If a <parent> is not supplied in the XML used to create a non-persistent
vHBA, then instead of failing, let's try to find a "vports" capable node
device and use that.
Signed-off-by: John Ferlan <jferlan@redhat.com>
Extract out code from virNodeDeviceGetParentHost into helpers - it's
going to be reused in upcoming patches to search on more fields
Create virNodeDeviceFindVPORTCapDef in order to return a virNodeDevCapsDefPtr
of the VPORT_OPS and virNodeDeviceFindFCParentHost to use the function and
generate an error message if the device doesn't have the capability.
Also clean up the processing in virNodeDeviceGetParentHost to remove
need for goto's.
Signed-off-by: John Ferlan <jferlan@redhat.com>
Commit 4b951d1e38 missed the fact that the
VM needs to be resumed after a live external checkpoint (memory
snapshot) where the cpus would be paused by the migration rather than
libvirt.
Again, not something that I'd hit, but there is a chance in
theory that this might bite us. Currently the way we decide
whether or not to create /dev entry for a device is by marching
first four characters of path with "/dev". This might be not
enough. Just imagine somebody has a disk image stored under
"/devil/path/to/disk". We ought to be matching against "/dev/".
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Not that I'd encounter any bug here, but the code doesn't look
100% correct. Imagine, somebody is trying to attach a device to a
domain, and the device's /dev entry already exists in the qemu
namespace. This is handled gracefully and the control continues
with setting up ACLs and calling security manager to set up
labels. Now, if any of these steps fail, control jump on the
'cleanup' label and unlink() the file straight away. Even when it
was not us who created the file in the first place. This can be
possibly dangerous.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
https://bugzilla.redhat.com/show_bug.cgi?id=1406837
Imagine you have a domain configured in such way that you are
assigning two PCI devices that fall into the same IOMMU group.
With mount namespace enabled what happens is that for the first
PCI device corresponding /dev/vfio/X entry is created and when
the code tries to do the same for the second mknod() fails as
/dev/vfio/X already exists:
2016-12-21 14:40:45.648+0000: 24681: error :
qemuProcessReportLogError:1792 : internal error: Process exited
prior to exec: libvirt: QEMU Driver error : Failed to make device
/var/run/libvirt/qemu/windoze.dev//vfio/22: File exists
Worse, by default there are some devices that are created in the
namespace regardless of domain configuration (e.g. /dev/null,
/dev/urandom, etc.). If one of them is set as backend for some
guest device (e.g. rng, chardev, etc.) it's the same story as
described above.
Weirdly, in attach code this is already handled.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Clang 3.9 refuses to compile the existing code with the
following error:
util/virfirewall.c:425:20: error: passing an object that undergoes
default argument promotion to 'va_start'
has undefined behavior [-Werror,-Wvarargs]
va_start(args, layer);
^
util/virfirewall.c:420:37: note: parameter of type 'virFirewallLayer'
is declared here
virFirewallLayer layer,
^
This happens because 'layer' is of type virFirewallLayer, which
is an enum type and not a standard type such as eg. void* or int.
To solve the issue, turn virFirewallAddRule() from a very thin
wrapper around virFirewallAddRuleFullV() to a macro that expands
to a call to virFirewallAddRuleFull() - itself a very thin wrapper
around the aforementioned virFirewallAddRuleFullV() - with no loss
of functionality or type safety.
https://bugzilla.redhat.com/show_bug.cgi?id=1405269
If a secret was not provided for what was determined to be a LUKS
encrypted disk (during virStorageFileGetMetadata processing when
called from qemuDomainDetermineDiskChain as a result of hotplug
attach qemuDomainAttachDeviceDiskLive), then do not attempt to
look it up (avoiding a libvirtd crash) and do not alter the format
to "luks" when adding the disk; otherwise, the device_add would
fail with a message such as:
"unable to execute QEMU command 'device_add': Property 'scsi-hd.drive'
can't find value 'drive-scsi0-0-0-0'"
because of assumptions that when the format=luks that libvirt would have
provided the secret to decrypt the volume.
Access to unlock the volume will thus be left to the application.
After 478ddedc12 a bug is fixed where we wrongly presumed loopack
device name on non-Linux systems. It's lo0. However, the fix is
not reflected in the tests which are failing now.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Due to nature of operations we do over the string list (more
precisely due to how virStringListRemove() works), it is not the
best idea to use dataFree callback. Problem is, on MAC address
remove, the string list remove function modifies the original
list in place. Then, virHashUpdateEntry() is called which frees
all the data stored in the list rendering @newMacsList point to
freed data.
==16002== Invalid read of size 8
==16002== at 0x50BC083: virFree (viralloc.c:582)
==16002== by 0x513DC39: virStringListFree (virstring.c:251)
==16002== by 0x51089B4: virMacMapHashFree (virmacmap.c:67)
==16002== by 0x50EF30B: virHashAddOrUpdateEntry (virhash.c:352)
==16002== by 0x50EF4FD: virHashUpdateEntry (virhash.c:415)
==16002== by 0x5108BED: virMacMapRemoveLocked (virmacmap.c:129)
==16002== by 0x51092D5: virMacMapRemove (virmacmap.c:346)
==16002== by 0x402F02: testMACRemove (virmacmaptest.c:107)
==16002== by 0x403F15: virTestRun (testutils.c:180)
==16002== by 0x4032C4: mymain (virmacmaptest.c:205)
==16002== by 0x405A3B: virTestMain (testutils.c:992)
==16002== by 0x403D87: main (virmacmaptest.c:237)
==16002== Address 0xdd5a4d0 is 0 bytes inside a block of size 24 free'd
==16002== at 0x4C2AD6F: realloc (vg_replace_malloc.c:693)
==16002== by 0x50BB99B: virReallocN (viralloc.c:245)
==16002== by 0x513DC0B: virStringListRemove (virstring.c:235)
==16002== by 0x5108BA6: virMacMapRemoveLocked (virmacmap.c:124)
==16002== by 0x51092D5: virMacMapRemove (virmacmap.c:346)
==16002== by 0x402F02: testMACRemove (virmacmaptest.c:107)
==16002== by 0x403F15: virTestRun (testutils.c:180)
==16002== by 0x4032C4: mymain (virmacmaptest.c:205)
==16002== by 0x405A3B: virTestMain (testutils.c:992)
==16002== by 0x403D87: main (virmacmaptest.c:237)
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
In virMacMapRemoveLocked() we have two variables: @macsList and
@newMacsList. Obviously, @newMacsList is supposed to hold pointer
to modified list but in fact it holds pointer to the old list.
It's confusing.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Before, boot devices information for CTs was always empty and we
didn't indicate that containers can boot from disk.
Signed-off-by: Maxim Nestratov <mnestratov@virtuozzo.com>
Virtuozzo SDK interface doesn't differ filesystems from disks and sees them as disks.
Before, we always mistakenly presented disks based on files as filesystems, which is
not completely correct. Now we are going to show either disks or filesystems depending
on a hint, which uses boot device section of VZ config. Though this information
doesn't change booting order of a CT, it is used by vz libvirt interface as a hint
for libvirt representation of disks. Since now, if we have filesystems in input xml,
then we add them to VZ booting devices list and rely on this information to show
corresponding libvirt xml.
Signed-off-by: Maxim Nestratov <mnestratov@virtuozzo.com>
Implicit devices like controllers are confusing for CTs and
function virDomainDefAddImplicitDevices never intended to be called
for CTs.
Signed-off-by: Maxim Nestratov <mnestratov@virtuozzo.com>
virQEMUCapsSupportsChardev existing checks returns true
for spapr-vty alone. Instead verify spapr-vty validity
and let the logic to return true for other device types
so that virtio-console passes.
The non-pseries machines dont have spapr-vio-bus. So, the
function always returned false for them before.
Fixes - https://bugzilla.redhat.com/show_bug.cgi?id=1257813
Signed-off-by: Shivaprasad G Bhat <sbhat@linux.vnet.ibm.com>
https://bugzilla.redhat.com/show_bug.cgi?id=1332019
This function will essentially be a wrapper to virStorageVolInfo in order
to provide a mechanism to have the "physical" size of the volume returned
instead of the "allocation" size. This will provide similar capabilities to
the virDomainBlockInfo which can return both allocation and physical of a
domain storage volume.
NB: Since we're reusing the _virStorageVolInfo and not creating a new
_virStorageVolInfoFlags structure, we'll need to generate the rpc APIs
remoteStorageVolGetInfoFlags and remoteDispatchStorageVolGetInfoFlags
(although both were originally created from gendispatch.pl and then
just copied into daemon/remote.c and src/remote/remote_driver.c).
The new API will allow the usage of a VIR_STORAGE_VOL_GET_PHYSICAL flag
and will make the decision to return the physical or allocation value
into the allocation field.
In order to get that physical value, virStorageBackendUpdateVolTargetInfoFD
adds logic to fill in physical value matching logic in qemuStorageLimitsRefresh
used by virDomainBlockInfo when the domain is inactive.
Signed-off-by: John Ferlan <jferlan@redhat.com>
Although the virStorageBackendUpdateVolTargetInfo will update the
target.physical value, there is no way to provide that information
via the virStorageGetVolInfo API since it only returns the capacity
and allocation of a volume. So as described in commit id '0282ca45',
it should be possible to generate an output only <physical> value
for that purpose.
This patch generates the <physical> value in the volume XML output
for the sole purpose of being able to view/see the value to allow
someone to parse the XML in order to obtain the value.
Update the documentation to describe the output only nature.
Signed-off-by: John Ferlan <jferlan@redhat.com>
According to commit id '0282ca45a' the 'physical' value should
essentially be the last offset of the image or the host physical
size in bytes of the image container. However, commit id '15fa84ac'
refactored the GetBlockInfo to use the same returned data as the
GetStatsBlock API for an active domain. For the 'entry->physical'
that would end up being the "actual-size" as set through the
qemuMonitorJSONBlockStatsUpdateCapacityOne (commit '7b11f5e5').
Digging deeper into QEMU code one finds that actual_size is
filled in using the same algorithm as GetBlockInfo has used for
setting the 'allocation' field when the domain is inactive.
The difference in values is seen primarily in sparse raw files
and other container type files (such as qcow2), which will return
a smaller value via the stat API for 'st_blocks'. Additionally
for container files, the 'capacity' field (populated via the
QEMU "virtual-size" value) may be slightly different (smaller)
in order to accomodate the overhead for the container. For
sparse files, the state 'st_size' field is returned.
This patch thus alters the allocation and physical values for
sparse backed storage files to be more appropriate to the API
contract. The result for GetBlockInfo is the following:
capacity: logical size in bytes of the image (how much storage
the guest will see)
allocation: host storage in bytes occupied by the image (such
as highest allocated extent if there are no holes,
similar to 'du')
physical: host physical size in bytes of the image container
(last offset, similar to 'ls')
NB: The GetStatsBlock API allows a different contract for the
values:
"block.<num>.allocation" - offset of the highest written sector
as unsigned long long.
"block.<num>.capacity" - logical size in bytes of the block device
backing image as unsigned long long.
"block.<num>.physical" - physical size in bytes of the container
of the backing image as unsigned long long.
This patch detects a misconfiguration between the disk bus type and disk
address type for controller based disk buses (SATA, SCSI, FDC and
IDE). The addresses of these bus types are all managed in common code so
it's possible to decide in common code whether the disk address and bus
type are compatible or not.
Signed-off-by: Marc Hartmayer <mhartmay@linux.vnet.ibm.com>
Reviewed-by: Bjoern Walk <bwalk@linux.vnet.ibm.com>
Disk->info is not live updatable so add a check for this. Otherwise
libvirt reports success even though no data was updated.
Signed-off-by: Marc Hartmayer <mhartmay@linux.vnet.ibm.com>
Reviewed-by: Bjoern Walk <bwalk@linux.vnet.ibm.com>
Reviewed-by: Boris Fiuczynski <fiuczy@linux.vnet.ibm.com>
This function will be needed by the QEMU driver in an upcoming
patch. Additionally, removed a useless empty line.
Signed-off-by: Marc Hartmayer <mhartmay@linux.vnet.ibm.com>
This patch reduces the complexity of the filtering algorithm in
virCgroupDetect by first correcting the controller mask and then
checking for potential co-mounts without any correlating
controller mask modifications.
If you agree that this patch removes complexity and improves
readability it could simply be squashed into the first patch
of this series.
Signed-off-by: Boris Fiuczynski <fiuczy@linux.vnet.ibm.com>
Reviewed-by: Bjoern Walk <bwalk@linux.vnet.ibm.com>
Reviewed-by: Marc Hartmayer <mhartmay@linux.vnet.ibm.com>
The cgroup controller filtering in virCgroupDetect does not work
properly if the following conditions are met:
1) the host system does not have a cgroup controller which
libvirt requests (unavailable controller) and
2) libvirt is configured to disable a controller (disabled controller) and
3) the disabled controller is located before the unavailable controller
in virCgroupController.
As an example: The memory controller is unavailable and the cpuset
controller is configured to be disabled.
In this scenario trying to start a domain results in the error
error: Controller 'cpuset' is not wanted, but 'memory' is co-mounted: Invalid argument
This error occurs when virCgroupDetect is called with a valid parent group.
The resulting group created by virCgroupCopyMounts holds for cpuset and
memory controller empty mount points. The filtering of disabled controllers
checks for co-mounts by comparing the mount points. The cpuset controller
causes the filtering to occur before the memory controller is marked as to be
ignored by modifying the controller mask since it is unavailable.
Therefore the co-mount detection logic compares the cpuset and memory controller
mount points and since both are empty the memory controller is regarded
erroneously as being co-mounted.
Signed-off-by: Boris Fiuczynski <fiuczy@linux.vnet.ibm.com>
Reviewed-by: Marc Hartmayer <mhartmay@linux.vnet.ibm.com>
Reviewed-by: Bjoern Walk <bwalk@linux.vnet.ibm.com>
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Similarly to localOnly DNS domain, localPtr attribute can be used to
tell the DNS server not to forward reverse lookups for unknown IPs which
belong to the virtual network.
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
The API creates PTR domain which corresponds to a given addr/prefix.
Both IPv4 and IPv6 addresses are supported, but the prefix must be
divisible by 8 for IPv4 and divisible by 4 for IPv6.
The generated PTR domain has the following format
IPv4: 1.2.3.4.in-addr.arpa
IPv6: 0.1.2.3.4.5.6.7.8.9.a.b.c.d.e.f.0.1.2.3.4.5.6.7.8.9.a.b.c.d.e.f.ip6.arpa
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
Iterating over all child nodes when we only support one instance of each
child is pretty weird. And it would even cause memory leaks if more
than one <tftp> element was specified.
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
The PERF_COUNT_HW_REF_CPU_CYCLES constant is not available
on all Linux distros libvirt targets, so its use must be
made conditional. Other constant have existed long enough
that we can assume they exist, as we don't support very
old distros like RHEL-5 any more.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Enable libvirt users to modify daemon's logging output settings from outside.
If either an empty string or NULL is passed, a default logging output will be
used the same way as it would be in case writing an empty string to the
libvirtd.conf
Signed-off-by: Erik Skultety <eskultet@redhat.com>
Now that virLog{Get,Set}DefaultOutput routines are introduced we can wire them
up to the daemon's logging initialization code. Also, change the order of
operations a bit so that we still strictly honor our precedence of settings:
cmdline > env > config now that outputs and filters are not appended anymore.
Signed-off-by: Erik Skultety <eskultet@redhat.com>
Along with an empty string, it should also be possible for users to pass
NULL to the public APIs which in turn would trigger a routine(future
work) responsible for defining an appropriate default logging output
given the current circumstances.
Signed-off-by: Erik Skultety <eskultet@redhat.com>
These helpers will manage the log destination defaults (fetch/set). The reason
for this is to stay consistent with the current daemon's behaviour with respect
to /etc/libvirt/<daemon>.conf file, since both assignment of an empty string
or not setting the log output variable at all trigger the daemon's decision on
the default log destination which depends on whether the daemon runs daemonized
or not.
This patch also changes the logic of the selection of the default
logging output compared to how it is done now. The main difference though is
that we should only really care if we're running daemonized or not, disregarding
the fact of (not) having a TTY completely (introduced by commit eba36a3878) as
that should be of the libvirtd's parent concern (what FD it will pass to it).
Before:
if (godaemon || !hasTTY):
if (journald):
use journald
if (godaemon):
if (privileged):
use SYSCONFIG/libvirtd.log
else:
use XDG_CONFIG_HOME/libvirtd.log
else:
use stderr
After:
if (godaemon):
if (journald):
use journald
else:
if (privileged):
use SYSCONFIG/libvirtd.log
else:
use XDG_CONFIG_HOME/libvirtd.log
else:
use stderr
Signed-off-by: Erik Skultety <eskultet@redhat.com>
External disk-only snapshots with recent enough qemu don't require
libvirt to pause the VM. The logic determining when to resume cpus was
slightly flawed and attempted to resume them even if they were not
paused by the snapshot code. This normally was not a problem, but with
locking enabled the code would attempt to acquire the lock twice.
The fallout of this bug would be a error from the API, but the actual
snapshot being created. The bug was introduced with when adding support
for external snapshots with memory (checkpoints) in commit f569b87.
Resolves problems described by:
https://bugzilla.redhat.com/show_bug.cgi?id=1403691
After qemu delivers the resume event it's already running and thus it's
too late to enter lockspaces since it may already have modified the
disk. The code only creates false log entries in the case when locking
is enabled. The lockspace needs to be acquired prior to starting cpus.
Given how intrusive previous patches are, it might happen that
there's a bug or imperfection. Lets give users a way out: if they
set 'namespaces' to an empty array in qemu.conf the feature is
suppressed.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
When attaching a device to a domain that's using separate mount
namespace we must maintain /dev entries in order for qemu process
to see them.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
When attaching a device to a domain that's using separate mount
namespace we must maintain /dev entries in order for qemu process
to see them.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
When attaching a device to a domain that's using separate mount
namespace we must maintain /dev entries in order for qemu process
to see them.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
When attaching a device to a domain that's using separate mount
namespace we must maintain /dev entries in order for qemu process
to see them.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Instead of trying to fix our security drivers, we can use a
simple trick to relabel paths in both namespace and the host.
I mean, if we enter the namespace some paths are still shared
with the host so any change done to them is visible from the host
too.
Therefore, we can just enter the namespace and call
SetAllLabel()/RestoreAllLabel() from there. Yes, it has slight
overhead because we have to fork in order to enter the namespace.
But on the other hand, no complexity is added to our code.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
When starting a domain and separate mount namespace is used, we
have to create all the /dev entries that are configured for the
domain.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
When starting a domain and separate mount namespace is used, we
have to create all the /dev entries that are configured for the
domain.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
When starting a domain and separate mount namespace is used, we
have to create all the /dev entries that are configured for the
domain.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
When starting a domain and separate mount namespace is used, we
have to create all the /dev entries that are configured for the
domain.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
When starting a domain and separate mount namespace is used, we
have to create all the /dev entries that are configured for the
domain.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
When starting a domain and separate mount namespace is used, we
have to create all the /dev entries that are configured for the
domain.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Prime time. When it comes to spawning qemu process and
relabelling all the devices it's going to touch, there's inherent
race with other applications in the system (e.g. udev). Instead
of trying convincing udev to not touch libvirt managed devices,
we can create a separate mount namespace for the qemu, and mount
our own /dev there. Of course this puts more work onto us as we
have to maintain /dev files on each domain start and device
hot(un-)plug. On the other hand, this enhances security also.
From technical POV, on domain startup process the parent
(libvirtd) creates:
/var/lib/libvirt/qemu/$domain.dev
/var/lib/libvirt/qemu/$domain.devpts
The child (which is going to be qemu eventually) calls unshare()
to create new mount namespace. From now on anything that child
does is invisible to the parent. Child then mounts tmpfs on
$domain.dev (so that it still sees original /dev from the host)
and creates some devices (as explained in one of the previous
patches). The devices have to be created exactly as they are in
the host (including perms, seclabels, ACLs, ...). After that it
moves $domain.dev mount to /dev.
What's the $domain.devpts mount there for then you ask? QEMU can
create PTYs for some chardevs. And historically we exposed the
host ends in our domain XML allowing users to connect to them.
Therefore we must preserve devpts mount to be shared with the
host's one.
To make this patch as small as possible, creating of devices
configured for domain in question is implemented in next patches.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
This is a list of devices that qemu needs for its run (apart from
what's configured for domain). The devices on the list are
enabled in the CGroups by default so they will be good candidates
for initial /dev for new qemu.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
We will need this function in near future so that we know what
/dev device corresponds to the SCSI device.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
We will need this function in near future so that we know what
/dev device corresponds to the SCSI device.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
We will need this function in near future so that we know what
/dev device corresponds to the USB device.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Namely, virFileGetACLs, virFileSetACLs, virFileFreeACLs and
virFileCopyACLs. These functions are going to be required when we
are creating /dev for qemu. We have copy anything that's in
host's /dev exactly as is. Including ACLs.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
libvirt libxl picks its own default with respect to the default NIC
to use. libxlMakeNic is the one responsible for this and on boot it
picks LIBXL_NIC_TYPE_VIF_IOEMU for HVM domains such that it accomodates
both PV and emulated one. The good behaving guest at boot will then
select the pv and unplug the emulated device.
Now, on HVM when attaching an interface it will pick the same default
that is LIBXL_NIC_TYPE_VIF_IOEMU which as a result will fail the attach
(see xen commit 32e9d0f ("libxl: nic type defaults to vif in hotplug for
hvm guest"). Xen doesn't yet support the hotplug of emulated devices,
but we don't want to rule out that case either, which might get support
in the future. Hence we simply reverse the defaults when we are
attaching the interface which allows libvirt to prefer the PV nic first
without adding "model='netfront'" following the same pattern as above
commit. Also to avoid ruling out the emulated one we set to
LIBXL_NIC_TYPE_IOEMU when setting a model type that is not 'netfront'.
Signed-off-by: Joao Martins <joao.m.martins@oracle.com>
Signed-off-by: Jim Fehlig <jfehlig@suse.com>
The virDomainSendProcessSignal method says the flags values
come from virDomainProcessSignalFlag, but this enum has
never existed. No flags are needed for this method.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Almost none of our virJSONValue*Get* functions accept const virJSONValue
pointers and it wouldn't even make sense since we sometimes modify what
we get. And because there is no reason for preventing callers of
virJSONValueObjectForeachKeyValue from modifying the values they get in
each iteration we can just stop doing it.
