We can now check for the error and not care about the return value as
it will be properly handled in virBufferContentAndReset() anyway.
Signed-off-by: Martin Kletzander <mkletzan@redhat.com>
The function is useful even without using the return value. And if
needed, the return value can be obtained by other calls as well. The
potential for clean-up can be seen in the following patch.
Signed-off-by: Martin Kletzander <mkletzan@redhat.com>
Otherwise longer domain names might generate paths that are too long
to be created. This follows what other parts of the code do as well.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1453194
Signed-off-by: Martin Kletzander <mkletzan@redhat.com>
We always truncated the name at 20 bytes instead of characters. In
case 20 bytes were in the middle of a multi-byte character, then the
string became invalid and various parts of the code would error
out (e.g. XML parsing of that string). Let's instead properly
truncate it after 20 characters instead.
We cannot test this in our test suite because we would need to know
what locales are installed on the system where the tests are ran and
if there is supported one (most probably there will be, but we cannot
be 100% sure), we could initialize gettext in qemuxml2argvtest, but
there would still be a chance of getting two different (both valid,
though) results.
In order to test this it is enough to start a machine with a name for
which trimming it after 20 bytes would create invalid sequence (e.g.
1234567890123456789č where č is any multi-byte character). Then start
the domain and restart libvirtd. The domain would disappear because
such illegal sequence will not go through the XML parser. And that's
not a bug of the parser, it should not be in the XML in the first
place, but since we don't use any sophisticated formatter, just
mash some strings together, the formatting succeeds.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1448766
Signed-off-by: Martin Kletzander <mkletzan@redhat.com>
The reconnect attribute for chardev devices in QEMU is used to
configure the reconnect timeout in seconds. Setting '0' value disables
the reconnect functionality thus we don't allow to set '0' for QEMU.
To disable the reconnect user should use <reconnect enabled='no'/>.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1254971
Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
This reverts commit 92840eb3a7.
More recent reviews/changes don't have the vir*ObjNew APIs
consuming the @def, so remove from Interface as well. Changes
needed to also deal with conflicts from commit id '46f5eca4'.
Signed-off-by: John Ferlan <jferlan@redhat.com>
Commit 94c465d0 refactored the logging setup phase but introduced an
issue, where the daemon ignores verbose mode when there are no outputs
defined and the default must be used. The problem is that the default
output was determined too early, thus ignoring the potential '--verbose'
option taking effect. This patch postpones the creation of the default
output to the very last moment when nothing else can change. Since the
default output is only created during the init phase, it's safe to leave
the pointer as NULL for a while, but it will be set eventually, thus not
affecting runtime.
Patch also adjusts both the other daemons.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1442947
Signed-off-by: Erik Skultety <eskultet@redhat.com>
We can't retrieve the isolation group of a device that's not present
in the system. However, it's very common for VFs to be created late
in the boot, so they might not be present yet when libvirtd starts,
which would cause the guests using them to disappear.
Moreover, for other architectures and even ppc64 before isolation
groups were introduced, it's considered perfectly fine to configure a
guest to use a device that's not yet (or no longer) available to the
host, with the obvious caveat that such a guest won't be able to
start before the device is available.
In order to be consistent, when a device's isolation group can't be
determined fall back to not isolating it rather than erroring out or,
worse, making the guest disappear.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1484254
Signed-off-by: Andrea Bolognani <abologna@redhat.com>
While formatting disk or chardev element they both uses
virDomainDiskSourceDefFormatSeclabel() function which also closes
the source element. This is not extendable.
Use the new virXMLFormatElement() to properly format the source
element with possible child elements.
As a side effect it fixes a bug in disk source formatting.
Reviewed-by: John Ferlan <jferlan@redhat.com>
Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
This helper allows you to better structurize the code if some element
may or may not contains attributes and/or child elements.
