This is the mail archive of the gcc-bugs@gcc.gnu.org mailing list for the GCC 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]

[Bug middle-end/44826] Mozilla build ICE at Invalid first operand of MEM_REF.



------- Comment #6 from rguenther at suse dot de  2010-07-06 11:45 -------
Subject: Re:  Mozilla build ICE at Invalid first operand
 of MEM_REF.

On Tue, 6 Jul 2010, hubicka at ucw dot cz wrote:

> 
> 
> ------- Comment #5 from hubicka at ucw dot cz  2010-07-06 11:41 -------
> Subject: Re:  Mozilla build ICE at Invalid first
>         operand of MEM_REF.
> 
> > What's the patch needed to trigger it?
> The DECL_BY_REFERENCE change.  THe one I sent to list should be enough.  So if
> you could look into it, it would be great.  On Mozilla this gimplify trick
> works, but later it causes different ICE.  Here regimplification splits
> 
> var = fncall ()
> 
> into
> 
> tmp.ssa_ver =fncall ()
> var = tmp.ssa_ver
> 
> that is completely wrong since type of tmp is aggregate and we die later
> checking
> SSA form.

Well, that sounds like caused by your needs_to_live_in_memory change.

There is asymmetry wrt caller / callee and DECL_BY_REFERENCE handling.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=44826


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