Differences between revisions 112 and 113
Revision 112 as of 2022-08-01 13:48:05
Size: 109357
Comment:
Revision 113 as of 2022-08-08 21:00:05
Size: 112160
Comment:
Deletions are marked like this. Additions are marked like this.
Line 33: Line 33:
== Next Meeting: 2022-08-08 @ 0900h EST5EDT == == Next Meeting: 2022-08-15 @ 0900h EST5EDT ==
Line 40: Line 40:
== Meeting: 2022-08-08 @ 0900h EST5EDT ==
<<Anchor(Update)>>
Video/Audio: https://bluejeans.com/9093064454

IRC: #glibc on OFTC.

Review new patches and attendee requested reviews.

 - Status NEW delegate NOBODY at 218 patches.
 - Starting 56589 (last start 56467)
 - [v1] elf: Replace `strcpy` call with `memcpy` [BZ #29454] (Noah)
  - Needs review.
 - [v1] x86: Fix `#define STRCPY` guard in strcpy-sse2.S (Noah)
  - HJ to review?
 - [RFC] elf: fine-grained output LD_DEBUG log when symbol lookup error
  - Florian to point out this doesn't do what they want.
 - stdlib: tests: don't double-define _FORTIFY_SOURCE (Sam)
  - Siddhesh to review.
  - Adhemerval: Do we support building glibc with _FORTIFY_SOURCE=2 on
    by default?
  - Tested by Gentoo and it's working.
 - [RFC] RISC-V: Implement {convert,round}toint() (Palmer)
 - [v3] Remove catomics (Wilco)
  - Adhemerval to comment on a strategy that can involve staggered removal
 - [v1] stdlib: Add more room for reuse of random bits in arc4random_uniform
  - Given Noah's comments marked as dropped.
 - x86-64: Restore LD_PREFER_MAP_32BIT_EXEC support [BZ #28656] (HJ)
  - Florian will continue discussion with HJ about kernel support.
 - feat: crypt support sm3 algorithm (Mingjun)
  - Carlos marked rejected because it needs to go to libxcrypt.
 - resolv: Set a timeout for TCP-based query read (bug 19643)
 - Skip to 56275 to review older bugs.
 - [V2] resolv: use arc4random_uniform in the generation of random ids
  - Carlos marked as changes requested.
 - longlong: Update the checks for "mul" on RISC-V (Palmer)
  - Reviewed by Andreas. Marked Changes Requested.
 - inet: Turn __ivaliduser into a compatibility symbol (Florian)
  - Needs a v2. Marked changes requested.
 - Stopped at 56223.
 - Szabolcs: There are licensing discrepancies between various start.S
   assembly files for architectures and this should be corrected in order
   to move towards the expected license behaviour.
  - Carlos: I agree. We can take this up with the FSF to adjust.
 - Florian: Fixing the mount.h header file conflict.
  - Adhemerval: Discussed this briefly.
  - Siddhesh: It's sys/fs and linux/fs that breaks things and I've
    filed several bugs for maintainers to move one way or the other.
  - Florian: qemu developers have had to copy things.
  - Ahdemerval: If the user needs the kernel interface then they should
    use syscalls directly and should use the linux headers directly.
    Why do we need to mix the header usage?
  - Florian: We don't have a wrapper header for linux/fs.h.
  - Adhemerval: We can work on a fix.
 - Adhemerval: How are we going to handle the posix_spawn issue?
  - Florian: Linux 6.0 will have the vfork+exec fix.
  - Florian: I'll file some backport requests for the kernel.
Line 41: Line 98:
<<Anchor(Update)>>

Patchwork Review Meetings

The community uses a Patchwork v3 instance to track community submitted patches.

Everyone in the community is invited to attend a weekly Monday patchwork review meeting.

The meetings will be held at the following times:

  • Monday at 0900h EST5EDT

Other meeting times may be added in the future to cover additional timezones.

The focus of the meeting is to review the patchwork queue here: https://patchwork.sourceware.org/project/glibc/list/

The intent is to find reviewers for patches and move patches forward along one of the states in the patchwork workflow: https://sourceware.org/glibc/wiki/Patch%20Review%20Workflow

Please set your Profile->Items per page to "500" for the meeting.

Please set your Profile->Show Patch IDs enabled to allow patch IDs to be used for meeting review.

Meetings will not be recorded, but notes from the meetings will be taken and recorded on this page.

Process:

  • Show patches with: State = NEW, Delegate = Nobody
  • Review newest patches looking for critical action items.
    • Record the patch at which review started.
  • Delegate patches for review throughout the meeting.
  • Record the patch at which review stopped.
    • This allows external observers to know which range of patches was reviewed and coordinate review.
  • At the 30 minute mark go through the attendees and request items that they would like to discuss for review.

Next Meeting: 2022-08-15 @ 0900h EST5EDT

Video/Audio: https://bluejeans.com/9093064454

IRC: #glibc on OFTC.

Review new patches and attendee requested reviews.

Meeting: 2022-08-08 @ 0900h EST5EDT

Video/Audio: https://bluejeans.com/9093064454

IRC: #glibc on OFTC.

Review new patches and attendee requested reviews.

  • - Status NEW delegate NOBODY at 218 patches. - Starting 56589 (last start 56467)

    - [v1] elf: Replace strcpy call with memcpy [BZ #29454] (Noah)

    • - Needs review.

    - [v1] x86: Fix #define STRCPY guard in strcpy-sse2.S (Noah)

    • - HJ to review?
    - [RFC] elf: fine-grained output LD_DEBUG log when symbol lookup error
    • - Florian to point out this doesn't do what they want.
    - stdlib: tests: don't double-define _FORTIFY_SOURCE (Sam)
    • - Siddhesh to review. - Adhemerval: Do we support building glibc with _FORTIFY_SOURCE=2 on
      • by default?
      - Tested by Gentoo and it's working.
    - [RFC] RISC-V: Implement {convert,round}toint() (Palmer) - [v3] Remove catomics (Wilco)
    • - Adhemerval to comment on a strategy that can involve staggered removal
    - [v1] stdlib: Add more room for reuse of random bits in arc4random_uniform
    • - Given Noah's comments marked as dropped.
    - x86-64: Restore LD_PREFER_MAP_32BIT_EXEC support [BZ #28656] (HJ)
    • - Florian will continue discussion with HJ about kernel support.
    - feat: crypt support sm3 algorithm (Mingjun)
    • - Carlos marked rejected because it needs to go to libxcrypt.
    - resolv: Set a timeout for TCP-based query read (bug 19643) - Skip to 56275 to review older bugs. - [V2] resolv: use arc4random_uniform in the generation of random ids
    • - Carlos marked as changes requested.
    - longlong: Update the checks for "mul" on RISC-V (Palmer)
    • - Reviewed by Andreas. Marked Changes Requested.

    - inet: Turn ivaliduser into a compatibility symbol (Florian)

    • - Needs a v2. Marked changes requested.
    - Stopped at 56223. - Szabolcs: There are licensing discrepancies between various start.S
    • assembly files for architectures and this should be corrected in order to move towards the expected license behaviour.
    • - Carlos: I agree. We can take this up with the FSF to adjust.
    - Florian: Fixing the mount.h header file conflict.
    • - Adhemerval: Discussed this briefly. - Siddhesh: It's sys/fs and linux/fs that breaks things and I've
      • filed several bugs for maintainers to move one way or the other.
      - Florian: qemu developers have had to copy things. - Ahdemerval: If the user needs the kernel interface then they should
      • use syscalls directly and should use the linux headers directly. Why do we need to mix the header usage?
      - Florian: We don't have a wrapper header for linux/fs.h. - Adhemerval: We can work on a fix.
    - Adhemerval: How are we going to handle the posix_spawn issue?
    • - Florian: Linux 6.0 will have the vfork+exec fix. - Florian: I'll file some backport requests for the kernel.

Meeting: 2022-08-01 @ 0900h EST5EDT

Video/Audio: https://bluejeans.com/9093064454

IRC: #glibc on OFTC.

Review new patches and attendee requested reviews.

  • NEW / NOBODY at 219 patches.
  • Release 2.36 branch fixes needed?
  • Starting 56467
  • Internal asserts cleanups (Florian)
    • Not immediately urgent.
  • malloc: Use getrandom_nocancel during tcache initiailization (Florian)

  • Remove spurious references to _dl_open_hook (Florian)
  • feat: crypt support sm3 algorithm (Yang)
    • Discussed rejection of patch due to deprecation of libcrypt in favour
      • of libxcrypt. Need to adjust library for 2.37 to remove crypt routines. This is similar to Sun RPC deprecation.
  • [v4] socket: Check lengths before advancing pointer in CMSG_NXTHDR (Arjun)
    • OK to commit. Has RB.
  • [v2] stdlib: Simplify arc4random_uniform
    • OK to commit. Has RB.
  • resolv: Set a timeout for TCP-based query read (bug 19643) (Sinan)
    • Florian to review.
  • [v4,1/1] memalign: Support scanning for aligned chunks. (DJ)
    • Carlos to review with DJ next week.
    • Florian noted we need to process tcache chunks.
  • Remove atomic-machine headers (Wilco)
    • Patch doesn't apply.
  • Use atomic_thread_fence (Wilco)
    • Patch doesn't apply.
  • GLIBC LOONGARCH PATCHES (Yu)
    • Needs review for gettimeofday vDSO addition.
  • [v2] NEWS: Add entry for x86-64 ISA level build (Noah)
    • Dropped.
  • [v2] nscd: Fix netlink cache invalidation if epoll is used [BZ #29415]
    • Reviewed by DJ. Can be committed with 2.37 open.
  • [V2] resolv: use arc4random_uniform in the generation of random ids (Cristian)
    • This needs a v3.
  • Stopped at 56275.
  • Any issues to raise for review:
  • Adhemerval: [v3] debug: test for more required cacellation points (BZ# 29274) (Adhemerval)
    • Carlos needs to review.

Meeting: 2022-07-25 @ 0900h EST5EDT

Video/Audio: https://bluejeans.com/9093064454

IRC: #glibc on OFTC.

Review new patches and attendee requested reviews.

Meeting: 2022-07-18 @ 0900h EST5EDT

Video/Audio: https://bluejeans.com/9093064454

IRC: #glibc on OFTC.

Review new patches and attendee requested reviews.

  • Status NEW / NOBODY is at 229 patches.
  • Starat at 56119.
  • Discussing ABI issues for glibc 2.36
    • LoongArch patches need a new version.

      • Very contained to just the architecture patches.
      • codonell: Will ask them to send a new version for testing.
    • arc4random needs review.
      • Testing has happened on all real hardware.
      • fweimer: Do we label the function as crypto secure?
      • fweimer: Anyone certifying OpenBSD for use in FIPS backends
        • will either have to rewrite the openbsd impl or
          • note that it should not be used crypto secure.
      • adhemerval: Note that POSIX has an issue out to add arc4random
  • noah: Looking at bug 29327 and impact on glibc 2.36?
  • adhemerval: Szabolcs, May you please check on optimzied arc4random?
  • codonell: I'm out of the office 22-28.
    • Discussed having replacement RM to ack patches into tree.
    • adhemerval accepted doing the role.
  • New reviews at 56119 (226 patches)
  • [v2] x86: Fix typo of Slow_SSE4_2 def in isa-level.h - skip

  • Always do locking when accessing streams (bug 15142, bug 14697)
    • codonell: This needs review for release.
  • [v1] x86: Continue building memmove-ssse3.S as ISA level V3 - skip
  • [v3] linux: return UNSUPPORTED from tst-mount if entering mount namespace fails
    • codonell: Check on push.
  • [v1,1/1] memalign: Support scanning for aligned chunks. (DJ)
    • Fails CI.
  • [v1] nptl: Continue use arch prefered atomic exchange in spinlock loop (need review)
  • [v1] nptl: Replace CAS with cheaper atomics for pthread_cancel logic (need review)
  • [2.37] x86-64: Use ENTRY_P2ALIGN and END in strcmp-XXX.S (waiting)
  • [v2] Update scripts/config.* files from upstream GNU config version
    • codonell: Will review.
  • [v2] debug: test for more required cacellation points (BZ# 29274) (need review)
  • [PATCHv4] tst-pidfd.c: UNSUPPORTED if we get EPERM on valid pidfd_getfd call
    • codonell: To review as RM.
  • elf.h: Add ELFCOMPRESS_ZSTD (Fangrui)
    • codonell: To comment.
  • 55882-55877 LoongArch sequence needs repost with new version.