Reviewed-by: John Ferlan <jferlan@redhat.com>
Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
To handle setting a default heads value. Convert callers that were
doing it by hand
Reviewed-by: John Ferlan <jferlan@redhat.com>
Signed-off-by: Cole Robinson <crobinso@redhat.com>
And into DeviceDefValidate which is the expected place
Reviewed-by: John Ferlan <jferlan@redhat.com>
Signed-off-by: Cole Robinson <crobinso@redhat.com>
The ram/vram = 0 bits aren't needed, and PostParse will fill in the
needed QXL default
Reviewed-by: John Ferlan <jferlan@redhat.com>
Signed-off-by: Cole Robinson <crobinso@redhat.com>
Both of these are dead code: qemu_command.c explicitly rejects
VIRT_XEN earlier in the call chain, and qemu_parse_command.c
will never set VIRT_XEN anymore
Reviewed-by: John Ferlan <jferlan@redhat.com>
Signed-off-by: Cole Robinson <crobinso@redhat.com>
This is more user-friendly because the error will be
displayed directly instead of being buried in the log.
Signed-off-by: Andrea Bolognani <abologna@redhat.com>
Reviewed-by: Pavel Hrdina <phrdina@redhat.com>
Using a variable named 'stat' clashes with the system function
'stat()' causing compiler warnings on some platforms:
libxl/libxl_driver.c: In function 'libxlDomainBlockStatsVBD':
libxl/libxl_driver.c:5387: error: declaration of 'stat' shadows a global declaration [-Wshadow]
/usr/include/sys/stat.h:455: error: shadowed declaration is here [-Wshadow]
Signed-off-by: George Dunlap <george.dunlap@citrix.com>
Reviewed-by: Andrea Bolognani <abologna@redhat.com>
When parsing the config, we look for the SCSI controllers one by one,
remembering their models, then let virDomainDefAddImplicitDevices
add them if any SCSI disk is using them.
Since these controllers are not really implicit (they are present
in the source config), add them explicitly.
This patch maintains the behavior of not adding a controller
if it was present in the config, but no disk was using it.
This also resolves the memory leak of virVMXParseConfig overwriting
the video device added by calling virDomainDefAddImplicitDevices
before the parsing is finished.
Reported-by: Michal Privoznik <mprivozn@redhat.com>
For selected hostdev types, we validate that the address type
matches the subsystem type when parsing the XML.
Move it to the validation phase, to allow extending the checks
to other subsystem types without making existing domains disappear.
At the time the check was written virtuozzo did not use disabled items in boot
order configuration. Boot items were always enabled. Now they can be disabled
as well. Supporting such items is easy - they just should be ignored.
When parsing bootable devices, we maintain a bitmap of used
<boot order=""> elements. Use it in the post-parse function
to figure out whether the user tried to mix per-device and
per-domain boot elements.
This removes the need to count them twice.
These functions contain the post-parse steps common for all drivers.
Rename it to use the 'Common' prefix, instead of the vagueness
of 'Internal', leaving 'Internal' available for other vague uses.
Since the source element is parsed only once for these type of
character devices we don't have to use temporary variable and
check whether the variable was already set.
Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
The extra check whether (connect|bind)(Host|Service) was set is
required because for UDP chardev there can be two source elements.
Without the check there could be a memory leak.
Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
In order to ensure that the default protocol is RAW, explicitly
assigning VIR_DOMAIN_CHR_TCP_PROTOCOL_RAW = 0.
Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
Currently we accept and correctly parse this chardev XML:
...
<channel type='tcp'>
<source mode='connect'/>
<source mode='bind' host='localhost'/>
<source service='4567'/>
<target type='virtio' name='test'/>
</channel>
...
The parsed formatted XML is:
...
<channel type='tcp'>
<source mode='connect' host='localhost' service='4567'/>
<target type='virtio' name='test'/>
</channel>
...
That behavior is super wrong and should not be allowed. If you notice
the current parse takes the first found attribute and uses that value,
so for example from the "<source mode='bind' host='localhost'/>" only
the "host" attribute is used. It works the same way for all possible
attributes that we are able to parse for source element.
This patch enforces providing only one source element for all character
devices, only for UDP type we allow to provide two source elements
since you can specify both modes.
Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
Since its introduction in commit 874e65aa, if someone requests:
<os><bios useserial="yes"/><os/>
we report an error if we cannot successfully count the number
of serial devices via an XPath query.
Instead of fixing the check (and moving it to the validation phase,
to prevent existing domains from disappearing), drop it completely.
For QEMU, the number of serials is checked when building the command
line.
When security drivers are active but confinement is not enabled,
there is no need to autogenerate <seclabel> elements when starting
a domain def that contains no <seclabel> elements. In fact,
autogenerating the elements can result in needless save/restore and
migration failures when the security driver is not active on the
restore/migration target.
