[Bug ipa/86279] New: [9 Regression] ICE in malloc_candidate_p, at ipa-pure-const.c:927

[marxin at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=86279

            Bug ID: 86279
           Summary: [9 Regression] ICE in malloc_candidate_p, at
                    ipa-pure-const.c:927
           Product: gcc
           Version: unknown
            Status: UNCONFIRMED
          Keywords: ice-on-valid-code
          Severity: normal
          Priority: P3
         Component: ipa
          Assignee: unassigned at gcc dot gnu.org
          Reporter: marxin at gcc dot gnu.org
                CC: hubicka at gcc dot gnu.org, marxin at gcc dot gnu.org
  Target Milestone: ---

Caused by my commit:

$ gcc /home/marxin/Programming/gcc/gcc/testsuite/gcc.dg/pr47991.c
-fipa-pure-const
during IPA pass: pure-const
/home/marxin/Programming/gcc/gcc/testsuite/gcc.dg/pr47991.c: In function
‘memset’:
/home/marxin/Programming/gcc/gcc/testsuite/gcc.dg/pr47991.c:25:1: internal
compiler error: in malloc_candidate_p, at ipa-pure-const.c:927
 }
 ^
0x1b168f5 malloc_candidate_p
        /home/marxin/Programming/gcc/gcc/ipa-pure-const.c:927
0x1b17729 analyze_function
        /home/marxin/Programming/gcc/gcc/ipa-pure-const.c:1096
0x1b1b07c pure_const_generate_summary
        /home/marxin/Programming/gcc/gcc/ipa-pure-const.c:1179
0xe01b4f execute_ipa_summary_passes(ipa_opt_pass_d*)
        /home/marxin/Programming/gcc/gcc/passes.c:2149
0x9c34e7 ipa_passes
        /home/marxin/Programming/gcc/gcc/cgraphunit.c:2428
0x9c34e7 symbol_table::compile()
        /home/marxin/Programming/gcc/gcc/cgraphunit.c:2539
0x9c5ab6 symbol_table::compile()
        /home/marxin/Programming/gcc/gcc/cgraphunit.c:2702
0x9c5ab6 symbol_table::finalize_compilation_unit()
        /home/marxin/Programming/gcc/gcc/cgraphunit.c:2699

I'm wondering whether we should do malloc propagation when optimize == false?
Honza?