Michal Privoznik [Tue, 25 Oct 2016 10:16:36 +0000 (12:16 +0200)]
qemuDomainAttachNetDevice: Don't overwrite error on rollback
If there is an error hotpluging a net device (for whatever
reason) a rollback operation is performed. However, whilst doing
so various helper functions that are called report errors on
their own. This results in the original error to be overwritten
and thus misleading the user.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Function qemuDomainAttachShmemDevice() steals the device data if the
hotplug was successful, but the condition checked for unsuccessful
execution otherwise.
Signed-off-by: Martin Kletzander <mkletzan@redhat.com>
qemu: conf: add option for tuning debug logging level
This helps in selecting log level of the gluster gfapi, output to stderr.
The option is 'gluster_debug_level', can be tuned by editing
'/etc/libvirt/qemu.conf'
Debug levels ranges 0-9, with 9 being the most verbose, and 0
representing no debugging output. The default is the same as it was
before, which is a level of 4. The current logging levels defined in
the gluster gfapi are:
Andrea Bolognani [Wed, 19 Oct 2016 15:36:45 +0000 (11:36 -0400)]
tests: qemucaps: Add QMP introspection data for qemu 2.6.0 on aarch64 gicv3
Add QMP schema data query for aarch64. The gic capabilities are
unfortunately queried after the QMP schema and thus this patch needs to
undo the temporary removal of the declared support for query-qmp-schema.
Note that as a gicv3 machine was not available the schema data is taken
from the gicv2 case. It should be identical since qemu would be built
from the same source.
Andrea Bolognani [Wed, 19 Oct 2016 15:36:45 +0000 (11:36 -0400)]
tests: qemucaps: Add QMP introspection data for qemu 2.6.0 on aarch64 gicv2
Add QMP schema data query for aarch64. The gic capabilities are
unfortunately queried after the QMP schema and thus this patch needs to
undo the temporary removal of the declared support for query-qmp-schema.
Peter Krempa [Wed, 19 Oct 2016 12:40:53 +0000 (14:40 +0200)]
qemu: capabilities: Add support for QMP schema introspection
Allow detecting capabilities according to the qemu QMP schema. This is
necessary as sometimes the availability of certain options depends on
the presence of a field in the schema.
This patch adds support for loading the QMP schema when detecting qemu
capabilities and adds a very simple query language to allow traversing
the schema and selecting a certain element from it.
The infrastructure in this patch uses a query path to set a specific
capability flag according to the availability of the given element in
the schema.
Peter Krempa [Thu, 20 Oct 2016 08:11:05 +0000 (10:11 +0200)]
tests: qemucaps: Temporarily remove 'query-qmp-schema' from test data
Remove the command from the supported commands list temporarily so that
QMP introspection code can be added without breaking tests and having to
tweak the test data in the same commit.
This will be later reverted and test data will be added. The aarch64
code is special as it calls additional commands and thus the test data
can't be added upfront.
test driver: Deny some operations on inactive domains
Some operations like reboot, save, coreDump, blockStats,
ifaceStats make sense iff domain is running. While it is
technically possible for our test driver to return success
regardless of domain state, we should copy constraints from
other drivers and thus deny these operations over inactive
domains.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Add check in qemuCheckDiskConfig for an invalid combination
of using the 'scsi' bus for a block 'lun' device and any disk
source format other than 'raw'.
Guido Günther [Fri, 4 Nov 2016 17:30:16 +0000 (18:30 +0100)]
Unbreak rebuilding docs with release tarballs
Release tarballs ship the include/libvirt/libvirt-common.h.
when srcdir != builddir we end up including libvirt-common.h twice: from
$top_srcdir/include/libvirt-common.h and from
$builddir/include/libvirt-common.h leading to
function virTypedParamsGetUInt from /tmp/buildd/libvirt-2.4.0/debian/build/docs/../include/libvirt/libvirt-common.h redeclared in /tmp/buildd/libvirt-2.4.0/docs/../include/libvirt/libvirt-common.h
function virTypedParamsAddBoolean from /tmp/buildd/libvirt-2.4.0/debian/build/docs/../include/libvirt/libvirt-common.h redeclared in /tmp/buildd/libvirt-2.4.0/docs/../include/libvirt/libvirt-common.h
…
Only add the builddir to the search list if there is no pregenerated
libvirt-common.h.
