This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [cfg-branch] liveness versus REG_EQUAL notes
- From: Graham Stott <grahams at redhat dot com>
- To: Jan Hubicka <jh at suse dot cz>
- Cc: gcc-patches at gcc dot gnu dot org, rth at cygnus dot com,gcc-pdo at atrey dot karlin dot mff dot cuni dot cz
- Date: Thu, 17 Jan 2002 12:48:48 +0000
- Subject: Re: [cfg-branch] liveness versus REG_EQUAL notes
- References: <20020117124126.GG13326@atrey.karlin.mff.cuni.cz>
Jan,
Jan Hubicka wrote:
>
> Hi,
> this patch makes flow analysis to optionally mark registers used by REG_EQUAL notes
> as live. THis makes the info valid for our RTL representation used in CSE.
> I can now do the dead code removal and friends safely.
>
> Honza
>
> Thu Jan 17 13:39:26 CET 2002 Jan Hubicka <jh@suse.cz>
> * basic-block.h (PROP_EQUAL_NOTES): New flag.
> * flow.c (propagate_one_insn): Use it.
> (mark_used_regs): Handle NIL.
> Index: basic-block.h
> ===================================================================
> RCS file: /cvs/gcc/egcs/gcc/basic-block.h,v
> retrieving revision 1.127.2.11
> diff -c -3 -p -r1.127.2.11 basic-block.h
> *** basic-block.h 2002/01/16 18:19:56 1.127.2.11
> --- basic-block.h 2002/01/17 12:31:42
> *************** enum update_life_extent
> *** 573,578 ****
> --- 573,579 ----
> #define PROP_ALLOW_CFG_CHANGES 32 /* Allow the CFG to be changed
> by dead code removal. */
> #define PROP_AUTOINC 64 /* Create autoinc mem references. */
> + #define PROP_EQUAL_NOTES 128 /* Take into account REG_EQUAL notes. */
> #define PROP_FINAL 127 /* All of the above. */
Umm ^This comment is no longer true!
If PROP_FINAL should include PROP_EQUAL_NOTES then update PROP_FINAL otherwise
move PROP_EQUAL_NOTES after PROP_FINAL so that the comment remains true.
Thinking about it is there any reason why we don't explicitly names the flags
included in PROP_FINAL ?
Graham