openbsd
6 weeks agodrm/i915: Do not attempt to load the GSC multiple times
jsg [Thu, 12 Sep 2024 23:26:26 +0000 (23:26 +0000)]
drm/i915: Do not attempt to load the GSC multiple times

From Daniele Ceraolo Spurio
337266ada863a4232c9f8634deedc298a145521c in linux-6.6.y/6.6.51
59d3cfdd7f9655a0400ac453bf92199204f8b2a1 in mainline linux

6 weeks agoRevert "drm/amdgpu: align pp_power_profile_mode with kernel docs"
jsg [Thu, 12 Sep 2024 23:22:10 +0000 (23:22 +0000)]
Revert "drm/amdgpu: align pp_power_profile_mode with kernel docs"

From Alex Deucher
94479011f4f551b4b1e010776a131512402b27bc in linux-6.6.y/6.6.51
1a8d845470941f1b6de1b392227530c097dc5e0c in mainline linux

6 weeks agoReintroduce check that CRL Number is in range
tb [Thu, 12 Sep 2024 10:33:25 +0000 (10:33 +0000)]
Reintroduce check that CRL Number is in range

The CRL number draft clarified what ignoring means and it includes checking
that the CRL number is well-formed again. So do this but continue to ignore
the value for any other purpose. This refactors x509_convert_seqnum() into
a couple of helpers. There's some duplication between crl_check_crl_number()
and crl_parse_crl_number() which could be removed if anyone cares.

tweaks/ok job

6 weeks agoDo a basic sanity check that dirents returned via fuse are kind of sane.
claudio [Thu, 12 Sep 2024 09:10:46 +0000 (09:10 +0000)]
Do a basic sanity check that dirents returned via fuse are kind of sane.

Ensure that file names passed back by readdir do not include a '/'
character. The '/' char is the path separator and is not allowed in
any filename. On top of this also check that d_reclen and d_namlen
are kind of sane and zero out the padding bytes after d_name.

OK beck@

6 weeks agomsdos already transfroms for Windows long names a '/' char into '?'.
claudio [Thu, 12 Sep 2024 09:07:28 +0000 (09:07 +0000)]
msdos already transfroms for Windows long names a '/' char into '?'.
Do the same for the 8.3 case as well.

This is not ideal since now it is possible that two files in the same
directory have the same name but the msdos code already does a lot of
this and so the problem already exists.

OK beck@ miod@

6 weeks agoEnsure that file names passed back by readdir do not include a '/'
claudio [Thu, 12 Sep 2024 09:04:51 +0000 (09:04 +0000)]
Ensure that file names passed back by readdir do not include a '/'
character. The '/' char is the path separator and is not allowed in
any filename.

NFS specific report by Apple Security Engineering and Architecture (SEAR).

Input from guenther@ and millert@
OK beck@ miod@

6 weeks agoRelax absolute path requirement back to what it was prior to OpenSSH 9.8,
djm [Thu, 12 Sep 2024 00:36:27 +0000 (00:36 +0000)]
Relax absolute path requirement back to what it was prior to OpenSSH 9.8,
which incorrectly required that sshd was started with an absolute path
in inetd mode. bz3717, patch from Colin Wilson

7 weeks agoMouse move keys are not useful as key bindings because we do not turn
nicm [Wed, 11 Sep 2024 19:12:33 +0000 (19:12 +0000)]
Mouse move keys are not useful as key bindings because we do not turn
them on unless the application requests them. Ignore them so they do not
cause the prefix to be canceled, GitHub issue 4111.

7 weeks agoImplement AMD SEV support in vmd(8).
bluhm [Wed, 11 Sep 2024 15:42:52 +0000 (15:42 +0000)]
Implement AMD SEV support in vmd(8).

To launch a guest with AMD SEV enabled, vmd needs to do a few things:
- retrieve ASID used by guest on VM creation
- provide ASID to psp(4)
- let psp(4) encrypt memory used intially by guest
- run guest
- release resources held by psp(4) on guest shutdown
To enable SEV for a guest use the parameter "sev" in the guest's vm
section in vm.conf.

from hshoexer@; OK mlarkin@

7 weeks agoparametes -> parameters
tb [Wed, 11 Sep 2024 15:04:16 +0000 (15:04 +0000)]
parametes -> parameters

7 weeks agoAfter calling m_freem() on nmi_mrep (or nmi_mreq) set the pointer to NULL.
claudio [Wed, 11 Sep 2024 12:22:34 +0000 (12:22 +0000)]
After calling m_freem() on nmi_mrep (or nmi_mreq) set the pointer to NULL.

Only do this if struct nfsm_info doesn't have local scope.
In some cases the caller would perfrom another m_freem and double free
the mbuf and Bad Things(TM) would happen.

Reported by Claes M Nyberg on bugs@; with & ok miod@

7 weeks agoWhile I can understand that ext2fs is using ufs_ihashget() I'm still
claudio [Wed, 11 Sep 2024 08:29:55 +0000 (08:29 +0000)]
While I can understand that ext2fs is using ufs_ihashget() I'm still
flabbergasted by the abuse from fuse.

For whatever dumb reason fuse uses the ufs inode instead of having its
own much simpler struct. Again this is a workaround to not crash when
fuse is used.

