https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65277
--- Comment #4 from Jan Hubicka <hubicka at gcc dot gnu.org> ---
get_untransformed_body should turn itself into noop in !in_lto_p compilation as
all function bodies have DECL_RESULT. (it probably deserves comment, but that
way we check if the body was read)
Formatting is off:
for (i = new_order_pos - 1; i >= 0; i--)
{
node = order[i];
if (node->process)
{
expanded_func_count++;
if(node->tp_first_run)
profiled_func_count++;
if (symtab->dump_file)
fprintf (symtab->dump_file,
"Time profile order in expand_all_functions:%s:%d\n",
node->asm_name (), node->tp_first_run);
node->process = 0;
node->expand ();
}
}
but actual code makes sense. Node in question is:
_ZTch0_v0_n40_N3Ent4ThisEv/7 (virtual Ent& Ent::_ZTch0_v0_n40_N3Ent4ThisEv())
@0x7ffff6c5d620
Type: function
Body removed by symtab_remove_unreachable_nodes
Visibility: externally_visible public weak comdat comdat_group:_ZN3Ent4ThisEv
one_only section:.text._ZN3Ent4ThisEv (implicit_section) virtual artificial
Same comdat group as: _ZN3Ent4ThisEv/5
References:
Referring:
Availability: not_available
First run: 0
Function flags: icf_merged
Called by:
Calls:
it was removed, so it is wrong that we try to output it.
The removal happens in ipa-icf and it is based on
cgraph_can_remove_if_no_direct_calls_p being wrong - if symbol is in comdat
group removing all direct calls to it may not make it removable if other
symbols hold it around... Will fix that.
