Sat, 17 Aug 2019 08:23:28 +0000 (08:23 +0000)]
drm/amd/display: Increase size of audios array

From Tai Man
8d641499bf969a284161d2682b71669c96389773 in linux 4.19.y/4.19.67
7352193a33dfc9b69ba3bf6a8caea925b96243b1 in mainline linux

5 years agodrm/amd/display: Only enable audio if speaker allocation exists
jsg [Sat, 17 Aug 2019 08:20:59 +0000 (08:20 +0000)]
drm/amd/display: Only enable audio if speaker allocation exists

From Alvin Lee
f9420bfa29f9ece0b02a5435ae95f1a48bc97723 in linux 4.19.y/4.19.67
6ac25e6d5b2fbf251e9fa2f4131d42c815b43867 in mainline linux

5 years agodrm/amd/display: Fix dc_create failure handling and 666 color depths
jsg [Sat, 17 Aug 2019 08:17:34 +0000 (08:17 +0000)]
drm/amd/display: Fix dc_create failure handling and 666 color depths

From Julian Parkin
3998e684463a7fa1721c171172ca085978d03a00 in linux 4.19.y/4.19.67
0905f32977268149f06e3ce6ea4bd6d374dd891f in mainline linux

5 years agodrm/amd/display: use encoder's engine id to find matched free audio device
jsg [Sat, 17 Aug 2019 08:14:34 +0000 (08:14 +0000)]
drm/amd/display: use encoder's engine id to find matched free audio device

From Tai Man
e7a8a794109c07e0b8d7bd55fbfcb3082991626a in linux 4.19.y/4.19.67
74eda776d7a4e69ec7aa1ce30a87636f14220fbb in mainline linux

5 years agodrm/amd/display: Wait for backlight programming completion in set backlight level
jsg [Sat, 17 Aug 2019 08:11:32 +0000 (08:11 +0000)]
drm/amd/display: Wait for backlight programming completion in set backlight level

From Sivapiriyan Kumarasamy
2a5e21adc71b46beec9232cc6418676fd2255bc5 in linux 4.19.y/4.19.67
c7990daebe71d11a9e360b5c3b0ecd1846a3a4bb in mainline linux

5 years agochange drm memory barriers to be closer to what linux does on
jsg [Sat, 17 Aug 2019 06:07:22 +0000 (06:07 +0000)]
change drm memory barriers to be closer to what linux does on
amd64 and i386

ok kettenis@

5 years agoMove play blocks count from the audio_softc to the audio_buf structure.
ratchov [Sat, 17 Aug 2019 05:04:56 +0000 (05:04 +0000)]
Move play blocks count from the audio_softc to the audio_buf structure.

As the audio_buf structure is per-direction, this makes play and
recording code similar. No behavior change.

5 years agoMove block size and block count calculations in their own routines.
ratchov [Sat, 17 Aug 2019 04:57:52 +0000 (04:57 +0000)]
Move block size and block count calculations in their own routines.

Makes the code easier to read, no behabior change.

5 years agosort; OK deraadt@
schwarze [Fri, 16 Aug 2019 22:23:41 +0000 (22:23 +0000)]
sort; OK deraadt@

5 years agosync
deraadt [Fri, 16 Aug 2019 20:11:07 +0000 (20:11 +0000)]
sync

5 years agocheck that software de/encrypt is possible: under hardware
procter [Fri, 16 Aug 2019 19:53:32 +0000 (19:53 +0000)]
check that software de/encrypt is possible: under hardware
offload, it needn't be. the stack must otherwise rely on every
offloading driver correctly handling all frames governed by a
given key.
ok stsp@

5 years agosync
deraadt [Fri, 16 Aug 2019 19:28:14 +0000 (19:28 +0000)]
sync

5 years agoSend snmpctl(8) to the great bitbucket in the sky. You've served us well.
martijn [Fri, 16 Aug 2019 17:03:55 +0000 (17:03 +0000)]
Send snmpctl(8) to the great bitbucket in the sky. You've served us well.

People are now expected to use snmp(1) for all their snmp client needs.

