Re: [PATCH v8 1/2] IOMMU/VT-d: wire common device reserved memory API

[Marek Marczykowski-Górecki <marmarek@xxxxxxxxxxxxxxxxxxxxxx>]

On Thu, Nov 03, 2022 at 02:55:31AM +0000, Henry Wang wrote:
> Hi Jan,
> 
> > -----Original Message-----
> > From: Jan Beulich <jbeulich@xxxxxxxx>
> > Subject: Re: [PATCH v8 1/2] IOMMU/VT-d: wire common device reserved
> > memory API
> > 
> > >>>> Signed-off-by: Marek Marczykowski-Górecki
> > >>> <marmarek@xxxxxxxxxxxxxxxxxxxxxx>
> > >>>> Reviewed-by: Kevin Tian <kevin.tian@xxxxxxxxx>
> > >>>
> > >>> Henry, can this be included in 4.17? The AMD counterpart went in
> > >>> earlier, but due to late review on Intel part, this one didn't.
> > >>
> > >> Thanks for the information. I agree this is a valid reason, but to be
> > >> safe I would like to hear opinions from the x86 maintainers (added
> > >> in CC).
> > >>
> > >> Andrew/Jan/Roger: May I have your feedback about this? Thanks!
> > >
> > > Hmm, not sure what to say here. Yes, it would be nice for things to end
> > > up consistent across vendors. And yes, the change here is largely
> > > mechanical (afaics) and to code most of which shouldn't typically be in
> > > use on systems anyway, and so should not pose an undue risk. But still
> > > it is quite a bit of code churn ...
> > 
> > Was this lost, did you decide against allowing this in, or were you hoping
> > for further responses by others?
> 
> Sorry for the confusion. Yeah I was hoping to see if we can have further
> responses from others, but it seems no responses so far...
> 
> I have the exact same opinion as yours so I am also not sure. But if you
> changed your mind and would like to commit the patch for completeness
> of the original series, please feel free to add my release-ack. I would not
> block this patch.

FWIW, most of the diff is just extracting loop body into a function, the
only functional difference is a new called for this function, and moving
one of the checks (MAX_USER_RMRR_PAGES enforcement) outside of it. So,
my (biased) opinion is, it's rather low risk.

-- 
Best Regards,
Marek Marczykowski-Górecki
Invisible Things Lab

Attachment: signature.asc
Description: PGP signature