  • [1/7] Use atomic_exchange_release/acquire (Wilco)
    • Not OK for 2.36. Can review early for 2.37.
  • soft-fp: Add fixhf[uns][di|si] and float[uns][di|si]hf (need review)
    • codonell: This isn't a public ABI/API but glibc is just canonical.
    • Not a release blocker.
  • Stopping new review at 55745.

Meeting: 2022-07-11 @ 0900h EST5EDT

Video/Audio: https://bluejeans.com/9093064454

IRC: #glibc on OFTC.

Review new patches and attendee requested reviews.

  • State NEW delegate NOBODY 217 patches
  • Switching to discussing ABI blockers for release.
  • ABI issues for glibc 2.36
    • arc4random
    • Loongarch
  • [codonell] Kernel syscalls we would add?
    • azanella: Checked to 5.17.
    • Latest two releases may have new syscalls.
    • What about libiouring? Don't want to be tied to libc itself.
  • The 5.19 rc3/4 for kernel ABI is in place for Loongarch.
  • [v8] Add arc4random support (Series 10509)
    • Florian will look at this.
  • [v6] LoongArch (13 patches)

    • fweimer: Is upstream binutils ready?
    • azanella: Built cross toolchain and it looks ready (2.38)
    • fweimer: Upstream binutils issues:
    • azanella: Want a committed upstream binutils and gcc.
      • codonell: Agree, want a committed linux, gcc, and binutils.
    • codonell: Will reach out to Nick Clifton to review LoongArch.

  • patrick: Can we submit regression fixes during the freeze?
    • codonell: Yes, we can fix bugs up until the release.
  • Update miscellaneous files from upstream sources (azanella)
    • Needs review. Built many arches and it was OK.
    • fweimer: config.guess/config.sub
  • elf.h: Add ELFCOMPRESS_ZSTD (fangrui)
    • Not a release blocker.
    • Blocked on Generic ABI or GNU ABI accepting.
  • x86-64: Remove redundant strcspn-generic/strpbrk-generic/strspn-generic
    • Reviewed by Noah.
  • [PATCHv3] Apply asm redirections in stdio.h before first use [BZ #27087]
    • With clang with long double == float128.
    • When Fedora was updated to clang14 we ran into this issue.
    • codonell: Not a release blocker *yet* because upstream defaults don't fail.
  • [1/7] Use atomic_exchange_release/acquire (Wilco)
    • codonell: Impact on aarch64? Can we wait until after the branch?
    • szabolcs: I think so.
  • Stopped at 55764.
  • Any patches to raise?
  • soft-fp: Add fixhf[uns][di|si] and float[uns][di|si]hf
    • codonell: Are we the canonical place for soft-fp?
    • Check with Joseph Myers about canonical place for soft-fp.
    • szabolcs: Confirmed. glibc is upstream.
    • Needs review.
  • 55628 [PATCHv3] tst-pidfd.c: UNSUPPORTED if we get EPERM on pidfd_open or pidfd_getfd
    • Rejected.
    • codonell: I'll respond again.
  • fweimer: Noted that Joseph Myers builder has been failing.

Meeting: 2022-07-04 @ 0900h EST5EDT

Video/Audio: https://bluejeans.com/9093064454

IRC: #glibc on OFTC.

Review new patches and attendee requested reviews.

  • State NEW delegate NOBODY at 237 patches.
  • Discussed having a workable backlog.
  • Started at 55340
  • v7 linux: Add new syscalls (Adhemveral)
    • Carlos to review.
  • build: report an error message when missing test name (Yann Droneaud)
    • Carlos to review.
  • At 55269.
  • Simplify internal single-threaded usage (Adhemerval)
  • At 55129.
  • [patch/idea,v2] Add register scrambling to testsuite (DJ)
    • Needs review.
  • vfprintf refactor (Florian Weimer)
  • Discsuss ABI for glibc 2.36
    • C++20 APIs need review.
    • arc4random adds ABIs.
    • LoongArch new ABI.

    • Predictable ELF destructor ordering (not ABI, but impacts behaviour)
    • Implementation RTLD_SHARED for dlmopen.
  • Stopped review and noted 227 patches in teh queue.

Meeting: 2022-06-27 @ 0900h EST5EDT

Video/Audio: https://bluejeans.com/9093064454

IRC: #glibc on OFTC.

Review new patches and attendee requested reviews.

  • State NEW delegate NOBODY is 243 patches.
  • Started review at 55440.
  • [PATCHv2] tst-pidfd.c: Test is UNSUPPORTED without PTRACE_MODE_ATTACH_REALCREDS (Mark)
    • Yes, we need the credentials. Need to review.
  • [PATCHv2] time/tst-clock2.c: clock_settime CLOCK_MONOTONIC might return EPERM (Mark)
    • It seems correct, but strange that we get EPERM.
  • build: report an error message when missing test name (Yann)
    • Carlos: I will commit this for Yann.
  • [1/4] time/tst-clock2.c: clock_settime CLOCK_MONOTONIC might return EPERM (Mark)
    • I guess this is correct, but it is superseeded.
  • [v5] x86: Fix backwards Prefer_No_VZEROUPPER check in ifunc-evex.h (Noah)
    • Accepted by HJ
  • [v7] linux: Add new syscalls (Adhemerval)
    • Carlos: I will review again.
  • 55381 [v2] x86: Add comment with ISA level for all targets support by GCC12.1
    • Not needed. Marked rejected.
  • 55380-55375 are Noah's patches for the x86 optimizations.
    • Marked as Accepted or Superceded with review from HJ.
  • Stopped at 55340 libc-symbols.h: remove unused macros.
  • 55269 aarch64: Optimize string functions with shrn instruction
    • Carlos: Will comment on Signed-off-by vs. Google copyright assignment.
  • Siddhesh: Patchwork and Django are up to date now.
    • Carlos: This enables us to respond to patchwork comments.
  • Adhemerval: Ping arc4random since it has been on the queue for some time.
    • Looking for an acknowledgement from Florian regarding the arc4random approach.
    • Florian Weimer was the previous author of an attempt to include arc4random.
    • Carlos: I'll follow up with Florian.
    • Siddhesh: I can talk to Florian about hand-off of the series to Siddhesh for review.
  • Carlos: Note that glibc 2.36 needs an RM.

Meeting: 2022-06-20 @ 0900h EST5EDT

Video/Audio: https://bluejeans.com/9093064454

IRC: #glibc on OFTC.

Review new patches and attendee requested reviews.

Meeting: 2022-06-13 @ 0900h EST5EDT

Video/Audio: https://bluejeans.com/9093064454

IRC: #glibc on OFTC.

Review new patches and attendee requested reviews.

Meeting: 2022-06-06 @ 0900h EST5EDT

Video/Audio: https://bluejeans.com/9093064454

IRC: #glibc on OFTC.

Review new patches and attendee requested reviews.

  • State NEW delegate NOBODY 238
  • Start at 54820
  • aarch64: Handle fewer relocations for RTLD_BOOTSTRAP
    • Arm should review for AEABI compliance?
  • riscv: Change the relocations handled for RTLD_BOOTSTRAP
    • Need machine maintainer review.
  • [v4,1/2] nss: add assert to DB_LOOKUP_FCT (BZ #28752)
    • Carlos: Will ask DJ to review.
  • [v1] x86: ZERO_UPPER_VEC_REGISTERS_RETURN_XTEST expect no transactions
    • Noah: Will resubmit after teh VEC classes gets approved.
    • Fails CI because previous series needs application.
  • [v3,1/8] x86: Create header for VEC classes in x86 strings library
    • Noah: HJ should have a look at these.
  • x86_64: Add strstr function with 512-bit EVEX
    • Noah: Looking into this.
  • [v2] nptl_db: disable DT_RELR on libthread_db.so
    • Carlos: Will review.
  • Avoid -Wstringop-overflow= warning in iconv module. \
    • Siddhesh: Will review.
  • resolv: Implement no-aaaa stub resolver option
    • Carlos: Already reviewed. Changes request.
  • New syscalls for mount and namespace
    • Carlos: Adhemerval can you please rebase and send the new syscalls?
  • Add PT_AARCH64_MEMTAG_MTE from Linux 5.18 to elf.h
    • Already reviewed.
  • [v3] elf: Remove ELF_RTYPE_CLASS_EXTERN_PROTECTED_DATA
    • Szabolcs: Semantics have changed with this patch and it needs more
      • serious review. Simple test cases previously worked and now it doesn't.
  • GLIBC LoongArch PATCHES (caiyinyu)

    • Needs review.
    • Carlos: What is the status of the ABI? It has been submitted to musl.
    • Adhemerval: I think musl's ABI is done, but there is kernel discussion.
  • nptl: Fix libc_cleanup_pop_restore asynchronous restore (BZ#29214) (Adhemerval)

    • Carlos: Will review.
    • Broken in 2.35.
  • v3: Fix various NPTL synchronization issues (Adhemerval)
    • Robustness issue, and fixes BZ #19951, which is a real issue.
  • Review ended 54535.
  • Noah: Why can't we use #define to define rseq offset?
    • Noah: Florian isn't here, and I was going to ask him the question.
    • Szabolcs: It's for maintenance reasons. The kernel wanted to grow
      • the size of rseq, and we didn't want to have to define a constant location for all arches, and we didn't want to define a different constant per arch, so we just made it another lookup. You can cache the values in globals if the two TLS accesses are slower than you expect.

Next Meeting: 2022-05-30 @ 0900h EST5EDT

Video/Audio: https://bluejeans.com/9093064454

IRC: #glibc on OFTC.

Review new patches and attendee requested reviews.

  • State NEW delegate NOBODY, 228 patches.
  • Starting at 54476
  • elf: Remove ELF_RTYPE_CLASS_EXTERN_PROTECTED_DATA (Fangrui Song)
    • Needs review. Szabolcs: Same as aarch64.
    • This never worked.
  • [v2] elf: Remove one-default-version check when searching an unversioned symbol (Fangrui Song)
    • Florian made a comment about the underlying semantics.
  • [v1] benchtests: Improve benchtests for strstr, memmem, and memchr (Noah)
    • Already pushed.
  • posix: Fix getconf symbolic constants defined in limits.h (BZ# 29147) (Adhemerval)
    • Review if the minimum impacts applications.
  • elf: Fix handling of symbol versions which hash to zero (bug 29190) (Florian Weimer)
    • Found all uppercase P's that hash to zero.
  • [1/1] x86_64: Add strstr function with 512-bit EVEX (Raghuveer Devulapalli)
    • Noah updated the benchmark for this.
    • Adhemerval: Can we make this more generic?
    • Noah: Wilco's version is simply better.
  • Remove _dl_skip_args (Adhemerval)
    • Needs review.
    • Some machime maintainers have reviewed.
    • i686,x86_64: Done, hppa: Reviewed, aarch64: Already done.
  • Stopped at 54394.
  • Discussed "linux: Add fallback for clone failure on posix_spawn (BZ #29115)"
    • The issues goes far beyond glibc.
    • If we make posix_spawn work then users can use this.
    • Adhemerval: Rewriting with shared mapping.
    • Florian: Always use a shared mapping the first time to detect vfork-as-fork.
      • Just a future direction for the code.
  • Reminder: GNU Tools Cauldron 2022

Meeting: 2022-05-16 @ 0900h EST5EDT

Video/Audio: https://bluejeans.com/9093064454

IRC: #glibc on OFTC.

