Re: [Xen-ia64-devel] [Patch 2/2] paravirtualize mmap hundlers of /proc/bus/pci

[Isaku Yamahata <yamahata@xxxxxxxxxxxxx>]

On Thu, Feb 22, 2007 at 06:31:39AM +0100, Tristan Gingold wrote:
> On Thu, Feb 22, 2007 at 02:23:38PM +0900, Isaku Yamahata wrote:
> > On Thu, Feb 22, 2007 at 05:55:13AM +0100, Tristan Gingold wrote:
> > > > add_window(), which registers io region and memory region,
> > > > calls ioremap with ACPI_IO_RANGE regions, but it doesn't
> > > > with ACPI_MEMORY_RANGE regions.
> > > > If the ioremap hypercall is called with ACPI_MEMORY_RANGE regions,
> > > > all pci devices are free from such paravirtualization.
> > > > Especially we can get rid of /dev/mem and pci mmap
> > > > paravirtualization to use X.
> > > Would be nice, but isn't the ACPI_MEMORY_RANGE very large on some machine 
> > > ?
> > 
> > Yes.
> > So I wondered what kind of devices had such large area.
> IMHO this is simply reserved memory.
> There may be a misunderstanding here because I don't know this code well
> enough.  Does add_window() registers region from the memmap or from pci
> devices BARs ?

It registers from acpi walk and the result can be seen
in /proc/ioport and /proc/iomem.


> > > > Presumably we don't want to map pci bridge which may be large 
> > > > so that we need to detect it somehow in add_window().
> > > Pci root bridge or pci-to-pci bridges ?
> > Yes.
> Yes means both ?

both.

-- 
yamahata

_______________________________________________
Xen-ia64-devel mailing list
Xen-ia64-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-ia64-devel