OK beck@

7 weeks agofstat(2) can't return an S_IFLNK, so delete that test.
guenther [Wed, 11 Sep 2024 03:57:14 +0000 (03:57 +0000)]
fstat(2) can't return an S_IFLNK, so delete that test.
Also, switch to S_IS*() tests and update the manpage
to reflect that POSIX-2024 has no substantive changes
for wc(1)

ok op@ millert@

7 weeks agoFix a typo
yasuoka [Wed, 11 Sep 2024 00:41:51 +0000 (00:41 +0000)]
Fix a typo

7 weeks agoDelete a test_ps() call that does not actually test
schwarze [Tue, 10 Sep 2024 19:46:01 +0000 (19:46 +0000)]
Delete a test_ps() call that does not actually test
what it was intended to test.

The intention was to test that src/bin/ps/utf8.c handles embedded NUL bytes
correctly, just as the other tests in the same group test the handling of
various other non-printable characters.  But testing that does not work
for multiple reasons.  Neither does the shell pass the NUL byte to the
called test program as intended, nor can argv[] in a C program contain
a NUL byte in the middle of an argument, simply because in the C language,
a string is defined to end at the first NUL byte.  On top of all that,
even the function mbswprint() that was supposed to be tested terminates
the processing of the multi-byte input string as soon as it encounters
a NUL byte (all of which is correct behaviour).

So this particular subtest was totally bogus and only worked by accident,
for reasons completely unrelated to the intended purpose.  I don't think
the test needs to be replaced by anything else.  NUL bytes in the middle
of a program argument just aren't a thing in the first place.

Thanks to deraadt@ for asking what the purpose of this test_ps() call was.

7 weeks agonfsm_srvnamesiz() may set up an NFSERR_NAMETOL error, which nfsm_reply() would
miod [Tue, 10 Sep 2024 18:44:04 +0000 (18:44 +0000)]
nfsm_srvnamesiz() may set up an NFSERR_NAMETOL error, which nfsm_reply() would
consider as not tragic enough to abort the operation, in order to batch error
replies.

This would end up invoking nfs_namei() using an uninitialized variable as
length, and Bad Things(TM) would happen.

Reported by Claes M Nyberg on bugs@; tweaks & ok claudio@

7 weeks agoUse doc/html link for consistency
tb [Tue, 10 Sep 2024 18:37:42 +0000 (18:37 +0000)]
Use doc/html link for consistency

7 weeks agoAdd reference to the manifest numbers draft
tb [Tue, 10 Sep 2024 18:34:19 +0000 (18:34 +0000)]
Add reference to the manifest numbers draft

ok job

7 weeks agolibcrypto hasn't had VIA padlock support compiled in for quite some time...
tb [Tue, 10 Sep 2024 17:01:09 +0000 (17:01 +0000)]
libcrypto hasn't had VIA padlock support compiled in for quite some time...

prompted by a question by jmc

7 weeks agoFix build of m_print_chain() on sparc64.
bluhm [Tue, 10 Sep 2024 14:52:42 +0000 (14:52 +0000)]
Fix build of m_print_chain() on sparc64.

Use %zu to print mbuf MHLEN and MLEN in ddb, otherwise gcc complains.

found by claudio@

7 weeks agoufs_ihashget() is also used by the ext2fs code but the DIP() makro does
claudio [Tue, 10 Sep 2024 12:14:26 +0000 (12:14 +0000)]
ufs_ihashget() is also used by the ext2fs code but the DIP() makro does
not handle that. So for now add an ugly hack here to support ext2fs.
This fixes access to ext2fs after the last commit. A better fix can be
implemented in tree.
Reported and OK anton@

7 weeks agoBe more careful with aspath that have 0 length (aka the empty AS_PATH).
claudio [Tue, 10 Sep 2024 09:38:45 +0000 (09:38 +0000)]
Be more careful with aspath that have 0 length (aka the empty AS_PATH).

Again malloc(0) is not portable and calling memcpy with a NULL pointer
and a 0 length is not allowed by the C standard.

OK tb@

7 weeks agocommunity_copy needs to check if nentries is 0 and handle that specially.
claudio [Tue, 10 Sep 2024 08:53:20 +0000 (08:53 +0000)]
community_copy needs to check if nentries is 0 and handle that specially.

Calling malloc / reallocarray with a 0 size is not portable and the
memcpy with a possible NULL pointer as source and 0 len is seen as UB
by newer C standards (grmbl).

OK tb@

7 weeks agoAID_VPN_IPv4 and AID_VPN_IPv6 require a labellen that is non-zero.
claudio [Tue, 10 Sep 2024 08:47:51 +0000 (08:47 +0000)]
AID_VPN_IPv4 and AID_VPN_IPv6 require a labellen that is non-zero.

OK tb@

7 weeks agoEnfroce proper encoding of ASPA announce/withdraw PDU.
claudio [Tue, 10 Sep 2024 08:41:13 +0000 (08:41 +0000)]
Enfroce proper encoding of ASPA announce/withdraw PDU.

An announce PDU requires at least one provider ASnum while a withdraw
must not include any provider ASnums. The first is mandated by the ASPA
profile and the 2nd by the 8210bis draft.

Further cleanup some leftovers from the old per-AFI split of ASPA.
OK tb@

7 weeks agoUse reallocarray() instead of recallocarray(), this code does not need the
claudio [Tue, 10 Sep 2024 08:37:52 +0000 (08:37 +0000)]
Use reallocarray() instead of recallocarray(), this code does not need the
extra security measures of recallocarray() which adds a lot of overhead.
OK tb@

7 weeks agoInclude CLUSTER_LIST in the bad attribute tests. Also add a 2nd session
claudio [Tue, 10 Sep 2024 08:27:00 +0000 (08:27 +0000)]
Include CLUSTER_LIST in the bad attribute tests. Also add a 2nd session
which is iBGP so we can test both CLUSTER_LIST behaviours.

