Cédric Bosdonnat [Wed, 11 Dec 2013 10:38:04 +0000 (11:38 +0100)]
Added default case with error for object event dispatching
Hitting this should be pretty rare, but at least developers will know
that they are providing a weird event ID. Otherwise for namespace that
are added in the normal way, gcc will raise a warning about unhandled
case in the switch.
Cédric Bosdonnat [Wed, 11 Dec 2013 10:37:58 +0000 (11:37 +0100)]
Added Network events API and virNetworkEventLifecycle.
Define the public API for (de-)registering network events
and the callbacks for receiving lifecycle events. The lifecycle
event includes a 'detail' parameter to match the domain lifecycle
event data, but this is currently unused.
The network events related code goes into its own set of internal
files src/conf/network_event.[ch]
Eric Blake [Tue, 10 Dec 2013 12:45:26 +0000 (05:45 -0700)]
storage: fix omitted slash in gluster volume URI
When doing 'virsh vol-dumpxml' on a gluster pool's volume, the
resulting URI incorrectly omitted a slash between hostname and
path: gluster://192.168.122.206rhsvol1/fedora-19.img
This is fallout from me rebasing earlier versions of my patch
that ended up as commit efee1af; I had originally played with
always requiring the gluster volume to have a leading slash,
but it was easier to use the gluster API if the gluster volume
name was guaranteed to have no slash. While I got the URI of
the pool correct, I forgot to fix the URI of a libvirt volume.
* src/storage/storage_backend_gluster.c
(virStorageBackendGlusterRefreshVol): Use correct starting point
since uri construction requires leading slash.
There's a window when starting a qemu process between fork() and exec()
during which we are doing things that may fail but not tunnelling the
error to the daemon. This is basically all within qemuProcessHook().
So whenever we fail in something, e.g. placing a process onto numa node,
users are left with:
error: Child quit during startup handshake: Input/output error
while the original error is thrown into the domain log:
libvirt: error : internal error: NUMA memory tuning in 'preferred'
mode only supports single node
Hence, we should read the log file and search for the error message and
report it to users.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Cédric Bosdonnat [Fri, 22 Nov 2013 14:55:12 +0000 (15:55 +0100)]
Add object event namespaces for the event IDs
Each unique event ID will thus be composed by 1 byte for the namespace
and 1 byte for a namespace-specific ID. The namespace for domain event
needs to be 0 for compatibility reasons.
Michael Chapman [Mon, 9 Dec 2013 06:23:30 +0000 (17:23 +1100)]
spec: clean up libvirtd and virtlockd service mgmt
- systemctl and the %systemd_* RPM macros can take multiple unit names
in the one invocation. Make use of this to avoid repeated systemd
daemon reloads.
- virtlockd was only properly enabled and disabled when using systemd,
but when systemd RPM macros were not available (e.g. on Fedora < 18).
Make sure it's enabled when systemd RPM macros are present, or when
using initscripts.
- Always use "reload" on virtlockd, not "condrestart". This allows it to
cleanly re-execute itself without losing running state. Ignore any
error should the reload fail.
- Move the reloading of virtlockd and libvirtd via their initscripts
into the daemon package's %postun scriptlet. These services must be
restarted after all of the libvirt-daemon-driver-* packages have
been upgraded during the same RPM transaction.
- Add a %triggerpostun executed only when upgrading an older
libvirt-daemon. As an older package would only reload libvirtd during
%post, and the newer package would only reload libvirtd during
%postun, such an upgrade would not reload libvirtd at all without the
trigger.
Signed-off-by: Michael Chapman <mike@very.puzzling.org>
Michael Chapman [Mon, 9 Dec 2013 06:23:27 +0000 (17:23 +1100)]
virtlockd: treat SIGHUP like SIGUSR1
SIGHUP is commonly used to instruct a daemon to reload its config. For
now we should handle it in virtlockd just like SIGUSR1, rather than
having it kill the process.
