Peter Krempa [Fri, 24 Nov 2017 11:04:14 +0000 (12:04 +0100)]
qemu: process: Move handling of non-backing files into qemuDomainDetermineDiskChain
Until now we would skip loading of the backing chain for files which
don't support backing chains only when starting up the VM. Move the
check from qemuProcessPrepareHostStorage with some adaptations so that's
always applied.
Pavel Hrdina [Thu, 30 Nov 2017 16:04:35 +0000 (17:04 +0100)]
conf: fix migratable XML for graphics if socket is generated based on config
The graphics code is complex and there are a lot of exceptions and
backward compatible combinations. One of them is the possibility
to configure "spice_auto_unix_socket" in qemu.conf which will convert
all spice graphics with listen type "address" without any address
specified to listen type "socket" when the guest is started.
We don't format this generated socket into migratable XML to make
migration work with older libvirt. However, spice has another
exception that if autoport='no' and there is no port configured
it is converted to listen type "none". Because of this we need
to format autoport='yes' to make sure that the listen type will
be the same as the offline XML.
Ján Tomko [Wed, 29 Nov 2017 14:05:23 +0000 (15:05 +0100)]
virQEMUCapsHasPCIMultiBus: assume true if we have no version information
In status XML, we do not store the QEMU version information, we only
format all the capabilities. We dropped QEMU_CAPS_PCI_MULTIBUS
in commit 5b783379 which was released in libvirt 3.2.0.
Therefore the only way of telling if the already running domain
at the time of daemon restart has been started with a QEMU that does
use 'pci.0' or not on PPC is to look at the pci-root controller's
alias. This is not an option if the domain has a user-specified alias
for the pci-root.
Instead of reintroducing the capability, assume 'pci.0' when we have
no version information. That way the only left broken use case would
be the combination of user aliases and very old QEMU.
Ján Tomko [Wed, 29 Nov 2017 13:38:52 +0000 (14:38 +0100)]
qemu: prefer the PCI bus alias from status XML
For some corner cases, virQEMUCapsHasPCIMultiBus depends on the QEMU
version, which is by design not stored in the status XML and therefore
it cannot be fixed for all existing running domains.
Prefer the controller alias read from the status XML when formatting
PCI addresses and only fall back to using virQEMUCapsHasPCIMultiBus
if the alias is a user alias.
This fixes hotplug after daemon restart for domains not using user
aliases.
The 'wiremode' attribute exists in a couple of Xen XML files, but no code has
ever parsed that value. It was later added to the RNG schema too, again despite
there not being any code which parses it.
Reviewed-by: Jim Fehlig <jfehlig@suse.com> Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
The libxlxml2domconfigdata directory was not covered in the RNG schema
tests. This hid a few bugs in both the libxl XML files and the RNG
schema itself.
Reviewed-by: Jim Fehlig <jfehlig@suse.com> Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Jiri Denemark [Wed, 29 Nov 2017 10:08:40 +0000 (11:08 +0100)]
spec: Make the build reproducible
If the spec file applies a patch which touches any file in the API XMLs
dependency tree, we need to regenerate the XMLs and consequently
recreate hvsupport.html. The file will contain a time stamp in a comment
which means it will be different every time the package is built. The
commit a54c9622860 which added the time stamp also added support for
SOURCE_DATE_EPOCH environment variable. Let's set it to the time stamp
of the spec file itself to make the build reproducible.
Jiri Denemark [Wed, 29 Nov 2017 10:08:15 +0000 (11:08 +0100)]
spec: Turn on verbose build
When building a package in a build system, such as koji or cbs, logs are
the only thing which can be used to diagnose failures. Make them verbose
since human friendly output of V=0 build doesn't really help when a
build fails.
Andrea Bolognani [Tue, 28 Nov 2017 10:34:55 +0000 (11:34 +0100)]
qemu: Require QEMU_CAPS_DEVICE_PL011 for pl011
Even though we never format the device on the QEMU command line,
as it's a platform serial device that's not user-instantiable,
we should still make sure it's available before using it.
