https://gcc.gnu.org/bugzilla/show_bug.cgi?id=79197
Jakub Jelinek <jakub at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|NEW |ASSIGNED
Assignee|unassigned at gcc dot gnu.org |jakub at gcc dot gnu.org
--- Comment #7 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Created attachment 40622
--> https://gcc.gnu.org/bugzilla/attachment.cgi?id=40622&action=edit
gcc7-pr79197.patch
So, looking at tmp-mddump.md, there is just a single define_insn that can match
(unsigned_fix:DI (operand:{SF,DF})), which is the one right below the expander
(others are either for SImode instead of DI, or for KF/TF instead of SF/DF).
So, IMNSHO the expander should then use the same predicates and same condition
as the define_insn, at which point it really doesn't make sense to have a
separate expander for it.
