public inbox for gcc-cvs@sourceware.org
help / color / mirror / Atom feed
* [gcc r12-6834] Update the type of control.base after changed
@ 2022-01-24  9:07 Jiu Fu Guo
  0 siblings, 0 replies; only message in thread
From: Jiu Fu Guo @ 2022-01-24  9:07 UTC (permalink / raw)
  To: gcc-cvs

https://gcc.gnu.org/g:634de54f9c421b7069865d0d7365ad97412f34bd

commit r12-6834-g634de54f9c421b7069865d0d7365ad97412f34bd
Author: Jiufu Guo <guojiufu@linux.ibm.com>
Date:   Fri Jan 21 17:03:50 2022 +0800

    Update the type of control.base after changed
    
    This patch correct the type of niter->control.base, when it is updated
    as a PLUS expr.
    During build PLUS expr, the result type should align with the type of
    the operands.
    
            PR tree-optimization/102087
    
    gcc/ChangeLog:
    
            * tree-ssa-loop-niter.cc (number_of_iterations_until_wrap):
            Correct PLUS result type.
    
    gcc/testsuite/ChangeLog:
    
            * gcc.dg/pr102087_1.c: New test.

Diff:
---
 gcc/testsuite/gcc.dg/pr102087_1.c | 13 +++++++++++++
 gcc/tree-ssa-loop-niter.cc        | 17 +++++++++++++++--
 2 files changed, 28 insertions(+), 2 deletions(-)

diff --git a/gcc/testsuite/gcc.dg/pr102087_1.c b/gcc/testsuite/gcc.dg/pr102087_1.c
new file mode 100644
index 00000000000..ba4efe3b412
--- /dev/null
+++ b/gcc/testsuite/gcc.dg/pr102087_1.c
@@ -0,0 +1,13 @@
+/* PR tree-optimization/102087 */
+/* { dg-do compile } */
+/* { dg-options "-O3 -fprefetch-loop-arrays -w" { target x86_64-*-* powerpc*-*-* } } */
+
+char **Gif_ClipImage_gfi_0;
+int Gif_ClipImage_gfi_1, Gif_ClipImage_y, Gif_ClipImage_shift;
+void Gif_ClipImage() {
+  Gif_ClipImage_y = Gif_ClipImage_gfi_1 - 1;
+  for (; Gif_ClipImage_y >= Gif_ClipImage_shift; Gif_ClipImage_y++)
+    Gif_ClipImage_gfi_0[Gif_ClipImage_shift] =
+        Gif_ClipImage_gfi_0[Gif_ClipImage_y];
+}
+
diff --git a/gcc/tree-ssa-loop-niter.cc b/gcc/tree-ssa-loop-niter.cc
index b767056aeb0..21cc257c91b 100644
--- a/gcc/tree-ssa-loop-niter.cc
+++ b/gcc/tree-ssa-loop-niter.cc
@@ -1579,8 +1579,21 @@ number_of_iterations_until_wrap (class loop *loop, tree type, affine_iv *iv0,
      { IVbase - STEP, +, STEP } != bound
      Here, biasing IVbase by 1 step makes 'bound' be the value before wrap.
      */
-  niter->control.base = fold_build2 (MINUS_EXPR, niter_type,
-				     niter->control.base, niter->control.step);
+  tree base_type = TREE_TYPE (niter->control.base);
+  if (POINTER_TYPE_P (base_type))
+    {
+      tree utype = unsigned_type_for (base_type);
+      niter->control.base
+	= fold_build2 (MINUS_EXPR, utype,
+		       fold_convert (utype, niter->control.base),
+		       fold_convert (utype, niter->control.step));
+      niter->control.base = fold_convert (base_type, niter->control.base);
+    }
+  else
+    niter->control.base
+      = fold_build2 (MINUS_EXPR, base_type, niter->control.base,
+		     niter->control.step);
+
   span = fold_build2 (MULT_EXPR, niter_type, niter->niter,
 		      fold_convert (niter_type, niter->control.step));
   niter->bound = fold_build2 (PLUS_EXPR, niter_type, span,


^ permalink raw reply	[flat|nested] only message in thread

only message in thread, other threads:[~2022-01-24  9:07 UTC | newest]

Thread overview: (only message) (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2022-01-24  9:07 [gcc r12-6834] Update the type of control.base after changed Jiu Fu Guo

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).