2011-09-02 21:34:48 +04:00
|
|
|
/*
|
2019-06-13 18:33:58 +03:00
|
|
|
* QEMU Management Protocol commands
|
2011-09-02 21:34:48 +04:00
|
|
|
*
|
|
|
|
* Copyright IBM, Corp. 2011
|
|
|
|
*
|
|
|
|
* Authors:
|
|
|
|
* Anthony Liguori <aliguori@us.ibm.com>
|
|
|
|
*
|
|
|
|
* This work is licensed under the terms of the GNU GPL, version 2. See
|
|
|
|
* the COPYING file in the top-level directory.
|
|
|
|
*
|
2012-01-13 20:44:23 +04:00
|
|
|
* Contributions after 2012-01-13 are licensed under the terms of the
|
|
|
|
* GNU GPL, version 2 or (at your option) any later version.
|
2011-09-02 21:34:48 +04:00
|
|
|
*/
|
|
|
|
|
2016-01-29 20:50:05 +03:00
|
|
|
#include "qemu/osdep.h"
|
2016-03-20 20:16:19 +03:00
|
|
|
#include "qemu/cutils.h"
|
2018-02-01 14:18:46 +03:00
|
|
|
#include "qemu/option.h"
|
2015-03-17 20:16:21 +03:00
|
|
|
#include "monitor/monitor.h"
|
2012-12-17 21:20:04 +04:00
|
|
|
#include "sysemu/sysemu.h"
|
2016-12-12 20:22:24 +03:00
|
|
|
#include "qemu/config-file.h"
|
2016-09-21 07:27:14 +03:00
|
|
|
#include "qemu/uuid.h"
|
2017-01-26 16:19:46 +03:00
|
|
|
#include "chardev/char.h"
|
2011-12-07 17:17:51 +04:00
|
|
|
#include "ui/qemu-spice.h"
|
2021-01-04 19:12:00 +03:00
|
|
|
#include "ui/console.h"
|
2021-10-09 23:16:57 +03:00
|
|
|
#include "ui/dbus-display.h"
|
2012-12-17 21:20:04 +04:00
|
|
|
#include "sysemu/kvm.h"
|
2019-08-12 08:23:59 +03:00
|
|
|
#include "sysemu/runstate.h"
|
2020-12-11 19:52:43 +03:00
|
|
|
#include "sysemu/runstate-action.h"
|
2012-12-17 21:20:04 +04:00
|
|
|
#include "sysemu/blockdev.h"
|
2015-10-19 18:53:22 +03:00
|
|
|
#include "sysemu/block-backend.h"
|
2018-02-01 14:18:31 +03:00
|
|
|
#include "qapi/error.h"
|
2020-09-13 22:53:47 +03:00
|
|
|
#include "qapi/qapi-commands-acpi.h"
|
2020-02-24 17:29:53 +03:00
|
|
|
#include "qapi/qapi-commands-block.h"
|
2020-01-29 13:22:37 +03:00
|
|
|
#include "qapi/qapi-commands-control.h"
|
2019-06-19 23:10:41 +03:00
|
|
|
#include "qapi/qapi-commands-machine.h"
|
2018-02-27 02:13:27 +03:00
|
|
|
#include "qapi/qapi-commands-misc.h"
|
qmp: Support for querying stats
Gathering statistics is important for development, for monitoring and
for performance measurement. There are tools such as kvm_stat that do
this and they rely on the _user_ knowing the interesting data points
rather than the tool (which can treat them as opaque).
The commands introduced in this commit introduce QMP support for
querying stats; the goal is to take the capabilities of these tools
and making them available throughout the whole virtualization stack,
so that one can observe, monitor and measure virtual machines without
having shell access + root on the host that runs them.
query-stats returns a list of all stats per target type (only VM
and vCPU to start); future commits add extra options for specifying
stat names, vCPU qom paths, and providers. All these are used by the
HMP command "info stats". Because of the development usecases around
statistics, a good HMP interface is important.
query-stats-schemas returns a list of stats included in each target
type, with an option for specifying the provider. The concepts in the
schema are based on the KVM binary stats' own introspection data, just
translated to QAPI.
There are two reasons to have a separate schema that is not tied to
the QAPI schema. The first is the contents of the schemas: the new
introspection data provides different information than the QAPI data,
namely unit of measurement, how the numbers are gathered and change
(peak/instant/cumulative/histogram), and histogram bucket sizes.
There's really no reason to have this kind of metadata in the QAPI
introspection schema (except possibly for the unit of measure, but
there's a very weak justification).
Another reason is the dynamicity of the schema. The QAPI introspection
data is very much static; and while QOM is somewhat more dynamic,
generally we consider that to be a bug rather than a feature these days.
On the other hand, the statistics that are exposed by QEMU might be
passed through from another source, such as KVM, and the disadvantages of
manually updating the QAPI schema for outweight the benefits from vetting
the statistics and filtering out anything that seems "too unstable".
Running old QEMU with new kernel is a supported usecase; if old QEMU
cannot expose statistics from a new kernel, or if a kernel developer
needs to change QEMU before gathering new info from the new kernel,
then that is a poor user interface.
The framework provides a method to register callbacks for these QMP
commands. Most of the work in fact is done by the callbacks, and a
large majority of this patch is new QAPI structs and commands.
Examples (with KVM stats):
- Query all VM stats:
{ "execute": "query-stats", "arguments" : { "target": "vm" } }
{ "return": [
{ "provider": "kvm",
"stats": [
{ "name": "max_mmu_page_hash_collisions", "value": 0 },
{ "name": "max_mmu_rmap_size", "value": 0 },
{ "name": "nx_lpage_splits", "value": 148 },
... ] },
{ "provider": "xyz",
"stats": [ ... ] }
] }
- Query all vCPU stats:
{ "execute": "query-stats", "arguments" : { "target": "vcpu" } }
{ "return": [
{ "provider": "kvm",
"qom_path": "/machine/unattached/device[0]"
"stats": [
{ "name": "guest_mode", "value": 0 },
{ "name": "directed_yield_successful", "value": 0 },
{ "name": "directed_yield_attempted", "value": 106 },
... ] },
{ "provider": "kvm",
"qom_path": "/machine/unattached/device[1]"
"stats": [
{ "name": "guest_mode", "value": 0 },
{ "name": "directed_yield_successful", "value": 0 },
{ "name": "directed_yield_attempted", "value": 106 },
... ] },
] }
- Retrieve the schemas:
{ "execute": "query-stats-schemas" }
{ "return": [
{ "provider": "kvm",
"target": "vcpu",
"stats": [
{ "name": "guest_mode",
"unit": "none",
"base": 10,
"exponent": 0,
"type": "instant" },
{ "name": "directed_yield_successful",
"unit": "none",
"base": 10,
"exponent": 0,
"type": "cumulative" },
... ]
},
{ "provider": "kvm",
"target": "vm",
"stats": [
{ "name": "max_mmu_page_hash_collisions",
"unit": "none",
"base": 10,
"exponent": 0,
"type": "peak" },
... ]
},
{ "provider": "xyz",
"target": "vm",
"stats": [ ... ]
}
] }
Signed-off-by: Mark Kanda <mark.kanda@oracle.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2022-02-15 18:04:31 +03:00
|
|
|
#include "qapi/qapi-commands-stats.h"
|
2018-02-27 02:13:27 +03:00
|
|
|
#include "qapi/qapi-commands-ui.h"
|
2021-09-08 12:35:43 +03:00
|
|
|
#include "qapi/type-helpers.h"
|
2015-03-17 19:22:46 +03:00
|
|
|
#include "qapi/qmp/qerror.h"
|
2021-09-08 12:35:43 +03:00
|
|
|
#include "exec/ramlist.h"
|
2018-04-23 19:51:16 +03:00
|
|
|
#include "hw/mem/memory-device.h"
|
2014-06-16 21:12:28 +04:00
|
|
|
#include "hw/acpi/acpi_dev_interface.h"
|
2021-09-08 12:35:43 +03:00
|
|
|
#include "hw/intc/intc.h"
|
2021-09-08 12:35:43 +03:00
|
|
|
#include "hw/rdma/rdma.h"
|
qmp: Support for querying stats
Gathering statistics is important for development, for monitoring and
for performance measurement. There are tools such as kvm_stat that do
this and they rely on the _user_ knowing the interesting data points
rather than the tool (which can treat them as opaque).
The commands introduced in this commit introduce QMP support for
querying stats; the goal is to take the capabilities of these tools
and making them available throughout the whole virtualization stack,
so that one can observe, monitor and measure virtual machines without
having shell access + root on the host that runs them.
query-stats returns a list of all stats per target type (only VM
and vCPU to start); future commits add extra options for specifying
stat names, vCPU qom paths, and providers. All these are used by the
HMP command "info stats". Because of the development usecases around
statistics, a good HMP interface is important.
query-stats-schemas returns a list of stats included in each target
type, with an option for specifying the provider. The concepts in the
schema are based on the KVM binary stats' own introspection data, just
translated to QAPI.
There are two reasons to have a separate schema that is not tied to
the QAPI schema. The first is the contents of the schemas: the new
introspection data provides different information than the QAPI data,
namely unit of measurement, how the numbers are gathered and change
(peak/instant/cumulative/histogram), and histogram bucket sizes.
There's really no reason to have this kind of metadata in the QAPI
introspection schema (except possibly for the unit of measure, but
there's a very weak justification).
Another reason is the dynamicity of the schema. The QAPI introspection
data is very much static; and while QOM is somewhat more dynamic,
generally we consider that to be a bug rather than a feature these days.
On the other hand, the statistics that are exposed by QEMU might be
passed through from another source, such as KVM, and the disadvantages of
manually updating the QAPI schema for outweight the benefits from vetting
the statistics and filtering out anything that seems "too unstable".
Running old QEMU with new kernel is a supported usecase; if old QEMU
cannot expose statistics from a new kernel, or if a kernel developer
needs to change QEMU before gathering new info from the new kernel,
then that is a poor user interface.
The framework provides a method to register callbacks for these QMP
commands. Most of the work in fact is done by the callbacks, and a
large majority of this patch is new QAPI structs and commands.
