[libvirt] [PATCH 7/7] parallels: don't prevent domain define if VIR_DOMAIN_NET_TYPE_BRIDGE
Dmitry Guryanov
dguryanov at parallels.com
Wed Mar 18 13:31:54 UTC 2015
On 03/17/2015 05:14 PM, Michal Privoznik wrote:
> On 13.03.2015 16:52, Maxim Nestratov wrote:
>> network adapter is used
>> ---
>> src/parallels/parallels_sdk.c | 3 ++-
>> 1 files changed, 2 insertions(+), 1 deletions(-)
>>
>> diff --git a/src/parallels/parallels_sdk.c b/src/parallels/parallels_sdk.c
>> index a0a2ba0..4c90a18 100644
>> --- a/src/parallels/parallels_sdk.c
>> +++ b/src/parallels/parallels_sdk.c
>> @@ -2246,7 +2246,8 @@ static int prlsdkCheckSerialUnsupportedParams(virDomainChrDefPtr chr)
>>
>> static int prlsdkCheckNetUnsupportedParams(virDomainNetDefPtr net)
>> {
>> - if (net->type != VIR_DOMAIN_NET_TYPE_NETWORK) {
>> + if (net->type != VIR_DOMAIN_NET_TYPE_NETWORK &&
>> + net->type != VIR_DOMAIN_NET_TYPE_BRIDGE) {
>> virReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s",
>> _("Specified network adapter type is not "
>> "supported by Parallels Cloud Server."));
>>
> I really, really hate how the whole 'is this configuration supported'
> thing is handled in parallels. Instead of enumerating what it does not
> know, it should do it the other way round. Do a positive checking, if
> configuration is supported. Then, even if we introduce yet another
> device, or device attribute in other hypervisors, we don't need to
> update parallels driver (which we permanently keep forgetting about).
I also don't like, that this function will fail, if some unsupported
option is provided. It was long time ago when Peter told me, that I
should report error, if there are some unsupported parameters:
https://www.redhat.com/archives/libvir-list/2012-July/msg00602.html.
It should be possible to define a domain if some unsupported parameters
are provided, which wouldn't affect domain behaviour very much.
> ACK though.
>
> Michal
>
> --
> libvir-list mailing list
> libvir-list at redhat.com
> https://www.redhat.com/mailman/listinfo/libvir-list
--
Dmitry Guryanov
More information about the libvir-list
mailing list