openbsd
5 months agoremove switch_exit() prototypes, replaced by sched_exit()
jsg [Tue, 21 May 2024 23:16:06 +0000 (23:16 +0000)]
remove switch_exit() prototypes, replaced by sched_exit()

5 months agoRefactor vmx(4) receive offloading.
jan [Tue, 21 May 2024 19:49:06 +0000 (19:49 +0000)]
Refactor vmx(4) receive offloading.

 - avoid combination of M_TCP_CSUM_IN_OK and M_UDP_CSUM_IN_OK
 - remove useless letoh32() calls
 - simplify the IPv4 checksum flag check

ok bluhm@

5 months agoRemove obsolete comment, and fix some spacing.
mglocker [Tue, 21 May 2024 18:19:22 +0000 (18:19 +0000)]
Remove obsolete comment, and fix some spacing.

5 months agoOnly enable BUS_DMA_64BIT if the controller supports it.
mglocker [Tue, 21 May 2024 16:09:00 +0000 (16:09 +0000)]
Only enable BUS_DMA_64BIT if the controller supports it.

5 months agoInform user land when vltime / pltime changes.
florian [Tue, 21 May 2024 15:12:25 +0000 (15:12 +0000)]
Inform user land when vltime / pltime changes.

Do not send a RTM_CHGADDRATTR route message when the address is
tentative because we will send one when DAD finishes.

To be used by rad(8) shortly.

OK bluhm

5 months agoMake igc(4) and ix(4) struct names unique.
bluhm [Tue, 21 May 2024 11:19:39 +0000 (11:19 +0000)]
Make igc(4) and ix(4) struct names unique.

In ddb show struct gets confused if the same struct name exists
twice.  Add a prefix to tx_ring and rx_ring in drivers.

OK miod@ jan@

5 months agoremove prototypes with no matching function
jsg [Tue, 21 May 2024 11:13:08 +0000 (11:13 +0000)]
remove prototypes with no matching function

5 months agoregen
jsg [Tue, 21 May 2024 07:14:20 +0000 (07:14 +0000)]
regen

5 months agomove entry to be ordered by device id
jsg [Tue, 21 May 2024 07:13:29 +0000 (07:13 +0000)]
move entry to be ordered by device id

5 months agoregen
jsg [Tue, 21 May 2024 07:04:36 +0000 (07:04 +0000)]
regen

5 months agomove some entries to be ordered by device id
jsg [Tue, 21 May 2024 07:03:55 +0000 (07:03 +0000)]
move some entries to be ordered by device id

5 months agoremove prototypes with no matching function; ok ratchov@
jsg [Tue, 21 May 2024 06:07:06 +0000 (06:07 +0000)]
remove prototypes with no matching function; ok ratchov@

5 months agoaucat.1: Fix needless line breaks in the -h and MIDI sections.
ratchov [Tue, 21 May 2024 05:09:29 +0000 (05:09 +0000)]
aucat.1: Fix needless line breaks in the -h and MIDI sections.

Diff from Jan Stary <hans at stare.cz>, thanks!

ok jmc

5 months agoremove prototypes with no matching function and externs with no var
jsg [Tue, 21 May 2024 05:00:47 +0000 (05:00 +0000)]
remove prototypes with no matching function and externs with no var
partly checked by millert@

5 months agoremove extern with no matching var
jsg [Mon, 20 May 2024 23:20:29 +0000 (23:20 +0000)]
remove extern with no matching var

5 months agoremove prototypes with no matching function
jsg [Mon, 20 May 2024 23:17:10 +0000 (23:17 +0000)]
remove prototypes with no matching function

5 months agoremove prototypes with no matching function; ok visa@
jsg [Mon, 20 May 2024 23:13:33 +0000 (23:13 +0000)]
remove prototypes with no matching function; ok visa@

5 months agoFix an off by one.
martijn [Mon, 20 May 2024 21:22:43 +0000 (21:22 +0000)]
Fix an off by one.