Reuse the existing check that predates the libvirt.h → libvirt-common.h
split and that probably was meant for exactly that.
Commit e8861f69714f changed our spec file to compile and run
tests in parallel. That's a very good step forward, but why
stop there? Let's run *all* make jobs in parallel and really
put those expensive cores to use!
Andrea Bolognani [Wed, 26 Oct 2016 14:27:05 +0000 (16:27 +0200)]
wireshark: Use ${exec_prefix} instead of ${prefix}
${exec_prefix} and ${prefix} point to the same directory in
most setups, but when that's not the case the former should
be used for architecture-dependent data such as shared objects,
which makes it the best fit for our Wireshark dissector.
While at it, change all uses of $(var) to ${var}: they are
absolutely identicaly as far as make's concerned, but autoconf
itself seems to prefer the latter form so we might as well
follow suit.
Andrea Bolognani [Thu, 27 Oct 2016 12:43:00 +0000 (14:43 +0200)]
wireshark: Make fallback path construction more reliable
We only need to strip $ws_prefix from $ws_plugindir if we've
retrieved it from pkg-config: if we're building it ourselves
from $libdir, we can just use it without further processing.
../../src/qemu/qemu_capabilities.c: In function 'virQEMUCapsIsValid':
../../src/qemu/qemu_capabilities.c:4085: error: declaration of 'ctime'
shadows a global declaration [-Wshadow]
/usr/include/time.h:258: error: shadowed declaration is here [-Wshadow]
After 06a7b1ff4 the @&opts_need_arg is not used anywhere. Well,
it is set but never read:
vsh.c: In function 'vshReadlineParse':
vsh.c:2658:14: warning: variable 'opts_need_arg' set but not used [-Wunused-but-set-variable]
uint64_t opts_need_arg, opts_seen;
^~~~~~~~~~~~~
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Jiri Denemark [Thu, 27 Oct 2016 09:13:28 +0000 (11:13 +0200)]
qemu: Unify cached caps validity checks
Let's keep all run time validation of cached QEMU capabilities in
virQEMUCapsIsValid and call it whenever we access the cache.
virQEMUCapsInitCached should keep only the checks which do not make
sense once the cache is loaded in memory.
Jiri Denemark [Thu, 27 Oct 2016 08:54:13 +0000 (10:54 +0200)]
qemu: Store loaded QEMU binary ctime in qemuCaps
virQEMUCapsLoadCache loads QEMU capabilities from a file, but strangely
enough it returns the loaded QEMU binary ctime in qemuctime parameter
instead of storing it in qemuCaps.
So far, the main code is built in parallel, which makes it pretty
fast. But with a lots of tests we have now I've noticed this part
takes too much time to build. The problem was that tests were
build and run in a single job.
Also, 'make' in the first hunk is useless. The test suite is not
built due to 'make all' because there's no .git in the sources
unpacked from a tar.xz archive. It's 'make check' which triggers
tests build.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
QEMU added support for ivshmem-plain and ivshmem-doorbell. Those are
reworked varians of legacy ivshmem that are compatible from the guest
POV, but not from host's POV and have sane specification and handling.
Details about the newer device type can be found in qemu's commit 5400c02b90bb:
We're keeping some things at default and that's not something we want to
do intentionally. Let's save some sensible defaults upfront in order to
avoid having problems later. The details for the defaults (of the newer
implementation) can be found in qemu's commit 5400c02b90bb:
Since we are merely saving the defaults it will not change the guest ABI
and thanks to the fact that we're doing it in the PostParse callback it
will not break the ABI stability checks.
