Re: [Xen-devel] [PATCH 2/2] x86/HVM: fix setting mem access to default

[Tim Deegan <tim@xxxxxxx>]

At 08:50 +0100 on 03 Apr (1396511454), Jan Beulich wrote:
> >>> On 03.04.14 at 02:13, <aravindp@xxxxxxxxx> wrote:
> >> commit 3b0bcb89 ("x86/mm/p2m: Move p2m code in
> >>HVMOP_[gs]et_mem_access into p2m.c") introduced an off-by-one mistake
> >>forcing an input of HVMMEM_access_default to always fail. Since related, 
> >>also
> >>eliminate the inefficient setup of an on-stack array for each function
> >>invocation.
> > 
> > Thanks for catching this. I do have a question. If the caller wants to set 
> > the default access by specifying pfn == ~0ul, the call could be made with 
> > access = HVMMEM_access_default. Should this be allowed?
> 
> I asked myself this question too, but wasn't really decided about the
> usefulness of this, so left the behavior unchanged (i.e. allow this
> pointless operation). I'll leave the decision whether to disable this to
> the mm maintainers.

Seems harmless enough, and has well-defined (if admittedly pointless)
semantics.  I'd accept a patch to disable it if anyone feels strongly
enough about it to write one. :)

Tim.

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