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
Using a variable named 'stat' clashes with the system function
'stat()' causing compiler warnings on some platforms
cc1: warnings being treated as errors
../../src/qemu/qemu_monitor_text.c: In function 'parseMemoryStat':
../../src/qemu/qemu_monitor_text.c:604: error: declaration of 'stat' shadows a global declaration [-Wshadow]
/usr/include/sys/stat.h:455: error: shadowed declaration is here [-Wshadow]
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
If the cpuset cgroup controller is disabled in /etc/libvirt/qemu.conf
QEMU virtual machines can in principle use all host CPUs, even if they
are hot plugged, if they have no explicit CPU affinity defined.
However, there's libvirt code supposed to handle the situation where
the libvirt daemon itself is not using all host CPUs. The code in
qemuProcessInitCpuAffinity attempts to set an affinity mask including
all defined host CPUs. Unfortunately, the resulting affinity mask for
the process will not contain the offline CPUs. See also the
sched_setaffinity(2) man page.
That means that even if the host CPUs come online again, they won't be
used by the QEMU process anymore. The same is true for newly hot
plugged CPUs. So we are effectively preventing that QEMU uses all
processors instead of enabling it to use them.
It only makes sense to set the QEMU process affinity if we're able
to actually grow the set of usable CPUs, i.e. if the process affinity
is a subset of the online host CPUs.
There's still the chance that for some reason the deliberately chosen
libvirtd affinity matches the online host CPU mask by accident. In this
case the behavior remains as it was before (CPUs offline while setting
the affinity will not be used if they show up later on).
Signed-off-by: Viktor Mihajlovski <mihajlov@linux.vnet.ibm.com>
Tested-by: Matthew Rosato <mjrosato@linux.vnet.ibm.com>
The functions to retrieve online and present host CPU information
are only supported on Linux for the time being.
This leads to runtime errors if these function are used on other
platforms. To avoid that, code in higher levels using the functions
must replicate the conditional compilation in higher level which
is error prone (and is plainly spoken ugly).
Adding a function virHostCPUHasBitmap that can be used to check
for host CPU bitmap support.
NB: There are other functions including the host CPU count that
are lacking support on all platforms, but they are too essential
in order to be bypassed.
Signed-off-by: Viktor Mihajlovski <mihajlov@linux.vnet.ibm.com>
virQEMUCapsFindTarget is supposed to find an alternative QEMU binary if
qemu-system-$GUEST_ARCH doesn't exist. The alternative is using host
architecture when it is compatible with $GUEST_ARCH. But a special
treatment has to be applied for ppc64le since the QEMU binary is always
called qemu-system-ppc64.
Broken by me in v2.2.0-171-gf2e71550d.
https://bugzilla.redhat.com/show_bug.cgi?id=1403745
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
qemuAgentNotifyEvent accesses monitor structure and is called on qemu
reset/shutdown/suspend events under domain lock. Other monitor
functions on the other hand take monitor lock and don't hold domain lock.
Thus it is possible to have risky simultaneous access to the structure
from 2 threads. Let's take monitor lock here to make access exclusive.
In case of 0 filesystems *info is not set while according
to virDomainGetFSInfo contract user should call free on it even
in case of 0 filesystems. Thus we need to properly set
it. NULL will be enough as free eats NULLs ok.
The libvirt-domain.h documentation indicates that for a qcow2 file
in a filesystem being used for a backing store should report the disk
space occupied by a file; however, commit id '15fa84ac' altered the
code to trust that the wr_highest_offset should be used whenever
wr_highest_offset_valid was set.
As it turns out this will lead to indeterminite results. For an active
domain when qemu hasn't yet had the need to find the wr_highest_offset
value, qemu will report 0 even though qemu-img will report the proper
disk size. This causes reporting of the following XML:
<disk type='file' device='disk'>
<driver name='qemu' type='qcow2'/>
<source file='/path/to/test-1g.qcow2'/>
to be as follows:
Capacity: 1073741824
Allocation: 0
Physical: 1074139136
with qemu-img indicating:
image: /path/to/test-1g.qcow2
file format: qcow2
virtual size: 1.0G (1073741824 bytes)
disk size: 1.0G
Once the backing source file is opened on the guest, then wr_highest_offset
is updated, but only to the high water mark and not the size of the file.
This patch will adjust the logic to check for the file backed qcow2 image
and enforce setting the allocation to the returned 'physical' value, which
is the 'actual-size' value from a 'query-block' operation.
NB: The other consumer of the wr_highest_offset output (GetAllDomainStats)
has a contract that indicates 'allocation' is the offset of the highest
written sector, so it doesn't need adjustment.
Signed-off-by: John Ferlan <jferlan@redhat.com>
Instead of having duplicated code in qemuStorageLimitsRefresh and
virStorageBackendUpdateVolTargetInfo to get capacity specific data
about the storage backing source or volume -- create a common API
to handle the details for both.
As a side effect, virStorageFileProbeFormatFromBuf returns to being
a local/static helper to virstoragefile.c
For the QEMU code - if the probe is done, then the format is saved so
as to avoid future such probes.
For the storage backend code, there is no need to deal with the probe
since we cannot call the new API if target->format == NONE.
Signed-off-by: John Ferlan <jferlan@redhat.com>
Instead of having duplicated code in qemuStorageLimitsRefresh and
virStorageBackendUpdateVolTargetInfoFD to fill in the storage backing
source or volume allocation, capacity, and physical values - create a
common API that will handle the details for both.
The common API will fill in "default" capacity values as well - although
those more than likely will be overridden by subsequent code. Having just
one place to make the determination of what the values should be will
make things be more consistent.
For the QEMU code - the data filled in will be for inactive domains
for the GetBlockInfo and DomainGetStatsOneBlock API's. For the storage
backend code - the data will be filled in during the volume updates.
Signed-off-by: John Ferlan <jferlan@redhat.com>
Commit id '8dc27259' introduced virStorageSourceUpdateBlockPhysicalSize
in order to retrieve the physical size for a block backed source device
for an active domain since commit id '15fa84ac' changed to use the
qemuMonitorGetAllBlockStatsInfo and qemuMonitorBlockStatsUpdateCapacity
API's to (essentially) retrieve the "actual-size" from a 'query-block'
operation for the source device.
However, the code only was made functional for a BLOCK backing type
and it neglected to use qemuOpenFile, instead using just open. After
the open the block lseek would find the end of the block and set the
physical value, close the fd and return.
Since the code would return 0 immediately if the source device wasn't
a BLOCK backed device, the physical would be displayed incorrectly,
such as follows in domblkinfo for a file backed source device:
Capacity: 1073741824
Allocation: 0
Physical: 0
This patch will modify the algorithm to get the physical size for other
backing types and it will make use of the qemuDomainStorageOpenStat
helper in order to open/stat the source file depending on its type.
The qemuDomainGetStatsOneBlock will no longer inhibit printing errors,
but it will still ignore them leaving the physical value set to 0.
Signed-off-by: John Ferlan <jferlan@redhat.com>
Split out the opening of the file and fetch of the stat buffer into a
helper qemuDomainStorageOpenStat. This will handle either opening the
local or remote storage.
Additionally split out the cleanup of that into a separate helper
qemuDomainStorageCloseStat which will either close the file or
call the virStorageFileDeinit function.
Signed-off-by: John Ferlan <jferlan@redhat.com>
Originally added by commit id '89646e69' prior to commit id '15fa84ac'
and '71d2c172' which ensured that qemuStorageLimitsRefresh was only called
for inactive domains.
Adjust the comment describing the need for FIXME and move all the text
to the function description.
Signed-off-by: John Ferlan <jferlan@redhat.com>
This flag is used in Virtuozzo backend implicitly, thus
we need to support it and don't fail if it's set.
Signed-off-by: Pavel Glushchak <pglushchak@virtuozzo.com>
This flag tells backend not to create instance
disks making behavior the same as in qemu driver.
Disk files have to be created beforehand on target
host manually or by upper management layer i.e.
OpenStack Nova.
Signed-off-by: Pavel Glushchak <pglushchak@virtuozzo.com>
When save/migrate a domain and we autogenerated a port, then if we
print the inactive domain config, write out a -1 for the socket value;
otherwise, it's possible that the subsequent start will fail if the
autogenerated websocket used conflicts with an existing running config
that also used autogenerated websockets.
Examples:
== A. Can not restore domain with autoconfigured websocket.
domain 1 and 2 have autoconfigured websocket.
1. domain 1 is started then, saved
2. domain 2 is started
3. domain 1 restoration is failed:
error: internal error: qemu unexpectedly closed the monitor: 2016-11-21T10:23:11.356687Z
qemu-kvm: -vnc 0.0.0.0:2,websocket=5700: Failed to start VNC server on `(null)':
Failed to bind socket: Address already in use
== B. Can not migrate domain with autoconfigured websocket.
domain 1 on host A, domain 2 on host B, both have autoconfigured websocket
1. domain 1 started, domain 2 started
2. domain 1 migration to host B is failed with the above error.
In preparation to the code move to virnetdevtap.c, this change:
* renames virNetInterfaceStats to virNetDevTapInterfaceStats
* changes 'path' to 'ifname', to use the same vocable as other
method in virnetdevtap.c.
* Add the attributes checker
When vhostuser interfaces are used, the interface statistics
are not available in /proc/net/dev.
This change looks at the openvswitch interfaces statistics
tables to provide this information for vhostuser interface.
Note that in openvswitch world drop/error doesn't always make sense
for some interface type. When these informations are not available we
set them to 0 on the virDomainInterfaceStats.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
There's nothing to compress if the requested snapshot memory format is
set to 'raw' explicitly. After commit 9e14689ea libvirt would try to
run /sbin/raw to process the memory stream if the qemu.conf option
snapshot_image_format is set.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1402726
Since its introduction in 2012 this internal API did nothing.
Moreover we have the same API that does exactly the same:
virSecurityManagerDomainSetPathLabel.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
If you've ever tried running a huge page backed guest under
different user than in qemu.conf, you probably failed. Problem is
even though we have corresponding APIs in the security drivers,
there's no implementation and thus we don't relabel the huge page
path. But even if we did, so far all of the domains share the
same path:
/hugepageMount/libvirt/qemu
Our only option there would be to set 0777 mode on the qemu dir
which is totally unsafe. Therefore, we can create dir on
per-domain basis, i.e.:
/hugepageMount/libvirt/qemu/domainName
and chown domainName dir to the user that domain is configured to
run under.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
So far this function takes virDomainObjPtr which:
1) is an overkill,
2) might be not available in all the places we will use it.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Since the great rework of how we store vcpu- and iothread-related
data, we have overly complex part of code that is trying to format the
scheduler tuning data in as less lines as possible by grouping
settings for multiple threads. That was designed as an input syntax
sugar for users, but we don't need to also use that when formatting
the XML. Switching to simple enumeration makes the code nicer,
shorter and more welcoming to future changes.
Signed-off-by: Martin Kletzander <mkletzan@redhat.com>
When redoing the website we deleted the libvirtLogo.png file
not remembering that the test driver screenshot API impl
relied on it.
Rather than having the test driver use the logo as a side
effect, give it its own dedicated image to use. This is
installed in /usr/share/libvirt/test-screenshot.png and
is taken from a NeXT Cube running WorldWideWeb[1]. The
very first web browser in existance, running on the
hardware it was originally written on.
[1] https://en.wikipedia.org/wiki/WorldWideWeb
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Qemu 2.8.0+ changes arguments structure for blockdev-add in the effort
to make it finally stable. Since libvirt recently added the detection of
gluster debug support relying on the old syntax we need to add the new
as well.
With current perf framework, this patch adds support and documentation
for the branch_instructions perf event.
Signed-off-by: Nitesh Konkar <nitkon12@linux.vnet.ibm.com>
Adjust the spacing a bit in order to generate 'cleaner' looking output.
This matches what virDomainMemoryStats does and it creates text/code boxes
in order to list each of the stats for each category.
With kernel 3.18 (since commit 3e32cb2e0a12b6915056ff04601cf1bb9b44f967)
the "unlimited" value for cgroup memory limits has changed once again as
its byte value is now computed from a page counter.
The new "unlimited" value reported by the cgroup fs is therefore 2**51-1
pages which is (VIR_DOMAIN_MEMORY_PARAM_UNLIMITED - 3072). This results
e.g. in virsh memtune displaying 9007199254740988 instead of unlimited
for the limits.
This patch uses the value of memory.limit_in_bytes from the cgroup
memory root which is the system's "real" unlimited value for comparison.
See also libvirt commit 231656bbeb for the
history for kernel 3.12 and before.
Signed-off-by: Viktor Mihajlovski <mihajlov@linux.vnet.ibm.com>
So far the NSS module looks up only hostnames as provided by
guests themselves. However, there are some cases where this is
not enough: e.g. when there's a fresh new guest being installed
(with some generic hostname) say from a live ISO image; or some
(older) systems don't advertise their hostname in DHCP
transactions at all.
In cases like that it would be helpful if we translate domain
name as seen by libvirt too so that users can:
# virsh start $dom && ssh $dom
In order to achieve that new libvirt-guest module is introduced,
while older libvirt module maintains its current behaviour (that
is translating guest provided names into IP addresses).
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Now that we have a module that's able to track
<domain, mac addres list> pairs, hook it up into
our network driver.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
This module will be used to track:
<domain, mac address list>
pairs. It will be important to know these mappings without
libvirt connection (that is from a JSON file), because NSS
module will use those to provide better host name translation.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
There are couple of places where we have a string and want to
save it to a file. Atomically. In all those places we use
virFileRewrite() but also implement the very same callback which
takes the string and write it into temp file. This makes no
sense. Unify the callbacks and move them to one place.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
In dd7bfb2cdc I've removed locking of the network driver upon
it's allocation. However, I forgot to remove one location of the
driver unlock.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Add in the block I/O throttling group parameter to the command line
if supported. If not supported, fail command creation.
Add the xml2argvtest for testing.
Signed-off-by: John Ferlan <jferlan@redhat.com>
Modify _virDomainBlockIoTuneInfo and rng schema to support the group_name
option for iotune throttling. Document the new value.
Signed-off-by: John Ferlan <jferlan@redhat.com>
Rather than have multiple bool values, create a single enum with bits
representing what fields are set. Fields are generally set in groups
of 3 (read, write, total).
Currently we build the JSON object for the "block_set_io_throttle"
command using the knowledge that a NULL for a support*Options boolean
would essentially ignore the rest of the arguments.
This may not work properly if some capability was backported, plus it just
looks rather ugly. So instead, build the "base" arguments and then if
the support*Option bool capability is set, add in the arguments on the fly.
Then append those arguments to the basic command and send to qemu.
Rather than using negative logic and setting the maxparams to a lesser
value based on which capabilities exist, alter the logic to modify the
maxparams based on a base value plus the found capabilities. Reduces the
chance that some backported feature produces an incorrect value.
These features are included:
AVX512DQ, AVX512IFMA, AVX512BW, AVX512VL, AVX512VBMI, AVX512_4VNNIW and
AVX512_4FMAPS.
qemu commits: cc728d14 and 95ea69fb
Signed-off-by: Lin Ma <lma@suse.com>
Commit id '03e750f3' added support for checking the PLOOP type; however,
it used 'target.type' which no storage code ever fills in, so it will
never be set. Change to just vol->type (could use vol->target.format
as well).
Add a global check for duplicate drive addresses. This will fix the
problem of duplicate disk and hostdev drive addresses.
Example for duplicate drive addresses:
<disk>
...
<target name='sda'/>
</disk>
<disk>
...
<target name='sdb'/>
<address type='drive' controller=0 bus=0 target=0 unit=0/>
</disk>
Another example:
<hostdev mode='subsystem' type='scsi' managed='no'>
<source>
...
</source>
<address type='drive' controller='0' bus='0' target='0' unit='0'/>
</hostdev>
<hostdev mode='subsystem' type='scsi' managed='no'>
<source>
...
</source>
<address type='drive' controller='0' bus='0' target='0' unit='0'/>
</hostdev>
Unfortunately the fixes (1b08cc170a,
8d46386bfe) weren't enough to catch these
cases and it isn't possible to add additional checks in
virDomainDeviceDefPostParseInternal() for SCSI hostdevs or
virDomainDiskDefAssignAddress() for SCSI/IDE/FDC/SATA disks without
adding another parse flag (virDomainDefParseFlags) to disable this
validation while updating or detaching a disk or hostdev.
Signed-off-by: Marc Hartmayer <mhartmay@linux.vnet.ibm.com>
Reviewed-by: Boris Fiuczynski <fiuczy@linux.vnet.ibm.com>
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Comparing the parameter 'type' against the member 'bus' instead of
against the member 'type' is quite confusing. Rename the parameter
'type' to 'bus_type' to clarify its meaning.
Signed-off-by: Marc Hartmayer <mhartmay@linux.vnet.ibm.com>
Reviewed-by: Boris Fiuczynski <fiuczy@linux.vnet.ibm.com>
Pass the virDomainDeviceDriveAddress as a struct instead of individual
arguments. Reworked the function descriptions.
Signed-off-by: Marc Hartmayer <mhartmay@linux.vnet.ibm.com>
Reviewed-by: Boris Fiuczynski <fiuczy@linux.vnet.ibm.com>
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Two reasons:
1.in none hotplug, we will pass it. We can see from libvirt function
qemuBuildVhostuserCommandLine
2.qemu will use this vetcor num to init msix table. If we don't pass, qemu
will use default value, this will cause VM can only use default value
interrupts at most.
Signed-off-by: gaohaifeng <gaohaifeng.gao@huawei.com>
Consider the following XML snippets:
$ cat scsicontroller.xml
<controller type='scsi' model='virtio-scsi' index='0'/>
$ cat scsihostdev.xml
<hostdev mode='subsystem' type='scsi'>
<source>
<adapter name='scsi_host0'/>
<address bus='0' target='8' unit='1074151456'/>
</source>
</hostdev>
If we create a guest that includes the contents of scsihostdev.xml,
but forget the virtio-scsi controller described in scsicontroller.xml,
one is silently created for us. The same holds true when attaching
a hostdev before the matching virtio-scsi controller.
(See qemuDomainFindOrCreateSCSIDiskController for context.)
Detaching the hostdev, followed by the controller, works well and the
guest behaves appropriately.
If we detach the virtio-scsi controller device first, any associated
hostdevs are detached for us by the underlying virtio-scsi code (this
is fine, since the connection is broken). But all is not well, as the
guest is unable to receive new virtio-scsi devices (the attach commands
succeed, but devices never appear within the guest), nor even be
shutdown, after this point.
While this is not libvirt's problem, we can prevent falling into this
scenario by checking if a controller is being used by any hostdev
devices. The same is already done for disk elements today.
Applying this patch and then using the XML snippets from earlier:
$ virsh detach-device guest_01 scsicontroller.xml
error: Failed to detach device from scsicontroller.xml
error: operation failed: device cannot be detached: device is busy
$ virsh detach-device guest_01 scsihostdev.xml
Device detached successfully
$ virsh detach-device guest_01 scsicontroller.xml
Device detached successfully
Signed-off-by: Eric Farman <farman@linux.vnet.ibm.com>
Reviewed-by: Bjoern Walk <bwalk@linux.vnet.ibm.com>
Reviewed-by: Boris Fiuczynski <fiuczy@linux.vnet.ibm.com>
Although nearly all host devices that are assigned to guests using
VFIO ("<hostdev>" devices in libvirt) are physically PCI Express
devices, until now libvirt's PCI address assignment has always
assigned them addresses on legacy PCI controllers in the guest, even
if the guest's machinetype has a PCIe root bus (e.g. q35 and
aarch64/virt).
This patch tries to assign them to an address on a PCIe controller
instead, when appropriate. First we do some preliminary checks that
might allow setting the flags without doing any extra work, and if
those conditions aren't met (and if libvirt is running privileged so
that it has proper permissions), we perform the (relatively) time
consuming task of reading the device's PCI config to see if it is an
Express device. If this is successful, the connect flags are set based
on the result, but if we aren't able to read the PCI config (most
likely due to the device not being present on the system at the time
of the check) we assume it is (or will be) an Express device, since
that is almost always the case anyway.
If libvirtd is running unprivileged, it can open a device's PCI config
data in sysfs, but can only read the first 64 bytes. But as part of
determining whether a device is Express or legacy PCI,
qemuDomainDeviceCalculatePCIConnectFlags() will be updated in a future
patch to call virPCIDeviceIsPCIExpress(), which tries to read beyond
the first 64 bytes of the PCI config data and fails with an error log
if the read is unsuccessful.
In order to avoid creating a parallel "quiet" version of
virPCIDeviceIsPCIExpress(), this patch passes a virQEMUDriverPtr down
through all the call chains that initialize the
qemuDomainFillDevicePCIConnectFlagsIterData, and saves the driver
pointer with the rest of the iterdata so that it can be used by
qemuDomainDeviceCalculatePCIConnectFlags(). This pointer isn't used
yet, but will be used in an upcoming patch (that detects Express vs
legacy PCI for VFIO assigned devices) to examine driver->privileged.
The path to the config file for a PCI device is conventiently stored
in a virPCIDevice object, but that object's contents aren't directly
visible outside of virpci.c, so we need to have an accessor function
for it if anyone needs to look at it.
This new function just calls fstat() (if provided with a valid fd) or
stat() (if fd is -1) and returns st_size (or -1 if there is an
error). We may decide we want this function to be more complex, and
handle things like block devices - this is a placeholder (that works)
for any more complicated function.
We can't change feature names for compatibility reasons even if they
contain typos or other software uses different names for the same
features. By adding alternative spellings in our CPU map we at least
allow anyone to grep for them and find the correct libvirt's name.
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
When virt-aa-helper parses xml content it can fail on security labels.
It fails by requiring to parse active domain content on seclabels that
are not yet filled in.
Testcase with virt-aa-helper on a minimal xml:
$ cat << EOF > /tmp/test.xml
<domain type='kvm'>
<name>test-seclabel</name>
<uuid>12345678-9abc-def1-2345-6789abcdef00</uuid>
<memory unit='KiB'>1</memory>
<os><type arch='x86_64'>hvm</type></os>
<seclabel type='dynamic' model='apparmor' relabel='yes'/>
<seclabel type='dynamic' model='dac' relabel='yes'/>
</domain>
EOF
$ /usr/lib/libvirt/virt-aa-helper -d -r -p 0 \
-u libvirt-12345678-9abc-def1-2345-6789abcdef00 < /tmp/test.xml
Current Result:
virt-aa-helper: error: could not parse XML
virt-aa-helper: error: could not get VM definition
Expected Result is a valid apparmor profile
Signed-off-by: Christian Ehrhardt <christian.ehrhardt@canonical.com>
Signed-off-by: Guido Günther <agx@sigxcpu.org>
Only the latest APIs are fully documented and the documentation of the
older variants (which are just limited versions of the new APIs anyway)
points to the newest APIs.
