Re: [PATCH] tree-optimization/104941: Actually assign the conversion result

[Jakub Jelinek <jakub@redhat.com>]

On Wed, Mar 16, 2022 at 08:24:24PM +0530, Siddhesh Poyarekar wrote:
> Assign the result of fold_convert to offset.
> 
> gcc/ChangeLog:
> 
> 	PR tree-optimization/104941
> 	* tree-object-size.cc (size_for_offset): Assign result of
> 	fold_convert to OFFSET.
> 
> gcc/testsuite/ChangeLog:
> 
> 	PR tree-optimization/104941
> 	* gcc.dg/builtin-dynamic-object-size-0.c (S1, S2): New structs.
> 	(test_alloc_nested_structs, g): New functions.
> 	(main): Call test_alloc_nested_structs.
> 
> Signed-off-by: Siddhesh Poyarekar <siddhesh@gotplt.org>

> --- a/gcc/tree-object-size.cc
> +++ b/gcc/tree-object-size.cc
> @@ -372,7 +372,7 @@ size_for_offset (tree sz, tree offset, tree wholesize = NULL_TREE)
>  
>    /* Safe to convert now, since a valid net offset should be non-negative.  */
>    if (!types_compatible_p (TREE_TYPE (offset), sizetype))
> -    fold_convert (sizetype, offset);
> +    offset = fold_convert (sizetype, offset);
>  
>    if (TREE_CODE (offset) == INTEGER_CST)
>      {

I think the types_compatible_p call is too heavy, you want to check if
emitting the fold_convert to sizetype is unnecessary, which is
  if (!useless_type_conversion_p (sizetype, TREE_TYPE (offset)))
    offset = fold_convert (sizetype, offset);
types_compatible_p is that the conversion is useless in both directions.

Ok for trunk with that change.

	Jakub