On Fri, Sep 28, 2001 at 03:15:32PM -0400, Mark Hahn wrote:
>
> seems like an interaction between your signal code and pthread,
> no obvious reason to blame the kernel:
I've updated my test program to address your issues:
1) I removed the SIGSEGV handler (you know with the fprintf())
2) I modified it to use separate memory for all thread initialization data
The segfault still occurs. (to reenable backtrace #define BACKTRACE to 1 at
the beginning of the program) core is not dumped, because threaded programs
do not drop core.
I still think it's a kernel issue, at least the SMP dependancy tells me so.
(remember SMP kernel on an UP is affected too)
I also attach the strace of the process, which also shows the segfault. (it
is made by strace 4.4 which can follow threads)
PS: again please Cc me, since I'm not subscribed.
-- Bazsi PGP info: KeyID 9AF8D0A9 Fingerprint CD27 CFB0 802C 0944 9CFD 804E C82C 8EB1
This archive was generated by hypermail 2b29 : Sun Oct 07 2001 - 21:00:14 EST