Cédric Bosdonnat [Mon, 18 Dec 2017 14:48:33 +0000 (15:48 +0100)]
lxc: set a hostname based on the container name
Set a transient hostname on containers. The hostname is computed from
the container name, only keeping the valid characters [a-zA-Z0-9-] in it.
This filtering is based on RFC 1123 and allows a digit to start the
hostname.
On live migration with --p2p like:
$ virsh migrate --live --p2p kvmguest-bionic-normal \
qemu+ssh://10.6.221.80/system
We hit an apparmor deny like:
apparmor="DENIED" operation="file_inherit"
profile="/usr/sbin/libvirtd" pid=23477 comm="ssh" family="unix"
sock_type="stream" protocol=0 requested_mask="send receive"
denied_mask="send" addr=none peer_addr=none peer="unconfined"
The rule is not perfect, but can't be restricted further at the moment
(new upstream kernel features needed). For now the lack of a profile on the
peer as well as comm not being a conditional on rules do not allow to filter
further.
Signed-off-by: Christian Ehrhardt <christian.ehrhardt@canonical.com>
Michal Privoznik [Tue, 12 Dec 2017 12:52:50 +0000 (13:52 +0100)]
src: Export virDomainMemoryAccessType*String
These are already exported at header file level because of
VIR_ENUM_DECL being in numa_conf.h. However, they are not being
exported at object level because of missing libvirt_private.syms
record.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: John Ferlan <jferlan@redhat.com>
QEMU 2.7 and newer don't allow guests to start unless the initial
vCPUs count is a multiple of the vCPU hotplug granularity, so
validate it and report an error if needed.
Andrea Bolognani [Thu, 14 Dec 2017 14:29:28 +0000 (15:29 +0100)]
qemu: Invert condition nesting in qemuDomainDefValidate()
While at the moment we're only performing a single check that is
connected to vCPU hotplugging, we're going to introduce a second
one soon. Move the topology check underneath the capability check
to make that easier; since, after this change, the 'topologycpus'
variable doesn't need to have function scope, we move its
declaration to the inner scope as well.
The comments around the check are modified in order to explain
the different QEMU versions involved.
Signed-off-by: Andrea Bolognani <abologna@redhat.com>
John Ferlan [Wed, 20 Dec 2017 11:36:26 +0000 (06:36 -0500)]
qemu: Move TCP and haveTLS checks into qemuDomainDelChardevTLSObjects
Similar to qemuDomainAddChardevTLSObjects let's move the chardev
source must be TCP and it has the @haveTLS flag set checks before
trying to delete the TLS objects.
For the Chr device this represents no change; however, for RNG device
this is an additionaly check that was missed in commit id '68808516'.
Before adding the objects, TCP and haveTLS are checked.
Cédric Bosdonnat [Wed, 20 Dec 2017 16:36:10 +0000 (17:36 +0100)]
netserver: close clients before stopping all drivers
So far clients were closed when disposing the daemon, after the state
driver cleanup. This was leading to libvirtd crashing at shutdown due
to missing driver.
Moving the client close in virNetServerClose() fixes the problem.
Cédric Bosdonnat [Mon, 11 Dec 2017 10:09:31 +0000 (11:09 +0100)]
virt-aa-helper: handle more disk images
virt-aa-helper needs read access to the disk image to resolve symlinks
and add the proper rules to the profile. Its profile whitelists a few
common paths, but users can place their images anywhere.
This commit helps users allowing access to their images by adding their
own rules in apparmor.d/local/usr.lib.libvirt.virt-aa-helper.