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
Restarting libvirtd on the source host at the end of migration when a
domain is already running on the destination would cause image labels to
be reset effectively killing the domain. Commit e8d0166e1d fixed similar
issue on the destination host, but kept the source always resetting the
labels, which was mostly correct except for the specific case handled by
this patch.
https://bugzilla.redhat.com/show_bug.cgi?id=1343858
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
Post-copy migration needs bi-directional communication between the
source and the destination QEMU processes, which is not supported by
tunnelled migration.
https://bugzilla.redhat.com/show_bug.cgi?id=1371358
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
We had a lot of rados_conf_set and check works.
Use helper virStorageBackendRBDRADOSConfSet for them.
Signed-off-by: Chen Hanxiao <chenhanxiao@gmail.com>
Thanks to the complex capability caching code virQEMUCapsProbeQMP was
never called when we were starting a new qemu VM. On the other hand,
when we are reconnecting to the qemu process we reload the capability
list from the status XML file. This means that the flag preventing the
function being called was not set and thus we partially reprobed some of
the capabilities.
The recent addition of CPU hotplug clears the
QEMU_CAPS_QUERY_HOTPLUGGABLE_CPUS if the machine does not support it.
The partial re-probe on reconnect results into attempting to call the
unsupported command and then killing the VM.
Remove the partial reprobe and depend on the stored capabilities. If it
will be necessary to reprobe the capabilities in the future, we should
do a full reprobe rather than this partial one.
QEMU 2.8.0 adds support for unavailable-features in
query-cpu-definitions reply. The unavailable-features array lists CPU
features which prevent a corresponding CPU model from being usable on
current host. It can only be used when all the unavailable features are
disabled. Empty array means the CPU model can be used without
modifications.
We can use unavailable-features for providing CPU model usability info
in domain capabilities XML:
<domainCapabilities>
...
<cpu>
<mode name='host-passthrough' supported='yes'/>
<mode name='host-model' supported='yes'>
<model fallback='allow'>Skylake-Client</model>
...
</mode>
<mode name='custom' supported='yes'>
<model usable='yes'>qemu64</model>
<model usable='yes'>qemu32</model>
<model usable='no'>phenom</model>
<model usable='yes'>pentium3</model>
<model usable='yes'>pentium2</model>
<model usable='yes'>pentium</model>
<model usable='yes'>n270</model>
<model usable='yes'>kvm64</model>
<model usable='yes'>kvm32</model>
<model usable='yes'>coreduo</model>
<model usable='yes'>core2duo</model>
<model usable='no'>athlon</model>
<model usable='yes'>Westmere</model>
<model usable='yes'>Skylake-Client</model>
...
</mode>
</cpu>
...
</domainCapabilities>
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
"host" CPU model is supported by a special host-passthrough CPU mode and
users is not allowed to specify this model directly with custom mode.
Thus we should not advertise "host" CPU model in domain capabilities.
This worked well on architectures for which libvirt provides a list of
supported CPU models in cpu_map.xml (since "host" is not in the list).
But we need to explicitly filter "host" model out for all other
architectures.
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
CPU models (and especially some additional details which we will start
probing for later) differ depending on the accelerator. Thus we need to
call query-cpu-definitions in both KVM and TCG mode to get all data we
want.
Tests in tests/domaincapstest.c are temporarily switched to TCG to avoid
having to squash even more stuff into this single patch. They will all
be switched back later in separate commits.
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
This patch moves the CPU models formatting code from
virQEMUCapsFormatCache into a separate function.
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
The function just returned cached capabilities without checking whether
they are still valid. We should check that and refresh the capabilities
to make sure we don't return stale data. In other words, we should do
what all other lookup functions do.
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
The function is made a little bit more readable and the code which
refreshes cached capabilities if they are not valid any more was moved
into a separate function (virQEMUCapsCacheValidate) so that it can be
reused in other places.
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
If a user asked for a KVM domain capabilities when KVM is not available,
we would happily return data we got when probing through TCG and
pretended they were relevant for KVM. Let's just report KVM is not
supported to avoid confusion.
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
When domain capabilities were introduced we did not have enough data to
decide whether KVM works on the host or not and thus working legacy/VFIO
device assignment was used as a witness. Now that we know whether KVM
was enabled when probing QEMU capabilities (and thus we know it's
working), we can use this knowledge to provide better default value for
virttype.
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
Since some may depend on the accelerator used when probing QEMU the
cache becomes invalid when KVM becomes available or if it is not
available anymore.
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
CPU related capabilities may differ depending on accelerator used when
probing. Let's use KVM if available when probing QEMU and fall back to
TCG. The created capabilities already contain all we need to distinguish
whether KVM or TCG was used:
- KVM was used when probing capabilities:
QEMU_CAPS_KVM is set
QEMU_CAPS_ENABLE_KVM is not set
- TCG was used and QEMU supports KVM, but it failed (e.g., missing
kernel module or wrong /dev/kvm permissions)
QEMU_CAPS_KVM is not set
QEMU_CAPS_ENABLE_KVM is set
- KVM was not used and QEMU does not support it
QEMU_CAPS_KVM is not set
QEMU_CAPS_ENABLE_KVM is not set
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
Let's set QEMU_CAPS_KVM and QEMU_CAPS_ENABLE_KVM early so that the rest
of the probing code can use these capabilities to handle KVM/TCG replies
differently.
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
Using -machine instead of -M for QMP probing is safe because any QEMU
binary which is capable of QMP probing supports -machine.
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
The code that runs a new QEMU process to be used for probing
capabilities is separated into four reusable functions so that any code
that wants to probe a QEMU process may just follow a few simple steps:
cmd = virQEMUCapsInitQMPCommandNew(...);
virQEMUCapsInitQMPCommandRun(cmd);
/* talk to the running QEMU process using its QMP monitor */
if (reprobeIsRequired) {
virQEMUCapsInitQMPCommandAbort(cmd, ...);
virQEMUCapsInitQMPCommandRun(cmd);
/* talk to the running QEMU process again */
}
virQEMUCapsInitQMPCommandFree(cmd);
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
This reverts commit 3a6cf6fc16.
Mistakenly this commit was pushed because I thought I missed the
corret one b880ff42dd while in fact I didn't.
Signed-off-by: Maxim Nestratov <mnestratov@virtuozzo.com>
We have couple of functions that operate over NULL terminated
lits of strings. However, our naming sucks:
virStringJoin
virStringFreeList
virStringFreeListCount
virStringArrayHasString
virStringGetFirstWithPrefix
We can do better:
virStringListJoin
virStringListFree
virStringListFreeCount
virStringListHasString
virStringListGetFirstWithPrefix
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
If libvirt is compiled without NUMACTL support starting libvirtd
reports a libvirt internal error "NUMA isn't available on this host"
without checking if NUMA support is compiled into the libvirt binaries.
This patch adds the missing NUMA support check to prevent the internal error.
It also includes a check if the cgroup controller cpuset is available before
using it.
The error was noticed when libvirtd was restarted with running domains and
on libvirtd start the qemuConnectCgroup gets called during qemuProcessReconnect.
Signed-off-by: Boris Fiuczynski <fiuczy@linux.vnet.ibm.com>
Reviewed-by: Bjoern Walk <bwalk@linux.vnet.ibm.com>
With the QEMU components in place, provide the XML parsing to
invoke that code when given the following XML snippet:
<hostdev mode='subsystem' type='scsi_host'>
<source protocol='vhost' wwpn='naa.501234567890abcd'/>
</hostdev>
An optional address element can be specified within the hostdev
(pick CCW or PCI as necessary):
<address type='ccw' cssid='0xfe' ssid='0x0' devno='0x0625'/>
<address type='pci' domain='0x0000' bus='0x00' slot='0x05' function='0x0'/>
Add basic vhost-scsi tests which were cloned from hostdev-scsi-virtio-scsi
in both xml2argv and xml2xml. Added ones for both vhost-scsi-ccw and
vhost-scsi-pci since the syntaxes are slightly different between them.
Also adjusted the docs to describe the changes.
Signed-off-by: Eric Farman <farman@linux.vnet.ibm.com>
Reviewed-by: Boris Fiuczynski <fiuczy@linux.vnet.ibm.com>
Adjust the device string that is built for vhost-scsi devices so that it
can be invoked from hotplug.
From the QEMU command line, the file descriptors are expect to be numeric only.
However, for hotplug, the file descriptors are expected to begin with at least
one alphabetic character else this error occurs:
# virsh attach-device guest_0001 ~/vhost.xml
error: Failed to attach device from /root/vhost.xml
error: internal error: unable to execute QEMU command 'getfd':
Parameter 'fdname' expects a name not starting with a digit
We also close the file descriptor in this case, so that shutting down the
guest cleans up the host cgroup entries and allows future guests to use
vhost-scsi devices. (Otherwise the guest will silently end.)
Signed-off-by: Eric Farman <farman@linux.vnet.ibm.com>
Open /dev/vhost-scsi, and record the resulting file descriptor, so that
the guest has access to the host device outside of the libvirt daemon.
Pass this information, along with data parsed from the XML file, to build
a device string for the qemu command line. That device string will be
for either a vhost-scsi-ccw device in the case of an s390 machine, or
vhost-scsi-pci for any others.
Signed-off-by: Eric Farman <farman@linux.vnet.ibm.com>
For a new hostdev type='scsi_host' we have a number of
required functions for managing, adding, and removing the
host device to/from guests. Provide the basic infrastructure
for these tasks.
The name "SCSIVHost" (and its variants) is chosen to avoid
conflicts with existing code named "SCSIHost" to refer to
a hostdev type='scsi' protcol='none'.
Signed-off-by: Eric Farman <farman@linux.vnet.ibm.com>
We already have a "scsi" hostdev subsys type, which refers to a single
LUN that is passed through to a guest. But what of things where
multiple LUNs are passed through via a single SCSI HBA, such as with
the vhost-scsi target? Create a new hostdev subsys type that will
carry this.
Signed-off-by: Eric Farman <farman@linux.vnet.ibm.com>
Do all the stuff for the vhost-scsi capability in QEMU,
so it's in place for our checks later.
Signed-off-by: Eric Farman <farman@linux.vnet.ibm.com>
Reviewed-by: Boris Fiuczynski <fiuczy@linux.vnet.ibm.com>
* add vboxDriver object to serve as a singleton global object that
holds references to IVirtualBox and ISession to be shared among
multiple connections. The vbox_driver is instantiated only once in
the first call vboxGetDriverConnection function that is guarded by
a mutex.
* call vbox API initialize only when the first connection is
established, and likewise uninitialize when last connection
disconnects. The prevents each subsequent connection from overwriting
IVirtualBox/ISession instances of any other active connection that
led to libvirtd segfaults. The virConnectOpen and virConnectClose
implementations are guarded by mutex on the global vbox_driver_lock
where the global vbox_driver object counts connectios and decides
when it's safe to call vbox's init/uninit routines.
* add IVirutalBoxClient to vboxDriver and use it to in tandem with newer
pfnClientInitialize/pfnClientUninitalize APIs for vbox versions that
support it, to avoid usage of the old pfnComInitialize/Uninitialize.
Removed the comment 'Set the migration source' as it isn't valid anymore
and 'start it up' isn't useful as qemuProcessStart() is already a
speaking name.
Signed-off-by: Marc Hartmayer <mhartmay@linux.vnet.ibm.com>
The path prefixes for sysfs trees are always prepended by paths
beginning with a slash, making the trailing slash in the prefix
redundant.
Signed-off-by: Bjoern Walk <bwalk@linux.vnet.ibm.com>
Reviewed-by: Marc Hartmayer <mhartmay@linux.vnet.ibm.com>
Reviewed-by: Boris Fiuczynski <fiuczy@linux.vnet.ibm.com>
Only generate a warning if there is something to report.
Signed-off-by: Marc Hartmayer <mhartmay@linux.vnet.ibm.com>
Reviewed-by: Bjoern Walk <bwalk@linux.vnet.ibm.com>
Reviewed-by: Boris Fiuczynski <fiuczy@linux.vnet.ibm.com>
Just like in the previous commit, we are not updating CGroups on
chardev hot(un-)plug and thus leaving qemu unable to access any
non-default device users are trying to hotplug.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
If users try to hotplug RNG device with a backend different to
/dev/random or /dev/urandom the whole operation fails as qemu is
unable to access the device. The problem is we don't update
device CGroups during the operation.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
qemuDomainObjExitAgent is unsafe.
First it accesses domain object without domain lock.
Second it uses outdated logic that goes back to commit 79533da1 of
year 2009 when code was quite different. (unref function
instead of unreferencing only unlocked and disposed object
in case of last reference and leaved unlocking to the caller otherwise).
Nowadays this logic may lead to disposing locked object
i guess.
Another problem is that the callers of qemuDomainObjEnterAgent
use domain object again (namely priv->agent) without domain lock.
This patch address these two problems.
qemuDomainGetAgent is dropped as unused.
Sometimes after domain restart agent is unavailabe even
if it is up and running in guest. Diagnostic message is
"QEMU guest agent is not available due to an error"
that is 'priv->agentError' is set. Investiagion shows that
'priv->agent' is not NULL, so error flag is set probably
during domain shutdown process and not cleaned up eventually.
The patch is quite simple - just clean up error flag unconditionally
upon domain stop.
Other hunks address other cases when error flag is not cleaned up.
1. processSerialChangedEvent. We need to clean error flag
unconditionally here too. For example if upon first 'connected' event we
fail to connect and set error flag and then connect on second
'connected' event then error flag will remain set erroneously
and make agent unavailable.
2. qemuProcessHandleAgentEOF. If error flag is set and we get
EOF we need to change state (and diagnostic) from 'error' to
'not connected'.
qemuConnectAgent return -1 or -2 in case of different errors.
A. -1 is a case of unsuccessuful connection to guest agent.
B. -2 is a case of destoyed domain during connection attempt.
All qemuConnectAgent callers handle the first error the same way
so let's move this logic into qemuConnectAgent itself. Patched
function returns 0 in case A and -1 in case B.
It is already discussed in "[RFC] daemon: remove hardcode dep on libvirt-guests" [1].
Mgmt can use means to save/restore domains on system shutdown/boot other than
libvirt-guests.service. Thus we need to specify appropriate ordering dependency between
libvirtd, domains and save/restore service. This patch takes approach suggested
in RFC and introduces a systemd target, so that ordering can be built next way:
libvirtd -> domain -> virt-guest-shutdown.target -> save-restore.service.
This way domains are decoupled from specific shutdown service via intermediate
target.
[1] https://www.redhat.com/archives/libvir-list/2016-September/msg01353.html
Use the util function virHostdevIsSCSIDevice() to simplify if
statements.
Signed-off-by: Marc Hartmayer <mhartmay@linux.vnet.ibm.com>
Reviewed-by: Bjoern Walk <bwalk@linux.vnet.ibm.com>
Reviewed-by: Boris Fiuczynski <fiuczy@linux.vnet.ibm.com>
Add the function virHostdevIsSCSIDevice() which detects whether a
hostdev is a SCSI device or not.
Signed-off-by: Marc Hartmayer <mhartmay@linux.vnet.ibm.com>
Reviewed-by: Bjoern Walk <bwalk@linux.vnet.ibm.com>
Reviewed-by: Boris Fiuczynski <fiuczy@linux.vnet.ibm.com>
Add missing checks if a hostdev is a subsystem/SCSI device before access
the union member 'subsys'/'scsi'. Also fix indentation and simplify
qemuDomainObjCheckHostdevTaint().
Signed-off-by: Marc Hartmayer <mhartmay@linux.vnet.ibm.com>
Reviewed-by: Bjoern Walk <bwalk@linux.vnet.ibm.com>
Reviewed-by: Boris Fiuczynski <fiuczy@linux.vnet.ibm.com>
New line character in name of storagepool is now forbidden because it
mess virsh output and can be confusing for users.
Validation of name is done in driver, after parsing XML to avoid
problems with dissappeared pools which was already created with
new-line char in name.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
New line character in name of domain is now forbidden because it
mess virsh output and can be confusing for users.
Validation of name is done in drivers, after parsing XML to avoid
problems with dissappeared domains which was already created with
new-line char in name.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Commit 3f71c79768 added 'qemu_id' field to track the id of the cpu
as reported by query-cpus. The patch did not include changes necessary
to propagate the id through the functions matching the data to the
libvirt cpu structures and thus all vcpus had id 0.
Don't use qemuMonitorGetCPUInfo which does a lot of matching to get the
full picture which is not necessary and would be mostly discarded.
Refresh only the vcpu halted state using data from query-cpus.
We don't need to call qemuMonitorGetCPUInfo which is very inefficient to
get data required to update the vcpu 'halted' state.
Add a monitor helper that will retrieve the halted state and return it
in a bitmap so that it can be indexed easily.
Whenever qemuMonitorJSONCheckError returns 0, the "return" object is
guaranteed to exist. Thus virJSONValueObjectGetObject will never fail to
get it. On the other hand, virJSONValueObjectGetArray may fail since the
"return" object may not be an array.
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
Libvirt's code relies on this fact so don't allow parsing a command line
which would have none.
Libvirtd would crash in the post parse callback on such config.
After a944bd92 we gained support for setting gluster debug level.
However, due to a space we haven't tested whether augeas file
actually works.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
After commit f2bf5fbb04, MinGW strikes again. Simply print pid as any
other place after commit b7d2d4af2b.
Signed-off-by: Martin Kletzander <mkletzan@redhat.com>
Commit 94cc577807 tried fixing build on systems that did not have
SCHED_BATCH or SCHED_IDLE defined. But instead of changing it to
conditional support, it rather completely disabled the support for
setting any scheduler. Since then, such old systems are not
supported, but rather than reverting that commit, let's change that to
the conditional support. That way any addition to the list of
schedulers can follow the same style so that we're consistent in the
future.
Signed-off-by: Martin Kletzander <mkletzan@redhat.com>
https://bugzilla.redhat.com/show_bug.cgi?id=1366460
When using the --overwrite switch on a pool-build or pool-create, the
The mkfs.ext{2|3|4} commands use mke2fs which requires using the '-F' switch
in order to force overwriting the current filesystem on the whole disk.
Likewise, the mkfs.vfat command uses mkfs.fat which requires using the '-I'
switch in order to force overwriting the current filesystem on the whole disk.
Old GCC on CentOS 6 thinks vendor and vendor_id might be used
uninitialized in virCPUDefStealModel. The compiler is wrong, though.
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
Implement in virtNetClient and VirNetSocket the needed functions to
expose a new libssh transport, providing all the options that the
libssh2 transport supports.
Implement a new libssh transport, which uses libssh to communicate with
remote hosts, and add all the build system stuff (search of libssh,
private symbols, etc) to built it.
This new transport supports all the common ssh authentication methods,
making use of libvirt's auth callbacks for interaction with the user.
Add a couple of helper functions to check whether one of the default
names of SSH keys (as documented in ssh-keygen(1)) exists, and use them
to specify a key for the libssh2 transport if none was passed.
Add an internal variable to mark the FD as "not owned" by the
virNetSocket, in case the internal implementation takes the actual
ownership of the descriptor; this avoids a warning when closing the
socket, as the FD would be invalid.
Guest CPU definitions with mode='custom' and missing <vendor> are
expected to run on a host CPU from any vendor as long as the required
CPU model can be used as a guest CPU on the host. But even though no CPU
vendor was explicitly requested we would sometimes force it due to a bug
in virCPUUpdate and virCPUTranslate.
The bug would effectively forbid cross vendor migrations even if they
were previously working just fine.
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
PPC driver needs to convert POWERx_v* legacy CPU model names into POWERx
to maintain backward compatibility with existing domains. This patch
adds a new step into the guest CPU configuration work flow which CPU
drivers can use to convert legacy CPU definitions.
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
It was introduced by commit 7a51d9ebb, which started to use
monitor commands without job acquiring, which is unsafe and leads
to simultaneous access to vm->mon structure by different threads.
Crash backtrace is the following (shortened):
Program received signal SIGSEGV, Segmentation fault.
qemuMonitorSend (mon=mon@entry=0x7f4ef4000d20, msg=msg@entry=0x7f4f18e78640) at qemu/qemu_monitor.c:1011
1011 while (!mon->msg->finished) {
0 qemuMonitorSend () at qemu/qemu_monitor.c:1011
1 0x00007f691abdc720 in qemuMonitorJSONCommandWithFd () at qemu/qemu_monitor_json.c:298
2 0x00007f691abde64a in qemuMonitorJSONCommand at qemu/qemu_monitor_json.c:328
3 qemuMonitorJSONQueryCPUs at qemu/qemu_monitor_json.c:1408
4 0x00007f691abcaebd in qemuMonitorGetCPUInfo g@entry=false) at qemu/qemu_monitor.c:1931
5 0x00007f691ab96863 in qemuDomainRefreshVcpuHalted at qemu/qemu_domain.c:6309
6 0x00007f691ac0af99 in qemuDomainGetStatsVcpu at qemu/qemu_driver.c:18945
7 0x00007f691abef921 in qemuDomainGetStats at qemu/qemu_driver.c:19469
8 qemuConnectGetAllDomainStats at qemu/qemu_driver.c:19559
9 0x00007f693382e806 in virConnectGetAllDomainStats at libvirt-domain.c:11546
10 0x00007f6934470c40 in remoteDispatchConnectGetAllDomainStats at remote.c:6267
(gdb) p mon->msg
$1 = (qemuMonitorMessagePtr) 0x0
This change fixes it by calling qemuDomainRefreshVcpuHalted only when job is acquired.
Signed-off-by: Maxim Nestratov <mnestratov@virtuozzo.com>
For machinetypes with a pci-root bus (all legacy PCI), libvirt will
make a "fake" reservation for one extra slot prior to assigning
addresses to unaddressed PCI endpoint devices in the domain. This will
trigger auto-adding of a pci-bridge for the final device to be
assigned an address *if that device would have otherwise instead been
the last device on the last available pci-bridge*; thus it assures
that there will always be at least one slot left open in the domain's
bus topology for expansion (which is important both for hotplug (since
a new pci-bridge can't be added while the guest is running) as well as
for offline additions to the config (since adding a new device might
otherwise in some cases require re-addressing existing devices, which
we want to avoid)).
It's important to note that for the above case (legacy PCI), we must
check for the special case of all slots on all buses being occupied
*prior to assigning any addresses*, and avoid attempting to reserve
the extra address in that case, because there is no free address in
the existing topology, so no place to auto-add a pci-bridge for
expansion (i.e. it would always fail anyway). Since that condition can
only be reached by manual intervention, this is acceptable.
