This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
[PATCH] Fix PR81373
- From: Richard Biener <rguenther at suse dot de>
- To: gcc-patches at gcc dot gnu dot org
- Cc: sebpop at gmail dot com
- Date: Tue, 19 Sep 2017 16:17:32 +0200 (CEST)
- Subject: [PATCH] Fix PR81373
- Authentication-results: sourceware.org; auth=none
The following forces scev analyzable SESE liveouts to be handed as
defs. Otherwise we end up forgetting to code-generate them. We
naturally expect SCEV-cprop to handle them but that pass has cost
cutoffs (and one can disable it).
Bootstrap and regtest running on x86_64-unknown-linux-gnu, ok?
Thanks,
Richard.
2017-09-19 Richard Biener <rguenther@suse.de>
PR tree-optimization/81373
* graphite-scop-detection.c (build_cross_bb_scalars_def):
Force SESE live-out defs to be handled even if they are
scev_analyzable_p.
* gcc.dg/graphite/pr81373.c: New testcase.
Index: gcc/graphite-scop-detection.c
===================================================================
--- gcc/graphite-scop-detection.c (revision 252968)
+++ gcc/graphite-scop-detection.c (working copy)
@@ -1717,17 +1717,20 @@ build_cross_bb_scalars_def (scop_p scop,
if (!def || !is_gimple_reg (def))
return;
- /* Do not gather scalar variables that can be analyzed by SCEV as they can be
- generated out of the induction variables. */
- if (scev_analyzable_p (def, scop->scop_info->region))
- return;
+ bool scev_analyzable = scev_analyzable_p (def, scop->scop_info->region);
gimple *use_stmt;
imm_use_iterator imm_iter;
FOR_EACH_IMM_USE_STMT (use_stmt, imm_iter, def)
- if ((def_bb != gimple_bb (use_stmt) && !is_gimple_debug (use_stmt))
- /* PHIs have their effect at "BBs" on the edges. See PR79622. */
- || gimple_code (SSA_NAME_DEF_STMT (def)) == GIMPLE_PHI)
+ /* Do not gather scalar variables that can be analyzed by SCEV as they can
+ be generated out of the induction variables. */
+ if ((! scev_analyzable
+ /* But gather SESE liveouts as we otherwise fail to rewrite their
+ exit PHIs. */
+ || ! bb_in_sese_p (gimple_bb (use_stmt), scop->scop_info->region))
+ && ((def_bb != gimple_bb (use_stmt) && !is_gimple_debug (use_stmt))
+ /* PHIs have their effect at "BBs" on the edges. See PR79622. */
+ || gimple_code (SSA_NAME_DEF_STMT (def)) == GIMPLE_PHI))
{
writes->safe_push (def);
DEBUG_PRINT (dp << "Adding scalar write: ";
Index: gcc/testsuite/gcc.dg/graphite/pr81373.c
===================================================================
--- gcc/testsuite/gcc.dg/graphite/pr81373.c (nonexistent)
+++ gcc/testsuite/gcc.dg/graphite/pr81373.c (working copy)
@@ -0,0 +1,40 @@
+/* { dg-options "-fno-tree-scev-cprop -fgraphite-identity -O -fdump-tree-graphite-all" } */
+
+void bar (void);
+
+int toto()
+{
+ int i, j, k;
+ int a[101][100];
+ int b[100];
+
+ for (i = 1; i < 100; i++)
+ {
+ for (j = 1; j < 100; j++)
+ for (k = 1; k < 100; k++)
+ a[j][k] = a[j+1][i-1] + 2;
+
+ b[i] = b[i-1] + 2;
+
+ bar ();
+
+ for (j = 1; j < 100; j++)
+ a[j][i] = a[j+1][i-1] + 2;
+
+ b[i] = b[i-1] + 2;
+
+ bar ();
+
+ for (j = 1; j < 100; j++)
+ a[j][i] = a[j+1][i-1] + 2;
+
+ b[i] = a[i-1][i] + 2;
+
+ for (j = 1; j < 100; j++)
+ a[j][i] = a[j+1][i-1] + 2;
+ }
+
+ return a[3][5] + b[1];
+}
+
+/* { dg-final { scan-tree-dump-times "number of SCoPs: 2" 1 "graphite"} } */