"Go ahead" deraadt@

5 years agolink to the new EC_KEY_METHOD_new(3) page
schwarze [Fri, 16 Aug 2019 16:20:23 +0000 (16:20 +0000)]
link to the new EC_KEY_METHOD_new(3) page
and mention a trap set by EC_KEY_copy(3)

5 years agodocument the EC_KEY_METHOD function table
schwarze [Fri, 16 Aug 2019 16:15:50 +0000 (16:15 +0000)]
document the EC_KEY_METHOD function table

5 years agodocument EC_KEY_get_ex_new_index(3), EC_KEY_get_ex_data(3),
schwarze [Fri, 16 Aug 2019 12:16:22 +0000 (12:16 +0000)]
document EC_KEY_get_ex_new_index(3), EC_KEY_get_ex_data(3),
and EC_KEY_set_ex_data(3)

5 years agoAdd explanation for the [IKE/ESP only] column of the transform table.
tobhe [Fri, 16 Aug 2019 12:11:07 +0000 (12:11 +0000)]
Add explanation for the [IKE/ESP only] column of the transform table.

Ok kn@

5 years agoAdd a flag to reverse sort in the various choose modes, from Benjamin
nicm [Fri, 16 Aug 2019 11:49:12 +0000 (11:49 +0000)]
Add a flag to reverse sort in the various choose modes, from Benjamin
Poirier in GitHub issue 1875.

5 years agosome cleanup:
schwarze [Fri, 16 Aug 2019 11:30:10 +0000 (11:30 +0000)]
some cleanup:
* do not jump back and forth among different arguments and flags
* add the missing STANDARDS section
* mark up ASN.1 type names
* and several wording improvements

5 years agoCall comma-separated lists as such to clarify semantics
kn [Fri, 16 Aug 2019 11:16:32 +0000 (11:16 +0000)]
Call comma-separated lists as such to clarify semantics

Options such as Ciphers take values that may be a list of ciphers;  the
complete list, not indiviual elements, may be prefixed with a dash or plus
character to remove from or append to the default list respectively.

Users might read the current text as if each elment took an optional prefix,
so tweak the wording from "values" to "list" to prevent such ambiguity for
all options supporting this semantics (those that provide a list of
available elements via "ssh -Q ...").

Input and OK jmc

5 years agominor cleanup:
schwarze [Fri, 16 Aug 2019 10:43:15 +0000 (10:43 +0000)]
minor cleanup:
* add the missing STANDARDS section
* mark up ASN.1 type names
* and minor wording improvements and typo fixes

5 years agogrid_view_delete_cells does need to clear, GitHub issue 1871.
nicm [Fri, 16 Aug 2019 08:52:25 +0000 (08:52 +0000)]
grid_view_delete_cells does need to clear, GitHub issue 1871.

5 years agoFix segfault in parser when specifying an invalid transform.
tobhe [Fri, 16 Aug 2019 07:42:13 +0000 (07:42 +0000)]
Fix segfault in parser when specifying an invalid transform.
For all transforms the error case only printed the error but
did not exit. YYERROR was added to exit gracefully instead of
segfaulting later.

ok benno@

5 years agoinclude sshbuf-misc.c in SRCS_BASE
djm [Fri, 16 Aug 2019 06:35:27 +0000 (06:35 +0000)]
include sshbuf-misc.c in SRCS_BASE

5 years agoswitch percent_expand() to use sshbuf instead of a limited fixed buffer;
djm [Fri, 16 Aug 2019 06:13:15 +0000 (06:13 +0000)]
switch percent_expand() to use sshbuf instead of a limited fixed buffer;
ok markus@

5 years agoknock out useless Pp;
jmc [Fri, 16 Aug 2019 05:33:50 +0000 (05:33 +0000)]
knock out useless Pp;

5 years agoifq_hdatalen should keep the mbuf it's looking at, not leak it.
dlg [Fri, 16 Aug 2019 04:09:02 +0000 (04:09 +0000)]
ifq_hdatalen should keep the mbuf it's looking at, not leak it.

ie, use ifq_deq_rollback after looking at the head mbuf instead of
ifq_deq_commit.

this is used in tun/tap, where it had the effect that you'd get the
datalen for the packet, and then when you try to read that many
bytes it had gone. cool and normal.

this was found by a student who was trying to do just that. i've
always just read the packet into a large buffer.

5 years agoSTP works fine on etherip(4), it was gif(4) that lacked a MAC address.
dlg [Fri, 16 Aug 2019 01:55:21 +0000 (01:55 +0000)]
STP works fine on etherip(4), it was gif(4) that lacked a MAC address.

getting etherip(4) to talk to a switch so i could verify this was
pretty fun.

remi@ pointed this bit of the manpage out somewhere, but it didn't
seem right anymore since etherip(4) was split out of gif(4). these
days if we have an interface that can be added to a bridge, we try
to make it look like an actual ethernet interface, which includes
having a mac address. we don't really have interfaces that change
modes and have these edge cases anymore.