7 weeks agochange bus notifier defines into inline functions
jsg [Tue, 10 Sep 2024 05:45:16 +0000 (05:45 +0000)]
change bus notifier defines into inline functions

7 weeks agomissing "Ar" in previous;
jmc [Tue, 10 Sep 2024 05:33:32 +0000 (05:33 +0000)]
missing "Ar" in previous;

7 weeks agoWhen accounting start the type attribute was added twice.
yasuoka [Mon, 9 Sep 2024 23:38:29 +0000 (23:38 +0000)]
When accounting start the type attribute was added twice.
from markus

7 weeks agoHandle the CLUSTER_LIST attribute as described in RFC7606
claudio [Mon, 9 Sep 2024 15:00:45 +0000 (15:00 +0000)]
Handle the CLUSTER_LIST attribute as described in RFC7606

Just drop the attribute if received from an external peer.
Treat as withdraw if the len is 0 or not % 4.
OK tb@

7 weeks agoThere is no need for an explicit size check for IMSG_RECONF_ASPA_TAS
claudio [Mon, 9 Sep 2024 14:58:47 +0000 (14:58 +0000)]
There is no need for an explicit size check for IMSG_RECONF_ASPA_TAS
imsg_get_data() does the same and produces the same error.
OK tb@

7 weeks agodocument the mlkem768x25519-sha256 key exchange algorithm
naddy [Mon, 9 Sep 2024 14:41:21 +0000 (14:41 +0000)]
document the mlkem768x25519-sha256 key exchange algorithm

7 weeks agoUse msgbuf_init() instead of hand initalizing the msgbuf.
claudio [Mon, 9 Sep 2024 12:59:49 +0000 (12:59 +0000)]
Use msgbuf_init() instead of hand initalizing the msgbuf.
OK tb@

7 weeks agoPrint mbuf size also for non cluster.
bluhm [Mon, 9 Sep 2024 11:27:03 +0000 (11:27 +0000)]
Print mbuf size also for non cluster.

Command "ddb> show /c mbuf" always prints mbuf data size.  In
uipc_mbuf.c include db_interface.h as it contains prototype for
m_print_chain().

OK mvs@

7 weeks agodrm/amd/display: Skip wbscl_set_scaler_filter if filter is null
jsg [Mon, 9 Sep 2024 10:02:33 +0000 (10:02 +0000)]
drm/amd/display: Skip wbscl_set_scaler_filter if filter is null

From Alex Hung
e3a95f29647ae45d1ec9541cd7df64f40bf2120a in linux-6.6.y/6.6.50
c4d31653c03b90e51515b1380115d1aedad925dd in mainline linux

7 weeks agodrm/amd/display: Check BIOS images before it is used
jsg [Mon, 9 Sep 2024 10:00:20 +0000 (10:00 +0000)]
drm/amd/display: Check BIOS images before it is used

From Alex Hung
e50bec62acaeec03afc6fa5dfb2426e52d049cf5 in linux-6.6.y/6.6.50
8b0ddf19cca2a352b2a7e01d99d3ba949a99c84c in mainline linux

7 weeks agodrm/amd/display: use preferred link settings for dp signal only
jsg [Mon, 9 Sep 2024 09:58:02 +0000 (09:58 +0000)]
drm/amd/display: use preferred link settings for dp signal only

From Wenjing Liu
67cf14c04ff0ea5c246f80ea3a044c2afb93c99d in linux-6.6.y/6.6.50
abf34ca465f5cd182b07701d3f3d369c0fc04723 in mainline linux

7 weeks agodrm/amd/display: Correct the defined value for AMDGPU_DMUB_NOTIFICATION_MAX
jsg [Mon, 9 Sep 2024 09:56:10 +0000 (09:56 +0000)]
drm/amd/display: Correct the defined value for AMDGPU_DMUB_NOTIFICATION_MAX

From Wayne Lin
800a5ab673c4a61ca220cce177386723d91bdb37 in linux-6.6.y/6.6.50
ad28d7c3d989fc5689581664653879d664da76f0 in mainline linux

7 weeks agodrm/amd/display: added NULL check at start of dc_validate_stream
jsg [Mon, 9 Sep 2024 09:53:32 +0000 (09:53 +0000)]
drm/amd/display: added NULL check at start of dc_validate_stream

From winstang
154a50bf4221a6a6ccf88d565b8184da7c40a2dd in linux-6.6.y/6.6.50
26c56049cc4f1705b498df013949427692a4b0d5 in mainline linux

7 weeks agodrm/amd/display: Don't use fsleep for PSR exit waits on dmub replay
jsg [Mon, 9 Sep 2024 09:51:03 +0000 (09:51 +0000)]
drm/amd/display: Don't use fsleep for PSR exit waits on dmub replay

From Nicholas Kazlauskas
da696cbb47941c40b01cc10ea5504dc5a4ebac6e in linux-6.6.y/6.6.50
b5236da757adc75d7e52c69bdc233d29249a0d0c in mainline linux

7 weeks agodrm/amdgpu: add lock in kfd_process_dequeue_from_device
jsg [Mon, 9 Sep 2024 09:49:12 +0000 (09:49 +0000)]
drm/amdgpu: add lock in kfd_process_dequeue_from_device

