Jonathon Jongsma [Tue, 27 Aug 2019 20:35:56 +0000 (15:35 -0500)]
virsh: alphabetize domain commands in man page
It appears that all commands were originally fully in alphabetical order
but as new commands were added, they were sometimes inserted out of
order. Fix up all domain commands so that they're in alphabetical order
again.
Signed-off-by: Jonathon Jongsma <jjongsma@redhat.com> Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Jonathon Jongsma [Tue, 27 Aug 2019 20:35:55 +0000 (15:35 -0500)]
virsh: add 'guestinfo' command
The 'guestinfo' command uses the new virDomainGetGuestInfo() API to
query information about the specified domain and print it out for the
user. The output is modeled roughly on the 'domstats' command.
Signed-off-by: Jonathon Jongsma <jjongsma@redhat.com> Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
When we're collecting guest information, older agents may not support
all agent commands. In the case where the user requested all info
types (i.e. types == 0), ignore unsupported command errors and gather as
much information as possible. If the agent command failed for some other
reason, or if the user explciitly requested a specific info type (i.e.
types != 0), abort on the first error.
Signed-off-by: Jonathon Jongsma <jjongsma@redhat.com> Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Jonathon Jongsma [Tue, 27 Aug 2019 20:35:53 +0000 (15:35 -0500)]
lib: minor fixes to virDomainGetGuestInfo docs
Due to a typo, some of the field names didn't have closing quotes,
the information about the hostname was omitted and there was an
empty line missing after filesystem info description (which helps
our docs generator produce better looking HTML).
Signed-off-by: Jonathon Jongsma <jjongsma@redhat.com> Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Michal Privoznik [Tue, 27 Aug 2019 15:21:32 +0000 (17:21 +0200)]
docs: Make anchors follow HTML5
According to HTML specification, <a name=''> works in HTML4, but
<a id=''> works in both HTML4 and HTML5. This is followed even in
docs/page.xsl where HTML bookmark links are generated only for
those anchors which have @id attribute.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Boris Fiuczynski [Tue, 27 Aug 2019 14:19:23 +0000 (16:19 +0200)]
qemu: support bootindex on vfio-ccw mdev devices
Add support to specify a boot order on vfio-ccw passthrough devices.
Reviewed-by: Marc Hartmayer <mhartmay@linux.ibm.com> Signed-off-by: Bjoern Walk <bwalk@linux.ibm.com> Signed-off-by: Boris Fiuczynski <fiuczy@linux.ibm.com> Reviewed-by: Ján Tomko <jtomko@redhat.com> Signed-off-by: Ján Tomko <jtomko@redhat.com>
Boris Fiuczynski [Tue, 27 Aug 2019 14:19:22 +0000 (16:19 +0200)]
qemu: refactor mdev validation method signatures
Refactoring the method signatures in preparation for
checking boot index of the mediated devices.
Reviewed-by: Ján Tomko <jtomko@redhat.com> Reviewed-by: Marc Hartmayer <mhartmay@linux.ibm.com> Signed-off-by: Bjoern Walk <bwalk@linux.ibm.com> Signed-off-by: Boris Fiuczynski <fiuczy@linux.ibm.com> Signed-off-by: Ján Tomko <jtomko@redhat.com>
Boris Fiuczynski [Tue, 27 Aug 2019 14:19:21 +0000 (16:19 +0200)]
qemu: make error messages device specific
Changing the error messages to report the problem encountered.
Reviewed-by: Ján Tomko <jtomko@redhat.com> Reviewed-by: Marc Hartmayer <mhartmay@linux.ibm.com> Signed-off-by: Bjoern Walk <bwalk@linux.ibm.com> Signed-off-by: Boris Fiuczynski <fiuczy@linux.ibm.com> Signed-off-by: Ján Tomko <jtomko@redhat.com>
Boris Fiuczynski [Tue, 27 Aug 2019 14:19:20 +0000 (16:19 +0200)]
qemu: move hostdev boot validation into domain validation
Moving the hostdev boot support validation from the command line
generator code into the domain validation code.
