Re: AW: [avr-libc-dev] Additional instruction patterns for or more efficient code using mixed QI/HI/SI expressions

[Denis Chertykov]

Björn Haase <address@hidden> writes:

> Hi,
> 
> I have been adding some specialized instruction patterns
> for dealing with expresions that mix QI,HI and SI objects. (Tedious work.)
> 
> In addition to a bunch of specialized "combined" patterns, I would like to
> suggest to
> change the "copy_to_mode_reg" within the movsi and movhi "define_expand"s.
> We could copy const_int values that could fit into 8 or 16 bit in QI or HI
> registers instead and use the appropriate extension operations for promoting
> the values to the HI/SI target modes on the fly.
> 
> In my opinion the main possible benefit of the suggested changes in the
> machine description is that less registers are required.

Are you have a signed FSF COPYING ASSIGNMENT FOR GCC ?
If not then I can't apply your patch to FSF's GCC CVS tree.

Some your new patterns are right some are wrong. I suggest to produce
the work step by step with small changes.

Your code must satisfy GNU coding standards !

> P.S.:
> 
> Could anyone give me an advice on how to write automized scripts for
> checking compilation result on a simulator? I have tried to use
> simulavr. I, however, did not succeed to build it under cygwin.

You must learn DEJAGNU.

> 
> 
> P.P.S.:
> 
> implementing at least plus:DI and minus:DI operations would be so easy. Any
> reason, not to extend
> the existing %Ax ... %Dx templates to the %Ax ... %Hx range and add the
> corresponding plus:DI and
> minus:DI patterns in the machine description?

I havn't such reasons. (Except my opinion in PS)

I think that implementing of movdi pattern most important.

Denis.

PS: Righ now I think that better to rewrite many insns for
    HI,SI,DI modes as splitters to QI mode (naitive AVR) insns.
    If you want to help then better to concentrate on this.