From Yunxiang Li
381113ef01e28282f5f5a3267584b075427529f0 in linux-6.6.y/6.6.50
d225960c2330e102370815367b877baaf8bb8b5d in mainline linux

7 weeks agodrm/amdgpu: add lock in amdgpu_gart_invalidate_tlb
jsg [Mon, 9 Sep 2024 09:47:46 +0000 (09:47 +0000)]
drm/amdgpu: add lock in amdgpu_gart_invalidate_tlb

From Yunxiang Li
ddfe95f2e1fcefeb9ab70476753e1fd0a083f59b in linux-6.6.y/6.6.50
18f2525d31401e5142db95ff3a6ec0f4147be818 in mainline linux

7 weeks agodrm/amdgpu: add skip_hw_access checks for sriov
jsg [Mon, 9 Sep 2024 09:46:24 +0000 (09:46 +0000)]
drm/amdgpu: add skip_hw_access checks for sriov

From Yunxiang Li
494b42f35fc2a858941757dbbf5619f6ff405562 in linux-6.6.y/6.6.50
b3948ad1ac582f560e1f3aeaecf384619921c48d in mainline linux

7 weeks agodrm/amdgu: fix Unintentional integer overflow for mall size
jsg [Mon, 9 Sep 2024 09:44:41 +0000 (09:44 +0000)]
drm/amdgu: fix Unintentional integer overflow for mall size

From Jesse Zhang
54a11ce4ff2a01fc2a231c546a864883633d2b97 in linux-6.6.y/6.6.50
c09d2eff81a997c169e0cacacd6b60c5e3aa33f2 in mainline linux

7 weeks agodrm/amdgpu: update type of buf size to u32 for eeprom functions
jsg [Mon, 9 Sep 2024 09:42:49 +0000 (09:42 +0000)]
drm/amdgpu: update type of buf size to u32 for eeprom functions

From Tao Zhou
0798e4330b5404c547e236c07a5a06309d0a2980 in linux-6.6.y/6.6.50
2aadb520bfacec12527effce3566f8df55e5d08e in mainline linux

7 weeks agodrm/kfd: Correct pinned buffer handling at kfd restore and validate process
jsg [Mon, 9 Sep 2024 09:41:07 +0000 (09:41 +0000)]
drm/kfd: Correct pinned buffer handling at kfd restore and validate process

From Xiaogang Chen
cdc65b5f99988f9679908cdc5204dd183260bed9 in linux-6.6.y/6.6.50
f326d7cc745683f53052b84382bd10567b45cd5d in mainline linux

7 weeks agodrm/amd/pm: check negtive return for table entries
jsg [Mon, 9 Sep 2024 09:39:49 +0000 (09:39 +0000)]
drm/amd/pm: check negtive return for table entries

From Jesse Zhang
7e8d106ca98aac5ba4fc8816e64fea994ab83d54 in linux-6.6.y/6.6.50
f76059fe14395b37ba8d997eb0381b1b9e80a939 in mainline linux

7 weeks agodrm/amdgpu: the warning dereferencing obj for nbio_v7_4
jsg [Mon, 9 Sep 2024 09:37:34 +0000 (09:37 +0000)]
drm/amdgpu: the warning dereferencing obj for nbio_v7_4

From Jesse Zhang
7d265772e44d403071a2b573eac0db60250b1c21 in linux-6.6.y/6.6.50
d190b459b2a4304307c3468ed97477b808381011 in mainline linux

7 weeks agodrm/amd/pm: check specific index for smu13
jsg [Mon, 9 Sep 2024 09:35:48 +0000 (09:35 +0000)]
drm/amd/pm: check specific index for smu13

From Jesse Zhang
025798f44b8d19c33d48ff85aa5c172dc6576d47 in linux-6.6.y/6.6.50
a3ac9d1c9751f00026c2d98b802ec8a98626c3ed in mainline linux

7 weeks agodrm/amd/pm: check specific index for aldebaran
jsg [Mon, 9 Sep 2024 09:34:36 +0000 (09:34 +0000)]
drm/amd/pm: check specific index for aldebaran

From Jesse Zhang
d0230b37206888c49df41639551802dce27fa26e in linux-6.6.y/6.6.50
0ce8ef2639c112ae203c985b758389e378630aac in mainline linux

7 weeks agodrm/amdgpu: fix the waring dereferencing hive
jsg [Mon, 9 Sep 2024 09:33:21 +0000 (09:33 +0000)]
drm/amdgpu: fix the waring dereferencing hive

From Jesse Zhang
4ab720b6aa1ef5e71db1e534b5b45c80ac4ec58a in linux-6.6.y/6.6.50
1940708ccf5aff76de4e0b399f99267c93a89193 in mainline linux

7 weeks agodrm/amdgpu: fix dereference after null check
jsg [Mon, 9 Sep 2024 09:32:06 +0000 (09:32 +0000)]
drm/amdgpu: fix dereference after null check

From Jesse Zhang
0aad97bf6d0bc7a34a19f266b0b9fb2861efe64c in linux-6.6.y/6.6.50
b1f7810b05d1950350ac2e06992982974343e441 in mainline linux

7 weeks agodrm/amdgpu: Fix the warning division or modulo by zero
jsg [Mon, 9 Sep 2024 09:30:22 +0000 (09:30 +0000)]
drm/amdgpu: Fix the warning division or modulo by zero

