Re: How to clean up i386 machine description?

[Joern Rennecke <amylaar at cygnus dot co dot uk>]

> Colin Douglas Howell writes:
> 
>  > This gives me a interesting thought.  We have several requirements
>  > here.  One is to write the instruction patterns so that we know what
>  > type of instruction has been selected for scheduling and output
>  > purposes.  A second is to tighten the operand predicates so we
>  > minimize the amount of work done by reload.  Finally, we must write
>  > the patterns so that reload can do a good job of fixing up operands at
>  > minimal cost.  Since in many cases the type of instruction we will
>  > select depends greatly on the operands, and since reload is modifying
>  > the operands, these requirements may be in conflict.
>  > 
>  > Suppose we solve them by replacing each pattern with three sets of
>  > patterns, one to be used before reload, one during reload, and one
>  > after.
> 
> Hmmm, this looks like a pattern explosion!  I haven't been following
> this thread too closely so I might have the wrong end of the stick.
> I gather you're wanting to tighten up the operand predicates for the
> x86 and are having trouble with the weird combination of valid
> operands.  
> 
> If this is the case, what I suggest you do is to use the extra
> condition field of the patterns to reject invalid combinations of the
> operands.  I did this with the C4x target where I wanted to minimise
> the amount of reloading.  On the x86 you'll probably find that the
> predicates match almost anything but that there are limited
> combinations of valid operands.

This is a good approach for machine-specific patterns, but you can't apply
it to standard patterns like movsi or addsi, since the predicates of these
must be evaluable independent of the instruction.