When deleting external snapshot we should remove the memory snapshot
file as well.
Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
'reconnect' parameter doesn't pass to qemu properly when
hotplug vhost-user device to vm. Fix this by making
'reconnect' to get correct value.
Signed-off-by: Zhenguo Yao <yaozhenguo1@gmail.com>
Reviewed-by: Jonathon Jongsma <jjongsma@redhat.com>
It makes sense to accept pvpanic-pci also without specified PCI
address and assign one if possible.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=1961326
Signed-off-by: Kristina Hanicova <khanicov@redhat.com>
Reviewed-by: Andrea Bolognani <abologna@redhat.com>
This capability detects the availability of the pvpanic-pci
device that is required in order to use pvpanic on Arm (original
pvpanic is an emulated ISA device, for which Arm does not have
support).
Signed-off-by: Kristina Hanicova <khanicov@redhat.com>
Reviewed-by: Andrea Bolognani <abologna@redhat.com>
It was used briefly and subsequently removed in 3592b81c4c71.
Signed-off-by: Martin Kletzander <mkletzan@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
The 'newapi.xsl' stylesheet was referencing non-existing paths to the
XML files holding ACL permission flags for individual APIs. Additionally
the 'document()' XSL function doesn't even allow concatenation of the
path as it was done via '{$builddir}/src..', but requires either direct
argument or use of the 'concat()' function.
This meant that the 'acls' variable was always empty and thus none of
our API documentation was actually generated with the 'acl' section.
Fix it by passing the path to the XML via an argument to the stylesheet
as the files differ based on which document is being generated.
Since the 'admin' API does not have ACL we need to handle it separately
now in the build system.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Both the object name and permission name in ACL use '-' instead of '_'
separator when referring to them in the docs or even when used inside of
polkit. Unfortunately the generators used for generating our docs don't
honour this in certain cases which would result in broken names in the
API docs (once they will be generated).
Rename both object and permission name to use dash and reflect that in
the anchor names in the documentation.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
In selinux driver there's virSecuritySELinuxSetFileconImpl()
which is responsible for actual setting of SELinux label on given
file and handling possible failures. In fhe failure handling code
we decide whether failure is fatal or not. But there is a bug:
depending on SELinux mode (Permissive vs. Enforcing) the ENOENT
is either ignored or considered fatal. This not correct - ENOENT
must always be fatal for couple of reasons:
- In virSecurityStackTransactionCommit() the seclabels are set
for individual secdrivers (e.g. SELinux first and then DAC),
but if one secdriver succeeds and another one fails, then no
rollback is performed for the successful one leaking remembered
labels.
- QEMU would fail opening the file anyways (if neither of
secdrivers reported error and thus cancelled domain startup)
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=2004850
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Martin Kletzander <mkletzan@redhat.com>
In virSecuritySELinuxSetFileconImpl() we have code that handles
setfilecon_raw() failure. The code consists of two blocks: one
for dealing with shared filesystem like NFS (errno is ENOTSUP or
EROFS) and the other block that's dealing with EPERM for
privileged daemon. Well, the order of these two blocks is a bit
confusing because the comment above them mentions the NFS case
but EPERM block follows. Swap these two blocks to make it less
confusing.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Martin Kletzander <mkletzan@redhat.com>
The way we start passt currently is: we use
virCommandSetPidFile() to use our virCommand machinery to acquire
the PID file and leak opened FD into passt. Then, we use
virPidFile*() APIs to read the PID file (which is needed when
placing it into CGroups or killing it). But this does not fly
really because passt daemonizes itself. Thus the process we
started dies soon and thus the PID file is closed and unlocked.
We could work around this by passing '--foreground' argument, but
that weakens passt as it can't create new PID namespace (because
it doesn't fork()).
