Re: [patch] Simply ia64-sigill.exp

[Pedro Alves <pedro at codesourcery dot com>]

On Thursday 23 June 2011 08:34:31, Yao Qi wrote:
> On 06/23/2011 02:20 AM, Pedro Alves wrote:
> > Why not fix pthread_cond_signal/pthread_cond_wait?
> > 
> 
> This is a good question :).  I tried but can't find the problem in
> pthread_cond_signal/pthread_cond_wait.  After reading ia64-sigill.c,
> find this test itself has few to do with
> pthread_cond_signal/pthread_cond_wait.  I don't want buggy
> pthread_cond_signal/pthread_cond_wait to make test fail, and write an
> alternative approach to replace pthread_cond_signal/pthread_cond_wait.
> 
> If it is unacceptable, please ignore this patch.

Well, what would you do if you had found that your port had a
broken pthread_mutex?  :-)  Or more generally, if your rework
triggered some compiler/runtime bug in some other port someone
else is working on?

IMO, we should leave it as is, and fix the port's libc, if
that's where the bug is.

-- 
Pedro Alves