https://bugzilla.redhat.com/show_bug.cgi?id=1200149
Even though we have a mutex mechanism so that two clients don't spawn
two daemons, it's not strong enough. It can happen that while one
client is spawning the daemon, the other one fails to connect.
Basically two possible errors can happen:
error: Failed to connect socket to '/home/mprivozn/.cache/libvirt/libvirt-sock': Connection refused
or:
error: Failed to connect socket to '/home/mprivozn/.cache/libvirt/libvirt-sock': No such file or directory
The problem in both cases is, the daemon is only starting up, while we
are trying to connect (and fail). We should postpone the connecting
phase until the daemon is started (by the other thread that is
spawning it). In order to do that, create a file lock 'libvirt-lock'
in the directory where session daemon would create its socket. So even
when called from multiple processes, spawning a daemon will serialize
on the file lock. So only the first to come will spawn the daemon.
Tested-by: Richard W. M. Jones <rjones@redhat.com>
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Two non-static functions in virjson.c were missing their export info in
libvirt_private.syms, so they couldn't be used anywhere it the code (and
that's about to get changed).
Signed-off-by: Martin Kletzander <mkletzan@redhat.com>
Luckily we are allocating structs as clean memory and
PTHREAD_MUTEX_INITIALIZER is "{ 0 }", so nothing happened, but it should
still be created as lockable object.
Signed-off-by: Martin Kletzander <mkletzan@redhat.com>
Check the proposed pool source host XML definition against existing gluster
pools to ensure the incoming definition doesn't use the same source dir and
soure host XML definition as an existing pool.
Check the proposed pool source host XML definition against existing sheepdog
pools to ensure the incoming definition doesn't use the same source host XML
definition as an existing pool.
Rather than have duplicate code doing the same check, have the netfs
matching processing code use the new virStoragePoolSourceMatchSingleHost.
Signed-off-by: John Ferlan <jferlan@redhat.com>
Create a separate iSCSI Source matching subroutine. Makes the calling
code a bit cleaner as well as sets up for future patches which need to
do better source hosts[0].name processing/checking.
As part of the effort the logic will be inverted from a multi-level
if statement to a series of single level checks for better readability
and further separation
Signed-off-by: John Ferlan <jferlan@redhat.com>
When acquiring resource via sanlock fails, we would report it as
VIR_ERR_INTERNAL_ERROR, which is not very friendly to applications using
libvirt. Moreover, the lockd driver would report the same failure as
VIR_ERR_RESOURCE_BUSY, which looks better.
Unfortunately, in sanlock driver we don't really know if acquiring the
resource failed because it was already locked or there was another
reason behind. But the end result is the same and I think using
VIR_ERR_RESOURCE_BUSY reason for all acquire failures is still better
than what we have now.
https://bugzilla.redhat.com/show_bug.cgi?id=1165119
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
Commit 49ed6cff is broken on mingw and other non-linux platforms:
CCLD libvirt.la
Cannot export virNetDevSysfsFile: symbol not defined
collect2: error: ld returned 1 exit status
* src/util/virnetdev.c: Provide virNetDevSysfsFile fallback.
Signed-off-by: Eric Blake <eblake@redhat.com>
Found by ./autobuild.sh during a mingw cross-compile:
Commit 8a96e87 was not innocuous - glibc happens to leak the
definition of time() through other headers, so that even without
<sys/select.h>, virrandom.c compiled just fine. But on mingw,
we were not so lucky; <sys/select.h> was important for its side
effect of dragging in <time.h>, and we now have nothing providing
the declaration of time():
../../src/util/virrandom.c: In function 'virRandomOnceInit':
../../src/util/virrandom.c:65:5: error: implicit declaration of function 'time' [-Werror=implicit-function-declaration]
unsigned int seed = time(NULL) ^ getpid();
^
../../src/util/virrandom.c:65:5: error: nested extern declaration of 'time' [-Werror=nested-externs]
Signed-off-by: Eric Blake <eblake@redhat.com>
Changing the prototype to not have "int *index" since we'll soon be
disallowing index as a name. Curiously the original commit (a4504ac)
for the function used 'int idx' in the function - so they didn't match.
Now they do.
It is there even with -nodefaults and -no-user-config, so count with
that so we can start sparc domains.