The solution is to let passt write the PID file, but since it
does not lock the file and closes it as soon as it is written, we
have to switch to those virPidFile APIs which don't expect PID
file to be locked.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Laine Stump <laine@redhat.com>
There are two places where we kill passt:
1) qemuPasstStop() - called transitively from qemuProcessStop(),
2) qemuPasstStart() - after failed start.
Now, the code from 2) lack error preservation (so if there's
another error during cleanup we might overwrite the original
error). Therefore, move the internals of qemuPasstStop() into a
separate function and call it from both places.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Laine Stump <laine@redhat.com>
When starting passt, it may write something onto its stderr
(convincing it to print even more is addressed later). Pass this
string we read to user.
Since we're not daemonizing passt anymore (see previous commit),
we can let virCommand module do all the heavy lifting and switch
to virCommandSetErrorBuffer() instead of reading error from an
FD.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Stefano Brivio <sbrivio@redhat.com>
Reviewed-by: Laine Stump <laine@redhat.com>
When passt is started, it daemonizes itself by default. There's
no point in having our virCommand module daemonize it too.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Stefano Brivio <sbrivio@redhat.com>
Reviewed-by: Laine Stump <laine@redhat.com>
Fetching whether a node-device is marked for autostart can be allowed
from read-only connections similarly to other objects.
Fixes: c6607a25b93
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
For all other objects we allow the 'read' permission for anonymous
users. In fact the idea is to allow all permissions users using the
readonly connection would have.
This impacts the following APIs (in terms of RPC procedure names):
$ git grep -A 3 node_device:read | grep REMOTE
src/remote/remote_protocol.x- REMOTE_PROC_NODE_DEVICE_GET_XML_DESC = 114,
src/remote/remote_protocol.x- REMOTE_PROC_NODE_DEVICE_GET_PARENT = 115,
src/remote/remote_protocol.x- REMOTE_PROC_NODE_DEVICE_NUM_OF_CAPS = 116,
src/remote/remote_protocol.x- REMOTE_PROC_NODE_DEVICE_LIST_CAPS = 117,
src/remote/remote_protocol.x- REMOTE_PROC_NODE_DEVICE_GET_AUTOSTART = 433,
src/remote/remote_protocol.x- REMOTE_PROC_NODE_DEVICE_IS_PERSISTENT = 435,
src/remote/remote_protocol.x- REMOTE_PROC_NODE_DEVICE_IS_ACTIVE = 436,
Fixes: a93cd08f
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
The way setting up CGroups for external helpers work, is:
qemuExtDevicesHasDevice() is called first to determine whether
there is a helper process running, the CGroup controller is
created and then qemuExtDevicesSetupCgroup() is called to place
helpers into the CGroup. But when one reads just
qemuExtDevicesSetupCgroup() it's easy to miss this hidden logic.
Therefore, add a warning at the beginning of the function.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Laine Stump <laine@redhat.com>
If qemuPasstGetPid() fails, or the passt's PID is -1 then
qemuPasstSetupCgroup() returns early without any error message
set. Report an appropriate error.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Laine Stump <laine@redhat.com>
We can have external helper processes running for domain
<interface/> too (e.g. slirp or passt). But this is not reflected
in qemuExtDevicesHasDevice() which simply ignores these.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Laine Stump <laine@redhat.com>
This reverts commit 0c4e716835eaf2a575bd063fde074c0fc7c4e4d4.
This patch was pushed by my mistake. Even though it got ACKed on
the list, I've raised couple of issues with it. They will be
fixed in next commits.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Laine Stump <laine@redhat.com>
Shutdown of virtlogd prints:
(process:54742): GLib-CRITICAL **: 11:00:40.873: g_regex_unref: assertion 'regex != NULL' failed
Use g_clear_pointer instead which prevents it in the NULL case.
Fixes: 69eeef5dfbf
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
The warning about max_client_requests is hit inside virtlogd every time
a VM starts which spams the logs.
Emit the warning only when the client request limit is not 1 and add a
warning into the daemon config to not configure it too low instead.