OK miod@, claudio@

5 months agoIntroduce some macros for static intr. aggr. and slot integers, and do
mglocker [Mon, 20 May 2024 20:08:04 +0000 (20:08 +0000)]
Introduce some macros for static intr. aggr. and slot integers, and  do
more strict checking on the slot count.

5 months agofix usage; Thomas Klausner will commit the same diff to NetBSD within the
sobrado [Mon, 20 May 2024 19:16:48 +0000 (19:16 +0000)]
fix usage; Thomas Klausner will commit the same diff to NetBSD within the
next hours, so our trees will remain as similar as possible.

ok jmc@

5 months ago.Bl -enum produces ordinal numbers, not cardinal numbers;
schwarze [Mon, 20 May 2024 18:40:53 +0000 (18:40 +0000)]
.Bl -enum produces ordinal numbers, not cardinal numbers;
patch from Jan Stary <hans at stare dot cz>.

We are both confident that the practical consequences of this
documentation bug are limited since you are not supposed to commit
manual pages containing infinite numbers of list items in the first
place (remember, OpenBSD's not GNU!) - but correctness is a virtue
in documentation nonetheless.

5 months agoRewrite the description of the .Os macro to match reality.
schwarze [Mon, 20 May 2024 18:12:23 +0000 (18:12 +0000)]
Rewrite the description of the .Os macro to match reality.

While BSD releases from the CSRG did indeed put "BSD" and an operating
system version number on .Os macro lines, that practice was discontinued
in OpenBSD 25 years ago because it causes maintenance effort without
providing any benefit.
Other operating systems tend to agree that .Os is usually best left blank.
For example, FreeBSD weeded out .Os arguments about two years after OpenBSD.

5 months agoRemove always-false condition in amap_free KASSERT.
dv [Mon, 20 May 2024 17:03:36 +0000 (17:03 +0000)]
Remove always-false condition in amap_free KASSERT.

The if-condition immediately preceding it already checks the value
so the first part of the KASSERT is redundant.

ok mpi@, miod@, jca@

5 months agoInstead of tracking certificates by SKI track them by an internal identifier.
claudio [Mon, 20 May 2024 15:51:43 +0000 (15:51 +0000)]
Instead of tracking certificates by SKI track them by an internal identifier.

The certificate SKI is not strictly unique so using it as a unique id is
problematic. It is also not really needed to do that since in theory we
already know the path (but this got lost in the privsep communication).
So add a cert id and pass this id back and forth between main process and
the parser. With this id we can lookup the authentication chain in the
parser and this even works with multiple paths to the same resource.
Since we no longer lookup by SKI the valid_aki_ski function is replaced
by find_issuer() which does the lookup by certid.

The loop protection is now extended to allow each TAL to reach each file
once but still triggers if a file is reaccessed by the tree of a TAL.

In filemode the lookup now uses an AIA uri based lookup tree. Again this
replaces the SKI based lookups from before.

Done together with tb@
OK tb@ job@

5 months agocmac: zero_iv should be const
tb [Mon, 20 May 2024 14:53:37 +0000 (14:53 +0000)]
cmac: zero_iv should be const

5 months agoAdd flag to switch between normal interrupts (one interrupt per completed
mglocker [Mon, 20 May 2024 12:42:45 +0000 (12:42 +0000)]
Add flag to switch between normal interrupts (one interrupt per completed
command) and interrupt aggregation (one interrupt per <count> completed
commands).  For now, enable normal interrupts by default, since it has
turned out that this works better for us currently (see comment in the diff
for more details).

Discussed with dlg@

5 months agopool_setlowat() may call pool_prime() and thus allocate memory
jca [Mon, 20 May 2024 11:12:08 +0000 (11:12 +0000)]
pool_setlowat() may call pool_prime() and thus allocate memory

ok dlg@

5 months agoRework interaction between sleep API and exit1() and start unlocking ps_threads
claudio [Mon, 20 May 2024 10:32:20 +0000 (10:32 +0000)]
Rework interaction between sleep API and exit1() and start unlocking ps_threads

This diff adjusts how single_thread_set() accounts the threads by using
ps_threadcnt as initial value and counting all threads out that are already
parked. In single_thread_check call exit1() before decreasing ps_singlecount
this is now done in exit1().

exit1() and thread_fork() ensure that ps_threadcnt is updated with the
pr->ps_mtx held and in exit1() also account for exiting threads since
exit1() can sleep.

