Commit Graph

31006 Commits

Author SHA1 Message Date
Katerina Koukiou
aa3e07caec qemu: qemuDomainChangeNet: validity checks should be done before XML autocompletion
This patch ensures that changes in attributes of interfaces will emit
errors except if they are missing from the XML.
Previously we were falsely reporting successful updates, because some
changed attributes got overwritten before the validity checks.

https://bugzilla.redhat.com/show_bug.cgi?id=1599513

Signed-off-by: Katerina Koukiou <kkoukiou@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
2018-08-27 17:02:02 +02:00
Christian Ehrhardt
8741b94351
apparmor: fix ptrace rules with kernel 4.18
Due to kernel upstream change 338d0be4 ("apparmor: fix ptrace read check")
libvirt now hits apparmor denies like:
  apparmor="DENIED" operation="ptrace" profile="/usr/sbin/libvirtd"
  pid=4409 comm="libvirtd" requested_mask="read" denied_mask="read"
  peer="libvirt-14e92a75-7668-4b97-8f92-322fc1b9c78a"

Extend the ptrace rule to also allow 'ptrace (read)' for libvirtd to work
with these newer kernels.

Fixes: https://bugs.launchpad.net/bugs/1788603

Reported-by: Thadeu Lima de Souza Cascardo <thadeu.cascardo@canonical.com>
Reviewed-by: Erik Skultety <eskultet@redhat.com>
Acked-by: Jamie Strandboge <jamie@canonical.com>
Signed-off-by: Christian Ehrhardt <christian.ehrhardt@canonical.com>
2018-08-27 10:46:45 +02:00
Roman Bolshakov
171aa72baa qemu: Start domain on a node without cpu affinity
qemuProcessInitCpuAffinity prevents a VM from getting started on a
platform that uses cpu affinity wrapper stubs e.g. macOS.

The patch adds qemuProcessInitCpuAffinity stub on all platforms without
HAVE_SCHED_GETAFFINITY or HAVE_BSD_CPU_AFFINITY.

Signed-off-by: Roman Bolshakov <r.bolshakov@yadro.com>
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
2018-08-27 10:42:38 +02:00
Roman Bolshakov
0041eda1e4 util: eventpoll: Survive EBADF on macOS
Fixes:
https://www.redhat.com/archives/libvir-list/2017-January/msg00978.html

QEMU is probed through monitor fd to check capabilities during libvirtd init.
The monitor fd is closed after probing by virQEMUCapsInitQMPCommandFree
that calls virQEMUCapsInitQMPCommandAbort that calls qemuMonitorClose,
the latter one notifies the event loop via an interrupt handle in
qemuMonitorUnregister and after then closes monitor fd.

There could be a case when interrupt is sent after eventLoop is unlocked
but before virEventPollRunOnce blocks in poll, shortly before file
descriptor is closed by qemuMonitorClose. Then poll receives closed monitor
fd in fdset and returns EBADF.

EBADF is not mentioned as a valid errno on macOS poll man-page but such
behaviour can appear release-to-release, according to cpython:
https://github.com/python/cpython/blob/master/Modules/selectmodule.c#L1161

The change also fixes the issue in qemucapabilitiestest. It returns
Bad file descriptor message 25 times without the fix.

Signed-off-by: Roman Bolshakov <r.bolshakov@yadro.com>
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
2018-08-27 10:38:26 +02:00
Peter Krempa
3b89e1f962 docs: api_extension: Don't encourage other tools than git
Save us hassle in the list if anybody would read this.

Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Erik Skultety <eskultet@redhat.com>
2018-08-24 16:27:23 +02:00
Peter Krempa
6be034a8c0 docs: api_extension: Remove example patches
Now that they are not linked any more remove them.

Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
2018-08-24 16:23:01 +02:00
Peter Krempa
3ce03abc1d docs: api_extension: Remove links to the stale example patches
The patches used as an example for the api_extension manual don't hold
up to the current standards any more. Carefully remove links and
mentions of the patches from the docs.

Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
2018-08-24 16:20:36 +02:00
Peter Krempa
f2019083de qemu: capabilities: Always assume QEMU_CAPS_ADD_FD
The capability was usable since qemu 1.3 so we can remove all the
detection code.

Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: John Ferlan <jferlan@redhat.com>
2018-08-24 15:58:34 +02:00
Peter Krempa
3411fd4db4 tests: qemuxml2argv: modernize TPM passthrough tests
All supported qemus support FD passing so modify the tests to test the
proper code path.

Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: John Ferlan <jferlan@redhat.com>
2018-08-24 15:58:34 +02:00
Peter Krempa
09e45e8a9b qemu: command: Extract opening of TPM backend FDs for mocking purposes
Allow mocking of the file descriptor numbers used for the TPM
passthrough mode by extracting the relevant code into an exported
function.

Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: John Ferlan <jferlan@redhat.com>
2018-08-24 15:58:34 +02:00
Peter Krempa
d06a8ebe8f tests: qemuxml2argvmock: Allow 'safe' file descriptors in mocked virCommandPassFD
Allow FDs which are marked as safe for FD passing.

Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: John Ferlan <jferlan@redhat.com>
2018-08-24 15:58:34 +02:00
Peter Krempa
22d8f55b21 qemu: capabilities: Detect active block commit via QMP schema probing if possible
For versions where we can probe that the arguments are optional we can
perform the probing by a schema query rather than sending a separate
command to do so.

Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: John Ferlan <jferlan@redhat.com>
2018-08-24 15:58:34 +02:00
Peter Krempa
139ee1f192 qemu: qapi: Allow selecting specifically optional schema entries in virQEMUQAPISchemaTraverse
Add a new modifier character '*' which will select given schema entry
only when it is optional (denoted by the presence of the 'default' key).

Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: John Ferlan <jferlan@redhat.com>
2018-08-24 15:58:33 +02:00
Peter Krempa
c59734e3f6 qemu: qapi: Split up virQEMUQAPISchemaObjectGetType
Split it into a function that returns the whole schema entry so that we
can do additional checks and a helper getting the type string from the
schema entry.

Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: John Ferlan <jferlan@redhat.com>
2018-08-24 15:58:33 +02:00
Peter Krempa
cbcee32350 qemu: qapi: Simplify value handling in virQEMUQAPISchemaTraverse
Introduce a few variables so that we can easily access the modifier
character and also don't have to do pointer arithmetic when selecting
the schema entries. This will simplify adding of new modifier
characters.

Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: John Ferlan <jferlan@redhat.com>
2018-08-24 15:58:33 +02:00
John Ferlan
b04d1b6a35 storage: Add --shrink to qemu-img command when shrinking vol
https://bugzilla.redhat.com/show_bug.cgi?id=1613746

When shrinking the capacity of a qcow2 or luks volume using
the qemu-img program, the --shrink qualifier must be added.

Signed-off-by: John Ferlan <jferlan@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
2018-08-24 09:14:13 -04:00
John Ferlan
b4833917f1 nwfilter: Add extra verbiage for binding create/delete
https://bugzilla.redhat.com/show_bug.cgi?id=1609454

Add some cautionary words related to the create and delete
NWFilter Binding use cases and possible issues that may result
to the virsh nwfilter-binding-{create|delete} descriptions
and the virNWFilterBinding{CreateXML|Delete) API descriptions.

Essentially summarizing commit 2d9318b6c without using the
shoot yourself in the foot wording.

Signed-off-by: John Ferlan <jferlan@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
2018-08-24 08:45:19 -04:00
John Ferlan
6ef65e3c96 access: Fix nwfilter-binding ACL access API name generation
https://bugzilla.redhat.com/show_bug.cgi?id=1611320

Generation of the ACL API policy is a "automated process"
based on this perl script which "worked" with the changes to
add nwfilter binding API's because they had the "nwfilter"
prefix; however, the generated output name was incorrect
based on the remote protocol algorithm which expected to
generate names such as 'nwfilter-binding.action' instead
of 'nwfilter.binding-action'.