Examples (with KVM stats):
- Query all VM stats:
{ "execute": "query-stats", "arguments" : { "target": "vm" } }
{ "return": [
{ "provider": "kvm",
"stats": [
{ "name": "max_mmu_page_hash_collisions", "value": 0 },
{ "name": "max_mmu_rmap_size", "value": 0 },
{ "name": "nx_lpage_splits", "value": 148 },
... ] },
{ "provider": "xyz",
"stats": [ ... ] }
] }
- Query all vCPU stats:
{ "execute": "query-stats", "arguments" : { "target": "vcpu" } }
{ "return": [
{ "provider": "kvm",
"qom_path": "/machine/unattached/device[0]"
"stats": [
{ "name": "guest_mode", "value": 0 },
{ "name": "directed_yield_successful", "value": 0 },
{ "name": "directed_yield_attempted", "value": 106 },
... ] },
{ "provider": "kvm",
"qom_path": "/machine/unattached/device[1]"
"stats": [
{ "name": "guest_mode", "value": 0 },
{ "name": "directed_yield_successful", "value": 0 },
{ "name": "directed_yield_attempted", "value": 106 },
... ] },
] }
- Retrieve the schemas:
{ "execute": "query-stats-schemas" }
{ "return": [
{ "provider": "kvm",
"target": "vcpu",
"stats": [
{ "name": "guest_mode",
"unit": "none",
"base": 10,
"exponent": 0,
"type": "instant" },
{ "name": "directed_yield_successful",
"unit": "none",
"base": 10,
"exponent": 0,
"type": "cumulative" },
... ]
},
{ "provider": "kvm",
"target": "vm",
"stats": [
{ "name": "max_mmu_page_hash_collisions",
"unit": "none",
"base": 10,
"exponent": 0,
"type": "peak" },
... ]
},
{ "provider": "xyz",
"target": "vm",
"stats": [ ... ]
}
] }
Signed-off-by: Mark Kanda <mark.kanda@oracle.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2022-02-15 18:04:31 +03:00
|
|
|
#include "monitor/stats.h"
|
2011-09-02 21:34:48 +04:00
|
|
|
|
|
|
|
NameInfo *qmp_query_name(Error **errp)
|
|
|
|
{
|
|
|
|
NameInfo *info = g_malloc0(sizeof(*info));
|
|
|
|
|
2022-11-04 19:06:59 +03:00
|
|
|
info->name = g_strdup(qemu_name);
|
2011-09-02 21:34:48 +04:00
|
|
|
return info;
|
|
|
|
}
|
2011-08-27 00:38:13 +04:00
|
|
|
|
2011-09-12 22:10:53 +04:00
|
|
|
KvmInfo *qmp_query_kvm(Error **errp)
|
|
|
|
{
|
|
|
|
KvmInfo *info = g_malloc0(sizeof(*info));
|
|
|
|
|
|
|
|
info->enabled = kvm_enabled();
|
2021-07-30 13:59:41 +03:00
|
|
|
info->present = accel_find("kvm");
|
2011-09-12 22:10:53 +04:00
|
|
|
|
|
|
|
return info;
|
|
|
|
}
|
|
|
|
|
2011-09-14 00:16:25 +04:00
|
|
|
UuidInfo *qmp_query_uuid(Error **errp)
|
|
|
|
{
|
|
|
|
UuidInfo *info = g_malloc0(sizeof(*info));
|
|
|
|
|
2016-09-21 07:27:22 +03:00
|
|
|
info->UUID = qemu_uuid_unparse_strdup(&qemu_uuid);
|
2011-09-14 00:16:25 +04:00
|
|
|
return info;
|
|
|
|
}
|
|
|
|
|
2014-05-02 15:26:31 +04:00
|
|
|
void qmp_quit(Error **errp)
|
2011-09-15 21:20:28 +04:00
|
|
|
{
|
2020-12-11 19:52:43 +03:00
|
|
|
shutdown_action = SHUTDOWN_ACTION_POWEROFF;
|
2018-12-05 14:01:31 +03:00
|
|
|
qemu_system_shutdown_request(SHUTDOWN_CAUSE_HOST_QMP_QUIT);
|
2011-09-15 21:20:28 +04:00
|
|
|
}
|
|
|
|
|
2011-09-15 21:34:39 +04:00
|
|
|
void qmp_stop(Error **errp)
|
|
|
|
{
|
2016-02-18 08:16:49 +03:00
|
|
|
/* if there is a dump in background, we should wait until the dump
|
|
|
|
* finished */
|
2022-03-23 18:57:31 +03:00
|
|
|
if (qemu_system_dump_in_progress()) {
|
2016-02-18 08:16:49 +03:00
|
|
|
error_setg(errp, "There is a dump in process, please wait.");
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
2012-10-23 16:54:21 +04:00
|
|
|
if (runstate_check(RUN_STATE_INMIGRATE)) {
|
|
|
|
autostart = 0;
|
|
|
|
} else {
|
|
|
|
vm_stop(RUN_STATE_PAUSED);
|
|
|
|
}
|
2011-09-15 21:34:39 +04:00
|
|
|
}
|
|
|
|
|
2011-09-15 21:41:46 +04:00
|
|
|
void qmp_system_reset(Error **errp)
|
|
|
|
{
|
2018-12-05 14:01:31 +03:00
|
|
|
qemu_system_reset_request(SHUTDOWN_CAUSE_HOST_QMP_SYSTEM_RESET);
|
2011-09-15 21:41:46 +04:00
|
|
|
}
|
2011-09-28 18:06:15 +04:00
|
|
|
|
2019-12-05 20:46:27 +03:00
|
|
|
void qmp_system_powerdown(Error **errp)
|
2011-09-28 18:06:15 +04:00
|
|
|
{
|
|
|
|
qemu_system_powerdown_request();
|
|
|
|
}
|
2011-10-06 21:31:39 +04:00
|
|
|
|
2011-11-22 23:58:31 +04:00
|
|
|
void qmp_cont(Error **errp)
|
|
|
|
{
|
2015-10-19 18:53:22 +03:00
|
|
|
BlockBackend *blk;
|
2019-06-06 18:41:30 +03:00
|
|
|
BlockJob *job;
|
2017-06-23 19:24:15 +03:00
|
|
|
Error *local_err = NULL;
|
2011-11-22 23:58:31 +04:00
|
|
|
|
2016-02-18 08:16:49 +03:00
|
|
|
/* if there is a dump in background, we should wait until the dump
|
|
|
|
* finished */
|
2022-03-23 18:57:31 +03:00
|
|
|
if (qemu_system_dump_in_progress()) {
|
2016-02-18 08:16:49 +03:00
|
|
|
error_setg(errp, "There is a dump in process, please wait.");
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
2013-04-26 07:24:40 +04:00
|
|
|
if (runstate_needs_reset()) {
|
2014-03-22 03:42:26 +04:00
|
|
|
error_setg(errp, "Resetting the Virtual Machine is required");
|
2011-11-22 23:58:31 +04:00
|
|
|
return;
|
2012-04-27 20:33:36 +04:00
|
|
|
} else if (runstate_check(RUN_STATE_SUSPENDED)) {
|
|
|
|
return;
|
2019-01-24 15:25:24 +03:00
|
|
|
} else if (runstate_check(RUN_STATE_FINISH_MIGRATE)) {
|
|
|
|
error_setg(errp, "Migration is not finalized yet");
|
|
|
|
return;
|
2011-11-22 23:58:31 +04:00
|
|
|
}
|
|
|
|
|
2015-10-19 18:53:22 +03:00
|
|
|
for (blk = blk_next(NULL); blk; blk = blk_next(blk)) {
|
|
|
|
blk_iostatus_reset(blk);
|
2014-05-02 15:26:42 +04:00
|
|
|
}
|
2016-03-22 20:58:50 +03:00
|
|
|
|
2022-09-26 12:32:04 +03:00
|
|
|
WITH_JOB_LOCK_GUARD() {
|
|
|
|
for (job = block_job_next_locked(NULL); job;
|
|
|
|
job = block_job_next_locked(job)) {
|
|
|
|
block_job_iostatus_reset_locked(job);
|
|
|
|
}
|
2019-06-06 18:41:30 +03:00
|
|
|
}
|
|
|
|
|
2015-12-22 16:07:08 +03:00
|
|
|
/* Continuing after completed migration. Images have been inactivated to
|
2017-05-04 19:52:36 +03:00
|
|
|
* allow the destination to take control. Need to get control back now.
|
|
|
|
*
|
|
|
|
* If there are no inactive block nodes (e.g. because the VM was just
|
|
|
|
* paused rather than completing a migration), bdrv_inactivate_all() simply
|
|
|
|
* doesn't do anything. */
|
2022-02-09 13:54:51 +03:00
|
|
|
bdrv_activate_all(&local_err);
|
2017-05-04 19:52:36 +03:00
|
|
|
if (local_err) {
|
|
|
|
error_propagate(errp, local_err);
|
|
|
|
return;
|
2015-12-22 16:07:08 +03:00
|
|
|
}
|
|
|
|
|
2012-10-23 16:54:21 +04:00
|
|
|
if (runstate_check(RUN_STATE_INMIGRATE)) {
|
|
|
|
autostart = 1;
|
|
|
|
} else {
|
|
|
|
vm_start();
|
|
|
|
}
|
2011-11-22 23:58:31 +04:00
|
|
|
}
|
2011-12-13 00:29:34 +04:00
|
|
|
|
2012-02-23 16:45:21 +04:00
|
|
|
void qmp_system_wakeup(Error **errp)
|
|
|
|
{
|
qmp hmp: Make system_wakeup check wake-up support and run state
The qmp/hmp command 'system_wakeup' is simply a direct call to
'qemu_system_wakeup_request' from vl.c. This function verifies if
runstate is SUSPENDED and if the wake up reason is valid before
proceeding. However, no error or warning is thrown if any of those
pre-requirements isn't met. There is no way for the caller to
differentiate between a successful wakeup or an error state caused
when trying to wake up a guest that wasn't suspended.
This means that system_wakeup is silently failing, which can be
considered a bug. Adding error handling isn't an API break in this
case - applications that didn't check the result will remain broken,
the ones that check it will have a chance to deal with it.
Adding to that, the commit before previous created a new QMP API called
query-current-machine, with a new flag called wakeup-suspend-support,
that indicates if the guest has the capability of waking up from suspended
state. Although such guest will never reach SUSPENDED state and erroring
it out in this scenario would suffice, it is more informative for the user
to differentiate between a failure because the guest isn't suspended versus
a failure because the guest does not have support for wake up at all.
All this considered, this patch changes qmp_system_wakeup to check if
the guest is capable of waking up from suspend, and if it is suspended.
After this patch, this is the output of system_wakeup in a guest that
does not have wake-up from suspend support (ppc64):
(qemu) system_wakeup
wake-up from suspend is not supported by this guest
(qemu)
And this is the output of system_wakeup in a x86 guest that has the
support but isn't suspended:
(qemu) system_wakeup
Unable to wake up: guest is not in suspended state
(qemu)
Reported-by: Balamuruhan S <bala24@linux.vnet.ibm.com>
Signed-off-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Message-Id: <20181205194701.17836-4-danielhb413@gmail.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Acked-by: Eduardo Habkost <ehabkost@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Markus Armbruster <armbru@redhat.com>
2018-12-05 22:47:01 +03:00
|
|
|
if (!qemu_wakeup_suspend_enabled()) {
|
|
|
|
error_setg(errp,
|
|
|
|
"wake-up from suspend is not supported by this guest");
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
|
|
|
qemu_system_wakeup_request(QEMU_WAKEUP_REASON_OTHER, errp);
|
2012-02-23 16:45:21 +04:00
|
|
|
}
|
|
|
|
|
2022-02-25 11:49:49 +03:00
|
|
|
void qmp_set_password(SetPasswordOptions *opts, Error **errp)
|
2011-12-07 17:17:51 +04:00
|
|
|
{
|
|
|
|
int rc;
|
|
|
|
|
2022-02-25 11:49:49 +03:00
|
|
|
if (opts->protocol == DISPLAY_PROTOCOL_SPICE) {
|
2015-01-13 19:07:15 +03:00
|
|
|
if (!qemu_using_spice(errp)) {
|
2011-12-07 17:17:51 +04:00
|
|
|
return;
|
|
|
|
}
|
2022-02-25 11:49:49 +03:00
|
|
|
rc = qemu_spice.set_passwd(opts->password,
|
|
|
|
opts->connected == SET_PASSWORD_ACTION_FAIL,
|
|
|
|
opts->connected == SET_PASSWORD_ACTION_DISCONNECT);
|
2022-02-25 11:49:48 +03:00
|
|
|
} else {
|
2022-02-25 11:49:49 +03:00
|
|
|
assert(opts->protocol == DISPLAY_PROTOCOL_VNC);
|
|
|
|
if (opts->connected != SET_PASSWORD_ACTION_KEEP) {
|
2011-12-07 17:17:51 +04:00
|
|
|
/* vnc supports "connected=keep" only */
|
2015-03-17 13:54:50 +03:00
|
|
|
error_setg(errp, QERR_INVALID_PARAMETER, "connected");
|
2011-12-07 17:17:51 +04:00
|
|
|
return;
|
|
|
|
}
|
|
|
|
/* Note that setting an empty password will not disable login through
|
|
|
|
* this interface. */
|
2022-02-25 11:49:49 +03:00
|
|
|
rc = vnc_display_password(opts->u.vnc.display, opts->password);
|
2011-12-07 17:17:51 +04:00
|
|
|
}
|
|
|
|
|
2020-11-13 11:26:20 +03:00
|
|
|
if (rc != 0) {
|
|
|
|
error_setg(errp, "Could not set password");
|
|
|
|
}
|
2011-12-07 17:17:51 +04:00
|
|
|
}
|
2011-12-07 17:47:57 +04:00
|
|
|
|
2022-02-25 11:49:49 +03:00
|
|
|
void qmp_expire_password(ExpirePasswordOptions *opts, Error **errp)
|
2011-12-07 17:47:57 +04:00
|
|
|
{
|
|
|
|
time_t when;
|
|
|
|
int rc;
|
2022-02-25 11:49:49 +03:00
|
|
|
const char *whenstr = opts->time;
|
2011-12-07 17:47:57 +04:00
|
|
|
|
|
|
|
if (strcmp(whenstr, "now") == 0) {
|
|
|
|
when = 0;
|
|
|
|
} else if (strcmp(whenstr, "never") == 0) {
|
|
|
|
when = TIME_MAX;
|
|
|
|
} else if (whenstr[0] == '+') {
|
|
|
|
when = time(NULL) + strtoull(whenstr+1, NULL, 10);
|
|
|
|
} else {
|
|
|
|
when = strtoull(whenstr, NULL, 10);
|
|
|
|
}
|
|
|
|
|
2022-02-25 11:49:49 +03:00
|
|
|
if (opts->protocol == DISPLAY_PROTOCOL_SPICE) {
|
2015-01-13 19:07:15 +03:00
|
|
|
if (!qemu_using_spice(errp)) {
|
2011-12-07 17:47:57 +04:00
|
|
|
return;
|
|
|
|
}
|
2020-10-19 10:52:16 +03:00
|
|
|
rc = qemu_spice.set_pw_expire(when);
|
2020-11-13 11:26:20 +03:00
|
|
|
} else {
|
2022-02-25 11:49:49 +03:00
|
|
|
assert(opts->protocol == DISPLAY_PROTOCOL_VNC);
|
|
|
|
rc = vnc_display_pw_expire(opts->u.vnc.display, when);
|
2011-12-07 17:47:57 +04:00
|
|
|
}
|
|
|
|
|
2020-11-13 11:26:20 +03:00
|
|
|
if (rc != 0) {
|
|
|
|
error_setg(errp, "Could not set password expire time");
|
|
|
|
}
|
2011-12-07 17:47:57 +04:00
|
|
|
}
|
2011-12-08 17:45:55 +04:00
|
|
|
|
2011-12-08 17:13:50 +04:00
|
|
|
#ifdef CONFIG_VNC
|
2011-12-08 17:45:55 +04:00
|
|
|
void qmp_change_vnc_password(const char *password, Error **errp)
|
|
|
|
{
|
|
|
|
if (vnc_display_password(NULL, password) < 0) {
|
2020-11-13 11:26:20 +03:00
|
|
|
error_setg(errp, "Could not set password");
|
2011-12-08 17:45:55 +04:00
|
|
|
}
|
|
|
|
}
|
qapi: add conditions to VNC type/commands/events on the schema
Add #if defined(CONFIG_VNC) in generated code, and adjust the
qmp/hmp code accordingly.
query-qmp-schema no longer reports the command/events etc as
available when disabled at compile.
