This is a partial revert of 87a43a907f
The change to use g_clear_pointer() in more places was accidentally
applied to cases involving vir_g_source_unref().
In some cases, the ordering of g_source_destroy() and
vir_g_source_unref() was reversed, which resulted in the source being
marked as destroyed, after it is already unreferenced. This
use-after-free case might work in many cases, but with versions of
glib older than 2.64.0 it may defer unref to run within the main
thread to avoid a race condition, which creates a large distance
between the g_source_unref() and g_source_destroy().
In some cases, the call to vir_g_source_unref() was replaced with a
second call to g_source_destroy(), leading to a memory leak or worse.
In our experience, the symptoms were that use of libvirt-python became
slower over time, with OpenStack nova-compute initially taking around
one second to periodically query the host PCI devices, and within an
hour it was taking over a minute to complete the same operation, until
it is was eventually running this query back-to-back, resulting in the
nova-compute process consuming 100% of one CPU thread, losing its
RabbitMQ connection frequently, and showing up as down to the control
plane.
Signed-off-by: Mark Mielke <mark.mielke@gmail.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Signed-off-by: Ján Tomko <jtomko@redhat.com>
This command tells QEMU to start listening for an incoming post-copy
recovery connection. Just like migrate-incoming is used for starting
fresh migration on the destination host.
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Pavel Hrdina <phrdina@redhat.com>
When connection breaks during post-copy migration, QEMU enters
'postcopy-paused' state. We need to handle this state and make the
situation visible to upper layers.
Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
None of the callers now uses the slirp fd passing feature, so it can be
removed.
At this point even the VIR_DEBUG doesn't make sense as it would only log
the pointer of 'props'.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Jonathon Jongsma <jjongsma@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
All callers now pass NULL/0 as arguments for vhostfd passing so we can
remove all the associated code.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Jonathon Jongsma <jjongsma@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
The only caller doesn't use the fdset info any more.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Jonathon Jongsma <jjongsma@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Add support for sending one FD from the client along with a monitor
command so that it's possible to use 'getfd' and 'add-fd' to use FDs
passed from the client with other QMP commands.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
There are some enums that are declared in qemu_monitor.h but
implemented in qemu_monitor_json.c. While from compiler and
linker POV it doesn't matter, the code is cleaner if an enum is
implemented in .c file that corresponds to .h file which declared
the enum.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Andrea Bolognani <abologna@redhat.com>
Add mode parameter to qemuDomainStartDirtyRateCalc API, 'mode'
option of 'calc-dirty-rate' command was introduced since
qemu >= 6.2.
Signed-off-by: Hyman Huang(黄勇) <huangy81@chinatelecom.cn>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
We don't want to be dealing with real FDs thus we mock
'qemuMonitorIOWriteWithFD' to do the same thing as when no FD is being
passed.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Adding an exception for the whole file usually defeats the purpose of a
syntax check and is also likely to get forgotten once the file is
removed.
In case of the suggestion of using 'safewrite' instead of write even the
comment for safewrite states that the function needs to be used only in
certain cases.
Remove the blanket exceptions for files and use an exclude string
instead. The only instance where we keep the full file exception is for
src/libvirt-stream.c as there are multiple uses in example code in
comments where I couldn't find a nicer targetted wapproach.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
In order to mock the SCM_RIGHTS sendmsg to simulate sending
filedescriptors to fake qemu in tests we need access to some fields of
'struct _qemuMonitor'. Move its declaration to the private header file.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
'qemuMonitorRemoveFdset' validates that the 'fdset' argument isn't less
than 0. We can turn it to unsigned and thus avoid the error message
completely.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Use VIR_AUTOFREE for the temp socket so that the 'error:' label can be
removed.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Tim Wiederhake <twiederh@redhat.com>
This change was generated using the following spatch:
@ rule1 @
expression a;
identifier f;
@@
<...
