The code is entirely correct, but it still managed to trip me
up when I first ran into it because I did not realize right away
that VIR_PCI_CONNECT_TYPES_ENDPOINT was not a single flag, but
rather a mask including both VIR_PCI_CONNECT_TYPE_PCI_DEVICE and
VIR_PCI_CONNECT_TYPE_PCIE_DEVICE.
In order to save the next distracted traveler in PCI Address Land
some time, document this fact with a comment. Add a test case for
the behavior as well.
It was missing... Also since I'm using the soft link from qemuxml2xmloutdata
to the qemuxml2argvdata file, modify the output file to have the necessary
<address> elements plus the mouse and keyboard.
Signed-off-by: John Ferlan <jferlan@redhat.com>
The x86 CPU driver translated each CPU definition from domain XML into
CPUID data and then back to CPU definition. This effectively sorted the
list of CPU features according to their CPUID values. Since this is
going to change, we need to reorder CPU features in a few test files to
make sure the generated QEMU command lines will not change.
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
Most of QEMU's PCI display device models, such as:
libvirt video/model/@type QEMU -device
------------------------- ------------
cirrus cirrus-vga
vga VGA
qxl qxl-vga
virtio virtio-vga
come with a linear framebuffer (sometimes called "VGA compatibility
framebuffer"). This linear framebuffer lives in one of the PCI device's
MMIO BARs, and allows guest code (primarily: firmware drivers, and
non-accelerated OS drivers) to display graphics with direct memory access.
Due to architectural reasons on aarch64/KVM hosts, this kind of
framebuffer doesn't / can't work in
qemu-system-(arm|aarch64) -M virt
machines. Cache coherency issues guarantee a corrupted / unusable display.
The problem has been researched by several people, including kvm-arm
maintainers, and it's been decided that the best way (practically the only
way) to have boot time graphics for such guests is to consolidate on
QEMU's "virtio-gpu-pci" device.
>From <https://bugzilla.redhat.com/show_bug.cgi?id=1195176>, libvirt
supports
<devices>
<video>
<model type='virtio'/>
</video>
</devices>
but libvirt unconditionally maps @type='virtio' to QEMU's "virtio-vga"
device model. (See the qemuBuildDeviceVideoStr() function and the
"qemuDeviceVideo" enum impl.)
According to the above, this is not right for the "virt" machine type; the
qemu-system-(arm|aarch64) binaries don't even recognize the "virtio-vga"
device model (justifiedly). Whereas "virtio-gpu-pci", which is a pure
virtio device without a compatibility framebuffer, is available, and works
fine.
(The ArmVirtQemu ("AAVMF") platform of edk2 -- that is, the UEFI firmware
for "virt" -- supports "virtio-gpu-pci", as of upstream commit
3ef3209d3028. See
<https://tianocore.acgmultimedia.com/show_bug.cgi?id=66>.)
Override the default mapping of "virtio", from "virtio-vga" to
"virtio-gpu-pci", if qemuDomainMachineIsVirt() evaluates to true.
Cc: Andrea Bolognani <abologna@redhat.com>
Cc: Drew Jones <drjones@redhat.com>
Cc: Marc-André Lureau <marcandre.lureau@redhat.com>
Cc: Martin Kletzander <mkletzan@redhat.com>
Suggested-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1372901
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Acked-by: Martin Kletzander <mkletzan@redhat.com>
Add a new TLS X.509 certificate type - "chardev". This will handle the
creation of a TLS certificate capability (and possibly repository) for
properly configured character device TCP backends.
Unlike the vnc and spice there is no "listen" or "passwd" associated. The
credentials eventually will be handled via a libvirt secret provided to
a specific backend.
Make use of the default verify option as well.
Signed-off-by: John Ferlan <jferlan@redhat.com>
When the user doesn't specify any model for a USB controller,
we use an architecture-dependent default, but we don't reflect
it in the guest XML.
Pick the default USB controller model when parsing the guest
XML instead of when creating the QEMU command line, so that
our choice is saved back to disk.
https://bugzilla.redhat.com/show_bug.cgi?id=1356937
Add the definitions to allow for viewing/setting cgroup period and quota
limits for IOThreads.
This is similar to the work done for emulator quota and period by
commit ids 'b65dafa' and 'e051c482'.
Being able to view/set the IOThread specific values is related to more
recent changes adding global period (commmit id '4d92d58f') and global
quota (commit id '55ecdae') definitions and qemu support (commit id
'4e17ff79' and 'fbcbd1b2'). With a global setting though, if somehow
the IOThread value in the cgroup hierarchy was set "outside of libvirt"
to a value that is incompatible with the global value.
Allowing control over IOThread specific values provides the capability
to alter the IOThread values as necessary.
To allow using failover with gluster it's necessary to specify multiple
volume hosts. Add support for starting qemu with such configurations.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
We were requiring a USB port path in the schema, but not enforcing it.