Commands made conditional:
* query-vnc, query-vnc-servers, change-vnc-password
Before the patch, the commands for !CONFIG_VNC are stubs that fail
like this:
{"error": {"class": "GenericError",
"desc": "The feature 'vnc' is not enabled"}}
Afterwards, they fail like this:
{"error": {"class": "CommandNotFound",
"desc": "The command FOO has not been found"}}
I call that an improvement, because it lets clients distinguish
between command unavailable (class CommandNotFound) and command failed
(class GenericError).
Events made conditional:
* VNC_CONNECTED, VNC_INITIALIZED, VNC_DISCONNECTED
HMP change:
* info vnc
Will return "unknown command: 'info vnc'" when VNC is compiled
out (same as error for spice when --disable-spice)
Occurrences of VNC (case insensitive) in the schema that aren't
covered by this change:
* add_client
Command has other uses, including "socket bases character devices".
These are unconditional as far as I can tell.
* set_password, expire_password
In theory, these commands could be used for managing any service's
password. In practice, they're used for VNC and SPICE services.
They're documented for "remote display session" / "remote display
server".
The service is selected by argument @protocol. The code special-cases
protocol-specific argument checking, then calls a protocol-specific
function to do the work. If it fails, the command fails with "Could
not set password". It does when the service isn't compiled in (it's a
stub then).
We could make these commands conditional on the conjunction of all
services [currently: defined(CONFIG_VNC) || defined(CONFIG_SPICE)],
but I doubt it's worthwhile.
* change
Command has other uses, namely changing media.
This patch inlines a stub; no functional change.
Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Gerd Hoffmann <kraxel@redhat.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <20180703155648.11933-14-marcandre.lureau@redhat.com>
Signed-off-by: Markus Armbruster <armbru@redhat.com>
2018-07-03 18:56:47 +03:00
|
|
|
#endif
|
2011-12-23 00:40:54 +04:00
|
|
|
|
2012-09-13 23:52:20 +04:00
|
|
|
void qmp_add_client(const char *protocol, const char *fdname,
|
|
|
|
bool has_skipauth, bool skipauth, bool has_tls, bool tls,
|
|
|
|
Error **errp)
|
|
|
|
{
|
2016-12-07 16:20:22 +03:00
|
|
|
Chardev *s;
|
2012-09-13 23:52:20 +04:00
|
|
|
int fd;
|
|
|
|
|
2020-10-05 18:58:44 +03:00
|
|
|
fd = monitor_get_fd(monitor_cur(), fdname, errp);
|
2012-09-13 23:52:20 +04:00
|
|
|
if (fd < 0) {
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (strcmp(protocol, "spice") == 0) {
|
2015-01-13 19:07:15 +03:00
|
|
|
if (!qemu_using_spice(errp)) {
|
2012-09-13 23:52:20 +04:00
|
|
|
close(fd);
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
skipauth = has_skipauth ? skipauth : false;
|
|
|
|
tls = has_tls ? tls : false;
|
2020-10-19 10:52:17 +03:00
|
|
|
if (qemu_spice.display_add_client(fd, skipauth, tls) < 0) {
|
2012-09-13 23:52:20 +04:00
|
|
|
error_setg(errp, "spice failed to add client");
|
|
|
|
close(fd);
|
|
|
|
}
|
|
|
|
return;
|
|
|
|
#ifdef CONFIG_VNC
|
|
|
|
} else if (strcmp(protocol, "vnc") == 0) {
|
|
|
|
skipauth = has_skipauth ? skipauth : false;
|
|
|
|
vnc_display_add_client(NULL, fd, skipauth);
|
|
|
|
return;
|
2021-10-09 23:16:57 +03:00
|
|
|
#endif
|
|
|
|
#ifdef CONFIG_DBUS_DISPLAY
|
|
|
|
} else if (strcmp(protocol, "@dbus-display") == 0) {
|
|
|
|
if (!qemu_using_dbus_display(errp)) {
|
|
|
|
close(fd);
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
if (!qemu_dbus_display.add_client(fd, errp)) {
|
|
|
|
close(fd);
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
return;
|
2012-09-13 23:52:20 +04:00
|
|
|
#endif
|
|
|
|
} else if ((s = qemu_chr_find(protocol)) != NULL) {
|
|
|
|
if (qemu_chr_add_client(s, fd) < 0) {
|
|
|
|
error_setg(errp, "failed to add client");
|
|
|
|
close(fd);
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
|
|
|
error_setg(errp, "protocol '%s' is invalid", protocol);
|
|
|
|
close(fd);
|
|
|
|
}
|
2013-12-21 02:21:09 +04:00
|
|
|
|
2013-12-21 02:21:10 +04:00
|
|
|
|
2014-06-16 21:12:25 +04:00
|
|
|
MemoryDeviceInfoList *qmp_query_memory_devices(Error **errp)
|
|
|
|
{
|
2018-04-23 19:51:16 +03:00
|
|
|
return qmp_memory_device_list();
|
2014-06-16 21:12:25 +04:00
|
|
|
}
|
2014-06-16 21:12:28 +04:00
|
|
|
|
|
|
|
ACPIOSTInfoList *qmp_query_acpi_ospm_status(Error **errp)
|
|
|
|
{
|
|
|
|
bool ambig;
|
|
|
|
ACPIOSTInfoList *head = NULL;
|
|
|
|
ACPIOSTInfoList **prev = &head;
|
|
|
|
Object *obj = object_resolve_path_type("", TYPE_ACPI_DEVICE_IF, &ambig);
|
|
|
|
|
|
|
|
if (obj) {
|
|
|
|
AcpiDeviceIfClass *adevc = ACPI_DEVICE_IF_GET_CLASS(obj);
|
|
|
|
AcpiDeviceIf *adev = ACPI_DEVICE_IF(obj);
|
|
|
|
|
|
|
|
adevc->ospm_status(adev, &prev);
|
|
|
|
} else {
|
|
|
|
error_setg(errp, "command is not supported, missing ACPI device");
|
|
|
|
}
|
|
|
|
|
|
|
|
return head;
|
|
|
|
}
|
2017-08-29 18:30:21 +03:00
|
|
|
|
|
|
|
MemoryInfo *qmp_query_memory_size_summary(Error **errp)
|
|
|
|
{
|
2022-03-15 17:41:56 +03:00
|
|
|
MemoryInfo *mem_info = g_new0(MemoryInfo, 1);
|
2020-10-28 13:24:22 +03:00
|
|
|
MachineState *ms = MACHINE(qdev_get_machine());
|
2017-08-29 18:30:21 +03:00
|
|
|
|
2020-10-28 13:24:22 +03:00
|
|
|
mem_info->base_memory = ms->ram_size;
|
2017-08-29 18:30:21 +03:00
|
|
|
|
|
|
|
mem_info->plugged_memory = get_plugged_memory_size();
|
|
|
|
mem_info->has_plugged_memory =
|
|
|
|
mem_info->plugged_memory != (uint64_t)-1;
|
|
|
|
|
|
|
|
return mem_info;
|
|
|
|
}
|
2021-03-16 10:58:45 +03:00
|
|
|
|
|
|
|
void qmp_display_reload(DisplayReloadOptions *arg, Error **errp)
|
|
|
|
{
|
|
|
|
switch (arg->type) {
|
|
|
|
case DISPLAY_RELOAD_TYPE_VNC:
|
|
|
|
#ifdef CONFIG_VNC
|
|
|
|
if (arg->u.vnc.has_tls_certs && arg->u.vnc.tls_certs) {
|
|
|
|
vnc_display_reload_certs(NULL, errp);
|
|
|
|
}
|
|
|
|
#else
|
|
|
|
error_setg(errp, "vnc is invalid, missing 'CONFIG_VNC'");
|
|
|
|
#endif
|
|
|
|
break;
|
|
|
|
default:
|
|
|
|
abort();
|
|
|
|
}
|
|
|
|
}
|
2021-09-08 12:35:43 +03:00
|
|
|
|
2022-04-01 17:39:35 +03:00
|
|
|
void qmp_display_update(DisplayUpdateOptions *arg, Error **errp)
|
|
|
|
{
|
|
|
|
switch (arg->type) {
|
|
|
|
case DISPLAY_UPDATE_TYPE_VNC:
|
|
|
|
#ifdef CONFIG_VNC
|
|
|
|
vnc_display_update(&arg->u.vnc, errp);
|
|
|
|
#else
|
|
|
|
error_setg(errp, "vnc is invalid, missing 'CONFIG_VNC'");
|
|
|
|
#endif
|
|
|
|
break;
|
|
|
|
default:
|
|
|
|
abort();
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2021-09-08 12:35:43 +03:00
|
|
|
static int qmp_x_query_rdma_foreach(Object *obj, void *opaque)
|
|
|
|
{
|
|
|
|
RdmaProvider *rdma;
|
|
|
|
RdmaProviderClass *k;
|
|
|
|
GString *buf = opaque;
|
|
|
|
|
|
|
|
if (object_dynamic_cast(obj, INTERFACE_RDMA_PROVIDER)) {
|
|
|
|
rdma = RDMA_PROVIDER(obj);
|
|
|
|
k = RDMA_PROVIDER_GET_CLASS(obj);
|
|
|
|
if (k->format_statistics) {
|
|
|
|
k->format_statistics(rdma, buf);
|
|
|
|
} else {
|
|
|
|
g_string_append_printf(buf,
|
|
|
|
"RDMA statistics not available for %s.\n",
|
|
|
|
object_get_typename(obj));
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
HumanReadableText *qmp_x_query_rdma(Error **errp)
|
|
|
|
{
|
|
|
|
g_autoptr(GString) buf = g_string_new("");
|
|
|
|
|
|
|
|
object_child_foreach_recursive(object_get_root(),
|
|
|
|
qmp_x_query_rdma_foreach, buf);
|
|
|
|
|
|
|
|
return human_readable_text_from_str(buf);
|
|
|
|
}
|
2021-09-08 12:35:43 +03:00
|
|
|
|
|
|
|
HumanReadableText *qmp_x_query_ramblock(Error **errp)
|
|
|
|
{
|
|
|
|
g_autoptr(GString) buf = ram_block_format();
|
|
|
|
|
|
|
|
return human_readable_text_from_str(buf);
|
|
|
|
}
|
2021-09-08 12:35:43 +03:00
|
|
|
|
|
|
|
static int qmp_x_query_irq_foreach(Object *obj, void *opaque)
|
|
|
|
{
|
|
|
|
InterruptStatsProvider *intc;
|
|
|
|
InterruptStatsProviderClass *k;
|
|
|
|
GString *buf = opaque;
|
|
|
|
|
|
|
|
if (object_dynamic_cast(obj, TYPE_INTERRUPT_STATS_PROVIDER)) {
|
|
|
|
intc = INTERRUPT_STATS_PROVIDER(obj);
|
|
|
|
k = INTERRUPT_STATS_PROVIDER_GET_CLASS(obj);
|
|
|
|
uint64_t *irq_counts;
|
|
|
|
unsigned int nb_irqs, i;
|
|
|
|
if (k->get_statistics &&
|
|
|
|
k->get_statistics(intc, &irq_counts, &nb_irqs)) {
|
|
|
|
if (nb_irqs > 0) {
|
|
|
|
g_string_append_printf(buf, "IRQ statistics for %s:\n",
|
|
|
|
object_get_typename(obj));
|
|
|
|
for (i = 0; i < nb_irqs; i++) {
|
|
|
|
if (irq_counts[i] > 0) {
|
|
|
|
g_string_append_printf(buf, "%2d: %" PRId64 "\n", i,
|
|
|
|
irq_counts[i]);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
} else {
|
|
|
|
g_string_append_printf(buf,
|
|
|
|
"IRQ statistics not available for %s.\n",
|
|
|
|
object_get_typename(obj));
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
HumanReadableText *qmp_x_query_irq(Error **errp)
|
|
|
|
{
|
|
|
|
g_autoptr(GString) buf = g_string_new("");
|
|
|
|
|
|
|
|
object_child_foreach_recursive(object_get_root(),
|
|
|
|
qmp_x_query_irq_foreach, buf);
|
|
|
|
|
|
|
|
return human_readable_text_from_str(buf);
|
|
|
|
}
|
qmp: Support for querying stats
Gathering statistics is important for development, for monitoring and
for performance measurement. There are tools such as kvm_stat that do
this and they rely on the _user_ knowing the interesting data points
rather than the tool (which can treat them as opaque).
