[libvirt] [PATCH 2/2] virDomainGetBlockJobInfo: Fix corner case when qemu reports no info

[Michal Privoznik]

On 13.09.2016 11:52, Michal Privoznik wrote:
> On 12.09.2016 21:19, Eric Blake wrote:
>> On 09/09/2016 04:30 PM, John Ferlan wrote:
>>
>>>>  
>>>> +    /* Fix job completeness reporting. If cur == end mgmt
>>>> +     * applications think job is completed. Except when both cur
>>>> +     * and end are zero, in which case qemu hasn't started the
>>>> +     * job yet. */
>>>> +    if (!info->cur && !info->end) {
>>
>> We get here if qemu reports 0/0 (or if qemu reports nothing, and we end
>> up with 0/0 because we 0-initialized the object)...
>>
>>>> +        if (rawInfo->ready > 0) {
>>>> +            info->cur = info->end = 1;
>>
>> if qemu reported done (on a no-op job), then we fudge to 1/1 and the
>> caller knows we are done...
>>
>>>> +        } else if (rawInfo->ready < 0) {
>>>> +            info->end = 1;
>>
>> if qemu didn't tell us it was ready, then we fudge to 0/1.
>>
>> I thought the original email thread was that if rawInfo->ready == 0
>> (qemu explicitly told us it is NOT done) that we want to fudge to 0/1,
>> and then the real question is that if qemu tells us nothing at all about
>> rawInfo->ready, then fudging MIGHT treat a no-op job as never ending, so
>> it was better to leave it at 0/0 (an application getting 0/0 when
>> talking to new-enough libvirt then knows it is talking to older qemu).
>> In other words, I think this condition is slightly better as
>> rawInfo->ready == 0, and leave the rawInfo->ready < 0 case as 0/0.
>>
>> Or am I misremembering the results of the earlier thread?
> 
> So, just to make it crystal clear, is this what you're saying?
> 
> ready | initial C/R |fudged C/R

Oh, This should have been C/E instead of C/R. Current/End. 'e' and 'r'
keys are just too close to each other :-)

> ------+-------------+----------
>  < 0  |     0/0     |   0/0
>  = 0  |	    0/0     |   0/1
>  > 0  |	    0/0     |   1/1
> 

Michal