Unlock ip6_sysctl().
Use temporary local buffer for lockless IO within ip6_sysctl_soiikey().
Use existing `sysctl_lock' rwlock(9) to protect `ip6_soiikey'. Also,
replace temporary `ip_sysctl_lock' with `sysctl_lock' rwlock(9). It was
introduced as temporary to avoid recursive lock acquisition in
ip*_sysctl() until upcoming unlocked. The usage paths are not the hot
paths, no reason to have dedicated locks for them.
ok bluhm
Rework PKCS7_simple_smimecap()
This is nearly identical to CMS_add_simple_smimecap(). We can reuse
its doc comment mutatis mutandis and use the same construction.
Maybe this wants deduplicating. Maybe not.
ok kenjiro
Rework PKCS7_add1_attrib_digest()
There's nothing really wrong here (at least when compared to the rest of
this file an hour or so ago), but we can make this look somewhat more like
code. That there's no bug here is not really related to the fact that it's
an add1 function, not an add0 one. In fact, it's kind of surprising that
the author had an uncharacteristic moment of lucidity and remembered to
free the last argument passed to PKCS7_add_signed_attribute() on failure.
ok kenjiro
Rewrite PKCS7_get_smimecap() to use d2i_X509_ALGORS()
Since we finally found a use for i2d_X509_ALGORS(), make use of its
sibling here. This avoids some ridiculous contortions in not quite
peak muppet code (obviously this was a first test run for the grand
finale in CMS).
ok kenjiro
Plug leaks due to misuse of PKCS7_add_signed_attribute()
set0/add0 functions that can fail are the worst. Without fail this trips
up both users and authors (by and large these are two identical groups
consisting of a single person), resulting in leaks and double frees.
In today's episode of spelunking in the gruesome gore provided by the
PKCS#7 and Time-Stamp protocol "implementations", we fix a couple of
leaks in PKCS7_add_attrib_smimecap() and ESS_add_signing_cert().
We do so by recalling that there is i2d_X509_ALGORS(), so we might
as well put it to use instead of inlining it poorly (aka, without
error checking). Normalize said error checking and ensure ownership
is handled correctly in the usual single-exit idiom.
ESS_add_signing_cert() can also make use of proper i2d handling, so
it's simpler and correct and in the end looks pretty much the same
as PKCS7_add_attrib_smimecap().
ok kenjiro
ensure some strings are actually sent to the parent process
some imsgs messages to the parent are not allowed to send an empty
string. if it happens, fatal() instead of possibly crashing later
on trying to process those strings.
diff from gilles@, committing on his behalf.
ok millert@
PIC code which needs to load data from the GOT and thus has code to setup
r25 for that purpose in the function prologue, needs to mark r1 as needing
to be saved, as the GOT setup relies on a local branch and a pc-relative
relocation.
The comment was correct, the code wasn't - and this was hidden by another bug,
soon to be removed.
Remove local in{,4}_cksum routines and use the MI routines instead. The
MI routines unroll more aggressively for large mbufs and are thus faster.
Be sure to rm in_cksum.d from the compile directory before rebuilding the
kernel.
Revert previous.
With help from gbranden@, the defective commit causing the regression was
found much more quickly than anticipated, and i drafted a patch than can
be used in our upcoming groff-1.23.0 port and that can be pushed upstream,
so SKIP_GROFF is no longer needed for the A1 and w tests.
See https://savannah.gnu.org/bugs/index.php?67372 .
Mark the Dd_opt test with SKIP_GROFF for now because groff-1.23.0 changed
behaviour with respect to groff-1.22.4. This is closely related to the
other .Dd regressions, so it is not yet clear whether mandoc(1) behaviour
should be changed.
drm/amd/display: Free memory allocation
From Clayton King
084eb54b820f737652a0f35ccae67a579b69b0c8 in linux-6.12.y/6.12.40
b2ee9fa0fe6416e16c532f61b909c79b5d4ed282 in mainline linux
drm/amd/display: Disable CRTC degamma LUT for DCN401
From Melissa Wen
74162dda80e7dacffa9642d113fbfdecac01fe79 in linux-6.12.y/6.12.40
97a0f2b5f4d4afcec34376e4428e157ce95efa71 in mainline linux
drm/amdgpu: Increase reset counter only on success
From Lijo Lazar
62f2a58a4cb087f7e06472ae915826485b2dd596 in linux-6.12.y/6.12.40
86790e300d8b7bbadaad5024e308c52f1222128f in mainline linux
drm/amdgpu/gfx8: reset compute ring wptr on the GPU on resume
From Eeli Haapalainen
228ad2ab5b333596ec58bdf925c0bdc8f68365e1 in linux-6.12.y/6.12.40
83261934015c434fabb980a3e613b01d9976e877 in mainline linux
Remove htole64() calls on GPT data. GPT is little-endian as
are all current OpenBSD architectures using UEFI booting.
In the unlikely event that BE architectures using UEFI
booting ever materialize this will be one of many things
that need attention.
Suggested by kettenis@ in response to my original diff.
Below STANDARDS, reference the two most relevant sections of RFC 5652.
Given that RFC 5652 does not override the earlier (and simpler)
standards but instead strives to remain compatible, referencing
both the original and the latest versions seems helpful.
OK tb@