Reviewed-by: Marc Hartmayer <mhartmay@linux.ibm.com> Signed-off-by: Bjoern Walk <bwalk@linux.ibm.com> Signed-off-by: Boris Fiuczynski <fiuczy@linux.ibm.com> Reviewed-by: Ján Tomko <jtomko@redhat.com> Signed-off-by: Ján Tomko <jtomko@redhat.com>
Boris Fiuczynski [Tue, 27 Aug 2019 14:19:19 +0000 (16:19 +0200)]
tests: add vhost scsi hostdev boot unsupported test
Adding a failure test for booting from a vhost scsi hostdev device.
Reviewed-by: Marc Hartmayer <mhartmay@linux.ibm.com> Signed-off-by: Bjoern Walk <bwalk@linux.ibm.com> Signed-off-by: Boris Fiuczynski <fiuczy@linux.ibm.com> Reviewed-by: Ján Tomko <jtomko@redhat.com> Signed-off-by: Ján Tomko <jtomko@redhat.com>
Boris Fiuczynski [Tue, 27 Aug 2019 14:19:18 +0000 (16:19 +0200)]
tests: add vfio-ap mdev tests
Reviewed-by: Marc Hartmayer <mhartmay@linux.ibm.com> Signed-off-by: Bjoern Walk <bwalk@linux.ibm.com> Signed-off-by: Boris Fiuczynski <fiuczy@linux.ibm.com> Signed-off-by: Ján Tomko <jtomko@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com>
Michal Privoznik [Tue, 27 Aug 2019 09:49:27 +0000 (11:49 +0200)]
remote_daemon_dispatch: Don't open code xdr_free()
At two places we are open coding xdr_free():
remoteRelayDomainEventTunable() and
remoteRelayDomainEventJobCompleted().
Bot of these functions use make_nonnull_domain() to put domain
IDs tuple into return structure and then continue encoding the
rest of structure. If that fails, they call VIR_FREE() directly.
While this okay, we should use xdr_free() which frees the whole
return structure for us.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Erik Skultety <eskultet@redhat.com>
Michal Privoznik [Tue, 27 Aug 2019 08:57:49 +0000 (10:57 +0200)]
remote: Check for limits when encoding typed params
The same way we check for limits when decoding typed parameters
(virTypedParamsDeserialize()) we should do the same check when
serializing them so that we don't put onto the wire more than our
limits allow. Surprisingly, we were doing so explicitly in some
places but not all of them.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Erik Skultety <eskultet@redhat.com> Reviewed-by: Jonathon Jongsma <jjongsma@redhat.com>
Michal Privoznik [Tue, 27 Aug 2019 09:05:21 +0000 (11:05 +0200)]
remote_daemon_dispatch: Check for limit properly in remoteDispatchConnectGetAllDomainStats
The return structure is a bit complicated and that's why it is
very easy to check for RPC limits incorrectly. The structure is
an array of remote_domain_stats_record structures with the limit
of REMOTE_DOMAIN_LIST_MAX. The latter structure then poses a
different limit on typed params:
REMOTE_CONNECT_GET_ALL_DOMAIN_STATS_MAX (which is what we are
checking for mistakenly).
Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Erik Skultety <eskultet@redhat.com> Reviewed-by: Jonathon Jongsma <jjongsma@redhat.com>
Michal Privoznik [Tue, 27 Aug 2019 14:08:12 +0000 (16:08 +0200)]
virhostdevtest: Reset libvirt error on expected failure
If a libvirt error occurred during a test, then virTestRun()
reports it (regardless of test returning success or failure).
For instance, in this specific case, a hostdev is detached twice
and the second attempt is expected to fail. It does fail and
libvirt error is reported which is then printed onto stderr.
