This is a small bug in the ext-dce code's handling of promoted subregs.
Essentially when we see a promoted subreg we need to make additional bit groups
live as various parts of the RTL path know that an extension of a suitably
promoted subreg can be trivially eliminated.
When I added support for dealing with this quirk I failed to account for the
larger modes properly and it ignored the case when the size of the inner object
was > 32 bits. Opps.
This does _not_ fix the outstanding x86 issue. That's caused by something
completely different and more concerning ;(
Bootstrapped and regression tested on x86. Obviously fixes the testcase on
riscv as well.
Pushing to the trunk.
PR rtl-optimization/116544
gcc/
* ext-dce.cc (ext_dce_process_uses): Fix thinko in promoted subreg
handling.
gcc/testsuite/
* gcc.dg/torture/pr116544.c: New test.
bitmap_set_bit (livenow, rn + 1);
if (size > 16)
bitmap_set_bit (livenow, rn + 2);
- if (size == 32)
+ if (size >= 32)
bitmap_set_bit (livenow, rn + 3);
iter.skip_subrtxes ();
}
--- /dev/null
+/* { dg-options "-fno-strict-aliasing -fwrapv" }
+/* { dg-do run { target longlong64 } } */
+
+extern void abort (void);
+long long a;
+signed char b[60];
+signed char c;
+long long d[60];
+int e[30];
+long long *f = d;
+static void g(long long *j, long k) { *j = k; }
+int main() {
+ d[5] = 0x100000000;
+ for (int h = 2; h < 7; h += 3)
+ for (int i = 0; i < (c || b[h]) + 10; i += 11)
+ e[2] = f[h];
+ g(&a, e[2]);
+ if (a != 0)
+ abort ();
+ return 0;
+}
+