This patch changes the virSecurityManagerGenLabel function in
src/security_manager.c to only autogenerate a <seclabel> element
if none is already defined for the domain *and* default
confinement is enabled. Otherwise the needless <seclabel>
autogeneration is skipped.
Resolves: https://bugzilla.opensuse.org/show_bug.cgi?id=1051017
I mistakenly thought pSeries guests supported 32 PHBs,
but it turns out they only support 31. Validate the
target index accordingly.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1479647
Signed-off-by: Andrea Bolognani <abologna@redhat.com>
Validation should happen after parsing, so the proper
location for it is virDomainControllerDefValidate()
rather than virDomainControllerDefParseXML().
Signed-off-by: Andrea Bolognani <abologna@redhat.com>
Use the new facility which allows to ignore failures in post parse
callbacks if they are not fatal so that VM configs are not lost if the
emulator binary is missing.
If qemuCaps can't be populated on daemon restart skip certain portions
of the post parse callbacks during config reload and re-run the callback
during VM startup.
This fixes VMs vanishing if the emulator binary was broken or
uninstalled and libvirtd was restarted.
qemuDomainControllerDefPostParse assigns the default USB controller
model when it was not specified by the user. Skip this step if @qemuCaps
is missing so that we don't fill wrong data. This will then be fixes by
re-running the post parse callback.
Report the given GIC version as unsupported if @qemuCapsi is NULL. This
will be helpful to run post parse callbacks even if qemu is not
currently installed.
If qemuCaps are not present, just return the original machine type name.
This will help in situations when qemuCaps is not available in the post
parse callback.
Some failures of the post parse callback can be tolerated. This is
specifically desired when loading the configs of existing VMs. In such
case the post parse callback should not really be modifying anything
in the definition.
This patch adds a parse flag VIR_DOMAIN_DEF_PARSE_ALLOW_POST_PARSE_FAIL
which will allow the callbacks to report non-fatal failures by returning
a positive return value. In such case the field 'postParseFailed' in the
domain definition is set to true, to notify the drivers that the
callback failed and possibly needs to be re-run.
Post parse callbacks will need to be able to signal that they failed
non-fatally. This means that we need to return the value returned by the
callback without modification.
The domain post parse callback, domain address callback and the domain
device callback (for every single device) would each grab qemuCaps for
the current emulator. This is quite wasteful. Use the new callback to do
this just once.
Some drivers use def-specific private data across callbacks (e.g.
qemuCaps in the qemu driver). Currently it's mostly allocated in every
single callback. This is rather wasteful, given that every single call
to the device callback allocates it.
The new callback will allocate the data (if not provided externally) and
then use it for the VM, address and device post parse callbacks.
Add yet another post parse callback, which is executed prior the real
one without @parseOpaque. This is meant to set basics before
@parseOpaque (in case of the qemu driver qemuCaps) can be allocated.
This callback will allow to optimize passing of custom parseOpaque
through the callbacks.
The helper returns true if a string contains any of the given chars.
virStringHasControlChars can be reimplemented using that helper.
Reviewed-by: Pavel Hrdina <phrdina@redhat.com>
Let this new method handle the device object we obtained from the
monitor in order to enhance readability.
Signed-off-by: Erik Skultety <eskultet@redhat.com>
So we have a sanity check for the udev monitor fd. Theoretically, it
could happen that the udev monitor fd changes (due to our own wrongdoing,
hence the 'sanity' here) and if that happens it means we are handling an
event from a different entity than we think, thus we should remove the
handle if someone somewhere somehow hits this hypothetical case.
Signed-off-by: Erik Skultety <eskultet@redhat.com>
It might happen that virFileResolveLinkHelper fails on the lstat system
call. virFileResolveLink expects the caller to report an error when it
fails, however this wasn't the case for udevProcessMediatedDevice.