Insert virResetLastError() calls on appropriate places to avoid
that.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com>
Michal Privoznik [Tue, 27 Aug 2019 14:00:55 +0000 (16:00 +0200)]
virhostdevtest: Drop most of 'cleanup' and 'out' labels
In this test there is this macro CHECK_LIST_COUNT() which checks
if a list of PCI devices contains expected count. If it doesn't
an error is reported and 'goto cleanup' is invoked. There's no
real reason for that as even since its introduction there is no
cleanup done and all 'cleanup' labels contain nothing but
'return'.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com>
Michal Privoznik [Tue, 27 Aug 2019 14:00:10 +0000 (16:00 +0200)]
virhostdevtest: Check for integer retval in more verbose way
There are few functions called from the test which return an
integer but their retval is compared as if it was a pointer.
Now, there is nothing wrong with that from machine POV, but
from readability perspective it's wrong.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com>
There is an inconsistency with VIR_TEST_DEBUG() calls. One half
(roughly) of calls does have the newline character the other one
doesn't. Well, it doesn't have it because it assumed blindly that
new line will be printed, which is not the case.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com>
Michal Privoznik [Tue, 27 Aug 2019 07:07:39 +0000 (09:07 +0200)]
qemu: Don't duplicate domain def in qemuDomainGetFSInfo
Introduced in v3.0.0-rc1~336, the commit message doesn't really
justifies the expensive domain def copy creation. Now, that
vm->def is guarded in this function by job acquirement we can use
vm->def directly.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com>
Michal Privoznik [Tue, 27 Aug 2019 06:53:53 +0000 (08:53 +0200)]
qemu: Acquire domain job in qemuDomainGetFSInfo and qemuDomainGetGuestInfo
These two functions work with vm->def in their critical sections
(i.e. after the job was acquired and before it is released). But
that means, they need QUERY domain job too to prevent vm->def
change.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com>
The remote client invokes the 'nc' binary on the remote server to tunnel
access to the socket. As such the 'nc' binary needs to be pulled in only
by the libvirt-daemon sub-RPM, not the libvirt-client sub-RPM.
Reviewed-by: Michal Privoznik <mprivozn@redhat.com> Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
The 'nc' RPM does not in fact exist anymore, this is a virtual provide
from the nmap-ncat RPM which the maintainer wishes to delete. Change the
dep to use the actual binary path we want to invoke.
Reviewed-by: Michal Privoznik <mprivozn@redhat.com> Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
rpm: don't enable socket activation in upgrade if --listen present
Currently during RPM upgrade we restart libvirtd and unconditionally
enable use of systemd socket activation for the UNIX sockets.
If the user had previously given the --listen arg to libvirtd though,
this will no longer be honoured if socket activation is used.
We could start libvirtd-tcp.socket or libvirtd-tls.socket for this,
but mgmt tools like puppet/ansible might not be expecting this.
So for now we silently disable socket activation if we see --listen
was previously set on the host.
Reviewed-by: Ján Tomko <jtomko@redhat.com> Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
remote: forbid the --listen arg when systemd socket activation
When using systemd socket activation the --listen arg has no
effect. This is confusing to users upgrading from previous versions of
libvirt as their config is silently ignored. Turn use of --listen into a
fatal error when sockets are passed from systemd.
This helps the admin discover the change in behaviour and thus decide
whether to stick with socket activation or revert to previous behaviour.
Reviewed-by: Ján Tomko <jtomko@redhat.com> Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
We need to give users the ability to customize the length of the
shutdown timeout, or even disable timeouts entirely. Thus we must move
the timeout arg into the sysconf file, instead of the service unit.
Reviewed-by: Ján Tomko <jtomko@redhat.com> Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
remote: use Wants instead of Requires for libvirtd sockets
To facilitate upgrades from earlier versions of libvirt which did not
use socket activation for libvirtd, we want to allow the libvirtd socket
units to be disabled (masked). This can only be supported if we use the
weaker Wants statement instead of Requires.
Reviewed-by: Ján Tomko <jtomko@redhat.com> Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
build: support customization of runstatedir variable with old autoconf
Many distros have moved /var/run to /run with the introduction of
systemd. /var/run still exists as a symlink to /run, but its usage
is deprecated.
autoconf added a --runstatedir option back in 2013 but there's still no
new release of autoconf that includes this.
gnulib meanwhile added support to propagate this arg's value to
configmake.h, but it falls back to $localstatedir/run for autoconf 2.69
and older, which is what every distro today has.