Omitting the USB port would lead to libvirt formatting it as (null).
Such domain cannot be started and will disappear after libvirtd restart
(since it cannot parse back the XML).
Only format the port if it has been specified and mark it as optional
in the XML schema.
Commit id's '9bbf0d7e6' and '2552fec24' added some XML parsing tests
for a LUKS volume to use a 'passphrase' secret format. After commit,
this was deemed to be incorrect, so covert the various tests to use
the volume usage format where the 'usage' is the path to the volume
rather than a user defined name string.
Also, removed the qemuxml2argv-luks-disk-cipher.xml since it was
just a duplicate of qemuxml2argv-luks-disks.xml.
Signed-off-by: John Ferlan <jferlan@redhat.com>
For type='ethernet' interfaces only.
(This patch had been pushed earlier in
commit 0b4645a7e0, but was reverted in
commit 84d47a3cce because it had been
accidentally pushed during the freeze for release 2.0.0)
In order to use more common code and set up for a future type, modify the
encryption secret to allow the "usage" attribute or the "uuid" attribute
to define the secret. The "usage" in the case of a volume secret would be
the path to the volume as dictated by the backwards compatibility brought
on by virStorageGenerateQcowEncryption where it set up the usage field as
the vol->target.path and didn't allow someone to provide it. This carries
into virSecretObjListFindByUsageLocked which takes the secret usage attribute
value from from the domain disk definition and compares it against the
usage type from the secret definition. Since none of the code dealing
with qcow/qcow2 encryption secrets uses usage for lookup, it's a mostly
cosmetic change. The real usage comes in a future path where the encryption
is expanded to be a luks volume and the secret will allow definition of
the usage field.
This code will make use of the virSecretLookup{Parse|Format}Secret common code.
Signed-off-by: John Ferlan <jferlan@redhat.com>
Make them work again... The xml2xml had been working, but the xml2argv
were not working. Making the xml2argv work required a few adjustments to
the xml to update to more recent times.
Signed-off-by: John Ferlan <jferlan@redhat.com>
There has been some progress lately in enabling virtio-pci on
aarch64 guests; however, guest OS support is still spotty at best,
so most guests are going to be using virtio-mmio instead.
Currently, mach-virt guests are closely modeled after q35 guests,
and that includes always adding a dmi-to-pci-bridge that's just
impossible to get rid of. While that's acceptable (if suboptimal)
for q35, where you will always need some kind of PCI device anyway,
mach-virt guests should be allowed to avoid it.
Until now, a Q35 domain (or arm/virt, or any other domain that has a
pcie-root bus) would always have a pci-bridge added, so that there
would be a hotpluggable standard PCI slot available to plug in any PCI
devices that might be added. This patch removes the explicit add,
instead relying on the pci-bridge being auto-added during PCI address
assignment (it will add a pci-bridge if there are no free slots).
This doesn't eliminate the dmi-to-pci-bridge controller that is
explicitly added whether or not a standard PCI slot is required (and
that is almost never used as anything other than a converter between
pcie.0's PCIe slots and standard PCI). That will be done separately.
This option allows or disallows detection of zero-writes if it is set to
"on" or "off", respectively. It can be also set to "unmap" in which
case it will try discarding that part of image based on the value of the
"discard" option.
Signed-off-by: Martin Kletzander <mkletzan@redhat.com>
This new listen type is currently supported only by spice graphics.
It's introduced to make it easier and clearer specify to not listen
anywhere in order to start a guest with OpenGL support.
The old way to do this was set spice graphics autoport='no' and don't
specify any ports. The new way is to use <listen type='none'/>. In
order to be able to migrate to old libvirt the migratable XML will be
generated without the listen element and with autoport='no'. Also the
old configuration will be automatically converted to the this listen
type.
Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
VNC graphics already supports sockets but only via 'socket' attribute.
This patch coverts that attribute into listen type 'socket'.
For backward compatibility we need to handle listen type 'socket' and 'socket'
attribute properly to support old XMLs and new XMLs. If both are provided they
have to match, if only one of them is provided we need to be able to parse that
configuration too.
To not break migration back to old libvirt if the socket is provided by user we
need to generate migratable XML without the listen element and use only 'socket'
attribute.
Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
Historically, we added heads=1 to videos, but for example for qxl, we
did not reflect that on the command line.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1283207
Signed-off-by: Martin Kletzander <mkletzan@redhat.com>
Hand-entering indexes for 20 PCI controllers is not as tedious as
manually determining and entering their PCI addresses, but it's still
annoying, and the algorithm for determining the proper index is
incredibly simple (in all cases except one) - just pick the lowest
unused index.
