Re: [PATCH] Add --core-pattern option to eu-stack

[Mark Wielaard <mark@klomp.org>]

[-- Attachment #1: Type: text/plain, Size: 1325 bytes --]

On Wed, 2014-11-26 at 15:32 +0100, Jan Kratochvil wrote:
> On Thu, 30 Oct 2014 22:55:41 +0100, Mark Wielaard wrote:
> > On Thu, 2014-10-09 at 23:25 +0200, Jan Kratochvil wrote:
> > > +      if (opt_core_pattern == true && show_one_tid == false)
> > > +	argp_error (state,
> > > +		    N_("--core-pattern requires -1"));
> > 
> > Why this restriction?
> 
> I was blindly following Oleg's note which was not so obvious to me, though:
> 
> On Wed, 03 Sep 2014 16:26:41 +0200, Oleg Nesterov wrote:
> # Obviously, this way you can only inspect the thread which dumps the core.
> 
> Therefore I have now tried to remove this limitation of
>   -1                         Show the backtrace of only one thread
> 
> But I have found the other threads end up with:
> 	wait4(13902, [{WIFSIGNALED(s) && WTERMSIG(s) == SIGSEGV && WCOREDUMP(s)}], __WALL, NULL) = 13902
> 
> Therefore they are dead at the core_pattern time, they cannot be ptraced and
> therefore they cannot be unwound. One could only find them in the core file
> itself but that is outside of the scope of this eu-stack feature.

Why do you think that is out of scope? We get both the actual tid
through the command line option and the core file through stdin. Can't
we combine those two to get all information we want/need?

Thanks,

Mark