On 11/29/2011 09:26 AM, Daniel P. Berrange wrote:
On Mon, Nov 14, 2011 at 09:30:00PM -0700, Eric Blake wrote:
> Without this, 'virsh blkiotune --live --config --weight=n'
> only affected live.
>
> * src/qemu/qemu_driver.c (qemuDomainSetBlkioParameters): Allow
> setting both configurations at once.
> ---
> src/qemu/qemu_driver.c | 5 ++++-
> 1 files changed, 4 insertions(+), 1 deletions(-)
>
> diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c
> index 5f4a18d..b0ce115 100644
> --- a/src/qemu/qemu_driver.c
> +++ b/src/qemu/qemu_driver.c
> @@ -5981,7 +5981,10 @@ static int qemuDomainSetBlkioParameters(virDomainPtr dom,
> ret = -1;
> }
> }
> - } else if (flags & VIR_DOMAIN_AFFECT_CONFIG) {
> + }
> + if (ret < 0)
> + goto cleanup;
> + if (flags & VIR_DOMAIN_AFFECT_CONFIG) {
> /* Clang can't see that if we get here, persistentDef was set. */
> sa_assert(persistentDef);
ACK
Thanks for the reviews. I've now pushed 1, 1.5, and 2. It looks like
Hu's suggestions to fix 3 and 4 to store 0 values in the DefPtr, but
omit it from the xml output, work without much further effort on my
part, and that both Hu and my contributions have all been acked by at
least one other reviewer, so I will finish my testing that I didn't
break anything when rebasing to merge Hu's touchups into my v4 patches,
then push those as well.
--
Eric Blake eblake(a)redhat.com +1-919-301-3266
Libvirt virtualization library
http://libvirt.org