This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Re: 2.26 release blockers?
On 04/07/2017 15:57, H.J. Lu wrote:
> On Tue, Jul 4, 2017 at 11:26 AM, Adhemerval Zanella
> <adhemerval.zanella@linaro.org> wrote:
>>
>>
>> On 04/07/2017 14:51, Siddhesh Poyarekar wrote:
>>> On Saturday 01 July 2017 02:21 AM, Adhemerval Zanella wrote:
>>>> I would like add both C11 threads support [1] and the final patches
>>>> to finally remove the sysdep-cancel.h asm hackery [2].
>>>>
>>>> For C11 I think it generated a consensus that there is no impending
>>>> technical issues about current approach using underlying POSIX
>>>> primitives and Tovarld also think it is a good idea to add it on
>>>> 2.26.
>>>
>>> Adhemerval and I discussed this on IRC earlier today and it looks to me
>>> like the C11 threads patches should not affect architecture testing
>>> beyond adding new symbols to the ABI. I'm inclined to allow it as a
>>> freeze exception but I know Joseph has his reservations about it.
>>> Joseph, if you still object, could you elaborate on why you think this
>>> would affect architecture testing? Overall the code impact is much less
>>> than some of the blockers currently approved given that it is a brand
>>> new ABI/API that is isolated from everything else in glibc.
>>>
>>> Either way, it would be nice if someone could help review these (I'll
>>> give it a shot too later in the week, but I may need a second pair of
>>> eyes on some of them) so that it could make it early into 2.27 if we all
>>> end up agreeing that 2.26 is not a suitable target.
>>>
>>> Siddhesh
>>>
>>
>> Thanks for bringing this up Siddhesh, if any is interested the latest
>> version of the patch is here [1].
>>
>> I would also like to push the sysdep-cancel.h asm removal, but due
>> x32 concerns raised by H.J.Lu and my plan to send an update revision
>> today I think it can wait. It still somewhat not intrusive, since the
>> changes I am planning is is just to create proper prototypes for
>> internal non-cancellable syscalls and it would be good to a simplified
>> 2.26 as base for new ports inclusions.
>>
>
> Please put wrapper changes into a separate patch so that I can
> compare glibc builds between before and after x32 macro changes.
> They must be equivalent.
>
> Thanks.
>
My understanding is latest sysdep-cancel.h patch for x32 [1] is ok.
Should I still refrain to adding is as a block?
[1] https://sourceware.org/ml/libc-alpha/2017-07/msg00105.html