https://gcc.gnu.org/bugzilla/show_bug.cgi?id=81848
--- Comment #2 from rguenther at suse dot de <rguenther at suse dot de> --- On Mon, 14 Aug 2017, jamborm at gcc dot gnu.org wrote: > https://gcc.gnu.org/bugzilla/show_bug.cgi?id=81848 > > Martin Jambor <jamborm at gcc dot gnu.org> changed: > > What |Removed |Added > ---------------------------------------------------------------------------- > Target| |x86_64-linux > CC| |rguenth at gcc dot gnu.org, > | |wschmidt at gcc dot gnu.org > Known to work| |7.1.0 > Host| |x86_64-linux > Known to fail| |8.0 > > --- Comment #1 from Martin Jambor <jamborm at gcc dot gnu.org> --- > Apparently, a phi statement is not present in a hash_map where expect > it is: > > Program received signal SIGSEGV, Segmentation fault. > 0x000000000190bdf4 in create_phi_basis (c=0x2921a80, from_phi=0x7ffff666c800, > basis_name=<ssa_name 0x7ffff669b240>, loc=2147483706, known_stride=false) > at /home/mjambor/gcc/bisect/src/gcc/gimple-ssa-strength-reduction.c:2356 > 2356 slsr_cand_t phi_cand = *stmt_cand_map->get (from_phi); > (gdb) p stmt_cand_map->get (from_phi) > $1 = (slsr_cand_d **) 0x0 > > > CCing Richi who committed the revision uncovering this and Bill who > added the lookup last year. Isn't this a dup of PR81354 and thus fixed?