Fixes: 031878c2364
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=2145188
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
virNetServerClientDispatchRead checked the return value but it's not
necessary any more as it can't return NULL nowadays.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
The 'pending' state needs to be handled by the blockjob code only when
the snapshot code requests a block-commit without auto-finalization.
If we always handle it we fail to properly remove the blockjob data for
the 'blockdev-create' job as that also transitions trhough 'pending' but
we'd never update it once it reaches 'concluded' as the code already
thinks that the job has finished and is no longer watching it.
Introduce a 'processPending' property into block job data and set it
only when we know that we need to process 'pending'.
Fixes: 90d9bc9d74a5157167548b26c00b1a016655e295
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=2168769
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Pavel Hrdina <phrdina@redhat.com>
Integrated PCI devices can be either PCIe (virtio-iommu) or
conventional PCI (pvpanic-pci). Right now libvirt will refuse
to assign an address on pcie.0 for the latter, but that's an
undesirable limitation that we can easily remove.
Signed-off-by: Andrea Bolognani <abologna@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
virDomainDefAddConsoleCompat in post parsing step appends a stub console
of type VIR_DOMAIN_CHR_TYPE_NULL to ch VMs' Domain XML. Cloud-hypervisor's
deviceValidateCallback (chValidateDomainDeviceDef) checks that the type of
stub console is not of type VIR_DOMAIN_CHR_TYPE_PTY and throws an error.
This commit introduces NO_STUB_CONSOLE feature check to Domain features and
uses it to skip adding stub console to ch VMs.
Signed-off-by: Praveen K Paladugu <prapal@linux.microsoft.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
We only set up host for VIR_DOMAIN_TPM_TYPE_EMULATOR and thus
similarly, we should do cleanup for the same type. This also
fixes a crasher, in which qemuTPMEmulatorCleanupHost() accesses
tpm->data.emulator.storagepath which is NULL for
VIR_DOMAIN_TPM_TYPE_EXTERNAL.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=2168762
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
When running "virsh domcapabilities" on a s390x host, all the CPU
models show up with vendor='unknown' - which sounds kind of weird
since the vendor of these mainframe CPUs is well known: IBM.
All CPUs starting with either "z" or "gen" match a real mainframe
CPU by IBM, so let's return the string "IBM" for those now.
The only remaining ones are now the artifical "qemu" and "max"
models from QEMU itself, so it should be OK to get an "unknown"
vendor for those two.
Reviewed-by: Jiri Denemark <jdenemar@redhat.com>
Signed-off-by: Boris Fiuczynski<fiuczy@linux.ibm.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
I initially had the passt process being started in an identical
fashion to the slirp-helper - libvirt was daemonizing the new process
and recording its pid in a pidfile. The problem with this is that,
since it is daemonized immediately, any startup error in passt happens
after the daemonization, and thus isn't seen by libvirt - libvirt
believes that the process has started successfully and continues on
its merry way. The result was that sometimes a guest would be started,
but there would be no passt process for qemu to use for network
traffic.
Instead, we should be starting passt in the same manner we start
dnsmasq - we just exec it as normal (along with a request that passt
create the pidfile, which is just another option on the passt
commandline) and wait for the child process to exit; passt then has a
chance to parse its commandline and complete all the setup prior to
daemonizing itself; if it encounters an error and exits with a non-0
code, libvirt will see the code and know about the failure. We can
then grab the output from stderr, log that so the "user" has some idea
of what went wrong, and then fail the guest startup.
Signed-off-by: Laine Stump <laine@redhat.com>
Commit 5ef2582646eb98 added emitting of even when refreshign disk state,
where it wanted to avoid sending the event if disk state didn't change.