To deal with this problem we add a --with-runstatedir arg that then sets
the $runstatedir env variable that future autoconf's --runstatedir arg
will also use. This finally enables $runstatedir to be pointed to /run.
Reviewed-by: Michal Privoznik <mprivozn@redhat.com> Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
All code using LOCALSTATEDIR "/run" is updated to use RUNSTATEDIR
instead. The exception is the remote driver client which still
uses LOCALSTATEDIR "/run". The client needs to connect to remote
machines which may not be using /run, so /var/run is more portable
due to the /var/run -> /run symlink.
Some duplicate paths in the apparmor code are also purged.
There's no functional change by default yet since both expressions
expand to the same value.
Reviewed-by: Michal Privoznik <mprivozn@redhat.com> Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Jim Fehlig [Fri, 23 Aug 2019 18:34:46 +0000 (12:34 -0600)]
xenconfig: move contents to libxl driver and remove directory
After the legacy xen driver was removed the libxl driver became
the only consumer of xenconfig. Move the few files in xenconfig
to the libxl driver and remove the directory.
Signed-off-by: Jim Fehlig <jfehlig@suse.com> Reviewed-by: Ján Tomko <jtomko@redhat.com>
Jonathon Jongsma [Fri, 23 Aug 2019 16:31:21 +0000 (11:31 -0500)]
qemu: add helper for getting full FSInfo
This function adds the complete filesystem information returned by the
qemu agent to an array of typed parameters with field names intended to
to be returned by virDomainGetGuestInfo()
Signed-off-by: Jonathon Jongsma <jjongsma@redhat.com> Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Michal Privoznik <mprivozn@redhat.com> Tested-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Jonathon Jongsma [Fri, 23 Aug 2019 16:31:20 +0000 (11:31 -0500)]
qemu: add support for new fields in FSInfo
Since version 3.0, qemu has returned disk usage statistics in
guest-get-fsinfo. And since 3.1, it has returned information about the
disk serial number and device node of disks that are targeted by the
filesystem.
Unfortunately, the public API virDomainGetFSInfo() returns the
filesystem info using a virDomainFSInfo struct, and due to API/ABI
guarantees it cannot be extended. So this new information cannot
easily be added to the public API. However, it is possible to add this
new filesystem information to a new virDomainGetGuestInfo() API which
will be based on typed parameters and is thus more extensible.
In order to support these two use cases, I added an internal struct
which the agent code uses to return all of the new data fields. This
internal struct can be converted to the public struct at a cost of some
extra memory allocation.
In a following commit, this additional information will be used within
virDomainGetGuestInfo().
Signed-off-by: Jonathon Jongsma <jjongsma@redhat.com> Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com> Tested-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Jonathon Jongsma [Fri, 23 Aug 2019 16:31:19 +0000 (11:31 -0500)]
qemu: add helper for querying timezone info
This function queries timezone information within the guest and adds
the information to an array of typed parameters with field names
intended to be returned to virDomainGetGuestInfo()
Signed-off-by: Jonathon Jongsma <jjongsma@redhat.com> Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com> Tested-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Jonathon Jongsma [Fri, 23 Aug 2019 16:31:18 +0000 (11:31 -0500)]
qemu: add helper function for querying OS info
This function queries the guest operating system information and adds
the returned information to an array of typed parameters with field
names intended to be returned in virDomainGetGuestInfo().
Signed-off-by: Jonathon Jongsma <jjongsma@redhat.com> Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com> Tested-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Jonathon Jongsma [Fri, 23 Aug 2019 16:31:17 +0000 (11:31 -0500)]
qemu: add helper for getting guest users
This function fetches the list of logged-in users from the qemu agent
and adds them to a list of typed parameters so that they can be used
internally in libvirt.
Also add some basic tests for the function.