The one exception is USB2 controllers because multiple controllers in
the same group have the same index. For these we look to see if 1) the
most recently added USB controller is also a USB2 controller, and 2)
the group *that* controller belongs to doesn't yet have a controller
of the exact model we're just now adding - if both are true, the new
controller gets the same index, but in all other cases we just assign
the lowest unused index.
With this patch in place and combined with the automatic PCI address
assignment, we can define a PCIe switch with several ports like this:
<controller type='pci' model='pcie-root-port'/>
<controller type='pci' model='pcie-switch-upstream-port'/>
<controller type='pci' model='pcie-switch-downstream-port'/>
<controller type='pci' model='pcie-switch-downstream-port'/>
<controller type='pci' model='pcie-switch-downstream-port'/>
<controller type='pci' model='pcie-switch-downstream-port'/>
<controller type='pci' model='pcie-switch-downstream-port'/>
...
These will each get a unique index, and PCI addresses that connect
them together appropriately with no pesky numbers required.
Add a new element to <domain> XML:
<os>
<acpi>
<table type="slic">/path/to/acpi/table/file</table>
</acpi>
</os>
To supply a path to a SLIC (Software Licensing) ACPI
table blob.
https://bugzilla.redhat.com/show_bug.cgi?id=1327537
Rather than only assigning a PCI address when no address is given at
all, also do it when the config says that the address type is 'pci',
but it gives no address (virDeviceInfoPCIAddressWanted()).
There are also several places after parsing but prior to address
assignment where code previously expected that any info with address
type='pci' would have a *valid* PCI address, which isn't always the
case - now we check not only for type='pci', but also for a valid
address (virDeviceInfoPCIAddressPresent()).
The test case added in this patch was directly copied from Cole's patch titled:
qemu: Wire up address type=pci auto_allocate
Commit 55320c23 introduced a new test for VNC to test if
vnc_auto_unix_socket is set in qemu.conf, but forget to enable it in
qemuxml2argvtest.c.
This patch also moves the code in qemuxml2xmltest.c next to other VNC
tests and refactor the test so we also check the case for parsing active
XML.
Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
The only case where the hardware capabilities influence the result
is when no <gic/> element was provided.
The test programs now ensure both that the correct GIC version is
picked in that case, and that hardware capabilities are not taken
into account when the user has already picked a GIC version.
We support omitting listen attribute of graphics element so we should
also support omitting address attribute of listen element. This patch
also updates libvirt to always add a listen element into domain XML
except for VNC graphics if socket attribute is specified.
Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
Add the ability to add an 'iothread' to the controller which will be how
virtio-scsi-pci and virtio-scsi-ccw iothreads have been implemented in qemu.
Describe the new functionality and add tests to parse/validate that the
new attribute can be added.
Similarly to what commit 7140807917 did with some internal paths,
clear vnc socket paths that were generated by us. Having such path in
the definition can cause trouble when restoring the domain. The path is
generated to the per-domain directory that contains the domain ID.
However, that ID will be different upon restoration, so qemu won't be
able to create that socket because the directory will not be prepared.
To be able to migrate to older libvirt, skip formatting the socket path
in migratable XML if it was autogenerated. And mark it as autogenerated
if it already exists and we're parsing live XML.
Best viewed with '-C'.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1326270
Signed-off-by: Martin Kletzander <mkletzan@redhat.com>
Currently we only allow /dev/random and /dev/hwrng as host input
for <rng><backend model='random'/> device. This was added after
various upstream discussions in commit 4932ef45
However this restriction has generated quite a few complaints over
the years, so a new discussion was initiated:
http://www.redhat.com/archives/libvir-list/2016-April/msg00987.html
Several people suggested removing the restriction, and nobody really
spoke up to defend it. So this patch drops the path restriction
entirely
https://bugzilla.redhat.com/show_bug.cgi?id=1074464
While working on the tests for the secret initialization vector, I found
that the existing iSCSI tests were lacking in how they defined the IQN.
Many had IQN's of just 'iqn.1992-01.com.example' for one disk while using
'iqn.1992-01.com.example/1' for the second disk (same for hostdevs - guess
how they were copied/generated).
Typically (and documented this way), IQN's would include be of the form
'iqn.1992-01.com.example:storage/1' indicating an IQN using "storage" for
naming authority specific string and "/1" for the iSCSI LUN.
So modify the input XML's to use the more proper format - this of course
has a ripple effect on the output XML and the args.
Also note that the "%3A" is generated by the virURIFormat/xmlSaveUri
to represent the colon.
Signed-off-by: John Ferlan <jferlan@redhat.com>
This controller provides a single PCIe port on a new root. It is
similar to pci-expander-bus, intended to provide a bus that can be
associated with a guest-identifiable NUMA node, but is for
machinetypes with PCIe rather than PCI (e.g. q35-based machinetypes).
