Commit Graph

26521 Commits

Author SHA1 Message Date
John Ferlan
82769c4fdc nwfilter: Use consistent naming for variables
When processing a virNWFilterPtr use 'nwfilter' as a variable name.

When processing a virNWFilterObjPtr use 'obj' as a variable name.

Signed-off-by: John Ferlan <jferlan@redhat.com>
2017-04-26 13:13:18 -04:00
Cédric Bosdonnat
b63de148a4 IPv6 route check: list devices only once
If several RA routes are found for the same device, only list that
device once in the error message.
2017-04-26 18:59:24 +02:00
Ján Tomko
8d704e6cf7 Use a separate buffer for <filesystem><driver>
Format the attributes in a separate buffer and only
print the element if it's not empty.
2017-04-26 16:29:38 +02:00
Ján Tomko
5e5fc76625 Use a separate buffer for <controller><driver>
Make adding new attributes easier.
2017-04-26 16:29:38 +02:00
Ján Tomko
d89803d27c Use a separate buffer for <disk><driver>
Eliminate the big condition at the start.
Instead use a buffer and only format the element if the buffer
is non-empty.
2017-04-26 16:29:38 +02:00
Ján Tomko
51219e11b8 Use a separate buffer for <input> subelements
Instead of figuring out upfront whether <input> will be a single
or a pair element, format the subelements into a separate buffer
and close <input/> early if this buffer is empty.
2017-04-26 16:29:38 +02:00
Erik Skultety
a94d431dc4 docs: Provide a nodedev driver stub documentation
There's lot more to document about the nodedev driver, besides PCI and
SR-IOV (even this might need to be extended), but let's start small-ish
and at least have a page for it linked from the drivers.html.

Signed-off-by: Erik Skultety <eskultet@redhat.com>
2017-04-26 15:56:11 +02:00
John Ferlan
3c4f2e3fb7 disk: Use virStorageBackendZeroPartitionTable
https://bugzilla.redhat.com/show_bug.cgi?id=1439132

During 'matrix' testing of all possible combinations I found that if
device is formated with "gpt" first, then an attempt is made to format
using "mac", a startup will fail.

Deeper analysis by Peter Krempa indicates that the "mac" table fits
into the first block on the disk. Since the GPT disklabel is stored
at LBA address 1 it is not overwritten at all. Thus it's apparent that
the (blkid) detection tool then prefers GPT over a older disklabel.

The GPT disklabel has also a secondary copy at the last LBA of the disk.

So, follow the same logic as the logical pool in clearing a 1MB swath
at the beginning and end of the device to avoid potential issues with
larger sector sizes for the device.

Also fixed a minor formatting nit in virStorageBackendDeviceIsEmpty call.
2017-04-26 07:28:08 -04:00
John Ferlan
d942bf6e9e logical: Increase the size of the data to wipe
Since a sector size may be larger than 512 bytes, let's just increase
the size to wipe to 1MB rather than 2KB
2017-04-26 07:28:08 -04:00
John Ferlan
c6aa81c65a logical: Use virStorageBackendZeroPartitionTable
Rather than open code it, use the new function which uses the wipe algorithm
in order to zero the front and tail of the partition.
2017-04-26 07:28:08 -04:00
John Ferlan
e8b0212458 storage: Introduce virStorageBackendZeroPartitionTable
Create a wrapper/helper that can be used to call the storage backend
wipe helper - storageBackendVolWipeLocalFile for future use by logical
and disk backends to clear out the partition table rather than having
each open code the same algorithm.
2017-04-26 07:28:08 -04:00
John Ferlan
859a2d162a storage: Modify storageBackendWipeLocal to allow zero from end of device
Add bool 'zero_end' and logic that would allow a caller to wipe specific
portions of a target device either from the beginning (the default) or
from the end when zero_end is true.

This will allow for this code to wipe out partition table information
from a device.
2017-04-26 07:28:08 -04:00
Daniel P. Berrange
1acce5c853 Update keycodemapdb submodule for python2 compat fixes
There were a few bugs in keycodemap tool that broke it
when run on python circa 2.7.5 or older, which affected
RHEL builds.

Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2017-04-26 11:53:40 +01:00
Daniel P. Berrange
e6c3b59c19 Add ability to generate man page describing key code names & values
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2017-04-25 21:20:43 +01:00
Daniel P. Berrange
02fb15fb60 util: switch over to use keycodemapdb GIT submodule
A long time ago we imported the keymaps.csv file from GTK-VNC so we
can do conversions between keycode sets. Meanwhile lots of bug fixes
have gone into this CSV file and libvirt hasn't kept in sync. The
keymaps.csv file and associated generator script has been pulled out
of GTK-VNC into a dedicated GIT repo for use as a submodule. This
allows GTK-VNC, SPICE-GTK, QEMU and libvirt to share the same master
database and tools and pushing updates merely requires a submodule
commit update as with gnulib.

The test suite is updated to cover some extra boundary conditions.

Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2017-04-25 21:14:18 +01:00
John Ferlan
ab54d5f152 secret: Clean up virSecretObjListExport logic
Shorten the time needed to keep the list lock and alter the cleanup
path to be more of an error path.

Utilize the the virObjectListFree function to handle the calls for
virObjectUnref on each list element and the VIR_FREE of the list
instead of open coding it.

Change the name of the virHashForEach callback to match the name
of the Export function with the Callback added onto it.

Signed-off-by: John Ferlan <jferlan@redhat.com>
2017-04-25 15:33:22 -04:00
John Ferlan
891d0a76b5 secret: Have virSecretObjNew return locked object
Rather than have caller need to do it, have the object returned locked.

Signed-off-by: John Ferlan <jferlan@redhat.com>
2017-04-25 15:33:22 -04:00
John Ferlan
c520f3f160 secret: Convert virsecretobjs.h to use "newer" formatting style
Alter the prototypes to use the newer formatting style

Signed-off-by: John Ferlan <jferlan@redhat.com>
2017-04-25 15:33:22 -04:00
John Ferlan
74face42b6 secret: Need to set data->error on VIR_ALLOC_N failure
Commit id 'bb1fba629' neglected to set when creating the function.

Signed-off-by: John Ferlan <jferlan@redhat.com>
2017-04-25 15:33:22 -04:00
John Ferlan
6b5ac3c71d test: Remove unnecessary unlocks in cleanup paths
Commit id '865f479da' altered the logic to use a common test*ObjFindByName
helpers which would lock/unlock the test driver; however, a few cleanup paths
in that cleanup missed removing the Unlock, so remove it now.

Signed-off-by: John Ferlan <jferlan@redhat.com>
2017-04-25 08:16:00 -04:00
John Ferlan
fe8f1c8b86 remote: Fix possible use-after-free when sending event message
Based upon an idea and some research by Wang King <king.wang@huawei.com>
and xinhua.Cao <caoxinhua@huawei.com>.

Since we're assigning the 'client' to our callback event lookaside list,
it's imperative that we grab a reference to the object; otherwise, when
the object is unref'd during virNetServerProcessClients when it's determined
that the virNetServerClientIsClosed and the memory is free'd before perhaps
the object event state callbacks are run.  When a virObjectLock() is run,
before sending the message the following trace occurs;

