Re: variable objects and registers

[Vladimir Prus <ghost@cs.msu.su>]

Daniel Jacobowitz wrote:

> On Wed, Nov 29, 2006 at 08:20:47PM +0300, Vladimir Prus wrote:
>> 
>> I was working on implementing a new MI command that creates variable
>> objects for registers, and I have something working good enough to
>> discuss. This patch lacks docs, but I wanted to make sure the interface
>> is fine with everybody before documenting it.
> 
> No one else had any comments, and it looks fine to me.  The patch looks
> OK too.  It'll want a testcase, naturally.

Ehm, note that the test for -data-list-register-names only works for Sparc.
I must admit I don't know how to test this behaviour in a generic fashion.
It might be possible to compare output of -data-list-register-names
and -var-registers, but given that the latter is implemented by cloning
some logic of the former, that's weak test.

>> As soon as we add the code to display memory-mapped registers, there will
>> be a problem that existing frontends might wish to show the memory-mapped
>> registers, but not wish (at the moment) to modify the code for displaying
>> regular registers. I plan to address this by either adding new
>> attribute "register-kind" to the output, that can be either "core"
>> or "memory-mapped", or by adding an option to -var-registers that says
>> what registers to show. But that's for future.
> 
> Or a different command to return just those?  I'm not sure they should
> be part of -var-registers; clients probably expect "registers" to be
> "the things that should go in a registers view", which won't include
> most MMIO registers.

FWIW, in my case MMIO registers will go straight to registers view. I think
that an interface with options to -var-registers is better that adding yet
another command.

- Volodya