This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Re: glibc 2.18 freeze!
- From: Siddhesh Poyarekar <siddhesh at redhat dot com>
- To: "Joseph S. Myers" <joseph at codesourcery dot com>
- Cc: Adhemerval Zanella <azanella at linux dot vnet dot ibm dot com>, libc-alpha at sourceware dot org
- Date: Thu, 13 Jun 2013 15:37:01 +0530
- Subject: Re: glibc 2.18 freeze!
- References: <51ACFE36 dot 5000603 at redhat dot com> <CAAKybw9ZuOKSpXgCGb5mFVHkKmrnUCKAdStg+i7+_ckUM9Qp2g at mail dot gmail dot com> <51ADFC0B dot 1060907 at linux dot vnet dot ibm dot com> <20130611183029 dot B62D52C0A2 at topped-with-meat dot com> <51B78382 dot 6000403 at linux dot vnet dot ibm dot com> <20130611213211 dot EA4E12C09F at topped-with-meat dot com> <51B893F9 dot 6010401 at linux dot vnet dot ibm dot com> <Pine dot LNX dot 4 dot 64 dot 1306130010000 dot 9276 at digraph dot polyomino dot org dot uk>
On Thu, Jun 13, 2013 at 12:15:57AM +0000, Joseph S. Myers wrote:
> On Wed, 12 Jun 2013, Adhemerval Zanella wrote:
>
> > Pushed as 6a97b62a5b4f18aea849d6f4d8de58d1469d2521. I'll fill a GCC bug
> > about this issue. I can think some possible solutions:
>
> I'm seeing a build failure on x86_64 (with GCC 4.7) caused by that commit
> (confirmed that the build passes as of the previous commit):
>
> readelf -s /home/jsm28/tmp/tmpdir/glibc/build-64/elf/ld.so \
> | gawk '($7 ~ /^UND(|EF)$/ && $1 != "0:" && $4 != "REGISTER") { print; p=1 } END { exit p != 0 }'
> 2: 0000000000000000 0 NOTYPE GLOBAL DEFAULT UND __GI_memmove
> 509: 0000000000000000 0 NOTYPE GLOBAL DEFAULT UND __GI_memmove
> make[2]: *** [/home/jsm28/tmp/tmpdir/glibc/build-64/elf/ld.so] Error 1
>
I don't see this in any of my tests with gcc-4.7.2 or with 4.8, with
or without -O3 in CFLAGS. Did you do a clean build? In fact the
patch resolves one of the testsuite breakages I had reported and
analyzed incorrectly (the __gettimeofday hidden_def removal - the
intent was correct, but not a resolution to the problem I was trying
to solve).
However, I've had the earlier patch to disable
tree-loop-distribute-patterns too, and I don't see how this would work
without it. Test cases on the Fedora rawhide build (-O3 with 4.8)
crash at random without that patch, and it seems obvious to me that
they would.
Siddhesh