public inbox for gcc-cvs@sourceware.org
help / color / mirror / Atom feed
* [gcc r14-8276] sccvn: Don't use SCALAR_INT_TYPE_MODE on BLKmode BITINT_TYPEs [PR113459]
@ 2024-01-19 9:02 Jakub Jelinek
0 siblings, 0 replies; only message in thread
From: Jakub Jelinek @ 2024-01-19 9:02 UTC (permalink / raw)
To: gcc-cvs
https://gcc.gnu.org/g:dcd5601c2b7298155c9a8e1bfb93ee8e952eca0b
commit r14-8276-gdcd5601c2b7298155c9a8e1bfb93ee8e952eca0b
Author: Jakub Jelinek <jakub@redhat.com>
Date: Fri Jan 19 10:00:16 2024 +0100
sccvn: Don't use SCALAR_INT_TYPE_MODE on BLKmode BITINT_TYPEs [PR113459]
sccvn uses GET_MODE_SIZE (SCALAR_INT_TYPE_MODE (type)) for INTEGER_TYPEs,
most likely because that is what native_{interpret,encode}_int used.
This obviously doesn't work for larger BITINT_TYPEs which have BLKmode
and the above ICEs on those. native_{interpret,encode}_int checks whether
the BITINT_TYPE is medium/large/huge (i.e. an array of 2+ ABI limbs)
and uses TYPE_SIZE_UNIT for that case, otherwise SCALAR_INT_TYPE_MODE like
for the INTEGER_TYPE case.
The following patch instead just uses SCALAR_INT_TYPE_MODE for non-BLKmode
TYPE_MODE and TYPE_SIZE_UNIT otherwise.
2024-01-19 Jakub Jelinek <jakub@redhat.com>
PR tree-optimization/113459
* tree-ssa-sccvn.cc (vn_walk_cb_data::push_partial_def): Use
TREE_INT_CST_LOW of TYPE_SIZE_UNIT rather than GET_MODE_SIZE
of SCALAR_INT_TYPE_MODE if type has BLKmode.
(vn_reference_lookup_3): Likewise. Formatting fix.
* gcc.dg/bitint-73.c: New test.
Diff:
---
gcc/testsuite/gcc.dg/bitint-73.c | 18 ++++++++++++++++++
gcc/tree-ssa-sccvn.cc | 20 ++++++++++++++++----
2 files changed, 34 insertions(+), 4 deletions(-)
diff --git a/gcc/testsuite/gcc.dg/bitint-73.c b/gcc/testsuite/gcc.dg/bitint-73.c
new file mode 100644
index 00000000000..d796134cc37
--- /dev/null
+++ b/gcc/testsuite/gcc.dg/bitint-73.c
@@ -0,0 +1,18 @@
+/* PR tree-optimization/113459 */
+/* { dg-do compile { target bitint } } */
+/* { dg-options "-std=c23 -O2" } */
+
+#if __BITINT_MAXWIDTH__ >= 129
+# define N 129
+#else
+# define N 63
+#endif
+
+_BitInt(N) a;
+
+_BitInt(N)
+foo (void)
+{
+ __builtin_memset (&a, 6, sizeof a);
+ return a;
+}
diff --git a/gcc/tree-ssa-sccvn.cc b/gcc/tree-ssa-sccvn.cc
index 97f5c7923db..f0fa718a723 100644
--- a/gcc/tree-ssa-sccvn.cc
+++ b/gcc/tree-ssa-sccvn.cc
@@ -2287,7 +2287,12 @@ vn_walk_cb_data::push_partial_def (pd_data pd,
BITS_PER_UNIT
- (maxsizei % BITS_PER_UNIT));
if (INTEGRAL_TYPE_P (type))
- sz = GET_MODE_SIZE (SCALAR_INT_TYPE_MODE (type));
+ {
+ if (TYPE_MODE (type) != BLKmode)
+ sz = GET_MODE_SIZE (SCALAR_INT_TYPE_MODE (type));
+ else
+ sz = TREE_INT_CST_LOW (TYPE_SIZE_UNIT (type));
+ }
if (sz > needed_len)
{
memcpy (this_buffer + (sz - needed_len), buffer, needed_len);
@@ -2967,8 +2972,10 @@ vn_reference_lookup_3 (ao_ref *ref, tree vuse, void *data_,
}
else
{
- unsigned buflen = TREE_INT_CST_LOW (TYPE_SIZE_UNIT (vr->type)) + 1;
- if (INTEGRAL_TYPE_P (vr->type))
+ unsigned buflen
+ = TREE_INT_CST_LOW (TYPE_SIZE_UNIT (vr->type)) + 1;
+ if (INTEGRAL_TYPE_P (vr->type)
+ && TYPE_MODE (vr->type) != BLKmode)
buflen = GET_MODE_SIZE (SCALAR_INT_TYPE_MODE (vr->type)) + 1;
unsigned char *buf = XALLOCAVEC (unsigned char, buflen);
memset (buf, TREE_INT_CST_LOW (gimple_call_arg (def_stmt, 1)),
@@ -3165,7 +3172,12 @@ vn_reference_lookup_3 (ao_ref *ref, tree vuse, void *data_,
offset + maxsize - 1. */
HOST_WIDE_INT sz = maxsizei / BITS_PER_UNIT;
if (INTEGRAL_TYPE_P (type))
- sz = GET_MODE_SIZE (SCALAR_INT_TYPE_MODE (type));
+ {
+ if (TYPE_MODE (type) != BLKmode)
+ sz = GET_MODE_SIZE (SCALAR_INT_TYPE_MODE (type));
+ else
+ sz = TREE_INT_CST_LOW (TYPE_SIZE_UNIT (type));
+ }
amnt = ((unsigned HOST_WIDE_INT) offset2i + size2i
- offseti - maxsizei) % BITS_PER_UNIT;
if (amnt)
^ permalink raw reply [flat|nested] only message in thread
only message in thread, other threads:[~2024-01-19 9:02 UTC | newest]
Thread overview: (only message) (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2024-01-19 9:02 [gcc r14-8276] sccvn: Don't use SCALAR_INT_TYPE_MODE on BLKmode BITINT_TYPEs [PR113459] Jakub Jelinek
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).