https://gcc.gnu.org/g:df3994a1be3565ad450d552dc94b696748a7807a
commit r13-8447-gdf3994a1be3565ad450d552dc94b696748a7807a
Author: Jakub Jelinek <ja...@redhat.com>
Date: Tue Mar 5 10:32:38 2024 +0100
lower-subreg: Fix ROTATE handling [PR114211]
On the following testcase, we have
(insn 10 7 11 2 (set (reg/v:TI 106 [ h ])
(rotate:TI (reg/v:TI 106 [ h ])
(const_int 64 [0x40]))) "pr114211.c":8:5 1042
{rotl64ti2_doubleword}
(nil))
before subreg1 and the pass decides to use
(reg:DI 127 [ h ]) / (reg:DI 128 [ h+8 ])
register pair instead of (reg/v:TI 106 [ h ]).
resolve_operand_for_swap_move_operator implements it by pretending it is
an assignment from
(concatn (reg:DI 127 [ h ]) (reg:DI 128 [ h+8 ]))
to
(concatn (reg:DI 128 [ h+8 ]) (reg:DI 127 [ h ]))
The problem is that if the rotate argument is the same as destination or
if there is even an overlap between the first half of the destination with
second half of the source we emit incorrect code, because the store to
(reg:DI 128 [ h+8 ]) overwrites what we need for source of the second
move. The following patch detects that case and uses a temporary pseudo
to hold the original (reg:DI 128 [ h+8 ]) value across the first store.
2024-03-05 Jakub Jelinek <ja...@redhat.com>
PR rtl-optimization/114211
* lower-subreg.cc (resolve_simple_move): For double-word
rotates by BITS_PER_WORD if there is overlap between source
and destination use a temporary.
* gcc.dg/pr114211.c: New test.
(cherry picked from commit aed445b0fd0c7ed16124c61e7eb732992426f103)
Diff:
---
gcc/lower-subreg.cc | 15 +++++++++++++++
gcc/testsuite/gcc.dg/pr114211.c | 23 +++++++++++++++++++++++
2 files changed, 38 insertions(+)
diff --git a/gcc/lower-subreg.cc b/gcc/lower-subreg.cc
index 481e1e85a26..98b18e27029 100644
--- a/gcc/lower-subreg.cc
+++ b/gcc/lower-subreg.cc
@@ -926,6 +926,21 @@ resolve_simple_move (rtx set, rtx_insn *insn)
SRC's operator. */
dest = resolve_operand_for_swap_move_operator (dest);
src = src_op;
+ if (resolve_reg_p (src))
+ {
+ gcc_assert (GET_CODE (src) == CONCATN);
+ if (reg_overlap_mentioned_p (XVECEXP (dest, 0, 0),
+ XVECEXP (src, 0, 1)))
+ {
+ /* If there is overlap between the first half of the
+ destination and what will be stored to the second one,
+ use a temporary pseudo. See PR114211. */
+ rtx tem = gen_reg_rtx (GET_MODE (XVECEXP (src, 0, 1)));
+ emit_move_insn (tem, XVECEXP (src, 0, 1));
+ src = copy_rtx (src);
+ XVECEXP (src, 0, 1) = tem;
+ }
+ }
}
else if (resolve_reg_p (src_op))
{
diff --git a/gcc/testsuite/gcc.dg/pr114211.c b/gcc/testsuite/gcc.dg/pr114211.c
new file mode 100644
index 00000000000..691dae5dbe8
--- /dev/null
+++ b/gcc/testsuite/gcc.dg/pr114211.c
@@ -0,0 +1,23 @@
+/* PR rtl-optimization/114211 */
+/* { dg-do run { target int128 } } */
+/* { dg-options "-O -fno-tree-coalesce-vars -Wno-psabi" } */
+
+typedef unsigned __int128 V __attribute__((__vector_size__ (16)));
+unsigned int u;
+V v;
+
+V
+foo (unsigned __int128 h)
+{
+ h = h << 64 | h >> 64;
+ h *= ~u;
+ return h + v;
+}
+
+int
+main ()
+{
+ V x = foo (1);
+ if (x[0] != (unsigned __int128) 0xffffffff << 64)
+ __builtin_abort ();
+}