Signed-off-by: Erik Skultety <eskultet@redhat.com>
Testing qemu-2.10-rc3 shows issues like:
qemu-system-aarch64: -drive file=/home/ubuntu/vm-start-stop/vms/
7936-0_CODE.fd,if=pflash,format=raw,unit=1: Failed to unlock byte 100
There is an apparmor deny due to qemu now locking those files:
apparmor="DENIED" operation="file_lock" [...]
name="/home/ubuntu/vm-start-stop/vms/7936-0_CODE.fd"
name="/var/lib/uvtool/libvirt/images/kvmguest-artful-normal.qcow"
[...] comm="qemu-system-aarch64" requested_mask="k" denied_mask="k"
The profile needs to allow locking for loader and nvram files via
the locking (k) rule.
Signed-off-by: Christian Ehrhardt <christian.ehrhardt@canonical.com>
Testing qemu-2.10-rc2 shows issues like:
qemu-system-x86_64: -drive file=/var/lib/uvtool/libvirt/images/kvmguest- \
artful-normal.qcow,format=qcow2,if=none,id=drive-virtio-disk0:
Failed to lock byte 100
It seems the following qemu commit changed the needs for the backing
image rules:
(qemu) commit 244a5668106297378391b768e7288eb157616f64
Author: Fam Zheng <famz@redhat.com>
file-posix: Add image locking to perm operations
The block appears as:
apparmor="DENIED" operation="file_lock" [...]
name="/var/lib/uvtool/libvirt/images/kvmguest-artful-normal.qcow"
[...] comm="qemu-system-x86" requested_mask="k" denied_mask="k"
With that qemu change in place the rules generated for the image
and backing files need the allowance to also lock (k) the files.
Disks are added via add_file_path and with this fix rules now get
that permission, but no other rules are changed, example:
- "/var/lib/uvtool/libvirt/images/kvmguest-artful-normal-a2.qcow" rw,
+ "/var/lib/uvtool/libvirt/images/kvmguest-artful-normal-a2.qcow" rwk
Signed-off-by: Christian Ehrhardt <christian.ehrhardt@canonical.com>
It's equivalent of calling virXPathString("string(.)", ctxt) but it
doesn't have to use the XPath resolving and parsing.
Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
The virXMLPropStringLimit is an equivalent of virXPathStringLimit
which should be preferred if you already have a XML dom node or
if you need to parse more than one property.
Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
Back in the day when I was implementing QoS for networks there
were no self inflating virBitmaps. Only the static ones.
Therefore, I had to allocate the whole 8KB of memory in order to
keep track of used/unused class IDs. This is rather wasteful
because nobody is ever gonna use that much classes (kernel
overhead would drastically lower the bandwidth). Anyway, now that
we have self inflating bitmaps we can start small and allocate
more if there's need for it.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: John Ferlan <jferlan@redhat.com>
Rename the variable, recent review requested just use of @filter,
so be consistent throughout.
NB: Also change the virNWFilterPtr to be @nwfilter to not conflict
with the renamed variable.
Use the structure names in the @data setup - makes it easier than
going back to find the struct fields to make sure the order of the
data is correct.
Signed-off-by: John Ferlan <jferlan@redhat.com>
To be consistent with the API definition, use the @maxnames instead
of @nnames when describing/comparing against the maximum names to
be provided for the *ConnectList[Defined]Networks APIs.
Signed-off-by: John Ferlan <jferlan@redhat.com>
Move the virObjectRef in virNetworkObjAssignDefLocked to after
the virHashAddEntry to make it "clearer" why the @ref is being
incremented. Upon return from the ObjNew we will have 1 ref on
the object already, adding it to the hash table requires the
increment.
Signed-off-by: John Ferlan <jferlan@redhat.com>
In preparation to privatize the virNetworkObj - create an accessor function
to get the current @persistent value. Also change the value to a bool rather
than an unsigned int (since that's how it's generated anyway).
Signed-off-by: John Ferlan <jferlan@redhat.com>
In order to privatize the virNetworkObj create accessors in virnetworkobj
in order to handle the get/set of the active value.
Also rather than an unsigned int, convert it to a boolean to match other
drivers representation and the reality of what it is.
Signed-off-by: John Ferlan <jferlan@redhat.com>
In preparation for making the object private, create a couple of API's
to get the obj->def & obj->newDef and set the obj->def.
While altering networkxml2conftest.c to use the virNetworkObjSetDef
API, fix the name of the variable from @dev to @def
Signed-off-by: John Ferlan <jferlan@redhat.com>
Change the variable name to be a bit more descriptive and less confusing
when used with the data.network.actual->class_id.
