Re: Freeze for the upcoming glibc 2.39 release - status

[Palmer Dabbelt <palmer@dabbelt.com>]

On Fri, 12 Jan 2024 09:19:14 PST (-0800), Carlos O'Donell wrote:
> On 1/11/24 17:24, Andreas K. Huettel wrote:
>> [PATCH v2] stdlib: Reinstate stable mergesort implementation on qsort
>> https://patchwork.sourceware.org/project/glibc/list/?series=29039
>> delegated, Florian is looking at this
>> release blocker
>
> Agreed. Florian is reviewing this.
>
>> [PATCH v2] linux: Fix fstat64 on alpha and sparc64
>> https://patchwork.sourceware.org/project/glibc/list/?series=29040
>> needs review
>> release blocker
>
> I have just reviewed this.
>
> Adhemerval, Please commit.
>
>> [PATCH v2] Make __getrandom_nocancel set errno and add a _nostatus version
>> https://patchwork.sourceware.org/project/glibc/list/?series=29130
>> R-B Adhemerval
>> bug fix - should go in now?
>
> Agreed.
>
>> [PATCH] debug/getwd_chk.c: warning should be emitted for the __getwd_chk symbol
>> https://patchwork.sourceware.org/project/glibc/list/?series=27941
>> R-B Adhemerval
>> bug fix - should go in now
>>
>
> Agreed.
>
>> [PATCH v3] RISC-V: Enable static-pie
>> https://patchwork.sourceware.org/project/glibc/list/?series=29034
>> R-B Adhemerval
>> feature - could be in already - now or after release?
>
> Agreed. It should go in ASAP.
>
> Adhemerval, Will you commit this?
>
>
>> [PATCH] math: remove exp10 wrappers
>> https://patchwork.sourceware.org/project/glibc/list/?series=29463
>> R-B Adhemerval
>> feature - could go in now, last minute for it
>
> Agreed. This should go in ASAP.
>
>> [PATCH v11 0/7] RISC-V: ifunced memcpy using new kernel hwprobe interface
>> https://patchwork.sourceware.org/project/glibc/list/?series=29471
>> needs another review?
>> lots of code, but already many iterations - soon or after release
>
> My feeling is that this should go in as soon as 2.40 opens.
>
> It introduces too many generic changes to make me comfortable for 2.39 right now.

OK, good timing -- my i686-gnu build just passed, so I was rebasing to 
commit stuff.  I'm not super opposed to holding off, it's pretty late in 
the cycle, but we'll probably end up with a bit of user headaches as 
folks will likely just call the syscall directly.

I'll hold off on commiting it unless folks feel like we should pick it 
up, just let me know...

>
>> [PATCH] x86-64: Check if mprotect works before rewriting PLT
>> https://patchwork.sourceware.org/project/glibc/list/?series=29560
>> needs review / discussion?
>
> Reviewed and needs a v2.
>
> HJ, Please review my comments.
>
>> [PATCH v3] rtld: Add glibc.rtld.enable_secure tunable
>> https://patchwork.sourceware.org/project/glibc/list/?series=29055
>> needs review, test failure
>> feature - after release
>
> Agreed. This is glibc 2.40 material at this point.