2.33 Hard Freeze

2.33 Hard Freeze

[Adhemerval Zanella]

Hi all,

Now that we handled all the release blockers [1], the master branch
is now at hard freeze.  The plan is still to set the release to next
week, Feb 1st (or maybe a couple days later). 

As usual, please do not push anything into master and if something
needs to go urgently please cc me with an explanation of why the fix
is necessary (broken build due a glibc issue, 0-day sev-1 security issue,
very important bugfix, etc.).

Currently we have status report for ARC, ARM (hard-float), C-SKY, MIPS
(LE hard-float and BE soft-float), PowerPC (32-bt soft-float, RISC-V
(rv64imac/lp64 and rv64imafdc/lp64) and x86.  I will try to get results
for alpha, hppa, sh, and sparc.

Regarding the current issues:

  - nios2 is failing to build with GCC11 [2].  From the bug report [3]
    it seems to be a binutils issues so I don't consider this a release
    blocker.

  - There are a couple of GCC11 regressions/blockers [4] [5].  For former
    it does seem a to be a gcc regression, while for later I think we
    should defer to next release.

  - The cond-var issue [6].  Carlos is reviewing it, although I do think
    this is kind late for a quite intrusive patchset.  I think it would
    be better to backport it once we it set to upstream.

[1] https://sourceware.org/glibc/wiki/Release/2.33
[2] https://sourceware.org/pipermail/libc-alpha/2021-January/121936.html
[3] https://gcc.gnu.org/bugzilla/show_bug.cgi?id=97683
[4] https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98512
[5] https://sourceware.org/pipermail/libc-alpha/2020-December/120560.html
[6] https://patchwork.sourceware.org/project/glibc/list/?series=1496

Re: 2.33 Hard Freeze

[Szabolcs Nagy]

The 01/26/2021 09:18, Adhemerval Zanella via Libc-alpha wrote:
> Hi all,
> 
> Now that we handled all the release blockers [1], the master branch
> is now at hard freeze.  The plan is still to set the release to next
> week, Feb 1st (or maybe a couple days later). 
> 
> As usual, please do not push anything into master and if something
> needs to go urgently please cc me with an explanation of why the fix
> is necessary (broken build due a glibc issue, 0-day sev-1 security issue,
> very important bugfix, etc.).
> 
> Currently we have status report for ARC, ARM (hard-float), C-SKY, MIPS
> (LE hard-float and BE soft-float), PowerPC (32-bt soft-float, RISC-V
> (rv64imac/lp64 and rv64imafdc/lp64) and x86.  I will try to get results
> for alpha, hppa, sh, and sparc.
> 
> Regarding the current issues:
> 
>   - nios2 is failing to build with GCC11 [2].  From the bug report [3]
>     it seems to be a binutils issues so I don't consider this a release
>     blocker.
> 
>   - There are a couple of GCC11 regressions/blockers [4] [5].  For former
>     it does seem a to be a gcc regression, while for later I think we
>     should defer to next release.
> 
>   - The cond-var issue [6].  Carlos is reviewing it, although I do think
>     this is kind late for a quite intrusive patchset.  I think it would
>     be better to backport it once we it set to upstream.


well there is a regression i could not reproduce yet:

https://sourceware.org/bugzilla/show_bug.cgi?id=27237
https://sourceware.org/pipermail/libc-alpha/2021-January/121916.html

the problematic patch can be reverted (it's not needed
for correctness) but i would like to understand what's
going on.

> 
> [1] https://sourceware.org/glibc/wiki/Release/2.33
> [2] https://sourceware.org/pipermail/libc-alpha/2021-January/121936.html
> [3] https://gcc.gnu.org/bugzilla/show_bug.cgi?id=97683
> [4] https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98512
> [5] https://sourceware.org/pipermail/libc-alpha/2020-December/120560.html
> [6] https://patchwork.sourceware.org/project/glibc/list/?series=1496

Re: 2.33 Hard Freeze

[DJ Delorie]

Adhemerval Zanella via Libc-alpha <libc-alpha@sourceware.org> writes:
> As usual, please do not push anything into master and if something
> needs to go urgently please cc me with an explanation of why the fix
> is necessary (broken build due a glibc issue, 0-day sev-1 security issue,
> very important bugfix, etc.).

