On 09/13/2012 04:52 PM, Michal Privoznik wrote:
On 13.09.2012 16:12, Martin Kletzander wrote:
This patch adds full support for EOI setting for domains. Because this is CPU feature (flag), the model needs to be added even when it's not specified. Fortunately this problem was already solved with kvmclock, so this patch simply abuses that.
And due to the size of the patch (17 lines) I dared to include the tests. --- src/qemu/qemu_command.c | 17 +++++++++++++ .../qemuxml2argv-cpu-eoi-disabled.args | 4 ++++ .../qemuxml2argv-cpu-eoi-disabled.xml | 28 ++++++++++++++++++++++ .../qemuxml2argv-cpu-eoi-enabled.args | 4 ++++ .../qemuxml2argv-cpu-eoi-enabled.xml | 28 ++++++++++++++++++++++ .../qemuxml2argv-eoi-disabled.args | 4 ++++ .../qemuxml2argvdata/qemuxml2argv-eoi-disabled.xml | 25 +++++++++++++++++++ .../qemuxml2argvdata/qemuxml2argv-eoi-enabled.args | 4 ++++ .../qemuxml2argvdata/qemuxml2argv-eoi-enabled.xml | 25 +++++++++++++++++++ tests/qemuxml2argvtest.c | 5 ++++ tests/qemuxml2xmltest.c | 6 +++++ 11 files changed, 150 insertions(+) create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-cpu-eoi-disabled.args create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-cpu-eoi-disabled.xml create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-cpu-eoi-enabled.args create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-cpu-eoi-enabled.xml create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-eoi-disabled.args create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-eoi-disabled.xml create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-eoi-enabled.args create mode 100644 tests/qemuxml2argvdata/qemuxml2argv-eoi-enabled.xml
diff --git a/src/qemu/qemu_command.c b/src/qemu/qemu_command.c index cd4ee93..4aed8f6 100644 --- a/src/qemu/qemu_command.c +++ b/src/qemu/qemu_command.c @@ -4193,6 +4193,18 @@ qemuBuildCpuArgStr(const struct qemud_driver *driver, } }
+ if (def->apic_eoi) { + char sign; + if (def->apic_eoi == VIR_DOMAIN_APIC_EOI_ON) + sign = '+'; + else + sign = '-'; + + virBufferAsprintf(&buf, "%s,%ckvm_pv_eoi", + have_cpu ? "" : default_model, + sign); + } + if (virBufferError(&buf)) goto no_memory;
@@ -7650,6 +7662,11 @@ qemuParseCommandLineCPU(virDomainDefPtr dom, } dom->clock.timers[i]->present = present; ret = 0; + } else if (STREQ(feature, "kvm_pv_eoi")) { + if (policy == VIR_CPU_FEATURE_REQUIRE) + dom->apic_eoi = VIR_DOMAIN_APIC_EOI_ON; + else + dom->apic_eoi = VIR_DOMAIN_APIC_EOI_OFF;
Hey, this is nice. I feel like we are forgetting this part sometimes.
} else { if (!cpu) { if (!(cpu = qemuInitGuestCPU(dom)))
ACK
Michal
Thanks, pushed. Martin