This commit also adds rules to allow reading files named:
- *.raw as this is a rather common disk image extension
- /run/libvirt/**[vd]d[a-z] as these are used by virt-sandbox
which is fine if on a host with GNU make, but on *BSD running
'make' will end in tears. We should tell users to run 'gmake'
on these platforms. If 'gmake' doesn't exist then we should
report an error too
"GNU make is required to build libvirt"
Reviewed-by: Andrea Bolognani <abologna@redhat.com> Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Jim Fehlig [Fri, 15 Dec 2017 22:28:49 +0000 (15:28 -0700)]
Improve filtering of Xen domain0 in libvirt-guests
The list_guests function in libvirt-guests uses 'grep -v' to filter
Xen domain0 from a list of guests. If domain0 is the only item in
the list, 'grep -v' returns 1, causing the 'stop' operation to fail
when action is 'suspend'. Improve the filtering by using sed to remove
domain0 from the list of guests.
Erik Skultety [Wed, 13 Dec 2017 13:39:40 +0000 (14:39 +0100)]
admin: Use the connection to determine a client is connected readonly
Prior to this change, we relied solely on the inherited readonly
attribute of a service's socket. This only worked for our UNIX sockets
(and only to some degree), but doesn't work for TCP sockets which are RW
by default, but such connections support RO as well. This patch forces
an update on the client object once we have established a connection to
reflect the nature of the connection itself rather than relying on the
underlying socket's attributes.
Clients connected to the admin server have always been connected as RW
only.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1524399 Signed-off-by: Erik Skultety <eskultet@redhat.com>
The use of [[:blank:]] was intended to cover tabs as well, but
it couldn't possibly work in its current form, so the regex was
tweaked in d09429abe826.
With the original reason for using [[:blank:]] now gone, we can
replace its usage with plain spaces. A comment about the purpose
of the check is added as well.
Signed-off-by: Andrea Bolognani <abologna@redhat.com>
Laine Stump [Tue, 12 Dec 2017 21:18:07 +0000 (16:18 -0500)]
qemu: log error on attempts to set filterref on an OVS-connected interface
ebtables/iptables processing is skipped for any interface connected to
Open vSwitch (they have their own packet filtering), likewise for
midonet (according to
http://blog.midokura.com/2016/04/midonet-rule-chains), but libvirt
would allow adding a <filterref> to interfaces connected in these
ways, so the user might mistakenly believe they were being protected.
This patch checks for a non-NULL <virtualport> element for an
interface (or its network) and logs an error if <virtualport> and
<filterref> are both present. This could cause some previously working
domains to no longer start, but that's really the whole point of this
patch - to warn people that their filterref isn't protecting them as
they might have thought.
I don't bother checking this during post-parse validation, because
such a check would be incomplete - it's possible that a network would
have a <virtualport> that would be applied to an interface, and you
can't know that until the domain is started.
Laine Stump [Mon, 11 Dec 2017 19:26:54 +0000 (14:26 -0500)]
qemu: delete exist bandwidth restrictions when they are removed from config
When the <bandwidth> of an interface is changed with update-device,
the old settings are cleared with tc, then new settings added with
tc. But if the <bandwidth has been removed, the old settings weren't
being removed, so the bandwidth restrictions would still be active on
the interface although the interface status in libvirt showed that
they had been removed.
This patch fixes it by calling virNetDevBandwidthClear() if the
"modification" to the interface bandwidth was to completely clear
it.
An alternative could have been to modify virNetDevBandwidthSet() to
always clear existing bandwith settings at the beginning of the
function (currently it short circuits in that case, doing nothing),
but that would have led to cases where virNetDevBandwidthClear() was
now being called in cases where it previously wasn't, and while many
of those cases would be NOPs, there could be cases where it would
cause an error. The way this patch works, the ...Clear() function is
only called in cases where the ...Set() function had previously been
called successfully, so the risk of regression is minimized.
Peter Krempa [Tue, 12 Dec 2017 16:55:03 +0000 (17:55 +0100)]
conf: Add infrastructure for disk source private data XML
VM drivers may need to store additional private data to the status XML
so that it can be restored after libvirtd restart. Since not everything
is needed add a callback infrastructure, where VM drivers can add only
stuff they need.
