Re: MTRR setting and framebuffer driver

Richard Gooch (rgooch@ras.ucalgary.ca)
Tue, 14 Sep 1999 12:20:52 -0600

Jeff Garzik writes:
> Jamie Lokier wrote:
> > I would like either vesafb, or the MTRR driver to realise that this
> > 2.5MB region is a sub-region of the 4MB region that won't be used by
> > anything else, so it is safe to allocate a 4MB write-combining MTRR.
> >
> > Richard, is this possible now that we have the resource tree to examine?
> > Perhaps it is possible to generically permit rounding an MTRR range,
> > provided the extra address space is not used in the resource tree, and
> > if this condition changes in the tree, to remove the MTRR.
>
> IMHO this should be done in vesafb, possibly based on the value passed
> to __request_region().

I agree. I don't like the idea of adding kludges to the MTRR code. It
should provide a minimal interface, making no assumptions about PCI
regions or anything like that.

That said, if a more convenient facility would be generally useful
(not just in vesafb), then that could be added somewhere else, and
vesafb and similar drivers could make use of it. It probably makes
sense to put such a facility in the x86-specific PCI support code,
since we are assuming PCI region rules.

Regards,

Richard....
Old: rgooch@atnf.csiro.au
Current: rgooch@ras.ucalgary.ca

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/