Signed-off-by: John Ferlan <jferlan@redhat.com>
In preparation for making the object private, create/use a couple of API's
to get/set the obj->dnsmasqPid and obj->radvdPid.
NB: Since the pid's can sometimes changed based on intervening functions,
be sure to always fetch the latest value.
Signed-off-by: John Ferlan <jferlan@redhat.com>
Since we can only ever have one reference to obj->macmap, rather
than only clearing obj->macmap during virNetworkObjUnrefMacMap
(e.g. virtual network from networkShutdownNetwork), let's just
unconditionally clear the obj->macmap to ensure that some future
change that created it's own reference to obj->macmap wouldn't
have that reference disappear if virNetworkObjDispose got called.
Signed-off-by: John Ferlan <jferlan@redhat.com>
In preparation for having a private virNetworkObj - let's create/move some
API's that handle the obj->macmap. The API's will be renamed to have a
virNetworkObj prefix to follow conventions and the arguments slightly
modified to accept what's necessary to complete their task.
Signed-off-by: John Ferlan <jferlan@redhat.com>
Move networkMacMgrFileName into src/util/virmacmap.c and rename to
virMacMapFileName. We're about to move some more MacMgr processing
files into virnetworkobj and it doesn't make sense to have this helper
in the driver or in virnetworkobj.
Signed-off-by: John Ferlan <jferlan@redhat.com>
If an environment specific _tls_x509_cert_dir is provided, then
do not VIR_STRDUP the defaultTLSx509secretUUID as that would be
for the "default" environment and not the vnc, spice, chardev, or
migrate environments. If the environment needs a secret to decode
it's certificate, then it must provide the secret. If the secrets
happen to be the same, then configuration would use the same UUID
as the default (but we cannot assume that nor can we assume that
the secret would be necessary).
Rather than assuming that what's passed to virObject{Ref|Unref}
would be a virObjectPtr as long as it's not NULL, let's do the
similar checks virObjectIsClass in order to prevent a possible
increment or decrement to some field at the obj->u.s.refs offset.
Signed-off-by: John Ferlan <jferlan@redhat.com>
The virObjectIsClass API has only ever checked object validity
based on if the @obj is not NULL and it was derived from some class.
While this has worked well in general, there is one additional
check that could be made prior to calling virClassIsDerivedFrom
which loops through the classes checking the magic number against
the klass expected magic number.
If by chance a non virObject is passed, rather than assuming the
void * @obj is a _virObject and thus offsetting to obj->klass,
obj->magic, and obj->parent, let's check that the void * @obj
has at least the "base part" of the magic number in the right
place and generate a more specific VIR_WARN message if not.
There are many consumers to virObjectIsClass, include the locking
primitives virObject{Lock|Unlock}, virObjectRWLock{Read|Write},
and virObjectRWUnlock. For those callers, the locking call will
not fail, but it also will not attempt a virMutex* call which
will "most likely" fail since the &obj->lock is used.
In order to avoid some possible future wrap on the 0xCAFExxxx
value, add a check during initialization that some new class
won't cause the wrap. Should be good for a few years at least!
It is still left up to the caller to handle the failed API calls
just as it would be if it passed a NULL opaque pointer anyobj.
If virObjectIsClass fails "internally" to virobject.c, create a
macro to generate the VIR_WARN describing what the problem is.
Also improve the checks and message a bit to indicate which was
the failure - whether the obj was NULL or just not the right class
Signed-off-by: John Ferlan <jferlan@redhat.com>
Rather than overload virObjectUnlock as commit id '77f4593b' has
done, create a separate virObjectRWUnlock API that will force the
consumers to make the proper decision regarding unlocking the
RWLock's. Similar to the RWLockRead and RWLockWrite, use the
virObjectGetRWLockableObj helper. This restores the virObjectUnlock
code to using the virObjectGetLockableObj.
Signed-off-by: John Ferlan <jferlan@redhat.com>
Introduce a helper to handle the error path more cleanly. The same
as virObjectGetLockableObj in order to essentially follow the original
logic of commit 'b545f65d' to ensure that the input argument at least
has some validity before using.
Signed-off-by: John Ferlan <jferlan@redhat.com>
Now that virObjectRWLockWrite exists to handle the virObjectRWLockable
objects, let's restore virObjectLock to only handle virObjectLockable
class locks. There still exists the possibility that the input @anyobj
isn't a valid object and the resource isn't truly locked, but that
also exists before commit id '77f4593b'.