From Jesse Zhang
d116bb921e8b104f45d1f30a473ea99ef4262b9a in linux-6.6.y/6.6.50
1a00f2ac82d6bc6689388c7edcd2a4bd82664f3c in mainline linux

7 weeks agodrm/amdgpu/pm: Check input value for CUSTOM profile mode setting on legacy SOCs
jsg [Mon, 9 Sep 2024 09:28:27 +0000 (09:28 +0000)]
drm/amdgpu/pm: Check input value for CUSTOM profile mode setting on legacy SOCs

From Ma Jun
58350786dbbddfb5ffcbb0ef976b5c9fef5870bc in linux-6.6.y/6.6.50
df0a9bd92fbbd3fcafcb2bce6463c9228a3e6868 in mainline linux

7 weeks agodrm/amdkfd: Reconcile the definition and use of oem_id in struct kfd_topology_device
jsg [Mon, 9 Sep 2024 09:26:32 +0000 (09:26 +0000)]
drm/amdkfd: Reconcile the definition and use of oem_id in struct kfd_topology_device

From Michael Chen
f163ba83e607f9fa71a342d37af5aeb83752ccdb in linux-6.6.y/6.6.50
10f624ef239bd136cdcc5bbc626157a57b938a31 in mainline linux

7 weeks agodrm/amdgpu: fix mc_data out-of-bounds read warning
jsg [Mon, 9 Sep 2024 09:24:57 +0000 (09:24 +0000)]
drm/amdgpu: fix mc_data out-of-bounds read warning

From Tim Huang
d0a43bf367ed640e527e8ef3d53aac1e71f80114 in linux-6.6.y/6.6.50
51dfc0a4d609fe700750a62f41447f01b8c9ea50 in mainline linux

7 weeks agodrm/amdgpu: fix ucode out-of-bounds read warning
jsg [Mon, 9 Sep 2024 09:22:20 +0000 (09:22 +0000)]
drm/amdgpu: fix ucode out-of-bounds read warning

From Tim Huang
f2b7a9f3839e92f43559b2795b34640ca8cf839f in linux-6.6.y/6.6.50
8944acd0f9db33e17f387fdc75d33bb473d7936f in mainline linux

7 weeks agodrm/amdgpu: Fix out-of-bounds read of df_v1_7_channel_number
jsg [Mon, 9 Sep 2024 09:19:49 +0000 (09:19 +0000)]
drm/amdgpu: Fix out-of-bounds read of df_v1_7_channel_number

From Ma Jun
f9267972490f9fcffe146e79828e97acc0da588c in linux-6.6.y/6.6.50
d768394fa99467bcf2703bde74ddc96eeb0b71fa in mainline linux

7 weeks agodrm/amdkfd: Check debug trap enable before write dbg_ev_file
jsg [Mon, 9 Sep 2024 09:17:44 +0000 (09:17 +0000)]
drm/amdkfd: Check debug trap enable before write dbg_ev_file

From Lin.Cao
e6ea3b8fe398915338147fe54dd2db8155fdafd8 in linux-6.6.y/6.6.50
547033b593063eb85bfdf9b25a5f1b8fd1911be2 in mainline linux

7 weeks agodrm/amdgpu: Fix out-of-bounds write warning
jsg [Mon, 9 Sep 2024 09:16:02 +0000 (09:16 +0000)]
drm/amdgpu: Fix out-of-bounds write warning

From Ma Jun
a60d1f7ff62e453dde2d3b4907e178954d199844 in linux-6.6.y/6.6.50
be1684930f5262a622d40ce7a6f1423530d87f89 in mainline linux

7 weeks agodrm/amdgpu: Fix the uninitialized variable warning
jsg [Mon, 9 Sep 2024 09:14:09 +0000 (09:14 +0000)]
drm/amdgpu: Fix the uninitialized variable warning

From Ma Jun
f71ef2bb69dde228e0788e8e309d3d9d73f458c3 in linux-6.6.y/6.6.50
7e39d7ec35883a168343ea02f40e260e176c6c63 in mainline linux

7 weeks agodrm/amdgpu/pm: Fix uninitialized variable agc_btc_response
jsg [Mon, 9 Sep 2024 09:12:21 +0000 (09:12 +0000)]
drm/amdgpu/pm: Fix uninitialized variable agc_btc_response

From Ma Jun
3ad44174a5767dbe1dede02b160a4acfa9899d4f in linux-6.6.y/6.6.50
df4409d8a04dd39d7f2aa0c5f528a56b99eaaa13 in mainline linux

7 weeks agodrm/amdgpu/pm: Fix uninitialized variable warning for smu10
jsg [Mon, 9 Sep 2024 09:10:54 +0000 (09:10 +0000)]
drm/amdgpu/pm: Fix uninitialized variable warning for smu10

From Ma Jun
3e04fa97077da6124b4c2a4f582d1e4aa3d704a0 in linux-6.6.y/6.6.50
336c8f558d596699d3d9814a45600139b2f23f27 in mainline linux

7 weeks agodrm/amd/pm: fix uninitialized variable warnings for vangogh_ppt
jsg [Mon, 9 Sep 2024 09:09:08 +0000 (09:09 +0000)]
drm/amd/pm: fix uninitialized variable warnings for vangogh_ppt

From Tim Huang
97667de35ba771ed50430c5541c217c035b37810 in linux-6.6.y/6.6.50
b2871de6961d24d421839fbfa4aa3008ec9170d5 in mainline linux