Signed-off-by: Jonathon Jongsma <jjongsma@redhat.com> Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com> Tested-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Jonathon Jongsma [Fri, 23 Aug 2019 16:31:15 +0000 (11:31 -0500)]
lib: add virDomainGetGuestInfo()
This API is intended to aggregate several guest agent information
queries and is ispired by stats API virDomainListGetStats(). It is
anticipated that this information will be provided by a guest agent
running within the domain.
Signed-off-by: Jonathon Jongsma <jjongsma@redhat.com> Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com> Tested-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Peter Krempa [Mon, 12 Aug 2019 11:48:51 +0000 (13:48 +0200)]
qemu: snapshot: Restrict file existence check only for local storage
Soon we'll allow more protocols and storage types with snapshots where
we in some cases can't check whether the storage already exists.
Restrict the sanity checks whether the destination images exist or not
for local storage where it's easy. For any other case we will fail
later.
Signed-off-by: Peter Krempa <pkrempa@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com>
Peter Krempa [Fri, 9 Aug 2019 13:15:07 +0000 (15:15 +0200)]
qemu: snapshot: Don't modify persistent XML if disk source is different
While the VM is running the persistent source of a disk might differ
e.g. as the 'newDef' was redefined. Our snapshot code would blindly
rewrite the source of such disk if it shared the 'target'. Fix this by
checking whether the source is the same in the first place.
Signed-off-by: Peter Krempa <pkrempa@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com>
Peter Krempa [Thu, 15 Aug 2019 17:29:43 +0000 (19:29 +0200)]
util: storagefile: Flag backing store strings with authentication
Using inline authentication for storage volumes will not work properly
as libvirt requires use of the secret driver for the auth data and
thus would not be able to represent the passwords stored in the backing
store string.
Make sure that the backing store parsers return 1 which is a sign for
the caller to not use the file in certain cases.
The test data include iscsi via a json pseudo-protocol string and URIs
with the userinfo part being present.
Signed-off-by: Peter Krempa <pkrempa@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com>
Peter Krempa [Fri, 16 Aug 2019 09:34:27 +0000 (11:34 +0200)]
util: storagefile: Don't traverse storage sources unusable by VM
virStorageFileGetMetadataRecurse would include files in the backing
chain which would not really be usable by libvirt directly e.g.
when such file would be promoted to the top layer by an active block
commit as for example inline authentication data can't be represented in
the VM xml file. The idea is to use secrets for this.
With the changes to the backing store string parsers we can report and
propagate if such a thing is present in the configuration and thus start
skipping those files in the backing chain traversal code. This approach
still allows to report the appropriate backing store string in the
storage driver which doesn't directly use the backing file.
Signed-off-by: Peter Krempa <pkrempa@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com>
Peter Krempa [Fri, 16 Aug 2019 09:28:03 +0000 (11:28 +0200)]
util: storagefile: Clarify docs for '@report_broken' of virStorageFileGetMetadata
virStorageFileGetMetadata does not report error if we can't interrogate
the file somehow. Clarify this in the description of the @report_broken
flag as it implies we should report an error in that case. The problem
is that we don't know whether there's a problem and unfortunately just
offload it to qemu.
Signed-off-by: Peter Krempa <pkrempa@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com>
Peter Krempa [Fri, 16 Aug 2019 09:14:31 +0000 (11:14 +0200)]
util: storagefile: Add handling of unusable storage sources
Introduce new semantics to virStorageSourceNewFromBacking and some
of the helpers used by it which propagate the return value from the
callers.
The new return value introduced by this patch allows to notify the
calller that the parsed virStorageSource correctly describes the source
but contains data such as inline authentication which libvirt does not
want to support directly. This means that such file would e.g. unusable
as a storage source (e.g. when actively commiting the overlay to it) or
would not work with blockdev.
The caller will then be able to decide whether to consider this backing
file as viable or just fall back to qemu dealing with it.
Signed-off-by: Peter Krempa <pkrempa@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com>
Peter Krempa [Thu, 15 Aug 2019 14:43:40 +0000 (16:43 +0200)]
util: storagefile: Modify arguments of virStorageSourceNewFromBackingAbsolue
Return the parsed storage source via an pointer in arguments and return
an integer from the function. Describe the semantics with a comment for
the function and adjust callers to the new semantics.
