BSD Commit Log Search

   Add Policy based routing by interface in NPF

   bump libdns

   new bind

   Merge changes between 9.20.18 and 9.20.22

   Import bind-9.20.22 (previous was 9.20.18)

   Notes for BIND 9.20.22
   Security Fixes
   Fix crash when reconfiguring zone update policy during active updates.
   We fixed a crash that could occur when running rndc reconfig to change a zone's
   update policy (e.g., from allow-update to update-policy) while DNS UPDATE
   requests were being processed for that zone. ISC would like to thank Vitaly
   Simonovich for bringing this issue to our attention. [GL #5817]

   Bug Fixes
   Fix intermittent named crashes during asynchronous zone operations.
   Asynchronous zone loading and dumping operations occasionally dispatched tasks
   to the wrong internal event loop. This threading violation triggered internal
   safety assertions that abruptly terminated named. Strict loop affinity is now
   enforced for these tasks, ensuring they execute on their designated threads and
   preventing the crashes. [GL #4882]

   Count temporal problems with DNSSEC validation as attempts.

    [160 lines not shown]

   s/numberered/numbered/ in comment.

   fix various typos, mainly in comments.

   blkdiscard.8: correct a detail in the history section

   s/referened/referenced/ in log message.

   libc/softfloat: prevent undefined behavior in floating point conversions

   See the previous commit for details.

   The same code pattern occur in the bits32 implementation, as well as in
   other copies of the same upstream code, for example in libkern.

   lint: add query for negation of minimum signed value

   This helped double-check that the remaining negations in the
   softfloat/bits64 code are fine. There are some false positives, due to
   int16 being typedef'd to int, thus being int32 realistically.

   tests/libc: document ldbl-to-uint64 conversion for more platforms

   Provide a weak do-nothing for module_init_md().

   tests/libc: test softfloat on sparc64 for converting ldbl to uint64

   Refactor cpu_exec_aout_makecmds().

   mac68k_init() -> machine_init(), and make it fit the pattern.

   bsd.own.mk: fix MKPROFILE comment for risc-v

   Fix test to match expected result on m68k.

   Tested on m68ksf and hardfloat (in qemu).

   Ok Roland.

   akbd(4): s/shoud/should/.

   s/mbr_info_to_partitition/mbr_info_to_partition/ in method name.

   approved by martin@.

   tty.4: trivial grammar/style corrections

   make: clean up, mostly redundant parentheses

   Use machine_init_common().

   Use machine_init_common().

   Use machine_init_common().

   Use machine_init_common().

   lint: don't rely on the floating point rounding mode

   When checking whether a floating point constant can be converted to an
   integer constant with the same value, ULLONG_MAX is not representable
   exactly in a 64-bit long double, so the comparison value could have been
   rounded down, leading to wrong warnings about lossy conversions.

   Perform exact calculations instead.

   Remove duplicate pmap_md_direct_map_paddr declaration.

   Simulate keyboard leds.

   Keep led state to set/get via ioclts.

   While here clean up adb probed led state (currently unused).

   lint: remove temporary code to identify floating point bug

   On sparc64, SoftFloat 2a invoked undefined behavior, returning wrong
   results. The code added here didn't trigger, as the ATF_SH environment
   variable was not set in the real test run.

   See tests/lib/libc/t_long_double.c for more specialized tests to
   pinpoint the exact problem.