Signed-off-by: Michael Chapman <mike@very.puzzling.org> Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Michael Chapman [Mon, 9 Dec 2013 06:23:26 +0000 (17:23 +1100)]
virtlockd: improve initscripts
- Use SIGUSR1, not SIGHUP, on reload. At present, virtlockd only
responds to the former.
- Fix PID file for virtlockd.
- Do not start virtlockd in any runlevels by default. It needs to be
explicitly selected in libvirt's qemu.conf anyway, so there is no
need to have it running on all systems regardless.
- Fix chkconfig priorities to ensure virtlockd is started before
libvirtd is started, and stopped after libvirtd is stopped.
- Add "Should-Start: virtlockd" to the libvirtd initscript's LSB header,
for the same reason.
- Add "Default-Stop" to both libvirtd and virtlockd initscripts. LSB
does not guarantee that this defaults to the inverse of
"Default-Start".
Signed-off-by: Michael Chapman <mike@very.puzzling.org> Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Michael Chapman [Mon, 9 Dec 2013 06:23:25 +0000 (17:23 +1100)]
virtlockd: improve systemd units
- Pass VIRTLOCKD_ARGS through to virtlockd.
- Use SIGUSR1, not SIGHUP, in ExecReload. At present, virtlockd only
responds to the former.
- Have "systemctl enable virtlockd.service" enable virtlockd.socket,
rather than throw an error.
- Make virtlockd.socket wanted by sockets.target, rather than
multi-user.target. This is consistent with other socket units in
Fedora, and it ensures that the socket is available before libvirtd is
started.
Signed-off-by: Michael Chapman <mike@very.puzzling.org>
When changing memtune limits to unlimited with AFFECT_CONFIG, the
values in virDomainDef are set to PARAM_UNLIMITED, which causes the
whole <memtune> to be formatted. This can be changed in all drivers,
but it also makes sense to use the default (0) as another value for
"unlimited", since zero memory limit makes no sense.
Signed-off-by: Martin Kletzander <mkletzan@redhat.com>
cgroups: Redefine what "unlimited" means wrt memory limits
Since kernel 3.12 (commit 34ff8dc08956098563989d8599840b130be81252 in
linux-stable.git in particular) the value for 'unlimited' in cgroup
memory limits changed from LLONG_MAX to ULLONG_MAX. Due to rather
unfortunate choice of our VIR_DOMAIN_MEMORY_PARAM_UNLIMITED constant
(which we transfer as an unsigned long long in Kibibytes), we ended up
with the situation described below (applies to x86_64):
- 2^64-1 (ULLONG_MAX) -- "unlimited" in kernel = 3.12
- 2^63-1 (LLONG_MAX) -- "unlimited" in kernel < 3.12
- 2^63-1024 -- our PARAM_UNLIMITED scaled to Bytes
- 2^53-1 -- our PARAM_UNLIMITED unscaled (in Kibibytes)
This means that when any number within (2^63-1, 2^64-1] is read from
memory cgroup, we are transferring that number instead of "unlimited".
Unfortunately, changing VIR_DOMAIN_MEMORY_PARAM_UNLIMITED would break
ABI compatibility and thus we have to resort to a different solution.
With this patch every value greater than PARAM_UNLIMITED means
"unlimited". Even though this may seem misleading, we are already in
such unclear situation when running 3.12 kernel with memory limits set
to 2^63.
One example showing most of the problems at once (with kernel 3.12.2):
# virsh memtune asdf --hard-limit 9007199254740991 --swap-hard-limit -1
# echo 12345678901234567890 >\
/sys/fs/cgroup/memory/machine/asdf.libvirt-qemu/memory.soft_limit_in_bytes
# virsh memtune asdf
hard_limit : 18014398509481983
soft_limit : 12056327051986884
swap_hard_limit: 18014398509481983
Signed-off-by: Martin Kletzander <mkletzan@redhat.com>
Cole Robinson [Thu, 5 Dec 2013 20:03:00 +0000 (15:03 -0500)]
qemu: hotplug: Fix double free on USB collision
If we hit a collision, we free the USB device while it is still part
of our temporary USBDeviceList. When the list is unref'd, the device
is free'd again.
