Commit Graph

25553 Commits

Author SHA1 Message Date
Roman Bogorodskiy
a69f3c9b63 docs: document bhyve e1000 support
* Add bhyve e1000 support entry to release notes
 * Update the bhyve driver page with usage sample
2017-02-13 17:25:43 +04:00
Ján Tomko
4a7773f7ea conf: check port range even for USB hubs
Move the range check introduced by commit 2650d5e into
virDomainUSBAddressFindPort. That way both virDomainUSBAddressRelease
and virDomainUSBAddressSetAddHub can benefit from it.

Reported-by: Michal Privoznik <mprivozn@redhat.com>
2017-02-13 13:08:36 +01:00
Ján Tomko
384504f7ba qemu: assign USB port on a selected hub for all devices
Due to a logic error, the autofilling of USB port when a bus is
specified:
    <address type='usb' bus='0'/>
does not work for non-hub devices on domain startup.

Fix the logic in qemuDomainAssignUSBPortsIterator to also
assign ports for USB addresses that do not yet have one.

https://bugzilla.redhat.com/show_bug.cgi?id=1374128
2017-02-13 09:46:15 +01:00
Marc Hartmayer
d6bc7622f0 rpc: Fix potentially segfaults
We have to allocate first and if, and only if, it was successful we
can set the count. A segfault has occurred in
virNetServerServiceNewPostExecRestart() when VIR_ALLOC_N(svc->socks,
n) has failed, but svc->nsocsk = n was already set. Thus
virObejectUnref(svc) was called and therefore it was possible that
virNetServerServiceDispose was called => segmentation fault.  For
safeness NULL pointer check were added in
virNetServerServiceDispose().

Signed-off-by: Marc Hartmayer <mhartmay@linux.vnet.ibm.com>
Reviewed-by: Boris Fiuczynski <fiuczy@linux.vnet.ibm.com>
Reviewed-by: Bjoern Walk <bwalk@linux.vnet.ibm.com>
2017-02-12 15:02:42 -05:00
Roman Bogorodskiy
5620c60959 bhyve: add e1000 nic support
Recently e1000 NIC support was added to bhyve; implement that in
the bhyve driver:

 - Add capability check by analyzing output of the 'bhyve -s 0,e1000'
   command
 - Modify bhyveBuildNetArgStr() to support e1000 and also pass
   virConnectPtr so it could call bhyveDriverGetCaps() to check if this
   NIC is supported
 - Modify command parsing code to add support for e1000 and adjust tests
 - Add net-e1000 test
2017-02-11 06:51:28 +04:00
Nitesh Konkar
ef41eda68a util: Fix indentation for virnetdevmacvlan
Signed-off-by: Nitesh Konkar <nitkon12@linux.vnet.ibm.com>
2017-02-10 14:13:50 -05:00
John Ferlan
4ab0c959e9 libxl: Resolve possible resource leak in dom0 maximum memory setting
If either the "if (STRPREFIX(mem_tokens[j], "max:"))" is never entered
or the "if (virStrToLong_ull(mem_tokens[j] + 4, &p, 10, maxmem) < 0)" break
is hit, control goes back to the outer loop processing 'cmd_tokens' and
it's possible that the 'mem_tokens' would be overwritten.

Found by Coverity
2017-02-10 14:11:04 -05:00
Erik Skultety
b2774db9c2 storage: Fix checking whether source filesystem is mounted
Right now, we use simple string comparison both on the source paths
(mount's output vs pool's source) and the target (mount's mnt_dir vs
pool's target). The problem are symlinks and mount indeed returns
symlinks in its output, e.g. /dev/mappper/lvm_symlink. The same goes for
the pool's source/target, so in order to successfully compare these two
replace plain string comparison with virFileComparePaths which will
resolve all symlinks and canonicalize the paths prior to comparison.

Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1417203

Signed-off-by: Erik Skultety <eskultet@redhat.com>
2017-02-10 17:01:12 +01:00
Erik Skultety
875894245a util: Introduce virFileComparePaths
So rather than comparing 2 paths (strings) as they are, which can very
easily lead to unnecessary errors (e.g. in storage driver) that the paths
are not the same when in fact they'd be e.g. just symlinks to the same
location, we should put our best effort into resolving any symlinks and
canonicalizing the path and only then compare the 2 paths for equality.

Signed-off-by: Erik Skultety <eskultet@redhat.com>
2017-02-10 17:01:12 +01:00
Erik Skultety
8fcf6330b6 storage: Fix reporting an error on an already mounted filesystem
When FS pool's source is already mounted on the target location instead
of just simply marking the pool as active, thus starting it we fail with
an error stating that the source is indeed already mounted on the target.

Signed-off-by: Erik Skultety <eskultet@redhat.com>
2017-02-10 17:01:12 +01:00
John Ferlan
e8a35271d4 tests: Link virusbmock with probes
Seeing similar error to commit id '997be5c27' with the inability
to find the libvirt_event_poll_purge_timeout_semaphore symbol
causing a virusbtest failure.
2017-02-10 10:55:26 -05:00
Boris Fiuczynski
d15b29be25 remote generator: Increase upper limit on lists of node devices
On a system with 697 SCSI disks each configured with 8 paths the command
virsh nodedev-list fails with
error: Failed to list node devices
error: internal error: Too many node_devices '16816' for limit '16384'
Increasing the upper limit on lists of node devices from 16K to 64K.

Signed-off-by: Boris Fiuczynski <fiuczy@linux.vnet.ibm.com>
2017-02-10 16:05:24 +01:00
Nitesh Konkar
f278a148e2 Mention the min duration for nodesuspend explicitly
Although currently this is documented in virsh man page
and virsh help, the expicit mention in the error message
is helful for tools using the API directly.

Signed-off-by: Nitesh Konkar <nitkon12@linux.vnet.ibm.com>
2017-02-10 09:13:30 -05:00
Michal Privoznik
732629dad3 qemuMonitorCPUModelInfoFree: Don't leak model_info->props
==11846== 240 bytes in 1 blocks are definitely lost in loss record 81 of 107
==11846==    at 0x4C2BC75: calloc (vg_replace_malloc.c:624)
==11846==    by 0x18C74242: virAllocN (viralloc.c:191)
==11846==    by 0x4A05E8: qemuMonitorCPUModelInfoCopy (qemu_monitor.c:3677)
==11846==    by 0x446E3C: virQEMUCapsNewCopy (qemu_capabilities.c:2171)
==11846==    by 0x437335: testQemuCapsCopy (qemucapabilitiestest.c:108)
==11846==    by 0x437CD2: virTestRun (testutils.c:180)
==11846==    by 0x437AD8: mymain (qemucapabilitiestest.c:176)
==11846==    by 0x4397B6: virTestMain (testutils.c:992)
==11846==    by 0x437B44: main (qemucapabilitiestest.c:188)

Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
2017-02-10 10:25:44 +01:00
Michal Privoznik
afecfc2bf4 testUSBList: don't leak @dev
==22187== 77 (56 direct, 21 indirect) bytes in 1 blocks are definitely lost in loss record 23 of 37
==22187==    at 0x4C2BC75: calloc (vg_replace_malloc.c:624)
==22187==    by 0x4E75685: virAlloc (viralloc.c:144)
==22187==    by 0x4F0613A: virUSBDeviceNew (virusb.c:332)
==22187==    by 0x4F05BA2: virUSBDeviceSearch (virusb.c:183)
==22187==    by 0x4F05F95: virUSBDeviceFind (virusb.c:296)
==22187==    by 0x403514: testUSBList (virusbtest.c:209)
==22187==    by 0x403BD8: virTestRun (testutils.c:180)
==22187==    by 0x4039E5: mymain (virusbtest.c:285)
==22187==    by 0x4056BC: virTestMain (testutils.c:992)
==22187==    by 0x403A4A: main (virusbtest.c:293)

Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
2017-02-10 10:25:44 +01:00
Michal Privoznik
eb58732c01 virusbmock: Link with libvirt_utils
We are using couple of functions from there (e.g. virStrdup) and
rely that the binary linking us has the libvirt_utils linked
already. Well, this makes valgrind sad.

Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
2017-02-10 10:25:44 +01:00
Michal Privoznik
a9595814ed tests: Trace re-executing processes with valgrind
A lot of our tests re-execute themeselves after loading their
mock library. This, however, makes valgrind sad because currently
we do not tell it to trace the process after exec().

Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
2017-02-10 10:25:44 +01:00
Nitesh Konkar
09a91f0528 Fix indentation in datatypes.h
Signed-off-by: Nitesh Konkar <nitkon12@linux.vnet.ibm.com>
2017-02-09 17:31:41 -05:00
Marc Hartmayer
fd98631cf0 remote generator: handle remoteDomainCreateWithFlags()
This commit removes the handcrafted code for
remoteDomainCreateWithFlags() and lets it auto generate.

A little bit of history repeating...
Commit 03d813bbcd removed the auto generation of
remoteDomainCreateWithFlags() because it was thought that the design
flaw in the remote protocol for virDomainCreate is also within the
remote protocol for virDomainCreateWithFlags. As the commit message of
ddaf15d7a3 mentions this is not the case therefore we
can auto generate the client part.

Even worse there was a typo in remoteDomainCreateWithFlags()

'remote_domain_create_with_flags_args ret;' but in fact it has to be
'remote_domain_create_with_flags_ret ret;'.

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>
2017-02-09 17:21:15 -05:00
Marc Hartmayer
c26fe44be5 util: reset the counters to zero
After freeing the data structures we have to reset the counters to
zero. This fixes a segmentation fault when virNetDevIPInfoClear is
called twice (e.g. this is possible in virDomainNetDefParseXML() if
virDomainNetIPInfoParseXML(...) fails with ret < 0 (this leads to the
first call of 'virNetDevIPInfoClear(&def->guestIP)') and the resulting
call of virDomainNetDefFree(def) in the error path of
virDomainNetDefParseXML() (this leads to the second call of
virNetDevIPInfoClear(&def->guestIP), and finally to the segmentation
fault).

Signed-off-by: Marc Hartmayer <mhartmay@linux.vnet.ibm.com>
Reviewed-by: Boris Fiuczynski <fiuczy@linux.vnet.ibm.com>
Reviewed-by: Bjoern Walk <bwalk@linux.vnet.ibm.com>
2017-02-09 14:20:42 -05:00
Marc Hartmayer
28dd54a5b9 conf: Fix libvirtd free() segfault if virDomainChrSourceDefNew(...) fails
If virDomainChrSourceDefNew(xmlopt) fails, it will lead to free()ing
the uninitialized pointer bus. The fix for this is to initialize bus
with NULL.

Signed-off-by: Marc Hartmayer <mhartmay@linux.vnet.ibm.com>
Reviewed-by: Boris Fiuczynski <fiuczy@linux.vnet.ibm.com>
Reviewed-by: Bjoern Walk <bwalk@linux.vnet.ibm.com>
2017-02-09 14:18:51 -05:00
Marc Hartmayer
62b2c2fcdd qemu: Check if virQEMUCapsNewCopy(...) has failed
Check if virQEMUCapsNewCopy(...) has failed, thus a segmentation fault
in virQEMUCapsFilterByMachineType(...) will be avoided.

Signed-off-by: Marc Hartmayer <mhartmay@linux.vnet.ibm.com>
Reviewed-by: Bjoern Walk <bwalk@linux.vnet.ibm.com>
2017-02-09 14:08:00 -05:00
David Dai
728c0e5df4 qemu: Fix live migration over RDMA with IPv6
Using libvirt to do live migration over RDMA via IPv6 address failed.

