[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Xen-devel] qemu/xendisk: properly update stats in ioreq_release()



>>> On 11.05.12 at 18:01, Stefano Stabellini <stefano.stabellini@xxxxxxxxxxxxx>
wrote:
> On Thu, 10 May 2012, Jan Beulich wrote:
>> While for the "normal" case (called from blk_send_response_all())
>> decrementing requests_finished is correct, doing so in the parse error
>> case is wrong; requests_inflight needs to be decremented instead.
>> 
>> Signed-off-by: Jan Beulich <jbeulich@xxxxxxxx>
>> 
>> --- a/hw/xen_disk.c
>> +++ b/hw/xen_disk.c
>> @@ -154,7 +154,7 @@ static void ioreq_finish(struct ioreq *i
>>      blkdev->requests_finished++;
>>  }
>>  
>> -static void ioreq_release(struct ioreq *ioreq)
>> +static void ioreq_release(struct ioreq *ioreq, bool finish)
>>  {
>>      struct XenBlkDev *blkdev = ioreq->blkdev;
>>  
>> @@ -162,7 +162,10 @@ static void ioreq_release(struct ioreq *
>>      memset(ioreq, 0, sizeof(*ioreq));
>>      ioreq->blkdev = blkdev;
>>      QLIST_INSERT_HEAD(&blkdev->freelist, ioreq, list);
>> -    blkdev->requests_finished--;
>> +    if (finish)
>> +        blkdev->requests_finished--;
>> +    else
>> +        blkdev->requests_inflight--;
>>  }
>>  
>>  /*
>> @@ -449,7 +452,7 @@ static void blk_send_response_all(struct
>>      while (!QLIST_EMPTY(&blkdev->finished)) {
>>          ioreq = QLIST_FIRST(&blkdev->finished);
>>          send_notify += blk_send_response_one(ioreq);
>> -        ioreq_release(ioreq);
>> +        ioreq_release(ioreq, true);
>>      }
>>      if (send_notify) {
>>          xen_be_send_notify(&blkdev->xendev);
>> @@ -505,7 +508,7 @@ static void blk_handle_requests(struct X
>>              if (blk_send_response_one(ioreq)) {
>>                  xen_be_send_notify(&blkdev->xendev);
>>              }
>> -            ioreq_release(ioreq);
>> +            ioreq_release(ioreq, false);
>>              continue;
>>          }
> 
> In case of an error returned by ioreq_parse requests_finished should
> also be decremented.

I don't think so - it gets incremented in ioreq_finish(), and that gets
called only when I/O was at least attempted (which isn't the case
when parse failed).

Jan


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


 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.