Aside from PCIe vs. PCI, the other main difference is that a
pci-expander-bus has a companion pci-bridge that is automatically
attached along with it, but pcie-expander-bus has only a single port,
and that port will only connect to a pcie-root-port, or to a
pcie-switch-upstream-port. In order for the bus to be of any use in
the guest, it must have either a pcie-root-port or a
pcie-switch-upstream-port attached (and one or more
pcie-switch-downstream-ports attached to the
pcie-switch-upstream-port).
This is backed by the qemu device "pxb".
The pxb device always includes a pci-bridge that is at the bus number
of the pxb + 1.
busNr and <node> from the <target> subelement are used to set the
bus_nr and numa_node options for pxb.
During post-parse we validate that the domain's machinetype is
440fx-based (since the pxb device only works on 440fx-based machines),
and <node> also gets a sanity check to assure that the NUMA node
specified for the pxb (if any - it's optional) actually exists on the
guest.
This is a standard PCI root bus (not a bridge) that can be added to a
440fx-based domain. Although it uses a PCI slot, this is *not* how it
is connected into the PCI bus hierarchy, but is only used for
control. Each pci-expander-bus provides 32 slots (0-31) that can
accept hotplug of standard PCI devices.
The usefulness of pci-expander-bus relative to a pci-bridge is that
the NUMA node of the bus can be specified with the <node> subelement
of <target>. This gives guest-side visibility to the NUMA node of
attached devices (presuming that management apps only assign a device
to a bus that has a NUMA node number matching the node number of the
device on the host).
Each pci-expander-bus also has a "busNr" attribute. The expander-bus
itself will take the busNr specified, and all buses that are connected
to this bus (including the pci-bridge that is automatically added to
any expander bus of model "pxb" (see the next commit)) will use
busNr+1, busNr+2, etc, and the pci-root (or the expander-bus with next
lower busNr) will use bus numbers lower than busNr.
When support for dmi-to-pci-bridge was added, it was assumed that,
just as with the pci-root bus, slot 0 was reserved. This is not the
case - it can be used to connect a device just like any other slot, so
remove the restriction and update the test cases that auto-assign an
address on a dmi-to-pci-bridge.
Commit dc98a5bc refactored the code a lot and forget about checking if
listen attribute is specified. This ensures that listen attribute and
first listen element are compared only if both exist.
Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
This patch adds support for "vpindex", "runtime", "synic",
"stimer", and "vendor_id" features available in qemu 2.5+.
- When Hyper-V "vpindex" is on, guest can use MSR HV_X64_MSR_VP_INDEX
to get virtual processor ID.
- Hyper-V "runtime" enlightement feature allows to use MSR
HV_X64_MSR_VP_RUNTIME to get the time the virtual processor consumes
running guest code, as well as the time the hypervisor spends running
code on behalf of that guest.
- Hyper-V "synic" stands for Synthetic Interrupt Controller, which is
lapic extension controlled via MSRs.
- Hyper-V "stimer" switches on Hyper-V SynIC timers MSR's support.
Guest can setup and use fired by host events (SynIC interrupt and
appropriate timer expiration message) as guest clock events
- Hyper-V "reset" allows guest to reset VM.
- Hyper-V "vendor_id" exposes hypervisor vendor id to guest.
Signed-off-by: Nikolay Shirokovskiy <nshirokovskiy@virtuozzo.com>
Signed-off-by: John Ferlan <jferlan@redhat.com>
We always place primary video device at first place, to make it easier
to create a qemu command or format an xml, but we should also set the
primary boolean for primary video device to 'true'.
Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
Add Spice graphics gl attribute. qemu 2.6 should have -spice gl=on argument to
enable opengl rendering context (patches on the ML). This is necessary to
actually enable virgl rendering.
Add a qemuxml2argv test for virtio-gpu + spice with virgl.
Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
This does nothing more than adding the new device and capability.
The device is present since QEMU 2.6.0.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Test all kinds of scenarios, including guests asking for GIC but
failing to specify a version, guests specifying an invalid version
and guests trying to use GIC with non-virt or even non-ARM machines.
Unify the naming to prepare for new test cases that will be added
later on.
Convert a couple of output XML files for the qemuxml2xml test to
symlinks while at it, since they were identical to the corresponding
input XML files anyways.
Moreover, since we're only interested in testing GIC support here,
simplify XML files by getting rid of the unrelevant bits.
We use the PreFormat callback for this. Many test cases need to be extended
to pass in proper qemuCaps flags so AssignAddresses doesn't throw errors.
One test case (pcie-root-port-too-many) is dropped, since it was meant
only for checking an error condition in qemuxml2argv, and one we add in
AssignAddresses it errors here too.
Long term I think AssignAddresses should be handled in qemu's PostParse
callback, but that's not entirely straightforward. Handling it here
means we can get the test suite churn over with.
When we unconditionally enable QEMU_CAPS_DEVICE, these tests need
some massaging, so do it ahead of time to not mix it in with the
big test refresh.