Note that the private data is formatted as a <privateData> sub-element
of the <disk> or <backingStore> <source> sub-element. This is done since
storing it out of band (in the VM private data) would require a complex
matching process to allow to put the data into correct place.
John Ferlan [Tue, 12 Dec 2017 13:33:48 +0000 (08:33 -0500)]
util: Fix error path in virSCSIVHostOpenVhostSCSI
We cannot be sure someone initialized the passed *vhostfd and we
certainly don't want or need to be calling VIR_FORCE_CLOSE on what
probably is -1. So let's just return -1 immediately.
John Ferlan [Tue, 12 Dec 2017 11:30:31 +0000 (06:30 -0500)]
qemu: Need to assign PCI address to vhost-scsi
Commit id '70249927b' neglected to cover this case because the test
had taken the "shortcut" to already add the <address>; however, when
the PCI address assignment code was adjusted by commit id '70249927'
the vhost-scsi (VIR_DOMAIN_HOSTDEV_SUBSYS_TYPE_SCSI_HOST) wasn't
covered thus returning a 0 for pciFlags. So I altered the tests too
to make sure it doesn't happen again.
Previously the qemuxml2xmloutdata was a softlink to the source
qemuxml2argvdata, so I unlinked and recreated the output file to
force generation of the adddress. Without the test changes, an
address generation returns:
libvirt: Domain Config error : internal error: Cannot automatically
add a new PCI bus for a device with connect flags 00
if an address was supplied in the test, a restart of libvirtd or
edit of a guest would display the following opaque message:
warning : qemuDomainCollectPCIAddress:1237 :
qemuDomainDeviceCalculatePCIConnectFlags() thinks that the device
with PCI address 0000:00:09.0 should not have a PCI address
where the address is related to the guest PCI address provided.
John Ferlan [Tue, 5 Dec 2017 21:03:34 +0000 (16:03 -0500)]
qemu: Tolerate storage source private data being NULL for hotplug SCSI hostdev
Commit id 'c5c96545' neglected to validate that the srcPriv was
non-NULL before dereferencing. Similar problem to what was fixed
by commit id '8056721c' but missed during multiple rebases and
code reworks.
John Ferlan [Mon, 9 Oct 2017 17:34:45 +0000 (13:34 -0400)]
storage: Convert virStoragePoolObjList to use virObjectRWLockable
Now that we have a private storage pool list, we can take the next
step and convert to using objects. In this case, we're going to use
RWLockable objects (just like every other driver) with two hash
tables for lookup by UUID or Name.
Along the way the ForEach and Search API's will be adjusted to use
the related Hash API's and the various FindBy functions altered and
augmented to allow for HashLookup w/ and w/o the pool lock already
taken.
After virStoragePoolObjRemove we will need to virObjectUnref(obj)
after to indicate the caller is "done" with it's reference. The
Unlock occurs during the Remove.
The NumOf, GetNames, and Export functions all have their own callback
functions to return the required data and the FindDuplicate code
can use the HashSearch function callbacks.
John Ferlan [Fri, 24 Nov 2017 13:17:02 +0000 (08:17 -0500)]
storage: Fix path check in storagePoolLookupByTargetPath
Commit id '5ab746b8' introduced the function as perhaps a copy
of storageVolLookupByPath; however, it did not use the @cleanpath
variable even though it used the virFileSanitizePath. So in essance
the only "check" being done for failure is whether it was possible
to strdup the path.
Looking at the virStoragePoolDefParseXML one will note that the
target.path is stored using the result of virFileSanitizePath.
Therefore, this function should sanitize and use the input @path
for the argument to storagePoolLookupByTargetPathCallback which
is comparing against stored target.path values.
Additionally, if there was an error we should use the proper error
of VIR_ERR_NO_STORAGE_POOL (instead of VIR_ERR_NO_STORAGE_VOL).
