There are couple of places where a domain with a VFIO device gets
special treatment: in CGroups when enabling/disabling access to
/dev/vfio/vfio, and when creating/removing nodes in domain mount
namespace. Well, a NVMe disk is a VFIO device too. Fortunately,
we have this qemuDomainNeedsVFIO() function which is the only
place that needs adjustment.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Cole Robinson <crobinso@redhat.com>
If a domain has an NVMe disk configured, then we need to create
/dev/vfio/* paths in domain's namespace so that qemu can open
them.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Cole Robinson <crobinso@redhat.com>
We have this beautiful function that does crystal ball
divination. The function is named
qemuDomainGetMemLockLimitBytes() and it calculates the upper
limit of how much locked memory is given guest going to need. The
function bases its guess on devices defined for a domain. For
instance, if there is a VFIO hostdev defined then it adds 1GiB to
the guessed maximum. Since NVMe disks are pretty much VFIO
hostdevs (but not quite), we have to do the same sorcery.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
ACKed-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Cole Robinson <crobinso@redhat.com>
Sometimes, we have a PCI address and not fully allocated
virPCIDevice and yet we still want to know its /dev/vfio/N path.
Introduce virPCIDeviceAddressGetIOMMUGroupDev() function exactly
for that.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Cole Robinson <crobinso@redhat.com>
Previous patches rendered some of 'cleanup' labels needless.
Drop them.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Cole Robinson <crobinso@redhat.com>
Now that all callers of qemuDomainGetHostdevPath() handle
/dev/vfio/vfio on their own, we can safely drop handling in this
function. In near future the decision whether domain needs VFIO
file is going to include more device types than just
virDomainHostdev.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Cole Robinson <crobinso@redhat.com>
There are several variables which could be automatically freed
upon return from the function. I'm not changing @tmpPaths (which
is a string list) because it is going to be removed in next
commit.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Cole Robinson <crobinso@redhat.com>
In near future, the decision what to do with /dev/vfio/vfio with
respect to domain namespace and CGroup is going to be moved out
of qemuDomainGetHostdevPath() because there will be some other
types of devices than hostdevs that need access to VFIO.
All functions that I'm changing (except qemuSetupHostdevCgroup())
assume that hostdev we are adding/removing to VM is not in the
definition yet (because of how qemuDomainNeedsVFIO() is written).
Fortunately, this assumption is true.
For qemuSetupHostdevCgroup(), the worst thing that may happen is
that we allow /dev/vfio/vfio which was already allowed.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Cole Robinson <crobinso@redhat.com>
qemuBuildSoundCodecStr() validates if a given QEMU binary
supports the sound codec. This validation can be moved to
qemu_domain.c to be executed in domain define time.
The codec validation was moved to the existing
qemuDomainDeviceDefValidateSound() function.
Reviewed-by: Cole Robinson <crobinso@redhat.com>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Move QEMU caps validation of QEMU_CAPS_OBJECT_USB_AUDIO and
QEMU_CAPS_DEVICE_ICH9_INTEL_HDA to a new function in qemu_domain.c,
qemuDomainDeviceDefValidateSound(). This function is called by
qemuDomainDeviceDefValidate() to validate the sound device
in domain define time.
qemuxml2xmltest.c was adjusted to add the now required caps for
domain definition.
Reviewed-by: Cole Robinson <crobinso@redhat.com>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>
qemuBuildTPMDevStr() does TPM model validation that can be moved to
qemu_domain.c, allowing validation in domain define time. This patch
moves it to the existing qemuDomainDeviceDefValidateTPM() function.
Reviewed-by: Cole Robinson <crobinso@redhat.com>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Console validation is currently being done by qemuBuildConsoleCommandLine().
This patch moves it to a new qemuDomainDefValidateConsole() function. This
new function is then called by qemuDomainDefValidate(), validating the
console in domain define time.
Reviewed-by: Cole Robinson <crobinso@redhat.com>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Move the SPICE caps validation from qemuBuildGraphicsSPICECommandLine()
to a new function called qemuDomainDeviceDefValidateSPICEGraphics().
