>>>>> Hiroyuki Machida writes:
> Hello Andreas,
> I had a experience that pthread_sighander() of current glibc-2.2
> breaks stack. I tracked down the problem, and finally found the
> mismatch between kenrel and glibc-2.2.
> Current kernel pass following args to the signal handler for the
> case of not SA_SIGINFO specified.
> a0 signal number
> a1 0 (cause code?)
> a2 pointer to sigcontext struct
> But, the pthread_sighander() of glibc-2.2 expects;
> 1st arg. signal number
> 2nd arg. sigcontext struct itself (not pointer)
> Patches attached below. Please apply.
Thanks, I've committed them,
SuSE Labs email@example.com