Re: [libvirt] [PATCH 04/13] qemu: Enable PCI multi bus for S390 guests

On Mon, Jun 04, 2018 at 01:57:04PM +0800, Yi Min Zhao wrote: > > > 在 2018/6/2 下午10:16, Ján Tomko 写道: >> On Thu, May 24, 2018 at 02:24:29PM +0200, Xiao Feng Ren wrote: >>> From: Yi Min Zhao <zyimin(a)linux.ibm.com&gt; >>> >>> QEMU on s390 supports PCI multibus since forever. But zPCI, as >>> extension >>> of PCI device on s390, is the significant capability. Only when zPCI >>> capability is existing, we consider QEMU supports PCI multibus >>> properly. >>> So let enable PCI multibus only if zPCI is supported. >>> >>> Signed-off-by: Yi Min Zhao <zyimin(a)linux.ibm.com&gt; >>> Reviewed-by: Boris Fiuczynski <fiuczy(a)linux.vnet.ibm.com&gt; >>> Reviewed-by: Stefan Zimmermann <stzi(a)linux.ibm.com&gt; >>> Reviewed-by: Bjoern Walk <bwalk(a)linux.vnet.ibm.com&gt; >>> --- >>> src/qemu/qemu_capabilities.c | 3 +++ >>> 1 file changed, 3 insertions(+) >>> >>> diff --git a/src/qemu/qemu_capabilities.c >>> b/src/qemu/qemu_capabilities.c >>> index e6fd7ee468..3d4f64ca09 100644 >>> --- a/src/qemu/qemu_capabilities.c >>> +++ b/src/qemu/qemu_capabilities.c >>> @@ -1716,6 +1716,9 @@ bool virQEMUCapsHasPCIMultiBus(virQEMUCapsPtr >>> qemuCaps, >>>         return false; >>>     } >>> >>> +    if (ARCH_IS_S390(def->os.arch)) >> >> This includes S390 as well as S390X, do we care about that distinction? > No need, at least from my point of view. On second thought, if this is the only way to get PCI to work on S390, this particular function should not be capability-based but always return true - it merely says whether to use pci.0 vs. pci for the 0th bus, not whether PCI is actually supported or not. Jano