On 09.07.21 10:55, Roger Pau Monné wrote:
> On Thu, Jul 08, 2021 at 02:43:44PM +0200, Juergen Gross wrote:
>> In order to avoid a malicious backend being able to influence the local
>> copy of a request build the request locally first and then copy it to
>> the ring page instead of doing it the other way round as today.
>>
>> Signed-off-by: Juergen Gross <jgross@suse.com>
>> Reviewed-by: Jan Beulich <jbeulich@suse.com>
> 
> Acked-by: Roger Pau Monné <roger.pau@citrx.com>
> 
> Thanks!
> 
> One unrelated comment below.
> 

...

>> @@ -798,7 +801,9 @@ static int blkif_queue_rw_req(struct request *req, struct blkfront_ring_info *ri
>>   		ring_req->u.rw.nr_segments = num_grant;
>>   		if (unlikely(require_extra_req)) {
>>   			extra_id = blkif_ring_get_request(rinfo, req,
>> -							  &extra_ring_req);
>> +							  &final_extra_ring_req);
>> +			extra_ring_req = &rinfo->shadow[extra_id].req;
> 
> I'm slightly confused about this extra request stuff because I cannot
> find any check that asserts we have two empty slots on the ring before
> getting here (I only see a RING_FULL check in blkif_queue_rq).
> 
> This is AFAIK only used on Arm when guest page size > 4KB.

I agree that this is a bug and should be fixed.


Juergen