This function is called by qemuDomainDeviceDefValidateGraphics(),
which in turn is called by qemuDomainDefValidate(), validating the graphics
parameters in domain define time.
This validation move exposed a flaw in the 'default-video-type' tests
for PPC64, AARCH64 and s390 archs. The XML was considering 'spice' as
the default video type, which isn't true for those architectures.
This was flying under the radar until now because the SPICE validation
was being made in 'virsh start' time, while the XML validation done in
qemuxml2xmltest.c considers define time.
All other tests were adapted to consider SPICE validation in this
earlier stage.
Reviewed-by: Cole Robinson <crobinso@redhat.com>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Move the VNC cap validation from qemuBuildGraphicsVNCCommandLine()
to qemuDomainDeviceDefValidateGraphics(). This function is called by
qemuDomainDefValidate(), validating the graphics parameters in domain
define time.
Tests were adapted to consider SDL validation in this earlier stage.
Reviewed-by: Cole Robinson <crobinso@redhat.com>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>
There are validations for SDL, VNC, SPICE and EGL_HEADLESS
around several BuildGraphics*CommandLine in qemu_command.c. This
patch starts to move all of them to qemu_domain.c, inside the
existent qemuDomainDeviceDefValidateGraphics() function. This
function is called by qemuDomainDefValidate(), validating the
graphics parameters in domain define time.
In this patch we'll move the SDL validation code from
qemuBuildGraphicsSDLCommandLine(). Tests were adapted to consider
SDL validation in this earlier stage.
Reviewed-by: Cole Robinson <crobinso@redhat.com>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Move the pcihole64 validation being done by
qemuBuildGlobalControllerCommandLine() to the existing function
qemuDomainDeviceDefValidateControllerPCI(), which provides
domain define time validation.
The existing pcihole64 validations in qemu_domain.c were replaced
by the ones moved from qemu_command.c. The reason is that they
are more specific, allowing VIR_DOMAIN_CONTROLLER_MODEL_PCI_ROOT
and VIR_DOMAIN_CONTROLLER_MODEL_PCIE_ROOT to have distinct validation,
with exclusive QEMU caps and machine types.
Tests were adapted to consider the new caps being needed in
this earlier stage.
Reviewed-by: Cole Robinson <crobinso@redhat.com>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Move the boot validation being done by qemuBuildBootCommandLine()
to to a new qemuDomainDefValidateBoot() function. This new function
is called by qemuDomainDefValidate(), allowing boot validation in
domain define time.
Tests were adapted to consider the new caps being needed in
this earlier stage.
Reviewed-by: Cole Robinson <crobinso@redhat.com>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Move the PM validation being done by qemuBuildPMCommandLine() to
to a new qemuDomainDefValidatePM() function. This new function
is called by qemuDomainDefValidate(), promoting PM validation in
domain define time.
Tests were adapted to consider the new caps being needed in
this earlier stage.
Reviewed-by: Cole Robinson <crobinso@redhat.com>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>
@def->clock validation is done by qemuBuildClockCommandLine() and
qemuBuildClockArgStr(). This patch centralize the validation done
in both these functions to a new qemuDomainDefValidateClockTimers()
function. This new function is then called by qemuDomainDefValidate(),
promoting clock validation in domain define time.
Tests were adapted to consider the new caps being needed in
this earlier stage.
Reviewed-by: Cole Robinson <crobinso@redhat.com>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>
QEMU_CAPS_DEVICE_VMGENID is now being validated by
qemuDomainDefValidate().
Reviewed-by: Cole Robinson <crobinso@redhat.com>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Move QEMU caps validation of qemuBuildHostdevCommandLine() to
qemuDomainDeviceDefValidateHostdev() and qemuDomainMdevDefValidate(),
allowing them to be validated at domain define time.
Tests were adapted to consider the new caps being needed in
this earlier stage.