For machinetypes with pcie-root (Q35, aarch64 virt), libvirt's
methodology for automatically expanding the bus topology is different
- pcie-root-ports are plugged into slots (soon to be functions) of
pcie-root as needed, and the new endpoint devices are assigned to the
single slot in each pcie-root-port. This is done so that the devices
are, by default, hotpluggable (the slots of pcie-root don't support
hotplug, but the single slot of the pcie-root-port does). Since
pcie-root-ports can only be plugged into pcie-root, and we don't
auto-assign endpoint devices to the pcie-root slots, this means
topology expansion doesn't compete with endpoint devices for slots, so
we don't need to worry about checking for all "useful" slots being
free *prior* to assigning addresses to new endpoint devices - as a
matter of fact, if we attempt to reserve the open slots before the
used slots, it can lead to errors.
Instead this patch just reserves one slot for a "future potential"
PCIe device after doing the assignment for actual devices, but only
if the only PCI controller defined prior to starting address
assignment was pcie-root, and only if we auto-added at least one PCI
controller during address assignment. This assures two things:
1) that reserving the open slots will only be done when the domain is
initially defined, never at any time after, and
2) that if the user understands enough about PCI controllers that they
are adding them manually, that we don't mess up their plan by
adding extras - if they know enough to add one pcie-root-port, or
to manually assign addresses such that no pcie-root-ports are
needed, they know enough to add extra pcie-root-ports if they want
them (this could be called the "libguestfs clause", since
libguestfs needs to be able to create domains with as few
devices/controllers as possible).
This is set to reserve a single free port for now, but could be
increased in the future if public sentiment goes in that direction
(it's easy to increase later, but essentially impossible to decrease)
Real Q35 hardware has an ICH9 chip that includes several integrated
devices at particular addresses (see the file docs/q35-chipset.cfg in
the qemu source). libvirt already attempts to put the first two sets
of ich9 USB2 controllers it finds at 00:1D.* and 00:1A.* to match the
real hardware. This patch does the same for the ich9 "HD audio"
device.
The main inspiration for this patch is that currently the *only*
device in a reasonable "workstation" type virtual machine config that
requires a legacy PCI slot is the audio device, Without this patch,
the standard Q35 machine created by virt-manager will have a
dmi-to-pci-bridge and a pci-bridge just for the sound device; with the
patch (and if you change the sound device model from the default
"ich6" to "ich9"), the machine definition constructed by virt-manager
has absolutely no legacy PCI controllers - any legacy PCI devices
(e.g. video and sound) are on pcie-root as integrated devices.
Previously we added a set of EHCI+UHCI controllers to Q35 machines to
mimic real hardware as closely as possible, but recent discussions
have pointed out that the nec-usb-xhci (USB3) controller is much more
virtualization-friendly (uses less CPU), so this patch switches the
default for Q35 machinetypes to add an XHCI instead (if it's
supported, which it of course *will* be).
Since none of the existing test cases left out USB controllers in the
input XML, a new Q35 test case was added which has *no* devices, so
ends up with only the defaults always put in by qemu, plus those added
by libvirt.
Now the a dmi-to-pci-bridge is automatically added just as it's needed
(when a pci-bridge is being added), we no longer have any need to
force-add one to every single Q35 domain.
Previously libvirt would only add pci-bridge devices automatically
when an address was requested for a device that required a legacy PCI
slot and none was available. This patch expands that support to
dmi-to-pci-bridge (which is needed in order to add a pci-bridge on a
machine with a pcie-root), and pcie-root-port (which is needed to add
a hotpluggable PCIe device). It does *not* automatically add
pcie-switch-upstream-ports or pcie-switch-downstream-ports (and
currently there are no plans for that).
Given the existing code to auto-add pci-bridge devices, automatically
adding pcie-root-ports is fairly straightforward. The
dmi-to-pci-bridge support is a bit tricky though, for a few reasons:
1) Although the only reason to add a dmi-to-pci-bridge is so that
there is a reasonable place to plug in a pci-bridge controller,
most of the time it's not the presence of a pci-bridge *in the
config* that triggers the requirement to add a dmi-to-pci-bridge.
Rather, it is the presence of a legacy-PCI device in the config,
which triggers auto-add of a pci-bridge, which triggers auto-add of
a dmi-to-pci-bridge (this is handled in
virDomainPCIAddressSetGrow() - if there's a request to add a
pci-bridge we'll check if there is a suitable bus to plug it into;
if not, we first add a dmi-to-pci-bridge).
2) Once there is already a single dmi-to-pci-bridge on the system,
there won't be a need for any more, even if it's full, as long as
there is a pci-bridge with an open slot - you can also plug
pci-bridges into existing pci-bridges. So we have to make sure we
don't add a dmi-to-pci-bridge unless there aren't any
dmi-to-pci-bridges *or* any pci-bridges.
3) Although it is strongly discouraged, it is legal for a pci-bridge
to be directly plugged into pcie-root, and we don't want to
auto-add a dmi-to-pci-bridge if there is already a pci-bridge
that's been forced directly into pcie-root.
Although libvirt will now automatically create a dmi-to-pci-bridge
when it's needed, the code still remains for now that forces a
dmi-to-pci-bridge on all domains with pcie-root (in
qemuDomainDefAddDefaultDevices()). That will be removed in a future
patch.
For now, the pcie-root-ports are added one to a slot, which is a bit
wasteful and means it will fail after 31 total PCIe devices (30 if
there are also some PCI devices), but helps keep the changeset down
for this patch. A future patch will have 8 pcie-root-ports sharing the
functions on a single slot.
Andrea had the right idea when he disabled the "reserve an extra
unused slot" bit for aarch64/virt. For *any* PCI Express-based
machine, it is pointless since 1) an extra legacy-PCI slot can't be
used for hotplug, since hotplug into legacy PCI slots doesn't work on
PCI Express machinetypes, and 2) even for "coldplug" expansion,
everybody will want to expand using Express controllers, not legacy
PCI.
This patch eliminates the extra slot reserve unless the system has a
pci-root (i.e. legacy PCI)
The nec-usb-xhci device (which is a USB3 controller) has always
presented itself as a PCI device when plugged into a legacy PCI slot,
and a PCIe device when plugged into a PCIe slot, but libvirt has
always auto-assigned it to a legacy PCI slot.
This patch changes that behavior to auto-assign to a PCIe slot on
systems that have pcie-root (e.g. Q35 and aarch64/virt).
Since we don't yet auto-create pcie-*-port controllers on demand, this
means a config with an nec-xhci USB controller that has no PCI address
assigned will also need to have an otherwise-unused pcie-*-port
controller specified:
<controller type='pci' model='pcie-root-port'/>
<controller type='usb' model='nec-xhci'/>
(this assumes there is an otherwise-unused slot on pcie-root to accept
the pcie-root-port)
The e1000e is an emulated network device based on the Intel 82574,
present in qemu 2.7.0 and later. Among other differences from the
e1000, it presents itself as a PCIe device rather than legacy PCI. In
order to get it assigned to a PCIe controller, this patch updates the
flags setting for network devices when the model name is "e1000e".
(Note that for some reason libvirt has never validated the network
device model names other than to check that there are no dangerous
characters in them. That should probably change, but is the subject of
another patch.)
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1343094
libvirt previously assigned nearly all devices to a "hotpluggable"
legacy PCI slot even on machines with a PCIe root bus (and even though
most such machines don't even support hotplug on legacy PCI slots!)
Forcing all devices onto legacy PCI slots means that the domain will
need a dmi-to-pci-bridge (to convert from PCIe to legacy PCI) and a
pci-bridge (to provide hotpluggable legacy PCI slots which, again,
usually aren't hotpluggable anyway).
To help reduce the need for these legacy controllers, this patch tries
to assign virtio-1.0-capable devices to PCIe slots whenever possible,
by setting appropriate connectFlags in
virDomainCalculateDevicePCIConnectFlags(). Happily, when that function
was written (just a few commits ago) it was created with a
"virtioFlags" argument, set by both of its callers, which is the
proper connectFlags to set for any virtio-*-pci device - depending on
the arch/machinetype of the domain, and whether or not the qemu binary
supports virtio-1.0, that flag will have either been set to PCI or
PCIe. This patch merely enables the functionality by setting the flags
for the device to whatever is in virtioFlags if the device is a
virtio-*-pci device.
NB: the first virtio video device will be placed directly on bus 0
slot 1 rather than on a pcie-root-port due to the override for primary
video devices in qemuDomainValidateDevicePCISlotsQ35(). Whether or not
to change that is a topic of discussion, but this patch doesn't change
that particular behavior.
NB2: since the slot must be hotpluggable, and pcie-root (the PCIe root
complex) does *not* support hotplug, this means that suitable
controllers must also be in the config (i.e. either pcie-root-port, or
pcie-downstream-port). For now, libvirt doesn't add those
automatically, so if you put virtio devices in a config for a qemu
that has PCIe-capable virtio devices, you'll need to add extra
pcie-root-ports yourself. That requirement will be eliminated in a
future patch, but for now, it's simple to do this:
<controller type='pci' model='pcie-root-port'/>
<controller type='pci' model='pcie-root-port'/>
<controller type='pci' model='pcie-root-port'/>
...
Partially Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1330024
This patch cleans up the connect flags for certain types/models of
devices that aren't PCI to return 0. In the future that may be used as
an indicator to the caller about whether or not a device needs a PCI
address. For now it's just ignored, except for in
virDomainPCIAddressEnsureAddr() - called during device hotplug - (and
in some cases actually needs to be re-set to PCI|HOTPLUGGABLE just in
case someone (in some old config) has manually set a PCI address for a
device that isn't PCI.
Before now, all the qemu hotplug functions assumed that all devices to
be hotplugged were legacy PCI endpoint devices
(VIR_PCI_CONNECT_TYPE_PCI_DEVICE). This worked out "okay", because all
devices *are* legacy PCI endpoint devices on x86/440fx machinetypes,
and hotplug didn't work properly on machinetypes using PCIe anyway
(hotplugging onto a legacy PCI slot doesn't work, and until commit
b87703cf any attempt to manually specify a PCIe address for a
hotplugged device would be erroneously rejected).
This patch makes all qemu hotplug operations honor the pciConnectFlags
set by the single all-knowing function
qemuDomainDeviceCalculatePCIConnectFlags(). This is done in 3 steps,
but in a single commit since we would have to touch the other points
at each step anyway:
1) add a flags argument to the hypervisor-agnostic
virDomainPCIAddressEnsureAddr() (previously it hardcoded
..._PCI_DEVICE)
2) add a new qemu-specific function qemuDomainEnsurePCIAddress() which
gets the correct pciConnectFlags for the device from
qemuDomainDeviceConnectFlags(), then calls
virDomainPCIAddressEnsureAddr().
3) in qemu_hotplug.c replace all calls to
virDomainPCIAddressEnsureAddr() with calls to
qemuDomainEnsurePCIAddress()
So in effect, we're putting a "shim" on top of all calls to
virDomainPCIAddressEnsureAddr() that sets the right pciConnectFlags.
Set pciConnectFlags in each device's DeviceInfo and then use those
flags later when validating existing addresses in
qemuDomainCollectPCIAddress() and when assigning new addresses with
qemuDomainPCIAddressReserveNextAddr() (rather than scattering the
logic about which devices need which type of slot all over the place).
Note that the exact flags set by
qemuDomainDeviceCalculatePCIConnectFlags() are different from the
flags previously set manually in qemuDomainCollectPCIAddress(), but
this doesn't matter because all validation of addresses in that case
ignores the setting of the HOTPLUGGABLE flag, and treats PCIE_DEVICE
and PCI_DEVICE the same (this lax checking was done on purpose,
because there are some things that we want to allow the user to
specify manually, e.g. assigning a PCIe device to a PCI slot, that we
*don't* ever want libvirt to do automatically. The flag settings that
we *really* want to match are 1) the old flag settings in
qemuDomainAssignDevicePCISlots() (which is HOTPLUGGABLE | PCI_DEVICE
for everything except PCI controllers) and 2) the new flag settings
done by qemuDomainDeviceCalculatePCIConnectFlags() (which are
currently exactly that - HOTPLUGGABLE | PCI_DEVICE for everything
except PCI controllers).
The lowest level function of this trio
(qemuDomainDeviceCalculatePCIConnectFlags()) aims to be the single
authority for the virDomainPCIConnectFlags to use for any given device
using a particular arch/machinetype/qemu-binary.
qemuDomainFillDevicePCIConnectFlags() sets info->pciConnectFlags in a
single device (unless it has no virDomainDeviceInfo, in which case
it's a NOP).
qemuDomainFillAllPCIConnectFlags() sets info->pciConnectFlags in all
devices that have a virDomainDeviceInfo
The latter two functions aren't called anywhere yet. This commit is
just making them available. Later patches will replace all the current
hodge-podge of flag settings with calls to this single authority.
dom xml generated on begin step should be passed
to perform step in VIR_MIGRATE_PARAM_DEST_XML parameter.
Otherwise 'XML error: failed to parse xml document' is
raised on destination host as dom xml is NULL.
Signed-off-by: Pavel Glushchak <pglushchak@virtuozzo.com>
Coverity identified that this variable might be leaked. And it's
right. If an error occurred and we have to roll back the control
jumps to try_remove label where we save the current error (see
0e82fa4c34 for more info). However, inside the code a jump onto
other label is possible thus leaking the error object.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Use the newly introduced close callback helpers to make the code look just a
bit cleaner and more importantly, to fix the following memleak regarding a
dangling virAdmConnect object reference caused by assigning NULL to the close
callback data once the catch-disconnect routine used the callback followed
by a comparison of NULL to the originally defined close callback (which at that
moment had already been NULL'd by remoteAdminClientCloseFunc) in
virAdmConnectCloseCallbackUnregister.
717 (88 direct, 629 indirect) bytes in 1 blocks are definitely lost record
110 of 141
at 0x4C2A988: calloc (vg_replace_malloc.c:711)
by 0x530696F: virAllocVar (viralloc.c:560)
by 0x53689E6: virObjectNew (virobject.c:193)
by 0x5368B5E: virObjectLockableNew (virobject.c:219)
by 0x4E3E7EE: virAdmConnectNew (datatypes.c:900)
by 0x4E398BB: virAdmConnectOpen (libvirt-admin.c:220)
by 0x10D3E3: vshAdmConnect (virt-admin.c:161)
by 0x10D624: vshAdmReconnect (virt-admin.c:215)
by 0x10DB0A: cmdConnect (virt-admin.c:353)
by 0x11288F: vshCommandRun (vsh.c:1313)
by 0x10FDB6: main (virt-admin.c:1439)
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1357358
Signed-off-by: Erik Skultety <eskultet@redhat.com>
Well, there were three different spots where closeCallback->freeCallback was
called, not looking the same --> potential for bugs - and there indeed is a bug
with refcounting of the @conn object. So this patch partially follows the path
set by commit 24dbb69f by introducing some close callback helpers both to
replace all the spots where we call clean the close callback data with a
dedicated function and to be able to fix the refcounting bug causing a memleak.
Signed-off-by: Erik Skultety <eskultet@redhat.com>
The only place we change the @conn object is actually virAdmConnectOpen
routine, thus at the moment we don't really need to lock it, given the fact that
what we're trying to do here is to change the closeCallback object which is a
lockable object itself, so that should be enough to avoid races.
Signed-off-by: Erik Skultety <eskultet@redhat.com>
As was suggested in an earlier review comment[1], we can
catch some additional code points by cleaning up how we use the
hostdev subsystem type in some switch statements.
[1] End of https://www.redhat.com/archives/libvir-list/2016-September/msg00399.html
Signed-off-by: Eric Farman <farman@linux.vnet.ibm.com>
Signed-off-by: John Ferlan <jferlan@redhat.com>
The memory device alias needs to be treated as machine ABI as qemu is
using it in the migration stream for section labels. To simplify this
generate the alias from the slot number unless an existing broken
configuration is detected.
With this patch the aliases are predictable and even certain
configurations which would not be migratable previously are fixed.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1359135
As with other devices assign the slot number right away when adding the
device. This will make the slot numbers static as we do with other
addressing elements and it will ultimately simplify allocation of the
alias in a static way which does not break with qemu.
Detect on reconnect to a running qemu VM whether the alias of a
hotpluggable memory device (dimm) does not match the dimm slot number
where it's connected to. This is necessary as qemu is actually
considering the alias as machine ABI used to connect the backend object
to the dimm device.
This will require us to keep them consistent so that we can reliably
restore them on migration. In some situations it was currently possible
to create a mismatched configuration and qemu would refuse to restore
the migration stream.
To avoid breaking existing VMs we'll need to keep the old algorithm
though.
Simplify handling of the 'dimm' address element by allowing to specify
the slot number only. This will allow libvirt to allocate slot numbers
before starting qemu.
We dropped support for RHEL-5 vintage Xen a while ago,
but forgot to remove some of the hacks for it.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
https://bugzilla.redhat.com/show_bug.cgi?id=1386976
We have everything ready. Actually the only limitation was our
check that denied hotplug of vhost-user.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
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>
Even though using /dev/shm/asdf as the backend, we still need to make
the mapping shared. The original patch forgot to add that parameter.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1392031
Signed-off-by: Martin Kletzander <mkletzan@redhat.com>
../../src/qemu/qemu_capabilities.c:3757: error: declaration of
'basename' shadows a global declaration [-Wshadow]
Signed-off-by: Pavel Hrdina <phrdina@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>
Propagate the selected or default level to qemu if it's supported.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1376009
Signed-off-by: Prasanna Kumar Kalever <prasanna.kalever@redhat.com>
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
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:
0 - None
1 - Emergency
2 - Alert
3 - Critical
4 - Error
5 - Warning
6 - Notice
7 - Info
8 - Debug
9 - Trace
Signed-off-by: Prasanna Kumar Kalever <prasanna.kalever@redhat.com>
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
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.
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>
https://bugzilla.redhat.com/show_bug.cgi?id=1379196
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'.
Fixes the behavior when destroying a domain more than once.
VIR_ERR_OPERATION_INVALID should be raised when destroying an
already destroyed domain.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Commit c29e6d4805 cause build failure on RHEL-6:
../../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]
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
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.
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
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.
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
This is needed in order to migrate a domain with shmem devices as that
is not allowed to migrate.
Signed-off-by: Martin Kletzander <mkletzan@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:
http://git.qemu.org/?p=qemu.git;a=commit;h=5400c02b90bb
Signed-off-by: Martin Kletzander <mkletzan@redhat.com>
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:
http://git.qemu.org/?p=qemu.git;a=commit;h=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>
The old ivshmem is deprecated in QEMU, so let's use the better
ivshmem-{plain,doorbell} variants instead.
Signed-off-by: Martin Kletzander <mkletzan@redhat.com>
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.
https://bugzilla.redhat.com/show_bug.cgi?id=1374718
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
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.
Reported here:
https://www.redhat.com/archives/libvirt-users/2016-October/msg00032.htmlhttps://www.redhat.com/archives/libvirt-users/2016-October/msg00045.html
If we failed to unlink old dom cfg file, we goto rollback.
But inside rollback, we fogot to unlink the new dom cfg file.
This patch fixes this issue.
Signed-off-by: Chen Hanxiao <chenhanxiao@gmail.com>
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
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
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.
Signed-off-by: Maxim Nestratov <mnestratov@virtuozzo.com>
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.
Signed-off-by: Chen Hanxiao <chenhanxiao@gmail.com>
When starting a new domain, we allocate the USB addresses and keep
an address cache in the domain object's private data.
However this data is lost on libvirtd restart.
Also generate the address cache if all the addresses have been
specified, so that devices hotplugged after libvirtd restart
also get theirs assigned.
https://bugzilla.redhat.com/show_bug.cgi?id=1387666
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.
This time do not require an address cache as a parameter.
Simplify qemuDomainAttachChrDeviceAssignAddr to not generate
the virtio serial address cache for devices of other types.
Partially reverts commit 925fa4b.
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.
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.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1009008
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.
https://bugzilla.redhat.com/show_bug.cgi?id=1300776
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.
Signed-off-by: John Ferlan <jferlan@redhat.com>
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.
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.
Signed-off-by: John Ferlan <jferlan@redhat.com>
Add in the block I/O throttling length/duration parameter to the command
line if supported. If not supported, fail command creation.
Add the xml2argvtest for testing.
Add support for a duration/length for the bps/iops and friends.
Modify the API in order to add the "blkdeviotune." specific definitions
for the iotune throttling duration/length options
total_bytes_sec_max_length
write_bytes_sec_max_length
read_bytes_sec_max_length
total_iops_sec_max_length
write_iops_sec_max_length
read_iops_sec_max_length
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.
This patch will also adjust the qemuMonitorJSONSetBlockIoThrottle error
procession so that rather than returning/displaying:
"error: internal error: Unexpected error"
Fetch the actual error message from qemu and display that
Create a macros to hide all the comparisons for each of the fields.
Add a 'continue;' for a compiler hint that we only need to find one
this should be similar enough to the if - elseif - elseif logic.
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)
Signed-off-by: John Ferlan <jferlan@redhat.com>
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.
Introduced by commit 68c70118.
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.
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>
Extended the qemuMonitorCPUInfo with a halted flag. Extract the halted
flag for both text and JSON monitor.
Signed-off-by: Viktor Mihajlovski <mihajlov@linux.vnet.ibm.com>
Signed-off-by: Boris Fiuczynski <fiuczy@linux.vnet.ibm.com>
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.
There is an existing virDomainPCIAddressReserveNextSlot() which will
reserve all functions of the next available PCI slot. One place in the
qemu PCI address assignment code requires reserving a *single*
function of the next available PCI slot. This patch modifies and
renames virDomainPCIAddressReserveNextSlot() so that it can fulfill
both the original purpose and the need to reserve a single function.
(This is being done so that the abovementioned code in qemu can have
its "kind of open coded" solution replaced with a call to this new
function).
Since TLS was introduced hostwide for libvirt 2.3.0 and a domain
configurable haveTLS was implemented for libvirt 2.4.0, we have to
modify the migratable XML for specific case where the 'tls' attribute
is based on setting from qemu.conf.
The "tlsFromConfig" is libvirt internal attribute and is stored only in
status XML to ensure that when libvirtd is restarted this internal flag
is not lost by the restart.
That flag is used to decide whether we should put *tls* attribute to
migratable XML or not.
Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
Add an optional "tls='yes|no'" attribute for a TCP chardev.