  • Carlos O'Donell sends his regrets.

Review new patches and attendee requested reviews.

  • Status NEW delgate NOBODY 224 patches.
  • Start at patch 54271.
  • dlfcn: Move RTLD_DEFAULT/RTLD_NEXT outside USE_GNU

    • - Looks ok, Florian will review.
  • linux: use statx for fstat if neither newfstatat nor fstatat64 is present
    • - Looks ok, Adhemerval will review.
  • x86-64: Ignore r_addend for R_X86_64_GLOB_DAT/R_X86_64_JUMP_SLOT
    • - Fangrui state is is already done on other environments, Florian will review.
  • dlsym: Make RTLD_NEXT prefer default version definition [#BZ #14932]
    • - Current status is too broken to keep a versioned symbol, change looks reasonable.
  • [v11,1/6] elf: Refactor dl_new_hash so it can be tested / benchmarked
    • - Accepted by Siddhesh.
  • [1/5] locale: Turn ADDC and ADDS into functions in linereader.c
    • - Not urgent, skipped for now.
  • Avoid RMW of flags2 outside lock (BZ #27842)
    • - Still need some discussion that lock can be skipped if flockfile is called.
  • [v2] x86_64: Implement evex512 version of strlen, strnlen, wcslen and wcsnlen
    • - Noah is reviewing it.
  • [v3] Fix deadlock when pthread_atfork handler calls pthread_atfork or dlclose
    • - Accepted by Adhemerval.
  • Add arc4random support
    • - Still need reviewers.
  • [v3] Remove special flags of libc.5.so and libc.4.so
    • - Patch looks ok, some changers required.
  • socket: Check lengths before advancing pointer in CMSG_NXTHDR
    • - Siddhesh will review it.
  • [v4,3/3] linux: Add fallback for clone failure on posix_spawn (BZ #29115)
    • - Skipped, most likely will result in WONTFIX and programs will need to use
      • any new kernel time namespace fix kernel will provide.
  • GLIBC LoongArch PATCHES

    • - Patchset looks ok, still requires some upstream fixes. Skipped due it will
      • require at least a v5.
  • [v3] elf: Rewrite long RESOLVE_MAP macro to a debug friendly function
    • - Siddhesh is ok with the always_inline change.

  • [v3] math: Add math-use-builtins-fabs (BZ#29027)
    • - The patch is ok, it does not change any code.
  • stdio-common: Add the fgetln function
    • - Still need review.
  • elf: Simplify version test when searching a versioned symbol
    • - Florian will review.
  • Simplify ELF_RTYPE_CLASS_EXTERN_PROTECTED_DATA and revert aarch64/arm's extern protected data handle
    • - Florian will review.
  • Optimize struct pthread size
    • - Skipped, Arjun will review.
  • libio: Remove _IO_adjust_column, _IO_adjust_wcolumn, column tracking
    • - Adhemerval will review.
  • powerpc64: small-model @got@tprel accesses
  • [v3] powerpc64: Enable static-pie
    • - skipped, waiting review from powerpc maintainers.
  • Stopped at 52777.
  • Siddhesh
    • - Nothing to review.
  • Szabolcs
    • - Nothing
  • Sunil
  • Noah
    • - Can we change pthread cond to use default mutex and increment the counter
      • outside? Is this a GNU extension or is related to some different mutex type?
  • Patrick
    • - Nothing
  • Florian
    • - Nothing

Meeting: 2022-05-16 @ 0900h EST5EDT

Video/Audio: https://bluejeans.com/9093064454

IRC: #glibc on OFTC.

  • Status NEW delgate NOBODY 252 patches.
  • Start at patch 53982
  • [v2] Remove special flags of libc.5.so and libc.4.so
    • Adhemerval will review.
    • Florian: Cleanup is valuable.
  • 53980 Enable DT_RELR automatically
    • Carlos: Make first commit message line more descriptive
  • 53963 Check for ISO C compilers should also allow C++
    • Fangrui and Florian reviewed.
  • 53886 fortify: Ensure that glibc_fortify condition is a constant [BZ #29141]

    • Florian noted Siddhesh is the subject matter expert.
  • [v8,1/6] elf: Refactor dl_new_hash so it can be tested / benchmarked
    • Adhemerval: Is this something we can get the compiler to do?
    • Carlos: Yes, it would be better if the compiler did it, but today it doesn't.
    • Ahemerval: Not a blocking objection, but don't like the asm.
    • Siddhesh: Elevate the objection about the asm. Will comment on the list.
  • 53756 Linux: Fix posix_spawn when user with time namespaces
    • Florian: Ask kernel for new way of entering time namespace
    • The kernel can't change the current behaviour.
    • Many C and C++ libraries have interfaces implemented from scratch.
    • All the interfaces would have to be changed to support this.
    • Florian: Reached out to Christian Brauner about this on the musl thread.
    • Do we need to speak with Andrei Vagin (CRIU)?
  • Stopped at 53612.
  • Adhemerval: Looking for arc4random review.
    • arc4random at v5 review.
    • Carlos will review kernel syscall patches.
    • Florian has not reviewed this as author of the previous attempt.
    • Carlos: We are talking about commiting to the ABI? Any objections to the ABI?
    • Florian: We should add the ABI.
  • Arjun: Have two patches to reivew. New patch to post v3 for pthread_atfork.
  • Florian:
    • brk change:
      • Not sure where I'm heading with the brk change.
      • Adhemerval: OK with the brk changes.
    • vfprintf rework still needs review of the buffer management.
    • UTF-8 locale source work:
      • Siddhesh can change one.
      • Carlos can change Spanish locales (22 of them)
  • HJ: Nothing to review.
    • Long term goal to get rid of copy relocation in x86.
    • There are a couple of platform that support copy relocations.
    • Many don't support copy relocations.
    • From a performance and security POV we want to get rid of them.
    • Especially with protected symbols they are supposed to improve performance.
    • Semantics of copy relocations break this.
    • gcc 12 has option to avoid copy relocations on x86.
    • Explore removing copy relocations?
    • Florian: From Fedora and downstream perspective copy relocations don't matter.
      • We use LTO and the builds localize the symbols.
      • The change to copy reocations is a NOP.
      • Likewise for SUSE.
      • HJ: Note in Windows with declspec(dllimport/dllexport) we can improve code gen.

        • HJ: Anything we can do here that is similar?
    • Florian: COPY relocations may still be relevant on some architectures.
      • For example cost to access global data through the GOT
  • Szabolcs: Outstanding patch was reviewed. Need to change and push.

Meeting: 2022-05-09 @ 0900h EST5EDT

Video/Audio: https://bluejeans.com/9093064454

IRC: #glibc on OFTC.

  • 255 patches in NEW/NOBODY
  • Started at 53612.
  • GLIBC LoongArch PATCHES (caiyinyu)

    • LoongArch patch series for glibc.

    • Joseph Myers was previously reviewing this.
    • Several other developers reviewed parts of the series.
    • Review against existing v2
    • LoongArch is not yet upstream in the kernel.

    • Need to have the port make its way into the kernel via linux-next.
    • With 5.19 slated for August, we need to consider if RC phase constants
      • are good enough for a glibc port.
    • Could we merge and then backout of the port doesn't make 5.19?
  • 53573: [v3] elf: Rewrite long RESOLVE_MAP macro to a debug friendly function (guriev)
    • Fangrui was looking at earlier versions of the cleanup.
  • 53613: [v2] wcrtomb: Make behavior POSIX compliant (poyarekar)
    • Looking for review for v2 from earlier reviewers.
    • Adhemerval: Patch itself is OK, but need to check if the patch has consensus.
    • Carlos: Manual should say what you have *now* in the impl.
    • Siddhesh: Main concern is bumping symbol version because it avoids backports of
      • the fix.
  • 53543: [v5] nptl: Add backoff mechanism to spinlock loop (Wangyang)
    • HJ reviewing. Should be committed today.
  • 53526: [v3] math: Add math-use-builtins-fabs (BZ#29027)
    • Joseph pointed out issues.
    • Followed up with suggestions. Just a code cleanup.
    • Should not change code generation.
  • 53523: [PATCH v4 1/3] ia64: Always define IA64_USE_NEW_STUB as a flag macro
  • 53519: Args adjustment with ./ld.so exe [BZ #23293] (v6)
    • Need 1 and 4 patches reviewed.
    • Florian: The first patch needs reivew.
  • Stopped at 53519.
  • Florian: vfprintf not that high priority.
    • Someone should look at the buffer abstraction.
    • Series 7969.
    • Patch 17 implements the abstraction.
  • Arjun: pthread_atfork v2.
  • Adhemerval:
    • Fallback clone on posix_spawn failure.
      • Ask Christian Brauner about the vDSO issue.
      • There is conflict here with the semantics.

Meeting: 2022-05-02 @ 0900h EST5EDT

Video/Audio: https://bluejeans.com/9093064454

IRC: #glibc on OFTC.

  • 259 patches in NEW/NOBODY.
  • 53383 powerpc32: Remove unused HAVE_PPC_SECURE_PLT (Fangrui)
    • Reviewed by Florian.
  • 53379 elf: Remove fallback to the start of DT_STRTAB for dladdr (Fangrui)
    • Needs review.
  • 53375 elf: Simplify version test when searching a versioned symbol (Fangrui)
    • Needs review.
  • Simplify ELF_RTYPE_CLASS_EXTERN_PROTECTED_DATA and revert aarch64/arm's extern protected data handl…
    • Ask HJ if he can provide review (worked on this before).
  • elf: Remove one-default-version check when searching an unversioned symbol (Fangrui)
    • Needs review.
  • S390: Enable static PIE (Stefan Liebler)
    • Needs quite a bit of kernel support.
    • Florian and Adhemerval in agreement that we should handle failures gracefully.
    • brk failures default back to mmap.
    • Should we just use mmap unconditionally?
  • time: Use nonnull to avoid null pointer (Xiaoming)

    • Under review from Siddhesh.
    • Fails CI/CD for 32-bit build.
  • [v4,1/6] elf: Refactor dl_new_hash so it can be tested / benchmarked (Noah)
  • [v2] Fix deadlock when pthread_atfork handler calls pthread_atfork or dlclose (Arjun)
    • Asking Adhemerval for review since he worked on the first patch.
  • Optimize struct pthread size (Adhemerval)
    • Two patches. Carlos to review.
  • Add arc4random support (Adhemerval)
    • Currently at v4.
    • Siddhesh raised questions about FIPS compliance.
      • Crypto stuff we add does not get exposed publicly as API.
      • No ChaCha20 API exposed then we should be OK.

      • Document in manual that arc4random is not intended for use in secure contexts.
      • Say "Not intended for cryptographic or security purposes."
    • v5 with manual change.
  • 53091 libio: Remove _IO_adjust_column, _IO_adjust_wcolumn, column tracking
    • Review after "vfprintf rework to remove vtables"
  • Stopped at 53091.
  • Sunil: Raised issues about backporting to release/2.32/master.
    • Problems caused by the AMD patch backport.
    • Florian: Could revert the AMD backport and then layer everything properly.
    • Consensus to revert the AMD patch, and then layer the commits correctly.
    • Florian: Can push revert to a branch for 2.32.
  • Florian: Would like to backport the dlinfo fixes.
    • Carlos: Only requires the addition of a header constant. OK with me.
  • Remove kernel version and distinguish library check (Adhemerval)
    • Florian to review later this week.
  • linux: Use /sys/devices/system/cpu on get_nprocs_conf (BZ#28991)

    • Needs a follow-on fix because of rseq.
    • Carlos: Is malloc affected?
    • Adhemerval: No, only the exported symbol is impacted.