This also restores some logic that commit id '77f4593b' removed
with respect to a common code path that commit id '10c2bb2b' had
introduced as virObjectGetLockableObj. This code path merely does
the same checks as the original virObjectLock commit 'b545f65d',
but in callable/reusable helper to ensure the @obj at least has
some validity before using.
Signed-off-by: John Ferlan <jferlan@redhat.com>
Instead of making virObjectLock be the entry point for two
different types of locks, let's create a virObjectRWLockWrite API
which will only handle the virObjectRWLockableClass objects.
Use the new virObjectRWLockWrite for the virdomainobjlist code
in order to handle the Add, Remove, Rename, and Load operations
that need to be very synchronous.
Signed-off-by: John Ferlan <jferlan@redhat.com>
Since the class it represents is based on virObjectRWLockableClass
and in order to make sure we differentiate just in case anyone somehow
believes they could use virObjectLockRead for a virObjectLockableClass,
let's rename the API to use the RW in the name. Besides the RW locks
refer to pthread_rwlock_{init|rdlock|wrlock|unlock|destroy} while the
other locks refer to pthread_mutex_{init|lock|unlock|destroy}.
Signed-off-by: John Ferlan <jferlan@redhat.com>
The correct lock order is:
nwfilter driver lock (not used in this code path)
nwfilter update lock
virt driver lock (not used in this code path)
domain object lock
but the current code have this order:
domain object lock
nwfilter update lock
Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
This way later patches can add another structures with virResctrl
prefix without the meaning being even more confusing than it needs to
be.
Signed-off-by: Martin Kletzander <mkletzan@redhat.com>
That means that returning negative values means error and non-negative
values differ in meaning, but are all successful.
Signed-off-by: Martin Kletzander <mkletzan@redhat.com>
It doesn't access anything from conf/ and ti will be needed to use
from other util/ places. This split makes the separation clearer.
Signed-off-by: Martin Kletzander <mkletzan@redhat.com>
Commit 81fb440b further qualified an if statement by adding the
boolean saveVlan to the condition. Coverity pointed out that this
change in the logic eliminated the need to check saveVlan in an
argument to virAsprintf().
Commit 9a94af6d restructured virHostdevReadNetConfig() so that it
would manually set ret = 0 after successfully reading the device's
config, but Coverity pointed out that "ret = 0" was erroneously placed
outside of an "else" clause, meaning that the the value of ret set in
the "if" clause was unnecessarily and incorrectly overwritten.
This patch moves ret = 0 into the else clause, which should silence
Coverity.
When using a VF from an SRIOV-capable network card in a guest (either
in macvtap passthrough mode, or via VFIO PCI device assignment), The
associated PF netdev must be online in order for the VF to be usable
by the guest. The guest, however, is not able to change the state of
the PF. And libvirt *could* set the PF online as needed, but that
could lead to the host receiving unexpected IPv6 traffic (since the
default for an unconfigured interface is to participate in IPv6
autoconf). For this reason, before assigning a VF to a guest, libvirt
verifies that the related PF netdev is online - if it isn't, then we
log an error and don't allow the guest startup to continue.
Until now, this check was done during virNetDevSetNetConfig(). This
works nicely because the same function is called both for macvtap
passthrough and for VFIO device assignment. But in the case of VFIO,
the VF has already been unbound from its netdev driver by the time we
get to virNetDevSetNetConfig(), and in the case of dual port Mellanox
NICs that have their VFs setup in single port mode, the *only* way to
determine the proper PF netdev to query for online status is via the
"phys_port_id" file that is in the VF netdev's sysfs directory. *BUT*
if we've unbound the VF from the netdev driver, then it doesn't *have*
a netdev sysfs directory.
So, in order to check the correct PF netdev for online status, this
patch moved the check earlier in the setup, into
virNetDevSaveNetConfig(), which is called *before* unbinding the VF
from its netdev driver.
