For 3 iterations of
unsigned char flagbits;
_877 = flagbits_832 + 254;
_879 = (int) _877;
# prephitmp_880 = PHI <_879(40), 6(41)>
_70 = _68 >> prephitmp_880;
The peeled converted IV handling added in r16-3562 incorrectly analyzes
it as [6, 6 + 254, 6 + 254 * 2] instead of [6, 4, 2]. Then VRP uses the
intersect of {6, 560, 514} and {2, 4, 6}, i.e. {6} as the possible value
range, and propagates the constant 6 for _70.
Extend the step (for example, 254 => -2) to fix the issue.
PR tree-optimization/125291
gcc/
* tree-scalar-evolution.cc (simplify_peeled_chrec): Sign-extend
the step for peeled converted IV.
gcc/testsuite/
* gcc.c-torture/execute/pr125291.c: New test.
---
Bootstrapped and regtested on x86_64-linux-gnu and
loongarch64-linux-gnu. Ok for trunk?
.../gcc.c-torture/execute/pr125291.c | 39 +++++++++++++++++++
gcc/tree-scalar-evolution.cc | 17 ++++++--
2 files changed, 52 insertions(+), 4 deletions(-)
create mode 100644 gcc/testsuite/gcc.c-torture/execute/pr125291.c
diff --git a/gcc/testsuite/gcc.c-torture/execute/pr125291.c
b/gcc/testsuite/gcc.c-torture/execute/pr125291.c
new file mode 100644
index 00000000000..073866a9b73
--- /dev/null
+++ b/gcc/testsuite/gcc.c-torture/execute/pr125291.c
@@ -0,0 +1,39 @@
+/* PR tree-optimization/125291 */
+
+char buf[1111];
+char *archive_le16dec_filename = buf;
+unsigned int archive_le16dec_end, archive_le16dec_fn_end,
+ archive_le16dec_filename_size, archive_le16dec_offset;
+char archive_le16dec_p[] = { 21, 0x7f };
+
+[[gnu::noipa]]
+void
+archive_le16dec ()
+{
+ archive_le16dec_filename_size = (short)archive_le16dec_filename_size;
+ unsigned char flagbits = 0, flagbyte;
+ archive_le16dec_end = archive_le16dec_filename_size;
+ archive_le16dec_fn_end = archive_le16dec_filename_size * 2;
+ archive_le16dec_filename_size = flagbits = 0;
+ while (archive_le16dec_offset < archive_le16dec_end
+ && archive_le16dec_filename_size < archive_le16dec_fn_end)
+ {
+ if (!flagbits)
+ {
+ flagbyte = archive_le16dec_p[archive_le16dec_offset++];
+ flagbits = 8;
+ }
+ flagbits -= 2;
+ if (!(flagbyte >> flagbits & 3))
+ archive_le16dec_filename_size++;
+ }
+}
+
+int
+main ()
+{
+ archive_le16dec_filename_size = 2;
+ archive_le16dec ();
+ if (archive_le16dec_filename_size != 1)
+ __builtin_trap ();
+}
diff --git a/gcc/tree-scalar-evolution.cc b/gcc/tree-scalar-evolution.cc
index f524786f33b..b27037cb02f 100644
--- a/gcc/tree-scalar-evolution.cc
+++ b/gcc/tree-scalar-evolution.cc
@@ -1391,10 +1391,19 @@ simplify_peeled_chrec (class loop *loop, tree arg, tree
init_cond)
&& wi::to_widest (init_cond) == wi::to_widest (left_before)
&& !scev_probably_wraps_p (NULL_TREE, left_before, right, NULL,
loop, false))
- return build_polynomial_chrec (loop->num, init_cond,
- chrec_convert (TREE_TYPE (ev),
- right, NULL,
- false, NULL_TREE));
+ {
+ tree tp = TREE_TYPE (right);
+
+ /* We need a sign-extension to make things like
+ u8(6, 4, 2) => i32(6, 4, 2), instead of i32(6, 260, 514). */
+ if (TYPE_UNSIGNED (tp))
+ right = fold_convert (signed_type_for (tp), right);
+
+ return build_polynomial_chrec (loop->num, init_cond,
+ chrec_convert (TREE_TYPE (ev),
+ right, NULL,
+ false, NULL_TREE));
+ }
return chrec_dont_know;
}
--
2.54.0