Meeting: 2022-04-25 @ 0900h EST5EDT

Video/Audio: https://bluejeans.com/9093064454

IRC: #glibc on OFTC.

Review new patches and restart review at the top.

  • NEW/NOBODY is at 324 patches.
  • Starting at 53184
  • v4 Add arc4random support (Adhemerval) 53184
  • 53169 elf: Replace PI_STATIC_AND_HIDDEN with opposite HIDDEN_VAR_NEEDS_DYNAMIC_RELOC
    • Adhemerval to look at this.
  • 53144 elf: Move post-relocation code of _dl_start into _dl_start_final
    • Adhemerval to look at this.
    • glibc lacks some clear distinctions around the loader startup.
    • Specific barriers are needed here.
  • 53127 Support DT_RELR relative relocation format (HJ)
    • Pretty much done.
    • Try to remove configure knob on this - Turn on DT_RELR by default.
  • 53117 INSTALL: Rephrase -with-default-link documentation
    • Carlos to respond.
  • 53094 Fix deadlock when pthread_atfork handler calls pthread_atfork or dlclose
    • Failed CI: Bad test.
    • Arjun: Reviewing Adhemerval's patches for this.
    • Arjun: What is the consensus for prepare/after handlers added during the fork
      • in a handler? Should they just become something to use later?
    • Adhemerval: When you fork, you use the list as-is.
    • Arjun to post new version.
    • Arjun to add manual to document pthread_atfork
    • Arjun to investigate what to say to the austingroup.
      • "If any handler causes a handler to be registered to deregistered then the behaviour of the handler is implementation defined"
  • Stopped at 53091.
  • Args adjustment with ./ld.so exe [BZ #23293] (szabolcs)
  • Sunil: Backporting code into the release branches.
  • [v2,1/6] elf: Refactor dl_new_hash so it can be tested / benchmarked (Noah)
    • Asking for Florian to review.
  • [v5] elf: Fix DFS sorting algorithm for LD_TRACE_LOADED_OBJECTS with missing libraries (BZ #28868)
    • Siddhesh to review patch 51584.

Meeting: 2022-04-18 @ 0900h EST5EDT

Video/Audio: https://bluejeans.com/9093064454

IRC: #glibc on OFTC.

No meeting (Easter Monday).

Meeting: 2022-04-11 @ 0900h EST5EDT

Video/Audio: https://bluejeans.com/9093064454

IRC: #glibc on OFTC.

Review new patches and restart review at the top.

  • NEW/NOBODY 266 patches.
  • Starting at patch 52775.
  • ELF tests in Python and libio vtables/RELRO interaction (fweimer)
    • Needs review.
  • [v4,1/6] x86: Remove {w}memcmp-ssse3 (noah)
    • Complaints about a specific machine was handled with reporter.
    • HJ to review.
  • nptl: Handle spurious EINTR when thread cancellation is disabled (BZ#29029) (fweimer)
    • Debian has tested (Aurelien Jarno)
    • Florian wants to review this one given his state on the changes.
  • [v2] math: Add math-use-builtins-fabs (BZ#29027) (adhemerval)
    • Needs review.
  • manual: Avoid name collision in libm ULP table [BZ #28956] (Coldrick)
    • Carlos to push.
  • [v1] sysdeps: Add 'get_fast_jitter' interace in fast-jitter.h (noah)
    • Jitter interface is for thread backoff.
    • This patch is for use by the in-progress changes to futex backoff.
  • [v6] Add .clang-format style file (noah)
    • Looking for 1 review.
    • Carlos to review.
  • [v3] elf: fixes compile error when both enable -Werror and -DNDEBUG (yang)
    • This needs the usual compiler pragmas to remove the unused warnings.
    • Carlos to respond.
  • 52556-52525 posted fixes for fuzzing by Dmitry Chestnyh (omp.ru)
    • Discussing options for reviewing fuzz patches.
  • 52304 - Superseded
  • linux: Use /sys/devices/system/cpu on get_nprocs_conf (BZ#28991)

    • Still up for review (week 2).
  • Refactor syslog implementation (adhemerval)
    • Just a refactor without timestamp change.
    • Improves code and testing.
  • Stopped at 52087: vfprintf rework to remove vtables (fweimer).
  • Florian: Discussed PT_GNU_RELRO and potential binutils issues with PT_LOAD alignment.
    • Final PT_LOAD doesn't start page aligned. Should it?
  • Noah: Discussed improvements to symbol lookup.
    • Noted dl lookup function and strcmp on the hot path.
    • Florian: Pointed out we can precompute some of the math operations per linkmap.
      • Some of the math on these lookups can be stored.
    • Noah: Can we know that we aren't on a page boundary and use a distinct strcmp?
    • Carlos: You know DT_STRTAB and STRSZ and can do some figuring out.

Meeting: 2022-04-04 @ 0900h EST5EDT

Video/Audio: https://bluejeans.com/9093064454

IRC: #glibc on OFTC.

Review new patches and restart review at the top.

  • NEW/Nobody 260 patches.
  • Starting review 52598
  • [v3] elf: fixes compile error when both enable -Werror and -DNDEBUG
    • Needs different structure.
  • Dmitry Chestnyh - Ran the static analyzer.
    • Need to follow up regarding assignment vs. DSO.
    • SVACE static analyzer results.
    • 12 patches need review.
  • Discussed briefly that string tests are taking a long time.
    • What can we do to improve developer experience?
    • Note: Split the tests at the make target level with different args.
    • Note: Split the tests at the do_test level with different threads.
    • Noted that any split up should help but not solve.
  • [v5] Add .clang-format style file
    • Carlos to review v5.
  • Remove -z combreloc and HAVE_Z_COMBRELOC
    • Different from Adhemerval's branch to enable clang.
    • Needs review.
  • elf: Simplify _dl_aux_init with inhibit_loop_to_libcall
    • Reviewed by Szabolcs. Needs fixes.
    • Adhemerval noted this can be fixed in a different way.
  • 52433: [v6] Allow for unpriviledged nested containers (DJ)
    • Carlos to review.
  • 52427: [v1,1/2] random-bits: Factor out entropy generating function
    • Needs review.
    • What is the status of Florian's arc4random?
  • 52428: [v1,2/2] x86: Use rdtsc for generating entropy for random_bits
    • Needs review.
  • Stopped at 52428.
  • Florian has a vfprintf patch series reviewed.
  • Adhemerval wants LD_AUDIT trace fixed.
    • Florian really wants to review.
  • linux: Use /sys/devices/system/cpu on get_nprocs_conf (BZ#28991)

    • Needs review.
  • [v2] linux: Fix closefrom_fallback iterates until max int (BZ#28993)

    • How do we proceed? -- Should be OK to proceed to fail fast.
    • Will post v3.
    • What about ENOMEM?

Meeting: 2022-03-28 @ 0900h EST5EDT

Video/Audio: https://bluejeans.com/9093064454

IRC: #glibc on OFTC.

Review new patches and restart review at the top.

  • State NEW, delegate NOBODY is 312 patches.
  • Adhemerval tries to push.

  "Enumerating objects: 1581, done.
Counting objects: 100% (1581/1581), done.
Delta compression using up to 24 threads
Compressing objects: 100% (807/807), done.
Writing objects: 100% (1093/1093), 162.32 KiB | 13.53 MiB/s, done.
Total 1093 (delta 950), reused 298 (delta 272), pack-reused 0
remote: Resolving deltas: 100% (950/950), completed with 415 local objects.
remote: *** This update introduces too many new commits (121), which would
remote: *** trigger as many emails, exceeding the current limit (100).
remote: *** Contact your repository adminstrator if you really meant
remote: *** to generate this many commit emails.
remote: error: hook declined to update refs/heads/azanella/clang"
  • Carlos reviewed and it's max-commit-emails, which can be set to 1000 to avoid
    • this problem and allow you to publish a branch for review.
  • [v2] nptl: Add backoff mechanism to spinlock loop (Wangyang Guo)
    • We can't use random-bits.h (30 cycles)
    • Can we use tid?
    • Can we use rdtsc? (18 cycles)
    • At least one microbench should be contributed.
  • [v1,1/2] x86: Small improvements for wcscpy-ssse3 (Noah)
  • [v2,1/6] x86: Remove {w}memcmp-ssse3 (Noah)
    • HJ reviewing.
  • rwlock for async application
    • This is an RFC for glibc to deviate from POSIX behaviour.
  • Add HWCAP2_AFP, HWCAP2_RPRES from Linux 5.17 to AArch64 bits/hwcap.h (Joseph)
    • Accepted by Szabolcs [Arm]
  • Note from Adhemvarl that POSIX will add strlcpy and strlcat
  • Stopped at 52304.
  • realpath: Bring back GNU extension on ENOENT and EACCES [BZ #28996]
    • Adhemerval will review.
  • [v5] Allow for unpriviledged nested containers
  • Siddhesh notes that DNS tests no longer depend on proc mount.
    • However we still have tests that use DNS and can fail.
    • They need to be ported to use the in-process test DNS resolver or changed.
  • Support DT_RELR relative relocation format review.
    • What is the status of DT_RELR for AArch64? Currently unimplemented.
    • Looking for review of v6 patches.
    • No technical blockers. Push April 11th if no further reviews.

Meeting: 2022-03-21 @ 0900h EST5EDT

Video/Audio: https://bluejeans.com/9093064454

IRC: #glibc on OFTC.

Review new patches and restart review at the top.

  • Carlos unable to start meeting until 930am due to technical issue.
  • Discussed with Adhemerval patches with outstanding review.

Meeting: 2022-03-14 @ 0900h EST5EDT

Video/Audio: https://bluejeans.com/9093064454

IRC: #glibc on OFTC.

Review new patches and restart review at the top.

  • 255 patches with NEW/NOBODY.
  • getaddrinfo facelift and fixes (Sidhesh)
    • Some feedback from Carlos regarding DNS access.
    • Feedback from Andreas about example.org.
  • [v5] Allow for unpriviledged nested containers (DJ)
    • Florian reviewing.
  • [v6] Support DT_RELR relative relocation format (HJ)
    • Added for x86_64, ppc64le looking at this. What about others? (Adhemerval)
    • What about aarch64 and s390x? Other arches like RSIC-V?
    • Confirmed for 2.38 we have x86_64, ppc64, ppc64le.
    • Would be nice to have aarch64.
  • scripts/dso-ordering-test.py: Fix C&P error in * callrefs processing

    • Adhemerval to review.
  • [v5] elf: Fix DFS sorting algorithm for LD_TRACE_LOADED_OBJECTS with missing libraries (BZ #28868)
    • Review from Florian?
  • Remove kernel version and distinguish library check (Adhemerval)
    • Waiting for Florian to finish review.
    • Discussed that 4.9 is the oldest LTS kernel we could support.
    • No large benefit to move to 4.9 and it would cause downstream problems.
    • Relaxing the version check is a step forward.
  • PowerPC64 static-pie (Alan Modra)
    • Tulio (IBM) to review.
  • : C++20 P0482R6 and C2X N2653: support for char8_t, mbrtoc8(), and c8rtomb().

    • No review.
  • [v4] localedata: Updates for Taiwanese locales [BZ #24409]
    • Carlos asking for split up and review.
  • Use RFC5424 for syslog (Adhemerval)
    • Adhemerval: Need to review upstream systemd thread.
  • linux: Add new syscalls
    • Added 12 syscalls.
    • Carlos: Do we need man pages for these?
    • Adhemerval: I think all of them have man pages.
    • man-pages currently slow to update.
      • Waiting for Alejandro to have access to tree?
  • Predictable ELF destructor ordering
    • Florian to repost.