Make the initial device freeing dependent on whether it is present
in the temporary list or not.
Cole Robinson [Thu, 5 Dec 2013 19:54:41 +0000 (14:54 -0500)]
qemu: hotplug: Only label hostdev after checking device conflicts
Similar to what Jiri did for cgroup setup/teardown in 05e149f94, push
it all into the device handler functions so we can do the necessary prep
work before claiming the device.
This also fixes hotplugging USB devices by product/vendor (virt-manager's
default behavior):
Jim Fehlig [Thu, 5 Dec 2013 23:15:55 +0000 (16:15 -0700)]
spec: move driver-specific files to driver subpackages
The libvirt-daemon package contains several driver-specific files,
directories, and script, which can be problematic when building the
package with multiple hypervisor support, e.g. both QEMU and Xen.
E.g. installing a QEMU+Xen enabled libvirt-daemon on a Xen-only system
will result in the creation of qemu and kvm groups and a qemu user.
Move the driver-specific files, directories, and script to the
respective driver subpackages.
Jim Fehlig [Thu, 5 Dec 2013 21:56:06 +0000 (14:56 -0700)]
spec: Fix files list when building without driver modules
The daemon-config-{network,nwfilter} subpackages are built regardless
of whether or not with_driver_modules is defined, therefore don't
conditionally define their files list.
Currently, if virFileMakePath() fails, the @ret is left initialized from
virAsprintf() just a few lines above leading to a wrong return value of
zero whereas -1 should be returned.
Signed-off-by: Chen Hanxiao <chenhanxiao@cn.fujitsu.com> Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
When attempting to enable more vCPUs in the guest than is currently
enabled in the guest but less than the maximum count for the VM we
currently reported an unhelpful message:
error: internal error: guest agent reports less cpu than requested
This patch changes it to:
error: invalid argument: requested vcpu count is greater than the count
of enabled vcpus in the domain: 3 > 2
When outputting the XML for the RNG device, the code didn't format the
PCI address info. Additionally the schema wasn't expecting the info
although it was being parsed and used internally. Fix those mistakes and
add test for the PCI info section.
Peter Krempa [Thu, 5 Dec 2013 14:58:31 +0000 (15:58 +0100)]
qemu: snapshot: Fix incorrect disk type for auto-generated disks
When changing the parsing and formatting functions in commit 43f2ccdc73090bd03f64de4d58d46ffa0134d705 I forgot to update the qemu
disk alignment function for snapshots that automatically adds snapshot
configs for disks that were not mentioned in the XML. The function
allocated a new disk snapshot definition but did not correctly
initialize the snapshot disk source type variable. This resulted into
the disks considered as block devices and invalid XML was generated.
virThreadPoolFree: Set n(Prio)Workers after the pool is locked
In 78839da I am trying to join the worker threads. However, I can't
sipmly reuse pool->nWorkers (same applies for pool->nPrioWorkers),
because of the following flow that is currently implemented:
1) the main thread executing virThreadPoolFree sets pool->quit = true,
wakes up all the workers and wait on pool->quit_cond.
2) A worker is woken up and see quit request. It immediately jumps of
the while() loop and decrements pool->nWorkers (or pool->nPrioWorkers in
case of priority worker). The last thread signalizes pool->quit_cond.
3) Main thread is woken up, with both pool->nWorkers and
pool->nPrioWorkers being zero.
So there's a need to copy the original value of worker thread counts
into local variables. However, these need to set *after* the check for
pool being NULL (dereferencing a NULL is no no). And for safety they can
be set right after the pool is locked.
