On 02/03/2012 09:09 AM, Laine Stump wrote:
On 01/31/2012 10:22 PM, Eric Blake wrote:
> On 01/31/2012 05:54 PM, Taku Izumi wrote:
>> Fix my typo at
>> commit 74e034964c32edb1732d0ff7642f3977f3587d72
>>
>> "disk->rawio == -1" indicates that this value is not
>> specified. So in case of this, domain must not
>> be tainted.
>>
>> Signed-off-by: Taku Izumi<izumi.taku(a)jp.fujitsu.com>
>> ---
>> src/qemu/qemu_domain.c | 2 +-
>> 1 file changed, 1 insertion(+), 1 deletion(-)
> ACK and pushed.
Sorry I didn't notice this mail earlier. Actually, in the version of the
code that was pushed, the concept of using "-1" as the default value was
scrapped in favor of having a "rawio_specified" bool in the structure
that defaulted to false. This way the structure doesn't need any special
initialization (the lack of which was causing multiple make check
failures with the original patch). So, rawio will only ever be set to 0
or 1 - it will never be -1.
This means that the code originally committed actually works correctly.
The modification in this "fix" also works, though.
And that still makes me wonder if we want to switch to a three-state
enum, default, off, and on, and compare disk->rawio ==
VIR_DOMAIN_DISK_RAWIO_ON, rather than tracking a secondary bool. Oh well.
--
Eric Blake eblake(a)redhat.com +1-919-301-3266
Libvirt virtualization library
http://libvirt.org