Signed-off-by: Martin Kletzander <mkletzan@redhat.com>
The variable 'last_processed_hostdev_vf' indicates index of the last
successfully configed vf. When resetvfnetconfig because of failure,
hostdevs[last_processed_hostdev_vf] should also be reset.
Signed-off-by: Huanle Han <hanxueluo@gmail.com>
1. 'last_good_net' indicates the index of last successfully configured
net. so def->nets[last_good_net] should also be clean up if error occurs.
2. if error occurs in 'virNetDevMacVLanVPortProfileRegisterCallback'
(second 'goto err_exit' in loop), we should also do
'virNetDevVPortProfileDisassociate' cleanup for the
'virNetDevVPortProfileAssociate'(first code block in loop). So we should
consider the net is successfully configured after first code block in
loop finishes.
Signed-off-by: Huanle Han <hanxueluo@gmail.com>
After set memory parameters for running domain, save the change to live
xml is needed otherwise it will disappear after restart libvirtd.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1211548
Signed-off-by: Shanzhi Yu <shyu@redhat.com>
Signed-off-by: Ján Tomko <jtomko@redhat.com>
Apparently for Xen-devel 'index' is a global and causes a build failure,
so just use the shortened 'idx' instead to avoid the conflict.
Signed-off-by: John Ferlan <jferlan@redhat.com>
QEMU does not abandon the mirror. The job carries on in the synchronised
phase and it might be either pivoted again or cancelled. The commit
hints that the described behavior was happening in a downstream version.
If the command returns false there are two possible options:
1) qemu did not reach the point where it would ask the block job to
pivot
2) pivotting failed in the actual qemu coroutine
If either of those would happen we return failure and reset the
condition that waits for the block job to complete. This makes the API
fail but in case where qemu would actually abandon the mirror the fact
is notified via the event and handled asynchronously.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1202704
qemuDomainBlockJobImpl become an unmaintainable mess over the years of
adding new stuff to it. This patch starts splitting up individual
functions from it until it can be killed entirely.
In bulk this will add lines of code rather than delete them but it will
be traded for maintainability.
My intention is to split qemuMonitorJSONBlockJob() into simpler separate
functions for every block job type. Since the error handling code is the
same for all block jobs, this patch extracts the code into a separate
function that will later be reused in more places.
With the new helper qemuMonitorJSONErrorIsClass we can save a few
function calls as we can extract the error object once.
Split out the function that checks the actual error class string into a
separate helper as it will be useful later and refactor
qemuMonitorJSONHasError to return bool type and remove few useless
checks.
Basically virJSONValueObjectHasKey are useless here since the next call
to virJSONValueObjectGet is checking the return value again (which can't
fail at that point). By removing the first check we save a function
call.
Previously we checked that the vcpu we are trying to set is in range of
the number of threads presented by qemu. The problem is that if the VM
is offline the count is 0. Since the condition subtracted 1 from the
count the number would overflow and the check would never trigger.
Change the condition for more sensible ones with specific error
messages.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1208434
Refactor the code to parse the vcpupin in a similar way the iothreadpin
code is now structured. This allows to get rid of some very strange
conditions and error messages.
Additionally since a existing bug
( https://bugzilla.redhat.com/show_bug.cgi?id=1208434 ) allows to add
vcpupin definitions for vcpus that don't exist, this patch makes the
parser to ignore all vcpupins that don't have a matching vCPU in the
definition rather than just offlined ones.
Defining a domain with the following config:
<domain ...>
...
<iothreads>1</iothreads>
<cputune>
<iothreadpin cpuset='1'/>
will result in the following config formatted back:
<domain type='kvm'>
...
<iothreads>1</iothreads>
<cputune>
<iothreadpin iothread='0' cpuset='1'/>
After restart the VM would vanish. Since our schema requires the
@iothread field to be present in <iothreadpin> make it required by the
code too.
When creating a snapshot with _REUSE_EXTERNAL when the pre-created image
does not directly link to the current active layer libvirt would
re-detect the backing chain incorrectly and it would not match with
qemu's view. Since the configuration is an operator mistake, document
that only the top layer image gets inserted.
Parallels Cloud Server supports block devices and virtual NIC
live attachment. I implemented that function for block devices so
OpenStack volume attachment is now works.