7 weeks agodrm/amd/amdgpu: Check tbo resource pointer
jsg [Mon, 9 Sep 2024 09:07:31 +0000 (09:07 +0000)]
drm/amd/amdgpu: Check tbo resource pointer

From Asad Kamal
4dfec5f5501a27e0a0da00e136d65ef9011ded4c in linux-6.6.y/6.6.50
6cd2b872643bb29bba01a8ac739138db7bd79007 in mainline linux

7 weeks agodrm/amd/display: Fix index may exceed array range within fpu_update_bw_bounding_box
jsg [Mon, 9 Sep 2024 09:05:54 +0000 (09:05 +0000)]
drm/amd/display: Fix index may exceed array range within fpu_update_bw_bounding_box

From Hersen Wu
4003bac784380fed1f94f197350567eaa73a409d in linux-6.6.y/6.6.50
188fd1616ec43033cedbe343b6579e9921e2d898 in mainline linux

7 weeks agodrm/amd/display: Skip inactive planes within ModeSupportAndSystemConfiguration
jsg [Mon, 9 Sep 2024 09:03:58 +0000 (09:03 +0000)]
drm/amd/display: Skip inactive planes within ModeSupportAndSystemConfiguration

From Hersen Wu
8406158a546441b73f0b216aedacbf9a1e5748fb in linux-6.6.y/6.6.50
a54f7e866cc73a4cb71b8b24bb568ba35c8969df in mainline linux

7 weeks agodrm/amd/display: Ensure index calculation will not overflow
jsg [Mon, 9 Sep 2024 09:02:02 +0000 (09:02 +0000)]
drm/amd/display: Ensure index calculation will not overflow

From Alex Hung
3dc6bb57dab36b38b7374af0ac916174c146b6ed in linux-6.6.y/6.6.50
8e2734bf444767fed787305ccdcb36a2be5301a2 in mainline linux

7 weeks agodrm/amd/display: Fix Coverity INTEGER_OVERFLOW within decide_fallback_link_setting_ma...
jsg [Mon, 9 Sep 2024 09:00:17 +0000 (09:00 +0000)]
drm/amd/display: Fix Coverity INTEGER_OVERFLOW within decide_fallback_link_setting_max_bw_policy

From Hersen Wu
94b0689984dd6a9e9f2303a1510333d18204ba88 in linux-6.6.y/6.6.50
83c0c8361347cf43937348e8ca0a487679c003ae in mainline linux

7 weeks agodrm/amd/display: Spinlock before reading event
jsg [Mon, 9 Sep 2024 08:57:58 +0000 (08:57 +0000)]
drm/amd/display: Spinlock before reading event

From Alex Hung
af43ed726fa517b19f900b56965ad29933820545 in linux-6.6.y/6.6.50
ae13c8a5cff92015b9a3eb7cee65ebc75859487f in mainline linux

7 weeks agodrm/amd/display: Fix Coverity INTEGER_OVERFLOW within dal_gpio_service_create
jsg [Mon, 9 Sep 2024 08:56:25 +0000 (08:56 +0000)]
drm/amd/display: Fix Coverity INTEGER_OVERFLOW within dal_gpio_service_create

From Hersen Wu
313d3dd4ca0e898897f4255a4c1f3d9c1f400fe2 in linux-6.6.y/6.6.50
c6077aa66fa230d12f37fef01161ef080d13b726 in mainline linux

7 weeks agodrm/amd/display: Fix Coverity INTERGER_OVERFLOW within construct_integrated_info
jsg [Mon, 9 Sep 2024 08:54:26 +0000 (08:54 +0000)]
drm/amd/display: Fix Coverity INTERGER_OVERFLOW within construct_integrated_info

From Hersen Wu
13faa9d401a5aa7a3b5211adb8ba0e9dddecaabb in linux-6.6.y/6.6.50
176abbcc71952e23009a6ed194fd203b99646884 in mainline linux

7 weeks agodrm/amd/display: Check msg_id before processing transcation
jsg [Mon, 9 Sep 2024 08:52:26 +0000 (08:52 +0000)]
drm/amd/display: Check msg_id before processing transcation

From Alex Hung
cb63090a17d3abb87f132851fa3711281249b7d2 in linux-6.6.y/6.6.50
fa71face755e27dc44bc296416ebdf2c67163316 in mainline linux

7 weeks agodrm/amd/display: Check num_valid_sets before accessing reader_wm_sets[]
jsg [Mon, 9 Sep 2024 08:51:04 +0000 (08:51 +0000)]
drm/amd/display: Check num_valid_sets before accessing reader_wm_sets[]

From Alex Hung
21f9cb44f8c60bf6c26487d428b1a09ad3e8aebf in linux-6.6.y/6.6.50
b38a4815f79b87efb196cd5121579fc51e29a7fb in mainline linux

7 weeks agodrm/amd/display: Add array index check for hdcp ddc access
jsg [Mon, 9 Sep 2024 08:49:00 +0000 (08:49 +0000)]
drm/amd/display: Add array index check for hdcp ddc access

From Hersen Wu
8b5ccf3d011969417be653b5a145c72dbd30472c in linux-6.6.y/6.6.50
4e70c0f5251c25885c31ee84a31f99a01f7cf50e in mainline linux

