Re: [libvirt] [PATCH for 1.2.7 8/8] qemu: Implement virConnectGetDomainCapabilities

On Mon, Jun 30, 2014 at 05:31:51PM +0200, Michal Privoznik wrote: > So far only information on disks and host devices are exposed in the > capabilities XML. Well, at least something. Even a new test is > introduced. The qemu capabilities are stolen from already existing > qemucapabilities test. There's one tricky point though. Functions that > checks host's KVM and VFIO capabilities, are impossible to mock > currently. So in the test, we are setting the capabilities by hand. > > Signed-off-by: Michal Privoznik <mprivozn(a)redhat.com&gt; > --- > src/libvirt_private.syms | 1 + > src/qemu/qemu_capabilities.c | 90 ++++++++++++++++++ > src/qemu/qemu_capabilities.h | 4 + > src/qemu/qemu_driver.c | 101 +++++++++++++++++++++ > tests/Makefile.am | 5 + > .../domaincaps-qemu_1.6.50-1.xml | 44 +++++++++ > tests/domaincapstest.c | 45 +++++++++ > 7 files changed, 290 insertions(+) > create mode 100644 tests/domaincapsschemadata/domaincaps-qemu_1.6.50-1.xml > > +static void > +virQEMUCapsFillDomainDeviceDiskCaps(virQEMUCapsPtr qemuCaps, > + virDomainCapsDeviceDiskPtr disk) > +{ > + disk->device.supported = true; > + /* QEMU supports all of these */ > + VIR_DOMAIN_CAPS_ENUM_SET(disk->diskDevice, > + VIR_DOMAIN_DISK_DEVICE_DISK, > + VIR_DOMAIN_DISK_DEVICE_CDROM, > + VIR_DOMAIN_DISK_DEVICE_FLOPPY); > + > + if (virQEMUCapsGet(qemuCaps, QEMU_CAPS_VIRTIO_BLK_SG_IO)) > + VIR_DOMAIN_CAPS_ENUM_SET(disk->diskDevice, VIR_DOMAIN_DISK_DEVICE_LUN); > + > + VIR_DOMAIN_CAPS_ENUM_SET(disk->bus, > + VIR_DOMAIN_DISK_BUS_IDE, > + VIR_DOMAIN_DISK_BUS_FDC, > + VIR_DOMAIN_DISK_BUS_SCSI, > + VIR_DOMAIN_DISK_BUS_VIRTIO, > + VIR_DOMAIN_DISK_BUS_SD); I have a feeling that 'SD' is not supported in all QEMU's we claim to work with, though perhaps we've never checked this before when building the CLI args.

> + > + if (virQEMUCapsGet(qemuCaps, QEMU_CAPS_DEVICE_USB_STORAGE)) > + VIR_DOMAIN_CAPS_ENUM_SET(disk->bus, VIR_DOMAIN_DISK_BUS_USB); > +} > + > + > + <hostdev supported='yes'> > + <enum name='mode'/> Hmm, so that's claiming we don't support any values for the mode attribute, but we support subsys.