On Thu, Dec 19, 2019 at 06:23:59PM +0100, Jakub Jelinek wrote:
> On Thu, Dec 19, 2019 at 04:50:40PM +0100, Jan Hubicka wrote:
> > Outputting the move as RIP relative movq would work.
> > LC12 is string "s" and has nothing to do with stack protecting.
>
> This should fix it by doing more carefully what *movdi_internal
> does. Will bootstrap/regtest it tonight.
Passed bootstrap/regtest on x86_64-linux and i686-linux. Ok for trunk?
>
> 2019-12-19 Jakub Jelinek <ja...@redhat.com>
>
> PR target/92841
> * config/i386/i386.md (*stack_protect_set_3): For pic_32bit_operand
> always use lea{q}, no matter what value which_alternative has.
>
> * gcc.target/i386/pr92841-2.c: New test.
Jakub