The commands introduced in this commit introduce QMP support for
querying stats; the goal is to take the capabilities of these tools
and making them available throughout the whole virtualization stack,
so that one can observe, monitor and measure virtual machines without
having shell access + root on the host that runs them.
query-stats returns a list of all stats per target type (only VM
and vCPU to start); future commits add extra options for specifying
stat names, vCPU qom paths, and providers. All these are used by the
HMP command "info stats". Because of the development usecases around
statistics, a good HMP interface is important.
query-stats-schemas returns a list of stats included in each target
type, with an option for specifying the provider. The concepts in the
schema are based on the KVM binary stats' own introspection data, just
translated to QAPI.
There are two reasons to have a separate schema that is not tied to
the QAPI schema. The first is the contents of the schemas: the new
introspection data provides different information than the QAPI data,
namely unit of measurement, how the numbers are gathered and change
(peak/instant/cumulative/histogram), and histogram bucket sizes.
There's really no reason to have this kind of metadata in the QAPI
introspection schema (except possibly for the unit of measure, but
there's a very weak justification).
Another reason is the dynamicity of the schema. The QAPI introspection
data is very much static; and while QOM is somewhat more dynamic,
generally we consider that to be a bug rather than a feature these days.
On the other hand, the statistics that are exposed by QEMU might be
passed through from another source, such as KVM, and the disadvantages of
manually updating the QAPI schema for outweight the benefits from vetting
the statistics and filtering out anything that seems "too unstable".
Running old QEMU with new kernel is a supported usecase; if old QEMU
cannot expose statistics from a new kernel, or if a kernel developer
needs to change QEMU before gathering new info from the new kernel,
then that is a poor user interface.
The framework provides a method to register callbacks for these QMP
commands. Most of the work in fact is done by the callbacks, and a
large majority of this patch is new QAPI structs and commands.
Examples (with KVM stats):
- Query all VM stats:
{ "execute": "query-stats", "arguments" : { "target": "vm" } }
{ "return": [
{ "provider": "kvm",
"stats": [
{ "name": "max_mmu_page_hash_collisions", "value": 0 },
{ "name": "max_mmu_rmap_size", "value": 0 },
{ "name": "nx_lpage_splits", "value": 148 },
... ] },
{ "provider": "xyz",
"stats": [ ... ] }
] }
- Query all vCPU stats:
{ "execute": "query-stats", "arguments" : { "target": "vcpu" } }
{ "return": [
{ "provider": "kvm",
"qom_path": "/machine/unattached/device[0]"
"stats": [
{ "name": "guest_mode", "value": 0 },
{ "name": "directed_yield_successful", "value": 0 },
{ "name": "directed_yield_attempted", "value": 106 },
... ] },
{ "provider": "kvm",
"qom_path": "/machine/unattached/device[1]"
"stats": [
{ "name": "guest_mode", "value": 0 },
{ "name": "directed_yield_successful", "value": 0 },
{ "name": "directed_yield_attempted", "value": 106 },
... ] },
] }
- Retrieve the schemas:
{ "execute": "query-stats-schemas" }
{ "return": [
{ "provider": "kvm",
"target": "vcpu",
"stats": [
{ "name": "guest_mode",
"unit": "none",
"base": 10,
"exponent": 0,
"type": "instant" },
{ "name": "directed_yield_successful",
"unit": "none",
"base": 10,
"exponent": 0,
"type": "cumulative" },
... ]
},
{ "provider": "kvm",
"target": "vm",
"stats": [
{ "name": "max_mmu_page_hash_collisions",
"unit": "none",
"base": 10,
"exponent": 0,
"type": "peak" },
... ]
},
{ "provider": "xyz",
"target": "vm",
"stats": [ ... ]
}
] }
Signed-off-by: Mark Kanda <mark.kanda@oracle.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2022-02-15 18:04:31 +03:00
|
|
|
|
|
|
|
typedef struct StatsCallbacks {
|
2022-04-26 12:49:33 +03:00
|
|
|
StatsProvider provider;
|
qmp: Support for querying stats
Gathering statistics is important for development, for monitoring and
for performance measurement. There are tools such as kvm_stat that do
this and they rely on the _user_ knowing the interesting data points
rather than the tool (which can treat them as opaque).
The commands introduced in this commit introduce QMP support for
querying stats; the goal is to take the capabilities of these tools
and making them available throughout the whole virtualization stack,
so that one can observe, monitor and measure virtual machines without
having shell access + root on the host that runs them.
query-stats returns a list of all stats per target type (only VM
and vCPU to start); future commits add extra options for specifying
stat names, vCPU qom paths, and providers. All these are used by the
HMP command "info stats". Because of the development usecases around
statistics, a good HMP interface is important.
query-stats-schemas returns a list of stats included in each target
type, with an option for specifying the provider. The concepts in the
schema are based on the KVM binary stats' own introspection data, just
translated to QAPI.
There are two reasons to have a separate schema that is not tied to
the QAPI schema. The first is the contents of the schemas: the new
introspection data provides different information than the QAPI data,
namely unit of measurement, how the numbers are gathered and change
(peak/instant/cumulative/histogram), and histogram bucket sizes.
There's really no reason to have this kind of metadata in the QAPI
introspection schema (except possibly for the unit of measure, but
there's a very weak justification).
Another reason is the dynamicity of the schema. The QAPI introspection
data is very much static; and while QOM is somewhat more dynamic,
generally we consider that to be a bug rather than a feature these days.
On the other hand, the statistics that are exposed by QEMU might be
passed through from another source, such as KVM, and the disadvantages of
manually updating the QAPI schema for outweight the benefits from vetting
the statistics and filtering out anything that seems "too unstable".
Running old QEMU with new kernel is a supported usecase; if old QEMU
cannot expose statistics from a new kernel, or if a kernel developer
needs to change QEMU before gathering new info from the new kernel,
then that is a poor user interface.
The framework provides a method to register callbacks for these QMP
commands. Most of the work in fact is done by the callbacks, and a
large majority of this patch is new QAPI structs and commands.
Examples (with KVM stats):
- Query all VM stats:
{ "execute": "query-stats", "arguments" : { "target": "vm" } }
{ "return": [
{ "provider": "kvm",
"stats": [
{ "name": "max_mmu_page_hash_collisions", "value": 0 },
{ "name": "max_mmu_rmap_size", "value": 0 },
{ "name": "nx_lpage_splits", "value": 148 },
... ] },
{ "provider": "xyz",
"stats": [ ... ] }
] }
- Query all vCPU stats:
{ "execute": "query-stats", "arguments" : { "target": "vcpu" } }
{ "return": [
{ "provider": "kvm",
"qom_path": "/machine/unattached/device[0]"
"stats": [
{ "name": "guest_mode", "value": 0 },
{ "name": "directed_yield_successful", "value": 0 },
{ "name": "directed_yield_attempted", "value": 106 },
... ] },
{ "provider": "kvm",
"qom_path": "/machine/unattached/device[1]"
"stats": [
{ "name": "guest_mode", "value": 0 },
{ "name": "directed_yield_successful", "value": 0 },
{ "name": "directed_yield_attempted", "value": 106 },
... ] },
] }
- Retrieve the schemas:
{ "execute": "query-stats-schemas" }
{ "return": [
{ "provider": "kvm",
"target": "vcpu",
"stats": [
{ "name": "guest_mode",
"unit": "none",
"base": 10,
"exponent": 0,
"type": "instant" },
{ "name": "directed_yield_successful",
"unit": "none",
"base": 10,
"exponent": 0,
"type": "cumulative" },
... ]
},
{ "provider": "kvm",
"target": "vm",
"stats": [
{ "name": "max_mmu_page_hash_collisions",
"unit": "none",
"base": 10,
"exponent": 0,
"type": "peak" },
... ]
},
{ "provider": "xyz",
"target": "vm",
"stats": [ ... ]
}
] }
Signed-off-by: Mark Kanda <mark.kanda@oracle.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2022-02-15 18:04:31 +03:00
|
|
|
StatRetrieveFunc *stats_cb;
|
|
|
|
SchemaRetrieveFunc *schemas_cb;
|
|
|
|
QTAILQ_ENTRY(StatsCallbacks) next;
|
|
|
|
} StatsCallbacks;
|
|
|
|
|
|
|
|
static QTAILQ_HEAD(, StatsCallbacks) stats_callbacks =
|
|
|
|
QTAILQ_HEAD_INITIALIZER(stats_callbacks);
|
|
|
|
|
2022-04-26 12:49:33 +03:00
|
|
|
void add_stats_callbacks(StatsProvider provider,
|
|
|
|
StatRetrieveFunc *stats_fn,
|
qmp: Support for querying stats
Gathering statistics is important for development, for monitoring and
for performance measurement. There are tools such as kvm_stat that do
this and they rely on the _user_ knowing the interesting data points
rather than the tool (which can treat them as opaque).
The commands introduced in this commit introduce QMP support for
querying stats; the goal is to take the capabilities of these tools
and making them available throughout the whole virtualization stack,
so that one can observe, monitor and measure virtual machines without
having shell access + root on the host that runs them.
query-stats returns a list of all stats per target type (only VM
and vCPU to start); future commits add extra options for specifying
stat names, vCPU qom paths, and providers. All these are used by the
HMP command "info stats". Because of the development usecases around
statistics, a good HMP interface is important.
query-stats-schemas returns a list of stats included in each target
type, with an option for specifying the provider. The concepts in the
schema are based on the KVM binary stats' own introspection data, just
translated to QAPI.
There are two reasons to have a separate schema that is not tied to
the QAPI schema. The first is the contents of the schemas: the new
introspection data provides different information than the QAPI data,
namely unit of measurement, how the numbers are gathered and change
(peak/instant/cumulative/histogram), and histogram bucket sizes.
There's really no reason to have this kind of metadata in the QAPI
introspection schema (except possibly for the unit of measure, but
there's a very weak justification).
Another reason is the dynamicity of the schema. The QAPI introspection
data is very much static; and while QOM is somewhat more dynamic,
generally we consider that to be a bug rather than a feature these days.
On the other hand, the statistics that are exposed by QEMU might be
passed through from another source, such as KVM, and the disadvantages of
manually updating the QAPI schema for outweight the benefits from vetting
the statistics and filtering out anything that seems "too unstable".
Running old QEMU with new kernel is a supported usecase; if old QEMU
cannot expose statistics from a new kernel, or if a kernel developer
needs to change QEMU before gathering new info from the new kernel,
then that is a poor user interface.
The framework provides a method to register callbacks for these QMP
commands. Most of the work in fact is done by the callbacks, and a
large majority of this patch is new QAPI structs and commands.
