https://gcc.gnu.org/bugzilla/show_bug.cgi?id=79521
Jakub Jelinek <jakub at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Priority|P3 |P1
Status|UNCONFIRMED |NEW
Last reconfirmed| |2017-02-15
CC| |law at gcc dot gnu.org
Target Milestone|--- |7.0
Ever confirmed|0 |1
--- Comment #1 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
In the IL we have:
(insn 30 13 26 2 (clobber (reg:TI 123)) "rq.i":21 -1
(nil))
(insn 26 30 27 2 (set (subreg:SI (reg:TI 123) 0)
(subreg:SI (reg/v:TF 122 [ p ]) 0)) "rq.i":21 82 {*movsi_internal}
(nil))
(insn 27 26 28 2 (set (subreg:SI (reg:TI 123) 4)
(subreg:SI (reg/v:TF 122 [ p ]) 4)) "rq.i":21 82 {*movsi_internal}
(nil))
(insn 28 27 29 2 (set (subreg:SI (reg:TI 123) 8)
(subreg:SI (reg/v:TF 122 [ p ]) 8)) "rq.i":21 82 {*movsi_internal}
(nil))
(insn 29 28 31 2 (set (subreg:SI (reg:TI 123) 12)
(subreg:SI (reg/v:TF 122 [ p ]) 12)) "rq.i":21 82 {*movsi_internal}
(nil))
(insn 31 29 36 2 (set (reg/v:DI 108 [ hp ])
(subreg:DI (reg:TI 123) 8)) "rq.i":21 81 {*movdi_internal}
(nil))
even when TImode actually isn't supported by the target for arithmetics.