Signed-off-by: Alexander Burluka <aburluka@parallels.com>
OpenStack needs disk serial number setup because
nova boot --block-device-mapping command generates that param in
libvirt xml. I took QEMU libvirt driver behavior as a base.
QEMU driver skips inability to set serial and continues work.
So Parallels driver will ignore this param too and let domain
boot.
This patch adds checks for empty bitmaps right after the calls of
virBitmapParse. These only include spots where set API's are called and
where domain's XML is parsed.
Also, it partially reverts commit 983f5a which added a check for
invalid nodeset "0,^0" into virBitmapParse function. This change broke
the logic, as an empty bitmap should not cause an error.
https://bugzilla.redhat.com/show_bug.cgi?id=1210545
On arm, we probe for virtio-*-pci devices, but use their
virtio-*-device variants.
Set the capabilities based on the -device variants as well,
to make them work with qemus with the PCI devices compiled out.
When pre-creating storage for domains, we need to find corresponding
disk in the XML on the destination (domain XML may differ there, e.g.
disk is accessible under different path). For better debugging, I'm
printing all info I received on a disk. But there was a typo when
printing the disk capacity: "%lluu" instead of "%llu".
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
The problem with the previous implementation is,
even when qemuMigrationUpdateJobStatus() detects a migration job
has completed, it will do a sleep for 50 ms (which is unnecessary
and only adds up to the VM pause time).
Signed-off-by: Xing Lin <xinglin@cs.utah.edu>
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Future IOThread setting patches would copy the code anyway, so create
and generalize the adding of pindef for the vcpu and the pinning of the
thread into their own APIs.
Add static virNetDevGetifaddrsAddress to attempt to get the interface
IP address. If getifaddrs is not supported, fall back to
virNetDevGetIPv4AddressIoctl to get the IP address.
This allows IPv6 addresses to be used for <listen type='network>
with device-backed networks.
https://bugzilla.redhat.com/show_bug.cgi?id=1192318
Signed-off-by: Luyao Huang <lhuang@redhat.com>
Signed-off-by: John Ferlan <jferlan@redhat.com>
Signed-off-by: Ján Tomko <jtomko@redhat.com>
Rename it to virNetDevGetIPv4AddressIoctl and make
virNetDevGetIPAddress a wrapper around it, allowing
other ways of getting the address to be implemented,
and still falling back to the old method.
Signed-off-by: John Ferlan <jferlan@redhat.com>
Signed-off-by: Ján Tomko <jtomko@redhat.com>
PCS doesn't store domain config in managed save state file.
It's forbidden to change config for VMs in this state.
It's possible to change config for containers, but after
restoring domain will have that new config, not a config,
which domain had at the moment of virDomainManagedSave.
So we need to handle this case differently from other states.
Let's forbid this operation, if config is changed and if it's
not changed - just do nothing.
Openstack/nova calls virDomainDefineXML on resume with
current domain config, so we can't forbid this operation
in managed save state.
Signed-off-by: Dmitry Guryanov <dguryanov@parallels.com>
We should add input devices with proper bus,
not VIR_DOMAIN_INPUT_BUS_XEN.
Signed-off-by: Dmitry Guryanov <dguryanov@parallels.com>
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Handle input devices in virDomainDefParseXML properly
in case of parallels containers and VMs.
Parallels containers support only
VIR_DOMAIN_INPUT_BUS_PARALLELS. And if VNC is enabled
we should add implicit mouse and keyboard.
For VMs we should add implicit PS/2 mouse and
keyboard.
BTW, is it worth to refactor code and move
all this code to drivers, to *DomainDefPostParse
functions?
Signed-off-by: Dmitry Guryanov <dguryanov@parallels.com>
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Add VIR_DOMAIN_INPUT_BUS_PARALLELS device type
to handle domain configuration properly for
parallels containers, when VNC is enabled.
When domain configuration has at least one
'graphics', there should be mouse and keyboard.
Signed-off-by: Dmitry Guryanov <dguryanov@parallels.com>
Fix function virDomainVideoDefaultType for
parallels VMs and containers. It should return
VGA for VMs and VIR_DOMAIN_VIDEO_TYPE_PARALLELS
for containers.