For QEMU, this will allow for disabling the host config setting of the
'chardev_tls' for a domain chardev channel by setting the value to "no" or
to attempt to use a host TLS environment when setting the value to "yes"
when the host config 'chardev_tls' setting is disabled, but a TLS environment
is configured via either the host config 'chardev_tls_x509_cert_dir' or
'default_tls_x509_cert_dir'
Signed-off-by: John Ferlan <jferlan@redhat.com>
Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
Currently the union has only one member so remove that union. If there
is a need to add a new type of source for new bus in the future this
will force the author to add a union and properly check bus type before
any access to union member.
Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
Commit id '2c322378' missed the nuance that the rng backend could be
using a TCP chardev and if TLS is enabled on the host, thus will need
to have the TLS object added.
Commit id '2c322378' missed the nuance that the redirdev backend could
be using a TCP chardev and if TLS is enabled on the host, thus will need
to have the TLS object added.
Rather than VIR_ALLOC() the data, use virDomainChrSourceDefNew in order
to get the private data if necessary.
Signed-off-by: John Ferlan <jferlan@redhat.com>
Use a pointer and the virDomainChrSourceDefNew() function in order to
allocate the structure for _virDomainRedirdevDef.
Signed-off-by: John Ferlan <jferlan@redhat.com>
Use a pointer and the virDomainChrSourceDefNew() function in order to
allocate the structure for _virDomainSmartcardDef.
Signed-off-by: John Ferlan <jferlan@redhat.com>
instead of:
virBufferAdd(buf, "arg1,");
virBufferAdd(buf, "arg2");
lets have:
virBufferAdd(buf, "arg1");
virBufferAdd(buf, ",arg2");
Because it's better. Consider we want to add conditionally arg3.
With this change, it's simple:
if (cond)
virBufferAdd(buf, ",arg3");
with current code there might be a comma hanging at EOL.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
from virDomainDefPtr to virDomainObjPtr so that the function has
access to the other parts of the virDomainObjPtr. Take advantage of
this by removing the "priv" arg and retrieving it from the
virDomainObjPtr instead.
No functional change.
For some reason the values of memballoon model are set using an
anonymous enum, making it impossible to perform nice tricks like
demanding there are cases for all possible values in a switch. This
patch turns the anonymous enum into virDomainMemballoonModel.
More occurences of repeatedly dereferencing the same pointer stored in
an array are replaced with the definition of a temporary pointer that
is then used directly. No functional change.
Commit id '5f2a132786' should have placed the data in the host source
def structure since that's also used by smartcard, redirdev, and rng in
order to provide a backend tcp channel. The data in the private structure
will be necessary in order to provide the secret properly.
This also renames the previous names from "Chardev" to "ChrSource" for
the private data structures and API's
Change the virDomainChrDef to use a pointer to 'source' and allocate
that pointer during virDomainChrDefNew.
This has tremendous "fallout" in the rest of the code which mainly
has to change source.$field to source->$field.
Signed-off-by: John Ferlan <jferlan@redhat.com>
When hotplugging networks with ancient QEMUs not supporting
QEMU_CAPS_NETDEV, we use space instead of a comma as the separator
between the network type and other options.
Except for "user", all the network types pass other options
and use up the first separator by the time we get to the section
that adds the alias (or vlan for QEMUs without CAPS_NETDEV).
Since the alias/vlan is mandatory, convert all preceding code to add
the separator at the end, removing the need to rewrite type_sep for
all types but NET_TYPE_USER.
Absent driver name attribute is invalid xml. Which in turn makes
unusable 'virsh edit' for example. The value does not make
much sense and ignored on input so nobody will hurt.
vz sdk supports setting serial number only for disk devices.
Getting serial upon cdrom(for example) is error however
setting is just ignored. Let's check for disk device
explicitly for clarity in both cases.
Setting serial number for other devices is ignored
with an info note just as before.
We need usual conversion from "" to NULL in direction
vz sdk -> libvirt, because "" is not valid for libvirt
and "" means unspecifiend in vz sdk which is NULL for libvirt.
New line character in name of network is now forbidden because it
mess virsh output and can be confusing for users. Validation of
name is done in network driver, after parsing XML to avoid
problems with disappeared network which was already created with
new-line char in name.
Closes-Bug: https://bugzilla.redhat.com/show_bug.cgi?id=818064
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
New util function virXMLCheckIllegalChars is now used to test if
parsed network contains illegal char '/' in it's name.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
This new function can be used to check if e.g. name of XML
node don't contains forbidden chars like "/" or "\n".
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Modeled after the qemuDomainHostdevPrivatePtr (commit id '27726d8c'),
create a privateData pointer in the _virDomainChardevDef to allow storage
of private data for a hypervisor in order to at least temporarily store
secret data for usage during qemuBuildCommandLine.
NB: Since the qemu_parse_command (qemuParseCommandLine) code is not
expecting to restore the secret data, there's no need to add code
code to handle this new structure there.
Signed-off-by: John Ferlan <jferlan@redhat.com>
Add a new qemu.conf variables to store the UUID for the secret that could
be used to present credentials to access the TLS chardev. Since this will
be a server level and it's possible to use some sort of default, introduce
both the default and chardev logic at the same time making the setting of
the chardev check for it's own value, then if not present checking whether
the default value had been set.
Signed-off-by: John Ferlan <jferlan@redhat.com>
When converting a domain xml containing a CDROM device without
any attached source, don't add a target=(null) to the libxl config
disk definition: xen doesn't like it at all and would fail to start
the domain.
There was inconsistency between alias used to create tls-creds-x509
object and alias used to link that object to chardev while hotpluging.
Hotplug ends with this error:
error: Failed to detach device from channel-tcp.xml
error: internal error: unable to execute QEMU command 'chardev-add':
No TLS credentials with id 'objcharchannel3_tls0'
In XML we have for example alias "serial0", but on qemu command line we
generate "charserial0".
The issue was that code, that creates QMP command to hotplug chardev
devices uses only the second alias "charserial0" and that alias is also
used to link the tls-creds-x509 object.
This patch unifies the aliases for tls-creds-x509 to be always generated
from "charserial0".
Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
Instead of typing the prefix every time we want to append parameters
to qemu command line use a variable that contains prefixed alias.
Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
We need to make sure that the chardev is TCP. Without this check we
may access different part of union and corrupt pointers.
Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
The code is entirely correct, but it still managed to trip me
up when I first ran into it because I did not realize right away
that VIR_PCI_CONNECT_TYPES_ENDPOINT was not a single flag, but
rather a mask including both VIR_PCI_CONNECT_TYPE_PCI_DEVICE and
VIR_PCI_CONNECT_TYPE_PCIE_DEVICE.
In order to save the next distracted traveler in PCI Address Land
some time, document this fact with a comment. Add a test case for
the behavior as well.
A pci-bridge has *almost* the same rules as a legacy PCI endpoint
device for where it can be automatically connected, and until now both
had been considered identical. There is one pairing that is okay when
specifically requested by the user (i.e. manual assignment), but we
want to avoid it when auto-assigning addresses - plugging a pci-bridge
directly into pcie-root (it is cleaner to plug in a dmi-to-pci-bridge,
then plug the pci-bridge into that).
In order to allow that difference, this patch makes a separate
CONNECT_TYPE for pci-bridge, and uses it to restrict auto-assigned
addresses for pci-bridges to be only on pci-root, pci-expander-bus,
dmi-to-pci-bridge, or on another pci-bridge.
NB: As with other discouraged-but-seem-to-work configurations
(e.g. plugging a legacy PCI device into a pcie-root-port) if someone
*really* wants to, they can still force a pci-bridge to be plugged
into pcie-root (by manually specifying its PCI address.)
https://bugzilla.redhat.com/show_bug.cgi?id=1357416
Rather than return a 0 or -1 and the *result string, return just the result
string to the caller. Alter all the callers to handle the different return.
As a side effect or result of this, it's much clearer that we cannot just
assign the returned string into the scsi_host wwnn, wwpn, and fabric_wwn
fields - rather we should fetch a temporary string, then as long as our
fetch was good, VIR_FREE what may have been there, and STEAL what we just got.
This fixes a memory leak in the virNodeDeviceCreateXML code path through
find_new_device and nodeDeviceLookupSCSIHostByWWN which will continually
call nodeDeviceSysfsGetSCSIHostCaps until the expected wwnn/wwpn is found
in the device object capabilities.
Signed-off-by: John Ferlan <jferlan@redhat.com>
https://bugzilla.redhat.com/show_bug.cgi?id=1366108
There are couple of things that needs to be done in order to
allow vhost-user hotplug. Firstly, vhost-user requires a chardev
which is connected to vhost-user bridge and through which qemu
communicates with the bridge (no acutal guest traffic is sent
through there, just some metadata). In order to generate proper
chardev alias, we must assign device alias way sooner.
Then, because we are plugging the chardev first, we need to do
the proper undo if something fails - that is remove netdev too.
We don't want anything to be left over in case attach fails at
some point.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
https://bugzilla.redhat.com/show_bug.cgi?id=1366505
So far, this function lacked support for
VIR_DOMAIN_NET_TYPE_VHOSTUSER leaving callers to hack around the
problem by constructing the command line on their own. This is
not ideal as it blocks hot plug support.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Currently, what we do for vhost-user network is generate the
following part of command line:
-netdev type=vhost-user,id=hostnet0,chardev=charnet0
There's no need for 'type=' it is the default. Drop it.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
There's no need to reinvent the wheel here. We already have a
function to format virDomainChrSourceDefPtr. It's called
qemuBuildChrChardevStr(). Use that instead of some dummy
virBufferAsprintf().
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
This alone makes not much sense. But the aim is to reuse this
function in qemuBuildVhostuserCommandLine() where 'nowait' is not
supported for vhost-user devices.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
We tend to prevent using 'default' in switches. And it is for a
good reason - control may end up in paths we wouldn't want for
new values. In this specific case, if qemuBuildHostNetStr is
called over VIR_DOMAIN_NET_TYPE_VHOSTUSER it would produce
meaningless output. Fortunately, there no such call yet.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Instead of blindly claim support for hot-plugging of every
interface type out there we should copy approach we have for
device types: white listing supported types and explicitly error
out on unsupported ones.
For instance, trying to hotplug vhostuser interface results in
nothing usable from guest currently. vhostuser typed interfaces
require additional work on our side.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
The idea is to have function that does some checking at its
beginning and then have one big switch for all the interface
types it supports.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
The idea is to have function that does some checking of the
arguments at its beginning and then have one big switch for all
the interface types it supports. Each one of them generating the
corresponding part of the command line.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
The idea is to have function that does some checking of the
arguments at its beginning and then have one big switch for all
the interface types it supports. Each one of them generating the
corresponding part of the command line.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
This function for some weird reason returns integer instead of
virDomainNetType type. It is important to return the correct type
so that we know what values we can expect.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Not every system out there has syslog, that's why we check for it
in our configure script. However, in 640b58abdf while fixing
another issue, some variables and functions are called that are
defined only when syslog.h is present. But these function
calls/variables were not guarded by #ifdef-s.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
qemuBuildSmbiosBiosStr and qemuBuildSmbiosSystemStr return NULL if
there's nothing to format on the commandline. Reporting errors from
buffer creation doesn't make sense since it would be ignored.
This initially started as a fix of some debug printing in
virCgroupDetect. However it turned out that other places suffer
from the similar problem. While dealing with pids, esp. in cases
where we cannot use pid_t for ABI stability reasons, we often
chose an unsigned integer type. This makes no sense as pid_t is
signed.
Also, new syntax-check rule is introduced so we won't repeat this
mistake.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
There are two video devices with models without VGA compatibility mode.
They are primary used as secondary video devices, but in some cases it
is required to use them also as primary video devices.
Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
This improves commit 706b5b6277 in a way that we check qemu capabilities
instead of what architecture we are running on to detect whether we can
use *virtio-vga* model or not. This is not a case only for arm/aarch64.
Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
Commit 21373feb added support for primary virtio-vga device but it was
checking for virtio-gpu. Let's check for existence of virtio-vga if we
want to use it.
Virtio video device is currently represented by three different models
*virtio-gpu-device*, *virtio-gpu-pci* and *virtio-vga*. The first two
models are tied together and if virtio video devices is compiled in they
both exist. However, the *virtio-vga* model doesn't have to exist on
some architectures even if the first two models exist. So we cannot
group all three together.
Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
Before this patch we've checked qemu capabilities for video devices
only while constructing qemu command line using "-device" option.
Since we support qemu only if "-device" option is present we can use
the same capabilities to check also video devices while using "-vga"
option to construct qemu command line.
Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
All definition validation that doesn't depend on qemu capabilities
and was allowed previously as valid definition should be placed into
qemuDomainDefValidate.
The check whether video type is supported or not was based on an enum
that translates type into model. Use switch to ensure that if new
video type is added, it will be properly handled.
Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
We generally uses QEMU_CAPS_DEVICE_$NAME to probe for existence of some
device and QEMU_CAPS_$NAME_$PROP to probe for existence of some property
of that device.
Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
If QEMU in question supports QMP, this capability is set if
QEMU_CAPS_DEVICE_QXL was set based on existence of "-device qxl". If
libvirt needs to parse *help*, because there is no QMP support, it
checks for existence of "-vga qxl", but it also parses output of
"-device ?" and sets QEMU_CAPS_DEVICE_QXL too.
Now that libvirt supports only QEMU that has "-device" implemented it's
safe to drop this capability and stop using it.
Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
This patch simplifies QEMU capabilities for QXL video device. QEMU
exposes this device as *qxl-vga* and *qxl* and they are both the same
device with the same set of parameters, the only difference is that
*qxl-vga* includes VGA compatibility.
Based on QEMU code they are tied together so it's safe to check only for
presence of only one of them.
This patch also removes an invalid test case "video-qxl-sec-nodevice"
where there is only *qxl-vga* device and *qxl* device is not present.
Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
Qemu supports *xen* video device only with XEN and this code was part
of xenner code. We dropped support for xenner in commit de9be0a.
Before this patch if you used 'xen' video type you ended up with
domain without any video device at all. Now we don't allow to start
such domain.
Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
It was never safe anyway and as such shouldn't have been enabled in the
first place. Future patches will allow hot-(un)pluging of some ivshmem
devices as a workaround.
Signed-off-by: Martin Kletzander <mkletzan@redhat.com>
Due to the switch of parameters in a call to virDomainShmemDefEquals()
no device was found when looking for device with all the information
except address. Also fix the indentation.
Signed-off-by: Martin Kletzander <mkletzan@redhat.com>
If the last event callback is unregistered while the event loop is
dispatching, it is only marked as deleted, but not removed. The number
of callbacks is more than zero in that case, so the timer is not
removed. Because it can be removed in this function now (but also
accessed afterwards so that we set 'isDispatching = false' and have it
locked), we need to temporarily increase the reference counter of the
state for the duration of this function.
Signed-off-by: Martin Kletzander <mkletzan@redhat.com>
There is a repeating pattern of code that removes the timer if it's not
needed. So let's move it to a new function. We'll also use it later.
Signed-off-by: Martin Kletzander <mkletzan@redhat.com>
There should be one more reference because it is being kept in the list
of callbacks as an opaque. We also unref it properly using
virObjectFreeCallback.
Signed-off-by: Martin Kletzander <mkletzan@redhat.com>
Currently Libvirt allows attempts to migrate read only disks. Qemu
cannot handle this as read only disks cannot be written to on the
destination system. The end result is a cryptic error message and a
failed migration.
This patch causes migration to fail earlier and provides a meaningful
error message stating that migrating read only disks is not supported.
Signed-off-by: Corey S. McQuay <csmcquay@linux.vnet.ibm.com>
Reviewed-by: Jason J. Herne <jjherne@linux.vnet.ibm.com>
Reviewed-by: Boris Fiuczynski <fiuczy@linux.vnet.ibm.com>
Make sure that the topology results into a sane number of cpus (up to
UINT_MAX) so that it can be sanely compared to the vcpu count of the VM.
Additionally the helper added in this patch allows to fetch the total
number the topology results to so that it does not have to be
reimplemented later.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1378290
In both virLogParseOutput and virLogParseFilter, rather than returning
NULL, goto cleanup since it's possible that for each the first condition
passes, but the || condition doesn't and thus we leak memory.
The dnsmasq man page recommends that dhcp-authoritative "should be
set when dnsmasq is definitely the only DHCP server on a network".
This is the case for libvirt-managed virtual networks.
The effect of this is that VMs that fail to renew their DHCP lease
in time (e.g. if the VM or host is suspended) will be able to
re-acquire the lease even if it's expired, unless the IP address has
been taken by some other host. This avoids various annoyances caused
by changing VM IP addresses.
Sometimes virObjectEventStateFlush can be called without timer (if the
last event was unregistered right when the timer fired). There is a
check for timer == -1, but that triggers warning and other log messages,
which is unnecessary.
Signed-off-by: Martin Kletzander <mkletzan@redhat.com>
Handling of outputs and filters has been changed in a way that splits
parsing and defining. Do the same thing for logging priority as well, this
however, doesn't need much of a preparation.
Signed-off-by: Erik Skultety <eskultet@redhat.com>
This is mainly virLogAddOutputTo* which were replaced by virLogNewOutputTo* and
the previously poorly named ones virLogParseAndDefine* functions. All of these
are unnecessary now, since all the original callers were transparently switched
to the new model of separate parsing and defining logic.
Signed-off-by: Erik Skultety <eskultet@redhat.com>
Similar to outputs, parser should do parsing only, thus the 'define' logic
is going to be stripped from virLogParseAndDefineFilters by replacing calls to
this method to virLogSetFilters instead.
Signed-off-by: Erik Skultety <eskultet@redhat.com>
Since virLogParseAndDefineOutputs is going to be stripped from 'output defining'
logic, replace all relevant occurrences with virLogSetOutputs call to make the
change transparent to all original callers (daemons mostly).
Signed-off-by: Erik Skultety <eskultet@redhat.com>
This method will eventually replace virLogParseAndDefineFilters which
currently does both parsing and defining.
Signed-off-by: Erik Skultety <eskultet@redhat.com>
This API is the entry point to output modification of the logger. Currently,
everything is done by virLogParseAndDefineOutputs. Parsing and defining will be
split into two operations both handled by this method transparently.
Signed-off-by: Erik Skultety <eskultet@redhat.com>
Abstraction added over parsing a single filter. The method parses potentially a
set of logging filters, while adding each filter logging object to a
caller-provided array.
Signed-off-by: Erik Skultety <eskultet@redhat.com>
Another abstraction added on the top of parsing a single logging output. This
method takes and parses the whole set of outputs, adding each single output
that has already been parsed into a caller-provided array. If the user-supplied
string contained duplicate outputs, only the last occurrence is taken into
account (all the others are removed from the list), so we silently avoid
duplicate logs.
Signed-off-by: Erik Skultety <eskultet@redhat.com>
Same as for outputs, introduce a new method, that is basically the same as
virLogParseAndDefineFilter with the difference that it does not define the
filter. It rather returns a newly created object that needs to be inserted into
a list and then defined separately.
Signed-off-by: Erik Skultety <eskultet@redhat.com>
Introduce a method to parse an individual logging output. The difference
compared to the virLogParseAndDefineOutput is that this method does not define
the output, instead it makes use of the virLogNewOutputTo* methods introduced
in the previous patch and just returns the virLogOutput object that has to be
added to a list of object which then can be defined as a whole via
virLogDefineOutputs. The idea remains still the same - split parsing and
defining of the logging primitives (outputs, filters).
Additionally, since virLogNewOutputTo* methods are now finally used,
ATTRIBUTE_UNUSED can be successfully removed from the methods' definitions,
since that was just to avoid compiler complaints about unused static functions.
Signed-off-by: Erik Skultety <eskultet@redhat.com>
Now that we're in the critical section, syslog connection can be re-opened
by issuing openlog, which is something that cannot be done beforehand, since
syslog keeps its file descriptor private and changing the tag earlier might
introduce a log inconsistency if something went wrong with preparing a new set
of logging outputs in order to replace the existing one.
Signed-off-by: Erik Skultety <eskultet@redhat.com>
Continuing with the effort to split output parsing and defining, these new
functions return a logging object reference instead of defining the output.
Eventually, these functions will replace the existing ones (virLogAddOutputTo*)
which will then be dropped.
Signed-off-by: Erik Skultety <eskultet@redhat.com>
Prepare a method that only defines a set of filters. It takes a list of
filters, preferably created by virLogParseFilters. The original set of filters
is reset and replaced by the new user-provided set of filters.
Signed-off-by: Erik Skultety <eskultet@redhat.com>
Prepare a method that only defines a set of outputs. It takes a list of
outputs, preferably created by virLogParseOutputs. The original set of outputs
is reset and replaced by the new user-provided set of outputs.
Signed-off-by: Erik Skultety <eskultet@redhat.com>
Outputs are a bit trickier than filters, since the user(config)-specified
set of outputs can contain duplicates. That would lead to logging the same
message twice. For compatibility reasons, we cannot just error out and forbid
the daemon to start if we find duplicate outputs which do not make sense.
Instead, we could silently take into account only the last occurrence of the
duplicate output and remove all the previous ones, so that the logger will not
try to use them when it is looping over all of its registered outputs.
Signed-off-by: Erik Skultety <eskultet@redhat.com>
In order to later split output parsing and output defining, introduce a new
function which will create a new virLogOutput object which the parser will
insert into a list with the list being eventually defined.
Signed-off-by: Erik Skultety <eskultet@redhat.com>
There is really no reason why we could not keep journald's fd within the
journald output object the same way as we do for regular file-based outputs.
By doing this we later won't have to special case the journald-based output
(due to the fd being globally shared) when replacing the existing set of outputs
with a new one. Additionally, by making this change, we don't need the
virLogCloseJournald routine anymore, plain virLogCloseFd will suffice.
Signed-off-by: Erik Skultety <eskultet@redhat.com>
Right now virLogParse* functions are doing both parsing and defining of filters
and outputs which should be two separate operations. Since the naming is
apparently a bit poor this patch renames these functions to
virLogParseAndDefine* which eventually will be replaced by virLogSet*.
Additionally, virLogParse{Filter,Output} will be later (after the split) reused,
so that these functions do exactly what the their name suggests.
Signed-off-by: Erik Skultety <eskultet@redhat.com>
During first stage of virlog.c refactor, commit 0b231195 forgot to remove the
macro definition along with its usage.
Signed-off-by: Erik Skultety <eskultet@redhat.com>
If the event is already disabled, then don't bother with setting it
disabled again. Causes unnecessary error on systems that don't support
the feature anyway.
The intel-iommu device has existed since QEMU 2.2.0, but
it was only possible to create it with -device since
QEMU 2.7.0, thanks to:
commit 621d983a1f9051f4cfc3f402569b46b77d8449fc
Author: Marcel Apfelbaum <marcel@redhat.com>
Date: Mon Jun 27 18:38:34 2016 +0300
hw/iommu: enable iommu with -device
Use the standard '-device intel-iommu' to create the IOMMU device.