Meeting: 2022-03-07 @ 0900h EST5EDT

Video/Audio: https://bluejeans.com/9093064454

IRC: #glibc on OFTC.

Review new patches and restart review at the top.

  • 260 patches in NEW/NOBODY
  • [v5] elf: Fix DFS sorting algorithm for LD_TRACE_LOADED_OBJECTS with missing libraries (BZ #28868)
    • Looking for review.
    • Florian will to review.
    • Backported just the fix downstream with no issues to report.
  • Remove kernel version and distinguish library check
    • Adressed Florian's review.
  • malloc: Exit early on test failure in tst-realloc
    • Need review.
  • v5: Support DT_RELR relative relocation format (HJ)
    • No reviewer right now.
  • v2: PowerPC64 static-pie
    • Tulio says IBM will review.
  • C++20 P0482R6 and C2X N2653: support for char8_t, mbrtoc8(), and c8rtomb().

    • Still in the same position as last week.
  • [v3] libio: Ensure output buffer for wchars (bug 28828)
    • Adhemerval reviewed.
    • Asked vor v4 with fixed tests.
    • Changes requested.
  • Use RFC5424 for syslog (Adhemerval)
  • linux: Add new syscalls (Adhemerval)
    • No reviewer yet.
  • [v2] x86: Reformat code in svml_s_atanhf16_core_avx512.S
    • Noah submitted. Working with Sunil on this issue.
  • v2: Predictable ELF destructor ordering (Florian)
    • Still looking for review.
    • Need to repost v3 after testing.
  • [v2,1/3] x86: Optimize svml_s_tanhf_core_avx512.S (Noah)
    • Working with Sunil to reformat code before optimizing.
  • Stopped at 50642.
  • New DSO dependency resolver buggyness with dlclose()
  • Fixes for getnameinfo() with NI_NOFQDN (race condition)
  • Adhemeral discussed symbol redirection and usage in gcc and clang
    • Will need some changes to glibc's headers if we support clang as a compiler.
    • Need to discuss this on list with a plan.
    • Florian notes that this is a large project.

Meeting: 2022-02-28 @ 0900h EST5EDT

Video/Audio: https://bluejeans.com/9093064454

IRC: #glibc on OFTC.

Review new patches and restart review at the top.

  • NEW/NOBODY 262.
  • PowerPC64 static-pie [v2 1/4-4/4]
    • Ask Tulio if he will review.
  • : C++20 P0482R6 and C2X N2653: support for char8_t, mbrtoc8(), and c8rtomb().

    • Accepted by WG14. Needs review. Repost.
    • Needs a reviewer to look at this against the N2653 implementation requirements.
  • [v4] localedata: Updates for Taiwanese locales [BZ #24409]
    • Carlos to look at this.
  • benchtests: make compare_strings.py accept string as attribute value
    • Patch failed to apply (CI/CD caught this). Siddhesh will cleanup
  • [v3] libio: Ensure output buffer for wchars (bug 28828)
    • Adhemerval reviewed the changes.
    • Adhemerval to respond again regarding test case.
  • [v3] elf: Fix DFS sorting algorithm for LD_TRACE_LOADED_OBJECTS with missing libraries (BZ #28868)
    • Adhemerval need to rewrite the test case using Florian's suggestions.
  • Remove distinguish library code (Adhemerval)
    • Drop kernel version check and drop arch review.
    • Florian to review.
  • [v3] x86: Test wcscmp RTM in the wcsncmp overflow case [BZ #28896]
    • HJ reviewing.
  • Use RFC5424 for syslog
  • Support DT_RELR relative relocation format
    • At v3. Latest version of binutils had DT_RELR support.
    • Note that Fedora Rawhide has binutils 2.38 and could be used to test DT_RELR.

Meeting: 2022-02-21 @ 0900h EST5EDT

Video/Audio: https://bluejeans.com/9093064454

IRC: #glibc on OFTC.

Review new patches and restart review at the top.

  • Discussed patches and how to make progress on them.

Meeting: 2022-02-14 @ 0900h EST5EDT

Video/Audio: https://bluejeans.com/9093064454

IRC: #glibc on OFTC.

Review new patches and restart review at the top.

  • 284 patches in NEW/NOBODY.
  • linux: Use socket-constants-time64.h on tst-socket-timestamp-compat
    • Carlos to review.
  • x86-64: Define memcmpeq in ld.so

    • Noah reviewing.
  • libio: Ensure output buffer for wchars (bug 28828)
    • Adhemerval reply to resend the patch.
    • Patrick: Was a whitespace issue. Lines that are not in the patch
      • all have extra whitespace.
  • resolv: Fix tst-resolv tests for 2.35 ABIs and later
    • Adhemerval said he'd look atthis.
  • rtld: Add new hook to initialize dp register on hppa
    • Reviews from Florian and HJ. Dave needs a v2.
  • inptl: Use 64 bit time_t stat internally for sem_open (BZ #28880)
    • Notes that running on a system with future >64-bit time fails a lot of tests.

    • Need to go fixing these issues.
  • Remove bcopy and bzero optimizations
    • HJ notes that rather than remove bzero we could rename to memset_zero

    • Note that memset_zero is not equivalent to bzero because of the return type.

    • Removing bcopy and bzero and replacing with memset_zero is arch specific.

    • Each maintainer needs to consider the memset_zero optimziation.

    • Carlos: Can we notify the arch maintainers before removal?
  • Stopped at 51018.
  • Support DT_RELR relative relocation format
    • HJ: Answered all of Joseph's comments.
    • Joseph was original reviewer.
    • Carlos to look at this on Friday.
  • Florian: Patch for aux vector changes.
    • Consensus to push but needs future work.

Meeting: 2022-02-07 @ 0900h EST5EDT

Video/Audio: https://bluejeans.com/9093064454

IRC: #glibc on OFTC.

Review new patches and restart review at the top.

  • NEW / NOBODY patches 355.
  • [v2] linux: fix accuracy of get_nprocs and get_nprocs_conf [BZ #28865]
    • Adhemerval to look at.
  • [v2] hppa: Fix bind-now audit (BZ #28857)
    • Adhemerval to send v3.
    • Carlos to review as hppa maintainer.
  • [v1] x86: Remove SSE3 instruction for broadcast in memset.S (SSE2 Only)
    • HJ provided feedback.
  • elf: Replace memcmp with memcmpeq for variable size

    • Using memcmp_eq limits the compiler optimizations for size.
      • There is a tradeoff for size vs. speed.
  • x86-64: Make bcmp an alias of memcmpeq

    • HJ will drop this.
    • Carlos will make a comment about bcmp.
  • [v1] x86: Improve vec generation in memset-vec-unaligned-erms.S
    • Already committed.
  • Support DT_RELR relative relocation format [v2]
    • Second version of DT_RELR.
  • Improve LC_MONETARY handling.
    • Ask DJ for review.
  • [v2] x86: Reformat code in svml_s_atanhf16_core_avx512.S
    • Sunil will review svml cleanups for reformatting.
  • Remove prelink support [v2]
    • Siddhesh has started to review.
  • linux: Add new syscalls
    • Carlos to review new syscalls.
  • HJ: During the 2.35 release cycle we removed a p_align check.
    • elf: Check invalid hole in PT_LOAD segments [BZ #28838]
      • Needs review.
  • Patrick: [v2] Remove upper limit on tunable MALLOC_MMAP_THRESHOLD
    • Cleaned up patch tracker state.
  • Siddhesh ran cleanup for patches.
    • Cleaned up to 264.
  • getaddrinfo: Fix leak with AI_ALL [BZ #28852]
    • May need a v2. Florian reviwed and asked for feedback.
  • [v2] setlocale: Fail if iconv module for charset is not present [BZ #27996]
    • Siddhes to post v3.

Meeting: 2022-01-31 @ 0900h EST5EDT

Video/Audio: https://bluejeans.com/9093064454

IRC: #glibc on OFTC.

With the release upcoming we've switched to reviewing blockers and desirable changes.

Meeting: 2022-01-24 @ 0900h EST5EDT

Video/Audio: https://bluejeans.com/9093064454

IRC: #glibc on OFTC.

With the release upcoming we've switched to reviewing blockers and desirable changes.

  • Review the release blockers for glibc 2.35.
  • PT_LOAD fix status?
    • Committed fix e22a4557eb39d7cba9a74d70f4582c13f1a7a83a
    • Adhemerval reviewing test case inclusion.
  • Fix glibc 2.34 ABI omssion.
  • linux: Fix ancillary 64-bit time timestamp conversion (BZ #28349, BZ #28350)
    • Florian sent review.
    • Adhemerval reposted patches with fixes.
  • posix: Add terminal control setting support for posix_spawn
    • Carlos to post review ASAP.
  • Multiple RTLD audit fixes.
    • Adhemerval sent updated versions.
    • Carlos to look at the fixes.
  • Use after free warning failures in gcc 12.
    • Carlos to follow up with Martin Sebor.
  • IBM long-double issues are release blockers we aren't tracking yet.
    • Blocks Fedora Rawhide builds.
    • We need to disable the flag in more places.
    • Florian to post the patches for this today.
  • Szabolcs: p_align issue.
    • Binutils switching on x86 to set p_align of load segments.
    • On AArch64 it would be nice too.
    • Old glibc would reject such a binary on a 64k page system.
    • Seems to me that I cannot switch this if there are existing 64k page sized systems.
    • Florian suggests fixing the bug in glibc and backport to old releases.
    • Carlos agrees with backport.
  • Adhemerval sent prelink removal patch for glibc 2.36.
    • Carlos: Post a NEWS update with a note saying 2.35 is the last relase with prelink.

Meeting: 2022-01-17 @ 0900h EST5EDT

Video/Audio: https://bluejeans.com/9093064454

IRC: #glibc on OFTC.

With the release upcoming we've switched to reviewing blockers and desirable changes.

  • NEW/NOBODY 290 patches
  • [v7] elf: Properly align all PT_LOAD segments [BZ #28676]
    • Adhemerval to review alignment patch.
  • linux: Fix ancillary 64-bit time timestamp conversion (BZ #28349, BZ #28350)
    • Florian to review.
  • posix: Add terminal control setting support for posix_spawn
    • Carlos to finish review Monday.
  • Multiple rtld-audit fixes
    • Adhemerval to rebase patch set.
    • Szabolcs looked at the aarch64 pieces and gave feedback.
  • Properly handle --disable-default-pie
    • Siddhesh is going to look at this.
  • Carlos and Adhemerval discussed 64-bit time_t testing.
    • No way to change library default time_t size, just testing can test it.
  • Prelink discussion
    • Adhemerval: May I submit patches to remove prelink?
    • Carlos: Yes.
    • Discussed Intel, Arm, Linaro involvement in prelink.
    • Carlos: IHVs and ISVs need to invest upstream if they want to see prelink maintained.
  • Paul sends his regrets and includes two items:
    • I have done a review of the accuracy of the libmvec functions (in master)
      • both with sse4.2, avx2 and avx512. Apart from a minor issue with atan2 in binary64 which was fixed (#28765), I found no error larger than 4 ulps, both in binary32 and binary64.
    • We are making good progress in the CORE-MATH project: 17 out of 26
      • functions are now available in single precision (binary32).

        See https://homepages.loria.fr/PZimmermann/CORE-MATH/. In most cases the correctly-rounded CORE-MATH code outperforms the GNU libc one (see the "perf" graphs). The only exceptions so far are the cosf function on i7, and the expf function.

Meeting: 2022-01-10 @ 0900h EST5EDT

Video/Audio: https://bluejeans.com/9093064454

IRC: #glibc on OFTC.

With the release upcoming we've switched to reviewing blockers and desirable changes.

