This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [PATCH] Fix PR42871, recursion in phi_translate
- From: Daniel Berlin <dberlin at dberlin dot org>
- To: Richard Guenther <rguenther at suse dot de>
- Cc: gcc-patches at gcc dot gnu dot org
- Date: Sun, 31 Jan 2010 17:03:08 -0500
- Subject: Re: [PATCH] Fix PR42871, recursion in phi_translate
- References: <alpine.LNX.2.00.1001271815570.12128@zhemvz.fhfr.qr>
On Wed, Jan 27, 2010 at 12:19 PM, Richard Guenther <rguenther@suse.de> wrote:
>
> We end up recursing in phi-translate because we have something
> like a[i_1] in the set we want to translate but not i_1 and
> we end up using a[i_1] as leader for i_1 (where both a[i_1]
> and i_1 have the same value-id).
??
Why do they have the same value id?
Unless their is i_1 = a[i_1]?
If so, yeah, we should always prefer NAME over non-NAME.
This gets fixed in the SCCVN replacement patch I sent you by doing
just that when figuring out leaders, so i don't see an issue doing the
same here :)