Reported-by: John Ferlan <jferlan@redhat.com> Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Peter Krempa [Wed, 4 Dec 2013 17:31:39 +0000 (18:31 +0100)]
domainsnapshotxml2xmltest: Add existing files as new tests
There were plenty snapshot XMLs in the tests/domainsnapshotxml2xmlin
directory that actually weren't used in XML testing. The upgraded
domainsnapshotxml2xml test now allows us to use them.
Peter Krempa [Wed, 4 Dec 2013 13:53:24 +0000 (14:53 +0100)]
domainsnapshotxml2xmltest: Allow for better testing of snapshots
Until now the test was only testing redefinition of snapshot XMLs stored
in tests/domainsnapshotxml2xmlout. This patch adds new infrastructure to
allow testing of files that may differ and will allow to utilize files
in tests/domainsnapshotxml2xmlin as new tests too.
Peter Krempa [Wed, 4 Dec 2013 12:49:40 +0000 (13:49 +0100)]
domainsnapshotxml2xmltest: Clean up labels and use bool instead of int
The 'internal' variable holds only two states; convert it to a boolean
and the 'fail' label should be called 'cleanup'. This patch also fixes a
minor memory leak of driver capabilities in case the XML config object
can't be allocated.
Adam Walters [Thu, 5 Dec 2013 04:09:14 +0000 (23:09 -0500)]
configure: Resolve compile issue in configure.ac
This patch resolves a compile issue caused by the
removal of examples/domsuspend code in commit 5eb4b04211ea379c58f735dee6c2852c8b80da89. This issue
is only seen in a fresh checkout, but causes the build
and configure to fail.
When an error occurred in qemuAgentIO, it will be saved in mon->lastError,
but it will not be freed at the end. Present since commit c160ce33;
and compare to commit 9cc8a5af fixing the same problem in qemu_monitor.c.
==22219== 54 bytes in 1 blocks are definitely lost in loss record 982 of 1,379
==22219== at 0x4C26B9B: malloc (vg_replace_malloc.c:263)
==22219== by 0x8520521: strdup (in /lib64/libc-2.11.3.so)
==22219== by 0x52E99CB: virStrdup (virstring.c:554)
==22219== by 0x52B44C4: virCopyError (virerror.c:195)
==22219== by 0x52B5123: virCopyLastError (virerror.c:312)
==22219== by 0x10905877: qemuAgentIO (qemu_agent.c:660)
==22219== by 0x52B6122: virEventPollDispatchHandles (vireventpoll.c:501)
==22219== by 0x52B7AEA: virEventPollRunOnce (vireventpoll.c:647)
==22219== by 0x52B5C1B: virEventRunDefaultImpl (virevent.c:274)
==22219== by 0x54181FD: virNetServerRun (virnetserver.c:1112)
==22219== by 0x11EF4D: main (libvirtd.c:1513)
Signed-off-by: Zhou Yimin <zhouyimin@huawei.com> Signed-off-by: Eric Blake <eblake@redhat.com>
This patch fixes memory leaks reported by valgrind on running
qemuxml2argvtest; introduced in commit 0df53f04.
Most of them are of the form:
==24777== 15 bytes in 1 blocks are definitely lost in loss record 39 of 129
==24777== at 0x4A0887C: malloc (vg_replace_malloc.c:270)
==24777== by 0x341F485E21: strdup (strdup.c:42)
==24777== by 0x4CADE5F: virStrdup (virstring.c:554)
==24777== by 0x4362B6: qemuBuildDriveStr (qemu_command.c:3848)
==24777== by 0x43EF73: qemuBuildCommandLine (qemu_command.c:8500)
==24777== by 0x426670: testCompareXMLToArgvHelper (qemuxml2argvtest.c:350)
==24777== by 0x427C01: virtTestRun (testutils.c:138)
==24777== by 0x41DDB5: mymain (qemuxml2argvtest.c:658)
==24777== by 0x4282A2: virtTestMain (testutils.c:593)
==24777== by 0x341F421A04: (below main) (libc-start.c:225)
==24777==
Peter Krempa [Thu, 14 Nov 2013 16:14:26 +0000 (17:14 +0100)]
maint: Kill usage of atoi()
Kill the use of atoi() and introduce syntax check to forbid it and it's
friends (atol, atoll, atof, atoq).