This effectively changes src/access/org.libvirt.api.policy entries:

  org.libvirt.api.nwfilter.binding-create ==>
      org.libvirt.api.nwfilter-binding.create

  org.libvirt.api.nwfilter.binding-delete ==>
      org.libvirt.api.nwfilter-binding.delete

  org.libvirt.api.nwfilter.binding-getattr ==>
      org.libvirt.api.nwfilter-binding.getattr

  org.libvirt.api.nwfilter.binding-read ==>
      org.libvirt.api.nwfilter-binding.read

Signed-off-by: John Ferlan <jferlan@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
2018-08-24 08:04:14 -04:00
Peter Krempa
3b7f11eb90 qemu: monitor: Separate probing for active block commit
Extract the code used to probe for the functionality so that it does not
litter the code used for actual work.

Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: John Ferlan <jferlan@redhat.com>
2018-08-24 13:52:44 +02:00
Peter Krempa
ee06cd36e3 qemu: monitor: Rename 'device' argument for block job control APIs
Starting from qemu 2.7 the 'device' argument is in fact a name of the
job itself. Change our APIs accordingly and adjust the error message.

Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: John Ferlan <jferlan@redhat.com>
2018-08-24 13:52:44 +02:00
Peter Krempa
29dd778d16 qemu: monitor: Use qemuMonitorJSONBlockJobError in qemuMonitorJSONDrivePivot
The API deals with a block job so use the common error reporting
function for it.

Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: John Ferlan <jferlan@redhat.com>
2018-08-24 13:52:44 +02:00
Peter Krempa
e94a4053dc qemu: monitor: Move qemuMonitorJSONDrivePivot together with block-job APIs
Move all relevant APIs dealing with existing jobs together.

Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: John Ferlan <jferlan@redhat.com>
2018-08-24 13:52:44 +02:00
Peter Krempa
7cead0af6a qemu: monitor: Use qemuMonitorJSONCheckError in qemuMonitorJSONBlockStream
The API does not report any special job-related error so the generic
error function should be used.

Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: John Ferlan <jferlan@redhat.com>
2018-08-24 13:52:44 +02:00
Peter Krempa
9c9c5766e1 qemu: monitor: Remove temporary variables
Now that the job name is used in single place in the respective
functions remove the temporary strings.

Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: John Ferlan <jferlan@redhat.com>
2018-08-24 13:52:44 +02:00
Peter Krempa
0ab2a65314 qemu: monitor: Use qemuMonitorJSONCheckError in qemuMonitorJSONBlockJobError
Report the generic errors using the existing function so that we don't
reimplement the same functionality multiple times.

Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: John Ferlan <jferlan@redhat.com>
2018-08-24 13:52:44 +02:00
Peter Krempa
f3b7607f6a qemu: monitor: Remove error classes not conforming to QAPI schema
Both were removed prior to qemu v1.2.0-rc0 when switching to the new
error format where almost all error types were converted to GenericError.

Relevant qemu commits are <de253f14912e> and <df1e608a01eb0>

Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: John Ferlan <jferlan@redhat.com>
2018-08-24 13:52:44 +02:00
Lubomir Rintel
4c833bf2d0 news: Add a mention of RISC-V guest support
Signed-off-by: Lubomir Rintel <lkundrak@v3.sk>
Reviewed-by: Andrea Bolognani <abologna@redhat.com>
2018-08-24 13:23:36 +02:00
Lubomir Rintel
097ff04948 tests: Add RISC-V guest
Signed-off-by: Lubomir Rintel <lkundrak@v3.sk>
Reviewed-by: Andrea Bolognani <abologna@redhat.com>
2018-08-24 13:23:36 +02:00
Lubomir Rintel
869cb692fa tests: Add RISC-V architectures
Signed-off-by: Lubomir Rintel <lkundrak@v3.sk>
Reviewed-by: Andrea Bolognani <abologna@redhat.com>
2018-08-24 13:23:36 +02:00
Lubomir Rintel
1ae7029870 qemu: assign addresses to virtio devices on RISC-V
Signed-off-by: Lubomir Rintel <lkundrak@v3.sk>
Reviewed-by: Andrea Bolognani <abologna@redhat.com>
2018-08-24 13:23:36 +02:00
Lubomir Rintel
db98a426a6 qemu: add qemuDomainAssignVirtioMMIOAddresses()
We're going to need to assign virtio-mmio addresses to non-ARM
guests soon, so let's create a generic wrapper that calls to
the architecture-specific implementation.

