[PATCH v6 5/9] nptl: Remove g_refs from condition variables

Carlos O'Donell carlos@redhat.com
Fri Jan 17 16:55:53 GMT 2025


On 1/17/25 11:22 AM, Florian Weimer wrote:
> * Carlos O'Donell:
> 
>> One potential benefit is if retaining the field help downstream in
>> some way with their implementation of an in-place upgrade strategy, so
>> let me discuss some of that strategy now.
>>
>> Upstream glibc considers in-place upgrades are unsupported, and doing
>> them is up to downstream to resolve.
> 
> I have not reviewed the technical points you make below (sorry), but in
> these statements above there is an implied assumption that we would not
> backport this to older release branches upstream.  I'm not sure this is
> a given because it's a bug that is quite widely encountered.

My statement in in-place upgrades is distinct from my position on backporing this fix.

I think we *will* backport it to older release branches, and I plan to do that for
release/2.39/master and release/2.40/master using the patch I developed.

> Of ourse, we could easily drop this patch from an upstream backport
> (even if we do not add other backwards compatibility features as part of
> that backport).  So I don't have any objections to including it in the
> upcoming release.

Correct.

-- 
Cheers,
Carlos.



More information about the Libc-alpha mailing list