Signed-off-by: Andrea Bolognani <abologna@redhat.com> Reviewed-by: Pavel Hrdina <phrdina@redhat.com>
Pino Toscano [Tue, 14 Nov 2017 17:00:30 +0000 (18:00 +0100)]
qemu: switch s390/s390x default console back to serial
Now that <serial> and <console> on s390/s390x behave a bit more like the
other architectures, remove this extra differentation, and use sclp
console by default for new guests. New virtio consoles can still be
added, and it is actually needed because of the limited number of
instances for sclp and sclplm.
Pino Toscano [Tue, 14 Nov 2017 15:27:04 +0000 (16:27 +0100)]
conf: add VIR_DOMAIN_CHR_SERIAL_TARGET_TYPE_SCLP
Introduce specific a target types with two models for the console
devices (sclp and sclplm) used in s390 and s390x guests, so isa-serial
is no more used for them.
This makes <serial> usable on s390 and s390x guests, with at most only
a single sclpconsole and one sclplmconsole devices usable in a single
guest (due to limitations in QEMU, which will enforce already at
runtime).
We can finally introduce a specific target model for the pl011 device
used by mach-virt guests, which means isa-serial will no longer show
up to confuse users.
We make sure migration works in both directions by interpreting the
isa-serial target type, or the lack of target type, appropriately
when parsing the guest XML, and skipping the newly-introduced type
when formatting if for migration. We also verify that pl011 is not
used for non-mach-virt guests and add a bunch of test cases.
Andrea Bolognani [Fri, 10 Nov 2017 16:57:53 +0000 (17:57 +0100)]
qemu: Support usb-serial and pci-serial on pSeries
The existing implementation set the address type for all serial
devices to spapr-vio, which made it impossible to use other devices
such as usb-serial and pci-serial; moreover, some decisions were
made based on the address type rather than the device type.
We can finally introduce a specific target model for the spapr-vty
device used by pSeries guests, which means isa-serial will no longer
show up to confuse users.
We make sure migration works in both directions by interpreting the
isa-serial target type, or the lack of target type, appropriately
when parsing the guest XML, and skipping the newly-introduced type
when formatting if for migration. We also verify that spapr-vty is
not used for non-pSeries guests and add a bunch of test cases.
Andrea Bolognani [Sun, 26 Nov 2017 11:30:10 +0000 (12:30 +0100)]
qemu: Remove redundancy in qemuBuildSerialChrDeviceStr()
Instead duplicating the capability check for each possible target
model, introduce a small helper that matches the target model with
the corresponding capability and collapse all existing checks into
a single one.
Signed-off-by: Andrea Bolognani <abologna@redhat.com> Reviewed-by: Pavel Hrdina <phrdina@redhat.com>
Andrea Bolognani [Tue, 21 Nov 2017 13:14:16 +0000 (14:14 +0100)]
qemu: Format targetModel for serial devices
Now that we've created a distinction between target type and target
model, with the latter being the concrete device name, it's time to
switch to formatting the model instead of the type.
Signed-off-by: Andrea Bolognani <abologna@redhat.com> Reviewed-by: Pavel Hrdina <phrdina@redhat.com>
Andrea Bolognani [Mon, 20 Nov 2017 11:05:17 +0000 (12:05 +0100)]
conf: Parse and format virDomainChrSerialTargetModel
This information will be used to select, and store in the guest
configuration in order to guarantee ABI stability, the concrete
(hypervisor-specific) model for serial devices.
Signed-off-by: Andrea Bolognani <abologna@redhat.com> Reviewed-by: Pavel Hrdina <phrdina@redhat.com>
Andrea Bolognani [Fri, 24 Nov 2017 16:52:26 +0000 (17:52 +0100)]
qemu: Improve qemuDomainChrTargetDefValidate()
Instead of validating each target type / address type combination
separately, create a small helper to perform the matching and
collapse all existing checks into a single one.
Signed-off-by: Andrea Bolognani <abologna@redhat.com> Reviewed-by: Pavel Hrdina <phrdina@redhat.com>
Andrea Bolognani [Mon, 20 Nov 2017 18:46:10 +0000 (19:46 +0100)]
qemu: Introduce qemuDomainChrTargetDefValidate()
Instead of waiting until we get to command line generation, we can
validate the target for a char device much earlier.