Signed-off-by: Martin Kletzander <mkletzan@redhat.com>
Jiri Denemark [Wed, 2 Nov 2016 13:20:42 +0000 (14:20 +0100)]
qemu: Reset post-copy capability after migration
Unlike other migration capabilities, post-copy is also set on the
destination host which means it doesn't disappear once domain is
migrated. As a result of that other functionality which internally uses
migration to a file (virDomainManagedSave, virDomainSave,
virDomainCoreDump) may fail after migration because the post-copy
capability is still set.
The library.xen file contains a braindump of thoughts dating
from the very first days of libvirt, when it was briefly
called libxen. This is not useful and potentially misleading
or confusing for people.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
John Ferlan [Fri, 28 Oct 2016 20:59:58 +0000 (16:59 -0400)]
vsh: Pass correct values for command line completion
Commit id 'dcfdf341' passes 'opts_need_arg' and 'opts_seen' to
vshCmddefGetData, but that seems to be incorrect as those values
are not initialized properly (something at least one compiler found).
Instead the static 'const_opts_need_arg' and 'const_opts_seen' values
should be passed.
By passing unitialized values leads to not finding possible options
for simpler commands (domfsfreeze for example), where if you're in
a virsh shell using command line completion - you'll get a list of
files in your current directory instead of two options --domain and
--mountpoint (as would happen with this patch applied.
Laine Stump [Fri, 28 Oct 2016 15:43:56 +0000 (11:43 -0400)]
network: fix endless loop when starting network with multiple IPs and no dhcp
commit 9065cfaa added the ability to disable DNS services for a
libvirt virtual network. If neither DNS nor DHCP is needed for a
network, then we don't need to start dnsmasq, so code was added to
check for this.
Unfortunately, it was written with a great lack of attention to detail
(I can say that, because I was the author), and the loop that checked
if DHCP is needed for the network would never end if the network had
multiple IP addresses and the first <ip> had no <dhcp> subelement
(which would have contained a <range> or <host> subelement, thus
requiring DHCP services).
This patch rewrites the check to be more compact and (more
importantly) finite.
This bug was present in release 2.2.0 and 2.3.0, so will need to be
backported to any relevant maintainence branches.
Michal Privoznik [Wed, 26 Oct 2016 10:56:01 +0000 (12:56 +0200)]
qemu: Minimalize global driver accesses
Whilst working on another issue, I've noticed that in some
functions we have a local @driver variable among with access to
global @qemu_driver variable. This makes no sense.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Rather than waiting until we've free'd up all the resources, cause the
'workerPool' thread pool to flush as soon as possible during stateCleanup.
Otherwise, it's possible something waiting to run will SEGV such as is the
case during race conditions of simultaneous exiting libvirtd and qemu process.
Resolves the following crash:
[1] crash backtrace: (bt is shortened a bit):
0 0x00007ffff7282f2b in virClassIsDerivedFrom
(klass=0xdeadbeef, parent=0x55555581d650) at util/virobject.c:169
1 0x00007ffff72835fd in virObjectIsClass
(anyobj=0x7fffd024f580, klass=0x55555581d650) at util/virobject.c:365
2 0x00007ffff7283498 in virObjectLock
(anyobj=0x7fffd024f580) at util/virobject.c:317
3 0x00007ffff722f0a3 in virCloseCallbacksUnset
(closeCallbacks=0x7fffd024f580, vm=0x7fffd0194db0,
cb=0x7fffdf1af765 <qemuProcessAutoDestroy>)
at util/virclosecallbacks.c:164
4 0x00007fffdf1afa7b in qemuProcessAutoDestroyRemove
(driver=0x7fffd00f3a60, vm=0x7fffd0194db0) at qemu/qemu_process.c:6365
5 0x00007fffdf1adff1 in qemuProcessStop
(driver=0x7fffd00f3a60, vm=0x7fffd0194db0, reason=VIR_DOMAIN_SHUTOFF_CRASHED,
asyncJob=QEMU_ASYNC_JOB_NONE, flags=0)
at qemu/qemu_process.c:5877
6 0x00007fffdf1f711c in processMonitorEOFEvent
(driver=0x7fffd00f3a60, vm=0x7fffd0194db0) at qemu/qemu_driver.c:4545
7 0x00007fffdf1f7313 in qemuProcessEventHandler
(data=0x555555832710, opaque=0x7fffd00f3a60) at qemu/qemu_driver.c:4589
8 0x00007ffff72a84c4 in virThreadPoolWorker
(opaque=0x555555805da0) at util/virthreadpool.c:167
Thread 1 (Thread 0x7ffff7fb1880 (LWP 494472)):
1 0x00007ffff72a7898 in virCondWait
(c=0x7fffd01c21f8, m=0x7fffd01c21a0) at util/virthread.c:154
2 0x00007ffff72a8a22 in virThreadPoolFree
(pool=0x7fffd01c2160) at util/virthreadpool.c:290
3 0x00007fffdf1edd44 in qemuStateCleanup ()
at qemu/qemu_driver.c:1102
4 0x00007ffff736570a in virStateCleanup ()
at libvirt.c:807
5 0x000055555556f991 in main (argc=1, argv=0x7fffffffe458) at libvirtd.c:1660
Do not dereference the 'dmn' until after the virStateCleanup is completed.
