This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
PATCH: [4.5 Regression] ICE with graphite enabled in cairo-1.8.10
- From: "H.J. Lu" <hongjiu dot lu at intel dot com>
- To: gcc-patches at gcc dot gnu dot org
- Date: Mon, 29 Nov 2010 11:24:46 -0800
- Subject: PATCH: [4.5 Regression] ICE with graphite enabled in cairo-1.8.10
- Reply-to: "H.J. Lu" <hjl dot tools at gmail dot com>
Hi,
This patch fixes a 4.5 regression. OK for 4.5?
Thanks.
H.J.
---
gcc/
2010-11-29 H.J. Lu <hongjiu.lu@intel.com>
Backport from mainline
PR middle-end/46651
2010-07-22 Sebastian Pop <sebastian.pop@amd.com>
* graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Use
SSA_NAME_DEF_STMT only on SSA_NAMEs.
gcc/testsuite/
2010-11-29 H.J. Lu <hongjiu.lu@intel.com>
Backport from mainline
2010-07-22 Sebastian Pop <sebastian.pop@amd.com>
PR middle-end/46651
* gcc.dg/graphite/id-24.c: New.
diff --git a/gcc/graphite-sese-to-poly.c b/gcc/graphite-sese-to-poly.c
index b12210b..b8f332a 100644
--- a/gcc/graphite-sese-to-poly.c
+++ b/gcc/graphite-sese-to-poly.c
@@ -2235,7 +2235,8 @@ rewrite_phi_out_of_ssa (gimple_stmt_iterator *psi)
/* Avoid the insertion of code in the loop latch to please the
pattern matching of the vectorizer. */
- if (e->src == bb->loop_father->latch)
+ if (TREE_CODE (arg) == SSA_NAME
+ && e->src == bb->loop_father->latch)
insert_out_of_ssa_copy (zero_dim_array, arg);
else
insert_out_of_ssa_copy_on_edge (e, zero_dim_array, arg);
diff --git a/gcc/testsuite/gcc.dg/graphite/id-24.c b/gcc/testsuite/gcc.dg/graphite/id-24.c
new file mode 100644
index 0000000..d466069
--- /dev/null
+++ b/gcc/testsuite/gcc.dg/graphite/id-24.c
@@ -0,0 +1,22 @@
+/* gcc.dg/tree-ssa/loadpre23.c used to ICE with Graphite. */
+
+struct {
+ int a;
+ int large[100];
+} x;
+
+int foo(int argc)
+{
+ int b;
+ int c;
+ int i;
+ int d, e;
+
+ for (i = 0; i < argc; i++)
+ {
+ e = x.a;
+ x.a = 9;
+ }
+ return d + e;
+}
+