(Note that this implies that if you are using VFIO device assignment
for the VFs of a Mellanox NIC that has the VFs programmed in single
port mode, you must let the VFs be bound to their net driver and use
"managed='yes'" in the device definition. To be more specific, this is
only true if the VFs in single port mode are using port *2* of the PF
- if the VFs are using only port 1, then the correct PF netdev will be
arrived at by default/chance))
This resolves: https://bugzilla.redhat.com/267191
virHostdevRestoreNetConfig() calls virNetDevReadNetConfig() to try and
read the "original config" of a netdev, and if that fails, it tries
again with a different directory/netdev name. This achieves the
desired effect (we end up finding the config wherever it may be), but
for each failure, virNetDevReadNetConfig() places a nice error message
in the system logs. Experience has shown that false-positive error
logs like this lead to erroneous bug reports, and can often mislead
those searching for *real* bugs.
This patch changes virNetDevReadNetConfig() to explicitly check if the
file exists before calling virFileReadAll(); if it doesn't exist,
virNetDevReadNetConfig() returns a success, but leaves all the
variables holding the results as NULL. (This makes sense if you define
the purpose of the function as "read a netdev's config from its config
file *if that file exists*).
To take advantage of that change, the caller,
virHostdevRestoreNetConfig() is modified to fail immediately if
virNetDevReadNetConfig() returns an error, and otherwise to try the
different directory/netdev name if adminMAC & vlan & MAC are all NULL
after the preceding attempt.
Mellanox ConnectX-3 dual port SRIOV NICs present a bit of a challenge
when assigning one of their VFs to a guest using VFIO device
assignment.
These NICs have only a single PCI PF device, and that single PF has
two netdevs sharing the single PCI address - one for port 1 and one
for port 2. When a VF is created it can also have 2 netdevs, or it can
be setup in "single port" mode, where the VF has only a single netdev,
and that netdev is connected either to port 1 or to port 2.
When the VF is created in dual port mode, you get/set the MAC
address/vlan tag for the port 1 VF by sending a netlink message to the
PF's port1 netdev, and you get/set the MAC address/vlan tag for the
port 2 VF by sending a netlink message to the PF's port 2 netdev. (Of
course libvirt doesn't have any way to describe MAC/vlan info for 2
ports in a single hostdev interface, so that's a bit of a moot point)
When the VF is created in single port mode, you can *set* the MAC/vlan
info by sending a netlink message to *either* PF netdev - the driver
is smart enough to understand that there's only a single netdev, and
set the MAC/vlan for that netdev. When you want to *get* it, however,
the driver is more accurate - it will return 00:00:00:00:00:00 for the
MAC if you request it from the port 1 PF netdev when the VF was
configured to be single port on port 2, or if you request if from the
port 2 PF netdev when the VF was configured to be single port on port
1.
Based on this information, when *getting* the MAC/vlan info (to save
the original setting prior to assignment), we determine the correct PF
netdev by matching phys_port_id between VF and PF.
(IMPORTANT NOTE: this implies that to do PCI device assignment of the
VFs on dual port Mellanox cards using <interface type='hostdev'>
(i.e. if you want the MAC address/vlan tag to be set), not only must
the VFs be configured in single port mode, but also the VFs *must* be
bound to the host VF net driver, and libvirt must use managed='yes')
By the time libvirt is ready to set the new MAC/vlan tag, the VF has
already been unbound from the host net driver and bound to
vfio-pci. This isn't problematic though because, as stated earlier,
when a VF is created in single port mode, commands to configure it can
be sent to either the port 1 PF netdev or the port 2 PF netdev.
When it is time to restore the original MAC/vlan tag, again the VF
will *not* be bound to a host net driver, so it won't be possible to
learn from sysfs whether to use the port 1 or port 2 PF netdev for the
netlink commands. And again, it doesn't matter which netdev you
use. However, we must keep in mind that we saved the original settings
to a file called "${PF}_${VFNUM}". To solve this problem, we just
check for the existence of ${PF1}_${VFNUM} and ${PF2}_${VFNUM}, and
use whichever one we find (since we know that only one can be there)
This patch updates functions in netdev.c to pay attention to
phys_port_id. It uses the new function virNetDevGetPhysPortID() to
learn the phys_port_id of a VF or PF, then sends that info to
virPCIGetNetName(), which has newly been modified to take an optional
phys_port_id.