During initialization, virStateInitialize requires/uses the "dmn" as the
argument to/for the daemonInhibitCallback functions. Thus, cleanup cannot
dereference 'dmn' until after calling the virStateCleanup which calls the
the daemonInhibitCallback using 'dmn'; otherwise, the following crash occurs:
backtrace (shortened a bit)
1 0x00007fd3a791b2e6 in virCondWait (c=<optimized out>, m=<optimized out>)
at util/virthread.c:154
2 0x00007fd3a791bcb0 in virThreadPoolFree (pool=0x7fd38024ee00)
at util/virthreadpool.c:266
3 0x00007fd38edaa00e in qemuStateCleanup () at qemu/qemu_driver.c:1116
4 0x00007fd3a79abfeb in virStateCleanup () at libvirt.c:808
5 0x00007fd3a85f2c9e in main (argc=<optimized out>, argv=<optimized out>)
at libvirtd.c:1660
Thread 1 (Thread 0x7fd38722d700 (LWP 32256)):
0 0x00007fd3a7900910 in virClassIsDerivedFrom
(klass=0xdfd36058d4853, parent=0x7fd3a8f394d0) at util/virobject.c:169
1 0x00007fd3a7900c4e in virObjectIsClass
(anyobj=anyobj@entry=0x7fd3a8f2f850, klass=<optimized out>)
at util/virobject.c:365
2 0x00007fd3a7900c74 in virObjectLock (anyobj=0x7fd3a8f2f850)
at util/virobject.c:317
3 0x00007fd3a7a24d5d in virNetDaemonRemoveShutdownInhibition
(dmn=0x7fd3a8f2f850) at rpc/virnetdaemon.c:547
4 0x00007fd38ed722cf in qemuProcessStop
(driver=driver@entry=0x7fd380103810, vm=vm@entry=0x7fd38025b6d0,
reason=reason@entry=VIR_DOMAIN_SHUTOFF_SHUTDOWN,
asyncJob=asyncJob@entry=QEMU_ASYNC_JOB_NONE, flags=flags@entry=0)
at qemu/qemu_process.c:5786
5 0x00007fd38edd9428 in processMonitorEOFEvent
(vm=0x7fd38025b6d0, driver=0x7fd380103810) at qemu/qemu_driver.c:4588
6 qemuProcessEventHandler (data=<optimized out>, opaque=0x7fd380103810)
at qemu/qemu_driver.c:4632
7 0x00007fd3a791bb55 in virThreadPoolWorker
(opaque=opaque@entry=0x7fd3a8f1e4c0) at util/virthreadpool.c:145
vz: support type=bridge network interface type correctly
Recently, libprlsdk got a separate flag PNA_BRIDGE corresponding to
type=bridge libvirt network interfaces. Let's use it and get rid of
all workarounds previously added to support it.
Chen Hanxiao [Sat, 22 Oct 2016 09:49:38 +0000 (17:49 +0800)]
qemu: Forbid pinning vCPUs for TCG domain
We don't support cpu pinning for TCG domains because QEMU runs them in
one thread only. But vcpupin command was able to set them, which
resulted in a failed startup, so make sure that doesn't happen.
