From: Paul H. Hargrove (PHHargrove_at_lbl_dot_gov)
Date: Mon Dec 01 2008 - 17:44:24 PST
Thanks, Neil.
When you have a chance, please see if the 1-line change below fixes the
prctl failure.
I believe the test was placing itself into SECCOMP mode and thus getting
killed when it tried to do much of anything.
In other words, the bug is probably in the test itself.
-Paul
--- tests/prctl.c 1 Dec 2008 00:42:20 -0000 1.4
+++ tests/prctl.c 2 Dec 2008 01:34:43 -0000
@@ -96,7 +96,7 @@
/* Not safe to change this one */
{"PR_GET_ENDIAN", my_type_int_ptr, PR_GET_ENDIAN, PR_SET_ENDIAN,
X, X, },
#endif
-#if defined(PR_GET_SECCOMP)
+#if defined(PR_GET_SECCOMP) && 0
{"PR_GET_SECCOMP", my_type_int, PR_GET_SECCOMP, PR_SET_SECCOMP,
1, 0, },
#endif
#if defined(PR_GET_TSC)
Neal Becker wrote:
> On Monday 01 December 2008, Paul H. Hargrove wrote:
>
>> Neal,
>>
>> As always, thanks for your assistance in testing.
>> The ptrace.ct bug has already been identified (BLCR bug 2455), with a
>> 2.6.26.6-49.fc8/x86_64 kernel.
>> However, the prctl failure is new to me (the test itself having been
>> added hours before 0.8.0_b1 was posted) and does not fail on my
>> 2.6.26.6-49.fc8/x86_64 platform. Any additional information you can
>> provide about this failure would be helpful (preferably in the form of a
>> new Bugzilla entry).
>> I will investigate both between now and the final 0.8.0 release.
>>
>>
>
> This is 2.6.27.5-41.fc9.x86_64
>
--
Paul H. Hargrove PHHargrove_at_lbl_dot_gov
Future Technologies Group Tel: +1-510-495-2352
HPC Research Department Fax: +1-510-486-6900
Lawrence Berkeley National Laboratory