Re: Mainline merge part 14 - profiler improvements

[Jan Hubicka <jh at suse dot cz>]

Uff, I didn't noticed there is review on the patch (for some purpose it is not
in my inbox).  I broke the patch already, so I will update the bits and send
anotehr version shortly.

> > > +       /* propagate counters of main thread to global ones.  */
> > > +       __bb_thread_end_func (GET_ARC_COUNTERS_TSD);
> > > + 
> > > +       /* If we are here for the first time (could be more times due
> > > +          to forking), move results from static counters.  */
> > > +       LOCK_ARC_COUNTERS;
> > 
> > What, exactly, are you locking here?  AFAICT, __bb_thread_end_func
> > is synchronous with this function.
> 
> Maybe, I'm not so sure about it. We probably won't spoil anything by leaving
> lock here.

I would not feel confortable to have bits of code no one knows the purpose,
pelase try to take a look at it.
> 
> > > +       UNLOCK_ARC_COUNTERS;
> > > +     }
> > > + 
> > > + #endif
> > > + 
> > > + #ifdef __GTHREADS
> > > +   if (arc_counters)
> > > +     {
> > > +       LOCK_ARC_COUNTERS;
> > 
> > Unlock followed immediately by lock?
> 
> Stupid, of course. I will fix it.
Looks like my merger error, sorry.
> 
> > > ! __bb_thread_start_func (void)
> > > ! {
> > > !   gcov_type *ret;
> > > ! 
> > > !   if (arc_counters && __gthreads_active)
> > > !     {
> > > !       ret = calloc (aofs, sizeof (gcov_type));
> > > !       SET_ARC_COUNTERS_TSD (ret);
> > 
> > Failure mode when (1) calloc fails,
> 
> I will think of what to do with this (I don't see completely correct solution
> just now).

I guess just abort the program with error message is sane.

Honza
> 
> >or (2) thread key not available?
> 
> I should test it __bb_init_func. I will fix it (but anyway, I'm probably the
> first one to ask for a key; so I really should obtain one :-)
> 
> Zdenek