On Thu, 2007-05-03 at 22:23 +0900, Isaku Yamahata wrote:
> On Wed, May 02, 2007 at 11:27:31PM -0600, Alex Williamson wrote:
> > --- linux-2.6-xen-sparse/arch/ia64/hp/common/sba_iommu.c 2006-09-19
> > 21:42:06.000000000 -0600
> > +++ linux-2.6-xen-sparse/arch/ia64/hp/common/sba_iommu.c 2007-05-02
> > 22:52:34.000000000 -0600
> > @@ -944,7 +943,7 @@
> >
> > while (size > 0) {
> > ASSERT(((u8 *)pdir_start)[7] == 0); /* verify availability */
> > - sba_io_pdir_entry(pdir_start, (unsigned long) addr);
> > + sba_io_pdir_entry(pdir_start, virt_to_bus(addr));
> >
> > DBG_RUN(" pdir 0x%p %lx\n", pdir_start, *pdir_start);
> >
>
> sba_io_pdir_entry() seems to expect virtual address as 2nd argument
> so that modifying sba_io_pdir_entry() itself would be better than
> modifying all callers argument.
> I don't know the sba_iommu spec, though.
Good point. sba_io_pdir_entry is taking a shortcut since it's only
mapping region 7 addresses. It would probably be more correct to do the
virt_to_bus there. Functionally the approaches end up with the same
result since a mask is used to do the virt to phys translation in
sba_io_pdir_entry. Thanks,
Alex
--
Alex Williamson HP Open Source & Linux Org.
_______________________________________________
Xen-ia64-devel mailing list
Xen-ia64-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-ia64-devel
|
Home