Re: [cft] subreg validation round 2

[Richard Henderson <rth@redhat.com>]

On Fri, Nov 19, 2004 at 12:22:28PM -0800, Richard Henderson wrote:
> I think you have a point.  I'll revert that part of the patch
> and fix this a different way.  And update the documentation.

I never did find why I made this change.  The only thing I can
think is that I was modifying something nearby and thought "this
obviously cannot be correct."

So I've applied this reversion.


r~


        * function.c (assign_parm_setup_block): Revert 11-13 change to
        forcably re-mode a (possibly BLKmode) hard register.

Index: function.c
===================================================================
RCS file: /cvs/gcc/gcc/gcc/function.c,v
retrieving revision 1.588
diff -c -p -d -u -r1.588 function.c
--- function.c	14 Nov 2004 06:22:56 -0000	1.588
+++ function.c	22 Nov 2004 18:52:41 -0000
@@ -2663,7 +2663,7 @@ assign_parm_setup_block (struct assign_p
 	    {
 	      rtx tem, x;
 	      int by = (UNITS_PER_WORD - size) * BITS_PER_UNIT;
-	      rtx reg = gen_lowpart (word_mode, entry_parm);
+	      rtx reg = gen_rtx_REG (word_mode, REGNO (entry_parm));
 
 	      x = expand_shift (LSHIFT_EXPR, word_mode, reg,
 				build_int_cst (NULL_TREE, by),