Signed-off-by: Dmitry Guryanov <dguryanov@parallels.com>
We support VNC for containers to have the same
interface with VMs. At this moment it just renders
linux text console.
Of course we don't pass any physical devices and
don't emulate virtual devices. Our VNC server
renders text from terminal master and sends
input events from VNC client to terminal.
So add special video type VIR_DOMAIN_VIDEO_TYPE_PARALLELS
for these pseudo-devices.
Signed-off-by: Dmitry Guryanov <dguryanov@parallels.com>
Network adapter model has no sense for container,
so we shouldn't set it to e1000 in
parallelsDomainDeviceDefPostParse.
Signed-off-by: Dmitry Guryanov <dguryanov@parallels.com>
We handle this parameter for VMs while defining
domains, so let's get this property from PCS and
set corresponding field of virDomainNetDef in
prlsdkLoadDomains function.
Signed-off-by: Dmitry Guryanov <dguryanov@parallels.com>
Call virDomainDefAddImplicitControllers to add disk
controllers, so virDomainDef, filled by this function
will look exactly like the one returned by virDomainDefParseString.
Signed-off-by: Dmitry Guryanov <dguryanov@parallels.com>
Implement virDomainManagedSave api function. In PCS
this feature called "suspend". You can suspend VM or
CT while it is in running or paused state. And after
resuming (or starting) it will have the same state, as
before suspend.
Signed-off-by: Dmitry Guryanov <dguryanov@parallels.com>
Split function prlsdkDomainChangeState into
prlsdkDomainChangeStateLocked and prlsdkDomainChangeState.
So it can be used from places, where virDomainObj already
found and locked.
Signed-off-by: Dmitry Guryanov <dguryanov@parallels.com>
Return value of functions prlsdkStart/Kill/Stop e.t.c.
is PRL_RESULT in parallels_sdk.c and int in parallels_sdk.h.
PRL_RESULT is int, so compiler didn't report errors.
Let's fix the difference.
Signed-off-by: Dmitry Guryanov <dguryanov@parallels.com>
Future IOThread setting patches would copy the code anyway, so create
and generalize a delete cgroup and pindef for the vcpu into its own API.
Signed-off-by: John Ferlan <jferlan@redhat.com>
Future IOThread setting patches would copy the code anyway, so create
and generalize the add the vcpu to a cgroup into its own API.
Signed-off-by: John Ferlan <jferlan@redhat.com>
Create a new common API to replace the virCgroupNew{Vcpu|Emulator|IOThread}
API's using an emum to generate the cgroup name
Signed-off-by: John Ferlan <jferlan@redhat.com>
https://bugzilla.redhat.com/show_bug.cgi?id=1206521
If the backend driver updates the pool available and/or allocation values,
then the storage_driver VolCreateXML, VolCreateXMLFrom, and VolDelete APIs
should not change the value; otherwise, it will appear as if the values
were "doubled" for each change. Additionally since unsigned arithmetic will
be used depending on the size and operation, either or both values could be
appear to be much larger than they should be (in the EiB range).
Currently only the disk pool updates the values, but other pools could.
Assume a "fresh" disk pool of 500 MiB using /dev/sde:
$ virsh pool-info disk-pool
...
Capacity: 509.88 MiB
Allocation: 0.00 B
Available: 509.84 MiB
$ virsh vol-create-as disk-pool sde1 --capacity 300M
$ virsh pool-info disk-pool
...
Capacity: 509.88 MiB
Allocation: 600.47 MiB
Available: 16.00 EiB
Following assumes disk backend updated to refresh the disk pool at deletion
of primary partition as well as extended partition:
$ virsh vol-delete --pool disk-pool sde1
Vol sde1 deleted
$ virsh pool-info disk-pool
...
Capacity: 509.88 MiB
Allocation: 9.73 EiB
Available: 6.27 EiB
This patch will check if the backend updated the pool values and honor that
update.
Commit id '471e1c4e' only considered updating the pool if the extended
partition was removed. As it turns out removing a primary partition
would also need to update the freeExtent list otherwise the following
sequence would fail (assuming a "fresh" disk pool for /dev/sde of 500M):
$ virsh pool-info disk-pool
...
