...
> ....
>
>> @@ -3298,6 +3325,21 @@ virDomainDefPostParseInternal(virDomainDefPtr def,
>> return -1;
>> }
>>
>> + /* Fully populate the IOThread ID list */
>> + if (def->iothreads && def->iothreads != def->niothreadids) {
>> + unsigned int iothread_id = 1;
>> + while (def->niothreadids != def->iothreads) {
>> + if (!virDomainIOThreadIDFind(def, iothread_id)) {
>> + virDomainIOThreadIDDefPtr iothrid;
>> +
>> + if (!(iothrid = virDomainIOThreadIDAdd(def, iothread_id)))
>> + return -1;
>
> Unfortunately, fixing the iothread list after you parse iothread pinning
> in patch 4 makes a loophole where you might force arbitrary iothread ID
> without using <iothreadids>.
>
> This code will probably need to be moved after the parsing code, despite
> the fact that the postparse callback is better place to do such checks.
>
>> + iothrid->autofill = true;
>> + }
>> + iothread_id++;
>> + }
>> + }
>> +
>> if (virDomainDefGetMemoryInitial(def) == 0) {
>> virReportError(VIR_ERR_XML_ERROR, "%s",
>> _("Memory size must be specified via <memory> or in
>> the "
>
> The rest of this patch looks good, but I'd like to see the above part
> fixed before my final ACK.
>
I'll move it and post v5 for 1/9 here as well as 4/9
I haven't yet got down to 8/9
I'd like to get this in before code freeze...
Tks -
John
--
libvir-list mailing list
[email protected]
https://www.redhat.com/mailman/listinfo/libvir-list