OK mpi@

5 months agoConvert more of the session engine parsers to the new ibuf API
claudio [Mon, 20 May 2024 10:01:52 +0000 (10:01 +0000)]
Convert more of the session engine parsers to the new ibuf API

This converts OPEN and it capability parser and RREFRESH and with that
all packet parser are kind of converted. There is still parse_header()
and the general packet handling that needs some work now so that all
the ugly ibuf_from_buffer() can be dropped.
OK tb@

5 months agoTabs on empty line
claudio [Mon, 20 May 2024 10:00:00 +0000 (10:00 +0000)]
Tabs on empty line

5 months agoDrop MNT_LOCAL flag in corresponding `vfsconflist' fuse(4) entry instead
mvs [Mon, 20 May 2024 09:11:21 +0000 (09:11 +0000)]
Drop MNT_LOCAL flag in corresponding `vfsconflist' fuse(4) entry instead
of cleaning it in fusefs_mount().

ok claudio

5 months agodrm/amd/display: Fix incorrect DSC instance for MST
jsg [Mon, 20 May 2024 05:26:04 +0000 (05:26 +0000)]
drm/amd/display: Fix incorrect DSC instance for MST

From Hersen Wu
af8af93c6a4f2c4772b353584b36bee88ebcbe81 in linux-6.6.y/6.6.3
892b41b16f6163e6556545835abba668fcab4eea in mainline linux

5 months agodrm/amd/display: Handle Y carry-over in VCP X.Y calculation
jsg [Mon, 20 May 2024 05:24:11 +0000 (05:24 +0000)]
drm/amd/display: Handle Y carry-over in VCP X.Y calculation

From George Shen
7be093e45215d4d9dfda515aebb7c31dc9413144 in linux-6.6.y/6.6.3
719564737a9ac3d0b49c314450b56cf6f7d71358 in mainline linux

5 months agodrm/i915/bios: Fix parsing backlight BDB data
jsg [Mon, 20 May 2024 05:22:41 +0000 (05:22 +0000)]
drm/i915/bios: Fix parsing backlight BDB data

From Karthikeyan Ramasubramanian
ca82830e453364d192e3c0c92a0f8823f621879c in linux-6.6.y/6.6.3
43b26bdd2ee5cfca80939be910d5b23a50cd7f9d in mainline linux

5 months agodrm/i915/gt: Automate CCS Mode setting during engine resets
jsg [Mon, 20 May 2024 05:20:41 +0000 (05:20 +0000)]
drm/i915/gt: Automate CCS Mode setting during engine resets

From Andi Shyti
be59b2d3d9750f784101d62cc0f25fb75e294f7e in linux-6.6.y/6.6.3
51c1b42a232f17743cd825be6790cb64735ff98f in mainline linux

5 months agodrm/i915/audio: Fix audio time stamp programming for DP
jsg [Mon, 20 May 2024 05:18:51 +0000 (05:18 +0000)]
drm/i915/audio: Fix audio time stamp programming for DP

From Chaitanya Kumar Borah
b0624c032a02a8c2141663e21961ade9d3424e3c in linux-6.6.y/6.6.3
c66b8356273c8d22498f88e4223af47a7bf8a23c in mainline linux

5 months agodrm/ttm: Print the memory decryption status just once
jsg [Mon, 20 May 2024 05:17:01 +0000 (05:17 +0000)]
drm/ttm: Print the memory decryption status just once

From Zack Rusin
306e99777886fd9fe3495a2b889d551dc509a820 in linux-6.6.y/6.6.3
27906e5d78248b19bcdfdae72049338c828897bb in mainline linux