Reviewed-by: Cole Robinson <crobinso@redhat.com>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Move QEMU caps validation of QEMU_CAPS_CHARDEV_FILE_APPEND and
QEMU_CAPS_CHARDEV_LOGFILE to qemuDomainChrSourceDefValidate().
Reviewed-by: Cole Robinson <crobinso@redhat.com>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Move QEMU caps validation of QEMU_CAPS_USB_HUB to a new function in
qemu_domain.c, qemuDomainDeviceDefValidateHub(). This function is
called by qemuDomainDeviceDefValidate() to validate the sound device
in domain define time.
qemuxml2xmltest.c was adjusted to add the now required caps for
domain definition.
Reviewed-by: Cole Robinson <crobinso@redhat.com>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>
A new function qemuDomainDeviceDefValidateNVRAM() was created
to validate the NVRAM in domain define time. Unit test was
adjusted to account for the extra QEMU_CAPS_DEVICE_NVRAM required
during domain define.
Reviewed-by: Cole Robinson <crobinso@redhat.com>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>
A new qemuDomainDefValidateNuma() function was created to host
all the QEMU caps validation being done inside qemuBuildNumaArgStr().
This new function is called by qemuDomainValidateCpuCount()
to allow NUMA validation in domain define time.
Tests were changed to account for the QEMU capabilities
that need to be present at domain define time.
Reviewed-by: Cole Robinson <crobinso@redhat.com>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Validation of MACHINE_KERNEL_IRQCHIP and MACHINE_KERNEL_IRQCHIP_SPLIT
QEMU caps are now being done in qemuDomainDefValidateFeatures().
Reviewed-by: Cole Robinson <crobinso@redhat.com>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>
virQEMUCapsSupportsVmport() is now being called inside
qemuDomainDefValidateFeatures() for VIR_DOMAIN_FEATURE_VMPORT
feature.
qemuxml2xmltest.c was changed to account for this caps being
now validated at domain define time.
Reviewed-by: Cole Robinson <crobinso@redhat.com>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Introduce a new function called qemuDomainDefValidatePSeriesFeature()
that will center all the PSeries validation done in qemu_command.c.
qemuDomainDefValidatePSeriesFeature() is then called during domain
define time, in qemuDomainDefValidateFeatures().
qemuxml2argvtest.c is also changed to include all the caps that now
are being validated in define time.
Reviewed-by: Cole Robinson <crobinso@redhat.com>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>
The use of the parseOpaque parameter was mistakenly removed in
commit 4a4132b462
Author: Daniel P. Berrangé <berrange@redhat.com>
Date: Tue Dec 3 10:49:49 2019 +0000
conf: don't use passed in caps in post parse method
causing the method to re-fetch qemuCaps that were already just
fetched and put into parseOpaque.
This is inefficient when parsing incoming XML, but for live
XML this is more serious as it means we use the capabilities
for the current QEMU binary on disk, rather than the running
QEMU.
That commit, however, did have a useful side effect of fixing
a crasher bug in the qemu post parse callback introduced by
commit 5e939cea89
Author: Jiri Denemark <jdenemar@redhat.com>
Date: Thu Sep 26 18:42:02 2019 +0200
qemu: Store default CPU in domain XML
The qemuDomainDefSetDefaultCPU() method in that patch did not
allow for the possibility that qemuCaps would be NULL and thus
resulted in a SEGV.
This shows a risk in letting each check in the post parse
callback look for qemuCaps == NULL. The safer option is to
check once upfront and immediately stop (postpone) further
validation.
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Don't check os type / virt type / arch in the post-parse callback
because we can't assume qemuCaps is non-NULL at this point. It
also conceptually belongs to the validation callback.
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
This commit aims to fix
https://bugzilla.redhat.com/show_bug.cgi?id=1610207
The cause was apparently incorrect handling of jobs in snapshot
revert code which allowed a thread executing snapshot delete to
begin job while snapshot revert was still running on another
thread. The snapshot delete thread then waited on a condition
variable in qemuMonitorSend() while the revert thread finished,
changing (and effectively corrupting) the qemuMonitor structure
under the delete thread which led to its crash.