Move all the checks out of qemuBuildSerialChrDeviceStr() and into
the new fuction. This will later allow us to validate the target
for platform devices.
Signed-off-by: Andrea Bolognani <abologna@redhat.com> Reviewed-by: Pavel Hrdina <phrdina@redhat.com>
Andrea Bolognani [Tue, 21 Nov 2017 12:03:19 +0000 (13:03 +0100)]
conf: Remove ATTRIBUTE_FALLTHROUGH from virDomainChrTargetDefFormat()
Formatting the <target/> element for serial devices will become a
bit more complicated later on, and leaving the fallthrough behavior
there would do nothing but complicate it further.
Signed-off-by: Andrea Bolognani <abologna@redhat.com> Reviewed-by: Pavel Hrdina <phrdina@redhat.com>
Andrea Bolognani [Fri, 10 Nov 2017 13:54:19 +0000 (14:54 +0100)]
conf: Drop virDomainChrDeviceType.targetTypeAttr
This attribute was used to decide whether to format the type
attribute of the <target> element, but the logic didn't take into
account all possible cases and as such could lead to unexpected
results. Moreover, it's one more thing to keep track of, and can
easily fall out of sync with other attributes.
Now that we have VIR_DOMAIN_CHR_SERIAL_TARGET_TYPE_NONE, we can
use that value to signal that no specific target type has been
configured for the serial device and as such the attribute should
not be formatted at all. All other values are now formatted.
Signed-off-by: Andrea Bolognani <abologna@redhat.com> Reviewed-by: Pavel Hrdina <phrdina@redhat.com>
conf: Run devicePostParse() again for the first serial device
The devicePostParse() callback is invoked for all devices so that
drivers have a chance to set their own specific values; however,
virDomainDefAddImplicitDevices() runs *after* the devicePostParse()
callbacks have been invoked and can add new devices, in which case
the driver wouldn't have a chance to customize them.
Work around the issue by invoking the devicePostParse() callback
after virDomainDefAddImplicitDevices(), only for the first serial
devices, which might have been added by it. The same was already
happening for the first video device for the very same reason.
This will become important later on, when we will change
virDomainDefAddConsoleCompat() not to set a targetType for
automatically added serial devices.
Signed-off-by: Andrea Bolognani <abologna@redhat.com> Reviewed-by: Pavel Hrdina <phrdina@redhat.com>
Having a separate function for char device handling is better than
adding even more code to qemuDomainDeviceDefPostParse().
Signed-off-by: Andrea Bolognani <abologna@redhat.com> Reviewed-by: Pavel Hrdina <phrdina@redhat.com> Reviewed-by: Marc Hartmayer <mhartmay@linux.vnet.ibm.com>
Andrea Bolognani [Wed, 15 Nov 2017 16:15:15 +0000 (17:15 +0100)]
docs: Improve documentation for serial consoles
Our current documentation is missing some information and doesn't
do a great job at explaining how the <serial> and <console> elements
are connected. Let's try to fix that.
Signed-off-by: Andrea Bolognani <abologna@redhat.com> Reviewed-by: Pavel Hrdina <phrdina@redhat.com>
Julio Faracco [Mon, 27 Nov 2017 00:18:11 +0000 (22:18 -0200)]
virsh: Fix virsh prompt when connection changes to readonly mode.
This commit fixes the virsh prompt when reconnection to the same URI is
called: `virsh # connect --readonly` (Reconnect). The problem is
happening because the code is considering URI (name) as a mandatory
parameter to change the prompt. This commit remove the assignment into
`priv->readonly` from `if (name)` conditional.
The blockdev-add code provides a mechanism to sanely provide user
and password-secret arguments for iscsi without placing them on the
command line to be viewable by a 'ps -ef' type command or needing
to create separate -iscsi devices for each disk/volume found.
So modify the iSCSI command line building to check for the presence
of the capability in order properly setup and use the domain master
secret object to encrypt the password in a secret object and alter
the parameters for the command line to utilize.
Modify the xml2argvtest to exhibit the syntax for both disk and
hostdev configurations.