Signed-off-by: Peter Krempa <pkrempa@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com>
Peter Krempa [Thu, 15 Aug 2019 12:49:49 +0000 (14:49 +0200)]
util: storage: Modify return value of virStorageSourceNewFromBacking
Return the storage source definition via a pointer in the arguments and
document the returned values. This will simplify the possibility to
ignore certain backing store types which are not representable by
libvirt.
Signed-off-by: Peter Krempa <pkrempa@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com>
Peter Krempa [Fri, 16 Aug 2019 15:01:10 +0000 (17:01 +0200)]
qemu: alias: Generate 'qomName' of disk with useraliases
Commit fb64e176f4f forgot to delete the check that short-circuits the
disk alias creation if the alias is already present. The side effect
of this is that the creation qomName which is necessary to be able to
refer to disk frontends when -blockdev is used was skipped when user
aliases are used.
Fix it by deleting the check. Also prevent any potential memory leaks
from calling this function repeatedly by creating the qomName only when
it's not present.
Michal Privoznik [Fri, 23 Aug 2019 11:37:25 +0000 (13:37 +0200)]
virsh: Allow graceful console shutdown
Currently, whenever there's a regular EOF on the console stream
or an error the virStreamAbort() is called regardless. While this
may not actually break anything, we should call virStreamFinish()
to let the daemon know we've successfully received all the data
and are shutting down the stream gracefully.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Roman Bolshakov <r.bolshakov@yadro.com>
Michal Privoznik [Fri, 23 Aug 2019 13:11:20 +0000 (15:11 +0200)]
storage_driver: Don't crash in storagePoolCreateXML
In my recent patches I've introduced
virStoragePoolObjIsStarting() which is then used to protect
storage pool definition when the pool object is locked and
unlocked during long running jobs. Well, my patches did not
anticipate that @obj can be NULL under 'cleanup' label in
storagePoolCreateXML() (for instance when parsing XML fails).
This imperfection is causing libvirtd to crash then.
Fixes: 13284a6b83 storage_driver: Protect pool def during startup and build Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Martin Kletzander <mkletzan@redhat.com>
Roman Bolshakov [Wed, 21 Aug 2019 13:33:18 +0000 (16:33 +0300)]
tools: console: Relax stream EOF handling
Regular VM shutdown triggers the error for existing session of virsh
console and it returns with non-zero exit code:
error: internal error: console stream EOF
The message and status code are misleading because there's no real
error. virStreamRecv returns 0 correctly when EOF is reached.
Existing implementations of esx, fd, and remote streams behave the same
for virStreamFinish and virStreamAbort: they close the stream. So, we
can continue to use virStreamAbort to handle EOF and errors from
virStreamRecv but additonally we can report error if virStreamAbort
fails.
Fixes: 29f2b5248c6 ("tools: console: pass stream/fd errors to user") Signed-off-by: Roman Bolshakov <r.bolshakov@yadro.com> Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com> Tested-by: Daniel Henrique Barboza <danielhb413@gmail.com> Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Roman Bolshakov [Wed, 21 Aug 2019 16:13:23 +0000 (19:13 +0300)]
tests: Make references to global symbols indirect in test drivers
A library has to be built with -flat_namespace to get all references to
global symbols indirected. That can also be achieved with two-level
namespace interposition but we're not using explicit symbol
interposition since it's more verbose and requires massive changes to
the mocks.
This provides a way to interpose a mock for virQEMUCapsProbeHostCPU from
qemucpumock and fixes domaincapstest on macOS.
Signed-off-by: Roman Bolshakov <r.bolshakov@yadro.com>
Roman Bolshakov [Wed, 21 Aug 2019 16:13:22 +0000 (19:13 +0300)]
tests: Avoid gnulib replacements in mocks
gnulib headers change stat, lstat and open to replacement functions,
even for function definitions. This effectively disables standard
library overrides in virfilewrapper and virmockstathelpers since they
are never reached.