- minimal-s390 is not a real world working config, so drop it
- disk-usb was testing for an old code path that will be removed.
instead use it to test lack of USB disk support, and rename it
to disk-usb-nosupport. Switch xml2xml to use disk-usb-device for
input.
- cputune-numatune was needlessly using q35, switch it to an older
machine type
Most qemuxml2xml tests expect that the input XML is unchanged after
parsing. This is unlike 99% of new qemu configs in the wild, which after
initial parsing end up with stable PCI device addresses. The xml2xml bit
doesn't currently hit that code path though, so most XML testing indeed
does not change.
Future patches will add that PCI address bits, which means most test cases
will have different output. So let's do away with the hardcoded same vs
different test split, and always track a separate output file. Tests can
still have same input and output, it just necessitates 2 separate XML files.
The current code was a little bit odd. At first we've removed all
possible implicit input devices from domain definition to add them later
back if there was any graphics device defined while parsing XML
description. That's not all, while formating domain definition to XML
description we at first ignore any input devices with bus different to
USB and VIRTIO and few lines later we add implicit input devices to XML.
This seems to me as a lot of code for nothing. This patch may look
to be more complicated than original approach, but this is a preferred
way to modify/add driver specific stuff only in those drivers and not
deal with them in common parsing/formating functions.
The update is to add those implicit input devices into config XML to
follow the real HW configuration visible by guest OS.
There was also inconsistence between our behavior and QEMU's in the way,
that in QEMU there is no way how to disable those implicit input devices
for x86 architecture and they are available always, even without graphics
device. This applies also to XEN hypervisor. VZ driver already does its
part by putting correct implicit devices into live XML.
Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
A future patch will refactor the storage of the pinning information in a
way where the ordering will be lost. Order them numerically to avoid
changing the tests later.
The real Q35 machine puts the first USB controller set (EHCI+(UHCIx4))
on bus 0 slot 0x1D, and the 2nd USB controller set on bus 0 slot 0x1A,
so let's attempt to make the virtual machine match that for
controllers with auto-assigned addresses when possible.
Three test cases were added to assure that the proper addresses are
assigned - one with a single set of unaddressed USB controllers, one
with 3 (to grab both preferred slots plus one more), and one with the
order of the controller definitions reordered, to assure that the
auto-assignment isn't mixed up by order.
Panic device type used depends on 'model' attribute.
If no model is specified then device type depends on hypervisor
and guest arch. 'pseries' model is used for pSeries guest and
'isa' model is used in other cases.
XML:
<devices>
<panic model='hyperv'/>
</devices>
QEMU command line:
qemu -cpu <cpu_model>,hv_crash
USB controllers can share the same 'index' which indicates, that there
is some sort of master-companion relationship. Reorder the controllers
in XML in to place the master controller before its companions. This is
required by QEMU to not fail with error message:
error: internal error: process exited while connecting to monitor:
2015-10-26T16:25:17.630265Z qemu-system-x86_64:
-device ich9-usb-uhci1,masterbus=usb.0,firstport=0,bus=pci.0,multifunction=on,addr=0x6:
USB bus 'usb.0' not found
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1166452
Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
https://bugzilla.redhat.com/show_bug.cgi?id=1260846
Introduced by 8fedbbdb, if we parse an unordered NUMA cell, will
get a segfault. This is because of a check for overlapping @cpus
sets we have there. However, since the array to hold guest NUMA
cells is allocated upfront and therefore it contains all zeros,
an out of order cell will break our assumption that cell IDs have
increasing character. At this point we try to access yet NULL
bitmap and therefore segfault.
Signed-off-by: Luyao Huang <lhuang@redhat.com>
There are some configuration options to some types of pci controllers
that are currently automatically derived from other parts of the
controller's configuration. For example, in qemu a pci-bridge
controller has an option that is called "chassis_nr"; up until now
libvirt has always set chassis_nr to the index of the pci-bridge. So
this:
<controller type='pci' model='pci-bridge' index='2'/>
will always result in:
-device pci-bridge,chassis_nr=2,...
on the qemu commandline. In the future we may decide there is a better
way to derive that option, but even in that case we will need for
existing domains to retain the same chassis_nr they were using in the
past - that is something that is visible to the guest so it is part of
the guest ABI and changing it would lead to problems for migrating
guests (or just guests with very picky OSes).
The <target> subelement has been added as a place to put the new
"chassisNr" attribute that will be filled in by libvirt when it
auto-generates the chassisNr; it will be saved in the config, then
reused any time the domain is started:
<controller type='pci' model='pci-bridge' index='2'>
<model type='pci-bridge'/>
<target chassisNr='2'/>
</controller>
The one oddity of all this is that if the controller configuration
is changed (for example to change the index or the pci address
where the controller is plugged in), the items in <target> will
*not* be re-generated, which might lead to conflict. I can't
really see any way around this, but fortunately if there is a
material conflict qemu will let us know and we will pass that on
to the user.