5 months agodrm/amdkfd: don't allow mapping the MMIO HDP page with large pages
jsg [Mon, 20 May 2024 05:15:34 +0000 (05:15 +0000)]
drm/amdkfd: don't allow mapping the MMIO HDP page with large pages

From Alex Deucher
4b4cff994a27ebf7bd3fb9a798a1cdfa8d01b724 in linux-6.6.y/6.6.3
be4a2a81b6b90d1a47eaeaace4cc8e2cb57b96c7 in mainline linux

5 months agodrm/amdgpu: once more fix the call oder in amdgpu_ttm_move() v2
jsg [Mon, 20 May 2024 05:14:07 +0000 (05:14 +0000)]
drm/amdgpu: once more fix the call oder in amdgpu_ttm_move() v2

From Christian Koenig
0c7ed3ed35eec9138b88d42217b5a6b9a62bda4d in linux-6.6.y/6.6.3
d3a9331a6591e9df64791e076f6591f440af51c3 in mainline linux

5 months agodrm/amdgpu: Fix comparison in amdgpu_res_cpu_visible
jsg [Mon, 20 May 2024 05:10:45 +0000 (05:10 +0000)]
drm/amdgpu: Fix comparison in amdgpu_res_cpu_visible

From Michel Daenzer
52c1af381cb8bf72f2600ea02ba03d05fee2d733 in linux-6.6.y/6.6.3
8d2c930735f850e5be6860aeb39b27ac73ca192f in mainline linux

5 months agodrm/amd/display: Atom Integrated System Info v2_2 for DCN35
jsg [Mon, 20 May 2024 05:08:47 +0000 (05:08 +0000)]
drm/amd/display: Atom Integrated System Info v2_2 for DCN35

From Gabe Teeger
7e3030774431eb093165a31baff040d35446fb8b in linux-6.6.y/6.6.31
9a35d205f466501dcfe5625ca313d944d0ac2d60 in mainline linux

5 months agodm/amd/pm: Fix problems with reboot/shutdown for some SMU 13.0.4/13.0.11 users
jsg [Mon, 20 May 2024 05:06:41 +0000 (05:06 +0000)]
dm/amd/pm: Fix problems with reboot/shutdown for some SMU 13.0.4/13.0.11 users

From Mario Limonciello
6c83a8f236ece78d5c2e60ae3dcfd1a64509410e in linux-6.6.y/6.6.31
cd94d1b182d2986378550c9087571991bfee01d4 in mainline linux

5 months agodrm/connector: Add \n to message about demoting connector force-probes
jsg [Mon, 20 May 2024 05:04:43 +0000 (05:04 +0000)]
drm/connector: Add \n to message about demoting connector force-probes

From Douglas Anderson
1b9e8de483bcc230f6e922bdfa9d1c186c27dd3b in linux-6.6.y/6.6.31
6897204ea3df808d342c8e4613135728bc538bcd in mainline linux

5 months agodrm/radeon: silence UBSAN warning (v3)
jsg [Mon, 20 May 2024 05:03:00 +0000 (05:03 +0000)]
drm/radeon: silence UBSAN warning (v3)

From Alex Deucher
aff9dba5e8cedddb32759e835304697661ccf3f1 in linux-6.6.y/6.6.31
781d41fed19caf900c8405064676813dc9921d32 in mainline linux

5 months agoamd/amdkfd: sync all devices to wait all processes being evicted
jsg [Mon, 20 May 2024 05:00:46 +0000 (05:00 +0000)]
amd/amdkfd: sync all devices to wait all processes being evicted

From Zhigang Luo
b6f6626528fe724b512c34f3fb5946c36a135f58 in linux-6.6.y/6.6.31
d06af584be5a769d124b7302b32a033e9559761d in mainline linux

5 months agodrm/amdgpu: Fix VCN allocation in CPX partition
jsg [Mon, 20 May 2024 04:58:57 +0000 (04:58 +0000)]
drm/amdgpu: Fix VCN allocation in CPX partition

