This is the mail archive of the libc-alpha@sourceware.org mailing list for the glibc project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: tst-pthread-getattr changes break sparc


On Fri, 27 Jul 2012 08:30:53 -0700, Richard wrote:

> On 07/27/2012 07:36 AM, Siddhesh Poyarekar wrote:
> > -  /* ??? STACK_POINTER_OFFSET is always defined now.  */
> > -#if defined (STACK_DYNAMIC_OFFSET) || defined
> > (STACK_POINTER_OFFSET) +#if defined (STACK_DYNAMIC_OFFSET) ||
> > STACK_POINTER_OFFSET must_align = true;
> >    extra_align = BITS_PER_UNIT;
> >  #endif
> 
> Ah, well, fyi, doing a test in the #if isn't going to work.
> 
> Also, I'll note that STACK_DYNAMIC_OFFSET has a default too, it's
> just not here in explow.c, but in  function.c.  Further, as the
> comment above says, we can't actually evaluate S_D_O at this time.
> Which is going to leave us with something like
> 
>   if (ACCUMULATE_OUTGOING_ARGS || STACK_POINTER_OFFSET)
>     must_align = true, extra_align = BITS_PER_UNIT;
> 
> We *could* probably arrange for STACK_DYNAMIC_OFFSET to be forced to
> be aligned to some value that gets computed from the maximum required
> by dynamic allocation though.  That way we wouldn't have to take
> S_D_O into account at all -- we'd know it was cooperating.
> 
> Wanna file a gcc bz enhancement request for this?
> 

Thanks, I'll do this today.

Regards,
Siddhesh


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]