Also fix a typo in variable name holding the cylinders count of a disk
pool (apparently unused).
examples/domsuspend/suspend.c will need a larger scale refactor as the
whole example file is broken thus it will be exempted from the syntax
check for now.
Peter Krempa [Wed, 4 Dec 2013 09:43:43 +0000 (10:43 +0100)]
examples: Remove broken bad example
The domsuspend example code is a really old and bad exmample of (how not
to use) the libvirt API. Remove it as it's apparent that nobody tried to
use it. It was broken and nobody complained.
Michal Privoznik [Wed, 13 Nov 2013 13:40:12 +0000 (14:40 +0100)]
virThreadPoolFree: Join worker threads
Even though currently we are freeing the pool of worker threads at the
daemon very end, nothing holds us back in joining the worker threads.
Moreover, we avoid leaks like this:
==26697== 1,680 bytes in 5 blocks are possibly lost in loss record 913 of 942
==26697== at 0x4C2BDE4: calloc (in /usr/lib64/valgrind/vgpreload_memcheck-amd64-linux.so)
==26697== by 0x4011131: allocate_dtv (in /lib64/ld-2.16.so)
==26697== by 0x401176D: _dl_allocate_tls (in /lib64/ld-2.16.so)
==26697== by 0x8499602: pthread_create@@GLIBC_2.2.5 (in /lib64/libpthread-2.16.so)
==26697== by 0x52F53E9: virThreadCreate (virthreadpthread.c:188)
==26697== by 0x52F5D4F: virThreadPoolNew (virthreadpool.c:221)
==26697== by 0x53F30DB: virNetServerNew (virnetserver.c:377)
==26697== by 0x11C6ED: main (libvirtd.c:1366)
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
qemu: fix live pinning to memory node on NUMA system
Ever since the subcpusets(vcpu,emulator) were introduced, the parent
cpuset cannot be modified to remove the nodes that are in use by the
subcpusets.
The fix is to break the memory node modification into three steps:
1. assign new nodes into the parent,
2. change the nodes in the child nodes,
3. remove the old nodes on the parent node.
The storageRegister() didn't check the return from the
virRegisterStorageDriver() like other callers did, so Coverity
flagged it. Just check the return and handle.
The x509dname is only set inside a WITH_GNUTLS conditional, so
when used/check later on for NULL, Coverity detects this is not
possible. Added WITH_GNUTLS around uses to remove message
John Ferlan [Mon, 2 Dec 2013 19:36:51 +0000 (14:36 -0500)]
nwfilter: Remove Coverity DEADCODE warning
The nwfilterStateInitialize() would only assign sysbus inside
a WITH_DBUS conditional, thus leaving a subsequent check for sysbus
and nwfilterDriverInstallDBusMatches() as a no-op
Rather than try to add WITH_DBUS conditions which ended up conflicting
with the usage of HAVE_FIREWALLD conditionals, just remove the WITH_DBUS
since virdbus.c has entry points for with and without conditions.