Ján Tomko [Fri, 21 Oct 2016 12:18:54 +0000 (14:18 +0200)]
Do not try to release virtio serial addresses
Return 0 instead of 1, so that qemuDomainAttachChrDevice does not
assume the address neeeds to be released on error.
No functional change, since qemuDomainReleaseDeviceAddress has been a noop
for virtio serial addresses since the address cache was removed
in commit 19a148b.
Ján Tomko [Fri, 21 Oct 2016 10:58:54 +0000 (12:58 +0200)]
Add 'FromCache' to virDomainVirtioSerialAddrAutoAssign
Commit 19a148b dropped the cache from QEMU's private domain object.
Assume the callers do not have the cache by default and use
a longer name for the internal ones that do.
This makes the shorter 'virDomainVirtioSerialAddrAutoAssign'
name availabe for a function that will not require the cache.
Sławek Kapłoński [Sat, 22 Oct 2016 10:30:01 +0000 (12:30 +0200)]
qemu, lxc: Raise error message when resuming running domain
When user tries to resume already running domain (Qemu or LXC)
VIR_ERR_OPERATION_INVALID error should be raised with message that
domain is already running.
Andrea Bolognani [Wed, 26 Oct 2016 12:00:16 +0000 (14:00 +0200)]
wireshark: Rename plugindir to ws_plugindir
Since we're using autoconf to substitute the right value in
Makefile.am now, we can use a less generic name without running
into circular dependencies.
Andrea Bolognani [Wed, 26 Oct 2016 11:56:52 +0000 (13:56 +0200)]
wireshark: Inject $(prefix) at the right time
Adding $(prefix) in Makefile.am, as we were doing, means that
it would be prepended even when using --with-ws-plugindir,
which is something we don't want to happen.
Instead, we add it beforehand but take care that it doesn't
get expanded until make is called.
Andrea Bolognani [Wed, 26 Oct 2016 11:52:08 +0000 (13:52 +0200)]
wireshark: Strip prefix correctly
Even when we're building $plugindir ourselves because we can't
retrieve it using pkg-config, we still want to strip the prefix,
except in that case it would be the same prefix we're using for
building libvirt.
The fact that $plugindir is missing also doesn't tell us
anything about $ws_prefix, so we have to handle the two variables
separately.
Gema Gomez [Sun, 9 Oct 2016 15:51:09 +0000 (16:51 +0100)]
qemu: Add support for using AES secret for SCSI hotplug
Support for virtio disks was added in commit id 'fceeeda', but not for
SCSI drives. Add the secret for the server when hotplugging a SCSI drive.
No need to make any adjustments for unplug since that's handled during
the qemuDomainDetachDiskDevice call to qemuDomainRemoveDiskDevice in
the qemuDomainDetachDeviceDiskLive switch.
Added a test to/for the command line processing to show the command line
options when adding a SCSI drive for the guest.
Complete the implementation of support for TLS encryption on
chardev TCP transports by adding the hotplug ability of a secret
to generate the passwordid for the TLS object for chrdev, RNG,
and redirdev.
Fix up the order of object removal on failure to be the inverse
of the attempted attach (for redirdev, chr, rng) - for each the
tls object was being removed before the chardev backend.
Likewise, add the ability to hot unplug that secret object as well
and be sure the order of unplug matches that inverse order of plug.
John Ferlan [Fri, 21 Oct 2016 23:02:35 +0000 (19:02 -0400)]
qemu: Add a secret object to/for a char source dev
Add the secret object so the 'passwordid=' can be added if the command line
if there's a secret defined in/on the host for TCP chardev TLS objects.
Preparation for the secret involves adding the secinfo to the char source
device prior to command line processing. There are multiple possibilities
for TCP chardev source backend usage.
Add test for at least a serial chardev as an example.
John Ferlan [Tue, 25 Oct 2016 19:32:27 +0000 (15:32 -0400)]
qemu: Fix depedency order in qemuRemoveDiskDevice
Need to remove the drive first, then the secobj and/or encobj if they exist.
