https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88926
Bug ID: 88926
Summary: ivopts with some NOP conversions
Product: gcc
Version: 9.0
Status: UNCONFIRMED
Keywords: missed-optimization
Severity: enhancement
Priority: P3
Component: tree-optimization
Assignee: unassigned at gcc dot gnu.org
Reporter: glisse at gcc dot gnu.org
Target Milestone: ---
Starting from gcc.dg/tree-ssa/ivopts-lt-2.c:
/* { dg-do compile } */
/* { dg-options "-O2 -fdump-tree-ivopts" } */
/* { dg-skip-if "PR68644" { hppa*-*-* powerpc*-*-* } } */
void
f1 (int *p, unsigned int i)
{
p += i;
do
{
*p = 0;
p += 1;
i++;
}
while (i < 100);
}
/* { dg-final { scan-tree-dump-times "PHI" 1 "ivopts" } } */
/* { dg-final { scan-tree-dump-times "PHI <p_" 1 "ivopts"} } */
/* { dg-final { scan-tree-dump-times "p_\[0-9\]* <" 1 "ivopts" } } */
ivopts normally manages to get rid of i in the loop. However, if I replace "p
+= i" with
__PTRDIFF_TYPE__ o = i;
o *= 4;
p = (char*)p + o;
(i.e. I compute i*4 in type 'long' instead of 'unsigned long')
ivopts doesn't find an invariant and doesn't eliminate i anymore.
This will be useful if I try to resurrect
https://gcc.gnu.org/ml/gcc-patches/2017-05/msg01641.html . I am not familiar
with ivopts (I don't see why the instruction p+=i is relevant, it only changes
the initial value of p, but without it we do not eliminate either) so this is
likely one of the last cases I'll try to fix.
