Re: RFD: More flexibility in defining addressing modes

[Bernd Schmidt <bernds_cb1 at t-online dot de>]

Ian Lance Taylor wrote:
> Bernd Schmidt <bernds_cb1@t-online.de> writes:
>
>
> > > At this point, would it be easier to understand if we instead passed
> > > the complete address to the target in RTL, and the target in the one
> > > case returned the register class, and in the other checked the regno?
> >
> > Well, you can have several registers in one address, and if you pass
> > the whole thing to the backend it won't know which part to look at.
>
>
> I was imagining that one would pass in the address of the REG in
> question.

One problem is that parts of the address may have had reloads applied to 
them, so the port may be looking at the wrong thing if it looks at the 
whole address.

> I don't have any particularly strong objection to your
> patch, it just seems to make an already confusing situation somewhat
> more confusing.  Adding the RTX code solves the nonorthoganality on
> the Blackfin, but there are other types of nonorthogonality in
> existing processors, like the MIPS16 in which different registers
> support different maximum offsets in unextended instructions.

Hmm, probably solvable by passing in even more information than the 
index_code, but it gets nasty.

The Blackfin also has some nonorthogonalities which aren't handled with 
this approach - which register you can load sometimes depends on which 
addressing mode you use.  I currently see no good way to teach reload 
about this kind of thing :-(

I now have a finished patch which I was going to check in, but then I 
remembered we're in stage 3.  Bummer.


Bernd