On 05/25/2018 04:24 AM, Michal Privoznik wrote:
> On 05/17/2018 02:42 PM, John Ferlan wrote:
>>
https://bugzilla.redhat.com/show_bug.cgi?id=1149445
>>
>> If the domain requests usage of the genid functionality,
>> then add the QEMU '-device vmgenid' to the command line
>> providing either the supplied or generated GUID value.
>>
>> Add tests for both a generated and supplied GUID value.
>>
>> Signed-off-by: John Ferlan <jferlan(a)redhat.com>
>> ---
>> src/qemu/qemu_command.c | 24 +++++++++++++++++
>> .../qemuxml2argvdata/genid-auto.x86_64-latest.args | 30 ++++++++++++++++++++++
>> tests/qemuxml2argvdata/genid.x86_64-latest.args | 30 ++++++++++++++++++++++
>> tests/qemuxml2argvtest.c | 4 +++
>> 4 files changed, 88 insertions(+)
>> create mode 100644 tests/qemuxml2argvdata/genid-auto.x86_64-latest.args
>> create mode 100644 tests/qemuxml2argvdata/genid.x86_64-latest.args
>>
>> diff --git a/src/qemu/qemu_command.c b/src/qemu/qemu_command.c
>> index c4237339bf..78bd685008 100644
>> --- a/src/qemu/qemu_command.c
>> +++ b/src/qemu/qemu_command.c
>> @@ -6002,6 +6002,27 @@ qemuBuildSmbiosCommandLine(virCommandPtr cmd,
>> }
>>
>>
>> +static int
>> +qemuBuildVMGenIDCommandLine(virCommandPtr cmd,
>> + const virDomainDef *def)
>> +{
>> + virBuffer opts = VIR_BUFFER_INITIALIZER;
>> + char guid[VIR_UUID_STRING_BUFLEN];
>> +
>> + if (!def->genidRequested)
>> + return 0;
>> +
>> + virUUIDFormat(def->genid, guid);
>> + virBufferAsprintf(&opts, "vmgenid,guid=%s,id=vmgenid0",
guid);
>> +
>> + virCommandAddArg(cmd, "-device");
>> + virCommandAddArgBuffer(cmd, &opts);
>> +
>> + virBufferFreeAndReset(&opts);
>> + return 0;
>
> As mentioned in previous patch this function must check for
> DEVICE_VMGENID capability. Otherwise we might put -device vmgenid on
> command line for qemu that doesn't support it.
>
Which path would do that? I believe I'm making "interesting" use of the
DO_TEST_CAPS_LATEST actually.
There are two ways to call qemuBuildCommandLine - one via the PretendCmd
logic and one via the Launch logic.
In the Launch logic prior to calling qemuBuildCommandLine, the
capability is checked *and* the "need" for generating a new GUID is
determined. It was a "design decision" to not check/change it earlier.
Is there somewhere else you believe that code should go? IDC, I can move
it, but understand the risk of doing so.
Now that we have DO_TEST_CAPS_LATEST there's no need to pass an explicit
flag for the PretendCmd line logic, so we get it for free for the
test/fake command line.
Well, qemuConnectDomainXMLToNative() for instance. So I suggest leaving
the capability check in qemuProcessGenID() and having it repeated here.
Michal