Examples (with KVM stats):
- Query all VM stats:
{ "execute": "query-stats", "arguments" : { "target": "vm" } }
{ "return": [
{ "provider": "kvm",
"stats": [
{ "name": "max_mmu_page_hash_collisions", "value": 0 },
{ "name": "max_mmu_rmap_size", "value": 0 },
{ "name": "nx_lpage_splits", "value": 148 },
... ] },
{ "provider": "xyz",
"stats": [ ... ] }
] }
- Query all vCPU stats:
{ "execute": "query-stats", "arguments" : { "target": "vcpu" } }
{ "return": [
{ "provider": "kvm",
"qom_path": "/machine/unattached/device[0]"
"stats": [
{ "name": "guest_mode", "value": 0 },
{ "name": "directed_yield_successful", "value": 0 },
{ "name": "directed_yield_attempted", "value": 106 },
... ] },
{ "provider": "kvm",
"qom_path": "/machine/unattached/device[1]"
"stats": [
{ "name": "guest_mode", "value": 0 },
{ "name": "directed_yield_successful", "value": 0 },
{ "name": "directed_yield_attempted", "value": 106 },
... ] },
] }
- Retrieve the schemas:
{ "execute": "query-stats-schemas" }
{ "return": [
{ "provider": "kvm",
"target": "vcpu",
"stats": [
{ "name": "guest_mode",
"unit": "none",
"base": 10,
"exponent": 0,
"type": "instant" },
{ "name": "directed_yield_successful",
"unit": "none",
"base": 10,
"exponent": 0,
"type": "cumulative" },
... ]
},
{ "provider": "kvm",
"target": "vm",
"stats": [
{ "name": "max_mmu_page_hash_collisions",
"unit": "none",
"base": 10,
"exponent": 0,
"type": "peak" },
... ]
},
{ "provider": "xyz",
"target": "vm",
"stats": [ ... ]
}
] }
Signed-off-by: Mark Kanda <mark.kanda@oracle.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2022-02-15 18:04:31 +03:00
|
|
|
SchemaRetrieveFunc *schemas_fn)
|
|
|
|
{
|
|
|
|
StatsCallbacks *entry = g_new(StatsCallbacks, 1);
|
2022-04-26 12:49:33 +03:00
|
|
|
entry->provider = provider;
|
qmp: Support for querying stats
Gathering statistics is important for development, for monitoring and
for performance measurement. There are tools such as kvm_stat that do
this and they rely on the _user_ knowing the interesting data points
rather than the tool (which can treat them as opaque).
The commands introduced in this commit introduce QMP support for
querying stats; the goal is to take the capabilities of these tools
and making them available throughout the whole virtualization stack,
so that one can observe, monitor and measure virtual machines without
having shell access + root on the host that runs them.
query-stats returns a list of all stats per target type (only VM
and vCPU to start); future commits add extra options for specifying
stat names, vCPU qom paths, and providers. All these are used by the
HMP command "info stats". Because of the development usecases around
statistics, a good HMP interface is important.
query-stats-schemas returns a list of stats included in each target
type, with an option for specifying the provider. The concepts in the
schema are based on the KVM binary stats' own introspection data, just
translated to QAPI.
There are two reasons to have a separate schema that is not tied to
the QAPI schema. The first is the contents of the schemas: the new
introspection data provides different information than the QAPI data,
namely unit of measurement, how the numbers are gathered and change
(peak/instant/cumulative/histogram), and histogram bucket sizes.
There's really no reason to have this kind of metadata in the QAPI
introspection schema (except possibly for the unit of measure, but
there's a very weak justification).
Another reason is the dynamicity of the schema. The QAPI introspection
data is very much static; and while QOM is somewhat more dynamic,
generally we consider that to be a bug rather than a feature these days.
On the other hand, the statistics that are exposed by QEMU might be
passed through from another source, such as KVM, and the disadvantages of
manually updating the QAPI schema for outweight the benefits from vetting
the statistics and filtering out anything that seems "too unstable".
Running old QEMU with new kernel is a supported usecase; if old QEMU
cannot expose statistics from a new kernel, or if a kernel developer
needs to change QEMU before gathering new info from the new kernel,
then that is a poor user interface.
The framework provides a method to register callbacks for these QMP
commands. Most of the work in fact is done by the callbacks, and a
large majority of this patch is new QAPI structs and commands.
Examples (with KVM stats):
- Query all VM stats:
{ "execute": "query-stats", "arguments" : { "target": "vm" } }
{ "return": [
{ "provider": "kvm",
"stats": [
{ "name": "max_mmu_page_hash_collisions", "value": 0 },
{ "name": "max_mmu_rmap_size", "value": 0 },
{ "name": "nx_lpage_splits", "value": 148 },
... ] },
{ "provider": "xyz",
"stats": [ ... ] }
] }
- Query all vCPU stats:
{ "execute": "query-stats", "arguments" : { "target": "vcpu" } }
{ "return": [
{ "provider": "kvm",
"qom_path": "/machine/unattached/device[0]"
"stats": [
{ "name": "guest_mode", "value": 0 },
{ "name": "directed_yield_successful", "value": 0 },
{ "name": "directed_yield_attempted", "value": 106 },
... ] },
{ "provider": "kvm",
"qom_path": "/machine/unattached/device[1]"
"stats": [
{ "name": "guest_mode", "value": 0 },
{ "name": "directed_yield_successful", "value": 0 },
{ "name": "directed_yield_attempted", "value": 106 },
... ] },
] }
- Retrieve the schemas:
{ "execute": "query-stats-schemas" }
{ "return": [
{ "provider": "kvm",
"target": "vcpu",
"stats": [
{ "name": "guest_mode",
"unit": "none",
"base": 10,
"exponent": 0,
"type": "instant" },
{ "name": "directed_yield_successful",
"unit": "none",
"base": 10,
"exponent": 0,
"type": "cumulative" },
... ]
},
{ "provider": "kvm",
"target": "vm",
"stats": [
{ "name": "max_mmu_page_hash_collisions",
"unit": "none",
"base": 10,
"exponent": 0,
"type": "peak" },
... ]
},
{ "provider": "xyz",
"target": "vm",
"stats": [ ... ]
}
] }
Signed-off-by: Mark Kanda <mark.kanda@oracle.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2022-02-15 18:04:31 +03:00
|
|
|
entry->stats_cb = stats_fn;
|
|
|
|
entry->schemas_cb = schemas_fn;
|
|
|
|
|
|
|
|
QTAILQ_INSERT_TAIL(&stats_callbacks, entry, next);
|
|
|
|
}
|
|
|
|
|
|
|
|
static bool invoke_stats_cb(StatsCallbacks *entry,
|
|
|
|
StatsResultList **stats_results,
|
2022-04-26 12:49:33 +03:00
|
|
|
StatsFilter *filter, StatsRequest *request,
|
qmp: Support for querying stats
Gathering statistics is important for development, for monitoring and
for performance measurement. There are tools such as kvm_stat that do
this and they rely on the _user_ knowing the interesting data points
rather than the tool (which can treat them as opaque).
The commands introduced in this commit introduce QMP support for
querying stats; the goal is to take the capabilities of these tools
and making them available throughout the whole virtualization stack,
so that one can observe, monitor and measure virtual machines without
having shell access + root on the host that runs them.
query-stats returns a list of all stats per target type (only VM
and vCPU to start); future commits add extra options for specifying
stat names, vCPU qom paths, and providers. All these are used by the
HMP command "info stats". Because of the development usecases around
statistics, a good HMP interface is important.
query-stats-schemas returns a list of stats included in each target
type, with an option for specifying the provider. The concepts in the
schema are based on the KVM binary stats' own introspection data, just
translated to QAPI.
There are two reasons to have a separate schema that is not tied to
the QAPI schema. The first is the contents of the schemas: the new
introspection data provides different information than the QAPI data,
namely unit of measurement, how the numbers are gathered and change
(peak/instant/cumulative/histogram), and histogram bucket sizes.
There's really no reason to have this kind of metadata in the QAPI
introspection schema (except possibly for the unit of measure, but
there's a very weak justification).
Another reason is the dynamicity of the schema. The QAPI introspection
data is very much static; and while QOM is somewhat more dynamic,
generally we consider that to be a bug rather than a feature these days.
On the other hand, the statistics that are exposed by QEMU might be
passed through from another source, such as KVM, and the disadvantages of
manually updating the QAPI schema for outweight the benefits from vetting
the statistics and filtering out anything that seems "too unstable".
Running old QEMU with new kernel is a supported usecase; if old QEMU
cannot expose statistics from a new kernel, or if a kernel developer
needs to change QEMU before gathering new info from the new kernel,
then that is a poor user interface.
The framework provides a method to register callbacks for these QMP
commands. Most of the work in fact is done by the callbacks, and a
large majority of this patch is new QAPI structs and commands.
Examples (with KVM stats):
- Query all VM stats:
{ "execute": "query-stats", "arguments" : { "target": "vm" } }
{ "return": [
{ "provider": "kvm",
"stats": [
{ "name": "max_mmu_page_hash_collisions", "value": 0 },
{ "name": "max_mmu_rmap_size", "value": 0 },
{ "name": "nx_lpage_splits", "value": 148 },
... ] },
{ "provider": "xyz",
"stats": [ ... ] }
] }
- Query all vCPU stats:
{ "execute": "query-stats", "arguments" : { "target": "vcpu" } }
{ "return": [
{ "provider": "kvm",
"qom_path": "/machine/unattached/device[0]"
"stats": [
{ "name": "guest_mode", "value": 0 },
{ "name": "directed_yield_successful", "value": 0 },
{ "name": "directed_yield_attempted", "value": 106 },
... ] },
{ "provider": "kvm",
"qom_path": "/machine/unattached/device[1]"
"stats": [
{ "name": "guest_mode", "value": 0 },
{ "name": "directed_yield_successful", "value": 0 },
{ "name": "directed_yield_attempted", "value": 106 },
... ] },
] }
- Retrieve the schemas:
{ "execute": "query-stats-schemas" }
{ "return": [
{ "provider": "kvm",
"target": "vcpu",
"stats": [
{ "name": "guest_mode",
"unit": "none",
"base": 10,
"exponent": 0,
"type": "instant" },
{ "name": "directed_yield_successful",
"unit": "none",
"base": 10,
"exponent": 0,
"type": "cumulative" },
... ]
},
{ "provider": "kvm",
"target": "vm",
"stats": [
{ "name": "max_mmu_page_hash_collisions",
"unit": "none",
"base": 10,
"exponent": 0,
"type": "peak" },
... ]
},
{ "provider": "xyz",
"target": "vm",
"stats": [ ... ]
}
] }
Signed-off-by: Mark Kanda <mark.kanda@oracle.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2022-02-15 18:04:31 +03:00
|
|
|
Error **errp)
|
|
|
|
{
|
2022-11-21 11:50:47 +03:00
|
|
|
ERRP_GUARD();
|
2022-04-26 15:59:44 +03:00
|
|
|
strList *targets = NULL;
|
qmp: add filtering of statistics by name
Allow retrieving only a subset of statistics. This can be useful
for example in order to plot a subset of the statistics many times
a second: KVM publishes ~40 statistics for each vCPU on x86; retrieving
and serializing all of them would be useless.
Another use will be in HMP in the following patch; implementing the
filter in the backend is easy enough that it was deemed okay to make
this a public interface.
Example:
{ "execute": "query-stats",
"arguments": {
"target": "vcpu",
"vcpus": [ "/machine/unattached/device[2]",
"/machine/unattached/device[4]" ],
"providers": [
{ "provider": "kvm",
"names": [ "l1d_flush", "exits" ] } } }
{ "return": {
"vcpus": [
{ "path": "/machine/unattached/device[2]"
"providers": [
{ "provider": "kvm",
"stats": [ { "name": "l1d_flush", "value": 41213 },
{ "name": "exits", "value": 74291 } ] } ] },
{ "path": "/machine/unattached/device[4]"
"providers": [
{ "provider": "kvm",
"stats": [ { "name": "l1d_flush", "value": 16132 },
{ "name": "exits", "value": 57922 } ] } ] } ] } }
Extracted from a patch by Mark Kanda.
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2022-05-24 20:13:16 +03:00
|
|
|
strList *names = NULL;
|
qmp: Support for querying stats
Gathering statistics is important for development, for monitoring and
for performance measurement. There are tools such as kvm_stat that do
this and they rely on the _user_ knowing the interesting data points
rather than the tool (which can treat them as opaque).
The commands introduced in this commit introduce QMP support for
querying stats; the goal is to take the capabilities of these tools
and making them available throughout the whole virtualization stack,
so that one can observe, monitor and measure virtual machines without
having shell access + root on the host that runs them.
query-stats returns a list of all stats per target type (only VM
and vCPU to start); future commits add extra options for specifying
stat names, vCPU qom paths, and providers. All these are used by the
HMP command "info stats". Because of the development usecases around
statistics, a good HMP interface is important.
query-stats-schemas returns a list of stats included in each target
type, with an option for specifying the provider. The concepts in the
schema are based on the KVM binary stats' own introspection data, just
translated to QAPI.
There are two reasons to have a separate schema that is not tied to
the QAPI schema. The first is the contents of the schemas: the new
introspection data provides different information than the QAPI data,
namely unit of measurement, how the numbers are gathered and change
(peak/instant/cumulative/histogram), and histogram bucket sizes.
There's really no reason to have this kind of metadata in the QAPI
introspection schema (except possibly for the unit of measure, but
there's a very weak justification).
Another reason is the dynamicity of the schema. The QAPI introspection
data is very much static; and while QOM is somewhat more dynamic,
generally we consider that to be a bug rather than a feature these days.