This is because the drive has a dependency on secobj (or the secret for
the networked storage server) and/or the encobj (or the secret for the
LUKS encrypted volume). Deleting either object first leaves an drive
without it's respective objects.
John Ferlan [Thu, 6 Oct 2016 19:08:00 +0000 (15:08 -0400)]
qemu: Introduce qemuDomainSetBlockIoTuneDefaults
Create a helper to set the bytes/iops iotune default values based on
the current qemu setting for both the live and persistent definitions.
NB: This also fixes an unreported bug where the persistent values for
*_max and size_iops_sec would be set back to 0 if unrelated persistent
values were set.
John Ferlan [Mon, 24 Oct 2016 19:47:56 +0000 (15:47 -0400)]
qemu: Move TLS object remove from DetachChr to RemoveChr
Commit id '2c32237' added the TLS object removal to the DetachChrDevice
all when it should have been added to the RemoveChrDevice since that's
the norm for similar processing (e.g. disk)
Ján Tomko [Mon, 24 Oct 2016 15:35:32 +0000 (17:35 +0200)]
virQEMUCapsReset: also clear out hostCPUModel
After succesfully reading an outdated caps cache from disk,
calling virQEMUCapsReset did not properly clear out the calculated
host CPU model. This lead to a memory leak when the host CPU model
pointer was overwritten later in virQEMUCapsNewForBinaryInternal.
Jim Fehlig [Fri, 14 Oct 2016 17:55:52 +0000 (11:55 -0600)]
libxl: fix leaking of allocated migration ports
Although the migration port is immediately released in the
finish phase of migration, it was never set in the domain
private object when allocated in the prepare phase. So
libxlDomainMigrationFinish() always released a 0-initialized
migrationPort, leaking any allocated port. After enough
migrations to exhaust the migration port pool, migration would
fail with
error: internal error: Unable to find an unused port in range
'migration' (49152-49216)
Fix it by setting libxlDomainObjPrivate->migrationPort to the
port allocated in the prepare phase. While at it, also fix
leaking an allocated port if the prepare phase fails.
doc: update virsh domstats documentation for vcpu statistics
Added description for new vcpu.<num>.halted statistics value.
While there, also added a description for vcpu.<num>.wait and
clarified the units displayed for time and wait.
Signed-off-by: Viktor Mihajlovski <mihajlov@linux.vnet.ibm.com> Reviewed-by: Bjoern Walk <bwalk@linux.vnet.ibm.com>
Extended qemuDomainGetStatsVcpu to include the per vcpu halted
indicator if reported by QEMU. The key for new boolean value
has the format "vcpu.<n>.halted".
Signed-off-by: Viktor Mihajlovski <mihajlov@linux.vnet.ibm.com> Reviewed-by: Bjoern Walk <bwalk@linux.vnet.ibm.com>
Adding a field to the domain's private vcpu object to hold the halted
state information.
Adding two functions in support of the halted state:
- qemuDomainGetVcpuHalted: retrieve the halted state from a
private vcpu object
- qemuDomainRefreshVcpuHalted: obtain the per-vcpu halted states
via qemu monitor and store the results in the private vcpu objects
Signed-off-by: Viktor Mihajlovski <mihajlov@linux.vnet.ibm.com> Reviewed-by: Bjoern Walk <bwalk@linux.vnet.ibm.com> Reviewed-by: Hao QingFeng <haoqf@linux.vnet.ibm.com> Signed-off-by: Boris Fiuczynski <fiuczy@linux.vnet.ibm.com>
Laine Stump [Fri, 2 Sep 2016 20:39:18 +0000 (16:39 -0400)]
qemu: replace calls to virDomainPCIAddressReserveNext*() with static function
An upcoming commit will remove the "flag" argument from all the calls
to reserve the next available address|slot, but I don't want to change
the arguments in the hypervisor-agnostic
virDomainPCIAddressReserveNext*() functions, so this patch places a
simple qemu-specific wrapper around those functions - the new
functions don't take a flags arg, but grab it from the device's
info->pciConnectFlags.
projects / libvirt.git / log