The incorrect handling of jobs in revert code was due to the fact
that although qemuDomainRevertToSnapshot() correctly begins a job
at the start, the job was implicitly ended when qemuProcessStop()
was called because the job lives in the QEMU driver's private
data (qemuDomainObjPrivate) that was purged during
qemuProcessStop().
This fix prevents qemuProcessStop() from clearing jobs as the
idea of qemuProcessStop() clearing jobs seems wrong in the first
place. It was (inadvertently) introduced in commit
888aa4b6b9, which is effectively reverted by
the second hunk of this commit. To preserve the desired effects
of the faulty commit, the first hunk is included as suggested by
Michal.
Signed-off-by: Pavel Mores <pmores@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
When QEMU uid/gid is set to non-root this is pointless as if we just
used a regular setuid/setgid call, the process will have all its
capabilities cleared anyway by the kernel.
When QEMU uid/gid is set to root, this is almost (always?) never
what people actually want. People make QEMU run as root in order
to access some privileged resource that libvirt doesn't support
yet and this often requires capabilities. As a result they have
to go find the qemu.conf param to turn this off. This is not
viable for libguestfs - they want to control everything via the
XML security label to request running as root regardless of the
qemu.conf settings for user/group.
Clearing capabilities was implemented originally because there
was a proposal in Fedora to change permissions such that root,
with no capabilities would not be able to compromise the system.
ie a locked down root account. This never went anywhere though,
and as a result clearing capabilities when running as root does
not really get us any security benefit AFAICT. The root user
can easily do something like create a cronjob, which will then
faithfully be run with full capabilities, trivially bypassing
the restriction we place.
IOW, our clearing of capabilities is both useless from a security
POV, and breaks valid use cases when people need to run as root.
This removes the clear_emulator_capabilities configuration
option from qemu.conf, and always runs QEMU with capabilities
when root. The behaviour when non-root is unchanged.
Reviewed-by: Cole Robinson <crobinso@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Since blockcommit is asynchronous, libvirtd can be restarted while the
operation runs. To ensure the information necessary to finish up the job
is not lost, serialisation to and deserialisation from the status XML is
added.
To unittest this, the new element was only added to the active commit test,
the non-active commit test doesn't have the new element so as to test its
absence.
Signed-off-by: Pavel Mores <pmores@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
A backup blockjob needs to be able to notify the parent backup job as
well as track all data to be able to clean up the bitmap and blockdev
used for the backup.
Add the data structure, job allocation function and status XML formatter
and parser.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Store the data of a backup job along with the index counter for new
backup jobs in the status XML. Currently we will support only one
backup job and thus there's no necessity to add arrays of jobs.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
A backup job may consist of many backup sub-blockjobs. Add the new
blockjob type and add all type converter strings.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
We will want to use the async job infrastructure along with all the APIs
and event for the backup job so add the backup job as a new async job
type.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Introduce QEMU_DOMAIN_JOB_STATS_TYPE_BACKUP and the convertors and other
plumbing to be able to report statistics for the backup job.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
$ cat f | grep -e arch -e emulator
<type arch='mipsel'>hvm</type>
$ sudo virsh define f
error: Failed to define domain from f
error: An error occurred, but the cause is unknown
After:
$ sudo virsh define f
error: Failed to define domain from f
error: unsupported configuration: No emulator found for arch 'mipsel'
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Cole Robinson <crobinso@redhat.com>
Annoyingly there was no existing constructor, and identifying all the
places which do a VIR_ALLOC(cpu) is a bit error prone. Hopefully this
has found & converted them all.
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Avoid grabbing the whole virCapsPtr object when we only need the
NUMA information.
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
The NUMA cells are stored directly in the virCapsHostPtr
struct. This moves them into their own struct allowing
them to be stored independantly of the rest of the host
capabilities. The change is used as an excuse to switch
the representation to use a GPtrArray too.
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Now that the domain XML APIs don't use virCapsPtr we can stop passing it
around many QEMU driver methods.
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