Signed-off-by: Lubomir Rintel <lkundrak@v3.sk>
Reviewed-by: Andrea Bolognani <abologna@redhat.com>
2018-08-24 13:23:36 +02:00
Lubomir Rintel
ee57f38c46 qemu: add qemuDomainIsRISCVVirt() and qemuDomainMachineIsRISCVVirt()
Signed-off-by: Lubomir Rintel <lkundrak@v3.sk>
Reviewed-by: Andrea Bolognani <abologna@redhat.com>
2018-08-24 13:23:36 +02:00
Lubomir Rintel
93f165331d qemu: RISC-V machines have no PCI
Signed-off-by: Lubomir Rintel <lkundrak@v3.sk>
Reviewed-by: Andrea Bolognani <abologna@redhat.com>
2018-08-24 13:23:36 +02:00
Lubomir Rintel
2faf932266 util: add RISC-V architectures
Signed-off-by: Lubomir Rintel <lkundrak@v3.sk>
Reviewed-by: Andrea Bolognani <abologna@redhat.com>
2018-08-24 13:23:36 +02:00
Lubomir Rintel
fa95035bd4 qemu: Rename qemuDomain*IsVirt() to qemuDomain*IsARMVirt()
They're ARM specific.

Signed-off-by: Lubomir Rintel <lkundrak@v3.sk>
Reviewed-by: Andrea Bolognani <abologna@redhat.com>
2018-08-24 13:23:36 +02:00
Simon Kobyda
95785c1c9d vsh: Added tests
For now, there are 9 test cases
- testVshTableNew: Creating table with empty header
- testVshTableHeader: Printing table with/without header
- testVshTableRowAppend: Appending row with various number of cells.
  Only row with same number of cells as in header is accepted.
- testUnicode: Printing table with unicode characters.
  Checking correct alignment.
- testUnicodeArabic: test opposite (right to left) writing
- testUnicodeZeroWidthChar
- testUnicodeCombiningChar
- testUnicodeNonPrintableChar,
- testNTables: Create and print varios types of tables - one column,
  one row table, table without content, standart table...

Signed-off-by: Simon Kobyda <skobyda@redhat.com>
2018-08-24 10:53:59 +02:00
Simon Kobyda
2e97450425 virsh: Implement new table API for virsh list
Instead of printing it straight in virsh, it creates table struct
which is filled with header and rows(domains). It allows us to know
more about table before printing to calculate alignment right.

Signed-off-by: Simon Kobyda <skobyda@redhat.com>
2018-08-24 10:53:20 +02:00
Simon Kobyda
9417f0b3f6 vsh: Add API for printing tables.
It solves problems with alignment of columns. Width of each column
is calculated by its biggest cell. Should solve unicode bug.
In future, it may be implemented in virsh, virt-admin...

This API has 5 public functions:
- vshTableNew - adds new table and defines its header
- vshTableRowAppend - appends new row (for same number of columns as in
header)
- vshTablePrintToStdout
- vshTablePrintToString
- vshTableFree

https://bugzilla.redhat.com/show_bug.cgi?id=1574624
https://bugzilla.redhat.com/show_bug.cgi?id=1584630

Signed-off-by: Simon Kobyda <skobyda@redhat.com>
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
2018-08-24 10:52:41 +02:00
Andrea Bolognani
d1a6d73ddf qemu: Make sure preferredMachines is not missing any entry
With the current implementation, adding a new architecture
and not updating preferredMachines accordingly will not
cause a build failure, making it very likely that subtle
bugs will be introduced in the process. Rework the code
so that such issues will be caught by the compiler.

Signed-off-by: Andrea Bolognani <abologna@redhat.com>
Reviewed-by: Erik Skultety <eskultet@redhat.com>
2018-08-24 09:17:39 +02:00
Shivaprasad G Bhat
1a94cbcd16 storage: fix the error message when encrypted raw volume resize
The vol-dumpxml shows the volume target format type as raw for
encrypted volumes. The error message when attempting to resize
with prealloc is confusing here.