7 weeks agodrm/amd/display: Check index for aux_rd_interval before using
jsg [Mon, 9 Sep 2024 08:47:23 +0000 (08:47 +0000)]
drm/amd/display: Check index for aux_rd_interval before using

From Alex Hung
48e0b68e2360b16edf2a0bae05c0051c00fbb48a in linux-6.6.y/6.6.50
9ba2ea6337b4f159aecb177555a6a81da92d302e in mainline linux

7 weeks agodrm/amd/display: Stop amdgpu_dm initialize when stream nums greater than 6
jsg [Mon, 9 Sep 2024 08:45:55 +0000 (08:45 +0000)]
drm/amd/display: Stop amdgpu_dm initialize when stream nums greater than 6

From Hersen Wu
94cb77700fa4ae6200486bfa0ba2ac547534afd2 in linux-6.6.y/6.6.50
84723eb6068c50610c5c0893980d230d7afa2105 in mainline linux

7 weeks agodrm/amd/display: Check gpio_id before used as array index
jsg [Mon, 9 Sep 2024 08:44:25 +0000 (08:44 +0000)]
drm/amd/display: Check gpio_id before used as array index

From Alex Hung
08e7755f754e3d2cef7d3a7da538d33526bd6f7c in linux-6.6.y/6.6.50
2a5626eeb3b5eec7a36886f9556113dd93ec8ed6 in mainline linux

7 weeks agodrm/amdgpu: avoid reading vf2pf info size from FB
jsg [Mon, 9 Sep 2024 08:42:39 +0000 (08:42 +0000)]
drm/amdgpu: avoid reading vf2pf info size from FB

From Zhigang Luo
30e60db4296c75f0d7a09c9beedcdab2e044cc2c in linux-6.6.y/6.6.50
3bcc0ee14768d886cedff65da72d83d375a31a56 in mainline linux

7 weeks agodrm/amd/pm: fix uninitialized variable warnings for vega10_hwmgr
jsg [Mon, 9 Sep 2024 08:41:10 +0000 (08:41 +0000)]
drm/amd/pm: fix uninitialized variable warnings for vega10_hwmgr

From Tim Huang
1a2c89396ddc2f8c4a5ae2dc0ab9618944c8c9f0 in linux-6.6.y/6.6.50
5fa7d540d95d97ddc021a74583f6b3da4df9c93a in mainline linux

7 weeks agodrm/amd/pm: fix the Out-of-bounds read warning
jsg [Mon, 9 Sep 2024 08:39:21 +0000 (08:39 +0000)]
drm/amd/pm: fix the Out-of-bounds read warning

From Jesse Zhang
f1e261ced9bcad772a45a2fcdf413c3490e87299 in linux-6.6.y/6.6.50
12c6967428a099bbba9dfd247bb4322a984fcc0b in mainline linux

7 weeks agodrm/amd/pm: Fix negative array index read
jsg [Mon, 9 Sep 2024 08:36:51 +0000 (08:36 +0000)]
drm/amd/pm: Fix negative array index read

From Jesse Zhang
4711b1347cb9f0c3083da6d87c624d75f9bd1d50 in linux-6.6.y/6.6.50
c8c19ebf7c0b202a6a2d37a52ca112432723db5f in mainline linux

7 weeks agodrm/amd/pm: fix warning using uninitialized value of max_vid_step
jsg [Mon, 9 Sep 2024 08:35:23 +0000 (08:35 +0000)]
drm/amd/pm: fix warning using uninitialized value of max_vid_step

From Jesse Zhang
59dd0d4b33269c5173c648cb2c62f9649618e766 in linux-6.6.y/6.6.50
17e3bea65cdc453695b2fe4ff26d25d17f5339e9 in mainline linux

7 weeks agodrm/amd/pm: fix uninitialized variable warning for smu8_hwmgr
jsg [Mon, 9 Sep 2024 08:33:40 +0000 (08:33 +0000)]
drm/amd/pm: fix uninitialized variable warning for smu8_hwmgr

From Tim Huang
1dbce92f325060551f98629e885e18fcc99fa012 in linux-6.6.y/6.6.50
86df36b934640866eb249a4488abb148b985a0d9 in mainline linux

7 weeks agodrm/amd/pm: fix uninitialized variable warning
jsg [Mon, 9 Sep 2024 08:31:19 +0000 (08:31 +0000)]
drm/amd/pm: fix uninitialized variable warning

From Jesse Zhang
fc0cb02efdfbf8f913417bb06ac16a1a985523e7 in linux-6.6.y/6.6.50
7c836905520703dbc8b938993b6d4d718bc739f3 in mainline linux

7 weeks agodrm/amdgpu/pm: Check the return value of smum_send_msg_to_smc
jsg [Mon, 9 Sep 2024 08:29:45 +0000 (08:29 +0000)]
drm/amdgpu/pm: Check the return value of smum_send_msg_to_smc

From Ma Jun
a2f2beaba783e5e99b05bb455b701257e6f1fa37 in linux-6.6.y/6.6.50
579f0c21baec9e7506b6bb3f60f0a9b6d07693b4 in mainline linux

7 weeks agodrm/amdgpu: fix overflowed array index read warning
jsg [Mon, 9 Sep 2024 08:27:57 +0000 (08:27 +0000)]
drm/amdgpu: fix overflowed array index read warning

From Tim Huang
da22d1b98d9d3a1bfd62bd291b10d57b0f19d6e5 in linux-6.6.y/6.6.50
ebbc2ada5c636a6a63d8316a3408753768f5aa9f in mainline linux

