BSD Commit Log Search

   bump x11/stumpwm after lang/sbcl update

   add regress test for servers that hang durig banner write;
   from Thomas Yiu

   openssh-10.5

   remove ml-dsa/x25519 from the default algorithm list for release
   (it will be back soon)

   make reading server banner non-blocking; prevents one stuck server
   from blocking a many-host keyscan; from Thomas Yiu

   ok dtucker@ markus@

   emulators/citra: switch aarch64 hwcap detection to elf_aux_info()

   Also, re-enable on aarch64.  It builds there with LLVM 22.

   hdestroy(3): stop freeing key values

   POSIX leaves ownership of ENTRY.key with the caller. Our current behaviour
   came from NetBSD when this code was imported in 2004. NetBSD and FreeBSD
   both stopped freeing key values in July 2014. NetBSD in commit fe0a1a6 and
   FreeBSD in commit e19f362.

   Remove the call to free(3) to align ourselves with NetBSD, FreeBSD and
   POSIX. Also adjust rsync for the change in semantics.

   Flagged by Sortix os-test.

   ok millert@, deraadt@

   update p5-Role-Tiny to 2.002005

   update p5-CryptX to 0.091

   rune obsolete /usr/libexec/cpp

   ok deraadt@

   update to 2026.08.151.1

   update to 151.0.7922.108

   Check HMAC() return value to avoid a later use of uninitialized data

   CID 25423

   ok tb@

   limits.h: bump _POSIX_THREAD_THREADS_MAX to 64

   This matches POSIX and the other BSDs. The mismatch was flagged by Sortix
   os-test.

   ok guenther@, asou@, deraadt@

   update p5-Module-Pluggable to 6.4

   lang/deno: Update to 2.9.5

   avoid bogus GnuTLS dep; from Brad, ok kurt@

   don't pick up gtest if present during configure, to avoid issues with junking
   found by naddy

   update p5-Crypt-OpenSSL-X509 to 2.1.5

   devel/difftastic: Update to 0.70.0

   */*: Sync up LZMA CpuArch patches with upstream LZMA SDK

   from Brad

   blake3 now looks for CXX_20 during configure. Shift to ports-gcc
   on base-gcc arches to fix build on sparc64

   sync

   acpidmar(4): Store PCI domain in softc for segment mapping

   Ensure we get a valid segement and enforce use of the IOMMU.

   While there:
    - tweak printf to use segment instead of PCI domain
    - use ACPI provided segment instead of hardcoded 0 in ivhd_showpage()

   ok kettenis@

   lang/lfe: Update to 2.2.2

   Update oxipng to 10.2.0.

   iked: avoid parameter overload

   Old ca_validate_cert() API interpreted data pointer as parsed X509
   structure if len == 0 and crashed if a zero size certificate buffer
   was passed instead.  Now we have an explicit certifcate and key
   parameter.

   from hshoexer@ markus@; ok bluhm@ tb@ tobhe@

   this is errata/7.9/009_iked.patch.sig

   iked: avoid parameter overload

   Old ca_validate_cert() API interpreted data pointer as parsed X509
   structure if len == 0 and crashed if a zero size certificate buffer
   was passed instead.  Now we have an explicit certifcate and key
   parameter.

   from hshoexer@ markus@; ok bluhm@ tb@ tobhe@

   this is errata/7.8/045_iked.patch.sig

   Fix LCP MRU option handing.  Pointed out by Arthur Chan.

   from yasuoka@; ok mvs deraadt

   this is errata/7.9/008_npppd.patch.sig

   Fix LCP MRU option handing.  Pointed out by Arthur Chan.

   from yasuoka@; ok mvs deraadt

   this in errata/7.8/044_npppd.patch.sig