I would like to get my nsswitch disable-reloading patch in as it's tiny
and a security issue:

https://sourceware.org/pipermail/libc-alpha/2021-January/121958.html
https://sourceware.org/pipermail/libc-alpha/2021-January/122023.html

Re: 2.33 Hard Freeze

[Adhemerval Zanella]

On 26/01/2021 12:52, DJ Delorie wrote:
> Adhemerval Zanella via Libc-alpha <libc-alpha@sourceware.org> writes:
>> As usual, please do not push anything into master and if something
>> needs to go urgently please cc me with an explanation of why the fix
>> is necessary (broken build due a glibc issue, 0-day sev-1 security issue,
>> very important bugfix, etc.).
> 
> I would like to get my nsswitch disable-reloading patch in as it's tiny
> and a security issue:
> 
> https://sourceware.org/pipermail/libc-alpha/2021-January/121958.html
> https://sourceware.org/pipermail/libc-alpha/2021-January/122023.html
> 

Ok for 2.33.

Re: 2.33 Hard Freeze

[Adhemerval Zanella]

On 26/01/2021 10:36, Szabolcs Nagy wrote:
> The 01/26/2021 09:18, Adhemerval Zanella via Libc-alpha wrote:
>> Hi all,
>>
>> Now that we handled all the release blockers [1], the master branch
>> is now at hard freeze.  The plan is still to set the release to next
>> week, Feb 1st (or maybe a couple days later). 
>>
>> As usual, please do not push anything into master and if something
>> needs to go urgently please cc me with an explanation of why the fix
>> is necessary (broken build due a glibc issue, 0-day sev-1 security issue,
>> very important bugfix, etc.).
>>
>> Currently we have status report for ARC, ARM (hard-float), C-SKY, MIPS
>> (LE hard-float and BE soft-float), PowerPC (32-bt soft-float, RISC-V
>> (rv64imac/lp64 and rv64imafdc/lp64) and x86.  I will try to get results
>> for alpha, hppa, sh, and sparc.
>>
>> Regarding the current issues:
>>
>>   - nios2 is failing to build with GCC11 [2].  From the bug report [3]
>>     it seems to be a binutils issues so I don't consider this a release
>>     blocker.
>>
>>   - There are a couple of GCC11 regressions/blockers [4] [5].  For former
>>     it does seem a to be a gcc regression, while for later I think we
>>     should defer to next release.
>>
>>   - The cond-var issue [6].  Carlos is reviewing it, although I do think
>>     this is kind late for a quite intrusive patchset.  I think it would
>>     be better to backport it once we it set to upstream.
> 
> 
> well there is a regression i could not reproduce yet:
> 
> https://sourceware.org/bugzilla/show_bug.cgi?id=27237
> https://sourceware.org/pipermail/libc-alpha/2021-January/121916.html
> 
> the problematic patch can be reverted (it's not needed
> for correctness) but i would like to understand what's
> going on.

I am not sure about this one, I am trying to replicate it on a powerpc
with the same build logs from the reported issue but I couldn't.
I even tried to run the tests contiguously for some time, but I couldn't
hit the issue. 

From you analysis it does not really seems related, thoughts? 

> 
>>
>> [1] https://sourceware.org/glibc/wiki/Release/2.33
>> [2] https://sourceware.org/pipermail/libc-alpha/2021-January/121936.html
>> [3] https://gcc.gnu.org/bugzilla/show_bug.cgi?id=97683
>> [4] https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98512
>> [5] https://sourceware.org/pipermail/libc-alpha/2020-December/120560.html
>> [6] https://patchwork.sourceware.org/project/glibc/list/?series=1496

Re: 2.33 Hard Freeze

[Szabolcs Nagy]