Marc Hartmayer [Wed, 13 Dec 2017 10:56:13 +0000 (11:56 +0100)]
audit: Log only an info message if audit_level < 2 and audit is not supported
Replace the error message during startup of libvirtd with an info
message if audit_level < 2 and audit is not supported by the
kernel. Audit is not supported by the current kernel if the kernel
does not have audit compiled in or if audit is disabled (e.g. by the
kernel cmdline).
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>
==25251== 5 bytes in 1 blocks are definitely lost in loss record 7 of 81
==25251== at 0x4C2BEDF: malloc (vg_replace_malloc.c:299)
==25251== by 0x967E379: strdup (in /lib64/libc-2.25.so)
==25251== by 0x5366F9F: virStrdup (virstring.c:941)
==25251== by 0x538BF1D: virDomainCapsNew (domain_capabilities.c:121)
==25251== by 0x10EACE: test_virDomainCapsFormat (domaincapstest.c:295)
==25251== by 0x10FBD2: virTestRun (testutils.c:180)
==25251== by 0x10F192: mymain (domaincapstest.c:457)
==25251== by 0x111C7F: virTestMain (testutils.c:1119)
==25251== by 0x10FA3C: main (domaincapstest.c:528)
Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: John Ferlan <jferlan@redhat.com>
Michal Privoznik [Tue, 12 Dec 2017 14:00:15 +0000 (15:00 +0100)]
xenMakeIPList: Don't leak @address_array
==32171== 32 bytes in 1 blocks are definitely lost in loss record 44 of 107
==32171== at 0x4C2DEF6: calloc (vg_replace_malloc.c:711)
==32171== by 0x55744A9: virAllocN (viralloc.c:191)
==32171== by 0x12CED2: xenMakeIPList (xen_common.c:1186)
==32171== by 0x12D0BE: xenFormatNet (xen_common.c:1221)
==32171== by 0x12F0D2: xenFormatVif (xen_common.c:1889)
==32171== by 0x12F2B4: xenFormatConfigCommon (xen_common.c:1944)
==32171== by 0x13BA32: xenFormatXL (xen_xl.c:1971)
==32171== by 0x1186CA: testCompareParseXML (xlconfigtest.c:105)
==32171== by 0x118A64: testCompareHelper (xlconfigtest.c:205)
==32171== by 0x119E36: virTestRun (testutils.c:180)
==32171== by 0x11970E: mymain (xlconfigtest.c:301)
==32171== by 0x11BEE3: virTestMain (testutils.c:1119)
Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: John Ferlan <jferlan@redhat.com>
==30399== 180 (144 direct, 36 indirect) bytes in 3 blocks are definitely lost in loss record 91 of 111
==30399== at 0x4C2E0FF: realloc (vg_replace_malloc.c:785)
==30399== by 0x5574572: virReallocN (viralloc.c:245)
==30399== by 0x5574668: virExpandN (viralloc.c:294)
==30399== by 0x55747AB: virResizeN (viralloc.c:352)
==30399== by 0x560074D: virStringSplitCount (virstring.c:115)
==30399== by 0x137A59: xenParseXLVnuma (xen_xl.c:442)
==30399== by 0x13952B: xenParseXL (xen_xl.c:1064)
==30399== by 0x11884D: testCompareFormatXML (xlconfigtest.c:152)
==30399== by 0x118A87: testCompareHelper (xlconfigtest.c:207)
==30399== by 0x119E36: virTestRun (testutils.c:180)
==30399== by 0x119186: mymain (xlconfigtest.c:274)
==30399== by 0x11BEE3: virTestMain (testutils.c:1119)
Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: John Ferlan <jferlan@redhat.com>
If domain is active, but the undefine API was called without the
VIR_DOMAIN_UNDEFINE_KEEP_NVRAM flag set, the following incorrect
error message is produced:
error: Requested operation is not valid: cannot delete inactive domain with nvram
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Move the IDE controller check from command line building to
controller def validation. Also explicitly include the avoidance
check for the implicit IDE controller from qemuBuildSkipController.
