This is the mail archive of the mailing list for the GCC project.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: Properly handle VIEW_CONVERT_EXPR in LHS (fixes 4.5 ACATS tests)

On Sun, Nov 28, 2004 at 04:40:05PM -0500, Richard Kenner wrote:
> There are two issues here.  First, if a VIEW_CONVERT_EXPR is on the LHS of
> MODIFY_EXPR, you can't automatically assume that this will be a V_MUST_DEF.
> Whether it's V_MUST_DEF or V_MAY_DEF still depends on whether there's a
> component (or array, ...) reference on the LHS.  So we need to strip it off
> before doing the test.

These are two unrelated changes, and therefore should have been 
submitted separately.  This part is fine, but...

> Second, in CCP, when you store a value, if the LHS had a VIEW_CONVERT_EXPR,
> the value being stored is the wrong type and needs to be view-converted to
> the proper type.

This bit is buggy.

> !     if (TREE_CODE (orig_lhs) == VIEW_CONVERT_EXPR
> ! 	&& val.lattice_val == CONSTANT)
> !       {
> ! 	val.const_val = build1 (VIEW_CONVERT_EXPR,
> ! 				TREE_TYPE (TREE_OPERAND (orig_lhs, 0)),
> ! 				val.const_val);

const_val must satisfy is_gimple_min_invariant, because it is
substituted into PHI nodes.  In some cases, you may be able to
construct a correct replacement, but in others it'll be tricky.

I suggest

	w = fold (build1 (VIEW_CONVERT_EXPR, ...))
	if (is_gimple_min_invariant (w))


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]