On the other hand, the statistics that are exposed by QEMU might be
passed through from another source, such as KVM, and the disadvantages of
manually updating the QAPI schema for outweight the benefits from vetting
the statistics and filtering out anything that seems "too unstable".
Running old QEMU with new kernel is a supported usecase; if old QEMU
cannot expose statistics from a new kernel, or if a kernel developer
needs to change QEMU before gathering new info from the new kernel,
then that is a poor user interface.
The framework provides a method to register callbacks for these QMP
commands. Most of the work in fact is done by the callbacks, and a
large majority of this patch is new QAPI structs and commands.
Examples (with KVM stats):
- Query all VM stats:
{ "execute": "query-stats", "arguments" : { "target": "vm" } }
{ "return": [
{ "provider": "kvm",
"stats": [
{ "name": "max_mmu_page_hash_collisions", "value": 0 },
{ "name": "max_mmu_rmap_size", "value": 0 },
{ "name": "nx_lpage_splits", "value": 148 },
... ] },
{ "provider": "xyz",
"stats": [ ... ] }
] }
- Query all vCPU stats:
{ "execute": "query-stats", "arguments" : { "target": "vcpu" } }
{ "return": [
{ "provider": "kvm",
"qom_path": "/machine/unattached/device[0]"
"stats": [
{ "name": "guest_mode", "value": 0 },
{ "name": "directed_yield_successful", "value": 0 },
{ "name": "directed_yield_attempted", "value": 106 },
... ] },
{ "provider": "kvm",
"qom_path": "/machine/unattached/device[1]"
"stats": [
{ "name": "guest_mode", "value": 0 },
{ "name": "directed_yield_successful", "value": 0 },
{ "name": "directed_yield_attempted", "value": 106 },
... ] },
] }
- Retrieve the schemas:
{ "execute": "query-stats-schemas" }
{ "return": [
{ "provider": "kvm",
"target": "vcpu",
"stats": [
{ "name": "guest_mode",
"unit": "none",
"base": 10,
"exponent": 0,
"type": "instant" },
{ "name": "directed_yield_successful",
"unit": "none",
"base": 10,
"exponent": 0,
"type": "cumulative" },
... ]
},
{ "provider": "kvm",
"target": "vm",
"stats": [
{ "name": "max_mmu_page_hash_collisions",
"unit": "none",
"base": 10,
"exponent": 0,
"type": "peak" },
... ]
},
{ "provider": "xyz",
"target": "vm",
"stats": [ ... ]
}
] }
Signed-off-by: Mark Kanda <mark.kanda@oracle.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2022-02-15 18:04:31 +03:00
|
|
|
|
2022-04-26 12:49:33 +03:00
|
|
|
if (request) {
|
|
|
|
if (request->provider != entry->provider) {
|
|
|
|
return true;
|
|
|
|
}
|
qmp: add filtering of statistics by name
Allow retrieving only a subset of statistics. This can be useful
for example in order to plot a subset of the statistics many times
a second: KVM publishes ~40 statistics for each vCPU on x86; retrieving
and serializing all of them would be useless.
Another use will be in HMP in the following patch; implementing the
filter in the backend is easy enough that it was deemed okay to make
this a public interface.
Example:
{ "execute": "query-stats",
"arguments": {
"target": "vcpu",
"vcpus": [ "/machine/unattached/device[2]",
"/machine/unattached/device[4]" ],
"providers": [
{ "provider": "kvm",
"names": [ "l1d_flush", "exits" ] } } }
{ "return": {
"vcpus": [
{ "path": "/machine/unattached/device[2]"
"providers": [
{ "provider": "kvm",
"stats": [ { "name": "l1d_flush", "value": 41213 },
{ "name": "exits", "value": 74291 } ] } ] },
{ "path": "/machine/unattached/device[4]"
"providers": [
{ "provider": "kvm",
"stats": [ { "name": "l1d_flush", "value": 16132 },
{ "name": "exits", "value": 57922 } ] } ] } ] } }
Extracted from a patch by Mark Kanda.
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2022-05-24 20:13:16 +03:00
|
|
|
if (request->has_names && !request->names) {
|
|
|
|
return true;
|
|
|
|
}
|
|
|
|
names = request->has_names ? request->names : NULL;
|
2022-04-26 12:49:33 +03:00
|
|
|
}
|
|
|
|
|
2022-04-26 15:59:44 +03:00
|
|
|
switch (filter->target) {
|
|
|
|
case STATS_TARGET_VM:
|
|
|
|
break;
|
|
|
|
case STATS_TARGET_VCPU:
|
|
|
|
if (filter->u.vcpu.has_vcpus) {
|
|
|
|
if (!filter->u.vcpu.vcpus) {
|
|
|
|
/* No targets allowed? Return no statistics. */
|
|
|
|
return true;
|
|
|
|
}
|
|
|
|
targets = filter->u.vcpu.vcpus;
|
|
|
|
}
|
|
|
|
break;
|
|
|
|
default:
|
|
|
|
abort();
|
|
|
|
}
|
|
|
|
|
qmp: add filtering of statistics by name
Allow retrieving only a subset of statistics. This can be useful
for example in order to plot a subset of the statistics many times
a second: KVM publishes ~40 statistics for each vCPU on x86; retrieving
and serializing all of them would be useless.
Another use will be in HMP in the following patch; implementing the
filter in the backend is easy enough that it was deemed okay to make
this a public interface.
Example:
{ "execute": "query-stats",
"arguments": {
"target": "vcpu",
"vcpus": [ "/machine/unattached/device[2]",
"/machine/unattached/device[4]" ],
"providers": [
{ "provider": "kvm",
"names": [ "l1d_flush", "exits" ] } } }
{ "return": {
"vcpus": [
{ "path": "/machine/unattached/device[2]"
"providers": [
{ "provider": "kvm",
"stats": [ { "name": "l1d_flush", "value": 41213 },
{ "name": "exits", "value": 74291 } ] } ] },
{ "path": "/machine/unattached/device[4]"
"providers": [
{ "provider": "kvm",
"stats": [ { "name": "l1d_flush", "value": 16132 },
{ "name": "exits", "value": 57922 } ] } ] } ] } }
Extracted from a patch by Mark Kanda.
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2022-05-24 20:13:16 +03:00
|
|
|
entry->stats_cb(stats_results, filter->target, names, targets, errp);
|
qmp: Support for querying stats
Gathering statistics is important for development, for monitoring and
for performance measurement. There are tools such as kvm_stat that do
this and they rely on the _user_ knowing the interesting data points
rather than the tool (which can treat them as opaque).
The commands introduced in this commit introduce QMP support for
querying stats; the goal is to take the capabilities of these tools
and making them available throughout the whole virtualization stack,
so that one can observe, monitor and measure virtual machines without
having shell access + root on the host that runs them.
query-stats returns a list of all stats per target type (only VM
and vCPU to start); future commits add extra options for specifying
stat names, vCPU qom paths, and providers. All these are used by the
HMP command "info stats". Because of the development usecases around
statistics, a good HMP interface is important.
query-stats-schemas returns a list of stats included in each target
type, with an option for specifying the provider. The concepts in the
schema are based on the KVM binary stats' own introspection data, just
translated to QAPI.
There are two reasons to have a separate schema that is not tied to
the QAPI schema. The first is the contents of the schemas: the new
introspection data provides different information than the QAPI data,
namely unit of measurement, how the numbers are gathered and change
(peak/instant/cumulative/histogram), and histogram bucket sizes.
There's really no reason to have this kind of metadata in the QAPI
introspection schema (except possibly for the unit of measure, but
there's a very weak justification).
Another reason is the dynamicity of the schema. The QAPI introspection
data is very much static; and while QOM is somewhat more dynamic,
generally we consider that to be a bug rather than a feature these days.
On the other hand, the statistics that are exposed by QEMU might be
passed through from another source, such as KVM, and the disadvantages of
manually updating the QAPI schema for outweight the benefits from vetting
the statistics and filtering out anything that seems "too unstable".
Running old QEMU with new kernel is a supported usecase; if old QEMU
cannot expose statistics from a new kernel, or if a kernel developer
needs to change QEMU before gathering new info from the new kernel,
then that is a poor user interface.
The framework provides a method to register callbacks for these QMP
commands. Most of the work in fact is done by the callbacks, and a
large majority of this patch is new QAPI structs and commands.
Examples (with KVM stats):
- Query all VM stats:
{ "execute": "query-stats", "arguments" : { "target": "vm" } }
{ "return": [
{ "provider": "kvm",
"stats": [
{ "name": "max_mmu_page_hash_collisions", "value": 0 },
{ "name": "max_mmu_rmap_size", "value": 0 },
{ "name": "nx_lpage_splits", "value": 148 },
... ] },
{ "provider": "xyz",
"stats": [ ... ] }
] }
- Query all vCPU stats:
{ "execute": "query-stats", "arguments" : { "target": "vcpu" } }
{ "return": [
{ "provider": "kvm",
"qom_path": "/machine/unattached/device[0]"
"stats": [
{ "name": "guest_mode", "value": 0 },
{ "name": "directed_yield_successful", "value": 0 },
{ "name": "directed_yield_attempted", "value": 106 },
... ] },
{ "provider": "kvm",
"qom_path": "/machine/unattached/device[1]"
"stats": [
{ "name": "guest_mode", "value": 0 },
{ "name": "directed_yield_successful", "value": 0 },
{ "name": "directed_yield_attempted", "value": 106 },
... ] },
] }
- Retrieve the schemas:
{ "execute": "query-stats-schemas" }
{ "return": [
{ "provider": "kvm",
"target": "vcpu",
"stats": [
{ "name": "guest_mode",
"unit": "none",
"base": 10,
"exponent": 0,
"type": "instant" },
{ "name": "directed_yield_successful",
"unit": "none",
"base": 10,
"exponent": 0,
"type": "cumulative" },
... ]
},
{ "provider": "kvm",
"target": "vm",
"stats": [
{ "name": "max_mmu_page_hash_collisions",
"unit": "none",
"base": 10,
"exponent": 0,
"type": "peak" },
... ]
},
{ "provider": "xyz",
"target": "vm",
"stats": [ ... ]
}
] }
Signed-off-by: Mark Kanda <mark.kanda@oracle.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2022-02-15 18:04:31 +03:00
|
|
|
if (*errp) {
|
|
|
|
qapi_free_StatsResultList(*stats_results);
|
|
|
|
*stats_results = NULL;
|
|
|
|
return false;
|
|
|
|
}
|
|
|
|
return true;
|
|
|
|
}
|
|
|
|
|
|
|
|
StatsResultList *qmp_query_stats(StatsFilter *filter, Error **errp)
|
|
|
|
{
|
|
|
|
StatsResultList *stats_results = NULL;
|
|
|
|
StatsCallbacks *entry;
|
2022-04-26 12:49:33 +03:00
|
|
|
StatsRequestList *request;
|
qmp: Support for querying stats
Gathering statistics is important for development, for monitoring and
for performance measurement. There are tools such as kvm_stat that do
this and they rely on the _user_ knowing the interesting data points
rather than the tool (which can treat them as opaque).
The commands introduced in this commit introduce QMP support for
querying stats; the goal is to take the capabilities of these tools
and making them available throughout the whole virtualization stack,
so that one can observe, monitor and measure virtual machines without
having shell access + root on the host that runs them.
query-stats returns a list of all stats per target type (only VM
and vCPU to start); future commits add extra options for specifying
stat names, vCPU qom paths, and providers. All these are used by the
HMP command "info stats". Because of the development usecases around
statistics, a good HMP interface is important.
query-stats-schemas returns a list of stats included in each target
type, with an option for specifying the provider. The concepts in the
schema are based on the KVM binary stats' own introspection data, just
translated to QAPI.
There are two reasons to have a separate schema that is not tied to
the QAPI schema. The first is the contents of the schemas: the new
introspection data provides different information than the QAPI data,
namely unit of measurement, how the numbers are gathered and change
(peak/instant/cumulative/histogram), and histogram bucket sizes.
There's really no reason to have this kind of metadata in the QAPI
introspection schema (except possibly for the unit of measure, but
there's a very weak justification).
Another reason is the dynamicity of the schema. The QAPI introspection
data is very much static; and while QOM is somewhat more dynamic,
generally we consider that to be a bug rather than a feature these days.
On the other hand, the statistics that are exposed by QEMU might be
passed through from another source, such as KVM, and the disadvantages of
manually updating the QAPI schema for outweight the benefits from vetting
the statistics and filtering out anything that seems "too unstable".
Running old QEMU with new kernel is a supported usecase; if old QEMU
cannot expose statistics from a new kernel, or if a kernel developer
needs to change QEMU before gathering new info from the new kernel,
then that is a poor user interface.
