This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [tree-ssa] Enhance statement checking & fix minor buglets
- From: Diego Novillo <dnovillo at redhat dot com>
- To: Richard Henderson <rth at redhat dot com>
- Cc: Jeff Law <law at redhat dot com>, "gcc-patches at gcc dot gnu dot org" <gcc-patches at gcc dot gnu dot org>, Andrew Pinski <pinskia at physics dot uc dot edu>
- Date: Mon, 05 Apr 2004 13:49:42 -0400
- Subject: Re: [tree-ssa] Enhance statement checking & fix minor buglets
- Organization: Red Hat Canada
- References: <200402280103.i1S13RnA030084@speedy.slc.redhat.com> <1081034160.3172.34.camel@localhost.localdomain> <20040405174052.GA26984@redhat.com>
On Mon, 2004-04-05 at 13:40, Richard Henderson wrote:
> I couldn't tell from your message what's breaking.
>
t.1_3 = &a - 24;
t.2_5 = p_3 + t.1_3;
Since &a - 24 is considered a valid gimple constant, we propagate it
into the use of t.1_3 getting the non-gimple code:
t.2_5 = p_3 + &a - 24;
the tree MINUS_EXPR <PLUS_EXPR, CONST_INT> is not a gimple expression.
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=14672
The quickest fix right now is to disallow
MINUS_EXPR<ADDR_EXPR,CONST_INT> in is_gimple_min_invariant. Unless
supporting it in fold_stmt was really trivial.
Diego.