This is the mail archive of the libc-alpha@sourceware.org mailing list for the glibc project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [PING][PATCH 1/8] Use gcc attribute ifunc in libc_ifunc macro instead of inline assembly due to false debuginfo.


PING

On 07/20/2016 04:05 PM, Stefan Liebler wrote:
PING

On 07/12/2016 04:57 PM, Stefan Liebler wrote:
PING

see latest patches:

[PATCH 1/8] Use gcc attribute ifunc in libc_ifunc macro instead of
inline assembly due to false debuginfo.
(This alters ppc and common code!)
https://www.sourceware.org/ml/libc-alpha/2016-07/msg00093.html

[PATCH 2/8] s390: Refactor ifunc resolvers due to false debuginfo.
https://www.sourceware.org/ml/libc-alpha/2016-07/msg00092.html

[PATCH 3/8] i386, x86: Use libc_ifunc macro for time, gettimeofday.
https://www.sourceware.org/ml/libc-alpha/2016-07/msg00091.html

[PATCH 4/8] ppc: Use libc_ifunc macro for time, gettimeofday.
https://www.sourceware.org/ml/libc-alpha/2016-06/msg00900.html

[PATCH 5/8] Use libc_ifunc macro for clock_* symbols in librt.
https://www.sourceware.org/ml/libc-alpha/2016-07/msg00090.html

[PATCH 6/8] Use libc_ifunc macro for system in libpthread.
https://www.sourceware.org/ml/libc-alpha/2016-07/msg00097.html

[PATCH 7/8] Use libc_ifunc macro for vfork in libpthread.
https://www.sourceware.org/ml/libc-alpha/2016-07/msg00098.html

[PATCH 8/8] Use libc_ifunc macro for siglongjmp, longjmp in libpthread.
https://www.sourceware.org/ml/libc-alpha/2016-07/msg00095.html


Hint:
These patches won't be commited before glibc 2.24 release.
The patch "[PATCH 03/14 v7] Do not stack-protect ifunc resolvers."
(https://www.sourceware.org/ml/libc-alpha/2016-07/msg00174.html) depends
on this series.

On 07/06/2016 01:45 PM, Florian Weimer wrote:
On 07/04/2016 03:44 PM, Stefan Liebler wrote:
These changes (i.e. patch 1/8) look reasonable to me.  Perhaps wait
another day if someone else objects.
Do I have to wait for an explicit okay from intel/power maintainer?

1/8 is power-only, and I assumed that you could deal with this question
internally. :)

The x86 changes need more substantial review from someone more familiar
with this part of glibc, I think.

Florian







Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]