This new subelement is used in PCI controllers: the toplevel
*attribute* "model" of a controller denotes what kind of PCI
controller is being described, e.g. a "dmi-to-pci-bridge",
"pci-bridge", or "pci-root". But in the future there will be different
implementations of some of those types of PCI controllers, which
behave similarly from libvirt's point of view (and so should have the
same model), but use a different device in qemu (and present
themselves as a different piece of hardware in the guest). In an ideal
world we (i.e. "I") would have thought of that back when the pci
controllers were added, and used some sort of type/class/model
notation (where class was used in the way we are now using model, and
model was used for the actual manufacturer's model number of a
particular family of PCI controller), but that opportunity is long
past, so as an alternative, this patch allows selecting a particular
implementation of a pci controller with the "name" attribute of the
<model> subelement, e.g.:
<controller type='pci' model='dmi-to-pci-bridge' index='1'>
<model name='i82801b11-bridge'/>
</controller>
In this case, "dmi-to-pci-bridge" is the kind of controller (one that
has a single PCIe port upstream, and 32 standard PCI ports downstream,
which are not hotpluggable), and the qemu device to be used to
implement this kind of controller is named "i82801b11-bridge".
Implementing the above now will allow us in the future to add a new
kind of dmi-to-pci-bridge that doesn't use qemu's i82801b11-bridge
device, but instead uses something else (which doesn't yet exist, but
qemu people have been discussing it), all without breaking existing
configs.
(note that for the existing "pci-bridge" type of PCI controller, both
the model attribute and <model> name are 'pci-bridge'. This is just a
coincidence, since it turns out that in this case the device name in
qemu really is a generic 'pci-bridge' rather than being the name of
some real-world chip)
Multi != One. And indeed, libvirt behaves the same way for queues='1'
as without such setting. Let's make it clear in the XML.
Signed-off-by: Martin Kletzander <mkletzan@redhat.com>
We have been formatting the first serial device also
as a console device, but only if there were no other consoles.
If there is a <serial> device present in the XML, but no serial
<console>, or if there isn't any <console> at all but the domain
definition hasn't gone through a parse->format->parse round-trip,
the <console> device would not be formatted.
Change the code to always add the stub device for the first
serial device.
Fixes https://bugzilla.redhat.com/show_bug.cgi?id=1089914
The guest firmware provides the same functionality as the pvpanic
device, and the relevant element should always be present in the
domain XML to reflect this fact, so add it after parsing the
definition if it wasn't there already.
We have previously effectively ignored all <controller type='ide'>
elements in a domain definition.
On the i440fx-based machinetypes there is an IDE controller that is
included in the chipset and can't be removed (which is the ide
controller with index='0'>), so it makes sense to ignore that one
controller. However, if an i440fx domain definition has a 2nd
controller, nothing catches this error (unless you also have a disk
attached to it, in which case qemu will complain that you're trying to
use the ide controller named "ide1", which doesn't exist), and if any
other type of domain has even a single controller defined, it will be
incorrectly ignored.
Ignoring a bogus controller definition isn't such a big problem, as
long as an error is logged when any disk is attached to that
non-existent controller. But in the case of q35-based machinetypes,
the hardcoded id ("alias" in libvirt terms) of its builtin SATA
controller is "ide", which happens to be the same id as the builtin
IDE controller on i440fx machinetypes. So libvirt creates a
commandline believing that it is connecting the disk to the builtin
(but actually nonexistent) IDE controller, qemu thinks that libvirt
wanted that disk connected to the builtin SATA controller, and
everybody is happy.
Until you try to connect a 2nd disk to the IDE controller. Then qemu
will complain that you're trying to set unit=1 on a controller that
requires unit=0 (SATA controllers are organized differently than IDE
controllers).
After this patch, if a domain has an IDE controller defined for a
machinetype that has no IDE controllers, libvirt will log an error
about the controller itself as it is building the qemu commandline
(rather than a (possible) error from qemu about disks attached to that
controller). This is done by adding IDE to the list of controller
types that are handled in the loop that creates controller command
strings in qemuBuildCommandline() (previously it would *always* skip
IDE controllers). Then qemuBuildControllerDevStr() is modified to log
an appropriate error in the case of IDE controllers.
In the future, if we add support for extra IDE controllers (piix3-ide
and/or piix4-ide) we can just add it into the IDE case in
qemuBuildControllerDevStr(). For now, nobody seems anxious to add
extra support for an aging and very slow controller, when there are so
many better options available.