For example:
    rhel73_host1_guest1 qemu+ssh://[deba::2222]/system --verbose
root@deba::2222's password:
error: internal error: unable to execute QEMU command 'migrate': RDMA
ERROR: could not rdma_getaddrinfo address deba

As we can see, the IPv6 address used by rdma_getaddrinfo() has only
"deba" part because we didn't properly enclose the IPv6 address in []
and passed rdma:deba::2222:49152 as the migration URI in
qemuMonitorMigrateToHost.

Signed-off-by: David Dai <zdai@linux.vnet.ibm.com>
2017-02-09 19:47:09 +01:00
Jim Fehlig
79692c3874 libxl: fix dom0 maximum memory setting
When the libxl driver is initialized, it creates a virDomainDef
object for dom0 and adds it to the list of domains. Total memory
for dom0 was being set from the max_memkb field of libxl_dominfo
struct retrieved from libxl, but this field can be set to
LIBXL_MEMKB_DEFAULT (~0ULL) if dom0 maximum memory has not been
explicitly set by the user.

This patch adds some simple parsing of the Xen commandline,
looking for a dom0_mem parameter that also specifies a 'max' value.
If not specified, dom0 maximum memory is effectively all physical
host memory.

Signed-off-by: Jim Fehlig <jfehlig@suse.com>
2017-02-09 10:02:19 -07:00
Jim Fehlig
d2b77608e9 libxl: fix reporting of maximum memory
The libxl driver reports different values of maximum memory depending
on state of a domain. If inactive, maximum memory value is reported
correctly. When active, maximum memory is derived from max_pages value
returned by the XEN_SYSCTL_getdomaininfolist sysctl operation. But
max_pages can be changed by toolstacks and does not necessarily
represent the maximum memory a domain can use during its active
lifetime.

A better location for determining a domain's maximum memory is the
/local/domain/<id>/memory/static-max node in xenstore. This value
is set from the libxl_domain_build_info.max_memkb field when creating
the domain. Currently it cannot be changed nor can its value be
exceeded by a balloon operation. From libvirt's perspective, always
reporting maximum memory with virDomainDefGetMemoryTotal() will produce
the same results as reading the static-max node in xenstore.

Signed-off-by: Jim Fehlig <jfehlig@suse.com>
2017-02-09 09:38:34 -07:00
Jim Fehlig
bd1168101a libxl: fix disk detach when <driver> not specified
When a user does not explicitly set a <driver> in the disk config,
libvirt defers selection of a default to libxl. This approach works
fine when starting a domain with such configuration or attaching a
disk to a running domain. But when detaching such a disk, libxl
will fail with "unrecognized disk backend type: 0". libxl makes no
attempt to recalculate a default backend (driver) on detach and
simply fails when uninitialized.

This patch updates the libvirt disk config with the backend selected
by libxl when starting a domain or attaching a disk to a running
domain. Another benefit of this approach is that the live XML is
also updated with the backend driver selected by libxl.
2017-02-09 09:24:44 -07:00
Jim Fehlig
321a28c6ae libxl: set default disk format in device post-parse
When starting a domian, a libxl_domain_config object is created from
virDomainDef. Any virDomainDiskDef devices with a format of
VIR_STORAGE_FILE_NONE are mapped to LIBXL_DISK_FORMAT_RAW in the
corresponding libxl_disk_device, but the virDomainDiskDef format is
never updated to reflect the change.

A better place to set a default format for disk devices is the
device post-parse callback, ensuring the virDomainDiskDef object
reflects the default format.
2017-02-09 09:24:44 -07:00
Boris Fiuczynski
c012052077 docs: Document new ovs_timeout configuration knob
Document the ovs_timeout introduced to libvirtd.conf in previous
patches.

Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
2017-02-09 17:22:31 +01:00
Boris Fiuczynski
4e036ab0c2 libvirtd: set openvswitch timeout value based on config data
Since a successful completion of the calls to openvswitch is expected
a longer timeout should be able to be chosen to account for loaded systems.
Therefore this patch provides the ability to specify the timeout value for
openvswitch calls in the libvirtd configuration file.