From Lijo Lazar
a624829bba2752a2b726f227574d6ca1d2e39671 in linux-6.6.y/6.6.31
f7e232de51bb1b45646e5b7dc4ebcf13510f2630 in mainline linux

5 months agodrm/amd/display: Skip on writeback when it's not applicable
jsg [Mon, 20 May 2024 04:56:48 +0000 (04:56 +0000)]
drm/amd/display: Skip on writeback when it's not applicable

From Alex Hung
951a498fa993c5501994ec2df97c9297b02488c7 in linux-6.6.y/6.6.31
ecedd99a9369fb5cde601ae9abd58bca2739f1ae in mainline linux

5 months agodrm/amdgpu: implement IRQ_STATE_ENABLE for SDMA v4.4.2
jsg [Mon, 20 May 2024 04:55:16 +0000 (04:55 +0000)]
drm/amdgpu: implement IRQ_STATE_ENABLE for SDMA v4.4.2

From Tao Zhou
c0beff4e35af8ef3f5e30ac3f55a6109f751d9c0 in linux-6.6.y/6.6.31
f886b49feaae30acd599e37d4284836024b0f3ed in mainline linux

5 months agodrm/amdgpu: Refine IB schedule error logging
jsg [Mon, 20 May 2024 04:53:54 +0000 (04:53 +0000)]
drm/amdgpu: Refine IB schedule error logging

From Lijo Lazar
1e2c28e74a823aa36821c5c68e7834ca7e4fbb9b in linux-6.6.y/6.6.31
4b18a91faf1752f9bd69a4ed3aed2c8f6e5b0528 in mainline linux

5 months agodrm/amdkfd: range check cp bad op exception interrupts
jsg [Mon, 20 May 2024 04:52:26 +0000 (04:52 +0000)]
drm/amdkfd: range check cp bad op exception interrupts

From Jonathan Kim
41dc6791596656dd41100b85647ed489e1d5c2f2 in linux-6.6.y/6.6.31
0cac183b98d8a8c692c98e8dba37df15a9e9210d in mainline linux

5 months agodrm/amdkfd: Check cgroup when returning DMABuf info
jsg [Mon, 20 May 2024 04:50:56 +0000 (04:50 +0000)]
drm/amdkfd: Check cgroup when returning DMABuf info

From Mukul Joshi
d644962423704441713be6669b9544beed5d4176 in linux-6.6.y/6.6.31
9d7993a7ab9651afd5fb295a4992e511b2b727aa in mainline linux

5 months agodrm/amdgpu: fix doorbell regression
jsg [Mon, 20 May 2024 04:49:32 +0000 (04:49 +0000)]
drm/amdgpu: fix doorbell regression

From Shashank Sharma
4fd8cd4154a213e944933622039d83184b31617c in linux-6.6.y/6.6.31
705d0480e6ae5a73ca3a9c04316d0678e19a46ed in mainline linux

5 months agoadd __counted_by() for 6.6.31 drm
jsg [Mon, 20 May 2024 03:21:58 +0000 (03:21 +0000)]
add __counted_by() for 6.6.31 drm

5 months agoremove xdr_ypreq_xfr() prototype that matches rpcsvc/yp.h
jsg [Mon, 20 May 2024 02:00:25 +0000 (02:00 +0000)]
remove xdr_ypreq_xfr() prototype that matches rpcsvc/yp.h

5 months agoMap MSI-X in addition to MSI and INTx on rge(4).
kevlo [Mon, 20 May 2024 01:51:32 +0000 (01:51 +0000)]
Map MSI-X in addition to MSI and INTx on rge(4).

Brad noticed on Orange Pi 5 Plus apparently the RTL8125 Ethernet controller
doesn't support MSIs, but does support MSI-X.  With this diff running on
Orange Pi 5+ we switch over from shared INTx to device-specific MSI-X
interrupts.