Cause the IDE case for command line building to generate a
failure if called to add an IDE since that shouldn't happen
if the Validate code did the right thing.
Migration never removes any configuration files on the destination host.
Thus when the domain is already defined on the destination, it will stay
persistent even after migration without --persist.
Jiri Denemark [Thu, 7 Dec 2017 10:23:50 +0000 (11:23 +0100)]
qemu: Make sure host-model uses CPU model supported by QEMU
When reconnecting to a running domain started by old libvirt, which did
not change host-model into a custom CPU definition, we replace the CPU
definition with a specific CPU model from host capabilities. However,
that CPU model may not be supported by the running qemu process. We need
to translate the CPU model to one of the models which libvirt could have
used when starting the domain.
Peter Krempa [Thu, 7 Dec 2017 15:11:18 +0000 (16:11 +0100)]
qemu: blockjob: Reset disk source index after pivot
Since we are re-detecting the backing chain after pivoting to the active
block commit target (or block copy target) the disk index needs to be
reset to 0. This is necessary since we move a member of the backing
chain to disk->src but clear indexes only starting from
disk->src->backingStore. The freshly detected images have indexes
starting from 1, but since we've pivoted into an image which was
previously a backing store it would have a non-0 index.
The lookup function would then return the top of the chain for queries
like 'vda[1]' instead of the first backing store.
This problem will not be present once we keep the disk indexes stable.
Peter Krempa [Wed, 6 Dec 2017 15:20:07 +0000 (16:20 +0100)]
storage: Don't dereference driver object if virStorageSource is not initialized
virStorageFileReportBrokenChain uses data from the driver private data
pointer to print the user and group. This would lead to a crash in call
paths where we did not initialize the storage backend as recently added
in commit 24e47ee2b93 to qemuDomainDetermineDiskChain.
Currently we only describe setting the CN field for server certs. This leads
to inevitable pain for users who set it to the fully qualified hostname and
then use a unqualified hostname or IP address to connect in the URI. Describe
the usage of Subject Alt Name extensions, to provide multiple hostnames and
IP addresses. This will help users avoid the classic mistake and is important
future proofing, since at least in browsers, TLS libraries no longer use the
CN field for validation, mandating use of SAN info instead.
Reviewed-by: John Ferlan <jferlan@redhat.com> Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
nwfilter: remove bogus 'protocolid' attribute on arp/rarp fields
Various example XML documents for arp/rarp filtering have a protocolid
XML attribute defined. This is never parsed or output by the libvirt XML
handling code, so shouldn't be present in example XML files either
Reviewed-by: John Ferlan <jferlan@redhat.com> Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
The contents of a <rule> are a choice of exactly one union member. The
RNG schema, however, was allowing an arbitrary number of instances of every
union member at once.
Reviewed-by: John Ferlan <jferlan@redhat.com> Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Lin Ma [Fri, 1 Dec 2017 11:31:30 +0000 (19:31 +0800)]
tests: Drop IDE controller in CCW
Adding an IDE controller for a machinetype that has no built-in IDE
controller, libvirt will log an error. Currently the machinetype list
which returns by qemuDomainMachineHasBuiltinIDE only includes 440fx,
malta, sun4u and g3beige.
Signed-off-by: Lin Ma <lma@suse.com> Signed-off-by: John Ferlan <jferlan@redhat.com>
Lin Ma [Fri, 1 Dec 2017 11:31:29 +0000 (19:31 +0800)]
tests: Remove use of IDE disk for pseries floppy test
Adding an IDE controller for a machinetype that has no built-in IDE
controller, libvirt will log an error. Currently the machinetype list
which returns by qemuDomainMachineHasBuiltinIDE only includes 440fx,
malta, sun4u and g3beige.