Rename the functions and provide a declartion that uses correct
assembler name for the mocks.
This fixes firmware lookup in domaincapstest on macOS.
Signed-off-by: Roman Bolshakov <r.bolshakov@yadro.com>
Roman Bolshakov [Wed, 21 Aug 2019 16:13:21 +0000 (19:13 +0300)]
tests: Use flat namespace on macOS
Test executables and mocks have assumption that any symbol can be
replaced with LD_PRELOAD. That's not a case for macOS unless flat
namespace is used, because every external symbol reference records the
library to be looked up. And the symbols cannot be replaced unless dyld
interposing is used.
Setting DYLD_FORCE_FLAT_NAMESPACE changes symbol lookup behaviour to be
similar to Linux dynamic linker. It's more lightweight solution than
explicitly decorating all mock symbols as interpositions and building
libvirt as interposable dynamic library.
This fixes vircryptotest and allows to proceed other tests that rely on
mocks a little bit further.
Signed-off-by: Roman Bolshakov <r.bolshakov@yadro.com>
Roman Bolshakov [Wed, 21 Aug 2019 16:13:20 +0000 (19:13 +0300)]
tests: Lookup extended stat/lstat in mocks
macOS syscall interface (/usr/lib/system/libsystem_kernel.dylib) has
three kinds of stat but only one of them can be used to fill
"struct stat": stat$INODE64.
virmockstathelpers looks up regular stat instead of stat$INODE64. That
causes a failure in qemufirmwaretest because "struct stat" is laid out
differently from the values returned by stat.
Introduce VIR_MOCK_REAL_INIT_ALIASED that can be used to lookup
stat$INODE64 and lstat$INODE64 and use it to setup real functions on
macOS.
Signed-off-by: Roman Bolshakov <r.bolshakov@yadro.com>
Roman Bolshakov [Wed, 21 Aug 2019 16:13:19 +0000 (19:13 +0300)]
build: Use flat namespace for libvirt on macOS
>From ld(1):
By default all references resolved to a dynamic library record the
library to which they were resolved. At runtime, dyld uses that
information to directly resolve symbols. The alternative is to use the
-flat_namespace option. With flat namespace, the library is not
recorded. At runtime, dyld will search each dynamic library in load
order when resolving symbols. This is slower, but more like how other
operating systems resolve symbols.
That fixes the set of tests that preload a mock library to replace
library symbols:
qemublocktest
qemumonitorjsontest
viriscsitest
virmacmaptest
virnetserverclienttest
Signed-off-by: Roman Bolshakov <r.bolshakov@yadro.com>
Roman Bolshakov [Wed, 21 Aug 2019 16:13:18 +0000 (19:13 +0300)]
tests: Drop /private CWD prefix in commandhelper
/tmp is a symbolic link to /private/tmp on macOS. That causes failures
in commandtest, because getcwd returns /private/tmp and the expected
output doesn't match to "CWD: /tmp".
Rathern than making a copy of commanddata solely for macOS, the /private
prefix is stripped.
Signed-off-by: Roman Bolshakov <r.bolshakov@yadro.com>
Roman Bolshakov [Wed, 21 Aug 2019 16:13:17 +0000 (19:13 +0300)]
tests: Remove -module flag for mocks
macOS has two kinds of loadable libraries: MH_BUNDLE, and MH_DYLIB.
bundle is used for plugins that are loaded with dlopen/dlsym/dlclose.
And there's no way to preload a bundle into an application. dynamic
linker (dyld) will reject it when finds it in DYLD_INSERT_LIBRARIES.
Unfortunately, a bundle is built if -module flag is provided to libtool.
The flag has been removed to build dylibs with ".dylib" suffix.
Signed-off-by: Roman Bolshakov <r.bolshakov@yadro.com>
Roman Bolshakov [Wed, 21 Aug 2019 16:13:16 +0000 (19:13 +0300)]
tests: Add lib- prefix to all mocks
In preparation libtool "-module" flag removal, add lib prefix to all
mock shared objects.
