> The attached patch (pr85667.patch) fixes the subjected issue for 32-bit.
> Please let me know your thoughts on the patch.

IMO you ought not to duplicate most of function_value_32 here.

-- 
Eric Botcazou

Reply via email to