This is another problem I found working on a new target. In fold-const,
when folding conversions, there are two instances where mode bitsizes
are compared to type precisions. This fails to do the right thing if
mode precision and bitsize differ (which I believe they currently don't
on any target).
Bootstrapped and regression tested on i686-linux. Ok?
Bernd
* fold-const.c (fold_unary_loc): Use GET_MODE_PRECISION for
comparisons against TYPE_PRECISION.
Index: gcc/fold-const.c
===
--- gcc/fold-const.c(revision 365109)
+++ gcc/fold-const.c(working copy)
@@ -8586,7 +8586,7 @@ fold_unary_loc (location_t loc, enum tre
inter_prec = inside_prec
(inter_float || inter_vec
|| inter_unsignedp == inside_unsignedp)
- ! (final_prec != GET_MODE_BITSIZE (TYPE_MODE (type))
+ ! (final_prec != GET_MODE_PRECISION (TYPE_MODE (type))
TYPE_MODE (type) == TYPE_MODE (inter_type))
! final_ptr
(! final_vec || inter_prec == inside_prec))
@@ -8620,7 +8620,7 @@ fold_unary_loc (location_t loc, enum tre
== (final_unsignedp final_prec inter_prec))
! (inside_ptr inter_prec != final_prec)
! (final_ptr inside_prec != inter_prec)
- ! (final_prec != GET_MODE_BITSIZE (TYPE_MODE (type))
+ ! (final_prec != GET_MODE_PRECISION (TYPE_MODE (type))
TYPE_MODE (type) == TYPE_MODE (inter_type)))
return fold_build1_loc (loc, code, type, TREE_OPERAND (op0, 0));
}