Re: [PATCH v3 12/13] conf: support manually specifying VFIO variant driver in <hostdev> XML

On Fri, Jan 05, 2024 at 03:20:15 -0500, Laine Stump wrote: > This patch makes it possible to manually specify which VFIO variant > driver to use for PCI hostdev device assignment, so that, e.g. you > could force use of a VFIO "variant" driver, with e.g. > > <driver model='mlx5_vfio_pci'/> > > or alternately to force use of the generic vfio-pci driver with > > <driver model='vfio-pci'/> > > when libvirt would have normally (after applying a subsequent patch) > found a "better match" for a device in the active kernel's > modules.alias file. (The main potential use of this manual override > would probably be to work around a bug in a new VFIO variant driver by > temporarily not using that driver). > > Signed-off-by: Laine Stump <laine(a)redhat.com&gt; > --- > > Change from V2: use new <driver model='blah'/> instead of repurposing > existing <driver name='blah'/> > > docs/formatdomain.rst | 54 +++++++++++++------ > docs/formatnetwork.rst | 21 ++++---- > src/conf/device_conf.c | 10 ++++ > src/conf/device_conf.h | 4 ++ > src/conf/domain_conf.c | 3 ++ > src/conf/network_conf.c | 2 + > src/conf/schemas/basictypes.rng | 21 +++++--- > src/conf/virnetworkportdef.c | 1 + > src/network/bridge_driver.c | 1 + > tests/networkxml2xmlin/hostdev-pf-old.xml | 8 +++ > tests/networkxml2xmlout/hostdev-pf-old.xml | 8 +++ > tests/networkxml2xmltest.c | 6 +++ > .../hostdev-vfio.x86_64-latest.args | 5 +- > tests/qemuxml2argvdata/hostdev-vfio.xml | 18 +++++++ > .../hostdev-vfio.x86_64-latest.xml | 23 +++++++- > .../plug-hostdev-pci-unmanaged.xml | 1 - > .../plug-hostdev-pci.xml | 1 - > 17 files changed, 148 insertions(+), 39 deletions(-) > create mode 100644 tests/networkxml2xmlin/hostdev-pf-old.xml > create mode 100644 tests/networkxml2xmlout/hostdev-pf-old.xml [...] > diff --git a/docs/formatnetwork.rst b/docs/formatnetwork.rst > index 5d300a035e..d4181ac029 100644 > --- a/docs/formatnetwork.rst > +++ b/docs/formatnetwork.rst > @@ -315,17 +315,14 @@ to the physical LAN (if at all). > guest, use the traditional ``<hostdev>`` device definition. :since:` Since > 0.10.0` > > - To force use of a particular type of device assignment, a <forward > - type='hostdev'> interface can have an optional ``driver`` sub-element with > - a ``name`` attribute set to either "vfio" (VFIO is a new method of device > - assignment that is compatible with UEFI Secure Boot) or "kvm" (the legacy > - device assignment handled directly by the KVM kernel module) :since:`Since > - 1.0.5 (QEMU and KVM only, requires kernel 3.6 or newer)` . When specified, > - device assignment will fail if the requested method of device assignment > - isn't available on the host. When not specified, the default is "vfio" on > - systems where the VFIO driver is available and loaded, and "kvm" on older > - systems, or those where the VFIO driver hasn't been loaded :since:`Since > - 1.1.3` (prior to that the default was always "kvm"). > + To force use of a particular device-specific VFIO driver when > + assigning the devices to a guest, a <forward type='hostdev'> > + interface can have an optional ``driver`` sub-element with a > + ``model`` attribute set to the name of the driver to use > + :since:`Since 10.0.0 (QEMU only)`. When not specified, libvirt > + will attempt to find a suitable VFIO variant driver for the > + device, and if not found it will use the generic driver > + "vfio-pci". There aren't any tests for this, the test added below just tests the old code

> > Note that this "intelligent passthrough" of network devices is very > similar to the functionality of a standard ``<hostdev>`` device, the > @@ -337,7 +334,7 @@ to the physical LAN (if at all). > to the guest domain), or if you are using a version of libvirt older than > 0.10.0, you should use a standard ``<hostdev>`` device definition in the > domain's configuration to assign the device to the guest instead of > - defining an ``<interface type='network'>`` pointing to a > + defining an ``<interface type='network'>`` pointing to a

> network with ``<forward mode='hostdev'/>``. > diff --git a/tests/networkxml2xmlin/hostdev-pf-old.xml b/tests/networkxml2xmlin/hostdev-pf-old.xml > new file mode 100644 > index 0000000000..5b8f59858c > --- /dev/null > +++ b/tests/networkxml2xmlin/hostdev-pf-old.xml > @@ -0,0 +1,8 @@ > +<network> > + <name>hostdev</name> > + <uuid>81ff0d90-c91e-6742-64da-4a736edb9a9b</uuid> > + <forward mode='hostdev' managed='yes'> > + <driver name='vfio'/> 'name'

> + <pf dev='eth2'/> > + </forward> > +</network> > diff --git a/tests/networkxml2xmlout/hostdev-pf-old.xml b/tests/networkxml2xmlout/hostdev-pf-old.xml > new file mode 100644 > index 0000000000..5b8f59858c > --- /dev/null > +++ b/tests/networkxml2xmlout/hostdev-pf-old.xml > @@ -0,0 +1,8 @@ > +<network> > + <name>hostdev</name> > + <uuid>81ff0d90-c91e-6742-64da-4a736edb9a9b</uuid> > + <forward mode='hostdev' managed='yes'> > + <driver name='vfio'/> 'name' > + <pf dev='eth2'/> > + </forward> > +</network> > diff --git a/tests/networkxml2xmltest.c b/tests/networkxml2xmltest.c > index b0814c7529..928f28b579 100644 > --- a/tests/networkxml2xmltest.c > +++ b/tests/networkxml2xmltest.c > @@ -146,6 +146,12 @@ mymain(void) > DO_TEST_FLAGS("passthrough-pf", VIR_NETWORK_XML_INACTIVE); > DO_TEST("hostdev"); > DO_TEST_FLAGS("hostdev-pf", VIR_NETWORK_XML_INACTIVE); > + > + /* libvirt pre-9.9.0 used "name='vfio'" which should be This version is no longer correct > + * automatically translated to "type='vfio'" by new parser > + */ And based on the files added above, this is not true.

> + DO_TEST_FLAGS("hostdev-pf-old", VIR_NETWORK_XML_INACTIVE); > + > DO_TEST("passthrough-address-crash"); > DO_TEST("nat-network-explicit-flood"); > DO_TEST("host-bridge-no-flood"); With the network testing stuff addressed: Reviewed-by: Peter Krempa <pkrempa(a)redhat.com&gt;