[gcc(refs/users/marxin/heads/marxin-gcc-benchmark-branch)] combine: Punt on out of range rotate counts [PR93505]
Martin Liska
marxin@gcc.gnu.org
Mon Mar 30 10:39:51 GMT 2020
https://gcc.gnu.org/g:56b92750f83724177d2c6eae30c208e935a56a37
commit 56b92750f83724177d2c6eae30c208e935a56a37
Author: Jakub Jelinek <jakub@redhat.com>
Date: Thu Jan 30 21:28:17 2020 +0100
combine: Punt on out of range rotate counts [PR93505]
What happens on this testcase is with the out of bounds rotate we get:
Trying 13 -> 16:
13: r129:SI=r132:DI#0<-<0x20
REG_DEAD r132:DI
16: r123:DI=r129:SI<0
REG_DEAD r129:SI
Successfully matched this instruction:
(set (reg/v:DI 123 [ <retval> ])
(const_int 0 [0]))
during combine. So, perhaps we could also change simplify-rtx.c to punt
if it is out of bounds rather than trying to optimize anything.
Or, but probably GCC11 material, if we decide that ROTATE/ROTATERT doesn't
have out of bounds counts or introduce targetm.rotate_truncation_mask,
we should truncate the argument instead of punting.
Punting is better for backports though.
2020-01-30 Jakub Jelinek <jakub@redhat.com>
PR middle-end/93505
* combine.c (simplify_comparison) <case ROTATE>: Punt on out of range
rotate counts.
* gcc.c-torture/compile/pr93505.c: New test.
Diff:
---
gcc/ChangeLog | 6 ++++++
gcc/combine.c | 3 ++-
gcc/testsuite/ChangeLog | 5 +++++
gcc/testsuite/gcc.c-torture/compile/pr93505.c | 15 +++++++++++++++
4 files changed, 28 insertions(+), 1 deletion(-)
diff --git a/gcc/ChangeLog b/gcc/ChangeLog
index 7132bf00d0e..f2cc0276600 100644
--- a/gcc/ChangeLog
+++ b/gcc/ChangeLog
@@ -1,3 +1,9 @@
+2020-01-30 Jakub Jelinek <jakub@redhat.com>
+
+ PR middle-end/93505
+ * combine.c (simplify_comparison) <case ROTATE>: Punt on out of range
+ rotate counts.
+
2020-01-30 Andrew Stubbs <ams@codesourcery.com>
* config/gcn/gcn.c (print_operand): Handle LTGT.
diff --git a/gcc/combine.c b/gcc/combine.c
index 0272e754297..d44b9c3bf95 100644
--- a/gcc/combine.c
+++ b/gcc/combine.c
@@ -12410,7 +12410,8 @@ simplify_comparison (enum rtx_code code, rtx *pop0, rtx *pop1)
bit. This will be converted into a ZERO_EXTRACT. */
if (const_op == 0 && sign_bit_comparison_p
&& CONST_INT_P (XEXP (op0, 1))
- && mode_width <= HOST_BITS_PER_WIDE_INT)
+ && mode_width <= HOST_BITS_PER_WIDE_INT
+ && UINTVAL (XEXP (op0, 1)) < mode_width)
{
op0 = simplify_and_const_int (NULL_RTX, mode, XEXP (op0, 0),
(HOST_WIDE_INT_1U
diff --git a/gcc/testsuite/ChangeLog b/gcc/testsuite/ChangeLog
index 269adb29e94..ce53e2c8bf8 100644
--- a/gcc/testsuite/ChangeLog
+++ b/gcc/testsuite/ChangeLog
@@ -1,3 +1,8 @@
+2020-01-30 Jakub Jelinek <jakub@redhat.com>
+
+ PR middle-end/93505
+ * gcc.c-torture/compile/pr93505.c: New test.
+
2020-01-30 Jeff Law <law@redhat.com
* gcc.dg/tree-ssa/ssa-dse-26.c: Make existing dg-final scan
diff --git a/gcc/testsuite/gcc.c-torture/compile/pr93505.c b/gcc/testsuite/gcc.c-torture/compile/pr93505.c
new file mode 100644
index 00000000000..0627962eae5
--- /dev/null
+++ b/gcc/testsuite/gcc.c-torture/compile/pr93505.c
@@ -0,0 +1,15 @@
+/* PR middle-end/93505 */
+
+unsigned a;
+
+unsigned
+foo (unsigned x)
+{
+ unsigned int y = 32 - __builtin_bswap64 (-a);
+ /* This would be UB (x << 32) at runtime. Ensure we don't
+ invoke UB in the compiler because of that (visible with
+ bootstrap-ubsan). */
+ x = x << y | x >> (-y & 31);
+ x >>= 31;
+ return x;
+}
More information about the Gcc-cvs
mailing list