The legacy '-machine,iommu=on' can still be used.
The libvirt capability check & command line formatting code
is thus broken for all QEMU versions 2.2.0 -> 2.6.0 inclusive.
This fixes it to use iommu=on instead.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Since introduction of chardev hotplug the code was wrong for the UDP
case and basically created a TCP socket instead. Use proper objects and
type for UDP.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1377602
Rather than copy-paste - use a macro
Unfortunately due to how the RNG schema was written keeping the 'value'
and 'value'_max next to each other in the XML causes a schema failure,
so the FORMAT has to write out singly rather than optimizing to write
out both values at once
Signed-off-by: John Ferlan <jferlan@redhat.com>
We're about to add more options, let's avoid having multiple if-then-else
which each try to set up the qemuMonitorJSONMakeCommand call with all the
parameters it knows about.
Instead, use the fact that when a NULL is found in the argument list that
processing of the remaining arguments stops and just have call.
Signed-off-by: John Ferlan <jferlan@redhat.com>
We're about to add 6 new options and it appears (from testing) one cannot
utilize both the shorthand (alias) and (much) longer names for the arguments.
So modify the command builder to use the longer name and of course alter the
test output .args to have the similarly innocuous long name.
Also utilize a macro to build that name makes it so much more visually
appealing and saves a few characters or potential cut-n-paste issues.
Signed-off-by: John Ferlan <jferlan@redhat.com>
When I added support for the pcie-expander-bus controller in commit
bc07251f, I incorrectly thought that it only had a single slot
available. Actually it has 32 slots, just like the root complex aka
pcie-root (the part that I *did* get correct is that unlike pcie-root
a pcie-expander-bus doesn't allow any integrated endpoint devices -
only pcie-root-ports and dmi-to-pci-controllers are allowed).
Reduce some cut-n-paste code by creating common helper. Make use of the
recently added virJSONValueObjectStealArray to grab the devices list as
part of the common code (we we can Free the reply) and return devices for
each of the callers to continue to parse.
NB: This also adds error checking to qemuMonitorJSONDiskNameLookup
Signed-off-by: John Ferlan <jferlan@redhat.com>
Provide the Steal API for any code paths that will desire to grab the
object array and then free it afterwards rather than relying to freeing
the whole chain from the reply.
Rather than use stack allocated state context pointers, let's allocate and
free the state context pointer. In doing so, we'll shrink the code a bit
since many routines perform the same initialization sequence.
Signed-off-by: John Ferlan <jferlan@redhat.com>
Since none of the callers check the status, let's just alter it to
a static void.
While we're at it - scrap the local runtime variable and just do the
math in the VIR_DEBUG directly.
Signed-off-by: John Ferlan <jferlan@redhat.com>
If attaching to a qemu process fails after opening the monitor socket
libvirt does not clean up the monitor. As the monitor also holds a
reference to the domain object the qemu attach API basically leaks it.
QEMU also does not interact on a second monitor connection and thus a
further attempt to attach to it would lock up.
Prevent libvirt from leaking the monitor by explicitly closing it.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1378401
Attaching to a existing qemu process allows to get us into a situation
when qemu is new enough to have JSON monitor and new vCPU hotplug but
the json monitor is not used. The vCPU detection code would require it
though. This broke attaching to qemu processes.
Make the condition less strict and just skip the vCPU hotplug detection
if JSON monitor is not available.
Resolves one of the symptoms in:
https://bugzilla.redhat.com/show_bug.cgi?id=1378401
Libvirt, on its own, shouldn't decide whether an expired lease should
stay in the custom leases database or not. It should rather rely on
the 'DEL' event from dnsmasq.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
We are about to add 6 new values to fetch. This will put us over the
current limit of 16 (we're at 13 now).
Once there are more than 16 parameters, this will affect existing clients
that attempt to fetch blockiotune config values for the domain from the
remote host since the server side has no mechanism to determine whether
the capability for the emulator exists and thus would attempt to return
all known values from the persistentDef. If attempting to fetch the
blockiotune values from a running domain, the code will check the emulator
capabilities and set maxparams (in qemuDomainGetBlockIoTune) appropriately.
On the client side of the remote connection, it uses this constant in
xdr_remote_domain_get_block_io_tune_ret and virTypedParamsDeserialize
calls, so if a remote server returns more than 16 parameters, then the
client will fail with "Unable to decode message payload".
Signed-off-by: John Ferlan <jferlan@redhat.com>
The REMOTE_DOMAIN_MEMORY_PARAMETERS_MAX was erroneously used in the
remoteDomainBlockStatsFlags and remoteDomainGetBlockIoTune calls. Change
the constant to be the right one.
Fortunately, all 3 are defined as 16.
Signed-off-by: John Ferlan <jferlan@redhat.com>
This breaks vCPU hotplug, because when starting a domain, we
create a copy of domain definition (which becomes live XML) and
during the post parse callbacks we might adjust some tunings so
that vCPU hotplug is possible.
This reverts commit 581b7756af.
This breaks vCPU hotplug, because when starting a domain, we
create a copy of domain definition (which becomes live XML) and
during the post parse callbacks we might adjust some tunings so
that vCPU hotplug is possible.
This reverts commit c0f90799bc.
Certain operations may make the vcpu order information invalid. Since
the order is primarily used to ensure migration compatibility and has
basically no other user benefits, clear the order prior to certain
operations and document that it may be cleared.
All the operations that would clear the order can still be properly
executed by defining a new domain configuration rather than using the
helper APIs.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1370357
virDomainDefSetVcpus was not designed to handle coldplug of vcpus now
that we can set state of vcpus individually.
Introduce qemuDomainSetVcpusConfig that properly handles state changes
of vcpus when coldplugging so that invalid configurations are not
created.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1375939
The current code that validates duplicate vcpu order would not work
properly if the order would exceed def->maxvcpus. Limit the order to the
interval described.
The bitmap indexes for the order duplicate check are shifted to 0 since
vcpu order 0 is not allowed. The error message doesn't need such
treating though.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1370360
https://bugzilla.redhat.com/show_bug.cgi?id=1292984
Hold on to your hats, because this is gonna be wild.
In bd3e16a3 I've tried to expose sanlock io_timeout. What I had
not realized (because there is like no documentation for sanlock
at all) was very unusual way their APIs work. Basically, what we
do currently is:
sanlock_add_lockspace_timeout(&ls, io_timeout);
which adds a lockspace to sanlock daemon. One would expect that
io_timeout sets the io_timeout for it. Nah! That's where you are
completely off the tracks. It sets timeout for next lockspace you
will probably add later. Therefore:
sanlock_add_lockspace_timeout(&ls, io_timeout = 10);
/* adds new lockspace with default io_timeout */
sanlock_add_lockspace_timeout(&ls, io_timeout = 20);
/* adds new lockspace with io_timeout = 10 */
sanlock_add_lockspace_timeout(&ls, io_timeout = 40);
/* adds new lockspace with io_timeout = 20 */
And so on. You get the picture.
Fortunately, we don't allow setting io_timeout per domain or per
domain disk. So we just need to set the default used in the very
first step and hope for the best (as all the io_timeout-s used
later will have the same value).
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Currently, we are checking for sanlock_add_lockspace_timeout
which is good for now. But in a subsequent patch we are going to
use sanlock_write_lockspace (which sets an initial value for io
timeout for sanlock). Now, there is no reason to check for both
functions in sanlock library as the sanlock_write_lockspace was
introduced in 2.7 release and the one we are currently checking
for in the 2.5 release. Therefore it is safe to assume presence
of sanlock_add_lockspace_timeout when sanlock_write_lockspace
is detected.
Moreover, the macro for conditional compilation is renamed to
HAVE_SANLOCK_IO_TIMEOUT (as it now encapsulates two functions).
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
If this reminds you of a commit message from around a year ago, it's
41c2aa729f and yes, we're dealing with
"the same thing" again. Or f309db1f4d and
it's similar.
There is a logic in place that if there is no real need for
memory-backend-file, qemuBuildMemoryBackendStr() returns 0. However
that wasn't the case with hugepage backing. The reason for that was
that we abused the 'pagesize' variable for storing that information, but
we should rather have a separate one that specifies whether we really
need the new object for hugepage backing. And that variable should be
set only if this particular NUMA cell needs special treatment WRT
hugepages.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1372153
Signed-off-by: Martin Kletzander <mkletzan@redhat.com>
Due to a copy and paste error, the scheduler 'cap' parameter
was over-writing the 'weight' parameter when preparing the
return parameters in libxlDomainGetSchedulerParametersFlags.
As a result, the scheduler weight was never shown when getting
schedinfo and setting the weight failed as well
virsh schedinfo testvm
Scheduler : credit
cap : 0
virsh schedinfo testvm --cap 50 --weight 500
Scheduler : credit
error: invalid scheduler option: weight
The obvious fix is to assign the 'caps' parameter to the correct
item in the parameter list.
Reported-by: Volo M. <vm@vovs.net>
Add support for formating/parsing libxl channels.
Syntax on xen libxl goes as following:
channel=["connection=pty|socket,path=/path/to/socket,name=XXX",...]
Signed-off-by: Joao Martins <joao.m.martins@oracle.com>
Acked-by: Jim Fehlig <jfehlig@suse.com>
Signed-off-by: Jim Fehlig <jfehlig@suse.com>
And allow libxl to handle channel element which creates a Xen
console visible to the guest as a low-bandwitdh communication
channel. If type is PTY we also fetch the tty after boot using
libxl_channel_getinfo to fetch the tty path. On socket case,
we autogenerate a path if not specified in the XML. Path autogenerated
is slightly different from qemu driver: qemu stores also on
"channels/target" but it creates then a directory per domain with
each channel target name. libxl doesn't appear to have a clear
definition of private files associated with each domain, so for
simplicity we do it slightly different. On qemu each autogenerated
channel goes like:
channels/target/<domain-name>/<target name>
Whereas for libxl:
channels/target/<domain-name>-<target name>
Should note that if path is not specified it won't persist,
existing only on live XML, unless user had initially specified it.
Since support for libxl channels only came on Xen >= 4.5 we therefore
need to conditionally compile it with LIBXL_HAVE_DEVICE_CHANNEL.
After this patch and having a qemu guest agent:
$ cat domain.xml | grep -a1 channel | head -n 5 | tail -n 4
<channel type='unix'>
<source mode='bind' path='/tmp/channel'/>
<target type='xen' name='org.qemu.guest_agent.0'/>
</channel>
$ virsh create domain.xml
$ echo '{"execute":"guest-network-get-interfaces"}' | socat
stdio,ignoreeof unix-connect:/tmp/channel
{"execute":"guest-network-get-interfaces"}
{"return": [{"name": "lo", "ip-addresses": [{"ip-address-type": "ipv4",
"ip-address": "127.0.0.1", "prefix": 8}, {"ip-address-type": "ipv6",
"ip-address": "::1", "prefix": 128}], "hardware-address":
"00:00:00:00:00:00"}, {"name": "eth0", "ip-addresses":
[{"ip-address-type": "ipv4", "ip-address": "10.100.0.6", "prefix": 24},
{"ip-address-type": "ipv6", "ip-address": "fe80::216:3eff:fe40:88eb",
"prefix": 64}], "hardware-address": "00:16:3e:40:88:eb"}, {"name":
"sit0"}]}
Signed-off-by: Joao Martins <joao.m.martins@oracle.com>
Signed-off-by: Jim Fehlig <jfehlig@suse.com>
So far only guestfwd and virtio were supported. Add an additional
for Xen as libxl channels create a Xen console visible to the guest.
Signed-off-by: Joao Martins <joao.m.martins@oracle.com>
Signed-off-by: Jim Fehlig <jfehlig@suse.com>
The qemucapsprobe helper calls virQEMUCapsNewForBinaryInternal with
caps == NULL, causing the following crash:
Program received signal SIGSEGV, Segmentation fault.
#0 0x00007ffff788775f in virQEMUCapsInitHostCPUModel
(qemuCaps=qemuCaps@entry=0x649680, host=host@entry=0x10) at
src/qemu/qemu_capabilities.c:2969
#1 0x00007ffff7889dbf in virQEMUCapsNewForBinaryInternal
(caps=caps@entry=0x0, binary=<optimized out>,
libDir=libDir@entry=0x4033f6 "/tmp", cacheDir=cacheDir@entry=0x0,
runUid=runUid@entry=4294967295, runGid=runGid@entry=4294967295,
qmpOnly=true) at src/qemu/qemu_capabilities.c:4039
#2 0x0000000000401702 in main (argc=2, argv=0x7fffffffd968) at
tests/qemucapsprobe.c:73
Caused by v2.2.0-182-g68c7011.
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
https://bugzilla.redhat.com/show_bug.cgi?id=1368417
So far, when it comes to 'virsh update-device --config' of disks
we are limiting ourselves for just the disk source update and
just for CDROMs and floppies. This makes no sense. Especially if
you look around and see that we already allow full update to
graphics and net devices. So let's just take whatever XML user
wants to have there and replace our internal definition with it.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
libxl events are delivered to libvirt via the libxlDomainEventHandler
callback registered with libxl. Documenation in
$xensrc/tools/libxl/libxl_event.h states that the callback "may occur
on any thread in which the application calls libxl". This can result
in deadlock since many of the libvirt callees of libxl hold a lock on
the virDomainObj they are working on. When the callback is invoked, it
attempts to find a virDomainObj corresponding to the domain ID provided
by libxl. Searching the domain obj list results in locking each obj
before checking if it is active, and its ID equals the requested ID.
Deadlock is possible when attempting to lock an obj that is already
locked further up the call stack. Indeed, Max Ustermann reported an
instance of this deadlock
https://www.redhat.com/archives/libvir-list/2015-November/msg00130.html
Guido Rossmueller also recently stumbled across it
https://www.redhat.com/archives/libvir-list/2016-September/msg00287.html
Fix the deadlock by moving the lookup of virDomainObj to the
libxlDomainShutdownThread. After this patch, libxl events are
enqueued on the libvirt side and processed by dedicated thread,
avoiding the described deadlock.
Reported-by: Max Ustermann <ustermann78@web.de>
Reported-by: Guido Rossmueller <Guido.Rossmueller@gdata.de>
enum types are unsigned and the qemuGetCompressionProgram
function can return -1 on error. It is therefore inappropriate
to return an enum type. This fixes a build error where the
internal 'ret' variable was used in a comparison with -1
../../src/qemu/qemu_driver.c: In function 'qemuGetCompressionProgram':
../../src/qemu/qemu_driver.c:3280:5: error: comparison of unsigned expression < 0 is always false [-Werror=type-limits]
../../src/qemu/qemu_driver.c:3289:5: error: comparison of unsigned expression < 0 is always false [-Werror=type-limits]
cc1: all warnings being treated as errors
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
When creating a copy of virDomainDef we save ourselves the
trouble of writing deep-copy functions and just format and parse
back domain/device XML. However, the XML we are parsing was
already fully formatted - there is no reason to run post parse
callbacks (which fill in blanks - there are none!).
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
This is an internal flag that prevents our two entry points to
XML parsing (virDomainDefParse and virDomainDeviceDefParse) from
running post parse callbacks. This is expected to be used in
cases when we already have full domain/device XML and we are just
parsing it back (i.e. virDomainDefCopy or virDomainDeviceDefCopy)
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Just like we did two commits ago, don't try to fetch capabilities
for non-existing binary. Re-use the ones we have for running
domain.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Just like we did two commits ago, don't try to fetch capabilities
for non-existing binary. Re-use the ones we have for running
domain.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
We can't rely on def->emulator path. It may be provided by user
as we give them opportunity to provide their own XML for
migration. Therefore the path may point to just whatever binary
(or even to a non-existent file). Moreover, this path is meant
for destination, but the capabilities lookup is done on source.
What we can do is to assume same capabilities for post parse
callbacks as the running domain has. They will be used just to
add some default models/controllers/devices/... anyway.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Just like virDomainDefPostParseCallback has gained new
parseOpaque argument, we need to follow the logic with
virDomainDeviceDefPostParse.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
We want to pass the proper opaque pointer instead of NULL to
virDomainDefParse and subsequently virDomainDefParseNode too.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
We want to pass the proper opaque pointer instead of NULL to
virDomainDefParseXML and subsequently virDomainDefPostParse too.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Some callers might want to pass yet another pointer to opaque
data to post parse callbacks. The driver generic one is not
enough because two threads executing post parse callback might
want to see different data (e.g. domain object pointer that
domain def belongs to).
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Based upon a patch from Chen Hanxiao <chenhanxiao@gmail.com>, rather than
need to call virFindFileInPath twice, let's just save the path and pass it
along with the compressed type. (NB: the second call would be in virExec as
called from virCommandRunAsync which is called from qemuMigrationToFile
using the argument 'compressor' which up to this point would be the string
from the cfg file that isn't the fully qualified path).
Since we now have the path, we can remove qemuCompressProgramName which
would return NULL or the string representation of the compress type.
There's only one caller and the code is duplicitous just converting the
recently converted cfg image name back into it's string value in order to
get/find the path to the image. A subsequent patch can return this path.
Let's do some more code reuse - there are 3 other callers that care to
check/get the compress program. Each of those though cares whether the
requested cfg image is valid and exists. So, add a parameter to handle
those cases.
NB: We won't need to initialize the returned value in the case where
the cfg image doesn't exist since the called program will handle that.
Add a new parameter 'styleFormat' to be used when printing the
warning message so that it's "clearer" what style of compression
call caused the error. Add that style to both messages as a paremter.
Also a VIR_WARN error message doesn't need to be translated
(e.g. inside _()), so remove the need for the translation.
There's only one caller now anyway... Besides it's just a shell for
getting the compress type. Subsequent patches will return the path
to the compression program.
Split out the guts of getCompressionType to perform the same functionality
in the new helper program with a subsequent patch goal to be reusable for
other callers making similar checks/calls to ensure the compression type
is valid and that the compression program cannot be found.
If passing an empty usbdevice_list to libxl, qemu will always get an
-usb parameter for HVM guests with only non-USB input devices. This
causes qemu to crash when passing pvusb device on HVM guests.
The solution is to allocate the list only when an item to put in it
is found.
Both cpuCompare* APIs are renamed to virCPUCompare*. And they should now
work for any guest CPU definition, i.e., even for host-passthrough
(trivial) and host-model CPUs. The implementation in x86 driver is
enhanced to provide a hint about -noTSX Broadwell and Haswell models
when appropriate.
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
The function is similar to virCPUDataCheckFeature, but it works directly
on CPU definition rather than requiring it to be transformed into CPU
data first.
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
The API is supposed to make sure the provided CPU definition does not
use a CPU model which is not supported by the hypervisor (if at all
possible, of course).
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
Keeping nfeatures_max set to 0 while nfeatures > 0 and some features are
already stored in features array is just asking for problems once we
want to add a new feature into the array.
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
The reworked API is now called virCPUUpdate and it should change the
provided CPU definition into a one which can be consumed by the QEMU
command line builder:
- host-passthrough remains unchanged
- host-model is turned into custom CPU with a model and features
copied from host
- custom CPU with minimum match is converted similarly to host-model
- optional features are updated according to host's CPU
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
x86ModelFromCPU is used to provide CPUID data for features matching
@policy. This patch allows callers to set @policy to -1 to get combined
CPUID for all CPU features (including those implicitly provided a CPU
model) specified in CPU def.
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
The domain capabilities XML is capable of showing whether each guest CPU
mode is supported or not with a possibility to provide additional
details. This patch enhances host-model capability to advertise the
exact CPU model which will be used as a host-model:
<cpu>
...
<mode name='host-model' supported='yes'>
<model fallback='allow'>Broadwell</model>
<vendor>Intel</vendor>
<feature policy='disable' name='aes'/>
<feature policy='require' name='vmx'/>
</mode>
...
</cpu>
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
The ARM CPU driver wrongly reported host CPU model as "host", which made
host-model to be just an alias for host-passthrough. Let's drop this
insanity.
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
Host capabilities provide libvirt's view of the host CPU, but for a
useful support for host-model CPUs we really need a hypervisor's view of
the CPU. And since the view can be differ with emulator, qemu
capabilities is the best place to store the host CPU model.
This patch just copies the CPU model from host capabilities, but this
will change in the future.
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
The function filters all CPU features through a given callback while
copying CPU model related parts of a CPU definition.
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
The function moves CPU model related parts from one CPU definition to
another. It can be used to avoid unnecessary copies from a temporary CPU
definitions which will be freed anyway.
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
Useful for copying a CPU definition without model related parts (i.e.,
without model name, feature list, vendor).
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
In case a hypervisor is able to tell us a list of supported CPU models
and whether each CPU models can be used on the current host, we can
propagate this to domain capabilities. This is a better alternative
to calling virConnectCompareCPU for each supported CPU model.
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
Listing all CPU models supported by QEMU in domain capabilities makes
little sense when libvirt will refuse any model it doesn't know about.
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
Some CPU drivers (such as arm) do not provide list of CPUs libvirt
supports and just pass any CPU model from domain XML directly to QEMU.
Such driver need to return models == NULL and success from cpuGetModels.
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
qemu_command.c should deal with translating our domain definition into a
QEMU command line and nothing else.
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
The list of supported CPU models in domain capabilities is stored in
virDomainCapsCPUModels. Let's use the same object for storing CPU models
in QEMU capabilities.
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
The patch adds <cpu> element to domain capabilities XML:
<cpu>
<mode name='host-passthrough' supported='yes'/>
<mode name='host-model' supported='yes'/>
<mode name='custom' supported='yes'>
<model>Broadwell</model>
<model>Broadwell-noTSX</model>
...
</mode>
</cpu>
Applications can use it to inspect what CPU configuration modes are
supported for a specific combination of domain type, emulator binary,
guest architecture and machine type.
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
Our internal APIs mostly use virArch rather than strings. Switching
cpuGetModels to virArch will save us from unnecessary conversions in the
future.
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
By default, virt-manager (and likely other libvirt-based apps) sets
the VIR_MIGRATE_PERSIST_DEST flag when invoking the migrate API, which
fails in a Xen setup since the libxl driver does not support the flag.
Persisting a domain is a trivial task in the grand scheme of migration,
so be nice to libvirt apps and add support for VIR_MIGRATE_PERSIST_DEST
in the libxl driver.
Signed-off-by: Jim Fehlig <jfehlig@suse.com>
We have a few of senarios that libvirtd would invoke qemuProcessStop
and leave a "shutting down" in /var/log/libvirt/qemu/$DOMAIN.log.
