Starting a KVM guest on s390 fails immediately. This is because "qemu --help" reports -no-acpi even for the s390(x) architecture but -no-acpi isn't supported there. Workaround is to remove QEMU_CAPS_NO_ACPI from the capability set after the version/capability extraction. Signed-off-by: Viktor Mihajlovski <mihajlov(a)linux.vnet.ibm.com&gt; --- src/qemu/qemu_capabilities.c | 5 +++++ 1 files changed, 5 insertions(+), 0 deletions(-) diff --git a/src/qemu/qemu_capabilities.c b/src/qemu/qemu_capabilities.c index 4308833..0c01cb0 100644 --- a/src/qemu/qemu_capabilities.c +++ b/src/qemu/qemu_capabilities.c @@ -1511,6 +1511,11 @@ int qemuCapsExtractVersionInfo(const char *qemu, const char *arch, qemuCapsSet(flags, QEMU_CAPS_PCI_MULTIBUS); } + /* S390 and probably other archs do not support no-acpi - + maybe the qemu option parsing should be re-thought. */ + if (STREQLEN(arch, "s390x", 5)) + qemuCapsClear(flags, QEMU_CAPS_NO_ACPI); + /* qemuCapsExtractDeviceStr will only set additional flags if qemu * understands the 0.13.0+ notion of "-device driver,". */ if (qemuCapsGet(flags, QEMU_CAPS_DEVICE) &&

On Fri, Jun 22, 2012 at 09:50:25AM +0200, Michal Privoznik wrote: > On 21.06.2012 18:15, Viktor Mihajlovski wrote: >> >> Starting a KVM guest on s390 fails immediately. This is because >> "qemu --help" reports -no-acpi even for the s390(x) architecture but >> -no-acpi isn't supported there. >> Workaround is to remove QEMU_CAPS_NO_ACPI from the capability set >> after the version/capability extraction. >> >> Signed-off-by: Viktor Mihajlovski&lt;mihajlov(a)linux.vnet.ibm.com&gt; >> --- >> src/qemu/qemu_capabilities.c | 5 +++++ >> 1 files changed, 5 insertions(+), 0 deletions(-) >> >> diff --git a/src/qemu/qemu_capabilities.c b/src/qemu/qemu_capabilities.c >> index 4308833..0c01cb0 100644 >> --- a/src/qemu/qemu_capabilities.c >> +++ b/src/qemu/qemu_capabilities.c >> @@ -1511,6 +1511,11 @@ int qemuCapsExtractVersionInfo(const char *qemu, const char *arch, >> qemuCapsSet(flags, QEMU_CAPS_PCI_MULTIBUS); >> } >> >> + /* S390 and probably other archs do not support no-acpi - >> + maybe the qemu option parsing should be re-thought. */ >> + if (STREQLEN(arch, "s390x", 5)) >> + qemuCapsClear(flags, QEMU_CAPS_NO_ACPI); >> + >> /* qemuCapsExtractDeviceStr will only set additional flags if qemu >> * understands the 0.13.0+ notion of "-device driver,". */ >> if (qemuCapsGet(flags, QEMU_CAPS_DEVICE)&& >> > > I don't think this is the right approach. I mean, if qemu's broken > shouldn't it be fixed? In principal I agree with you, but this aspect of QEMU is not likely to be changed in the forseeable future, not least because QEMU don't like us using -help in the first place. So, ACK to this patch on that basis Daniel