Meeting: 2022-01-03 @ 0900h EST5EDT

Video/Audio: https://bluejeans.com/9093064454

IRC: #glibc on OFTC.

Review new patches and restart review at the top.

Meeting: 2021-12-27 @ 0900h EST5EDT

Video/Audio: https://bluejeans.com/9093064454

IRC: #glibc on OFTC.

Review new patches and restart review at the top.

Meeting: 2021-12-20 @ 0900h EST5EDT

Video/Audio: https://bluejeans.com/9093064454

IRC: #glibc on OFTC.

Review new patches and restart review at the top.

Meeting: 2021-12-13 @ 0900h EST5EDT

Video/Audio: https://bluejeans.com/9093064454

IRC: #glibc on OFTC.

Review new patches and restart review at the top.

  • Carlos asks what are we looking at putting into the release.
  • Adhemerval:
    • Would like to get hugepage support into 2.35.
      • Siddhesh: Reviewed the last patch set.
      • Huge page support aligns with what we are doing with the text segment
        • Makes it easier to support THP with .text.
    • Carlos: What is the status of the LD_AUDIT series?
      • Adhemerval: Florian has reviewed half of the patches.
      • Florian: Ran out of time last week. I should be able to review this week.
  • Florian:
    • _dl_find_object needs review.
      • Adhemerval: Will review.
    • Minor fix to rseq. No ABI implications.
      • Removes not-needed relocations.
    • Add membarrier support (syscall wrapper)
      • Need to do some digging for programming model.
      • May be an additional step required for membarrier setup.
  • HJ:
    • libmvec patches are very large (42 patches)
      • Submit one function. Work with community to address concerns.
        • If the community agrees with the approach then the rest of the patches can be automated and ready.
      • Intent is to include the functions in 2.35.
    • DT_RUNPATH v3 posted.
      • Carlos to review v3.
  • Sunil:
    • libmvec needs review.
      • CFI escape? Compiler generated. Encodes the DFA expression.
        • Can we remove it? What are they?
  • Siddhesh:
    • No additional reviews required.
  • Szabolcs:
    • No additional patches for review.
  • Carlos:
    • pthread cond var fix and review. Bug fix.
    • Licked the cookie. MUST BE FIXED.
  • Series 5728: Fix 64-bit time_t ancillary socket conversion
    • Posted v3. Needs review from Florian.
  • Series 5725: [v4] elf: Add _dl_find_object function
    • Adhemerval to review.
  • Series 5714: fix p_align on PT_LOAD segment in DSO isn't honored
    • Checked in.
  • Series 2348: [v2] Use LFS and 64 bit time for installed programs (BZ #15333)
    • Needed.
  • Discussed alignment issues with some architectures from the PT_LOAD changes.
    • Still need this if p_align is large.
    • We can solve this in the loader.
    • Once we fix hte linker the normal binary will have 4KiB p_align.
    • Will only get the large alignment if we have a variable that needs it.
    • For huge page we still want the large p_align even if there is no variables.
    • TODO: Must fix unmapping for 2.35.
  • Series 5703: intl/plural.y: define macros to avoid multiple conflicting declarations of yyerror and yylex
    • Ask Arjun to review this change.
  • Series 5688: [v3] Support target specific ALIGN for variable alignment test [BZ #28676]
    • Needs review.
  • Series 5695: [v7] elf: Don't execute shared object directly [BZ #28453]
    • Needs review.
  • What is the state of the OpenRISC or1k port?
    • Joseph has reviewed some of the state.
    • Is this ready for glibc 2.35?
    • State of the port is looking good.
    • Need to check build-many-glibcs and with Jospeh for the state of the port.
    • Doesn't look like there are any large blockers for 2.35.
  • New ld sort algorithm is failing on OpenRISC.
    • Needs review.
  • New ld sort default should be flipped to new algorithm.
    • Need to flip the new sort.

Meeting: 2021-12-06 @ 0900h EST5EDT

Video/Audio: https://bluejeans.com/9093064454

IRC: #glibc on OFTC.

Review new patches and restart review at the top.

Meeting: 2021-11-29 @ 0900h EST5EDT

Video/Audio: https://bluejeans.com/9093064454

IRC: #glibc on OFTC.

Review new patches and restart review at the top.

Meeting: 2021-11-22 @ 0900h EST5EDT

Video/Audio: https://bluejeans.com/9093064454

IRC: #glibc on OFTC.

Review new patches and restart review at the top.

  • 306 patches in NEW/NOBODY
  • linux: Add generic ioctl implementation
    • Needed for ork1 port.
  • linux: Add generic syscall implementation
    • Needed for ork1 port.
  • Add GLIBC_PRIVATE_RELR as proxy for DT_RELR
    • What about toolchains that don't use symbol versions?
    • Why does *_RELR exist in the name?
    • Is Google going to rebase 2.27 to a more current version?
    • Ask Chrome OS developers?
    • Ask Google server branch developers?
    • What will be the generic ABI mechanism for feature lockout?
      • Does this impact the decisions we make?
      • The BSDs don't view it as an issue overall
        • Except that if we decide on this markup they may need to change things.
    • Context: Just glibc. Parallel feature lockout via symbol@version.
      • Add linker support for empty node.
      • Use dummy symbol in glibc.
    • Add concrete semantics to that version.

Meeting: 2021-11-15 @ 0900h EST5EDT

Cancelled.

Meeting: 2021-11-08 @ 0900h EST5EDT

Video/Audio: https://bluejeans.com/9093064454

IRC: #glibc on OFTC.

Review new patches and restart review at the top.

  • State NEW, Delegate Nobody, 289 patches.
  • 47212: io: Refactor close_range and closefrom
    • Adhemerval: Pointed out a discrepancy for patch review and refactoring.
  • 47211: powerpc: Define USE_PPC64_NOTOC iff compiler and linker also supports it
    • Adhemerval: These are further fixes for lld. Discussing with IBM.
  • Discussed the current gABI issues that relate to ensuring that old ld.so
    • can safely fail when new binaries are run on old system.
    • Waiting for Cary Coutant to publish the gABI repo.
  • Series 4535: Update Makefile fragments for DSO sorting tests
    • Florain: Needs changes. Post v2.
  • 47161: Remove the unnecessary +make-deps
    • Needs review.
  • 47065: powerpc: Use FLAG_ELF_LIBC6 for 32-bit known libraries
    • Needs review from Florian or Tulio.
  • 47064: Fix memmove call in vfprintf-internal.c:group_number
    • Florian: Given that I produced the bug I should review it.
  • 47061: [v3] x86: Optimize atomic_compare_and_exchange_[val|bool]_acq [BZ #28537]
    • Needs review.
    • Carlos: As x86 maintainer you could just wait and push. There is a lot of internal CPU details that this depends upon.
    • Carlos: We need a comment here to expalin the redundant read.
  • 47048: [v2] elf: Use the minimal malloc on tunables_strdup
    • Siddhesh should review.
  • https://sourceware.org/pipermail/libc-alpha/2021-November/132688.html

    • Florian needs review.
  • Adhemerval: 64-bit time_t recvmsg review.
    • Florian: Assign that to me for review.
  • Florian: May you please repost your LD_AUDIT patches? (to Adhemerval)

Meeting: 2021-11-01 @ 0900h EST5EDT

Video/Audio: https://bluejeans.com/9093064454

IRC: #glibc on OFTC.

Review new patches and restart review at the top.

  • State NEW, Delegate NOBODY, 317 patches.
  • Series 4330: Enable EVEX strcmp
    • Carlos noted that with Noah and HJ working together this is enough review
      • along with the microbenchmark to commit this code.
  • Series 4329: Fix lld build for armhf
    • HJ: Binary with ARM_EXIDX inbetween to relocation sections looks odd.
    • We fixed a similar issue with fa19d5c48a6b36518ce68412e3bdde6bfa8aa4a6.
    • Rewrite as cleanup and provide timing cycle information.
  • Series 4318: [v1,1/5] string: Make tests birdirectional test-memcpy.c
    • HJ to review.
  • Series 4294: benchtests: Add atan2f, expm1f, log1pf and tanf to bench-math (4 patches)
  • Series 4292: benchtests: Add float version of math functions to bench-math (12 patches)
    • Convert double input to float input.
  • Series 4208: regex: Unnest nested functions in regcomp.c
    • Carlos to look at it.
  • 46656: [PiNg,9] localedata: add new locale ab_GE

  • 46551: powerpc: Use the correct flag for 32-bit known libraries
    • Needs a new commit message.
  • 46481: manual: remove an obsolete requirement on aligned_alloc() usage
    • Status? Suggest version the interface, return error without alignment.
  • Enable hard-coded path in tests.
    • 46294: elf: Also try DT_RUNPATH for internal dlopen [BZ #28455]
    • 46295: Add and use link-test-modules-rpath-link [BZ #28455]
    • Carlos to review.
  • Series 2732: malloc: Improve Huge Page support
  • Maxim Kuvyrkov attended. Hello Maxim!
    • Linaro looking at benchmarking toolchains.
    • Patrick: libmicro has many mirobenchmarks.
    • Stream benchmarks were showing regression for the L3 memcpy issue we fixed before.
    • It is very important that you must find the patch that makes the change.
    • We are doing post-commit testing.
    • Pre-commit testing is the point where it is easiest to prevent a patch going.

Meeting: 2021-10-25 @ 0900h EST5EDT

Video/Audio: https://bluejeans.com/9093064454

IRC: #glibc on OFTC.

Review new patches and restart review at the top.

  • State NEW, Delegate NOBODY, 304 patches.
  • Series 4140, Improve hypot(), Adhemerval to rebase.
    • Needs rebase after HJs benchtests/Makefile change.
  • Series 4116, [v14,1/4] String: Add support for memcmpeq() ABI on all targets

    • Looks OK.
  • Series 4105, powerpc: Use the correct flag for 32-bit known libraries
    • This looks like old a.out code is blocking this.
    • Carlos & Florian to bring up with IBM.

  • Series 4101, Add missing math functions to benchtests (Sunil Pandey)
    • Adding functions to libmvec
    • Adding benchmark data for existing libmvec functions.
    • For each function we have both double and float.
    • In order to generate the microbenchmarks for float we need float and double.
    • Sunil copied the input from double to float.
  • Series 4056, manual: remove an obsolete requirement on aligned_alloc() usage
  • Series 3961, [PING,8] localedata: add new locale ab_GE
  • Series 3949, regex: fix buffer read overrun in search [BZ#28470]
    • Stuck on review. Needs v2. Marked Changes Requested given Andreas' comments.
  • Series 3941, Benchtests: Improve bench-memcpy-random
    • Siddhesh to look at improvement and reveiw.
  • Series 3895, [v2] elf: Support DT_RELR relative relocation format [BZ #27924]
    • Unknown dynamic tags are ignored by old glibc.
      • Old glibc must fail when running the new binary.
      • Need to add symbol dependency or bump ELF OSABI.
    • lld is there for review, but additional patches are needed to build glibc for tests.
      • More work is required in the tree to get lld working.
      • If the testsuite fails to build we can't talk about regressions easily.
      • What is the status of lld building with glibc?
        • Carlos: Ask on thread about status of lld to build glibc.
  • HJ: Testing gcc with glibc 2.34 issues.
  • Adhemerval: 64-bit time_t receive message fix.

Meeting: 2021-10-18 @ 0900h EST5EDT

Video/Audio: https://bluejeans.com/9093064454

IRC: #glibc on OFTC.

Review new patches and restart review at the top.