This was achieved by using 'continue' in the loop filling the
information. Unfortunately this skips extraction of whether the device
has a tray which is propagated into internal structures, which in turn
broke cdrom media change as the code thought there's no tray for the
device.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=2166411
Fixes: 5ef2582646eb98af208ce37355f82bdef39931fa
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Kristina Hanicova <khanicov@redhat.com>
In recent commit of v9.0.0-191-gc71c159248 I've introduced
remoteConnectFormatURI() function and in the function @query
variable. Even though, the variable is used, clang-13 fails to
see it. Surprisingly, newer clang is not affected. Fortunately,
swapping the order in which variables are set makes clang happy
again.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
When handling virConnectOpen(), we parse given URI, specifically
all those parameters we know, like ?mode, ?socket, ?name, etc.
ignoring those we don't recognize yet. Then, we reconstruct the
URI back, but ignoring all parameters we've parsed. In other
words:
qemu:///system?mode=legacy&foo=bar
becomes:
qemu:///system?foo=bar
The reconstructed URI is then passed to the corresponding driver
(QEMU in our example) with intent of it parsing parameters
further (or just ignoring them). But for some transport modes,
where virt-ssh-helper is ran on the remote host (libssh, libssh2,
ssh) we need to pass ?mode and ?socket parameters, so that it can
do the right thing, e.g. for 'mode=legacy' start the monolithic
daemon, or for 'socket=' connect to the given socket.
Resolves: https://gitlab.com/libvirt/libvirt/-/issues/433
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
The aim of this helper is to manipulate the .ignore value for
given list of parameters. For instance:
virURIParamsSetIgnore(uri, false, {"mode", "socket", NULL});
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
There's a piece of code in doRemoteOpen() that is going to be
called twice. Instead of duplicating the code, move it into a
function that will be called twice, later on.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
Similarly to the previous commit, let's accept "socket" parameter
in the connection URI. This change will allow us to use
virt-ssh-helper instead of 'nc' in all cases (done in one of
future commits).
Please note, when the parameter is used it effectively disables
automatic daemon spawning and an error is reported. But this is
intentional - so that the helper behaves just like regular
virConnectOpen() with different transport than ssh, e.g. unix.
But this 'change' is acceptable - there's no way for users to
make our remote code pass the argument to virt-ssh-helper, yet.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
When split daemons were introduced, we also made connection URI
accept new parameter: mode={auto,legacy,direct} so that a client
can force connecting to either old, monolithic daemon, or to
split daemon (see v5.7.0-rc1~257 for more info).
Now, the change was done to the remote driver, but not to
virt-ssh-helper. True, our remote driver code still does not pass
the 'mode' parameter, but that will be addressed in next commits.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
Our own coding style suggest not inventing new names for labels
and stick with 'cleanup' (when the path is used in both,
successful and unsuccessful returns), or 'error' (when the code
below the label is used only upon error). Well, 'failed' label
falls into the latter category. Rename it then.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
The virURIFormat() function either returns a string, or aborts
(on OOM). There's no way this function can return NULL (as of
v7.2.0-rc1~277). Therefore, it doesn't make sense to check its
retval against NULL.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
Our URI handling code (doRemoteOpen() specifically), uses case
insensitive parsing of query part of URI. For instance:
qemu:///system?socket=/some/path
qemu:///system?SoCkEt=/some/path
are the same URI. Even though the latter is probably not used
anywhere, let's switch to STRCASEEQ() instead of STREQ() at two
places: virURIGetParam() and virURICheckUnixSocket().
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
In one of recent commits (v9.0.0-rc1~106) I've made our QEMU
namespace code umount the original /dev. One of the reasons was
enhanced security, because previously we just mounted a tmpfs
over the original /dev. Thus a malicious QEMU could just
umount("/dev") and it would get to the original /dev with all
nodes.
Now, on some systems this introduced a regression:
failed to umount devfs on /dev: Device or resource busy
But how this could be? We've moved all file systems mounted under
/dev to a temporary location. Or have we? As it turns out, not
quite. If there are two file systems mounted on the same target,
e.g. like this:
mount -t tmpfs tmpfs /dev/shm/ && mount -t tmpfs tmpfs /dev/shm/
then only the top most (i.e. the last one) is moved. See
qemuDomainUnshareNamespace() for more info.
