BSD Commit Log Search

   s/sucess/success/

   s/excute/execute/
   s/exection/execution/
   s/exectuables/executables/

   s/microseqence/microsequence/ in documentation and comments.
   s/synchroneous/synchronous/ in comments.

   refactor: make sure tx_info is initilized before dereferencing. reviewed by christos

   fix wrong range from computing the MIN of tx and rx queues.


   needs pullup

   new unbound

   bump libunbound

   merge conflicts, bump

   Import unbound-1-26-2 (previous was 1-26-1)

   Unbound 1.26.2

   This release has a number of security fixes.

   The release is signed with the OpenPGP software signing key that is
   in use since Jan 1st 2026:

   User ID: NLnet Labs releases signing key G2 <releases at nlnetlabs.nl>
   Key ID: A144 323D EAAC DF45
   Fingerprint: 2310 1869 0C4D 903E F419  146A A144 323D EAAC DF45
   The key is available from https://nlnetlabs.nl/signing-keys .

   This release consolidates security fixes for issues reported over
   a period of time. There are fixes for CVE-2026-35501, CVE-2026-50106,
   CVE-2026-96405, CVE-2026-97206 and CVE-2026-109125.

   Bug Fixes:

    [25 lines not shown]

   use thr right size of cahr buffer


   buflen passes the right lengh of the char buffer nd not sizeof(buf)

   certctl(8): Gracefully handle deletion of untrusted certs.

   PR bin/60916: certctl(8) doesn't gracefully handle deleting untrusted
   certs

   certctl(8): Test response to updates that delete untrusted certs.

   PR bin/60916: certctl(8) doesn't gracefully handle deleting untrusted
   certs

   i2c(4): Avoid leaking kernel heap memory to userland.

   PR kern/60913: iic(4): I2C_IOCTL_EXEC can copy out an uninitialised
   buffer

   use PR_ZERO on on pool allcoated memory and get rid of memset.

   currently, xfer is zeroed using the wrong size.
   remove it and simplify code.


   reviewed by christos@

   use the right sizeof struct blake2s instead of sizeof pointer

   alpha: Align kernel stack pointer to 16-byte boundaries

   as ABI requires. With this fix, full ATF run successfully completed
   without any regression.

   PR port-alpha/60865 NetBSD-current/alpha panics running getcwd test

   OK thorpej@

   unbound 1.26.2 is out

   fix a HAVE_GCC test to be >= 12, not just == 12.

   should fix m68ksf build finally.

   Tickets #1368 - #1375

   dtb: add left-out earmv7eb to update-sets

   earmv7eb accidentally did not get updated when regenerating set lists, while earmv7hf got included twice.

   Reported by Elias Howell.

   Pull up following revision(s) (requested by riastradh in ticket #1375):

        common/lib/libc/string/memmem.c: revision 1.6

   memmem(3): Avoid undefined behaviour in shift.

   PR lib/58906: memmem.c: sync with musl upstream that fixes UB on
   signed overflow

   Pull up following revision(s) (requested by riastradh in ticket #1374):

        sys/dev/video.c: revision 1.46

   When dropping lock around scatter_io_uiomove(), use busy flag
   to serialize operations.

   Hold lock while accessing video_stream and video_buffer.

   Pull up following revision(s) (requested by riastradh in ticket #1373):

        crypto/external/bsd/heimdal/Makefile.rules.inc: revision 1.11

   heimdal: fix parallel make includes

   Split out the rules that generate SRC_asn1.h and SRC_asn1-priv.h
   from SRC.asn1 into their own explicit targets that create those files
   from the intermediate SRC.ts, which may help with parallel builds.

   Add MAKEVERBOSE messages to improve debugging.

   PR toolchain/58771: missing dependency breaks parallel build

   Pull up following revision(s) (requested by riastradh in ticket #1372):

        bin/df/df.c: revision 1.104

   PR/58718: Christof Meerwald: "total files" and "free files" values swapped
   in df -G output

   Pull up following revision(s) (requested by riastradh in ticket #1371):

        tests/lib/libm/t_modf.c: revision 1.6
        lib/libm/src/s_modfl.c: revision 1.2
        tests/lib/libm/t_fe_round.c: revision 1.20

   modfl(3): Fix conversion from FreeBSD.

   LDBL_MANL_SIZE is spelled EXT_FRACLBITS -- and not EXT_FRACHBITS.

   PR lib/58237: modfl returns wrong answers on ld128 architectures

   Pull up following revision(s) (requested by riastradh in ticket #1370):

        sys/kern/sys_futex.c: revision 1.20
        sys/kern/sys_futex.c: revision 1.21
        sys/kern/sys_futex.c: revision 1.22
        tests/lib/libc/sys/t_futex_ops.c: revision 1.6
        tests/lib/libc/sys/t_futex_ops.c: revision 1.7
        tests/lib/libc/sys/t_futex_ops.c: revision 1.8
        tests/lib/libc/sys/t_futex_ops.c: revision 1.9
        tests/lib/libc/sys/t_futex_ops.c: revision 1.10

   sys_futex.c: Fix illustration of futex(2).

   In this illustration, we need to _set_ bit 1 to claim ownership, not
   _clear_ bit 1 to claim ownership.

   No functional change intended -- comment only.



    [63 lines not shown]

   Pull up following revision(s) (requested by riastradh in ticket #1369):

        share/man/man4/unix.4: revision 1.31
        share/man/man4/unix.4: revision 1.32
        share/man/man4/unix.4: revision 1.33
        share/man/man4/unix.4: revision 1.34
        share/man/man4/unix.4: revision 1.35


   unix(4): Document options more legibly.

   1. Use `.Bl -tag' rather than just paragraphs upon paragraphs.
   2. Specify the type of each option and how it is to be used.
   3. Add LOCAL_CONNWAIT.
   4. Slim down the comment spacing a little to fit.

   PR kern/32844: unix domain sockets documentation is incomplete



    [30 lines not shown]

   Pull up following revision(s) (requested by riastradh in ticket #1368):

        share/man/man3/CMSG_DATA.3: revision 1.4
        share/man/man3/CMSG_DATA.3: revision 1.5
        share/man/man3/CMSG_DATA.3: revision 1.6

   cmsg(3): Don't use malloc in example.

   Since CMSG_SPACE (and CMSG_LEN) is required to expand to an integer
   constant expression by POSIX, and it does in NetBSD as of 2012, there
   is no reason to use malloc here -- we can just allocate the buffer on
   the stack without tripping over variable-length arrays.

   PR lib/59054: cmsg(3) uses malloc(3) unnecessarily


   cmsg(3): Handle as many descriptors as the sender sends in example.

   Due to a quirk of alignment, sizing the control buffer for

    [17 lines not shown]

   Fix compilation warning -Wdiscarded-qualifiers

   Ammend ticket #1383 for additional pullup