    #0  0x00007fda223d66d8 in virClassIsDerivedFrom
        (klass=0xdeadbeef, parent=0x7fda24c81b40)
         at util/virobject.c:169
    #1  0x00007fda223d6a1e in virObjectIsClass
        (anyobj=anyobj@entry=0x7fd9e575b400, klass=<optimized out>)
         at util/virobject.c:365
    #2  0x00007fda223d6a44 in virObjectLock
        (anyobj=0x7fd9e575b400)
        at util/virobject.c:317
    #3  0x00007fda22507f71 in virNetServerClientSendMessage
        (client=client@entry=0x7fd9e575b400, msg=msg@entry=0x7fd9ec30de90)
        at rpc/virnetserverclient.c:1422
    #4  0x00007fda230d714d in remoteDispatchObjectEventSend
        (client=0x7fd9e575b400, program=0x7fda24c844e0, procnr=348,
         proc=0x7fda2310e5e0 <xdr_remote_domain_event_callback_tunable_msg>,
         data=0x7ffc3857fdb0)
        at remote.c:3803
    #5  0x00007fda230dd71b in remoteRelayDomainEventTunable
        (conn=<optimized out>, dom=0x7fda27cd7660, params=0x7fda27f3aae0,
         nparams=1,opaque=0x7fd9e6c99e00)
        at remote.c:1033
    #6  0x00007fda224484cb in virDomainEventDispatchDefaultFunc
        (conn=0x7fda27cd0120, event=0x7fda2736ea00, cb=0x7fda230dd610
         <remoteRelayDomainEventTunable>, cbopaque=0x7fd9e6c99e00)
        at conf/domain_event.c:1910
    #7  0x00007fda22446871 in virObjectEventStateDispatchCallbacks
        (callbacks=<optimized out>, callbacks=<optimized out>,
         event=0x7fda2736ea00,state=0x7fda24ca3960)
        at conf/object_event.c:722
    #8  virObjectEventStateQueueDispatch
        (callbacks=0x7fda24c65800, queue=0x7ffc3857fe90, state=0x7fda24ca3960)
        at conf/object_event.c:736
    #9  virObjectEventStateFlush (state=0x7fda24ca3960)
        at conf/object_event.c:814
    #10 virObjectEventTimer (timer=<optimized out>, opaque=0x7fda24ca3960)
        at conf/object_event.c:560
    #11 0x00007fda223ae8b9 in virEventPollDispatchTimeouts ()
        at util/vireventpoll.c:458
    #12 virEventPollRunOnce ()
        at util/vireventpoll.c:654
    #13 0x00007fda223ad1d2 in virEventRunDefaultImpl ()
        at util/virevent.c:314
    #14 0x00007fda225046cd in virNetDaemonRun (dmn=0x7fda24c775c0)
        at rpc/virnetdaemon.c:818
    #15 0x00007fda230d6351 in main (argc=<optimized out>, argv=<optimized out>)
        at libvirtd.c:1623

Signed-off-by: John Ferlan <jferlan@redhat.com>
2017-04-25 07:26:36 -04:00
John Ferlan
2033e8cc11 daemon: Rework remoteClientFreeFunc cleanup loops into C macro
Rather than 'n' repetitive code segments, let's create a single macro
which will make the code easier to read.

Signed-off-by: John Ferlan <jferlan@redhat.com>
2017-04-25 07:26:36 -04:00
Wang King
0023cc039d daemon: Remove unnecessary goto error
Freeing the dst is unnecessary if the VIR_STRDUP fails, and therefore
we need to remove the error label as well.
2017-04-25 11:30:41 +02:00
Wang King
f69340472c daemon: Fix domain name leak in error path
Domain name is duplicated in make_nonnull_domain, but not freed when
virTypedParamsSerialize returns error.
2017-04-25 11:30:26 +02:00
ZhiPeng Lu
c77bc47f43 qemu: fix argument of virDomainNetGetActualDirectMode
it should be a comparison of modes between new and old devices. So
the argument of the second virDomainNetGetActualDirectMode should be
newdev.

Signed-off-by: ZhiPeng Lu <lu.zhipeng@zte.com.cn>
2017-04-25 10:12:31 +02:00
Andrea Bolognani
6d53255377 autogen.sh: Improve and generalize
The goal is twofold: firstly, we want to extend the script so
that it can deal with more than a single git submodule, and
secondly we'd like to reduce the amount of duplicated code.
Moreover, since we're making heavy changes to the code anyway,
we might as well make sure it follows a somewhat consistent
coding style too.

To reduce code duplication, we introduce a new --dry-run
option, which can be used by third parties to figure out
whether calling autogen.sh is necessary or not: this allows
us to get rid of the reimplementation of part of the logic
in cfg.mk and guarantee they'll never get out of sync.