The 01/26/2021 14:11, Adhemerval Zanella wrote:
> On 26/01/2021 10:36, Szabolcs Nagy wrote:
> > The 01/26/2021 09:18, Adhemerval Zanella via Libc-alpha wrote:
> >> Hi all,
> >>
> >> Now that we handled all the release blockers [1], the master branch
> >> is now at hard freeze.  The plan is still to set the release to next
> >> week, Feb 1st (or maybe a couple days later). 
> >>
> >> As usual, please do not push anything into master and if something
> >> needs to go urgently please cc me with an explanation of why the fix
> >> is necessary (broken build due a glibc issue, 0-day sev-1 security issue,
> >> very important bugfix, etc.).
> >>
> >> Currently we have status report for ARC, ARM (hard-float), C-SKY, MIPS
> >> (LE hard-float and BE soft-float), PowerPC (32-bt soft-float, RISC-V
> >> (rv64imac/lp64 and rv64imafdc/lp64) and x86.  I will try to get results
> >> for alpha, hppa, sh, and sparc.
> >>
> >> Regarding the current issues:
> >>
> >>   - nios2 is failing to build with GCC11 [2].  From the bug report [3]
> >>     it seems to be a binutils issues so I don't consider this a release
> >>     blocker.
> >>
> >>   - There are a couple of GCC11 regressions/blockers [4] [5].  For former
> >>     it does seem a to be a gcc regression, while for later I think we
> >>     should defer to next release.
> >>
> >>   - The cond-var issue [6].  Carlos is reviewing it, although I do think
> >>     this is kind late for a quite intrusive patchset.  I think it would
> >>     be better to backport it once we it set to upstream.
> > 
> > 
> > well there is a regression i could not reproduce yet:
> > 
> > https://sourceware.org/bugzilla/show_bug.cgi?id=27237
> > https://sourceware.org/pipermail/libc-alpha/2021-January/121916.html
> > 
> > the problematic patch can be reverted (it's not needed
> > for correctness) but i would like to understand what's
> > going on.
> 
> I am not sure about this one, I am trying to replicate it on a powerpc
> with the same build logs from the reported issue but I couldn't.
> I even tried to run the tests contiguously for some time, but I couldn't
> hit the issue. 
> 
> From you analysis it does not really seems related, thoughts? 

i cant reproduce this either but i noticed that some
tests run with -DMODULE_NAME=libc for some reason:

test-mutexattr-printers.c
test-mutex-printers.c
test-condattr-printers.c
test-cond-printers.c
test-rwlockattr-printers.c
test-rwlock-printers.c
../sysdeps/unix/sysv/linux/tst-sysconf-iov_max-uapi.c

which can cause problem (though not related to the
particular reported issue).

Re: 2.33 Hard Freeze

[Adhemerval Zanella]

On 26/01/2021 15:30, Szabolcs Nagy wrote:
> The 01/26/2021 14:11, Adhemerval Zanella wrote:
>> On 26/01/2021 10:36, Szabolcs Nagy wrote:
>>> The 01/26/2021 09:18, Adhemerval Zanella via Libc-alpha wrote:
>>>> Hi all,
>>>>
>>>> Now that we handled all the release blockers [1], the master branch
>>>> is now at hard freeze.  The plan is still to set the release to next
>>>> week, Feb 1st (or maybe a couple days later). 
>>>>
>>>> As usual, please do not push anything into master and if something
>>>> needs to go urgently please cc me with an explanation of why the fix
>>>> is necessary (broken build due a glibc issue, 0-day sev-1 security issue,
>>>> very important bugfix, etc.).
>>>>
>>>> Currently we have status report for ARC, ARM (hard-float), C-SKY, MIPS
>>>> (LE hard-float and BE soft-float), PowerPC (32-bt soft-float, RISC-V
>>>> (rv64imac/lp64 and rv64imafdc/lp64) and x86.  I will try to get results
>>>> for alpha, hppa, sh, and sparc.
>>>>
>>>> Regarding the current issues:
>>>>
>>>>   - nios2 is failing to build with GCC11 [2].  From the bug report [3]
>>>>     it seems to be a binutils issues so I don't consider this a release
>>>>     blocker.
>>>>
>>>>   - There are a couple of GCC11 regressions/blockers [4] [5].  For former
>>>>     it does seem a to be a gcc regression, while for later I think we
>>>>     should defer to next release.
>>>>
>>>>   - The cond-var issue [6].  Carlos is reviewing it, although I do think
>>>>     this is kind late for a quite intrusive patchset.  I think it would
>>>>     be better to backport it once we it set to upstream.
>>>
>>>
>>> well there is a regression i could not reproduce yet:
>>>
>>> https://sourceware.org/bugzilla/show_bug.cgi?id=27237
>>> https://sourceware.org/pipermail/libc-alpha/2021-January/121916.html
>>>
>>> the problematic patch can be reverted (it's not needed
>>> for correctness) but i would like to understand what's
>>> going on.
>>
>> I am not sure about this one, I am trying to replicate it on a powerpc
>> with the same build logs from the reported issue but I couldn't.
>> I even tried to run the tests contiguously for some time, but I couldn't
>> hit the issue. 
>>
>> From you analysis it does not really seems related, thoughts? 
> 
> i cant reproduce this either but i noticed that some
> tests run with -DMODULE_NAME=libc for some reason:
> 
> test-mutexattr-printers.c
> test-mutex-printers.c
> test-condattr-printers.c
> test-cond-printers.c
> test-rwlockattr-printers.c
> test-rwlock-printers.c
> ../sysdeps/unix/sysv/linux/tst-sysconf-iov_max-uapi.c
> 
> which can cause problem (though not related to the
> particular reported issue).
> 