- f(*a);
... when != a;
- *a = NULL;
+ g_clear_pointer(a, f);
...>
@ rule2 @
expression a;
identifier f;
@@
<...
- f(a);
... when != a;
- a = NULL;
+ g_clear_pointer(&a, f);
...>
Then, I left some of the changes out, like tools/nss/ (which
doesn't link with glib) and put back a comment in
qemuBlockJobProcessEventCompletedActiveCommit() which coccinelle
decided to remove (I have no idea why).
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Convert all code using the 'QOM_CPU_PATH' macro to accept the QOM path
as an argument.
For now the new helper for fetching the path 'qemuProcessGetVCPUQOMPath'
will always return the same hard-coded value.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
This change adds the domain name in the error and debug logs during
monitor IO processing so that we may infer which VM experienced
errors such as IO or socket hangup. This may help in debugging
monitor IO errors.
Signed-off-by: Rohit Kumar <rohit.kumar3@nutanix.com>
Reviewed-by: Ani Sinha <ani@anisinha.ca>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
Set a launch secret in guest memory using the sev-inject-launch-secret
QMP API. Only supported with qemu >= 6.0.0 and SEV-enabled guests in a
paused state.
Signed-off-by: Jim Fehlig <jfehlig@suse.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
We're only returning the set of fields needed to perform an
attestation, per the SEV API docs.
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Forces the data to be written synchronously to both the original and the
mirrored images which ensures that the job will reach synchronized
phase.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
We use this approach for other APIs which take a virJSONValue as
argument and the logic is also simpler.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Recent cleanup of snapshot revert code made these function unused.
Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
virJSONValueObjectAdd now works identically to virJSONValueObjectCreate
when used with a NULL argument. Replace all callers.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Pass in the double pointer from the wrappers directly to
virJSONValueObjectAddVArgs, which will allow us to directly allocate the
new objects inside the function.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
The upcoming QEMU 6.2.0 implements a new event called
DEVICE_UNPLUG_GUEST_ERROR, a new event that reports generic device
unplug errors that were detected by the guest and reported back to QEMU.
This new event is going to be specially useful for pseries guests that
uses newer kernels (must have kernel commit 29c9a2699e71), which is the
case for Fedora 34 at this moment. These guests have the capability of
reporting CPU removal errors back to QEMU which, starting in 6.2.0, will
emit the DEVICE_UNPLUG_GUEST_ERROR event. Libvirt can use this event to
abort the device removal immediately instead of waiting for 'setvcpus'
timeout.
QEMU 6.2.0 is also going to emit DEVICE_UNPLUG_GUEST_ERROR for memory
hotunplug errors, both in pseries and ACPI guests. QEMU 6.1.0 reports
memory removal errors using the MEM_UNPLUG_ERROR event, which is going to
be deprecated by DEVICE_UNPLUG_GUEST_ERROR in 6.2.0. Given that
Libvirt wasn't handling the MEM_UNPLUG_ERROR event we don't need to
worry about it - adding support to DEVICE_UNPLUG_GUEST_ERROR will be
enough to cover all future cases.
This patch adds support to DEVICE_UNPLUG_GUEST_ERROR by adding the
minimal wiring required for Libvirt to be aware of it. The monitor
callback for this event will abort the pending removal operation of the
device reported by the "device" property of the event. Most of the heavy
lifting is already done by existing code that handles
QEMU_DOMAIN_UNPLUGGING_DEVICE_STATUS_GUEST_REJECTED, making our life
easier to abort the pending removal operation.
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Similarly to one of previous commits, there's no need to
increment domain object refcounter before unlocking it. Any
number of lock and unlock calls over domain object has no effect
on the refcounter.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Use the 'query-nodes' flag to return all stats. The flag was introduced
prior to qemu-2.11 so we can always use it, but we invoke it only when
querying stats. The other invocation is used for detecting the nodenames
which is fragile code.
