brace_subst: fix single NUL byte overflow after reallocarray()
The check for sufficient space when performing brace substitution
did not take into account space for the terminating NUL byte.
From Thomas Habets
find: Fix mix of character block size check
If multiple -size primaries were specified and any of the values
had the 'c' suffix, it was applied to all values and not just the
one that ended with 'c'. The divsize setting is now stored in the
plan, not a global.
From Thomas Habets
usr.sbin/ifstated: timestamp sync external tests
When external_exec() runs a test synchronously, it refreshes prevstatus
but left lastexec unchanged. A later status change could therefore be
ignored by eval_state(), because that code gates body evaluation on the
first external test timestamp for the current state.
Set lastexec together with prevstatus for synchronous runs; this keeps
the cached test status and execution time consistent across state
re-entry.
OK: deraadt@
Introduce a maximum size for a single flowspec NLRI of 4000 bytes.
Enforce this in both in the parser but also in the RDE.
In the RDE flowspec_valid() will error out if the lenght is too long
but also pt_get_flow() and pt_add_flow() will error out.
The fixed buffer in pt_get_flow() is now sized appropriately instead
of using a arbitrary size.
OK tb@
rpki-client: switch subordinates in CCR manifests to a simple queue
Using a simple list meant that we reversed the order on parsing, resulting
in confusing output in file mode. This is output cosmetics: the CCR DER is
unaffected.
Problem spotted by/ok job (who also wrote the exact same diff)
Various fixes to the flowspec parser
- handle PROTO proto_item like PROTO { proto_item }.
- The LENGTH spec needs to be adjusted, curflow->type needs to be set
before evaluating lengthspec.
- Twice $$ is used in error checking before $$ was set. Use $1 instead.
- extra semicolon after flags
OK tb@
Change CH_LOCATE equality function to be named _eq
Reduce risk of using e.g. memcmp there. _eq returns 1 for match and
0 for no match. Also make all arguemnts const. This makes it a bit
unclear which argument is which. First argument is from the hash table
the second argument is the arg passed to CH_LOCATE.
OK tb@
Avoid calling bus_dmamap_destroy() from interrupt context in qwx(4).
Keep Rx TID DMA memory allocated until the interface goes down,
rather than freeing DMA memory in interrupt context.
Fixes splassert failures reported in dmesg because of qwx(4), such as:
splassert: vm_map_assert_wrlock_ln: want 0 have 4
splassert: uvm_mapent_alloc: want 0 have 4
splassert: assertwaitok: want 0 have 4
Root cause of problem in qwx(4) pointed out by kettenis@, who also
reviewed several diffs until I arrived at a correct fix.
ok kettenis@ phessler@
x509_addr: do not call memcmp() on NULL
If the minimum length is 0, either a->data or b->data could be NULL, so
do not call memcmp() and let the length comparison decide. Doing it this
way preserves the RFC 3779, section 2.2.3.3 semantics and avoids the UB.
A valid IPAddressFamily has an addressFamily element of 2 or 3 octets:
2 octets for the AFI and 1 octet for the optional SAFI. The check as
it is written compares the AFIs and, if they're equal, lets absent SAFI
be smaller than any other SAFI. So IPv4 (0x0001) sorts before IPv4
unicast (0x000101) and that in turn sorts before IPv6 (0x0002).
Found by beck while breaking OpenSSL
ok kenjiro
chacha: avoid -Wunterminated-string-initialization warning
The sizes of sigma[] and tau[] aren't used, so include a trailing NUL and
thereby avoid upsetting modern compilers about use of dangerous, valid C.
ok deraadt djm
drm/amdgpu/vcn4: Avoid overflow on msg bound check
From Benjamin Cheng
271cd5429513ff9b364a9bf8903e5b65b687eb25 in linux-6.18.y/6.18.32
65bce27ea6192320448c30267ffc17ffa094e713 in mainline linux
drm/amdgpu/vcn3: Avoid overflow on msg bound check
From Benjamin Cheng
016b64a0313ea5346cf526e30c8d3e66aca10175 in linux-6.18.y/6.18.32
e6e9faba8100628990cccd13f0f044a648c303cf in mainline linux
drm/amdgpu: rework how we handle TLB fences
From Alex Deucher
f0d3bb6f5125c69ce872a7574afd6d9e1f7ea349 in linux-6.18.y/6.18.32
69c5fbd2b93b5ced77c6e79afe83371bca84c788 in mainline linux
Revert "drm/amdgpu: don't attach the tlb fence for SI"
From Prike Liang
13e9ea445191390269a6f092d009b68e086e601b in linux-6.18.y/6.18.32
9163fe4d790fb4e16d6b0e23f55b43cddd3d4a65 in mainline linux
drm/amdgpu: Fix validating flush_gpu_tlb_pasid()
From Timur Kristof
576b73b246564c8fa75572d45c1ed9b053b4be5b in linux-6.18.y/6.18.32
e3a6eff92bbd960b471966d9afccb4d584546d17 in mainline linux
drm/amdgpu: validate the flush_gpu_tlb_pasid()
From Prike Liang
fd78e63459ab6fe99b9fd026c420808d09eaa8be in linux-6.18.y/6.18.32
f4db9913e4d3dabe9ff3ea6178f2c1bc286012b8 in mainline linux
drm/amdgpu/pm: align Hawaii mclk workaround with radeon
From Alex Deucher
f8f546883e8777493157ccc4fac5794c502fcbd2 in linux-6.18.y/6.18.32
1987c79b4fe5789dfa14423e78b5c25f6acf3e9d in mainline linux
