Re: [libvirt] [PATCH 07/13] qemu: Add hotpluging support for PCI devices on S390 guests
On Fri, 1 Jun 2018 12:43:09 +0800
Yi Min Zhao <zyimin(a)linux.ibm.com> wrote:
在 2018/5/29 下午5:04, Cornelia Huck 写道:
> On Mon, 28 May 2018 16:26:57 +0800
> Xiao Feng Ren <renxiaof(a)linux.vnet.ibm.com> wrote:
>> As the model of Shmem is not support in qemu, the rng device
doesn't
>> support MSI-X. So remove them.
> So, should qemu add support for msi-x with virtio-rng-pci in the
> future, we'd want to update libvirt as well, correct?
This is a little bit hard to answer. Our patch adds the code calling the
function which is to
attach zpci address in rng attachment function. As the logic in libvirt,
zpci could be generated
for rng device. But it actually can't be supported in s390 qemu binary.
If it is being plugged,
the error should be reported.
Thus, if we remove the attachement code for zpci in
qemuDomainAttachRNGDevice(),
we should update libvirt when rng supports msix in qemu. If we keep it
here, then update for libvirt
is not needed.
General question to the libvirt folks: is it better to just assume that
qemu can handle a device and deal with failure if it turns out that it
doesn't work after all, or to proactively fence off devices we assume
to not work? [Making virtio-rng-pci use msi-x in qemu does not look too
hard to do.]