The shutoff reason showing in debug log is also very important
for us to know why VM shutting down in domain log,
as we seldom enable debug log of libvirtd.
Signed-off-by: Chen Hanxiao <chenhanxiao@gmail.com>
https://bugzilla.redhat.com/show_bug.cgi?id=1322717
During offline migration, no storage is copied. Nor disks, nor
NVRAM file, nor anything. We use qemu for that and because domain
is not running there's nobody to copy that for us.
We should document this to avoid confusing users.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Calling virDomainGetEmulatorPinInfo on a live VM with automatic NUMA
pinning and VIR_DOMAIN_AFFECT_CONFIG would return the automatic pinning
data in some cases which is bogus. Use the autoCpuset property only when
called on a live definition.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1365779
Calling virDomainGetVcpuPinInfo on a live VM with automatic NUMA pinning
and VIR_DOMAIN_AFFECT_CONFIG would return the automatic pinning data
in some cases which is bogus. Use the autoCpuset property only when
called on a live definition.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1365779
Sometimes adding a separate variable to access vm->privateData is not
necessary. Add a macro that will do the typecasting rather than having
to add a temp variable to force the compiler to typecast it.
Old libvirt represents
<graphics type='spice'>
<listen type='none'/>
</graphics>
as
<graphics type='spice' autoport='no'/>
In this mode, QEMU doesn't listen for SPICE connection anywhere and
clients have to use virDomainOpenGraphics* APIs to attach to the domain.
That is, the client has to run on the same host where the domains runs
and it's impossible to tell the client to reconnect to the destination
QEMU during migration (unless there is some kind of proxy on the host).
While current libvirt correctly ignores such graphics devices when
creating graphics migration cookie, old libvirt just sends
<graphics type='spice' port='0' listen='0.0.0.0' tlsPort='-1'/>
in the cookie. After seeing this cookie, we happily would call
client_migrate_info QMP command and wait for SPICE_MIGRATE_COMPLETED
event, which is quite pointless since the doesn't know where to connecti
anyway. We should just ignore such cookies.
https://bugzilla.redhat.com/show_bug.cgi?id=1376083
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
Checking if a domain's definition or if it is active before we got a job
is pointless since the domain might have changed in the meantime.
Luckily libvirtd didn't crash when the API tried to talk to an inactive
domain:
debug : qemuDomainObjBeginJobInternal:2914 : Started job: modify
(async=none vm=0x7f8f340140c0 name=ble)
debug : qemuDomainObjEnterMonitorInternal:3137 : Entering monitor
(mon=(nil) vm=0x7f8f340140c0 name=ble)
warning : virObjectLock:319 : Object (nil) ((unknown)) is not a
virObjectLockable instance
debug : qemuMonitorOpenGraphics:3505 : protocol=spice fd=27
fdname=graphicsfd skipauth=1
error : qemuMonitorOpenGraphics:3508 : invalid argument: monitor must
not be NULL
debug : qemuDomainObjExitMonitorInternal:3160 : Exited monitor
(mon=(nil) vm=0x7f8f340140c0 name=ble)
debug : qemuDomainObjEndJob:3068 : Stopping job: modify (async=none
vm=0x7f8f340140c0 name=ble)
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
We can receive NULL as sync reply in two situations. First
is garbage sync reply and this situation is handled by
resending sync message. Second is different cases
of rebooting guest, destroing domain etc and we can
give more meaningful error message. Actually we have
this error message in qemuAgentCommand already which checks
for the same sitatuion. AFAIK case with mon->running
is just to be safe on adding some future(?) cases of
returning NULL reply.
We can easily handle receiving garbage on sync. We don't
have to make client deal with this situation. We just
need to resend sync command but this time garbage is
not be possible.
When we wait for sync reply we can receive delayed
reply to syncs or commands that were sent erlier. We can
safely skip them until we receive sync reply with correct id.
There is no much sense report this situation to client.
Actually with a bit of "luck" if we involve client into
this the play can go on forever: send sync 0, receive
sync reply -1, send sync 1, receive reply 0 ...
After sync is sent we can receive garbare and this is not error.
Consider next regular case:
1. libvirtd sends sync
2. qga sends partial sync reply and die
3. libvirtd sends sync
4. qga sends sync reply
5. libvirtd receives garbage
(half of first reply and second reply together)
We should handle this situation as it is recoverable.
Next sync can succeed. Let's report reply is NULL,
it will be converted to the VIR_ERR_AGENT_UNSYNCED
which signals client to retry.
Errors in qemuAgentIOProcessLine stop agent IO processing just
like any regular IO error, however some of current errors
that this functions spawns are false positives. Consider
next case for example:
1. send sync (unsynced state)
2. receive sync reply (sync established)
3. command send, but timeout occured (unsynced state)
4. receive command reply
Last IO triggers error because current code ignores
only delayed syncs when unsynced
We should not treat any delayed reply as error in unsynced
state. Until client and qga are not in sync delayed reply to any
command is possible. msg == NULL is the exact criterion
that we are not in sync.
Put it into qemuDomainPrepareShmemChardev() so it can be used later.
Also don't fill in the path unless the server option is enabled.
Signed-off-by: Martin Kletzander <mkletzan@redhat.com>
Some checks will need to be performed for newer device types as well, so
let's not duplicate them.
Signed-off-by: Martin Kletzander <mkletzan@redhat.com>
Commit 839a060 tied the lifecycle of virtlogd more
closely to that of libvirtd. Unfortunately, while starting
virtlogd when libvirtd is started is definitely a good idea,
restarting virtlogd or shutting it down at any time outside
of system poweroff is not.
Revert part of that commit by removing the PartOf= lines,
meaning that only startup requests will be propagated from
libvirtd to virtlogd.
Resolves: https://bugzilla.redhat.com/1372576
Now that we have two same implementations for getting path for
huge pages backed guest memory, lets merge them into one function.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
When trying to migrate a huge page enabled guest, I've noticed
the following crash. Apparently, if no specific hugepages are
requested:
<memoryBacking>
<hugepages/>
</memoryBacking>
and there are no hugepages configured on the destination, we try
to dereference a NULL pointer.
Program received signal SIGSEGV, Segmentation fault.
0x00007fcc907fb20e in qemuGetHugepagePath (hugepage=0x0) at qemu/qemu_conf.c:1447
1447 if (virAsprintf(&ret, "%s/libvirt/qemu", hugepage->mnt_dir) < 0)
(gdb) bt
#0 0x00007fcc907fb20e in qemuGetHugepagePath (hugepage=0x0) at qemu/qemu_conf.c:1447
#1 0x00007fcc907fb2f5 in qemuGetDefaultHugepath (hugetlbfs=0x0, nhugetlbfs=0) at qemu/qemu_conf.c:1466
#2 0x00007fcc907b4afa in qemuBuildMemoryBackendStr (size=4194304, pagesize=0, guestNode=0, userNodeset=0x0, autoNodeset=0x0, def=0x7fcc70019070, qemuCaps=0x7fcc70004000, cfg=0x7fcc5c011800, backendType=0x7fcc95087228, backendProps=0x7fcc95087218,
force=false) at qemu/qemu_command.c:3297
#3 0x00007fcc907b4f91 in qemuBuildMemoryCellBackendStr (def=0x7fcc70019070, qemuCaps=0x7fcc70004000, cfg=0x7fcc5c011800, cell=0, auto_nodeset=0x0, backendStr=0x7fcc70020360) at qemu/qemu_command.c:3413
#4 0x00007fcc907c0406 in qemuBuildNumaArgStr (cfg=0x7fcc5c011800, def=0x7fcc70019070, cmd=0x7fcc700040c0, qemuCaps=0x7fcc70004000, auto_nodeset=0x0) at qemu/qemu_command.c:7470
#5 0x00007fcc907c5fdf in qemuBuildCommandLine (driver=0x7fcc5c07b8a0, logManager=0x7fcc70003c00, def=0x7fcc70019070, monitor_chr=0x7fcc70004bb0, monitor_json=true, qemuCaps=0x7fcc70004000, migrateURI=0x7fcc700199c0 "defer", snapshot=0x0,
vmop=VIR_NETDEV_VPORT_PROFILE_OP_MIGRATE_IN_START, standalone=false, enableFips=false, nodeset=0x0, nnicindexes=0x7fcc95087498, nicindexes=0x7fcc950874a0, domainLibDir=0x7fcc700047c0 "/var/lib/libvirt/qemu/domain-1-fedora") at qemu/qemu_command.c:9547
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Both qemu monitor and agent print the same
log on HUANGUP event, which would be confusing
when reading libvirtd log.
This patch will give a different log message to them.
Signed-off-by: Chen Hanxiao <chenhanxiao@gmail.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Most of QEMU's PCI display device models, such as:
libvirt video/model/@type QEMU -device
------------------------- ------------
cirrus cirrus-vga
vga VGA
qxl qxl-vga
virtio virtio-vga
come with a linear framebuffer (sometimes called "VGA compatibility
framebuffer"). This linear framebuffer lives in one of the PCI device's
MMIO BARs, and allows guest code (primarily: firmware drivers, and
non-accelerated OS drivers) to display graphics with direct memory access.
Due to architectural reasons on aarch64/KVM hosts, this kind of
framebuffer doesn't / can't work in
qemu-system-(arm|aarch64) -M virt
machines. Cache coherency issues guarantee a corrupted / unusable display.
The problem has been researched by several people, including kvm-arm
maintainers, and it's been decided that the best way (practically the only
way) to have boot time graphics for such guests is to consolidate on
QEMU's "virtio-gpu-pci" device.
>From <https://bugzilla.redhat.com/show_bug.cgi?id=1195176>, libvirt
supports
<devices>
<video>
<model type='virtio'/>
</video>
</devices>
but libvirt unconditionally maps @type='virtio' to QEMU's "virtio-vga"
device model. (See the qemuBuildDeviceVideoStr() function and the
"qemuDeviceVideo" enum impl.)
According to the above, this is not right for the "virt" machine type; the
qemu-system-(arm|aarch64) binaries don't even recognize the "virtio-vga"
device model (justifiedly). Whereas "virtio-gpu-pci", which is a pure
virtio device without a compatibility framebuffer, is available, and works
fine.
(The ArmVirtQemu ("AAVMF") platform of edk2 -- that is, the UEFI firmware
for "virt" -- supports "virtio-gpu-pci", as of upstream commit
3ef3209d3028. See
<https://tianocore.acgmultimedia.com/show_bug.cgi?id=66>.)
Override the default mapping of "virtio", from "virtio-vga" to
"virtio-gpu-pci", if qemuDomainMachineIsVirt() evaluates to true.
Cc: Andrea Bolognani <abologna@redhat.com>
Cc: Drew Jones <drjones@redhat.com>
Cc: Marc-André Lureau <marcandre.lureau@redhat.com>
Cc: Martin Kletzander <mkletzan@redhat.com>
Suggested-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1372901
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Acked-by: Martin Kletzander <mkletzan@redhat.com>
There is nothing Linux-specific in that function. Also since commit
8c3b5bf481 mingw build is broken due to
the fact that this function is not compiled in the library.
Signed-off-by: Martin Kletzander <mkletzan@redhat.com>
In testOpenDefault we create a virtual computer that is later
presented to user. We also pretend to have NUMA cells and
initialize them somehow. But whilst doing so a magical constant
is used. Drop it.
Signed-off-by: Tomáš Ryšavý <tom.rysavy.0@gmail.com>
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
We will need this function shortly when implementing
nodeGetCPUStats in the test driver.
Signed-off-by: Tomáš Ryšavý <tom.rysavy.0@gmail.com>
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Use the state information (online, hotpluggable) provided by the monitor
code rather than trying to infer it. This fixes an issue where on
architectures that require hotplug of multiple threads at once the
sub-cores would get updated as offline on daemon restart thus creating
an invalid configuration.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1375783
Return whether a vcpu entry is hotpluggable or online so that upper
layers don't have to infer the information from other data.
Advantage is that this code can be tested by unit tests.
The algorithm that matches data from query-cpus and
query-hotpluggable-cpus is quite complex. Start using descriptive
iterator names to avoid confusion.
https://bugzilla.redhat.com/show_bug.cgi?id=1372613
Apparently, some management applications use the following code
pattern when waiting for a block job to finish:
while (1) {
virDomainGetBlockJobInfo(dom, disk, info, flags);
if (info.cur == info.end)
break;
sleep(1);
}
Problem with this approach is in its corner cases. In case of
QEMU, libvirt merely pass what has been reported on the monitor.
However, if the block job hasn't started yet, qemu reports cur ==
end == 0 which tricks mgmt apps into thinking job is complete.
The solution is to mangle cur/end values as described here [1].
1: https://www.redhat.com/archives/libvir-list/2016-September/msg00017.html
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Even though we merely just pass to users whatever qemu provided
on the monitor, we still do some translation. For instance we
turn bytes into mebibytes, or fix job type if needed. However, in
the future there is more fixing to be done so this code deserves
its own function.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Name it virNumaGetHostMemoryNodeset and return only NUMA nodes which
have memory installed. This is necessary as the kernel is not very happy
to set the memory cgroup setting for nodes which do not have any memory.
This would break vcpu hotplug with following message on such
configruation:
Invalid value '0,8' for 'cpuset.mems': Invalid argument
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1375268
In a full domain config, libvirt allows overriding the normal PCI
vs. PCI Express rules when a device address is explicitly provided
(so, e.g., you can force a legacy PCI device to plug into a PCIe port,
although libvirt would never do that on its own). However, due to a
bug libvirt doesn't give this same leeway when hotplugging devices. On
top of that, current libvirt assumes that *all* devices are legacy
PCI. The result of all this is that it's impossible to hotplug a
device into a PCIe port, even if you manually add the PCI address.
This can all be traced to the function
virDomainPCIAddressEnsureAddr(), and the fact that it calls
virDomainPCIaddressReserveSlot() for manually set addresses, and that
function hardcodes the argument "fromConfig" to false (meaning "this
address was auto-assigned, so it should be subject to stricter
validation").
Since virDomainPCIAddressReserveSlot() is just a one line simple
wrapper around virDomainPCIAddressReserveAddr() (adding in a hardcoded
reserveEntireSlot = true and fromConfig = false), all that's needed to
solve the problem with no unwanted side effects is to replace that
call for virDomainPCIAddressReserveSlot() with a direct call to
virDomainPCIAddressReserveAddr(), but with reserveEntireSlot = true,
fromConfig = true. That's what this patch does.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1337490
virQEMUDriverConfigNew() always initializes the bitmap in its
cgroupControllers member to -1 (i.e. all 1's).
Prior to commit a9331394, if qemu.conf had a line with
"cgroup_controllers", cgroupControllers would get reset to 0 before
going through a loop setting a bit for each named cgroup controller.
commit a9331394 left out the "reset to 0" part, so cgroupControllers
would always be -1; if you didn't want a controller included, there
was no longer a way to make that happen.
This was discovered by users who were using qemu commandline
passthrough to use the "input-linux" method of directing
keyboard/mouse input to a virtual machine:
https://www.redhat.com/archives/vfio-users/2016-April/msg00105.html
Here's the first report I found of the problem encountered after
upgrading libvirt beyond v2.0.0:
https://www.redhat.com/archives/vfio-users/2016-August/msg00053.html
Thanks to sL1pKn07 SpinFlo <sl1pkn07@gmail.com> for bringing the
problem up in IRC, and then taking the time to do a git bisect and
find the patch that started the problem.
previous commit:
commit 2c3223785c
Author: John Ferlan <jferlan@redhat.com>
Date: Mon Jun 13 12:30:34 2016 -0400
qemu: Add the ability to hotplug the TLS X.509 environment
added a parameter "bool listen" in some methods. This
unfortunately clashes with the listen() method, causing
compile failures on certain platforms (RHEL-6 for example)
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Commit id 'a48c7141' altered how to determine if a volume was encrypted
by adding a peek at an offset into the file at a specific buffer location.
Unfortunately, all that was compared was the first "char" of the buffer
against the expect "int" value.
Restore the virReadBufInt32BE to get the complete field in order to
compare against the expected value from the qcow2EncryptionInfo or
qcow1EncryptionInfo "modeValue" field.
This restores the capability to create a volume with encryption, then
refresh the pool, and still find the encryption for the volume.
A LUKS volume uses the volume secret type just like the QCOW2 secret, so
adjust the loading of the default secrets to handle any volume that the
virStorageFileGetMetadataFromBuf code has deemed to be an encrypted volume
to search for the volume's secret. This lookup is done by volume usage
where the usage is expected to be the path to volume.
When a new filter is being defined, the return code is not handled properly,
thus triggering OOM error reporting routine (bug introduced by 51b2606f).
Signed-off-by: Erik Skultety <eskultet@redhat.com>
When migration fails, we need to poke QEMU monitor to check for a reason
of the failure. We did this using query-migrate QMP command, which is
not supposed to return any meaningful result on the destination side.
Thus if the monitor was still functional when we detected the migration
failure, parsing the answer from query-migrate always failed with the
following error message:
"info migration reply was missing return status"
This irrelevant message was then used as the reason for the migration
failure replacing any message we might have had.
Let's use harmless query-status for poking the monitor to make sure we
only get an error if the monitor connection is broken.
https://bugzilla.redhat.com/show_bug.cgi?id=1374613
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
Commit id 'b00d7f29' shifted the opening of the /sys/devices/intel_cqm/type
file from event enable to perf event initialization. If the file did not
exist, then an error would be written to the domain log:
2016-09-06 20:51:21.677+0000: 7310: error : virFileReadAll:1360 : Failed to open file '/sys/devices/intel_cqm/type': No such file or directory
Since the error is now handled in virPerfEventEnable by checking if the
event_attr->attrType == 0 for CMT, MBML, and MBMT events - we can just
use the Quiet API in order to not log the error we're going to throw away.
Additionally, rather than using virReportSystemError, use virReportError
and VIR_ERR_ARGUMENT_UNSUPPORTED in order to signify that support isn't there
for that type of perf event - adjust the error message as well.
Akin to previous commit but for "virsh cpu-baseline" which
computes a baseline CPU for a set of host cpu elements.
Signed-off-by: Joao Martins <joao.m.martins@oracle.com>
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Implement support for "virsh cpu-compare" so that we can calculate
common cpu element between a pool of hosts, which had a requirement
of providing host cpu description.
Signed-off-by: Joao Martins <joao.m.martins@oracle.com>
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Parse libxl_hwcap accounting for versions since Xen 4.4 - Xen 4.7.
libxl_hwcaps is a set of cpuid leaves output that is described in [0] or
[1] in Xen 4.7. This is a collection of CPUID leaves that we version
in libvirt whenever feature words are reordered or added. Thus we keep the
common ones in one struct and others for each version. Since
libxl_hwcaps doesn't appear to have a stable format across all supported
versions thus we need to keep track of changes as a compromise until it's
exported in xen libxl API. We don't fail in initializing the driver in case
parsing of hwcaps failed for that reason. In addition, change the notation
on PAE feature such that is easier to read which bit it corresponds.
[0] xen/include/asm-x86/cpufeature.h
[1] xen/include/public/arch-x86/cpufeatureset.h
Signed-off-by: Joao Martins <joao.m.martins@oracle.com>
Add support for describing cpu topology in host cpu element. In doing
so, refactor hwcaps part to its own helper namely libxlCapsInitCPU to
handle all host cpu related operations, including topology.
Signed-off-by: Joao Martins <joao.m.martins@oracle.com>
Qemu always opens the tray if forced to. Skip the waiting step in such
case.
This also helps if qemu does not report the tray change event when
opening the cdrom forcibly (the documentation says that the event will
not be sent although qemu in fact does trigger it even if @force is
selceted).
This is a workaround for a qemu issue where qemu does not send the tray
change event in some cases (after migration with empty closed locked
drive) and thus renders the cdrom useless from libvirt's point of view.
Partially resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1368368
When a source image is dropped when missing due to startup policy the
policy needs to be cleared since it was relevant only for the given
storage source. New sources need to update it if needed.
Just like in the previous commit, teach qemu driver to detect
whether qemu supports this configuration knob or not.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Add a new secret usage type known as "tls" - it will handle adding the
secret objects for various TLS objects that need to provide some sort
of passphrase in order to access the credentials.
The format is:
<secret ephemeral='no' private='no'>
<description>Sample TLS secret</description>
<usage type='tls'>
<name>mumblyfratz</name>
</usage>
</secret>
Once defined and a passphrase set, future patches will allow the UUID
to be set in the qemu.conf file and thus used as a secret for various
TLS options such as a chardev serial TCP connection, a NBD client/server
connection, and migration.
Signed-off-by: John Ferlan <jferlan@redhat.com>
If the incoming XML defined a path to a TLS X.509 certificate environment,
add the necessary 'tls-creds-x509' object to the VIR_DOMAIN_CHR_TYPE_TCP
character device.
Likewise, if the environment exists the hot unplug needs adjustment as
well. Note that all the return ret were changed to goto cleanup since
the cfg needs to be unref'd
Signed-off-by: John Ferlan <jferlan@redhat.com>
When building a chardev device string for tcp, add the necessary pieces to
access provide the TLS X.509 path to qemu. This includes generating the
'tls-creds-x509' object and then adding the 'tls-creds' parameter to the
VIR_DOMAIN_CHR_TYPE_TCP command line.
Finally add the tests for the qemu command line. This test will make use
of the "new(ish)" /etc/pki/qemu setting for a TLS certificate environment
by *not* "resetting" the chardevTLSx509certdir prior to running the test.
Also use the default "verify" option (which is "no").
Signed-off-by: John Ferlan <jferlan@redhat.com>
Add a new TLS X.509 certificate type - "chardev". This will handle the
creation of a TLS certificate capability (and possibly repository) for
properly configured character device TCP backends.
Unlike the vnc and spice there is no "listen" or "passwd" associated. The
credentials eventually will be handled via a libvirt secret provided to
a specific backend.
Make use of the default verify option as well.
Signed-off-by: John Ferlan <jferlan@redhat.com>
Rather than specify perhaps multiple TLS X.509 certificate directories,
let's create a "default" directory which can then be used if the service
(e.g. for now vnc and spice) does not supply a default directory.
Since the default for vnc and spice may have existed before without being
supplied, the default check will first check if the service specific path
exists and if so, set the cfg entry to that; otherwise, the default will
be set to the (now) new defaultTLSx509certdir.
Additionally add a "default_tls_x509_verify" entry which can also be used
to force the peer verification option (for vnc it's a x509verify option).
Add/alter the macro for the option being found in the config file to accept
the default value.