John Ferlan [Fri, 1 Sep 2017 19:09:29 +0000 (15:09 -0400)]
qemu: Get capabilities to use iscsi password-secret argument
Detect the capability via the query-qmp-schema for blockdev-add
to find the 'password-secret' parameter that will allow the iSCSI
code to use the master secret object to encrypt the secret for an
and only need to provide the object id of the secret on the command
line thus obsfuscating the passphrase.
Jason J. Herne [Fri, 24 Nov 2017 08:02:02 +0000 (09:02 +0100)]
s390: qemu-capabilities: Avoid error message when missing non-kvm host cpu info
Libvirt prints an error on startup when it is missing host cpu model
information for any queried qemu binary. On s390 we only have host cpu model
information for kvm enabled qemu instances. So when virt type is not kvm, this
is actually not an error on s390.
This patch adds virt type as a parameter to virQEMUCapsInitCPUModelS390, and a
new return code 2 for virQEMUCapsInitCPUModel and virQEMUCapsInitCPUModelS390.
If the virt type is not kvm then we skip printing the scary error message
and return 2 because this case is actually expected behavior. The new return
code is meant to differentiate between the failure case and the case where we
simply expect the cpu model information to be unattainable.
Signed-off-by: Jason J. Herne <jjherne@linux.vnet.ibm.com> Reviewed-by: Boris Fiuczynski <fiuczy@linux.vnet.ibm.com> Reviewed-by: Marc Hartmayer <mhartmay@linux.vnet.ibm.com> Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
John Ferlan [Sun, 8 Oct 2017 12:44:31 +0000 (08:44 -0400)]
storage: Introduce virStoragePoolObjListSearch
Create an API to search through the storage pool objects looking for
a specific truism from a callback API in order to return the specific
storage pool object that is desired.
John Ferlan [Sun, 8 Oct 2017 12:14:56 +0000 (08:14 -0400)]
storage: Introduce virStoragePoolObjListForEach
Create an API to walk the pools->objs[] list in order to perform a
callback function for each element of the objs array that doesn't care
about whether the action succeeds or fails as the desire is to run the
code over every element in the array rather than fail as soon as or if
one fails.
John Ferlan [Sun, 8 Oct 2017 13:09:09 +0000 (09:09 -0400)]
storage: Introduce virStoragePoolObjEndAPI
For now it'll just call the virStoragePoolObjUnlock, but a future
adjustment will do something different. Since the new API will check
for a NULL object before the Unlock call, callers no longer need to
check for NULL before calling.
The virStoragePoolObjUnlock is now private/static to virstorageobj.c
with a short term forward reference.
Peter Krempa [Thu, 23 Nov 2017 16:15:17 +0000 (17:15 +0100)]
qemu: command: Properly format disk 'debug' attribute
Move the setup of the disk attribute to the disk source prepare function
which will allow proper usage with JSON props and move the fallback
(legacy) generating code into the block which is executed with legacy
options.
As a side-effect of this change we can clean up propagation of 'cfg'
into the command generator.
Also it's nice to see that the test output is the same even when the
value is generated in a different place.
Peter Krempa [Thu, 23 Nov 2017 15:06:33 +0000 (16:06 +0100)]
util: storage: Add fields for debug options for disk drivers
Some drive backends allow output of debugging information which can be
configured using properties of the image. Add fields to virStorageSource
which will allow configuring them.
Peter Krempa [Thu, 23 Nov 2017 14:37:37 +0000 (15:37 +0100)]
qemu: command: Inject password-secret only when not using JSON props
The 'file.password-secret' injection should be used only if we are using
the old formatter. When formatting the source string from the JSON
properties, the property should be added there.
Also drop the comment which refers to stuff that will not be used in
libvirt since -blockdev is the way to go.
Michal Privoznik [Thu, 23 Nov 2017 17:00:33 +0000 (18:00 +0100)]
tools: Work around ancient readline
My latest commit of a785186446de785d uncovered a problem we fixed
in 9eb23fe2 but then reverted in 834c5720e443. Turns out, some
systems (I'm looking at you OS X) have ancient readline with
broken header file.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Peter Krempa [Wed, 15 Nov 2017 14:21:14 +0000 (15:21 +0100)]
qemu: command: Mark <shared/> disks as such in qemu
Qemu has now an internal mechanism for locking images to fix specific
cases of disk corruption. This requires libvirt to mark the image as
shared so that qemu lifts certain restrictions.