Signed-off-by: Boris Fiuczynski <fiuczy@linux.vnet.ibm.com>
Reviewed-by: Bjoern Walk <bwalk@linux.vnet.ibm.com>
2017-02-09 14:34:08 +01:00
Boris Fiuczynski
f4d06ca8fd network: allow to specify timeout for openvswitch calls
This patchs allows to set the timeout value used for all
openvswitch calls. The default timeout value remains as
before at 5 seconds.

Signed-off-by: Boris Fiuczynski <fiuczy@linux.vnet.ibm.com>
Reviewed-by: Bjoern Walk <bwalk@linux.vnet.ibm.com>
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
2017-02-09 14:34:08 +01:00
Boris Fiuczynski
66583c0cf7 libvirtd: add openvitch timeout value
Provide the ability to specify a default timeout value for
successful completion of openvswitch calls in the libvirtd
configuration file.

Signed-off-by: Boris Fiuczynski <fiuczy@linux.vnet.ibm.com>
Reviewed-by: Bjoern Walk <bwalk@linux.vnet.ibm.com>
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
2017-02-09 14:34:08 +01:00
Jaroslav Safka
0857a3bf5c docs: Document new <memoryBacking/> elements
Document the extensions made to domain <memoryBacking/> element
made in previous commits.

Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
2017-02-09 14:27:19 +01:00
Jaroslav Safka
1c4f3b56f8 qemu: Add args generation for file memory backing
This patch add support for file memory backing on numa topology.

The specified access mode in memoryBacking can be overriden
by specifying token memAccess in numa cell.
2017-02-09 14:27:19 +01:00
Jaroslav Safka
bc6d3121a4 conf: Add new xml elements for file memorybacking support
This part introduces new xml elements for file based
memorybacking support and their parsing.
(It allows vhost-user to be used without hugepages.)

New xml elements:
<memoryBacking>
  <source type="file|anonymous"/>
  <access mode="shared|private"/>
  <allocation mode="immediate|ondemand"/>
</memoryBacking>
2017-02-09 14:27:19 +01:00
Jaroslav Safka
48d9e6cdcc qemu_conf: Add param memory_backing_dir
Add new parameter memory_backing_dir where files will be stored when memoryBacking
source is selected as file.

Value is stored inside char* memoryBackingDir
2017-02-09 14:27:19 +01:00
Jaroslav Safka
7c0c5f6d4b qemu, conf: Rename virNumaMemAccess to virDomainMemoryAccess
Rename to avoid duplicate code. Because virDomainMemoryAccess will be
used in memorybacking for setting default behaviour.

NOTE: The enum cannot be moved to qemu/domain_conf because of headers
dependency
2017-02-09 14:27:19 +01:00
Maxim Nestratov
c3ee75e5aa cpu: fix typo: rename __kvm_hv_spinlock to __kvm_hv_spinlocks
Strings associated with virDomainHyperv values in domain_conf.c are used to
construct HyperV CPU features names to be compared with names defined in
cpu_x86_data.h and the names for HyperV "spinlocks" feature don't match.
This leads to a misleading warning:
"host doesn't support hyperv 'spinlocks' feature" even when it's supported.
Let's fix it and rename along with it VIR_CPU_x86_KVM_HV_SPINLOCK to
VIR_CPU_x86_KVM_HV_SPINLOCKS.

Signed-off-by: Maxim Nestratov <mnestratov@virtuozzo.com>
2017-02-09 13:52:16 +01:00
Jiri Denemark
644804765b qemu_command: Fix check for gluster disks
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
2017-02-09 11:48:10 +01:00
Jiri Denemark
2cc317b1f5 qemu_blockjob: Avoid dereferencing NULL on OOM
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
2017-02-09 11:48:10 +01:00
Jiri Denemark
b97839b835 cpu_x86: Fix memory leak in virCPUx86Translate
virCPUDefStealModel is called with keepVendor == true which means the
cpu structure will keep its original vendor/vendor_id values. Thus it
makes no sense to copy them to the translated definition as they won't
be used there anyway. Except that the translated->vendor pointer might
get lost in x86Decode.

Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
2017-02-09 11:48:10 +01:00
Maxim Nestratov
eda4ec94ff vz: cleanup: remove unused constant
PARALLELS_STATISTICS_DROP_COUNT isn't used anymore

Signed-off-by: Maxim Nestratov <mnestratov@virtuozzo.com>
2017-02-09 13:06:15 +03:00
Maxim Nestratov
c52f5bea0d vz: fix event handle leak in prlsdkHandlePerfEvent
When we happen to lose a domain but still get a performance event
for it, we should also free the event handle.

Signed-off-by: Maxim Nestratov <mnestratov@virtuozzo.com>
2017-02-09 13:06:15 +03:00
Maxim Nestratov
05456cc97e vz: fix handle leak in prlsdkHandleVmStateEvent
Every successful call of PrlEvent_GetParamByName allocates a handle,
which has to be freed.

Signed-off-by: Maxim Nestratov <mnestratov@virtuozzo.com>
2017-02-09 13:06:15 +03:00
Michal Privoznik
c2130c0d47 qemu_security: Introduce ImageLabel APIs
Just like we need wrappers over other virSecurityManager APIs, we
need one for virSecurityManagerSetImageLabel and
virSecurityManagerRestoreImageLabel. Otherwise we might end up
relabelling device in wrong namespace.

Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
2017-02-09 08:04:57 +01:00
Jim Fehlig
c89a6e7878 libxl: use init and dispose functions with libxl_physinfo
The typical pattern when calling libxl functions that populate a
structure is

  libxl_foo foo;
  libxl_foo_init(&foo);
  libxl_get_foo(ctx, &foo);
  ...
  libxl_foo_dispose(&foo);

Fix several instances of libxl_physinfo missing the init and
dispose calls.

Signed-off-by: Jim Fehlig <jfehlig@suse.com>
2017-02-08 09:12:00 -07:00
Jim Fehlig
ff225538d4 libxl: honor autoballoon setting in libxl.conf
libxlGetAutoballoonConf is supposed to honor user-specified
autoballoon setting in libxl.conf. As written, the user-specified
setting could be overwritten by the subsequent logic to check
dom0_mem parameter. If user-specified setting is present and
correct, accept it. Only fallback to checking Xen dom0_mem
command line parameter if user-specfied setting is not present.

Signed-off-by: Jim Fehlig <jfehlig@suse.com>
2017-02-08 09:10:49 -07:00
Joao Martins
91ac80a986 xenconfig: fix xml to xl.cfg conversion with no graphics
If no graphics element is in XML xenFormatXLSpice will access
graphics without checking it has one in the first place, leading to a
segmentation fault.

Signed-off-by: Joao Martins <joao.m.martins@oracle.com>
2017-02-08 08:41:13 -07:00
Michal Privoznik
b7feabbfdc qemuDomainNamespaceSetupDisk: Simplify disk check
Firstly, instead of checking for next->path the
virStorageSourceIsEmpty() function should be used which also
takes disk type into account.
Secondly, not every disk source passed has the correct type set
(due to our laziness). Therefore, instead of checking for
virStorageSourceIsBlockLocal() and also S_ISBLK() the former can
be refined to just virStorageSourceIsLocalStorage().

Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
2017-02-08 15:56:21 +01:00
Michal Privoznik
786d8d91b4 qemuDomainDiskChainElement{Prepare,Revoke}: manage /dev entry
Again, one missed bit. This time without this commit there is no
/dev entry  in the namespace of the qemu process when doing disk
snapshots or block-copy.

Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
2017-02-08 15:56:13 +01:00
Michal Privoznik
18ce9d139d qemuDomainNamespace{Setup,Teardown}Disk: Don't pass pointer to full disk
These functions do not need to see the whole virDomainDiskDef.
Moreover, they are going to be called from places where we don't
have access to the full disk definition. Sticking with
virStorageSource is more than enough.

Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
2017-02-08 15:56:05 +01:00