5 years agopfctl_reset() must set syncookies settings back to default
sashan [Thu, 15 Aug 2019 18:44:53 +0000 (18:44 +0000)]
pfctl_reset() must set syncookies settings back to default
(bug found and fix tested by Jesper Wallin)

OK deraadt OK kn

5 years agominor cleanup:
schwarze [Thu, 15 Aug 2019 15:46:51 +0000 (15:46 +0000)]
minor cleanup:
* add the missing STANDARDS section
* mark up ASN.1 type names
* and minor wording improvements

5 years agoNMEA 4.1 has one more field in the RMC message; ok kettenis@
otto [Thu, 15 Aug 2019 15:08:14 +0000 (15:08 +0000)]
NMEA 4.1 has one more field in the RMC message; ok kettenis@

5 years agosome cleanup:
schwarze [Thu, 15 Aug 2019 11:42:53 +0000 (11:42 +0000)]
some cleanup:
* do not jump back and forth between different arguments
* display the flags in a proper list
* add the missing STANDARDS section
* mark up ASN.1 type names
* and minor wording improvements

5 years agoauich_freemem() returns void. while here fix a typo in an error message. audio playba...
miko [Thu, 15 Aug 2019 09:38:56 +0000 (09:38 +0000)]
auich_freemem() returns void. while here fix a typo in an error message. audio playback still works under virtualbox. ok ratchov@

5 years agodocument but do not recommend EVP_Cipher(3);
schwarze [Thu, 15 Aug 2019 09:36:29 +0000 (09:36 +0000)]
document but do not recommend EVP_Cipher(3);
jsing@ provided crucial help regarding the content;
tweaks and OK tb@

5 years agoSerialize access to the vnode pointers associated with acct(2) system
anton [Thu, 15 Aug 2019 07:29:21 +0000 (07:29 +0000)]
Serialize access to the vnode pointers associated with acct(2) system
accounting. Prevents a race where the acct thread and the acct(2)
syscall both tries to close a vnode.

ok visa@

Reported-by: syzbot+bf2ac4d4fa9ee92903b8@syzkaller.appspotmail.com
5 years agoRevert previous. There were some users of the quirky behaviour that were
martijn [Thu, 15 Aug 2019 06:11:18 +0000 (06:11 +0000)]
Revert previous. There were some users of the quirky behaviour that were
missed during code scan.

5 years agoAdd test case for ber_scanf_elements.
rob [Wed, 14 Aug 2019 23:17:08 +0000 (23:17 +0000)]
Add test case for ber_scanf_elements.

Significant tweak by and ok martijn@

5 years agorevert the inet6 format diff, regression observed by florian@
gilles [Wed, 14 Aug 2019 21:11:25 +0000 (21:11 +0000)]
revert the inet6 format diff, regression observed by florian@

5 years agoscsi_[add|remove]_link() are local functions so move their
krw [Wed, 14 Aug 2019 21:02:02 +0000 (21:02 +0000)]
scsi_[add|remove]_link() are local functions so move their
declarations.

5 years agoWhitespace nit. Add {} around body of SLIST_FOREACH().
krw [Wed, 14 Aug 2019 20:53:55 +0000 (20:53 +0000)]
Whitespace nit. Add {} around body of SLIST_FOREACH().

5 years agoExplain how to take over syspatch builds from another machine.
tb [Wed, 14 Aug 2019 20:30:19 +0000 (20:30 +0000)]
Explain how to take over syspatch builds from another machine.

ok ajacoutot robert

5 years agoClean up clean_old():
naddy [Wed, 14 Aug 2019 19:32:41 +0000 (19:32 +0000)]
Clean up clean_old():
* Remove syspatch files from the installed system and not the ramdisk.
* Directly try to remove the old files without superfluous check.
* Use an extended glob to match /usr/lib/gcc-lib contents that don't
belong to this OS version.
* For /usr/lib/clang, set an array to the expansion of the glob, which is
an alphabetically sorted list, drop the last element, and rm the rest.

ok tb@

5 years agoFix argument list for ber_set_writecallback
martijn [Wed, 14 Aug 2019 19:08:03 +0000 (19:08 +0000)]
Fix argument list for ber_set_writecallback

OK claudio@

5 years agowonky comma;
jmc [Wed, 14 Aug 2019 17:36:04 +0000 (17:36 +0000)]
wonky comma;

5 years agoDocument that ber_scanf_elements' 'p' and 't' attribute don't eat the
martijn [Wed, 14 Aug 2019 17:21:42 +0000 (17:21 +0000)]
Document that ber_scanf_elements' 'p' and 't' attribute don't eat the
current ber element.

