Re: [PATCH] Fix unnecessary register spill that depends on function ordering

2020-07-09 Thread Richard Sandiford
Sorry for the slow reply. Omar Tahir writes: >> Omar Tahir writes: >> > Hi Richard, >> > >> > From: Richard Sandiford >> >> > @@ -3719,6 +3722,7 @@ static unsigned int rest_of_handle_sched (void) >> >> > { #ifdef INSN_SCHEDULING >> >> > + first_moveable_pseudo = last_moveable_pseudo; >> >> >

RE: [PATCH] Fix unnecessary register spill that depends on function ordering

2020-07-02 Thread Omar Tahir
> Omar Tahir writes: > > Hi Richard, > > > > From: Richard Sandiford > >> > @@ -3719,6 +3722,7 @@ static unsigned int rest_of_handle_sched (void) > >> > { #ifdef INSN_SCHEDULING > >> > + first_moveable_pseudo = last_moveable_pseudo; > >> >if (flag_selective_scheduling > >> >&& ! may

Re: [PATCH] Fix unnecessary register spill that depends on function ordering

2020-07-01 Thread Richard Sandiford
Omar Tahir writes: > Hi Richard, > > From: Richard Sandiford >> > @@ -3719,6 +3722,7 @@ static unsigned int rest_of_handle_sched (void) >> > { #ifdef INSN_SCHEDULING >> > + first_moveable_pseudo = last_moveable_pseudo; >> >if (flag_selective_scheduling >> >&& ! maybe_skip_selective_

RE: [PATCH] Fix unnecessary register spill that depends on function ordering

2020-07-01 Thread Omar Tahir
Hi Richard, From: Richard Sandiford > > @@ -3719,6 +3722,7 @@ static unsigned int rest_of_handle_sched (void) > > { #ifdef INSN_SCHEDULING > > + first_moveable_pseudo = last_moveable_pseudo; > >if (flag_selective_scheduling > >&& ! maybe_skip_selective_scheduling ()) > > run_se

Re: [PATCH] Fix unnecessary register spill that depends on function ordering

2020-06-30 Thread Richard Sandiford
Hi, Thanks for the patch. Omar Tahir writes: > Hi, > > The variables first_moveable_pseudo and last_moveable_pseudo aren't reset > after compiling a function, which means they leak into the first scheduler > pass of the following function. In some cases, this can cause an extra spill > during re