Remove the disk and the .args file since the expectation is the test
will fail in qemuxml2argvtest because floppy is not supported on pseries
and thus no disk is necessary and no .args file would be created to
compare against.
Signed-off-by: Lin Ma <lma@suse.com> Signed-off-by: John Ferlan <jferlan@redhat.com>
This patch pass event error up to the place where we can
use it. Error is passed only for sync blockjob event mode
as we can't use the error in async mode. In async mode we
just pass the event details to the client thru event API
but current blockjob event API can not carry extra parameter.
nwfilter: don't crash listing filters in unprivileged daemon
The unprivileged libvirtd does not support nwfilter config, by leaves the
driver active. It is supposed to result in all APIs being an effective
no-op, but several APIs rely on driver->nwfilters being non-NULL, or they
will reference a NULL pointer. Rather than adding checks for NULL in many
places, just make sure driver->nwfilters is always initialized.
Reviewed-by: John Ferlan <jferlan@redhat.com> Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
gettid() is Linux-specific and should not be used in programs
that are intended to be portable.
Unfortunately, it looks like macOS implemented the functionality
and defined SYS_gettid accordingly, only to deprecate syscall()
altogether with 10.12 (Sierra), released last late year.
To avoid compilation errors, call gettid() on Linux only.
Signed-off-by: Andrea Bolognani <abologna@redhat.com> Reviewed-by: Daniel P. Berrange <berrange@redhat.com>
the next release will happen in the middle of January 2018 and,
being the first release of a new year, will bring a brand new
major version number with it.
Signed-off-by: Andrea Bolognani <abologna@redhat.com>
tests: Drop qemuxml2argv- prefix for qemuxml2argv-*.xml test cases
These XMLs live in a separate directory, there's no need for them
to have a special prefix in addition. It also doesn't play nicely
with ':e' completion in Vim, finding proper file based on
qemuxml2argvtest.c is also needlessly complicated.
The files were renamed using the following commands. From
qemuxml2argvdata:
for i in qemuxml2argv-*.xml; do mv $i ${i#qemuxml2argv-}; done
and then (to fix broken symlinks) from qemuxml2argvdata and
qemuxml2xmloutdata:
for i in $(find . -xtype l); do \
ln -sf $(readlink $i | sed 's/qemuxml2argv-//') $i;
done
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
John Ferlan [Tue, 21 Nov 2017 23:13:43 +0000 (18:13 -0500)]
conf: Clean up virDomainNumaDefNodeDistanceParseXML
Clean up the style a bit w/r/t to not using a unary operator on an
integer value that could be zero - compare vs. zero instead.
Set the def->mem_nodes[*].distances to rdist or ldist inside the
if condition - no need to set outside since the value being set
to is what was fetched.
During cleanup, be sure to initialize the ndistances on error and
use the < 0 comparison not the unary one.
==899== 39 bytes in 1 blocks are definitely lost in loss record 732 of 1,003
==899== at 0x4C2AEDF: malloc (vg_replace_malloc.c:299)
==899== by 0x8B68CE7: vasprintf (in /lib64/libc-2.25.so)
==899== by 0x55498D2: virVasprintfInternal (virstring.c:708)
==899== by 0x55499E7: virAsprintfInternal (virstring.c:729)
==899== by 0x2BECFFF0: qemuGetMemoryBackingBasePath (qemu_conf.c:1757)
==899== by 0x2BF23225: qemuStateInitialize (qemu_driver.c:893)
==899== by 0x563073D: virStateInitialize (libvirt.c:770)
==899== by 0x124CC4: daemonRunStateInit (libvirtd.c:834)
==899== by 0x55521CD: virThreadHelper (virthread.c:206)
==899== by 0x88D9686: start_thread (in /lib64/libpthread-2.25.so)
==899== by 0x8BEAEFE: clone (in /lib64/libc-2.25.so)
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
projects / libvirt.git / log