OK claudio@
Seems sensible to deraadt@

5 years agoMake ber_scanf_elements's 'e' attribute eat the element.
martijn [Wed, 14 Aug 2019 17:20:41 +0000 (17:20 +0000)]
Make ber_scanf_elements's 'e' attribute eat the element.
Right now all consumers use 'e' at the end of the list, so no regressions
should be introduced.

OK claudio@
Seems sensible to deraadt@

5 years agomake usage() and SYNOPSIS more alike
deraadt [Wed, 14 Aug 2019 14:40:23 +0000 (14:40 +0000)]
make usage() and SYNOPSIS more alike
discussed with martijn

5 years agodocument EVP_Digest(3)
schwarze [Wed, 14 Aug 2019 14:05:32 +0000 (14:05 +0000)]
document EVP_Digest(3)

5 years agoTweak some comments.
krw [Wed, 14 Aug 2019 12:56:20 +0000 (12:56 +0000)]
Tweak some comments.

5 years agoRework the way ribs are stored in the RDE. Instead of a flat array that
claudio [Wed, 14 Aug 2019 11:57:21 +0000 (11:57 +0000)]
Rework the way ribs are stored in the RDE. Instead of a flat array that
gets enlarged use an array of pointers, so pointers to struct rib entries
remain valid after adding new RIBs. Also remove the global ribs pointer
and rib_valid() since they are no longer used since all the code uses now
rib_byid() instead.
OK benno@

5 years agosome cleanup:
schwarze [Wed, 14 Aug 2019 11:54:23 +0000 (11:54 +0000)]
some cleanup:
* avoid repetitions and jumping back and forth between the functions
* more precision regarding which ASN.1 types and fields are involved
* mark up the ASN.1 type and field names
* explain CMS_REUSE_DIGEST more precisely
* move the discussion of attributes to CMS_NOATTR where it belongs
* GOST does not need an ENGINE in LibreSSL, so don't use it as an example
* add the missing STANDARDS section
* and minor wording improvements

5 years agoDefault to previous search string for search-forward and
nicm [Wed, 14 Aug 2019 10:02:24 +0000 (10:02 +0000)]
Default to previous search string for search-forward and
search-backward, from Leah Neukirchen.

5 years agoRevert a change I did not mean to commit yet.
nicm [Wed, 14 Aug 2019 09:59:43 +0000 (09:59 +0000)]
Revert a change I did not mean to commit yet.

5 years agoAdd -Z flag to rotate-window, select-pane, swap-pane, switch-client to
nicm [Wed, 14 Aug 2019 09:58:31 +0000 (09:58 +0000)]
Add -Z flag to rotate-window, select-pane, swap-pane, switch-client to
preserve zoomed state. GitHub issue 1839.

5 years agoFix NAT traversal detection bug when "local" option is not explicitly
tobhe [Wed, 14 Aug 2019 08:35:46 +0000 (08:35 +0000)]
Fix NAT traversal detection bug when "local" option is not explicitly
set.

ok patrick@

5 years agoread() returns -1 on failure
asou [Wed, 14 Aug 2019 08:31:00 +0000 (08:31 +0000)]
read() returns -1 on failure

ok yasuoka@

5 years agoAdd ikev2_ikesa_info dummy definition for parser test.
tobhe [Wed, 14 Aug 2019 08:03:42 +0000 (08:03 +0000)]
Add ikev2_ikesa_info dummy definition for parser test.