Peter Krempa [Tue, 14 Nov 2017 14:37:09 +0000 (15:37 +0100)]
qemu: domain: Reject shared disk access if backing format does not support it
Disk sharing between two VMs may corrupt the images if the format driver
does not support it. Check that the user declared use of a supported
storage format when they want to share the disk.
Peter Krempa [Tue, 14 Nov 2017 14:34:46 +0000 (15:34 +0100)]
qemu: block: Add function to check if storage source allows concurrent access
Storage source format backing a shared device (e.g. running a cluster
filesystem) needs to support the sharing so that metadata are not
corrupted. Add a central function for checking this.
Julio Faracco [Fri, 17 Nov 2017 21:27:40 +0000 (19:27 -0200)]
tests: changing network interface types when backend tag is defined.
Some test cases have the backend tag inside wrong interfaces. The backend xml
tag does not support <interface type='user|direct|hostdev'>. So this commit
changes some network types inside the interfaces that have backend defined.
Signed-off-by: Julio Faracco <jcfaracco@gmail.com>
Michal Privoznik [Tue, 14 Nov 2017 13:29:20 +0000 (14:29 +0100)]
qemu_capabilities: Introcude QEMU_CAPS_NUMA_DIST
This capability says if qemu is capable of specifying distances
between NUMA nodes on the command line. Unfortunately, there's no
real way to check this and thus we have to go with version check.
QEMU introduced this in 0f203430dd8 (and friend) which was
released in 2.10.0.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: John Ferlan <jferlan@redhat.com>
There's no point in checking if numa->mem_nodes[node].ndistances
is set if we check for numa->mem_nodes[node].distances. However,
it makes sense to check if the sibling node (@cellid) caller
passed falls within boundaries.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: John Ferlan <jferlan@redhat.com>
Jiri Denemark [Wed, 22 Nov 2017 12:12:00 +0000 (13:12 +0100)]
qemu: Use the end of QEMU log for reporting errors
When QEMU dies, we read its output stored in a log file and use it for
reporting a hopefully useful error. However, virReportError will trim
the message to (VIR_ERROR_MAX_LENGTH - 1) characters, which means the
end of the log (which likely contains the error message we want to
report) may get lost. We should trim the beginning of the log instead.
Jiri Denemark [Wed, 22 Nov 2017 09:19:12 +0000 (10:19 +0100)]
qemu: Properly skip "char device redirected to" in QEMU log
When reading QEMU log for reporting it as an error message, we want to
skip "char device redirected to" line. However, this string is not
printed at the beginning of a line, which means STRPREFIX will never
find it.
Signed-off-by: Jiri Denemark <jdenemar@redhat.com> Reviewed-by: Pavel Hrdina <phrdina@redhat.com>
xinhua.Cao [Mon, 13 Nov 2017 13:07:58 +0000 (21:07 +0800)]
remote: Move the call to remoteClientFreePrivateCallbacks from FreeFunc to CloseFunc
Still because of commit id 'fe8f1c8b' where we generate a REF for the
Register and that's transparent to the consumer (e.g. how would they
know they need to ensure that Deregister is called), thus the purpose of
this patch is to find a way to Deregister if it's determined that the
consumer hasn't by the time of the "last" REF we'd have.
This solution to this problem is to alter the processing to have the
remoteClientCloseFunc handle performing the Deregister calls instead of
the remoteClientFreeFunc because there's no way FreeFunc would be called
unless the Deregister was already called.
We have a check for backslash alignment that checks for two blanks preceding a
backslash. However there can be alignment done using a tabulator and in some
cases one might be enough. There are none currently, but I found out that was
the case before. at some point. so let's check for \t preceding the final
backslash as well.
Signed-off-by: Martin Kletzander <mkletzan@redhat.com>
projects / libvirt.git / log