On Tue, 2018-05-15 at 12:53 +0200, Lubomir Rintel wrote:
> Other "virt" type machines don't support it.
>
> Signed-off-by: Lubomir Rintel <lkundrak@xxxxx>
> ---
> src/qemu/qemu_domain.c | 1 +
> 1 file changed, 1 insertion(+)
>
> diff --git a/src/qemu/qemu_domain.c b/src/qemu/qemu_domain.c
> index 97633bfbb9..db73f45204 100644
> --- a/src/qemu/qemu_domain.c
> +++ b/src/qemu/qemu_domain.c
> @@ -3044,6 +3044,7 @@ qemuDomainDefEnableDefaultFeatures(virDomainDefPtr def,
> * was not included in the domain XML, we need to choose a suitable
> * GIC version ourselves */
> if ((def->features[VIR_DOMAIN_FEATURE_GIC] == VIR_TRISTATE_SWITCH_ABSENT &&
> + def->os.arch == VIR_ARCH_AARCH64 &&
> qemuDomainIsVirt(def)) ||
> (def->features[VIR_DOMAIN_FEATURE_GIC] == VIR_TRISTATE_SWITCH_ON &&
> def->gic_version == VIR_GIC_VERSION_NONE)) {
The check should include VIR_ARCH_ARMV7L, but that's beside the
point.
The thing is, up until now qemuDomainIsVirt() has been used to
check whether the guest uses the armv7/aarch64 virt machine type,
but now that RISC-V introduced its own virt machine type there is
a lot of potential for confusion.
IMHO the only sane way to deal with it is to rename the existing
function to qemuDomainIsArmVirt(), and introduce a new
qemuDomainIsRiscvVirt() specific to RISC-V. That will lead to a
lot of places where you have to call both, but it will also ensure
changes intended to one architecture don't end up affecting the
other.
--
Andrea Bolognani / Red Hat / Virtualization
--
libvir-list mailing list
libvir-list@xxxxxxxxxx
https://www.redhat.com/mailman/listinfo/libvir-list
