Re: [Xen-devel] [PATCH 16/18 V2]: PVH xen: elf changes to pref for dom0 PVH.

["Jan Beulich" <JBeulich@xxxxxxxx>]

>>> On 19.03.13 at 02:13, Mukesh Rathor <mukesh.rathor@xxxxxxxxxx> wrote:
> On Mon, 18 Mar 2013 12:43:33 +0000
> "Jan Beulich" <JBeulich@xxxxxxxx> wrote:
> 
>> >>> On 16.03.13 at 02:04, Mukesh Rathor <mukesh.rathor@xxxxxxxxxx>
>> >>> wrote:
>> > +    /* raw_copy_to_guest -> copy_to_user_hvm -> __hvm_copy needs
>> > curr to
>> > +     * point to the hvm/pvh vcpu. Hence for PVH dom0 we can't use
>> > that. For now
>> > +     * just use dbg_rw_mem(). */
>> 
>> Again - definitely not outside of an RFC patch.
> 
> What, the "for now" comment, or the use of dbg_rw_mem()? There are fixme's 
> in
> xen already. dbg_rw_mem() is perfectly fine to use IMO, but in future we
> may look at a faster copy, hence the comment.

No, this is a debugger interface, and should hence only be used
for that purpose. And actually I would have expected that
debugger code to only be built conditionally only anyway, the
more that we have a HAS_GDBSX construct. That not being the
case is no excuse imo, as much as pointing at other fixme
comments elsewhere isn't.

The fundamental problem here is that you'd have to prove that
an interface that isn't intended to be secure is now being used
in a way that doesn't open security holes.

Jan


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