Other changes include: making dirty submodules checking and
cleaning entirely independent of other operations; removing
the use of 'set -e' and handling errors explicitly instead;
better parsing of command line arguments.
2017-04-25 09:52:37 +02:00
Yuri Chornoivan
5efa7f2a4b Fix minor typos 2017-04-24 14:40:00 +02:00
Roman Bogorodskiy
5e010605d1 util: relax virNetDevSetCoalesce() stub
Currently, virNetDevSetCoalesce() stub is always returning error. As
it's used by virNetDevTapCreateInBridgePort(), it essentially breaks
bridged networking if coalesce is not supported.

To make it work, relax the stub to trigger error only when its
coalesce argument is not NULL, otherwise report success.
2017-04-24 15:57:54 +04:00
Erik Skultety
9c25d7a4ca docs: Utilize our XSLT list generating template more
Since we do have this template at hand, why not using it wherever
possible (list of supported pool types and remote access section).
Also, perform some stylistic micro adjustments.

Signed-off-by: Erik Skultety <eskultet@redhat.com>
2017-04-24 12:37:37 +02:00
Erik Skultety
9347652142 nodedev: udevProcessPCI: Drop syspath variable
Since we have that information provided by @def which is not a private
object, there is really no need for the variable.

Signed-off-by: Erik Skultety <eskultet@redhat.com>
2017-04-24 12:37:37 +02:00
Erik Skultety
bfaaaf108d conf: nodedev: Split virNodeDeviceDefFormat into more functions
Make the code look cleaner by moving the capability specific bits into
separate functions.

Signed-off-by: Erik Skultety <eskultet@redhat.com>
2017-04-24 12:37:37 +02:00
Erik Skultety
364c912cf3 nodedev: Make use of the compile-time missing enum in switch error
So udevGetDeviceDetails was one those functions using an enum in a
switch, but since it had a 'default' case, compiler didn't warn about an
unhandled enum. Moreover, the error about an unsupported device type
reported in the default case is unnecessary, since by the time we get
there, udevGetDeviceType (which was called before) already made sure
that any unrecognized device types had been handled properly.

Signed-off-by: Erik Skultety <eskultet@redhat.com>
2017-04-24 12:37:37 +02:00
Yi Wang
ab5bb6f346 rpc: fix keep alive timer segfault
ka maybe have been freeed in virObjectUnref, application using
virKeepAliveTimer will segfault when unlock ka. We should keep
ka's refs positive before using it.

#0  0x00007fd8f79970e8 in virClassIsDerivedFrom (klass=0xdeadbeef, parent=0x7fd8e8001b80) at util/virobject.c:169
#1  0x00007fd8f799742e in virObjectIsClass (anyobj=anyobj entry=0x7fd8e800b9c0, klass=<optimized out>) at util/virobject.c:365
#2  0x00007fd8f79974e4 in virObjectUnlock (anyobj=0x7fd8e800b9c0) at util/virobject.c:338
#3  0x00007fd8f7ac477e in virKeepAliveTimer (timer=<optimized out>, opaque=0x7fd8e800b9c0) at rpc/virkeepalive.c:177
#4  0x00007fd8f7e5c9cf in libvirt_virEventInvokeTimeoutCallback () from /usr/lib64/python2.7/site-packages/libvirtmod.so
#5  0x00007fd8ff64db94 in PyEval_EvalFrameEx () from /lib64/libpython2.7.so.1.0
#6  0x00007fd8ff64f1ad in PyEval_EvalCodeEx () from /lib64/libpython2.7.so.1.0
#7  0x00007fd8ff64d85f in PyEval_EvalFrameEx () from /lib64/libpython2.7.so.1.0
#8  0x00007fd8ff64d950 in PyEval_EvalFrameEx () from /lib64/libpython2.7.so.1.0
#9  0x00007fd8ff64d950 in PyEval_EvalFrameEx () from /lib64/libpython2.7.so.1.0
#10 0x00007fd8ff64f1ad in PyEval_EvalCodeEx () from /lib64/libpython2.7.so.1.0
#11 0x00007fd8ff5dc098 in function_call () from /lib64/libpython2.7.so.1.0
#12 0x00007fd8ff5b7073 in PyObject_Call () from /lib64/libpython2.7.so.1.0
#13 0x00007fd8ff5c6085 in instancemethod_call () from /lib64/libpython2.7.so.1.0
#14 0x00007fd8ff5b7073 in PyObject_Call () from /lib64/libpython2.7.so.1.0
#15 0x00007fd8ff648ff7 in PyEval_CallObjectWithKeywords () from /lib64/libpython2.7.so.1.0
#16 0x00007fd8ff67d7e2 in t_bootstrap () from /lib64/libpython2.7.so.1.0
#17 0x00007fd8ff358df3 in start_thread () from /lib64/libpthread.so.0
#18 0x00007fd8fe97d3ed in clone () from /lib64/libc.so.6