Tested by Brad.
ok kettenis@

5 months agoremove setproctitle() prototype that matches the one in stdlib.h
jsg [Mon, 20 May 2024 01:05:17 +0000 (01:05 +0000)]
remove setproctitle() prototype that matches the one in stdlib.h

5 months agoUse aggregation interrupts like stated in the documentation; Setup values
mglocker [Sun, 19 May 2024 20:24:02 +0000 (20:24 +0000)]
Use aggregation interrupts like stated in the documentation;  Setup values
(counter and timeout) during device init, reset values once commands have
been completed.  This also let us get rid of the 'sc_intraggr_enabled'
variable.

5 months agoAdd missing kex-names.c source file required since the ssh split.
anton [Sun, 19 May 2024 19:10:01 +0000 (19:10 +0000)]
Add missing kex-names.c source file required since the ssh split.

5 months agoremove prototypes with no matching function
jsg [Sun, 19 May 2024 10:39:40 +0000 (10:39 +0000)]
remove prototypes with no matching function

5 months agoremove prototypes with no matching function
jsg [Sun, 19 May 2024 10:30:43 +0000 (10:30 +0000)]
remove prototypes with no matching function

5 months agounwrap a line
tb [Sun, 19 May 2024 08:26:03 +0000 (08:26 +0000)]
unwrap a line

5 months agoAdd space after commas
tb [Sun, 19 May 2024 08:22:40 +0000 (08:22 +0000)]
Add space after commas

5 months agoKNF for dh_err and dsa_err
tb [Sun, 19 May 2024 08:21:13 +0000 (08:21 +0000)]
KNF for dh_err and dsa_err

5 months agoremove prototypes with no matching function
jsg [Sun, 19 May 2024 07:12:50 +0000 (07:12 +0000)]
remove prototypes with no matching function
feedback and ok tb@

5 months agoremove prototypes with no matching function
jsg [Sun, 19 May 2024 03:31:05 +0000 (03:31 +0000)]
remove prototypes with no matching function

5 months agoremove prototype with no matching function
jsg [Sun, 19 May 2024 03:27:58 +0000 (03:27 +0000)]
remove prototype with no matching function

5 months agoremove prototypes with no matching function; ok ratchov@
jsg [Sun, 19 May 2024 00:05:43 +0000 (00:05 +0000)]
remove prototypes with no matching function; ok ratchov@

5 months agoremove sun3/4 bits missed when sparc was removed
jsg [Sun, 19 May 2024 00:03:10 +0000 (00:03 +0000)]
remove sun3/4 bits missed when sparc was removed
ok miod@

5 months agoprefixing flags to ps(1) by a hyphen is optional; while here, make synopsis
sobrado [Sat, 18 May 2024 13:08:09 +0000 (13:08 +0000)]
prefixing flags to ps(1) by a hyphen is optional; while here, make synopsis
fit in a 80-column display.

ok jmc@

5 months agoremove prototypes with no matching function; ok claudio@
jsg [Sat, 18 May 2024 11:17:30 +0000 (11:17 +0000)]
remove prototypes with no matching function; ok claudio@

5 months agosync libc bump
tb [Sat, 18 May 2024 09:51:44 +0000 (09:51 +0000)]
sync libc bump

5 months agoremove prototypes with no matching function
jsg [Sat, 18 May 2024 09:02:34 +0000 (09:02 +0000)]
remove prototypes with no matching function

5 months agoremove prototypes with no matching function; ok nicm@
jsg [Sat, 18 May 2024 08:51:26 +0000 (08:51 +0000)]
remove prototypes with no matching function; ok nicm@

5 months agoremove externs with no matching var; ok nicm@
jsg [Sat, 18 May 2024 08:50:11 +0000 (08:50 +0000)]
remove externs with no matching var; ok nicm@

5 months agoremove extern with no matching var; ok tb@
jsg [Sat, 18 May 2024 08:48:31 +0000 (08:48 +0000)]
remove extern with no matching var; ok tb@

5 months agoremove prototypes with no matching function; ok tb@
jsg [Sat, 18 May 2024 08:47:13 +0000 (08:47 +0000)]
remove prototypes with no matching function; ok tb@

