[Intel-gfx] [RFC] make i830_debug handle different chip types

Jesse Barnes jbarnes at virtuousgeek.org
Wed Feb 11 01:03:13 CET 2009


On Tuesday, February 10, 2009 3:21 pm Eric Anholt wrote:
> On Mon, 2009-02-09 at 14:30 -0800, Jesse Barnes wrote:
> > With KMS in the tree now, I expect to have to debug mode some mode
> > setting issues, suspend/resume, and various other bits, so I thought it
> > might be a good time to update i830_debug.c to handle different chip
> > types.
> >
> > So this patch splits the register blocks into per-chipset lists, and
> > updates the code accordingly.  The snapshotting code ends up being a bit
> > different (doesn't capture later regs), but that's probably fine anyway.
> >
> > Comments?  With this split in place I could fix up the 830/915/965 fence
> > reg debug output, and also allow the reg dumper to be more selective
> > about what it dumps by default.
>
> Or we could just remove the static lists and programmatically add regs
> to be dumped.  Seems like that'll avoid duplication and let us take
> advantage of our existing macros for determining some things like fbc
> regs being present.

Yeah, that's fine too, but doesn't play as well with the snapshotting...  A 
sed script would make the conversion pretty easy.

--
Jesse Barnes, Intel Open Source Technology Center



More information about the Intel-gfx mailing list