[G]CSE of extended asm statements

Bernd Schmidt bernds@cygnus.co.uk
Wed Aug 30 00:28:00 GMT 2000


On Tue, 29 Aug 2000, Richard Henderson wrote:
> 
> Actually, the best way to represent this is with
> 
>   (set (parallel [(out1) (out2) (out3)])
>        (asm_operand ...))

I strongly dislike that.  We have an existing way to describe
parallel sets, and I don't see a benefit in adding another one.
I have a feeling that allowing the above will create a whole new
set of bugs because most likely there are several dozen places
in the compiler that would not be able to handle this construct.


Bernd



More information about the Gcc-patches mailing list