Monday Patch Queue Review update (2025-09-08)

[Carlos O'Donell]

Most recent meeting status is always here:
https://sourceware.org/glibc/wiki/PatchworkReviewMeetings#Update

Meeting: 2025-09-09 @ 0900h EST5EDT

Video/Audio: https://bbb.linuxfoundation.org/room/adm-alk-1uu-7fu

IRC: #glibc on OFTC.

Review new patches and restart review at the top.

​ * State NEW delegate NOBODY at 533.
  * Starting 119762.
  * atomic: Switch to builtin atomics (Wilco)
   * Reviewer Pool: Adhemerval.
  * [1/5] x86: Remove unused atomics (Uros)
   * https://sourceware.org/cgit/glibc/commit/?id=60d5e40ab200033a982a9fd7594a1f83dcdb94a0
x86: Remove low-level lock optimization
   * No low-level sing-threaded optimization, so this can be reviewed.
   * HJ noted he will review. Assigned.
  * [v6] malloc: Mark pages with MADV_DONTNEED to shrink and grow rather than mremap() (William)
   * Reviewer Pool: Wilco.
   * Discussed Florian have a look specifically at the overcommit piece given previous v5 comments.
  * alpha: Fix missing inexact-flag raising for lround/lrint (Adhemerval)
   * Reviewer Pool: Carlos.
  * math: Fix x86_64 build for -Os (BZ 33367) (Adhemerval)
   * Review assigned to HJ.
  * [v2] i686: Compile .op files and gmon tests with -mfentry (HJ)
   * Leaving for Joseph to comment on.
  * aarch64:Modify the copy_long function in the SVE memcpy implementation for 32-byte aligned access (Longwei)
   * Changes requested by Wilco to add glibc microbenchmark data on the runs of Kunpeng hardware.
  * [v4] malloc: avoid need for tcache == NULL checks (DJ)
   * Wilco: If this is an intermediate step and doesn't impact performance then that would be great.
   * Florian: May you please check the performance issues on the hardware you care about?
   * Next step is for Wilco to check performance.
  * Stopping at 119506.
  * Discussed aarch64 outstanding patch from Yury.
  * Discussed syscall and clone and manual update.
  * Discussed running testsuite with clang20.
   * HJ filed https://sourceware.org/bugzilla/show_bug.cgi?id=33322 for review.
   * Discussed splitting test issues from test build issues (issues building the tests with clang20).
  * Discussed CORE-MATH update from Adhemerval.
   * Standard describes cr* variants.
   * Use non-cr versions for performance?
   * If we have really really bad precision we should update e.g. lgamma/tgamma.

-- 
Cheers,
Carlos.