Signed-off-by: Shivaprasad G Bhat <sbhat@linux.vnet.ibm.com>
Reviewed-by: John Ferlan <jferlan@redhat.com>
2018-08-23 12:14:31 -04:00
Eric Blake
a4c1cba28e qemu: fix error message on directory creation
Minor copy-and-paste bug present since commit 462c74c3, in Apr 2010.

Signed-off-by: Eric Blake <eblake@redhat.com>
2018-08-22 16:24:06 -05:00
Peter Krempa
2cad501bcb qemu: monitor: Fix device matching in qemuMonitorJSONBlockIoThrottleInfo
We should compare the alias/qdev id only when it was provided by the
caller and when it was found in the reply. Otherwise we could
dereference a NULL pointer. STRNEQ_NULLABLE is not appropriate since
it would return 'true' if the string was not present in the JSON output.

Found by Coverity.

Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: John Ferlan <jferlan@redhat.com>
2018-08-22 13:55:27 +02:00
Peter Krempa
8aced2ea32 qemu: hotplug: Don't leak 'nodename' in qemuDomainChangeMediaBlockdev
qemuDomainDiskGetBackendAlias allocates a copy of the nodename string so
we need to free it at the end.

Found by Coverity.

Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: John Ferlan <jferlan@redhat.com>
2018-08-22 13:55:26 +02:00
Peter Krempa
c86735e2d8 qemu: hotplug: Don't generate alias when detaching controllers
qemuDomainDetachControllerDevice contained code which implied that alias
might be NULL when detaching the disk and tried to generate it. This is
no longer possible so we can remove the code.

Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: John Ferlan <jferlan@redhat.com>
2018-08-22 13:55:26 +02:00
Dan Kenigsberg
d88a87d6c4 src: storage: Provide a better explanation of virStoragePoolSetAutostart
The former documentation was an unhelpful tautology. The suggested doc
borrows the wording from virDomainSetAutostart.

Signed-off-by: Dan Kenigsberg <danken@redhat.com>
Reviewed-by: Erik Skultety <eskultet@redhat.com>
2018-08-22 09:46:10 +02:00
Christian Ehrhardt
9a4e4b942d
process: wait longer 5->30s on hard shutdown
In cases where virProcessKillPainfully already reailizes that
SIGTERM wasn't enough we are partially on a bad path already.
Maybe the system is overloaded or having serious trouble to free and
reap resources in time.

In those case give the SIGKILL that was sent after 10 seconds some more
time to take effect if force was set (only then we are falling back to
SIGKILL anyway).

Signed-off-by: Christian Ehrhardt <christian.ehrhardt@canonical.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
2018-08-21 17:37:10 +02:00
Christian Ehrhardt
be2ca04447
process: wait longer on kill per assigned Hostdev
It was found that in cases with host devices virProcessKillPainfully
might be able to send signal zero to the target PID for quite a while
with the process already being gone from /proc/<PID>.

That is due to cleanup and reset of devices which might include a
secondary bus reset that on top of the actions taken has a 1s delay
to let the bus settle. Due to that guests with plenty of Host devices
could easily exceed the default timeouts.

To solve that, this adds an extra delay of 2s per hostdev that is associated
to a VM.

Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Christian Ehrhardt <christian.ehrhardt@canonical.com>
2018-08-21 17:37:10 +02:00
Peter Krempa
c9247c8e26 qemu: driver: Prepare qemuDomainGetStatsBlock (bulk disk stats) for -blockdev
Add code paths which call into the new functions to gather the data on a
per-node-name basis and tweak the aliases used for extracting the data.

Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
2018-08-21 15:46:06 +02:00
Peter Krempa
6d736e292c qemu: driver: Allow using blockdev with qemuDomainBlocksStatsGather
Use the 'qdev' instead of the disk alias to lookup the stats and
transfer the capacity from the appropriate node name so that the
function works with -blockdev.

Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
2018-08-21 15:46:06 +02:00
Peter Krempa
49510e4191 qemu: Use QOM path with query-block when using -blockdev
Switch to using the QOM/qdev handles in all calls to
qemuMonitorGetBlockInfo when using -blockdev. The callers also need to
make sure to use the correct handle afterwards to extract the data.

Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
2018-08-21 15:46:06 +02:00