From BZ#27237 [1] report it seems it would be better to revert
this last patch and check if we can re-enable on 2.34 once we
fix the regression.

[1] https://sourceware.org/bugzilla/show_bug.cgi?id=27237

Re: 2.33 Hard Freeze

[H.J. Lu]

On Tue, Jan 26, 2021 at 4:46 AM Adhemerval Zanella via Libc-alpha
<libc-alpha@sourceware.org> wrote:
>
> Hi all,
>
> Now that we handled all the release blockers [1], the master branch
> is now at hard freeze.  The plan is still to set the release to next
> week, Feb 1st (or maybe a couple days later).
>
> As usual, please do not push anything into master and if something
> needs to go urgently please cc me with an explanation of why the fix
> is necessary (broken build due a glibc issue, 0-day sev-1 security issue,
> very important bugfix, etc.).
>
> Currently we have status report for ARC, ARM (hard-float), C-SKY, MIPS
> (LE hard-float and BE soft-float), PowerPC (32-bt soft-float, RISC-V
> (rv64imac/lp64 and rv64imafdc/lp64) and x86.  I will try to get results
> for alpha, hppa, sh, and sparc.
>
> Regarding the current issues:
>
>   - nios2 is failing to build with GCC11 [2].  From the bug report [3]
>     it seems to be a binutils issues so I don't consider this a release
>     blocker.
>
>   - There are a couple of GCC11 regressions/blockers [4] [5].  For former
>     it does seem a to be a gcc regression, while for later I think we
>     should defer to next release.
>
>   - The cond-var issue [6].  Carlos is reviewing it, although I do think
>     this is kind late for a quite intrusive patchset.  I think it would
>     be better to backport it once we it set to upstream.
>
> [1] https://sourceware.org/glibc/wiki/Release/2.33
> [2] https://sourceware.org/pipermail/libc-alpha/2021-January/121936.html
> [3] https://gcc.gnu.org/bugzilla/show_bug.cgi?id=97683
> [4] https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98512
> [5] https://sourceware.org/pipermail/libc-alpha/2020-December/120560.html
> [6] https://patchwork.sourceware.org/project/glibc/list/?series=1496

I discovered a 2.33 regression:

https://sourceware.org/bugzilla/show_bug.cgi?id=26625

A patch is posted at

https://sourceware.org/pipermail/binutils/2021-January/115088.html

It should be fixed before 2.33 is released.

-- 
H.J.

Re: 2.33 Hard Freeze

[DJ Delorie]

Adhemerval Zanella <adhemerval.zanella@linaro.org> writes:
>> I would like to get my nsswitch disable-reloading patch in as it's tiny
>> and a security issue:
>> 
>> https://sourceware.org/pipermail/libc-alpha/2021-January/121958.html
>> https://sourceware.org/pipermail/libc-alpha/2021-January/122023.html
>> 
>
> Ok for 2.33.

Thanks, it's in.