A single PCI device may have multiple netdevs associated with it. Each
of those netdevs will have a different phys_port_id entry in
sysfs. This patch modifies virPCIGetNetName() to allow selecting one
of the potential many netdevs in two different ways:
1) by setting the "idx" argument, the caller can select the 1st (0),
2nd (1), etc. netdev from the PCI device's net subdirectory.
2) If the physPortID arg is set (to a null-terminated string) then
virPCIGetNetName() returns the netdev that has that phys_port_id in
the sysfs file of the same name in the netdev's directory.
On Linux each network device *can* (but not necessarily *does*) have
an attribute called phys_port_id which can be read from the file of
that name in the netdev's sysfs directory. The examples I've seen have
been a many-digit hexadecimal number (as an ASCII string).
This value can be useful when a single PCI device is associated with
multiple netdevs (e.g a dual port Mellanox SR-IOV NIC - this card has
a single PCI Physical Function (PF), and that PF has two netdevs
associated with it (the "net" subdirectory of the PF in sysfs has two
links rather than the usual single link to a netdev directory). Each
of the PF netdevs has a different phys_port_id. The Virtual Functions
(VF) are similar - the PF (a PCI device) has "n" VFs (also each of
these is a PCI device), each VF has two netdevs, and each of the VF
netdevs points back to the VF PCI device (with the "device" entry in
its sysfs directory) as well as having a phys_port_id matching the PF
netdev it is associated with.
virNetDevGetPhysPortID() simply attempts to read the phys_port_id for
the given netdev and return it to the caller. If this particular
netdev driver doesn't support phys_port_id, it returns NULL (*not* a
NULL-terminated string, but a NULL pointer) but still counts it as a
success.
https://bugzilla.redhat.com/show_bug.cgi?id=1458638
This code is so complicated because we allow enabling the same
bits at many places. Just like in this case: huge pages can be
enabled by global <hugepages/> element under <memoryBacking> or
on per <memory/> basis. To complicate things a bit more, users
are allowed to omit the page size which case the default page
size is used. And this is what is causing this bug. If no page
size is specified, @pagesize is keeping value of zero throughout
whole function. Therefore we need yet another boolean to hold
[use, don't use] information as we can't sue @pagesize for that.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Martin Kletzander <mkletzan@redhat.com>
In virDomainNetDefParseXML() the def->coalesce is parsed and
allocated by virDomainNetDefCoalesceParseXML() but in fact it's
never freed .
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
https://bugzilla.redhat.com/show_bug.cgi?id=1467245
Currently, there's a bug when undefining a domain with NVRAM
store. Basically, the unlink() of the NVRAM store file happens
during the undefine procedure iff domain is inactive. So, if
domain is running and undefine is called the file is left behind.
It won't be removed in the domain cleanup process either
(qemuProcessStop). One of the solutions is to remove if
regardless of the domain state and rely on qemu having the file
opened. This still has a downside that if the domain is defined
back the NVRAM store file is going to be new, any changes to the
current one are lost (just like with any other file that is
deleted while a process has it opened). But is it really a
downside?
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Slight refactor of the WMI serialization code to minimize mixing
openwsman and libxml2 APIs that triggered clang alignment warnings.
The only usage of libxml2 APIs now is in creating CDATA blocks,
because the openwsman API does not provide that functionality. The
clang alignment warning in this case is silenced by casting to a
void pointer first.
The API docs for the various vir$OBJECTGetConnect functions
contain a warning
WARNING: When writing libvirt bindings in other languages, do
not use this function. Instead, store the connection and
the domain object together.
There is no reason why language bindings should not use this
method, and indeed the Perl, Python, and Go bindings all use
these methods.
This warning was originally added back in
commit 3edb4bc9fb
Author: Daniel Veillard <veillard@redhat.com>
Date: Tue Jul 24 15:32:55 2007 +0000
* libvirt.spec.in NEWS docs/* po/*: preparing release 0.3.1
* src/libvirt.c python/generator.py: some cleanup and warnings
from Richard W.M. Jones
IIUC, the rational was that these APIs do not need to be
directly exposed to the non-C language, as the language
can expose the same concept itself by storing the original
virConnectPtr object alongside the virDomainPtr. There's
no reason to mandate such an approach though - it is valid
for languages to expose this directly if that suits their
needs better.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
We're storing the machine name in @priv but free it just in
qemuProcessStop, Therefore this may leak.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
