Re: [Xen-devel] [PATCH] xen/arm: map_dev_mmio_region: The iomem permission check should be done on MFN

[Stefano Stabellini <sstabellini@xxxxxxxxxx>]

On Wed, 15 Jun 2016, Shannon Zhao wrote:
> Hi Julien,
> 
> On 2016/6/14 19:50, Julien Grall wrote:
> > The helper iomem_access_permitted expects MFNs in parameters and not
> > GNFs. Thankfully only the hardware domain can call this function and
> > it will always be with GFNS == MFNs for now.
> > 
> > Also, fix the printf to use the MFN range and not the GFN one.
> > 
> > Signed-off-by: Julien Grall <julien.grall@xxxxxxx>
> > Cc: Shannon Zhao <shannon.zhao@xxxxxxxxxx>
> > 
> Reviewed-by: Shannon Zhao <shannon.zhao@xxxxxxxxxx>

Shannon, thanks for your help reviewing this.

Acked-by: Stefano Stabellini <sstabellini@xxxxxxxxxx>

I'll commit.

> > ---
> >     This patch is a good candidate to backport to Xen 4.7. Without
> >     it, the hardware domain can map any MMIO because the permission
> >     check is done on the GPFNs and not the MNFs.
> > ---
> >  xen/arch/arm/p2m.c | 4 ++--
> >  1 file changed, 2 insertions(+), 2 deletions(-)
> > 
> > diff --git a/xen/arch/arm/p2m.c b/xen/arch/arm/p2m.c
> > index 6a19c57..4c6547d 100644
> > --- a/xen/arch/arm/p2m.c
> > +++ b/xen/arch/arm/p2m.c
> > @@ -1275,14 +1275,14 @@ int map_dev_mmio_region(struct domain *d,
> >  {
> >      int res;
> >  
> > -    if ( !(nr && iomem_access_permitted(d, start_gfn, start_gfn + nr - 1)) 
> > )
> > +    if ( !(nr && iomem_access_permitted(d, mfn, mfn + nr - 1)) )
> >          return 0;
> >  
> >      res = map_mmio_regions(d, start_gfn, nr, mfn);
> >      if ( res < 0 )
> >      {
> >          printk(XENLOG_G_ERR "Unable to map [%#lx - %#lx] in Dom%d\n",
> > -               start_gfn, start_gfn + nr - 1, d->domain_id);
> > +               mfn, mfn + nr - 1, d->domain_id);
> >          return res;
> >      }
> >  
> > 
> 
> -- 
> Shannon
> 

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
https://lists.xen.org/xen-devel