5 months agoremove prototypes with no matching function
jsg [Sat, 18 May 2024 07:10:16 +0000 (07:10 +0000)]
remove prototypes with no matching function

5 months agoremove prototypes with no matching function
jsg [Sat, 18 May 2024 06:45:00 +0000 (06:45 +0000)]
remove prototypes with no matching function

5 months agoremove prototypes with no matching function
jsg [Sat, 18 May 2024 06:34:46 +0000 (06:34 +0000)]
remove prototypes with no matching function

5 months agoWhen comparing mtimes for the -u and -Z options and the target is
guenther [Sat, 18 May 2024 05:21:38 +0000 (05:21 +0000)]
When comparing mtimes for the -u and -Z options and the target is
'too old', use pathconfat(_PC_TIMESTAMP_RESOLUTION, AT_SYMLINK_NOFOLLOW)
to get the timestamp resolution to which the _source_ timestamp
should be truncated for a stable comparison.

Problem reported by Walter Alejandro Iglesias (wai(at)roquesor.com)
ok millert@

5 months agoRegen
guenther [Sat, 18 May 2024 05:21:02 +0000 (05:21 +0000)]
Regen

5 months agoAdd pathconfat(2): pathconf(2) but with at-fd and flags arguments,
guenther [Sat, 18 May 2024 05:20:22 +0000 (05:20 +0000)]
Add pathconfat(2): pathconf(2) but with at-fd and flags arguments,
the latter supporting the ability to get timestamp resolution of
symlinks.

ok deraadt@ millert@

5 months agoremove prototype with no matching function
jsg [Sat, 18 May 2024 02:44:22 +0000 (02:44 +0000)]
remove prototype with no matching function

5 months agoremove prototypes with no matching function
jsg [Sat, 18 May 2024 01:51:44 +0000 (01:51 +0000)]
remove prototypes with no matching function

5 months agoremove externs for removed vars
jsg [Sat, 18 May 2024 00:08:06 +0000 (00:08 +0000)]
remove externs for removed vars

5 months agoremove prototypes with no matching function; ok florian@
jsg [Fri, 17 May 2024 23:56:19 +0000 (23:56 +0000)]
remove prototypes with no matching function; ok florian@

5 months agoIPv6 has to use ip6_defhlim, not ip_defttl.
bluhm [Fri, 17 May 2024 20:44:36 +0000 (20:44 +0000)]
IPv6 has to use ip6_defhlim, not ip_defttl.

OK claudio@

5 months agoMake sure `row' is not used uninitialized in wsmouse_matching().
miod [Fri, 17 May 2024 20:11:58 +0000 (20:11 +0000)]
Make sure `row' is not used uninitialized in wsmouse_matching().

5 months agoCast alpha_pal_swpipl return value to int in splx, to make sure both sides
miod [Fri, 17 May 2024 20:07:33 +0000 (20:07 +0000)]
Cast alpha_pal_swpipl return value to int in splx, to make sure both sides
of the ternary operator have the same type. NFC

5 months agoConstify the first argument of getdevunit(). NFCI
miod [Fri, 17 May 2024 20:05:08 +0000 (20:05 +0000)]
Constify the first argument of getdevunit(). NFCI

5 months agoUpdate comments to match reality.
miod [Fri, 17 May 2024 20:03:13 +0000 (20:03 +0000)]
Update comments to match reality.

5 months agoThe INTSTS register is read-only, so the attempt to clear interrupts
kettenis [Fri, 17 May 2024 19:43:45 +0000 (19:43 +0000)]
The INTSTS register is read-only, so the attempt to clear interrupts
by writing to it doesn't actually work.  Therefore we have to keep
processing interrupts until all the bits corresponding to enabled
interrupts are cleared.  This makes MSIs actually work, also on the AMD
implementations for which we currently disable MSIs.

Based on some digging by martijn@

ok ratchov@, martijn@, stsp@

5 months agoTurn sblock() to `sb_lock' rwlock(9) wrapper for all sockets.
mvs [Fri, 17 May 2024 19:11:14 +0000 (19:11 +0000)]
Turn sblock() to `sb_lock' rwlock(9) wrapper for all sockets.