Now, we could enhance our code to deal with these "doubled" mount
points. Or, since it is the top most file system that is
accessible anyways (and this one is preserved), we can
umount("/dev") in a recursive fashion.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=2167302
Fixes: 379c0ce4bfed8733dfbde557c359eecc5474ce38
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Jim Fehlig <jfehlig@suse.com>
When going through debug log of a domain startup process, one can
meet the following line:
debug : qemuProcessLaunch:7668 : Building mount namespace
But this is in fact wrong. Firstly, domain namespaces are just
enabled in domain's privateData. Secondly, the debug message says
nothing about actual state of namespace - whether it was enabled
or not.
Therefore, move the debug printing into
qemuProcessEnableDomainNamespaces() and tweak it so that the
actual value is reflected.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Jim Fehlig <jfehlig@suse.com>
Similar to other error paths in qemuDomainUnshareNamespace(), jump to
the cleanup label on umount error instead of directly returning -1.
Signed-off-by: Jim Fehlig <jfehlig@suse.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
When starting a guest, helper processes are started first. But
they need a bit of special handling. Just consider a regular cold
boot and an incoming migration. For instance, in case of swtpm
with its state on a shared volume, we want to set label on the
state for the cold boot case, but don't want to touch the label
in case of incoming migration (because the source very
specifically did not restore it either).
Until now, these two cases were differentiated by testing
@incoming against NULL. And while that makes sense for other
aspects of domain startup, for external devices we need a bit
more, because a restore from a save file is also 'incoming
migration'.
Now, there is a difference between regular migration and restore
from a save file. In the former case we do not want to set
seclabels in the save state. BUT, in the latter case we do need
to set them, because the code that saves the machine restored
seclabels.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=2161557
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
When stopping swtpm we can restore the label either on just the
swtpm's domain specific logfile (/var/log/swtpm/libvirt/qemu/...),
or on the logfile and the state too (/var/lib/libvirt/swtpm/...).
The deciding factor is whether the guest is stopped because of
outgoing migration OR the state is on a shared filesystem.
But this is not correct condition, because for instance saving the
guest into a file (virsh save) is also an outgoing migration.
Alternatively, when the swtpm state is stored on a shared
filesystem, but the guest is destroyed (virsh destroy), i.e.
stopped because of different reason than migration, we want to
restore the seclabels.
The correct condition is: skip restoring the state on outgoing
migration AND shared filesystem.
Resolves: https://bugzilla.redhat.com/show_bug.cgi?id=2161557
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
When cleaning up host in qemuProcessStop(), our external helper
processes (e.g. swtpm) want to know whether the domain is being
migrated out or not (so that they restore seclabels on a device
state that's on a shared storage).
This fact is reflected in the @outgoingMigration variable which
is set to true if asyncJob is anything but
VIR_ASYNC_JOB_MIGRATION_IN. Well, we have a specific job for
outgoing migration (VIR_ASYNC_JOB_MIGRATION_OUT) and thus we
should check for that.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Actually use the log cleaner introduced by previous commit.
Signed-off-by: Oleg Vasilev <oleg.vasilev@virtuozzo.com>
Reviewed-by: Martin Kletzander <mkletzan@redhat.com>
Before, logs from deleted machines have been piling up, since there were
no garbage collection mechanism. Now, virtlogd can be configured to
periodically scan the log folder for orphan logs with no recent modifications
and delete it.
A single chain of recent and rotated logs is deleted in a single transaction.
Signed-off-by: Oleg Vasilev <oleg.vasilev@virtuozzo.com>
Reviewed-by: Martin Kletzander <mkletzan@redhat.com>
We want to specify the folder to clean and how much time can a log
chain live.
Signed-off-by: Oleg Vasilev <oleg.vasilev@virtuozzo.com>
Reviewed-by: Martin Kletzander <mkletzan@redhat.com>
