[libvirt] [PATCH 3/4] domain_conf: Add support for iothreads in disk definition

[John Ferlan]

On 08/26/2014 12:04 PM, Ján Tomko wrote:
> On 08/26/2014 02:38 AM, John Ferlan wrote:
>> Add a new disk "driver" attribute "iothread" to be parsed as the thread
>> number for the disk to use. In order to more easily facilitate the usage
>> and configuration of the iothread, a "zero" for the attribute indicates
>> iothreads are not supported for the device and a positive value indicates
>> the specific thread to try and use.
>>
>> Signed-off-by: John Ferlan <jferlan at redhat.com>
>> ---
>>  docs/formatdomain.html.in     |  8 ++++++++
>>  docs/schemas/domaincommon.rng | 13 +++++++++++++
>>  src/conf/domain_conf.c        | 24 +++++++++++++++++++++++-
>>  src/conf/domain_conf.h        |  2 ++
>>  4 files changed, 46 insertions(+), 1 deletion(-)
>>
> 
>> @@ -11968,6 +11980,14 @@ virDomainDefParseXML(xmlDocPtr xml,
>>              goto error;
>>          }
>>          def->iothreads = count;
>> +
>> +        /* Create a bitmap for inuse threads - noting that entries are
>> +         * numbered 1..def->iothreads since 0 (zero) iothreads means
>> +         * nothing and assigning a disk to an IOThread requires at least a
>> +         * thread# > 0 since a zero would indicate no IOThread for the disk
>> +         */
>> +        if (!(def->iothreadmap = virBitmapNew(def->iothreads+1)))
>> +            goto error;
>>      }
> 
> Do you plan to do multiple disks on a single iothread in the followup patches?
> Or is that a configuration that doesn't make sense?
> 

I would say it doesn't make sense and from the bz I have the comment is
"2. 1 IOThread per device" (bz 1101574 if you care to read although it's
mostly private comments though which is why I didn't put it in the patches).


John