[PATCH 17/24] qemucapabilitiesdata: Add (fake) '+hvf' variant of x86_64 data
Peter Krempa
pkrempa at redhat.com
Thu Mar 30 11:55:43 UTC 2023
On Thu, Mar 30, 2023 at 10:40:37 +0100, Daniel P. Berrangé wrote:
> On Tue, Mar 28, 2023 at 04:53:13PM +0200, Peter Krempa wrote:
> > On Tue, Mar 21, 2023 at 16:13:22 +0100, Martin Kletzander wrote:
> > > On Thu, Mar 16, 2023 at 05:11:46PM +0100, Peter Krempa wrote:
> > > > The data is based on the generic variant of the 7.2.0 data on x86_64.
> > > >
> > > > Only modification to the '.replies' file is that KVM is reported as
> > > > unavailable/unsupported.
> > > >
> > > > Ideally this will be replaced by a dump captured from a real system.
> > > >
> > > > Signed-off-by: Peter Krempa <pkrempa at redhat.com>
> > > > ---
> > > > .../qemu_7.2.0-hvf.x86_64+hvf.xml | 183 +
> > > > .../qemu_7.2.0-tcg.x86_64+hvf.xml | 264 +
> > > > .../caps_7.2.0_x86_64+hvf.replies | 39399 ++++++++++++++++
> > > > .../caps_7.2.0_x86_64+hvf.xml | 3250 ++
> > > > tests/qemucaps2xmloutdata/caps.x86_64+hvf.xml | 28 +
> > > > 5 files changed, 43124 insertions(+)
> > > > create mode 100644 tests/domaincapsdata/qemu_7.2.0-hvf.x86_64+hvf.xml
> > > > create mode 100644 tests/domaincapsdata/qemu_7.2.0-tcg.x86_64+hvf.xml
> > > > create mode 100644 tests/qemucapabilitiesdata/caps_7.2.0_x86_64+hvf.replies
> > > > create mode 100644 tests/qemucapabilitiesdata/caps_7.2.0_x86_64+hvf.xml
> > > > create mode 100644 tests/qemucaps2xmloutdata/caps.x86_64+hvf.xml
> > > >
> > > > diff --git a/tests/domaincapsdata/qemu_7.2.0-hvf.x86_64+hvf.xml b/tests/domaincapsdata/qemu_7.2.0-hvf.x86_64+hvf.xml
> > > > new file mode 100644
> > > > index 0000000000..d6dc93d957
> > > > --- /dev/null
> > > > +++ b/tests/domaincapsdata/qemu_7.2.0-hvf.x86_64+hvf.xml
> > > > @@ -0,0 +1,183 @@
> > > > +<domainCapabilities>
> > > > + <path>/usr/bin/qemu-system-x86_64</path>
> > > > + <domain>hvf</domain>
> > > > + <machine>pc-i440fx-7.2</machine>
> > > > + <arch>x86_64</arch>
> > >
> > > Wow, x86_64 with ...
> > >
> > > > + <iothreads supported='yes'/>
> > > > + <os supported='yes'>
> > > > + <enum name='firmware'>
> > > > + <value>bios</value>
> > > > + <value>efi</value>
> > > > + </enum>
> > > > + <loader supported='yes'>
> > > > + <value>/usr/share/AAVMF/AAVMF_CODE.fd</value>
> > > > + <value>/usr/share/AAVMF/AAVMF32_CODE.fd</value>
> > >
> > > ... AAVMF =)
> >
> > Note that this is a real caps dump we have for 7.2. If you install
> > qemu-system-aarch64 on your x86 box it comes with the firmware which we
> > display here.
>
> But this caps dump is referring to <path>/usr/bin/qemu-system-x86_64</path>,
> not qemu-system-aarch64, so the loader must be OVMF, not AAVMF.
This is an artifcat of how domaincapstest does stuff:
'fillQemuCaps' has:
if (fillStringValues(&loader->values,
"/usr/share/AAVMF/AAVMF_CODE.fd",
"/usr/share/AAVMF/AAVMF32_CODE.fd",
"/usr/share/OVMF/OVMF_CODE.fd",
NULL) < 0)
So the output files have this fake data without any consideration.
Arguably it would be a better idea to simply fill this with obviously
fake strings.
More information about the libvir-list
mailing list