Signed-off-by: Yi Wang <wang.yi59@zte.com.cn>
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
2017-04-24 12:19:13 +02:00
Martin Kletzander
b2763f189c Remove pointless check for !ret in virDomainNetDefCoalesceParseXML
It was left there after removing a macro it was part of in first
version or so.  Now it will always be NULL.

Signed-off-by: Martin Kletzander <mkletzan@redhat.com>
2017-04-22 18:36:07 +02:00
Martin Kletzander
f08e26a19a Don't leak str in virDomainNetDefCoalesceParseXML
Signed-off-by: Martin Kletzander <mkletzan@redhat.com>
2017-04-22 18:33:21 +02:00
Laine Stump
997134fb8b util: allow ignoring SIOCSIFHWADDR when errno is EPERM
Commit f4ef3a71 made a variation of virNetDevSetMAC that would return
without logging an error message if errno was set to
EADDRNOTAVAIL. This errno is set by some SRIOV VF drivers (in
particular igbvf) when they fail to set the device's MAC address due
to the PF driver refusing the request. This is useful if we want to
try a different method of setting the VF MAC address before giving up
(Commit 86556e16 actually does this, setting the desired MAC address
to the "admin MAC in the PF, then detaching and reattaching the VF
netdev driver to force a reinit of the MAC address).

During testing of Bug 1442040 t was discovered that the ixgbe driver
returns EPERM in this situation, so this patch changes the exception
case for silent+non-terminal failure to account for this difference.

Completes resolution to: https://bugzilla.redhat.com/1415609 (RHEL 7.4)
                         https://bugzilla.redhat.com/1442040 (RHEL 7.3.z)
2017-04-21 10:37:13 -04:00
Daniel P. Berrange
e6625ed410 util: fix virNetDevSetCoalesce fallback on Win32/FreeBSD
The current fallback stub for virNetDevSetCoalesce is inside an
earlier conditional block. This deals with the feature being
missing on older Linux platforms. We need a second fallback stub
though, outside the top level conditional, to ensure builds work
on Win32/FreeBSD platforms too.

Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2017-04-21 14:57:31 +01:00
Daniel P. Berrange
dfda3a8ed9 docs: fix typo in closing HTML element
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2017-04-21 13:07:17 +01:00
Martin Kletzander
fcef44728d Set coalesce settings for domain interfaces
This patch makes use of the virNetDevSetCoalesce() function to make
appropriate settings effective for devices that support them.

Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1414627

Signed-off-by: Martin Kletzander <mkletzan@redhat.com>
2017-04-21 13:35:04 +02:00
Martin Kletzander
523c996062 conf, docs: Add support for coalesce setting(s)
We are currently parsing only rx/frames/max because that's the only
value that makes sense for us.  The tun device just added support for
this one and the others are only supported by hardware devices which
we don't need to worry about as the only way we'd pass those to the
domain is using <hostdev/> or <interface type='hostdev'/>.  And in
those cases the guest can modify the settings itself.