The framework provides a method to register callbacks for these QMP
commands. Most of the work in fact is done by the callbacks, and a
large majority of this patch is new QAPI structs and commands.
Examples (with KVM stats):
- Query all VM stats:
{ "execute": "query-stats", "arguments" : { "target": "vm" } }
{ "return": [
{ "provider": "kvm",
"stats": [
{ "name": "max_mmu_page_hash_collisions", "value": 0 },
{ "name": "max_mmu_rmap_size", "value": 0 },
{ "name": "nx_lpage_splits", "value": 148 },
... ] },
{ "provider": "xyz",
"stats": [ ... ] }
] }
- Query all vCPU stats:
{ "execute": "query-stats", "arguments" : { "target": "vcpu" } }
{ "return": [
{ "provider": "kvm",
"qom_path": "/machine/unattached/device[0]"
"stats": [
{ "name": "guest_mode", "value": 0 },
{ "name": "directed_yield_successful", "value": 0 },
{ "name": "directed_yield_attempted", "value": 106 },
... ] },
{ "provider": "kvm",
"qom_path": "/machine/unattached/device[1]"
"stats": [
{ "name": "guest_mode", "value": 0 },
{ "name": "directed_yield_successful", "value": 0 },
{ "name": "directed_yield_attempted", "value": 106 },
... ] },
] }
- Retrieve the schemas:
{ "execute": "query-stats-schemas" }
{ "return": [
{ "provider": "kvm",
"target": "vcpu",
"stats": [
{ "name": "guest_mode",
"unit": "none",
"base": 10,
"exponent": 0,
"type": "instant" },
{ "name": "directed_yield_successful",
"unit": "none",
"base": 10,
"exponent": 0,
"type": "cumulative" },
... ]
},
{ "provider": "kvm",
"target": "vm",
"stats": [
{ "name": "max_mmu_page_hash_collisions",
"unit": "none",
"base": 10,
"exponent": 0,
"type": "peak" },
... ]
},
{ "provider": "xyz",
"target": "vm",
"stats": [ ... ]
}
] }
Signed-off-by: Mark Kanda <mark.kanda@oracle.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2022-02-15 18:04:31 +03:00
|
|
|
|
|
|
|
QTAILQ_FOREACH(entry, &stats_callbacks, next) {
|
2022-04-26 12:49:33 +03:00
|
|
|
if (filter->has_providers) {
|
|
|
|
for (request = filter->providers; request; request = request->next) {
|
|
|
|
if (!invoke_stats_cb(entry, &stats_results, filter,
|
|
|
|
request->value, errp)) {
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
} else {
|
|
|
|
if (!invoke_stats_cb(entry, &stats_results, filter, NULL, errp)) {
|
|
|
|
break;
|
|
|
|
}
|
qmp: Support for querying stats
Gathering statistics is important for development, for monitoring and
for performance measurement. There are tools such as kvm_stat that do
this and they rely on the _user_ knowing the interesting data points
rather than the tool (which can treat them as opaque).
The commands introduced in this commit introduce QMP support for
querying stats; the goal is to take the capabilities of these tools
and making them available throughout the whole virtualization stack,
so that one can observe, monitor and measure virtual machines without
having shell access + root on the host that runs them.
query-stats returns a list of all stats per target type (only VM
and vCPU to start); future commits add extra options for specifying
stat names, vCPU qom paths, and providers. All these are used by the
HMP command "info stats". Because of the development usecases around
statistics, a good HMP interface is important.
query-stats-schemas returns a list of stats included in each target
type, with an option for specifying the provider. The concepts in the
schema are based on the KVM binary stats' own introspection data, just
translated to QAPI.
There are two reasons to have a separate schema that is not tied to
the QAPI schema. The first is the contents of the schemas: the new
introspection data provides different information than the QAPI data,
namely unit of measurement, how the numbers are gathered and change
(peak/instant/cumulative/histogram), and histogram bucket sizes.
There's really no reason to have this kind of metadata in the QAPI
introspection schema (except possibly for the unit of measure, but
there's a very weak justification).
Another reason is the dynamicity of the schema. The QAPI introspection
data is very much static; and while QOM is somewhat more dynamic,
generally we consider that to be a bug rather than a feature these days.
On the other hand, the statistics that are exposed by QEMU might be
passed through from another source, such as KVM, and the disadvantages of
manually updating the QAPI schema for outweight the benefits from vetting
the statistics and filtering out anything that seems "too unstable".
Running old QEMU with new kernel is a supported usecase; if old QEMU
cannot expose statistics from a new kernel, or if a kernel developer
needs to change QEMU before gathering new info from the new kernel,
then that is a poor user interface.
The framework provides a method to register callbacks for these QMP
commands. Most of the work in fact is done by the callbacks, and a
large majority of this patch is new QAPI structs and commands.
Examples (with KVM stats):
- Query all VM stats:
{ "execute": "query-stats", "arguments" : { "target": "vm" } }
{ "return": [
{ "provider": "kvm",
"stats": [
{ "name": "max_mmu_page_hash_collisions", "value": 0 },
{ "name": "max_mmu_rmap_size", "value": 0 },
{ "name": "nx_lpage_splits", "value": 148 },
... ] },
{ "provider": "xyz",
"stats": [ ... ] }
] }
- Query all vCPU stats:
{ "execute": "query-stats", "arguments" : { "target": "vcpu" } }
{ "return": [
{ "provider": "kvm",
"qom_path": "/machine/unattached/device[0]"
"stats": [
{ "name": "guest_mode", "value": 0 },
{ "name": "directed_yield_successful", "value": 0 },
{ "name": "directed_yield_attempted", "value": 106 },
... ] },
{ "provider": "kvm",
"qom_path": "/machine/unattached/device[1]"
"stats": [
{ "name": "guest_mode", "value": 0 },
{ "name": "directed_yield_successful", "value": 0 },
{ "name": "directed_yield_attempted", "value": 106 },
... ] },
] }
- Retrieve the schemas:
{ "execute": "query-stats-schemas" }
{ "return": [
{ "provider": "kvm",
"target": "vcpu",
"stats": [
{ "name": "guest_mode",
"unit": "none",
"base": 10,
"exponent": 0,
"type": "instant" },
{ "name": "directed_yield_successful",
"unit": "none",
"base": 10,
"exponent": 0,
"type": "cumulative" },
... ]
},
{ "provider": "kvm",
"target": "vm",
"stats": [
{ "name": "max_mmu_page_hash_collisions",
"unit": "none",
"base": 10,
"exponent": 0,
"type": "peak" },
... ]
},
{ "provider": "xyz",
"target": "vm",
"stats": [ ... ]
}
] }
Signed-off-by: Mark Kanda <mark.kanda@oracle.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2022-02-15 18:04:31 +03:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
return stats_results;
|
|
|
|
}
|
|
|
|
|
2022-04-26 12:49:33 +03:00
|
|
|
StatsSchemaList *qmp_query_stats_schemas(bool has_provider,
|
|
|
|
StatsProvider provider,
|
|
|
|
Error **errp)
|
qmp: Support for querying stats
Gathering statistics is important for development, for monitoring and
for performance measurement. There are tools such as kvm_stat that do
this and they rely on the _user_ knowing the interesting data points
rather than the tool (which can treat them as opaque).
The commands introduced in this commit introduce QMP support for
querying stats; the goal is to take the capabilities of these tools
and making them available throughout the whole virtualization stack,
so that one can observe, monitor and measure virtual machines without
having shell access + root on the host that runs them.
query-stats returns a list of all stats per target type (only VM
and vCPU to start); future commits add extra options for specifying
stat names, vCPU qom paths, and providers. All these are used by the
HMP command "info stats". Because of the development usecases around
statistics, a good HMP interface is important.
query-stats-schemas returns a list of stats included in each target
type, with an option for specifying the provider. The concepts in the
schema are based on the KVM binary stats' own introspection data, just
translated to QAPI.
There are two reasons to have a separate schema that is not tied to
the QAPI schema. The first is the contents of the schemas: the new
introspection data provides different information than the QAPI data,
namely unit of measurement, how the numbers are gathered and change
(peak/instant/cumulative/histogram), and histogram bucket sizes.
There's really no reason to have this kind of metadata in the QAPI
introspection schema (except possibly for the unit of measure, but
there's a very weak justification).
Another reason is the dynamicity of the schema. The QAPI introspection
data is very much static; and while QOM is somewhat more dynamic,
generally we consider that to be a bug rather than a feature these days.
On the other hand, the statistics that are exposed by QEMU might be
passed through from another source, such as KVM, and the disadvantages of
manually updating the QAPI schema for outweight the benefits from vetting
the statistics and filtering out anything that seems "too unstable".
Running old QEMU with new kernel is a supported usecase; if old QEMU
cannot expose statistics from a new kernel, or if a kernel developer
needs to change QEMU before gathering new info from the new kernel,
then that is a poor user interface.
The framework provides a method to register callbacks for these QMP
commands. Most of the work in fact is done by the callbacks, and a
large majority of this patch is new QAPI structs and commands.
Examples (with KVM stats):
- Query all VM stats:
{ "execute": "query-stats", "arguments" : { "target": "vm" } }
{ "return": [
{ "provider": "kvm",
"stats": [
{ "name": "max_mmu_page_hash_collisions", "value": 0 },
{ "name": "max_mmu_rmap_size", "value": 0 },
{ "name": "nx_lpage_splits", "value": 148 },
... ] },
{ "provider": "xyz",
"stats": [ ... ] }
] }
- Query all vCPU stats:
{ "execute": "query-stats", "arguments" : { "target": "vcpu" } }
{ "return": [
{ "provider": "kvm",
"qom_path": "/machine/unattached/device[0]"
"stats": [
{ "name": "guest_mode", "value": 0 },
{ "name": "directed_yield_successful", "value": 0 },
{ "name": "directed_yield_attempted", "value": 106 },
... ] },
{ "provider": "kvm",
"qom_path": "/machine/unattached/device[1]"
"stats": [
{ "name": "guest_mode", "value": 0 },
{ "name": "directed_yield_successful", "value": 0 },
{ "name": "directed_yield_attempted", "value": 106 },
... ] },
] }
- Retrieve the schemas:
{ "execute": "query-stats-schemas" }
{ "return": [
{ "provider": "kvm",
"target": "vcpu",
"stats": [
{ "name": "guest_mode",
"unit": "none",
"base": 10,
"exponent": 0,
"type": "instant" },
{ "name": "directed_yield_successful",
"unit": "none",
"base": 10,
"exponent": 0,
"type": "cumulative" },
... ]
},
{ "provider": "kvm",
"target": "vm",
"stats": [
{ "name": "max_mmu_page_hash_collisions",
"unit": "none",
"base": 10,
"exponent": 0,
"type": "peak" },
... ]
},
{ "provider": "xyz",
"target": "vm",
"stats": [ ... ]
}
] }
Signed-off-by: Mark Kanda <mark.kanda@oracle.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2022-02-15 18:04:31 +03:00
|
|
|
{
|
2022-11-21 11:50:47 +03:00
|
|
|
ERRP_GUARD();
|
qmp: Support for querying stats
Gathering statistics is important for development, for monitoring and
for performance measurement. There are tools such as kvm_stat that do
this and they rely on the _user_ knowing the interesting data points
rather than the tool (which can treat them as opaque).
The commands introduced in this commit introduce QMP support for
querying stats; the goal is to take the capabilities of these tools
and making them available throughout the whole virtualization stack,
so that one can observe, monitor and measure virtual machines without
having shell access + root on the host that runs them.
query-stats returns a list of all stats per target type (only VM
and vCPU to start); future commits add extra options for specifying
stat names, vCPU qom paths, and providers. All these are used by the
HMP command "info stats". Because of the development usecases around
statistics, a good HMP interface is important.
query-stats-schemas returns a list of stats included in each target
type, with an option for specifying the provider. The concepts in the
schema are based on the KVM binary stats' own introspection data, just
translated to QAPI.
There are two reasons to have a separate schema that is not tied to
the QAPI schema. The first is the contents of the schemas: the new
introspection data provides different information than the QAPI data,
namely unit of measurement, how the numbers are gathered and change
(peak/instant/cumulative/histogram), and histogram bucket sizes.
There's really no reason to have this kind of metadata in the QAPI
introspection schema (except possibly for the unit of measure, but
there's a very weak justification).
Another reason is the dynamicity of the schema. The QAPI introspection
data is very much static; and while QOM is somewhat more dynamic,
generally we consider that to be a bug rather than a feature these days.
On the other hand, the statistics that are exposed by QEMU might be
passed through from another source, such as KVM, and the disadvantages of
manually updating the QAPI schema for outweight the benefits from vetting
the statistics and filtering out anything that seems "too unstable".
