NPTL bug? 2.6.11.12 on an AMD64

[jstipins]

Hi there,

I apologize in advance if this question is inappropriate for this mailing
list.  I welcome any suggestions for a more appropriate forum.

I am running kernel 2.6.11.12 on an AMD Athlon 64, but the kernel is just
the straightforward i486 compilation -- no optimizations, no 64-bit.  The
distribution is LFS 6.1, which apparently runs without any problem.  The
glibc version is 2.3.4, built with gcc 3.4.3.

When I build glibc 2.3.4, 2.3.5, or 2.3.6, using gcc 3.4.3 or 4.0.2, the
"make check" test suite fails exactly one test, which I think may be due
to the kernel.  The "nptl/tst-clock2.c" test fails (in every configuration
listed), and we have:

/sources/glibc-build# cat /sources/glibc-build/nptl/tst-clock2.out
difference between thread 0 and 1 too small (0.053511687)

Here is a link to a person who has reported exactly the same problem,
although he does not indicate if he is using an AMD64 processor:
http://archives.linuxfromscratch.org/mail-archives/lfs-support/2005-August/028065.html

Does anyone have any insight into this problem?  There is next to nothing
on google about it (search on "tst-clock2"), which makes me think it's some
rare combination of processor and kernel.

Again, I welcome any suggestions for a more appropriate forum.  I wanted to
ask here before trying the glibc mailing lists, because they seem pretty clear
about not entertaining build error questions.

Thanks very much for your help,
-Janis
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/