  • 339 NEW/NOBODY patches in queue.
  • [v2] elf: Support DT_RELR relative relocation format [BZ #27924]
    • Need official gABI assignment (Carry Coutant status?)
      • Needs to be reserved verbally by Carry (Action item)
      • Required for patch acceptance.
    • Need BFD port or a released llvm/lld that supports DT_RELR to review patches.
      • Minimally a 32-bit and 64-bit port that can build glibc and passes regression testing.
      • Required for patch review.
    • Official glibc relese with feature?
      • Need BFD port to support build-many-glibcs to provide testing confidence.
    • HJ: Would like to see a BFD port before a glibc release with the features.
      • Sufficient to have a released binutils with DT_RELR also.
    • Release date of released static linkers with support should be close to glibc release with feature.
      • When glibc is released users can pick a static linker to test with (binutils or lld).
  • 317 NEW/NOBODY patches in queue.
  • Add and use link-test-modules-rpath-link [BZ #28455]
  • elf: Also try DT_RUNPATH for internal dlopen [BZ #28455]
    • Needed for change in linker default.
  • elf: Make global.out depend on reldepmod4.so [BZ #28457]
    • Existing defect.
  • 46290: elf: Fix elf_get_dynamic_info() for bootstrap
    • Needs review.
  • 45506: [v2] ld.so: Initialize bootstrap_map.l_ld_readonly [BZ #28340]
    • Needs review to fix valgrind issue with ppc64le.
  • _FORTIFY_SOURCE=3 improvements
    • Siddhesh looking for review.
  • Series 3601: Improve hypot()
    • Changes requested by Paul Zimmerman.

Meeting: 2021-10-11 @ 0900h EST5EDT

Cancelled.

Meeting: 2021-10-04 @ 0900h EST5EDT

Video/Audio: https://bluejeans.com/9093064454

IRC: #glibc on OFTC.

Review new patches and restart review at the top.

Meeting: 2021-09-27 @ 0900h EST5EDT

Video/Audio: https://bluejeans.com/9093064454

IRC: #glibc on OFTC.

Review new patches and restart review at the top.

  • 365 patches in the NEW/NOBODY queue.
  • 45462: [PING,5] localedata: add new locale ab_GE
  • 45457: linux: Fix a non-constant expression in _Static_assert
    • Accepted. Under review from Paul Eggert.
  • [v1,1/3] benchtests: Add medium cases and increase iters in bench-memset.c
    • Should be looked at by H.J. for correctness.
    • Results attached. Look OK.
  • 45437 [powerpc] Fix unrecognized instruction errors with recent binutils
    • Is this a binutils bug?
    • Paul should commit. What review is he expecting?
  • 45436 puts: indentation
    • Florian to follow up.
  • 45434 pthread/tst-cancel28: Fix barrier re-init race condition
    • Adhemerval has this on his list.
  • 45430 elf: Add libc_get_static_tls_bounds [BZ #16291]

    • Outstanding question from Florian.
  • 45429 [v3] elf: Avoid nested functions in the loader (all ports) [BZ #27220]
    • Complexity looks reasonable.
    • No global data introduced.
    • Should we test on hppa and ia64?
    • Florian: Will test ia64.
    • Adhemerval: Will test hppa.
  • Down to 344 patches.
  • 45423 Disable symbol hack in libc_nonshared.a
    • Florian to ACK. Agreement that it looks correct.
  • 45422 [v2] i386: Port elf_machine_{load_address,dynamic} from x86-64
    • Accepted by HJ
  • 45407 powerpc: Delete unneeded ELF_MACHINE_BEFORE_RTLD_RELOC
    • Accepted by Rafael
  • Break to review status from meeting participants.
  • Siddhesh: Inputs to math benchmarks?
    • Paul Z. wanted to add an extra sets.
    • Agreed that workloads could be added.
  • 44622 [v3] Fixed inaccuracy of j0f (BZ #28185)
    • Carlos to review.
  • Patrick: Recommend that new users should login to patchwork because
    • you can't set your settings until you have a login.
    • Carlos will amend wiki with notes.
  • Patrick: Is OFTC #glibc used for hte review meeting?
    • Carlos: It can be used if you want. It is the official IRC, and we will
      • discuss patch review via IRC.
  • Florian: 3 things to raise.
  • Adhemerval:

Meeting: 2021-09-20 @ 0900h EST5EDT

Video/Audio: https://bluejeans.com/9093064454

IRC: #glibc on OFTC.

Review new patches and restart review at the top.

Meeting: 2021-09-13 @ 0900h EST5EDT

Video/Audio: https://bluejeans.com/9093064454

IRC: #glibc on OFTC.

Review new patches and restart review at the top.

  • 334 patches to review.
  • 44860 [v2] nptl: Fix type of pthread_mutexattr_getrobust_np, pthread_mutexattr_setrobust_np (bug 28036)
    • Carlos will review.
  • Florian got feedback on pthread_cancel ESRCH changes.
    • Pushed to master and relase/2.34/master.
  • sysconf revert patches?
    • Florian reviewed the patches.
    • Final patches have some issues in the test case.
    • The rest is OK.
  • Siddhesh has a couple of locale realted patches that need review.
    • 44430 - v2 setlocale needs review (codonell)
    • 44607 - iconvconfig: Fix behaviour with --prefix [BZ #28199] (arjun)
  • Carlos to raise locale source file header review with Mike Frysinger.
    • Attempt to get consensus on default headers to use.
  • Szabolcs: No patches, but regression report of the latest glibc release has dynamic tls fixes.
  • Carlos is looking to review 2767
  • C.UTF-8 is in master, but will not be backported to release/2.34/master due to distribution impact.
    • Distros can either pick it up manually, or when 2.35 releases.
  • RISC-V looking at extending this to 4K vector.
    • Arm SVE is passing down linked-list kind of structure that links up the extensions.
    • Kernel dumps register states on the stack.
    • Headers point to extra register state.
    • AT_MINSIGSTKSZ is used there.
    • Florian will post a note about the Arm SVE case as an example to review.
  • 44926 - [v9,1/1] Extend struct r_debug to support multiple namespaces [BZ #15971]
    • Florian to review.
  • 44924 - [v2] posix: Fix attribute access mode on getcwd [BZ #27476]
    • Florain to review (since he reviewed v1)
  • 44897 - Add MADV_POPULATE_READ and MADV_POPULATE_WRITE from Linux 5.14 to bits/mman-linux.h
  • 32-bit Arm atomic issue.
  • Stopped at 44884.

Meeting: 2021-09-06 @ 0900h EST5EDT

Video/Audio: https://bluejeans.com/9093064454

IRC: #glibc on OFTC.

Review new patches and restart review at the top.

  • Reviewed status of pthread_cancel ESRCH patches.
  • Looked at sysconf revert. Adhemerval reviewing.

Meeting: 2021-08-30 @ 0900h EST5EDT

Video/Audio: https://bluejeans.com/9093064454

IRC: #glibc on OFTC.

Review new patches and restart review at the top.

Meeting: 2021-08-23 @ 0900h EST5EDT

Cancelled.

Meeting: 2021-08-16 @ 0900h EST5EDT

Video/Audio: https://bluejeans.com/9093064454

IRC: #glibc on OFTC.

Review new patches and restart review at the top.

  • Upstream patches 271 (New/Nobody)
  • Namespace patch with glibc/binutils/gdb and r_debug
    • Discussed keeping r_debug and bumping r_version
    • Export r_debug_v2 only via DT_DEBUG
  • v2 version of glibc change for removing COPY relocations.
  • Add THP support in glibc's malloc
    • https://patchwork.sourceware.org/project/glibc/list/?series=2660

      • Looking for architecture feedback.
      • Looking for Distro feedback.
    • Enable madvise syscall after successful mmap
    • How do we get feedback from the workload testing?
    • What about MADV_FREE?
      • Not reviewed, but could be related.
    • What about heap_trim? malloc_trim?
    • Can have a follow-up patch to directly use huge pages.
    • On some workloads THP needs a lot of tuning.
      • Can we provide a THP pool?
      • Be as simple as possible.

Meeting: 2021-08-09 @ 0900h EST5EDT

Video/Audio: https://bluejeans.com/9093064454

IRC: #glibc on OFTC.

Review new patches and restart review at the top.

Meeting: 2021-08-02 @ 0900h EST5EDT

Video/Audio: https://bluejeans.com/9093064454

IRC: #glibc on OFTC.

Review new patches and restart review at the top.

Meeting: 2021-07-26 @ 0900h EST5EDT

Video/Audio: https://bluejeans.com/9093064454

IRC: #glibc on OFTC.

Review new patches and restart review at the top.

Meeting: 2021-07-19 @ 0900h EST5EDT

Video/Audio: https://bluejeans.com/9093064454

IRC: #glibc on OFTC.

Review new patches and restart review at the top.

Meeting: 2021-07-12 @ 0900h EST5EDT

Video/Audio: https://bluejeans.com/9093064454

IRC: #glibc on OFTC.

Review new patches and restart review at the top.

Meeting: 2021-07-05 @ 0900h EST5EDT

Video/Audio: https://bluejeans.com/9093064454

IRC: #glibc on OFTC.

Review new patches and restart review at the top.

Meeting: 2021-07-05 @ 0900h EST5EDT

Video/Audio: https://bluejeans.com/9093064454

IRC: #glibc on OFTC.

Review new patches and restart review at the top.

No notes for this meeting.

Meeting: 2021-06-28 @ 0900h EST5EDT

Video/Audio: https://bluejeans.com/9093064454

IRC: #glibc on OFTC.

Meeting: 2021-06-21 @ 0900h EST5EDT

Video/Audio: https://bluejeans.com/9093064454

IRC: #glibc on OFTC.

There was a quick discussion about the public LD_AUDIT request after patch review.

  • Any detailed discussions need to happen on libc-alpha in public.
  • Discussion of the Rice University post here:
  • LD_BIND_NOW? How do we handle it.
    • Input how to fix it?
    • We need BIND_NOW to have something to work with SVE.
      • What do we do with SVE
    • NEON support broken on aarch64.
      • Bump the LA_VERSION on Arm64
      • Only supports version 2.
      • Don't know how to fix this for SVE.
      • Fixing it for NEON and SVE both is desirable.
    • Is BIND_NOW just an optimization or semantic issue?
    • Bump LAV_CURRENT to 2 on all arches.
      • Start calling la_symbind in more cases.
      • Call early in the bind now scenario.
        • For the bindnow call sites we add a check afterwards that the callback didn't request PLT_ENTER/PLT_EXIT hooks.
        • If the hook check fails, we report a dlsym error or relocation error.
        • Can we call the callbacks that early?
        • Yes, already relocated.

Meeting: 2021-06-14 @ 0900h EST5EDT

Video/Audio: https://bluejeans.com/9093064454

IRC: #glibc on OFTC.

Meeting: 2021-06-07 @ 0900h EST5EDT

Video/Audio: https://bluejeans.com/9093064454

IRC: #glibc on freenode.

Review new patches and restart review at the top.

Notes:

  • HJ raised upstream issue requiring review: Removal of COPY relocs.
    • Brief discussion about COPY relocs and where the conversation will take place.
    • Carlos suggested we need an upstream meeting between ABI stakeholders.
    • Discussed status of Non-PIE issues briefly and directed conversation to upstream.
  • Discussed briefly GNU Tools Cauldron at LPC since this would present an opportunity to discuss feature and patch status.
  • Suggested DJ review the iconv split up patches from Siddhesh.
  • Adhemerval noted that for 43664 that Szabolcs may want to double check compiler builtin status for Wilco's patch.
  • Stopped at 43663 which will be reviewed by Florian since v4 fixes all previous comments.

Meeting: 2021-05-31 @ 0900h EST5EDT

Video/Audio: https://bluejeans.com/9093064454

IRC: #glibc on freenode.

Review new patches and restart review at the top.