Signed-off-by: John Ferlan <jferlan@redhat.com>
If a migration of a domain which is already defined on the destination
host failed early (before we tried to start QEMU), we would forget to
remove the incoming transient definition. Later on when someone starts
the domain on the destination host, we will use the stale incoming
definition and the persistent definition will just be ignored.
https://bugzilla.redhat.com/show_bug.cgi?id=1368774
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
The code for replacing domain's transient definition with the persistent
one is repeated in several places and we'll need to add one more. Let's
make a nice helper for it.
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
There is an issue with a wrong label inside vah_add_path().
The compilation fails with the error:
make[3]: Entering directory '/tmp/libvirt/src'
CC security/virt_aa_helper-virt-aa-helper.o
security/virt-aa-helper.c: In function 'vah_add_path':
security/virt-aa-helper.c:769:9: error: label 'clean' used but not defined
goto clean;
This patch moves 'clean' label to 'cleanup' label.
Signed-off-by: Julio Faracco <jcfaracco@gmail.com>
This patch fixes a segfault in virt-aa-helper caused by attempting to
modify a static string literal. It is triggered when a domain has a
<filesystem> with type='mount' configured read-only and libvirt is
using the AppArmor security driver for sVirt confinement. An "R" is
passed into the function and converted to 'r'.
Similarly to vcpu hotplug the emulator thread cgroup numa mapping needs
to be relaxed while hot-adding vcpus so that the threads can allocate
data in the DMA zone.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1370084
When hot-adding vcpus qemu needs to allocate some structures in the DMA
zone which may be outside of the numa pinning. Extract the code doing
this in a set of helpers so that it can be reused.
There is a possibility that qemu driver frees by unreferencing its
closeCallbacks pointer as it has the only reference to the object,
while in fact not all users of CloseCallbacks called thier
virCloseCallbacksUnset.
Backtrace is the following:
Thread #1:
0 in pthread_cond_wait@@GLIBC_2.3.2 () from /lib64/libpthread.so.0
1 in virCondWait (c=<optimized out>, m=<optimized out>)
at util/virthread.c:154
2 in virThreadPoolFree (pool=0x7f0810110b50)
at util/virthreadpool.c:266
3 in qemuStateCleanup () at qemu/qemu_driver.c:1116
4 in virStateCleanup () at libvirt.c:808
5 in main (argc=<optimized out>, argv=<optimized out>)
at libvirtd.c:1660
Thread #2:
0 in virClassIsDerivedFrom (klass=0xdeadbeef, parent=0x7f0837c694d0) at util/virobject.c:169
1 in virObjectIsClass (anyobj=anyobj@entry=0x7f08101d4760, klass=<optimized out>) at util/virobject.c:365
2 in virObjectLock (anyobj=0x7f08101d4760) at util/virobject.c:317
3 in virCloseCallbacksUnset (closeCallbacks=0x7f08101d4760, vm=vm@entry=0x7f08101d47b0, cb=cb@entry=0x7f081d078fc0 <qemuProcessAutoDestroy>) at util/virclosecallbacks.c:163
4 in qemuProcessAutoDestroyRemove (driver=driver@entry=0x7f081018be50, vm=vm@entry=0x7f08101d47b0) at qemu/qemu_process.c:6368
5 in qemuProcessStop (driver=driver@entry=0x7f081018be50, vm=vm@entry=0x7f08101d47b0, reason=reason@entry=VIR_DOMAIN_SHUTOFF_SHUTDOWN, asyncJob=asyncJob@entry=QEMU_ASYNC_JOB_NONE, flags=flags@entry=0) at qemu/qemu_process.c:5854
6 in processMonitorEOFEvent (vm=0x7f08101d47b0, driver=0x7f081018be50) at qemu/qemu_driver.c:4585
7 qemuProcessEventHandler (data=<optimized out>, opaque=0x7f081018be50) at qemu/qemu_driver.c:4629
8 in virThreadPoolWorker (opaque=opaque@entry=0x7f0837c4f820) at util/virthreadpool.c:145
9 in virThreadHelper (data=<optimized out>) at util/virthread.c:206
10 in start_thread () from /lib64/libpthread.so.0
Let's reference CloseCallbacks object in virCloseCallbacksSet and
unreference in virCloseCallbacksUnset.
Signed-off-by: Maxim Nestratov <mnestratov@virtuozzo.com>
In the latest glibc, major() and minor() functions are marked as
deprecated (glibc commit dbab6577):
CC util/libvirt_util_la-vircgroup.lo
util/vircgroup.c: In function 'virCgroupGetBlockDevString':
util/vircgroup.c:768:5: error: '__major_from_sys_types' is deprecated:
In the GNU C Library, `major' is defined by <sys/sysmacros.h>.
For historical compatibility, it is currently defined by
<sys/types.h> as well, but we plan to remove this soon.
To use `major', include <sys/sysmacros.h> directly.
If you did not intend to use a system-defined macro `major',
you should #undef it after including <sys/types.h>.
[-Werror=deprecated-declarations]
if (virAsprintf(&ret, "%d:%d ", major(sb.st_rdev), minor(sb.st_rdev)) < 0)
^~
In file included from /usr/include/features.h:397:0,
from /usr/include/bits/libc-header-start.h:33,
from /usr/include/stdio.h:28,
from ../gnulib/lib/stdio.h:43,
from util/vircgroup.c:26:
/usr/include/sys/sysmacros.h:87:1: note: declared here
__SYSMACROS_DEFINE_MAJOR (__SYSMACROS_FST_IMPL_TEMPL)
^
Moreover, in the glibc commit, there's suggestion to keep
ordering of including of header files as implemented here.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Current implementation uses the dev.cpu.0.freq sysctl that is
provided by the cpufreq(4) framework and returns the actual
CPU frequency. However, there are environments where it's not available,
e.g. when running nested in KVM. In this case fall back to hw.clockrate
that reports CPU frequency at the boot time.
Resolves (hopefully):
https://bugzilla.redhat.com/show_bug.cgi?id=1369964
We already guarantee that virtlogd.socket is enabled/disabled
along with libvirtd.service, but if libvirtd.service has just
been installed and is started before rebooting, then
virtlogd.socket will not be running and guest startup will
fail.
Add Requires=virtlogd.socket to libvirtd.service to make sure
virtlogd.socket is always started along with libvirtd.service,
and add Before=libvirtd.service to both virtlogd.socket and
virtlogd.service so that virtlogd never disappears before
libvirtd has exited.
Also add PartOf=libvirtd.service to both virtlogd.socket and
virtlogd.service, so that virtlogd can be shut down when not
needed.
Resolves: https://bugzilla.redhat.com/1372576
We already have the ability to turn off dumping of guest
RAM via the domain XML. This is not particularly useful
though, as it is under control of the management application.
What is needed is a way for the sysadmin to turn off guest
RAM defaults globally, regardless of whether the mgmt app
provides its own way to set this in the domain XML.
So this adds a 'dump_guest_core' option in /etc/libvirt/qemu.conf
which defaults to false. ie guest RAM will never be included in
the QEMU core dumps by default. This default is different from
historical practice, but is considered to be more suitable as
a default because
a) guest RAM can be huge and so inflicts a DOS on the host
I/O subsystem when dumping core for QEMU crashes
b) guest RAM can contain alot of sensitive data belonging
to the VM owner. This should not generally be copied
around inside QEMU core dumps submitted to vendors for
debugging
c) guest RAM contents are rarely useful in diagnosing
QEMU crashes
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Currently the QEMU processes inherit their core dump rlimit
from libvirtd, which is really suboptimal. This change allows
their limit to be directly controlled from qemu.conf instead.
With current perf framework, this patch adds support and documentation
for more perf events, including cache misses, cache references, cpu cycles,
and instructions.
Signed-off-by: Qiaowei Ren <qiaowei.ren@intel.com>
Introduce a static attr table and refactor virPerfEventEnable() for
general purpose usage.
This patch creates a static table/matrix that converts the VIR_PERF_EVENT_*
events into their respective "attr.type" and "attr.config" so that
virPerfEventEnable doesn't have the switch the calling function passes
by value the 'type'.
Signed-off-by: Qiaowei Ren <qiaowei.ren@intel.com>
Since the domain lock is not held during preparation of an external XML
config, it is possible that the value can change resulting in unexpected
failures during ABI consistency checking for some save and migrate
operations.
This patch adds a new flag to skip the checking of the cur_balloon value
and then sets the destination value to the source value to ensure
subsequent checks without the skip flag will succeed.
This way it is protected from forges and is keeped up to date too.
Signed-off-by: Nikolay Shirokovskiy <nshirokovskiy@virtuozzo.com>
Add support for multi serial devices, after this patch virsh can be used to
connect different serial devices of running domains. E.g.
vish # console <xxx> --devname serial<xxx>
Note:
This depends on a xen/libxl bug fix to have libxl_console_get_tty(...) correctly
returning the tty path (as opposed to always returning the first one).
[0] https://lists.xen.org/archives/html/xen-devel/2016-08/msg00438.html
Signed-off-by: Bob Liu <bob.liu@oracle.com>
libvirt uses the new_id PCI sysfs interface to bind a PCI stub driver
to a PCI device. The new_id interface is known to be buggy and racey,
hence a more deterministic interface was introduced in the 3.12 kernel:
driver_override. For more details see
https://www.redhat.com/archives/libvir-list/2016-June/msg02124.html
For more details about the driver_override interface and examples of
its usage, see
https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/drivers/pci/pci-driver.c?h=v3.12&id=782a985d7af26db39e86070d28f987cad21313c0
This patch adds support for the driver_override interface by
- adding new virPCIDevice{BindTo,UnbindFrom}StubWithOverride functions
that use the driver_override interface
- renames the existing virPCIDevice{BindTo,UnbindFrom}Stub functions
to virPCIDevice{BindTo,UnbindFrom}StubWithNewid to perserve existing
behavior on new_id interface
- changes virPCIDevice{BindTo,UnbindFrom}Stub function to call one of
the above depending on availability of driver_override
The patch includes a bit of duplicate code, but allows for easily
dropping the new_id code once support for older kernels is no
longer desired.
Signed-off-by: Jim Fehlig <jfehlig@suse.com>
libxl only has API to address the host USB devices by bus/device.
Find the bus/device if the user only provided the vendor/product
of the USB device.
Signed-off-by: Jim Fehlig <jfehlig@suse.com>
Finding an USB device from the vendor/device values will be needed
by libxl driver to convert from vendor/device to bus/dev addresses.
Signed-off-by: Jim Fehlig <jfehlig@suse.com>
The 'multi' element in PCI address struct used as 'virTristateSwitch',
and its default value is 'VIR_TRISTATE_SWITCH_ABSENT'. Current PCI
process use 'false' to initialization 'multi', which is ambiguously
for assignment or comparison. This patch use '{0}' to initialize
the whole PCI address struct, which fix the 'multi' initialization
and makes code more simplify and explicitly.
Signed-off-by: Xian Han Yu <xhyubj@linux.vnet.ibm.com>
The libxl driver has long supported migration V3 but has never
indicated so in the connectSupportsFeature API. As a result, apps
such as virt-manager that use the more generic virDomainMigrate API
fail with
libvirtError: this function is not supported by the connection driver:
virDomainMigrate
Add VIR_DRV_FEATURE_MIGRATION_V3 to the list of features marked as
supported in the connectSupportsFeature API.
Test 12 from objecteventtest (createXML add event) segaults on FreeBSD
with bus error.
At some point it calls testNodeDeviceDestroy() from the test driver. And
it fails when it tries to unlock the device in the "out:" label of this
function.
Unlocking fails because the previous step was a call to
virNodeDeviceObjRemove from conf/node_device_conf.c. This function
removes the given device from the device list and cleans up the object,
including destroying of its mutex. However, it does not nullify the pointer
that was given to it.
As a result, we end up in testNodeDeviceDestroy() here:
out:
if (obj)
virNodeDeviceObjUnlock(obj);
And instead of skipping this, we try to do Unlock and fail because of
malformed mutex.
Change virNodeDeviceObjRemove to use double pointer and set pointer to
NULL.
As bhyve currently doesn't use controller addressing and simply
uses 1 implicit controller for 1 disk device, the scheme looks the
following:
pci addrees -> (implicit controller) -> disk device
So in fact we identify disk devices by pci address of implicit
controller and just pass it this way to bhyve in a form:
-s pci_addr,ahci-(cd|hd),/path/to/disk
Therefore, we cannot use virDeviceInfoPCIAddressWanted() because it
does not expect that disk devices might need PCI address assignment.
As a result, if a disk was specified without address, it will not be
generated and domain will to start.
Until proper controller addressing is implemented in the bhyve
driver, force each disk to have PCI address generated if it was not
specified by user.
Setting vcpu count when cpu topology is specified may result into an
invalid configuration. Since the topology can't be modified, reject the
setting if it doesn't match the requested topology. This will allow
fixing the topology in case it was broken.
Partially fixes: https://bugzilla.redhat.com/show_bug.cgi?id=1370066
Validating the vcpu count is more intricate and doing it in the XML
parser will make previously valid configs (with older qemus) vanish.
Now that we have a very similar check in the qemu domain validation
callback we can do it in a more appropriate place.
This basically reverts commit b54de0830a.
Partially resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1370066
ce43cca0e refactored the helper to prepare it for sparse topologies but
forgot to fix the iterator used to fill the structures. This would
result into a weirdly sparse populated array and possible out of bounds
access and crash once sparse vcpu topologies were allowed.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1369988
While dettaching/attaching device in OpenStack, nova
calls vzDomainDettachDevice twice, because the update of the internal
configuration of the ct comes a bit latter than the update event.
As the result, we suffer from the second call to dettach the same device.
Signed-off-by: Olga Krishtal <okrishtal@virtuozzo.com>
libvirt-python passes parameter bandwidth = 0
by default. This means that bandwidth is unlimited.
VZ driver doesn't support bandwidth rate limiting,
but we still need to handle it and fail if bandwidth > 0.
Signed-off-by: Pavel Glushchak <pglushchak@virtuozzo.com>
* Added VIR_MIGRATE_LIVE, VIR_MIGRATE_UNDEFINE_SOURCE and
VIR_MIGRATE_PERSIST_DEST to supported migration flags
Signed-off-by: Pavel Glushchak <pglushchak@virtuozzo.com>
When support for auto-creating tap devices was added to <interface
type='ethernet'> in commit 9c17d6, the code assumed that
virNetDevTapCreate() would honor the VIR_NETDEV_TAP__CREATE_IFUP flag
that is supported by virNetDevTapCreateInBridgePort(). That isn't the
case - the latter function performs several operations, and one of
them is setting the tap device online. But virNetDevTapCreate() *only*
creates the tap device, and relies on the caller to do everything
else, so qemuInterfaceEthernetConnect() needs to call
virNetDevSetOnline() after the device is successfully created.
The linkstate setting of an <interface> is only meant to change the
online status reported to the guest system by the emulated network
device driver in qemu, but when support for auto-creating tap devices
for <interface type='ethernet'> was added in commit 9717d6, a chunk of
code was also added to qemuDomainChangeNetLinkState() that sets the
online status of the tap device (i.e. the *host* side of the
interface) for type='ethernet'. This was never done for tap devices
used in type='bridge' or type='network' interfaces, nor was it done in
the past for tap devices created by external scripts for
type='ethernet', so we shouldn't be doing it now.
This patch removes the bit of code in qemuDomainChangeNetLinkState()
that modifies online status of the tap device.
The call to virNetDevIPInfoAddToDev() that sets up tap device IP
addresses and routes was somehow incorrectly placed in
qemuInterfaceStopDevice() instead of qemuInterfaceStartDevice() in
commit fe8567f6. This fixes that error by moving the call to
virNetDevIPInfoAddToDev() to qemuInterfaceStartDevice().
Signed-off-by: Vasiliy Tolstov <v.tolstov@selfip.ru>
This patch removes the old vcpu unplug code completely and replaces it
with the new code using device_del. The old hotplug code basically never
worked with any recent qemu and thus is useless.
As the new code is using device_del all the implications of using it
are present. Contrary to the device deletion code, the vcpu deletion
code fails if the unplug request is not executed in time.
To allow unplugging the vcpus, hotplugging of vcpus on platforms which
require to plug multiple logical vcpus at once or plugging them in an
arbitrary order it's necessary to use the new device_add interface for
vcpu hotplug.
This patch adds support for the device_add interface using the old
setvcpus API by implementing an algorithm to select the appropriate
entities to plug in.
Add support for using the new approach to hotplug vcpus using device_add
during startup of qemu to allow sparse vcpu topologies.
There are a few limitations imposed by qemu on the supported
configuration:
- vcpu0 needs to be always present and not hotpluggable
- non-hotpluggable cpus need to be ordered at the beginning
- order of the vcpus needs to be unique for every single hotpluggable
entity
Qemu also doesn't really allow to query the information necessary to
start a VM with the vcpus directly on the commandline. Fortunately they
can be hotplugged during startup.
The new hotplug code uses the following approach:
- non-hotpluggable vcpus are counted and put to the -smp option
- qemu is started
- qemu is queried for the necessary information
- the configuration is checked
- the hotpluggable vcpus are hotplugged
- vcpus are started
This patch adds a lot of checking code and enables the support to
specify the individual vcpu element with qemu.
The vcpu order information is extracted only for hotpluggable entities,
while vcpu definitions belonging to the same hotpluggable entity need
to all share the order information.
We also can't overwrite it right away in the vcpu info detection code as
the order is necessary to add the hotpluggable vcpus enabled on boot in
the correct order.
The helper will store the order information in places where we are
certain that it's necessary.
Introduce a new migration cookie flag that will be used for any
configurations that are not compatible with libvirt that would not
support the specific vcpu hotplug approach. This will make sure that old
libvirt does not fail to reproduce the configuration correctly.
Individual vCPU hotplug requires us to track the state of any vCPU. To
allow this add the following XML:
<domain>
...
<vcpu current='2'>3</vcpu>
<vcpus>
<vcpu id='0' enabled='yes' hotpluggable='no' order='1'/>
<vcpu id='1' enabled='yes' hotpluggable='yes' order='2'/>
<vcpu id='1' enabled='no' hotpluggable='yes'/>
</vcpus>
...
The 'enabled' attribute allows to control the state of the vcpu.
'hotpluggable' controls whether given vcpu can be hotplugged and 'order'
allows to specify the order to add the vcpus.
Similarly to devices the guest may allow unplug of the VCPU if libvirt
is down. To avoid problems, refresh the vcpu state on reconnect. Don't
mess with the vcpu state otherwise.
Now that the monitor code gathers all the data we can extract it to
relevant places either in the definition or the private data of a vcpu.
As only thread id is broken for TCG guests we may extract the rest of
the data and just skip assigning of the thread id. In case where qemu
would allow cpu hotplug in TCG mode this will make it work eventually.
For hotplug purposes it's necessary to retrieve data using
query-hotpluggable-cpus while the old query-cpus API report thread IDs
and order of hotplug.
This patch adds code that merges the data using a rather non-trivial
algorithm and fills the data to the qemuMonitorCPUInfo structure for
adding to appropriate place in the domain definition.
Add support for retrieving information regarding hotpluggable cpu units
supported by qemu. Data returned by the command carries information
needed to figure out the granularity of hotplug, the necessary cpu type
name and the topology information.
Note that qemu doesn't specify any particular order of the entries thus
it's necessary sort them by socket_id, core_id and thread_id to the
order libvirt expects.
To allow matching up the data returned by query-cpus to entries in the
query-hotpluggable-cpus reply for CPU hotplug it's necessary to extract
the QOM path as it's the only link between the two.
QEMU reports whether 'query-hotpluggable-cpus' is supported for a given
machine type. Extract and cache the information using the capability
cache.
When copying the capabilities for a new start of qemu, mask out the
presence of QEMU_CAPS_QUERY_HOTPLUGGABLE_CPUS if the machine type
doesn't support hotpluggable cpus.
As of qemu commit:
commit a32ef3bfc12c8d0588f43f74dcc5280885bbdb30
Author: Thomas Huth <thuth@redhat.com>
Date: Wed Jul 22 15:59:50 2015 +0200
vl: Add another sanity check to smp_parse() function
v2.4.0-952-ga32ef3b
configuration where the maximum CPU count doesn't match the topology is
rejected. Prior to that only configurations where the topology would
contain more cpus than the maximum count would be rejected.
Use QEMU_CAPS_QUERY_HOTPLUGGABLE_CPUS as a relevant recent enough
witness to avoid breaking old configs.
Prepare to extract more data by returning an array of structs rather than
just an array of thread ids. Additionally report fatal errors separately
from qemu not being able to produce data.
vzDomainMigrateConfirm3Params is whitelisted. Otherwise we need to
move removing domain from domain list from perform to confirm
step. This would further imply adding a flag and check that migration
is in progress to prohibit mistakenly (maliciously) removing domains
on confirm step. vz version of p2p also need to be fixed to include confirm step.
One would also need to add means to cleanup pending migration
on client disconnect as now is has state across several API
calls.
On the other hand current version of confirm step is totaly
harmless thus it is easier to whitelist it at the moment.
Signed-off-by: Nikolay Shirokovskiy <nshirokovskiy@virtuozzo.com>
Signed-off-by: Maxim Nestratov <mnestratov@virtuozzo.com>
This way we make naming consistent to API calls and make subsequent
ACL checks possible (otherwise ACL check would discover name
discrepancies).
Signed-off-by: Nikolay Shirokovskiy <nshirokovskiy@virtuozzo.com>
ACL check on perform step should be in API call itself to make ACL
checking script pass. Thus we need to reorganize code to obtain
domain object in perform API itself. Most of this is straight
forward, the only nuance is dropping locks on lengthy remote
operations.
The other motivation is to have only perform step ACL checks for
p2p migration instead of both begin in perform if we can leave
ACL check in vzDomainMigratePerformStep.
Signed-off-by: Nikolay Shirokovskiy <nshirokovskiy@virtuozzo.com>
We need it to prepare the calls for ACL checks otherwise ACL checking
script will fail.
Signed-off-by: Nikolay Shirokovskiy <nshirokovskiy@virtuozzo.com>
This action deserves its own function and makes main API call
structure much cleaner.
Signed-off-by: Nikolay Shirokovskiy <nshirokovskiy@virtuozzo.com>
The original motivation is to expand API calls like start/stop etc so that
the ACL checks could be added. But this patch has its own befenits.
1. functions like prlsdkStart/Stop use common routine to wait for
job without domain lock. They become more self contained and do
not return intermediate PRL_RESULT.
2. vzDomainManagedSave do not update cache twice.
Signed-off-by: Nikolay Shirokovskiy <nshirokovskiy@virtuozzo.com>
Signed-off-by: Maxim Nestratov <mnestratov@virtuozzo.com>