5 years agoThere is no longer a reason to use two structs for RIBs where one is part
claudio [Wed, 14 Aug 2019 07:39:04 +0000 (07:39 +0000)]
There is no longer a reason to use two structs for RIBs where one is part
of the other. Just merge struct rib_desc into struct rib. Makes code simpler.
OK benno@

5 years agoImprove the error message when supplying an invalid template to vmctl
anton [Wed, 14 Aug 2019 07:34:49 +0000 (07:34 +0000)]
Improve the error message when supplying an invalid template to vmctl
start. Favoring 'invalid template' over 'permission denied' should give
the user a better hint on what went wrong.

ok kn@ mlarkin@

5 years agomove CONFIG_DRM_FBDEV_EMULATION and CONFIG_BACKLIGHT_CLASS_DEVICE
jsg [Wed, 14 Aug 2019 05:32:56 +0000 (05:32 +0000)]
move CONFIG_DRM_FBDEV_EMULATION and CONFIG_BACKLIGHT_CLASS_DEVICE
defines to kconfig.h with most of the others

5 years agoMake sure that ber in ber_scanf_elements is not NULL before parsing format
martijn [Wed, 14 Aug 2019 04:48:13 +0000 (04:48 +0000)]
Make sure that ber in ber_scanf_elements is not NULL before parsing format
where ber is utilized. This also allows us to remove the ber->be_next
check, which can cause weird behaviour, because a NULL be_next would result
in parsing the last element twice.

OK claudio@ on previous version
OK rob@

5 years agoFix ber_scanf_elements for traphandler:
martijn [Wed, 14 Aug 2019 04:43:32 +0000 (04:43 +0000)]
Fix ber_scanf_elements for traphandler:
- pdu header has 3 elements, not 4
- additional varbinds are optional.

This is needed to make ber_scanf_elements stricter.

Note that people using "trap handle" in their snmpd.conf and expect a trap
without additional varbinds to show the trapoid to appear twice will have
to adjust their "command".

OK rob@

5 years agoenable amdgpu on amd64 requested by a few people
jsg [Wed, 14 Aug 2019 00:08:46 +0000 (00:08 +0000)]
enable amdgpu on amd64 requested by a few people

there are still issues:
- running piglit will quickly trigger a gpu hang
- hardware cursor (the default) isn't always correct,
  can be worked around by specifying SWcursor
- suspend/resume doesn't work

ok patrick@ kettenis@

5 years agosome cleanup:
schwarze [Tue, 13 Aug 2019 23:54:16 +0000 (23:54 +0000)]
some cleanup:
* state the common, general purpose up front
* more precision regarding which ASN.1 types and fields are involved
* mark up the ASN.1 type and field names
* add the missing STANDARDS section
* and minor wording improvements

5 years agoImplement backlight control for amdgpu(4). This allows me to set
patrick [Tue, 13 Aug 2019 22:02:28 +0000 (22:02 +0000)]
Implement backlight control for amdgpu(4).  This allows me to set
the backlight of my X395 using wsconsctl(8).  Unfortunately the
backlight hotkeys don't yet allow adjusting the brightness.

ok jsg@ kettenis@

5 years agoconfig -eu uses a second copy of the nlist array[], which should never
deraadt [Tue, 13 Aug 2019 21:36:18 +0000 (21:36 +0000)]
config -eu uses a second copy of the nlist array[], which should never
become different or the #define's for the slots could diverge.  To avoid
this, make one a copy of the other at runtime.

5 years agoReorganize the ipmi(4) code a bit in anticipation of adding SSIF support:
kettenis [Tue, 13 Aug 2019 18:31:23 +0000 (18:31 +0000)]
Reorganize the ipmi(4) code a bit in anticipation of adding SSIF support:
- Put function prototypes in more logical places.
- Inroduce a ipmi_attach_common() function.
- Move all the SMBIOS related code to the end of the file and only
  compile it in on amd64 & i386.

ok jmatthew@ and deraadt@

5 years agouse proper format to display inet6 addresses in logs as well as to store in
gilles [Tue, 13 Aug 2019 16:02:33 +0000 (16:02 +0000)]
use proper format to display inet6 addresses in logs as well as to store in
disk envelopes. smtpd used the Received and helo response format that isn't
valid in a relay url and looks wrong in logs.

spotted and tested ok@ by semarie

5 years agoazalia: add quirk for speaker routing on ThinkPad X1C7
jcs [Tue, 13 Aug 2019 15:28:12 +0000 (15:28 +0000)]
azalia: add quirk for speaker routing on ThinkPad X1C7

The X1C7 has 4 speakers and the speaker2 set needs to be routed to a
different DAC.

Also add the Intel 300 Series HDA to the list of devices on which to
enable snooping.