Notes:

  • Start at top of list.
  • Current patch queue count is 251 patches needding review.
  • Started at 43622.
  • Series 2241 "nptl: pthread cancellation refactor" needs review before the thread stack changes can get reworked.
  • 43558 to review for Carlos.
  • 43497, Carlos to ask Jonathan Wakely to repost v3 with a different title and the same in-reply-to chain.
  • 43389, 33388 - Superseded and removed.
  • Stopped at 41649. 197 patches in queue.

Meeting: 2021-05-24 @ 0900h EST5EDT

Video/Audio: https://bluejeans.com/9093064454

IRC: #glibc on freenode.

Review new patches and restart review at the top.

Notes:

  • Reviewed libpthread merge. Two patches left to review. Adhemerval doing review.
  • Two x86 patches for misaligned stack need review.
  • 64-bit time_t patches should be reviewed this week, 25 patches + 1 patch for header export.
  • One RFC patch for slow tls after dlopen, should review before 2.34 release.
  • Two 64-bit time_t test patches need possible rework, to look at after 64_bit time_t changes.

Meeting: 2021-05-17 @ 0900h EST5EDT

Video/Audio: https://bluejeans.com/9093064454

IRC: #glibc on freenode.

Review new patches and restart review at the top.

Notes:

  • 308 patches in queue.
  • Meeting URL has changed to avoid Jitsi server compatibility issues.
  • Lukasz's task is to provided y2038 support for 32-bit processors.
  • Last year some customers were now requiring to have a y2038 solution.
  • Some system designs are so ahead they need to support y2038.
  • Some problems with the refactor with ABI symbols.
  • Newest version is compilable is on github.
  • Not fixing utmp.
  • Review Adhemerval's changes.
  • 7 weeks.
    • 20 patches for 64-bit time_t
    • 8 patches for libpthread removal.
      • Deal with librt removal.
      • Some impact on timer handling.

Meeting: 2021-05-10 @ 0900h EST5EDT

Video/Audio: https://jitsi.member.fsf.org/glibc-patch-review

IRC: #glibc on freenode.

Review new patches and restart review at the top.

Notes:

  • Total of 313 patches in the queue (new/nobody).
  • Discussed making 64-bit time_t (15-16 patches) a serious 2.34 feature.
  • Reviewed up to patch 42023.
  • Paul would like a quick discussion about the match component.

Meeting: 2021-05-03 @ 0900h EST5EDT

Cancelled.

Meeting: 2021-04-26 @ 0900h EST5EDT

Video/Audio: https://jitsi.member.fsf.org/glibc-patch-review

IRC: #glibc on freenode.

Notes:

  • 333 - state new, delegate nobody.
  • 43062 execveat system call wrapper.
  • 42197 fix italian locale capitalization
  • Paul wants to discuss plans for math component.
  • Szabolcs - reviews complete for dynamic tls fixes, need comment reviews.

Meeting: 2021-04-19 @ 0900h EST5EDT

Video/Audio: https://jitsi.member.fsf.org/glibc-patch-review

IRC: #glibc on freenode.

Meeting: 2021-04-12 @ 0900h EST5EDT

Video/Audio: https://jitsi.member.fsf.org/glibc-patch-review

IRC: #glibc on freenode.

Review new patches and restart review at 42561.

Meeting: 2021-04-05 @ 0900h EST5EDT

Cancelled.

Meeting: 2021-03-29 @ 0900h EST5EDT

Video/Audio: https://jitsi.member.fsf.org/glibc-patch-review

IRC: #glibc on freenode.

Review new patches and restart review at 42244.

Notes:

  • Total patch list is at 388 patches, so approximately +30 patches from last time.
  • Discussed libpthread merge and the status of the patches. Moving forward well.

Meeting: 2021-03-22 @ 0900h EST5EDT

Cancelled.

Meeting: 2021-03-15 @ 0900h EST5EDT

Video/Audio: https://jitsi.member.fsf.org/glibc-patch-review

IRC: #glibc on freenode.

Review new patches and restart review at 42023.

Notes:

  • 352 patches in State=New Delegate=Nobody.
  • Discussed auto-posting committed patches to list using patchwork to drive that.
  • Asked DJ to look at https://patchwork.sourceware.org/project/glibc/list/?series=1817

  • 42282 - 42334: Any review? Some sparse review. Settle the filesystem 2038 issue? On disk file format issue for out-of-range dates. Repost test as a distinct series.
  • 42277 - Review x86-64 Avoid RTM abort inside a RTM region. Still need reviewer.
  • 42244. Stopped here.

Meeting: 2021-03-08 @ 0900h EST5EDT

Video/Audio: https://jitsi.member.fsf.org/glibc-patch-review

IRC: #glibc on freenode.

Review new patches and restart review at 41481.

Notes:

  • Manually review State=Under Review, Delegate=Nobody (15 patches).
  • Reviewed the patches without delegate and moved back to NEW for review.
  • 356 patches in queue for current state=new delegate=nobody.
  • 15 patches were delegate=nobody state=under review.
  • Stopped at 42023 "powerpc: backtrace cleanup."
  • Guest made a comment "https://sourceware.org/glibc/wiki/ThreadPropertiesAPI" which sparked a discussion about TLS issues that need fixing.

Meeting: 2021-03-01 @ 0900h EST5EDT

Video/Audio: https://jitsi.member.fsf.org/glibc-patch-review

IRC: #glibc on freenode.

Review new patches and restart review at 41785.

Notes:

  • IBM: Talk to Tulio and Paul Murphy about patch review for IBM.
  • Changes requested for 41576, 41572 - Raise with Fangrui Song discuss coverage and what support looks like.
  • Stopped at 41481.

Meeting: 2021-02-22 @ 0900h EST5EDT

Video/Audio: https://jitsi.member.fsf.org/glibc-patch-review

IRC: #glibc on freenode.

Review new patches and restart review at 41628.

Notes:

  • No reviewers found for TLS P&C issues yet, review next week.

  • Next week the 64-bit time_t patches for final support will be posted for review.

Meeting: 2021-02-15 @ 0900h EST5EDT

Cancelled.

Meeting: 2021-02-08 @ 0900h EST5EDT

Video/Audio: https://jitsi.member.fsf.org/glibc-patch-review

IRC: #glibc on freenode.

Notes:

  • Patch queue was 430 patches at the start of the meeting.
  • Closed the meeting with 412 patches. Up +5 patches from last week after review.
  • Ended meeting at 41628 (though Szabolcs was going to close that patch).

Meeting: 2021-02-01 @ 0900h EST5EDT

Video/Audio: https://jitsi.member.fsf.org/glibc-patch-review

IRC: #glibc on freenode.

Notes:

  • All release blockers reviewed. There were none left.
  • Started reviewing the queue.
  • Stopped at patch ID 41767. There were 407 patches left in the queue.

Meeting: 2021-01-25 @ 0900h EST5EDT

Video/Audio: https://jitsi.member.fsf.org/glibc-patch-review

IRC: #glibc on freenode.

Will review release blocker patches and desirable patches.

Notes:

  • All release blockers reviewed.
  • Decided that hard freeze should happen.
  • Moving SIGSTKSZ issues to be pushed as soon as 2.34 opens. Needs warning NEWS item for upcoming deprecation.
  • Discussed upcoming features.

Meeting: 2021-01-18 @ 0900h EST5EDT

Video/Audio: https://jitsi.member.fsf.org/glibc-patch-review

IRC: #glibc on freenode.

Will review release blocker patches and desirable patches.

Notes:

  • Moved DSO sorting to glibc 2.34 to resolve issues with dual sorting transition.
  • Carlos to review _SC_MINSIGSTKSZ and deprecate _SC_SIGSTKSZ_SOURCE.
  • Florian to review COMMON_CPUID_INDEX_MAX handshake.
  • Florian to review Fix getdents64 fallback on mips64-n32
  • Fix static PIE ifunc resolvers.
  • Adhemerval to email HJ and Szabolcs on status.
  • Florian Weimer to ping Martin Sebor on PR/98512.
  • Florian to raise issue around check for x86 ISA v2 and v3 and virt.
  • What about deadlines for the release?
    • Keep only those issues that are ABI regressions.
  • How do we shorten the release window?
    • Discuss on libc-alpha with new thread topic.

Meeting: 2021-01-11 @ 0900h EST5EDT

Video/Audio: https://jitsi.member.fsf.org/glibc-patch-review

IRC: #glibc on freenode.

Will review release blocker patches and desirable patches.

Notes:

  • Hosted by Siddhesh
  • Adhemerval to run the call as Release Manager
  • AArch64 static PIE needs a fix and patch reviewer should be found.
  • Adhemerval Notes

Meeting: 2021-01-04 @ 0900h EST5EDT

Video/Audio: https://jitsi.member.fsf.org/glibc-patch-review

IRC: #glibc on freenode.

Queue is clean and we are reviewing patches.

Notes:

  • Currently reviewing patches starting at patch ID 41361
  • Last meeting NEW queue size was 424 patches.
  • Reviewed patches for the release.
    • Release blockers patchwork patch review complete.
    • Desirable for release pathcwork patch review complete.
    • All patches have delegates for review or are in progress.
  • Marking LibOS RFC patchwork patches as dropped
    • GNU Cauldron 2019 feedback was the patches need updating.
    • LibOS implementation needs reworking to be supported by upstream.
    • Marking all patches as dropped.
  • Build failures with gcc 11 / glibc mainline need fixing before release.

Meeting: 2020-12-14 @ 0900h EST5EDT

Video/Audio: https://jitsi.member.fsf.org/glibc-patch-review

IRC: #glibc on freenode.

The goal is to review the entire queue for a clean and manageable state before returning to start tackling the patches. This ensures that we can run automation across the queue without worrying about patches that don't need to be handled, and to make people aware of our efforts.

Notes:

  • Currently at patch ID 33866, with only 558 patches in the queue.
  • We are 72% of the way through the full patch queue.
  • After we walk the full queue once doing cleanup we'll start handling patch-by-patch review.
  • Completed the full queue review.
  • Started working the patch queue from the top of the list downward. Made it to patch ID 41361.

Meeting: 2020-12-07 @ 0900h EST5EDT

Video/Audio: https://jitsi.member.fsf.org/glibc-patch-review

IRC: #glibc on freenode.

The goal is to review the entire queue for a clean and manageable state before returning to start tackling the patches. This ensures that we can run automation across the queue without worrying about patches that don't need to be handled, and to make people aware of our efforts.

Notes:

  • Currently at patch ID 40823, which is 10% of the way through the patch list.
  • Estimate that in 9 more meetings we get through the queue. Completion on 2021-02-15 (after glibc 2.33 release).
  • No meetings on December 21st, or December 28th, taking two weeks off for the holidays.
  • Ended meeting at patch ID 33866.

Meeting: 2020-11-30 @ 0900h EST5EDT

Video/Audio: https://jitsi.member.fsf.org/glibc-patch-review

IRC: #glibc on freenode.

The initial goal of several of the first meetings is to get the patchwork queue down to a clean manageable state. Along with writing automation to keep it in a clean state.

Notes:

  • Added "Patch IDs" to the list of things to enable in patchwork for review.
  • Discussed the process and the problems with using patchwork.
  • Started reviewing the list of patches and got the queue down to 1239.
  • Process is to skip patches from people who are known to actively maintain their queue.
  • Dug deeper into queues from submitters with fewer patches or people who are not expected to be deeply involved in the process.
  • Delegated some patches to reviewers that were on the call.
  • Ended meeting at patch ID 40823, reviewed 125 of 1239 (10% review).
  • Discussed patch sending problems and issues with patches sent to the list which are not understood by patchwork.
  • Arjun noted git imap-send to send patches to your imap drafts folder and then send them on to the list e.g. git format-patch --stdout HEAD^ | git imap-send

None: PatchworkReviewMeetings (last edited 2022-08-08 21:00:05 by CarlosODonell)