Signed-off-by: Martin Kletzander <mkletzan@redhat.com>
2017-04-21 13:34:41 +02:00
Martin Kletzander
652ef9bc8c util: Add virNetDevSetCoalesce function
That function is able to configure coalesce settings for an interface,
similarly to 'ethtool -C'.  This function also updates back the
structure so that it contains actual data on the device (if the device
doesn't support some settings kernel might just return 0 and not set
whatever is not supported), so this way we'll have up-to-date
information in the live domain XML.

Signed-off-by: Martin Kletzander <mkletzan@redhat.com>
2017-04-21 13:29:39 +02:00
Pavel Hrdina
42000bf7e5 util: check ifa_addr pointer before accessing its elements
Reported by Rafał Wojciechowski <it@rafalwojciechowski.pl>.

Thread 1 (Thread 0x7f194b99d700 (LWP 5631)):
0  virNetDevGetifaddrsAddress (addr=0x7f194b99c7c0, ifname=0x7f193400e2b0 "ovirtmgmt") at util/virnetdevip.c:738
1  virNetDevIPAddrGet (ifname=0x7f193400e2b0 "ovirtmgmt", addr=addr@entry=0x7f194b99c7c0) at util/virnetdevip.c:795
2  0x00007f19467800d6 in networkGetNetworkAddress (netname=<optimized out>, netaddr=netaddr@entry=0x7f1924013f18) at network/bridge_driver.c:4780
3  0x00007f193e43a33c in qemuProcessGraphicsSetupNetworkAddress (listenAddr=0x7f19340f7650 "127.0.0.1", glisten=0x7f1924013f10) at qemu/qemu_process.c:4062
4  qemuProcessGraphicsSetupListen (vm=<optimized out>, graphics=0x7f1924014f10, cfg=0x7f1934119f00) at qemu/qemu_process.c:4133
5  qemuProcessSetupGraphics (flags=17, vm=0x7f19240155d0, driver=0x7f193411f1d0) at qemu/qemu_process.c:4196
6  qemuProcessPrepareDomain (conn=conn@entry=0x7f192c00ab50, driver=driver@entry=0x7f193411f1d0, vm=vm@entry=0x7f19240155d0, flags=flags@entry=17) at qemu/qemu_process.c:4969
7  0x00007f193e4417c0 in qemuProcessStart (conn=conn@entry=0x7f192c00ab50, driver=driver@entry=0x7f193411f1d0, vm=0x7f19240155d0,asyncJob=asyncJob@entry=QEMU_ASYNC_JOB_START, migrateFrom=migrateFrom@entry=0x0, migrateFd=migrateFd@entry=-1, migratePath=migratePath@entry=0x0,snapshot=snapshot@entry=0x0, vmop=vmop@entry=VIR_NETDEV_VPORT_PROFILE_OP_CREATE, flags=17, flags@entry=1) at qemu/qemu_process.c:5553

Man page for getifaddrs also states that the "ifa_addr" may contain
a null pointer which happens if there is an existing network interface
on the host without IP address.

Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
2017-04-21 11:05:18 +02:00
Jim Fehlig
6b3fb86b39 Increase default task limit for libvirtd
libvirtd can spawn threads/tasks when creating new domains for
some hypervisors such as Xen's libxl driver, quickly reaching
the cgroups pids controller default TasksMax setting of 512. When
the limit is reached, attempting to create additional domains
results in an error from the cgroups pids controller, e.g.

kernel: [71282.213347] cgroup: fork rejected by pids controller in
/system.slice/libvirtd.service

Depending on domain type and configuration, anywhere from 4-7
threads/tasks may be created by libxl when starting a domain.
In order to support 4096 domains, similar to commit 27cd763500,
increase the TasksMax setting in libvirtd.service to
4096 * 8 = 32768 tasks.
2017-04-20 09:13:34 -06:00
Daniel P. Berrange
234ac4e18d Fix error reporting when poll returns POLLHUP/POLLERR
In the RPC client event loop code, if poll() returns only a POLLHUP
or POLLERR status, then we end up reporting a bogus error message:

  error: failed to connect to the hypervisor
  error: An error occurred, but the cause is unknown

We do actually report an error, but we virNetClientMarkClose method
has already captured the error status before we report it, so the
real error gets thrown away. The key fix is to report the error
before calling virNetClientMarkClose(). In changing this, we also
split out reporting of POLLHUP vs POLLERR to make any future bugs
easier to diagnose.

Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2017-04-20 10:15:46 +01:00
Peter Krempa
355f5ab998 qemu: hotplug: Don't save status XML when monitor is closed
In the vcpu hotplug code if exit from the monitor failed we would still
attempt to save the status XML. When the daemon is terminated the
monitor socket is closed. In such case, the written status XML would not
contain the monitor path and thus be invalid.

Avoid this issue by only saving status XML on success of the monitor
command.

Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1439452
2017-04-20 10:46:44 +02:00
Peter Krempa
f24dc5e2c2 qemu: hotplug: Unexport qemuDomainHotplugDelVcpu
The function is used only in the hotplug module.
2017-04-20 10:46:44 +02:00
Pavel Hrdina
90acbc76ec qemu_domain: use correct default USB controller on ppc64
The history of USB controller for ppc64 guest is complex and goes
back to libvirt 1.3.1 where the fun started.

Prior Libvirt 1.3.1 if no model for USB controller was specified
we've simply passed "-usb" on QEMU command line.

Since Libvirt 1.3.1 there is a patch (8156493d8d) that fixes this
issue by using "-device pci-ohci,..." but it breaks migration with
older Libvirts which was agreed that's acceptable.  However this
patch didn't reflect this change in the domain XML and the model
was still missing.

Since Libvirt 2.2.0 there is a patch (f55eaccb0c) that fixes the
issue with not setting the USB model into domain XML which we need
to know about to not break the migration and since the default
model was *pci-ohci* it was used as default in this patch as well.

This patch tries to take all the previous changes into account and
also change the default for newly defined domains that don't specify
any model for USB controller.

The VIR_DOMAIN_DEF_PARSE_ABI_UPDATE is set only if new domain is
defined or new device is added into a domain which means that in
all other cases we will use the old *pci-ohci* model instead of the
better and not broken *nec-usb-xhci* model.

Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1373184

Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
2017-04-20 09:03:53 +02:00
Pavel Hrdina
5c7d88085a conf: add a new parse flag VIR_DOMAIN_DEF_PARSE_ABI_UPDATE_MIGRATION
So far there is probably no change that is allowed to be done
by the VIR_DOMAIN_DEF_PARSE_ABI_UPDATE flag that would break
guest ABI but this may change in the future.

This introduces new VIR_DOMAIN_DEF_PARSE_ABI_UPDATE_MIGRATION
which should be used only for ABI updates that are "safe" for
persistent migration.

Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
2017-04-20 09:03:53 +02:00
Daniel P. Berrange
dd81f56842 Refresh translations from zanata
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2017-04-19 16:58:12 +01:00
Jiri Denemark
5b4a6adb5c qemu: Use more data for comparing CPUs
With QEMU older than 2.9.0 libvirt uses CPUID instruction to determine
what CPU features are supported on the host. This was later used when
checking compatibility of guest CPUs. Since QEMU 2.9.0 we ask QEMU for
the host CPU data. But the two methods we use usually provide disjoint
sets of CPU features because QEMU/KVM does not support all features
provided by the host CPU and on the other hand it can enable some
feature even if the host CPU does not support them.

So if there is a domain which requires a CPU features disabled by
QEMU/KVM, libvirt will refuse to start it with QEMU > 2.9.0 as its guest
CPU is incompatible with the host CPU data we got from QEMU. But such
domain would happily start on older QEMU (of course, the features would
be missing the guest CPU). To fix this regression, we need to combine
both CPU feature sets when checking guest CPU compatibility.

https://bugzilla.redhat.com/show_bug.cgi?id=1439933

Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
2017-04-19 16:36:38 +02:00