Unify behaviour to all sockets. Now sblock() should be always
taken before solock() in all involved paths as sosend(), soreceive(),
sorflush() and sosplice(). sblock() is fine-grained lock which
serializes socket send and receive routines on `so_rcv' or `so_snd'
buffers. There is no big problem to wait netlock while holding sblock().

This unification removes a lot of temporary "sb_flags & SB_MTXLOCK" code
from sockets layer. This unification makes straight "solock()" and
"sblock()" lock order, no more solock() -> sblock() -> sounlock() ->
solock() -> sbunlock() -> sounlock() chains in sosend() and soreceive()
paths. This unification brings witness(4) support for sblock(), include
NFS involved sockets, which is useful.

Since the witness(4) support was introduced to sblock() with this diff,
some new witness reports appeared.

bulk(1) tests by tb, ok bluhm

5 months agoSwitch AF_KEY sockets to the new locking scheme.
mvs [Fri, 17 May 2024 19:02:04 +0000 (19:02 +0000)]
Switch AF_KEY sockets to the new locking scheme.

The simplest case. Nothing to change in sockets layer, only set
SB_MTXLOCK on socket buffers.

ok bluhm

5 months agoFix uninitialized memory access in pfkeyv2_sysctl().
mvs [Fri, 17 May 2024 18:58:26 +0000 (18:58 +0000)]
Fix uninitialized memory access in pfkeyv2_sysctl().

pfkeyv2_sysctl() reads the SA type from uninitialized memory if it is
not provided by the caller of sysctl(2) because of a missing length
check.

From Carsten Beckmann.

ok bluhm

5 months agoFix possible segmentation fault in Perl 5.38.2.
bluhm [Fri, 17 May 2024 18:11:58 +0000 (18:11 +0000)]
Fix possible segmentation fault in Perl 5.38.2.

When calling POSIX::setlocale() with unknown locale category, Perl
could crash due to NULL dereference.

found by sthen@ in smokeping build; OK afresh1@

5 months agovio: Fix signal handling and locking in sysctl path
sf [Fri, 17 May 2024 16:37:10 +0000 (16:37 +0000)]
vio: Fix signal handling and locking in sysctl path

Commits f0b002d01d5 "Release the netlock when sleeping for control
messages in in vioioctl()" and 126b881f71 "Insert a workaround for
per-ifp ioctl being called w/o NET_LOCK()." in vio(4) fixed a deadlock
but may cause a crash with a protection fault trap if addresses are
added/removed concurrently.

The actual issue is that signals are not handled correctly while
sleeping. After a signal, there is a race condition where sc_ctrl_inuse
is first set to FREE and then the interrupt handler sets it to DONE,
causing a hang in the next vio_wait_ctrl() call.

To fix it:

* Revert the NET_LOCK unlocking work-around.

* Remove PCATCH from the sleep call when we wait for control queue,
  avoiding the race with vio_ctrleof(). To ensure that we don't hang
  forever, use a 5 second timeout.

* If the timeout is hit, or if the hypervisor has set the
  DEVICE_NEEDS_RESET status bit, do not try to use the control queue
  until the next ifconfig down/up which resets the device.

* In order to allow reading the device status from device drivers, add a
  new interface to the virtio transport drivers.

* Avoid a crash if there is outgoing traffic while doing ifconfig down.

OK bluhm@

5 months agoremove duplicate copy of relink kit for sshd-session
naddy [Fri, 17 May 2024 14:42:00 +0000 (14:42 +0000)]
remove duplicate copy of relink kit for sshd-session

5 months agodig: make function definitions match their prototypes
tb [Fri, 17 May 2024 09:36:48 +0000 (09:36 +0000)]
dig: make function definitions match their prototypes

ok florian

5 months agodig: garbage collect the unused 'added'
tb [Fri, 17 May 2024 09:36:15 +0000 (09:36 +0000)]
dig: garbage collect the unused 'added'

ok florian