Re: [libvirt] [RFC] qemu: Use virtio-pci by default for mach-virt guests

(When I first saw your mail I didn't realize it was a patch, because it didn't have "PATCH" in the subject) On 10/22/2016 05:50 PM, Martin Kletzander wrote: > On Fri, Oct 21, 2016 at 07:24:28PM +0200, Andrea Bolognani wrote: >> virtio-pci is the way forward for aarch64 guests: it's faster >> and less alien to people coming from other architectures. >> Now that guest support is finally getting there, we'd like to >> start using it by default instead of virtio-mmio. >> >> Users and applications can already opt-in by explicitly using >> >> <address type='pci'/> >> >> inside the relevant elements, but that's kinda cumbersome and >> requires all users and management applications to adapt, which >> we'd really like to avoid. >> >> What we can do instead is use virtio-mmio only if the guest >> already has at least one virtio-mmio device, and use virtio-pci >> in all other situations. >> >> That means existing virtio-mmio guests will keep using the old >> addressing scheme, and new guests will automatically be created >> using virtio-pci instead. Users can still override the default >> in either direction. >> --- >> Sending this as an RFC for the time being because it clearly >> needs some more polish, but I wanted to get the idea out >> there sooner rather than later. >> > > Makes sense for the non-user of this (or rather not-yet-user maybe). So > I mention only few details inline. I like that this makes pci truly the default in a simple manner, but still allows switching back to mmio if necessary. On the other hand, it puts the potential "switch" to decide whether or not to use mmio for all devices down into the config of a single device, which is a bit weird to explain. (On the other hand, how often will mmio be used in the future? Maybe it doesn't matter if it's weird to explain...) > >> It needs to be applied on top of Laine's PCI series[1]. >> >> >> [1] >> https://www.redhat.com/archives/libvir-list/2016-October/msg00699.html >> >> src/qemu/qemu_domain_address.c | 128 >> ++++++++++++++++++++- >> ...l2argv-aarch64-virt-2.6-virtio-pci-default.args | 14 ++- >> .../qemuxml2argv-aarch64-virtio-pci-default.args | 14 ++- >> .../qemuxml2xmlout-aarch64-virtio-pci-default.xml | 24 +++- >> 4 files changed, 162 insertions(+), 18 deletions(-) >> >> diff --git a/src/qemu/qemu_domain_address.c >> b/src/qemu/qemu_domain_address.c >> index f27d1e3..7f07764 100644 >> --- a/src/qemu/qemu_domain_address.c >> +++ b/src/qemu/qemu_domain_address.c >> @@ -265,6 +265,118 @@ >> qemuDomainAssignSpaprVIOAddresses(virDomainDefPtr def, >> } >> >> >> +static bool >> +qemuDomainAnyDeviceHasAddressOfType(virDomainDefPtr def, >> + virDomainDeviceAddressType type) >> +{ >> + size_t i; >> + > > It's super-easy to miss something here, moreover it's easy to forget > adding stuff here in the future. You should either use > virDomainDeviceInfoIterate() or at least (not a fan of that) copy the > check from virDomainDeviceInfoIterateInternal() here, so that people are > forced to add new device types here. I agree with this (and I wish that the address assignment used virDomainDeviceInfoIterate() when assigning addresses for the same reasons (for brevity and to be sure new device types aren't forgotten); the problem is that the order of devices during address assignment is different, which would result in different PCI addresses for the same input XML if we were to changeit, so we're stuck with that particular extra manual enumeration of all the devices. But definitely let's not make another.)