Resolves:
https://bugzilla.redhat.com/show_bug.cgi?id=1176071 (Fedora)
My commit 747761a79 (v1.2.15 only) dropped this bit of logic when filling
in a default arch in the XML:
- /* First try to find one matching host arch */
- for (i = 0; i < caps->nguests; i++) {
- if (caps->guests[i]->ostype == ostype) {
- for (j = 0; j < caps->guests[i]->arch.ndomains; j++) {
- if (caps->guests[i]->arch.domains[j]->type == domain &&
- caps->guests[i]->arch.id == caps->host.arch)
- return caps->guests[i]->arch.id;
- }
- }
- }
That attempt to match host.arch is important, otherwise we end up
defaulting to i686 on x86_64 host for KVM, which is not intended.
Duplicate it in the centralized CapsLookup function.
Additionally add some testcases that would have caught this.
https://bugzilla.redhat.com/show_bug.cgi?id=1219191
My commit 7b9de914 added some aarch64 CPU test cases. I wanted to test
two different code paths but inadvertently added two of the same test
cases.
The second code path (using <cpu><model>host</model</cpu>) isn't easily
exercised via the qemu tests anyways, I'll need to look elsewhere.
Regardless, remove the redundant tests for now
Remove the iothreadspin array from cputune and replace with a cpumask
to be stored in the iothreadids list.
Adjust the test output because our printing goes in order of the iothreadids
list now.
- Make sure aarch64 host-passthrough works correctly
- Make sure libvirt doesn't choke on cpu model=host, which is what
virt-install/virt-manager were incorrectly specifying up until recently.
Instead of always using controller 0 and incrementing port number,
respect the maximum port numbers of controllers and use all of them.
Ports for virtio consoles are quietly reserved, but not formatted
(neither in XML nor on QEMU command line).
Also rejects duplicate virtio-serial addresses.
https://bugzilla.redhat.com/show_bug.cgi?id=890606https://bugzilla.redhat.com/show_bug.cgi?id=1076708
Test changes:
* virtio-auto.args
Filling out the port when just the controller is specified.
switched from using
maxport + 1
to:
first free port on the controller
* virtio-autoassign.args
Filling out the address when no <address> is specified.
Started using all the controllers instead of 0, also discards
the bus value.
* xml -> xml output of virtio-auto
The port assignment is no longer done as a part of XML parsing,
so the unspecified values stay 0.
All the devices we have format their address as its last sub-element, so
let's change memballoon to follow suit. Also adjust RNG to allow any
order of them so 'virsh edit' doesn't shout at us.
Signed-off-by: Martin Kletzander <mkletzan@redhat.com>
Now that the size of guest's memory can be inferred from the NUMA
configuration (if present) make it optional to specify <memory>
explicitly.
To make sure that memory is specified add a check that some form of
memory size was specified. One side effect of this change is that it is
no longer possible to specify 0KiB as memory size for the VM, but I
don't think it would be any useful to do so. (I can imagine embedded
systems without memory, just registers, but that's far from what libvirt
is usually doing).
Forbidding 0 memory for guests also fixes a few corner cases where 0 was
not interpreted correctly and caused failures. (Arguments for numad when
using automatic placement, size of the balloon). This fixes problems
described in https://bugzilla.redhat.com/show_bug.cgi?id=1161461
Test case changes are added to verify that the schema change and code
behave correctly.
We don't usually do tests purely for one change, but one change was
special because when users will migrate to OVMF/AAVMF, commit 18f9f69b
makes their lives easier by allowing them to interleave <type/> inside
<os/>. It would be nice of us to keep the possibility of them pasting
the loader and nvram elements wherever it is valid, hence this test.
Signed-off-by: Martin Kletzander <mkletzan@redhat.com>
There was a mess in the way how we store unlimited value for memory
limits and how we handled values provided by user. Internally there
were two possible ways how to store unlimited value: as 0 value or as
VIR_DOMAIN_MEMORY_PARAM_UNLIMITED. Because we chose to store memory
limits as unsigned long long, we cannot use -1 to represent unlimited.
It's much easier for us to say that everything greater than
VIR_DOMAIN_MEMORY_PARAM_UNLIMITED means unlimited and leave 0 as valid
value despite that it makes no sense to set limit to 0.
Remove unnecessary function virCompareLimitUlong. The update of test
is to prevent the 0 to be miss-used as unlimited in future.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1146539
Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
Since the APIs support just one element per namespace and while
modifying an element all duplicates would be removed, let's do this
right away in the post parse callback.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1190590
https://bugzilla.redhat.com/show_bug.cgi?id=1142631
This patch resolves a situation where the same "<target dev='$name'...>"
can be used for multiple disks in the domain.
While the $name is "mostly" advisory regarding the expected order that
the disk is added to the domain and not guaranteed to map to the device
name in the guest OS, it still should be unique enough such that other
domblk* type operations can be performed.