Capacity: 509.88 MiB
Allocation: 0.00 B
Available: 509.84 MiB
$ virsh vol-create-as disk-pool sde1 --capacity 300M
$ virsh vol-delete --pool disk-pool sde1
$ virsh vol-create-as disk-pool sde1 --capacity 300M
error: Failed to create vol sde1
error: internal error: no large enough free extent
$
This patch will refresh the pool, rereading the partitions, and
return
https://bugzilla.redhat.com/show_bug.cgi?id=1073305
When creating a volume in a pool, the creation allows the 'capacity'
value to be larger than the available space in the pool. As long as
the 'allocation' value will fit in the space, the volume will be created.
However, resizing the volume checks were made with the new absolute
capacity value against existing capacity + the available space without
regard for whether the new absolute capacity was actually allocating
space or not. For example, a pool with 75G of available space creates
a volume of 10G using a capacity of 100G and allocation of 10G will succeed;
however, if the allocation used a capacity of 10G instead and then tried
to resize the allocation to 100G the code would fail to allow the backend
to try the resize.
Furthermore, when updating the pool "available" and "allocation" values,
the resize code would just "blindly" adjust them regardless of whether
space was "allocated" or just "capacity" was being adjusted. This left
a scenario whereby a resize to 100G would fail; however, a resize to 50G
followed by one to 100G would both succeed. Again, neither was adjusting
the allocation value, just the "capacity" value.
This patch adds more logic to the resize code to understand whether the
new capacity value is actually "allocating" space as well and whether it
shrinking or expanding. Since unsigned arithmatic is involved, the possibility
that we adjust the pool size values incorrectly is probable.
This patch also ensures that updates to the pool values only occur if we
actually performed the allocation.
NB: The storageVolDelete, storageVolCreateXML, and storageVolCreateXMLFrom
each only updates the pool allocation/availability values by the target
volume allocation value.
Support for drive-reopen was never present in the upstream code so we
don't need to pause the VM when doing the block pivot. Kill all the
code related to this semi-upstream artifact.
lxc-enter-namespace stopped working on recent kernels (at least 3.19+)
due to /proc/PID/ns/* file descriptors being opened RW. From outside
the namespace these can only be opened RO.
/var/run may reside on a tmpfs and we fail to create the PID file if
/var/run/lxc does not exist.
Since commit 0a8addc1, the lxc driver's state directory isn't
automatically created before starting a domain. Now, the lxc driver
makes sure the state directory exists when it initializes.
Signed-off-by: Lubomir Rintel <lkundrak@v3.sk>
rfc3986 states that the separator in URI path is a single slash.
Multiple slashes may potentially lead to different resources and thus we
should not remove them.
Like we are doing in qemu driver (ea576ee543), lets call
virNumaSetupMemoryPolicy() only if really needed. Problem is, if
we numa_set_membind() child, there's no way to change it from the
daemon afterwards. So any later attempts to change the pinning
will fail. But in very weird way - CGroups will be set, but due
to membind child will not allocate memory from any other node.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
131,088 bytes in 16 blocks are definitely lost in loss record 2,174 of 2,176
at 0x4C29BFD: malloc (in /usr/lib64/valgrind/vgpreload_memcheck-amd64-linux.so)
by 0x4C2BACB: realloc (in /usr/lib64/valgrind/vgpreload_memcheck-amd64-linux.so)
by 0x52A026F: virReallocN (viralloc.c:245)
by 0x52BFCB5: saferead_lim (virfile.c:1268)
by 0x52C00EF: virFileReadLimFD (virfile.c:1328)
by 0x52C019A: virFileReadAll (virfile.c:1351)
by 0x52A5D4F: virCgroupGetValueStr (vircgroup.c:763)
by 0x1DDA0DA3: qemuRestoreCgroupState (qemu_cgroup.c:805)
by 0x1DDA0DA3: qemuConnectCgroup (qemu_cgroup.c:857)
by 0x1DDB7BA1: qemuProcessReconnect (qemu_process.c:3694)
by 0x52FD171: virThreadHelper (virthread.c:206)
by 0x82B8DF4: start_thread (pthread_create.c:308)
by 0x85C31AC: clone (clone.S:113)
Signed-off-by: Luyao Huang <lhuang@redhat.com>
