On Mon, Nov 27, 2023 at 04:55:16PM +0800, Zhenzhong Duan wrote:
TDX guest requires some special parameters to boot, They are:
"-machine pc-q35-*" "kernel_irqchip=split"
Signed-off-by: Zhenzhong Duan <zhenzhong.duan@intel.com> --- src/qemu/qemu_validate.c | 10 ++++++++++ 1 file changed, 10 insertions(+)
diff --git a/src/qemu/qemu_validate.c b/src/qemu/qemu_validate.c index 5a9173e8ff..c4f386fe99 100644 --- a/src/qemu/qemu_validate.c +++ b/src/qemu/qemu_validate.c @@ -1329,6 +1329,16 @@ qemuValidateDomainDef(const virDomainDef *def, _("INTEL TDX launch security is not supported with this QEMU binary")); return -1; } + if (!qemuDomainIsQ35(def)) { + virReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s", + _("Intel TDX is supported with q35 machine types only")); + return -1; + }
Ideally QMP 'MachineInfo' struct would report whether TDX is supported so we don't need to hardcode that.
+ if (def->features[VIR_DOMAIN_FEATURE_IOAPIC] != VIR_DOMAIN_IOAPIC_QEMU) { + virReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s", + _("INTEL TDX launch security needs split kernel irqchip"));
s/INTEL/Intel/ Ideally QEMU would automatically use the correct ioapic impl when no args are given to QEMU. That would let us do if (def->features[VIR_DOMAIN_FEATURE_IOAPIC] == VIR_DOMAIN_IOAPIC_KVM) { thus allowing IOAPIC_NONE (ie QEMU's default) or IOAPIC_QEMU (explicitly requested config). This will make TDX guest "just work" in more scenarios.
+ return -1; + } break; case VIR_DOMAIN_LAUNCH_SECURITY_NONE: case VIR_DOMAIN_LAUNCH_SECURITY_LAST: -- 2.34.1
With regards, Daniel -- |: https://berrange.com -o- https://www.flickr.com/photos/dberrange :| |: https://libvirt.org -o- https://fstop138.berrange.com :| |: https://entangle-photo.org -o- https://www.instagram.com/dberrange :|