Without the patch, the domblklist will list the same Target twice:
$ virsh domblklist $dom
Target Source
------------------------------------------------
sda /var/lib/libvirt/images/file.qcow2
sda /var/lib/libvirt/images/file.img
Additionally, getting domblkstat, domblkerror, domblkinfo, and other block*
type calls will not be able to reference the second target.
Fortunately, hotplug disallows adding a "third" sda value:
$ qemu-img create -f raw /var/lib/libvirt/images/file2.img 10M
$ virsh attach-disk $dom /var/lib/libvirt/images/file2.img sda
error: Failed to attach disk
error: operation failed: target sda already exists
$
BUT, it since 'sdb' doesn't exist one would get the following on the same
hotplug attempt, but changing to use 'sdb' instead of 'sda'
$ virsh attach-disk $dom /var/lib/libvirt/images/file2.img sdb
error: Failed to attach disk
error: internal error: unable to execute QEMU command 'device_add': Duplicate ID 'scsi0-0-1' for device
$
Since we cannot fix this issue at parsing time, the best that can be done so
as to not "lose" a domain is to make the check prior to starting the guest
with the results as follows:
$ virsh start $dom
error: Failed to start domain $dom
error: XML error: target 'sda' duplicated for disk sources '/var/lib/libvirt/images/file.qcow2' and '/var/lib/libvirt/images/file.img'
$
Running 'make check' found a few more instances in the tests where this
duplicated target dev value was being used. These also exhibited some
duplicated 'id=' values (negating the uniqueness argument of aliases) in
the corresponding .args file and of course the *xmlout version of a few
input XML files.
Since s390 does not support usb the default creation of a usb controller
for a domain should not occur.
Also adjust s390 test cases by removing usb device instances since
usb devices are no longer created by default for s390 the s390
test cases need to be adjusted.
Signed-off-by: Stefan Zimmermann <stzi@linux.vnet.ibm.com>
Reviewed-by: Boris Fiuczynski <fiuczy@linux.vnet.ibm.com>
The mask was stored both as a bitmap and as a string. The string is used
for XML output only. Remove the string, as it can be reconstructed from
the bitmap.
The test change is necessary as the bitmap formatter doesn't "optimize"
using the '^' operator.
In order for QEMU vCPU (and other) threads to run with RT scheduler,
libvirt needs to take care of that so QEMU doesn't have to run privileged.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1178986
Signed-off-by: Martin Kletzander <mkletzan@redhat.com>
It is only supported for virtio adapters.
Silently drop it if it was specified for other models,
as is done for other virtio attributes.
Also mention this in the documentation.
https://bugzilla.redhat.com/show_bug.cgi?id=1147195
Add attribute to set vgamem_mb parameter of QXL device for QEMU. This
value sets the size of VGA framebuffer for QXL device. Default value in
QEMU is 8MB so reuse it also in libvirt to not break things.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1076098
Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
The vram attribute was introduced to set the video memory but it is
usable only for few hypervisors excluding QEMU/KVM and the old XEN
driver. Only in case of QEMU the vram was used for QXL.
This patch updates the documentation to reflect current code in libvirt
and also changes the cases when we will set the default vram attribute.
It also fixes existing strange default value for VGA devices 9MB to 16MB
because the video ram should be rounded to power of two.
The change of default value could affect migrations but I found out that
QEMU always round the video ram to power of two internally so it's safe
to change the default value to the next closest power of two and also
silently correct every domain XML definition. And it's also safe because
we don't pass the value to QEMU.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1076098
Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
To track state of virtio channels this patch adds a new output-only
attribute called 'state' to the <target> element of virtio channels.
This will be later populated with the guest state of the channel.
CPU numa topology implicitly allows memory specification in 'KiB'.
Enabling this to accept the 'unit' in which memory needs to be specified.
This now allows users to specify memory in units of choice, and
lists the same in 'KiB' -- just like other 'memory' elements in XML.
<numa>
<cell cpus='0-3' memory='1024' unit='MiB' />
<cell cpus='4-7' memory='1024' unit='MiB' />
</numa>
Also augment test cases to correctly model NUMA memory specification.
This adds the tag 'unit="KiB"' for memory attribute in NUMA cells.
Signed-off-by: Prerna Saxena <prerna@linux.vnet.ibm.com>
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Right now when building the qemu command line, we try to do various
unconditional validations of the guest CPU against the host CPU. However
this checks are overly applied. The only time we should use the checks
are:
- The user requests host-model/host-passthrough, or
- When KVM is requsted. CPU features requested in TCG mode are always
emulated by qemu and are independent of the host CPU, so no host CPU
checks should be performed.
Right now if trying to specify a CPU for arm on an x86 host, it attempts
to do non-sensical validation and falls over.
Switch all the test cases that were intending to test CPU validation to
use KVM, so they continue to test the intended code.
Amend some aarch64 XML tests with a CPU model, to ensure things work
correctly.
