Re: cleanup alter_subreg; fix simplify_subreg

[Jan Hubicka <jh at suse dot cz>]

>  > > The change
>  > > 
>  > >   Wed Nov 14 14:17:05 CET 2001  Jan Hubicka  <jh@suse.cz>
>  > > 
>  > >         * dbxout.c (dbxout_symbol_location): Update call of
>  > > alter_subreg.
>  > >         * final.c (walk_alter_subreg): Take pointer; update call of
>  > > alter_subreg.
>  > >         (final_scan_insn): Update call of alter_subreg.
>  > >         (cleanup_subreg_operands): Likewise.
>  > >         (alter_subreg): Rewrite using simplify_subreg.
>  > >         (output_address, output_operand): Update call of
>  > > alter_subreg.
>  > >         * output.h (alter_subreg): Update prototype.
>  > >         * sdbout.c (sdbout_symbol): Update call of alter_subreg.
>  > >         * simplify-rtx.c (simplify_subreg): Copy ORIGINAL_REGNO.
>  > >         * Makefile.in (final.o): Add depdendancy on expr.h
>  > > 
>  > > causes a segmentation fault during bootstrap for
>  > > arm-unknown-netbsd1.5Y.
>  > 
>  > Thanks,
>  > I will try to debug this at monday (I can't do much work till then).
>  > The problem appears to be that some ports construct subregs they claim
>  > to be invalid and thus simplify_subreg returns NULL.  I plan to solve
>  > it by adding simplify_subreg_force that will do the simplification
>  > even when the results does not validate.
>  > Honza
> 
> Jan,
> 
> I'm pretty sure your change also broke solaris2 bootstraps.  The sparc
> port has a bunch of calls to alter_subreg and none of them appear to
> have been updated as per your parameter change.  See:
> http://gcc.gnu.org/ml/gcc-bugs/2001-11/msg00519.html
> 
> 
> In fact, there are about 10 port subdirectories which call
> alter_subreg, and none of them seem to have gotten this change
> propagated.  IMHO this patch shouldn't have been checked in without
> catching all callers, "grep" is your friend.
Yes, I did grep only on the base directory.  My apologizes for that - I didn't
expected the alter subreg to be used elsewhere.  IMO the uses should be
slowly converted to simplify_subreg instead.
> 
> When do you expect to address this problem?
I am just working on it - I do have sparc tree bootstrapping again.

Honza
> 
> 		Thanks,
> 		--Kaveh
> --
> Kaveh R. Ghazi			Engagement Manager / Project Services
> ghazi@caip.rutgers.edu		Qwest Internet Solutions