Running old QEMU with new kernel is a supported usecase; if old QEMU
cannot expose statistics from a new kernel, or if a kernel developer
needs to change QEMU before gathering new info from the new kernel,
then that is a poor user interface.
The framework provides a method to register callbacks for these QMP
commands. Most of the work in fact is done by the callbacks, and a
large majority of this patch is new QAPI structs and commands.
Examples (with KVM stats):
- Query all VM stats:
{ "execute": "query-stats", "arguments" : { "target": "vm" } }
{ "return": [
{ "provider": "kvm",
"stats": [
{ "name": "max_mmu_page_hash_collisions", "value": 0 },
{ "name": "max_mmu_rmap_size", "value": 0 },
{ "name": "nx_lpage_splits", "value": 148 },
... ] },
{ "provider": "xyz",
"stats": [ ... ] }
] }
- Query all vCPU stats:
{ "execute": "query-stats", "arguments" : { "target": "vcpu" } }
{ "return": [
{ "provider": "kvm",
"qom_path": "/machine/unattached/device[0]"
"stats": [
{ "name": "guest_mode", "value": 0 },
{ "name": "directed_yield_successful", "value": 0 },
{ "name": "directed_yield_attempted", "value": 106 },
... ] },
{ "provider": "kvm",
"qom_path": "/machine/unattached/device[1]"
"stats": [
{ "name": "guest_mode", "value": 0 },
{ "name": "directed_yield_successful", "value": 0 },
{ "name": "directed_yield_attempted", "value": 106 },
... ] },
] }
- Retrieve the schemas:
{ "execute": "query-stats-schemas" }
{ "return": [
{ "provider": "kvm",
"target": "vcpu",
"stats": [
{ "name": "guest_mode",
"unit": "none",
"base": 10,
"exponent": 0,
"type": "instant" },
{ "name": "directed_yield_successful",
"unit": "none",
"base": 10,
"exponent": 0,
"type": "cumulative" },
... ]
},
{ "provider": "kvm",
"target": "vm",
"stats": [
{ "name": "max_mmu_page_hash_collisions",
"unit": "none",
"base": 10,
"exponent": 0,
"type": "peak" },
... ]
},
{ "provider": "xyz",
"target": "vm",
"stats": [ ... ]
}
] }
Signed-off-by: Mark Kanda <mark.kanda@oracle.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2022-02-15 18:04:31 +03:00
|
|
|
StatsSchemaList *stats_results = NULL;
|
|
|
|
StatsCallbacks *entry;
|
|
|
|
|
|
|
|
QTAILQ_FOREACH(entry, &stats_callbacks, next) {
|
2022-04-26 12:49:33 +03:00
|
|
|
if (!has_provider || provider == entry->provider) {
|
|
|
|
entry->schemas_cb(&stats_results, errp);
|
|
|
|
if (*errp) {
|
|
|
|
qapi_free_StatsSchemaList(stats_results);
|
|
|
|
return NULL;
|
|
|
|
}
|
qmp: Support for querying stats
Gathering statistics is important for development, for monitoring and
for performance measurement. There are tools such as kvm_stat that do
this and they rely on the _user_ knowing the interesting data points
rather than the tool (which can treat them as opaque).
The commands introduced in this commit introduce QMP support for
querying stats; the goal is to take the capabilities of these tools
and making them available throughout the whole virtualization stack,
so that one can observe, monitor and measure virtual machines without
having shell access + root on the host that runs them.
query-stats returns a list of all stats per target type (only VM
and vCPU to start); future commits add extra options for specifying
stat names, vCPU qom paths, and providers. All these are used by the
HMP command "info stats". Because of the development usecases around
statistics, a good HMP interface is important.
query-stats-schemas returns a list of stats included in each target
type, with an option for specifying the provider. The concepts in the
schema are based on the KVM binary stats' own introspection data, just
translated to QAPI.
There are two reasons to have a separate schema that is not tied to
the QAPI schema. The first is the contents of the schemas: the new
introspection data provides different information than the QAPI data,
namely unit of measurement, how the numbers are gathered and change
(peak/instant/cumulative/histogram), and histogram bucket sizes.
There's really no reason to have this kind of metadata in the QAPI
introspection schema (except possibly for the unit of measure, but
there's a very weak justification).
Another reason is the dynamicity of the schema. The QAPI introspection
data is very much static; and while QOM is somewhat more dynamic,
generally we consider that to be a bug rather than a feature these days.
On the other hand, the statistics that are exposed by QEMU might be
passed through from another source, such as KVM, and the disadvantages of
manually updating the QAPI schema for outweight the benefits from vetting
the statistics and filtering out anything that seems "too unstable".
Running old QEMU with new kernel is a supported usecase; if old QEMU
cannot expose statistics from a new kernel, or if a kernel developer
needs to change QEMU before gathering new info from the new kernel,
then that is a poor user interface.
The framework provides a method to register callbacks for these QMP
commands. Most of the work in fact is done by the callbacks, and a
large majority of this patch is new QAPI structs and commands.
Examples (with KVM stats):
- Query all VM stats:
{ "execute": "query-stats", "arguments" : { "target": "vm" } }
{ "return": [
{ "provider": "kvm",
"stats": [
{ "name": "max_mmu_page_hash_collisions", "value": 0 },
{ "name": "max_mmu_rmap_size", "value": 0 },
{ "name": "nx_lpage_splits", "value": 148 },
... ] },
{ "provider": "xyz",
"stats": [ ... ] }
] }
- Query all vCPU stats:
{ "execute": "query-stats", "arguments" : { "target": "vcpu" } }
{ "return": [
{ "provider": "kvm",
"qom_path": "/machine/unattached/device[0]"
"stats": [
{ "name": "guest_mode", "value": 0 },
{ "name": "directed_yield_successful", "value": 0 },
{ "name": "directed_yield_attempted", "value": 106 },
... ] },
{ "provider": "kvm",
"qom_path": "/machine/unattached/device[1]"
"stats": [
{ "name": "guest_mode", "value": 0 },
{ "name": "directed_yield_successful", "value": 0 },
{ "name": "directed_yield_attempted", "value": 106 },
... ] },
] }
- Retrieve the schemas:
{ "execute": "query-stats-schemas" }
{ "return": [
{ "provider": "kvm",
"target": "vcpu",
"stats": [
{ "name": "guest_mode",
"unit": "none",
"base": 10,
"exponent": 0,
"type": "instant" },
{ "name": "directed_yield_successful",
"unit": "none",
"base": 10,
"exponent": 0,
"type": "cumulative" },
... ]
},
{ "provider": "kvm",
"target": "vm",
"stats": [
{ "name": "max_mmu_page_hash_collisions",
"unit": "none",
"base": 10,
"exponent": 0,
"type": "peak" },
... ]
},
{ "provider": "xyz",
"target": "vm",
"stats": [ ... ]
}
] }
Signed-off-by: Mark Kanda <mark.kanda@oracle.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2022-02-15 18:04:31 +03:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
return stats_results;
|
|
|
|
}
|
|
|
|
|
|
|
|
void add_stats_entry(StatsResultList **stats_results, StatsProvider provider,
|
|
|
|
const char *qom_path, StatsList *stats_list)
|
|
|
|
{
|
|
|
|
StatsResult *entry = g_new0(StatsResult, 1);
|
|
|
|
|
|
|
|
entry->provider = provider;
|
2022-11-04 19:07:06 +03:00
|
|
|
entry->qom_path = g_strdup(qom_path);
|
qmp: Support for querying stats
Gathering statistics is important for development, for monitoring and
for performance measurement. There are tools such as kvm_stat that do
this and they rely on the _user_ knowing the interesting data points
rather than the tool (which can treat them as opaque).
The commands introduced in this commit introduce QMP support for
querying stats; the goal is to take the capabilities of these tools
and making them available throughout the whole virtualization stack,
so that one can observe, monitor and measure virtual machines without
having shell access + root on the host that runs them.
query-stats returns a list of all stats per target type (only VM
and vCPU to start); future commits add extra options for specifying
stat names, vCPU qom paths, and providers. All these are used by the
HMP command "info stats". Because of the development usecases around
statistics, a good HMP interface is important.
query-stats-schemas returns a list of stats included in each target
type, with an option for specifying the provider. The concepts in the
schema are based on the KVM binary stats' own introspection data, just
translated to QAPI.
There are two reasons to have a separate schema that is not tied to
the QAPI schema. The first is the contents of the schemas: the new
introspection data provides different information than the QAPI data,
namely unit of measurement, how the numbers are gathered and change
(peak/instant/cumulative/histogram), and histogram bucket sizes.
There's really no reason to have this kind of metadata in the QAPI
introspection schema (except possibly for the unit of measure, but
there's a very weak justification).
Another reason is the dynamicity of the schema. The QAPI introspection
data is very much static; and while QOM is somewhat more dynamic,
generally we consider that to be a bug rather than a feature these days.
On the other hand, the statistics that are exposed by QEMU might be
passed through from another source, such as KVM, and the disadvantages of
manually updating the QAPI schema for outweight the benefits from vetting
the statistics and filtering out anything that seems "too unstable".
Running old QEMU with new kernel is a supported usecase; if old QEMU
cannot expose statistics from a new kernel, or if a kernel developer
needs to change QEMU before gathering new info from the new kernel,
then that is a poor user interface.
The framework provides a method to register callbacks for these QMP
commands. Most of the work in fact is done by the callbacks, and a
large majority of this patch is new QAPI structs and commands.
Examples (with KVM stats):
- Query all VM stats:
{ "execute": "query-stats", "arguments" : { "target": "vm" } }
{ "return": [
{ "provider": "kvm",
"stats": [
{ "name": "max_mmu_page_hash_collisions", "value": 0 },
{ "name": "max_mmu_rmap_size", "value": 0 },
{ "name": "nx_lpage_splits", "value": 148 },
... ] },
{ "provider": "xyz",
"stats": [ ... ] }
] }
- Query all vCPU stats:
{ "execute": "query-stats", "arguments" : { "target": "vcpu" } }
{ "return": [
{ "provider": "kvm",
"qom_path": "/machine/unattached/device[0]"
"stats": [
{ "name": "guest_mode", "value": 0 },
{ "name": "directed_yield_successful", "value": 0 },
{ "name": "directed_yield_attempted", "value": 106 },
... ] },
{ "provider": "kvm",
"qom_path": "/machine/unattached/device[1]"
"stats": [
{ "name": "guest_mode", "value": 0 },
{ "name": "directed_yield_successful", "value": 0 },
{ "name": "directed_yield_attempted", "value": 106 },
... ] },
] }
- Retrieve the schemas:
{ "execute": "query-stats-schemas" }
{ "return": [
{ "provider": "kvm",
"target": "vcpu",
"stats": [
{ "name": "guest_mode",
"unit": "none",
"base": 10,
"exponent": 0,
"type": "instant" },
{ "name": "directed_yield_successful",
"unit": "none",
"base": 10,
"exponent": 0,
"type": "cumulative" },
... ]
},
{ "provider": "kvm",
"target": "vm",
"stats": [
{ "name": "max_mmu_page_hash_collisions",
"unit": "none",
"base": 10,
"exponent": 0,
"type": "peak" },
... ]
},
{ "provider": "xyz",
"target": "vm",
"stats": [ ... ]
}
] }
Signed-off-by: Mark Kanda <mark.kanda@oracle.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
2022-02-15 18:04:31 +03:00
|
|
|
entry->stats = stats_list;
|
|
|
|
|
|
|
|
QAPI_LIST_PREPEND(*stats_results, entry);
|
|
|
|
}
|
|
|
|
|
|
|
|
void add_stats_schema(StatsSchemaList **schema_results,
|
|
|
|
StatsProvider provider, StatsTarget target,
|
|
|
|
StatsSchemaValueList *stats_list)
|
|
|
|
{
|
|
|
|
StatsSchema *entry = g_new0(StatsSchema, 1);
|
|
|
|
|
|
|
|
entry->provider = provider;
|
|
|
|
entry->target = target;
|
|
|
|
entry->stats = stats_list;
|
|
|
|
QAPI_LIST_PREPEND(*schema_results, entry);
|
|
|
|
}
|
2022-04-26 15:59:44 +03:00
|
|
|
|
|
|
|
bool apply_str_list_filter(const char *string, strList *list)
|
|
|
|
{
|
|
|
|
strList *str_list = NULL;
|
|
|
|
|
|
|
|
if (!list) {
|
|
|
|
return true;
|
|
|
|
}
|
|
|
|
for (str_list = list; str_list; str_list = str_list->next) {
|
|
|
|
if (g_str_equal(string, str_list->value)) {
|
|
|
|
return true;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
return false;
|
|
|
|
}
|