While at it, introduce VIR_TEST_MOCK macros that makes path out of mock
name to be used with VIR_TEST_PRELOAD or VIR_TEST_MAIN_PRELOAD. That,
hopefully, improves readability, reduces line length and allows to
tailor VIR_TEST_MOCK for specific platform if it has shared library
suffix different from ".so".
Signed-off-by: Roman Bolshakov <r.bolshakov@yadro.com>
Roman Bolshakov [Wed, 21 Aug 2019 16:13:15 +0000 (19:13 +0300)]
tests: Preload mocks with DYLD_INSERT_LIBRARIES on macOS
LD_PRELOAD has no effect on macOS. Instead, dyld(1) provides a way for
symbol hooking via DYLD_INSERT_LIBRARIES. The variable should contain
colon-separated paths to the dylibs to be inserted.
Signed-off-by: Roman Bolshakov <r.bolshakov@yadro.com>
Roman Bolshakov [Wed, 21 Aug 2019 16:13:14 +0000 (19:13 +0300)]
tests: Avoid IPv4-translated IPv6 address in sockettest
getnameinfo on macOS formats certain IPv6 addresses as IPv4-translated
addresses. The following pattern has been observed:
::ffff is formated as ::0.0.255.255
::fffe is formated as ::0.0.255.254
::ffff:0 is formated as ::255.255.0.0
::fffe:0 is formated as ::255.254.0.0
::ffff:0:0 is formated as ::ffff:0.0.0.0
::fffe:0:0 is formated as ::fffe:0:0
::ffff:0:0:0 is formated as ::ffff:0:0:0
The getnameinfo behavior causes a failure for:
DO_TEST_PARSE_AND_FORMAT("::ffff", AF_UNSPEC, true);
Use non-ambigious IPv6 for parse/format testing.
Signed-off-by: Roman Bolshakov <r.bolshakov@yadro.com>
Roman Bolshakov [Wed, 21 Aug 2019 16:13:13 +0000 (19:13 +0300)]
tests: Don't test octal localhost IP in sockettest on macOS
getaddrinfo on macOS doesn't interpret octal IPv4 addresses. Only
inet_aton can be used for that. Therefore, from macOS standpoint
"0177.0.0.01" is not the same as "127.0.0.1".
The issue was also discovered by python and dotnet core:
https://bugs.python.org/issue27612
https://github.com/dotnet/corefx/issues/8362
Signed-off-by: Roman Bolshakov <r.bolshakov@yadro.com>
After my previous patches we have virPCIDeviceBindToStub() and
virPCIDeviceUnbindFromStub() which really do nothing but call
virPCIDeviceBindToStubWithOverride() and
virPCIDeviceUnbindFromStubWithOverride() respectively.
Drop "WithOverride" from the names and drop the thin wrappers.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Michal Privoznik [Tue, 20 Aug 2019 14:22:08 +0000 (16:22 +0200)]
news: Document KVM assignment removal
Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com> Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com> Tested-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Michal Privoznik [Tue, 20 Aug 2019 11:52:55 +0000 (13:52 +0200)]
virpcimock: Drop @driverActions enum
This enum was introduced to model how RHEL-7 kernel behaves - for
some reason going with the old way (via new_id + bind) fails but
using driver_override succeeds. Well, we don't need to care about
that anymore since we don't create new_id file.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com> Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com> Tested-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Michal Privoznik [Tue, 20 Aug 2019 11:49:54 +0000 (13:49 +0200)]
virpcimock: Don't create new_id or remove_id files
Now that PCI attach/detach happens solely via driver_override
these two files are no longer needed.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com> Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com> Tested-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Michal Privoznik [Tue, 20 Aug 2019 11:31:27 +0000 (13:31 +0200)]
virpcimock: Don't create "pci-stub" driver
Now that nothing supports "pci-stub" driver (aka KVM style of PCI
device assignment) there is no need for virpcimock to create it.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com> Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com> Tested-by: Daniel Henrique Barboza <danielhb413@gmail.com>
projects / libvirt.git / log