[PATCH] Re: PATCH: fix for problem in i386.md ashlqi3_1_lea pattern

Richard Henderson rth@cygnus.com
Thu Aug 3 16:00:00 GMT 2000


On Wed, Aug 02, 2000 at 01:17:24PM +0200, Jakub Jelinek wrote:
> 2) Replicates the lea_general conditionals for the splitter conditions:
>    although Richard said you you don't have to replicate them, it seems
>    the splitters are otherwise matched and I saw multiple times that e.g.
>    a SImode lea_0 valid instruction is splitted into the identical lea_0
>    instruction etc.
[...]
> -  "reload_completed"
> +  "reload_completed
> +   && (GET_MODE (operands[0]) == QImode || GET_MODE (operands[0]) == HImode)
> +   && (!TARGET_PARTIAL_REG_STALL || optimize_size)
> +   && GET_MODE (operands[0]) == GET_MODE (operands[1])
> +   && GET_MODE (operands[0]) == GET_MODE (operands[2])
> +   && (GET_MODE (operands[0]) == GET_MODE (operands[3])
> +       || GET_MODE (operands[3]) == VOIDmode)"

That's because I said use "&& reload_completed".  The initial "&&"
is critical.


r~


More information about the Gcc-patches mailing list