Re: [rfc][4/13] Eliminate read_register: read_register in deprecated_mips_set_processor_regs_hack

[Daniel Jacobowitz <drow@false.org>]

On Tue, Jun 12, 2007 at 03:49:57PM +0100, Maciej W. Rozycki wrote:
>  Worse yet -- the approach that we have taken is generic.  We can handle 
> arbitrary MIPS32/MIPS64 processors as conforming to the current revisions 
> of the base architecture specs and the application-specific extensions 
> (ASEs) by decoding feature bits defined in cp0 config registers -- there 
> four config registers defined so far; additional registers may need to be 
> read for variable length register subsets (e.g. watch and performance 
> counter registers).

I see - that sounds pretty nice; we just need to come up with a way to
support it in the rest of GDB.  Perhaps we should simply bite the
bullet and have a late target description hook in addition to the
early one.  We can't do without the early one; I added it e.g. for
existing MIPS remote targets, where the stub might send back 32-bit or
64-bit registers.

Or maybe I'm barking up the wrong tree entirely and we need more than
just a target description.  The description could report which
registers are available and how big they are, and then GDB could
gather other details from the config registers.

> > I don't have an easy answer for this.  Simplest would be to keep it
> > local to remote-mips.c (as it currently is), but change how it works;
> > move it from common_open to a new remote_mips_read_description, fetch
> > the PRID without going through GDB's register cache at all, and then
> > create an appropriate target description which specifies the processor
> > based on the PRID.  It would be nicer if we could make it work for
> > remote.c too though.
> 
>  Well, it's actually in mips-tdep.c, so it should work for any MIPS 
> target.

Have you moved the call?  Right now, that function is only reached
through remote-mips.c.

>  I guess this is unfeasible -- there are too many possibilites which are 
> neither fixed nor easy to predict as you can see from the above.  Unless 
> the XML tags provide means for subsetting the architecture.  Please note 
> that to make the matter more exciting the subsets do overlap.

The XML tags provide whatever you want them to - it's very easy to add
new ones :-)  We could provide the values of the config registers
directly in a description, for instance, and let mips-tdep.c interpret
them there.  But it might be better to have GDB query them using
normal register methods.

-- 
Daniel Jacobowitz
CodeSourcery