The images without a frontend don't have the device field so the
extraction code checks need to be relaxed.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
In qemuMonitorJSONCommandWithFd() given command (represented by
virJSONValue struct) is translated to string (represented by
virBuffer). The ownership of the string is then transferred to
the message which is then sent. The downside of this approach is
we have to have an explicit call to free the string from the
message. But if the message just "borrowed" the string (which it
can safely do because it is just reading from the string) then
automatic free of the buffer takes care of freeing the string.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Tim Wiederhake <twiederh@redhat.com>
In a few places we declare a variable (which is optionally
followed by a code not touching it) then set the variable to a
value and return the variable immediately. It's obvious that the
variable is needless and the value can be returned directly
instead.
This patch was generated using this semantic patch:
@@
type T;
identifier ret;
expression E;
@@
- T ret;
... when != ret
when strict
- ret = E;
- return ret;
+ return E;
After that I fixed couple of formatting issues because coccinelle
formatted some lines differently than our coding style.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
The qemuMonitorJSONGetMemoryDeviceInfo() command executes
'query-memory-devices' command and returns early if QEMU
doesn't know the command. Well, the command was introduced in
QEMU release 2.1 (specifically in commit v2.1.0-rc0~41^2~9) and
since the minimum required version is 2.11.0 we can be sure that
command will always exist.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
The qemuMonitorJSONGetMigrationParams() function executes
'query-migrate-parameters' command and returns early if QEMU
doesn't know the command. Well, the command was introduced in
QEMU release 2.4 (specifically in commit v2.4.0-rc0~147^2~3) and
since the minimum required version is 2.11.0 we can be sure that
the command will always exist.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Always fetch the stats for all backing chain members. Callers from
qemu_driver.c already always passed 'true' and the caller from the
migration code won't mind when we fetch all stats.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
All (proper) callers pass true so we can remove the argument.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
There have been countless reports from users concerned about the following
error reported by libvirtd when qemu domains are shutdown
internal error: End of file from qemu monitor
While the error is harmless, users often mistaken it for real problem with
their deployments. EOF from the monitor can't be entirely ignored since
other threads may be using the monitor and must be able to detect the EOF
condition.
One potential fix is to delay reporting EOF until the monitor is used
after EOF is detected. This patch adds a 'goteof' member to the
qemuMonitor structure, which is set when EOF is detected on the monitor
socket. If another thread later tries to send data on the monitor, the
EOF error is reported.
Signed-off-by: Jim Fehlig <jfehlig@suse.com>
Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
Remove the old-style 'device_add' helpers which parse the commandline
arguments to JSON since we now coverted all usage to use JSON directly.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
We commonly use 'props' for the JSON object describing something. Rename
the monitor device addition code.
Additionally the common approach is to clear the pointer if it was
consumed so the arguments are adjusted to do so.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
Unify it with the upcoming capabilities for -netdev and -device.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
As advertised in previous commit, this event is delivered to us
when virtio-mem module changes the allocation inside the guest.
It comes with one attribute - size - which holds the new size of
the virtio-mem (well, allocated size), in bytes.
Mind you, this is not necessarily the same number as 'requested
size'. It almost certainly will be when sizing the memory up, but
it might not be when sizing the memory down - the guest kernel
might be unable to free some blocks.
This current size is reported in the domain XML as an output
element only.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
As advertised in one of previous commits, we want to be able to
change 'requested-size' attribute of virtio-mem on the fly. This
commit does exactly that. Changing anything else is checked for
and forbidden.
Once guest has changed the allocation, QEMU emits an event which
we will use to track the allocation. In the next commit.
Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
All supported QEMU versions have all the fields so we can remove the
booleans controlling which fields are used on the monitor.
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
Reviewed-by: Martin Kletzander <mkletzan@redhat.com>
Reviewed-by: Ján Tomko <jtomko@redhat.com>
It was added by commit c2121602 and later removed by 5a4c2374a
Signed-off-by: Ján Tomko <jtomko@redhat.com>
Reviewed-by: Peter Krempa <pkrempa@redhat.com>