7 weeks agodrm/amd/display: Assign linear_pitch_alignment even for VM
jsg [Mon, 9 Sep 2024 08:26:08 +0000 (08:26 +0000)]
drm/amd/display: Assign linear_pitch_alignment even for VM

From Alvin Lee
d2fe7ac613a1ea8c346c9f5c89dc6ecc27232997 in linux-6.6.y/6.6.50
984debc133efa05e62f5aa1a7a1dd8ca0ef041f4 in mainline linux

7 weeks agodrm/amdgpu: Fix uninitialized variable warning in amdgpu_afmt_acr
jsg [Mon, 9 Sep 2024 08:22:35 +0000 (08:22 +0000)]
drm/amdgpu: Fix uninitialized variable warning in amdgpu_afmt_acr

From Ma Jun
761964b756c6e760e8af25e383b7f0a8cc473ac0 in linux-6.6.y/6.6.50
c0d6bd3cd209419cc46ac49562bef1db65d90e70 in mainline linux

7 weeks agodrm: panel-orientation-quirks: Add quirk for OrangePi Neo
jsg [Mon, 9 Sep 2024 08:21:03 +0000 (08:21 +0000)]
drm: panel-orientation-quirks: Add quirk for OrangePi Neo

From Philip Mueller
512bd0cd535bf9b819e638585d5fc69e97d2663b in linux-6.6.y/6.6.50
d60c429610a14560085d98fa6f4cdb43040ca8f0 in mainline linux

7 weeks agodrm/fb-helper: Don't schedule_work() to flush frame buffer during panic()
jsg [Mon, 9 Sep 2024 08:18:38 +0000 (08:18 +0000)]
drm/fb-helper: Don't schedule_work() to flush frame buffer during panic()

From Qiuxu Zhuo
d5618eaea8868e2534c375b8a512693658068cf8 in linux-6.6.y/6.6.50
833cd3e9ad8360785b6c23c82dd3856df00732d9 in mainline linux

7 weeks agoMake error 235 resolve to "no application protocol"
tb [Mon, 9 Sep 2024 07:40:03 +0000 (07:40 +0000)]
Make error 235 resolve to "no application protocol"

We accidentally have two errors 235 since we didn't notice that OpenSSL
removed the unused SSL_R_TRIED_TO_USE_UNSUPPORTED_CIPHER and later that
becamse SSL_R_NO_APPLICATION_PROTOCOL. Getting an "unsupported cipher"
error when fiddling with ALPN is confusing, so fix that.

ok jsing

7 weeks agoDocument 'socket' and 'sockbuf' structures locking.
mvs [Mon, 9 Sep 2024 07:38:45 +0000 (07:38 +0000)]
Document 'socket' and 'sockbuf' structures locking.

`so_oobmark' marked as [mr]. It is accessed in reception path and
corresponding SS_RCVATMARK flag belongs `so_rcv' buffer. However, it is
still protected by exclusive solock()/netlock.

No functional changes.

ok bluhm

7 weeks agoDon't take netlock while setting `if_description'.
mvs [Mon, 9 Sep 2024 07:37:47 +0000 (07:37 +0000)]
Don't take netlock while setting `if_description'.

net/if_pppx.c is the only place where `if_description' accessed outside
ifioctl() path and there is no reason to take netlock here. SIOCSIFDESCR
case of ifioctl() modifies `if_description' with the only kernel lock.

ok bluhm

7 weeks agoAdd triple-dots to synopsis as multiple name[=value] arguments may be given
kn [Mon, 9 Sep 2024 05:36:17 +0000 (05:36 +0000)]
Add triple-dots to synopsis as multiple name[=value] arguments may be given

OK jmc sobrado

7 weeks agoFix alert callback in the QUIC layer
tb [Mon, 9 Sep 2024 03:55:55 +0000 (03:55 +0000)]
Fix alert callback in the QUIC layer

Only close_notify and user_cancelled are warning alerts. All others
should be fatal. In order for the lower layers to behave correctly,
the return code for fatal alerts needs to be TLS13_IO_ALERT instead
of TLS13_IO_SUCCESS.

Failure to signal handshake failure in the public API led to a crash
in HAProxy when forcing the tls cipher to TLS_AES_128_CCM_SHA256 as
found by haproxyfred while investigating
https://github.com/haproxy/haproxy/issues/2569

Kenjiro Nakayama found misbehavior of ngtcp2-based servers, wrote a
similar patch and tested this version.

Fixes https://github.com/libressl/portable/issues/1093

ok jsing

7 weeks agomove some server only functions from nfs_subs.c to nfs_srvsubs.c
jsg [Mon, 9 Sep 2024 03:50:14 +0000 (03:50 +0000)]
move some server only functions from nfs_subs.c to nfs_srvsubs.c
saves space on ramdisks that have option NFSCLIENT

initial diff from and ok miod@

7 weeks agoAdd and use tls13_record_layer_alert_sent()
tb [Mon, 9 Sep 2024 03:32:29 +0000 (03:32 +0000)]
Add and use tls13_record_layer_alert_sent()

This is a small refactoring that wraps a direct call to the record layer's
alert_sent() callback into a handler for upcoming reuse in the QUIC code.

No functional change.

ok jsing

7 weeks agotest mlkem768x25519-sha256
djm [Mon, 9 Sep 2024 03:13:39 +0000 (03:13 +0000)]
test mlkem768x25519-sha256