Re: [Xen-devel] [PATCH v3] xen: handle paged gfn in wrmsr_hypervisor_regs

["Jan Beulich" <JBeulich@xxxxxxxx>]

>>> On 03.05.13 at 17:58, Keir Fraser <keir.xen@xxxxxxxxx> wrote:
> On 03/05/2013 16:30, "Jan Beulich" <JBeulich@xxxxxxxx> wrote:
> 
>>>          if ( idx > 0 )
>>>          {
>>>              gdprintk(XENLOG_WARNING,
>>>                       "Out of range index %u to MSR %08x\n",
>>>                       idx, 0x40000000);
>>> -            return 0;
>>> +            return -EINVAL;
>> 
>> But I'd stay away from converting to actual errors both here ...
>> 
>>>          }
>>>  
>>> -        page = get_page_from_gfn(d, gmfn, NULL, P2M_ALLOC);
>>> +        page = get_page_from_gfn(d, gmfn, &t, P2M_ALLOC);
>>>  
>>>          if ( !page || !get_page_type(page, PGT_writable_page) )
>>>          {
>>>              if ( page )
>>>                  put_page(page);
>>> +
>>> +            if ( p2m_is_paging(t) )
>>> +            {
>>> +                p2m_mem_paging_populate(d, gmfn);
>>> +                return -EAGAIN;
>>> +            }
>>> +
>>>              gdprintk(XENLOG_WARNING,
>>>                       "Bad GMFN %lx (MFN %lx) to MSR %08x\n",
>>> -                     gmfn, page_to_mfn(page), base + idx);
>>> -            return 0;
>>> +                     gmfn, page ? page_to_mfn(page) : -1UL, base);
>>> +            return -EINVAL;
>> 
>> ... and here. If at all these ought to go into a separate patch
>> (which we'd likely postpone until after 4.3).
> 
> But they are errors?

Right. But they weren't handled as such so far, and doing the
conversion isn't the purpose of the patch.

> Could agree with postponing to post-4.3 though.

Hence the request to split off that part of the patch. I'd be happy
to apply that after the trees got branched.

Jan


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