https://gcc.gnu.org/bugzilla/show_bug.cgi?id=97350
--- Comment #12 from Jan Hubicka <hubicka at gcc dot gnu.org> ---
Aha, the code in question is:
# USE = nonlocal null { D.8330 D.22051 D.22054 D.22059 D.22060 } (nonlocal,
escaped, interposable)
# CLB = nonlocal null { D.8330 D.22051 D.22054 D.22059 D.22060 } (nonlocal,
escaped, interposable)
find_placeholder_in_expr (size_8(D), &self_refs);
# PT = nonlocal escaped null
_30 = self_refs.m_vec;
if (_30 != 0B)
goto <bb 10>; [100.00%]
else
goto <bb 11>; [0.00%]
<bb 10> [count: 7690]:
_31 = MEM[(const struct vec *)_30].m_vecpfx.m_num;
if (_31 == 0)
goto <bb 11>; [0.00%]
else
goto <bb 12>; [100.00%]
What we seem to optimize out is the to m_vec, here alias set 12 makes more
sense.
and indeed it seems that this is missing in the summary. Smells like a bug in
ipa_merge_modref_summary_after_inlining since the function is split and
re-merged by inliner.