GEN access/viraccessapichecklxc.h
GEN hyperv/hyperv_wmi.generated.h
GEN access/viraccessapichecklxc.c
GEN hyperv/hyperv_wmi.generated.c
GEN hyperv/hyperv_wmi_classes.generated.typedef
GEN hyperv/hyperv_wmi_classes.generated.h
GEN hyperv/hyperv_wmi_classes.generated.c
GEN libvirt_access_qemu.xml
GEN libvirt_access.syms
GEN libvirt_access_lxc.xml
GEN libvirt_access_qemu.syms
GEN libvirt_access_lxc.syms
GEN libvirt_qemu.def
GEN esx/esx_vi_types.generated.typedef
GEN esx/esx_vi_types.generated.typeenum
GEN esx/esx_vi_types.generated.typetostring
GEN esx/esx_vi_types.generated.typefromstring
GEN esx/esx_vi_types.generated.h
GEN esx/esx_vi_types.generated.c
GEN esx/esx_vi_methods.generated.h
GEN esx/esx_vi_methods.generated.c
GEN esx/esx_vi_methods.generated.macro
GEN esx/esx_vi.generated.h
GEN esx/esx_vi.generated.c
GEN libvirt_lxc.def
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
The BZ asked for the capability to change the number of queues used by
a virtio-net device while the device is in use. Because the number of
queues can only be set at the time the device is created, that isn't
possible. However, libvirt also shouldn't be silently reporting
success when someone tries to change the number of queues. So this
patch flags that as an error (just as attempts to change any of the
other virtio-specific parameters already do).
Currently, initialization of drivers is done in a separate thread. This
is done for several reasons: a driver that is initialized may require
running event loop, it may take ages to initialize driver (e.g. due to
autostarting domains). While the thread is spawn and run, the main()
continues its execution. However, if something goes bad, or the event
loop is just exited (e.g. due to a --timeout or SIGINT) we try to
cleanup all the drivers. So we have two threads running Initialize() and
Cleanup() concurrently. This may result in accessing stale pointers -
e.g. netcf driver will free() itself in stateCleanup callback, while the
init thread may come, open a dummy connection in order to autostart some
domains and voilà: do_open() iterates over interface drivers and
accesses stale netcf driver.
The fix consists in not running stateCleanup if the init thread is still
running.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
(which was already closed as CANTFIX because the qemu "-boot strict"
commandline option wasn't available at the time).
Problem: you couldn't have a domain that used PXE to boot, but also
had an un-bootable disk device *even if that disk wasn't listed in the
boot order*, because if PXE timed out (e.g. due to the bridge
forwarding delay), the BIOS would move on to the next target, which
would be the unbootable disk device (again - even though it wasn't
given a boot order), and get stuck at a "BOOT DISK FAILURE, PRESS ANY
KEY" message until a user intervened.
The solution available since sometime around QEMU 1.5, is to add
"-boot strict=on" to *every* qemu command. When this is done, if any
devices have a boot order specified, then QEMU will *only* attempt to
boot from those devices that have an explicit boot order, ignoring the
rest.
Commit f094aaac48a6 changed the PCI device assignment in qemu domains
to default to using VFIO rather than legacy KVM device assignment
(when VFIO is available). It didn't change which driver was used by
default for virNodeDeviceDetachFlags(), though, so that API (and the
virsh nodedev-detach command) was still binding to the pci-stub
driver, used by legacy KVM assignment, by default.
This patch publicizes (only within the qemu module, though, so no
additions to the symbol exports are needed) the functions that check
for presence of KVM and VFIO device assignment, then uses those
functions to decide what to do when no driver is specified for
virNodeDeviceDetachFlags(); if the vfio driver is loaded, the device
will be bound to vfio-pci, or if legacy KVM assignment is supported on
this system, the device will be bound to pci-stub; if neither method
is available, the detach will fail.
Peter Krempa [Mon, 25 Nov 2013 16:51:04 +0000 (17:51 +0100)]
qemu: snapshots: Declare supported and unsupported snapshot configs
Currently the snapshot code did not check if it actually supports
snapshots on various disk backends for domains. To avoid future problems
add checkers that whitelist the supported configurations.
Peter Krempa [Thu, 21 Nov 2013 17:43:59 +0000 (18:43 +0100)]
qemu: Refactor disk source string formatting
This patch adds function qemuGetDriveSourceString to produce
qemu-compatible disk source strings that will enable to reuse the code
and refactors building of the qemu commandline of disks to use this new
helper.