qemu-e2k/monitor/qmp.c
Stefan Hajnoczi effd60c878 monitor: only run coroutine commands in qemu_aio_context
monitor_qmp_dispatcher_co() runs in the iohandler AioContext that is not
polled during nested event loops. The coroutine currently reschedules
itself in the main loop's qemu_aio_context AioContext, which is polled
during nested event loops. One known problem is that QMP device-add
calls drain_call_rcu(), which temporarily drops the BQL, leading to all
sorts of havoc like other vCPU threads re-entering device emulation code
while another vCPU thread is waiting in device emulation code with
aio_poll().

Paolo Bonzini suggested running non-coroutine QMP handlers in the
iohandler AioContext. This avoids trouble with nested event loops. His
original idea was to move coroutine rescheduling to
monitor_qmp_dispatch(), but I resorted to moving it to qmp_dispatch()
because we don't know if the QMP handler needs to run in coroutine
context in monitor_qmp_dispatch(). monitor_qmp_dispatch() would have
been nicer since it's associated with the monitor implementation and not
as general as qmp_dispatch(), which is also used by qemu-ga.

A number of qemu-iotests need updated .out files because the order of
QMP events vs QMP responses has changed.

Solves Issue #1933.

Cc: qemu-stable@nongnu.org
Fixes: 7bed89958b ("device_core: use drain_call_rcu in in qmp_device_add")
Buglink: https://bugzilla.redhat.com/show_bug.cgi?id=2215192
Buglink: https://bugzilla.redhat.com/show_bug.cgi?id=2214985
Buglink: https://issues.redhat.com/browse/RHEL-17369
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-ID: <20240118144823.1497953-4-stefanha@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Tested-by: Fiona Ebner <f.ebner@proxmox.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
2024-01-26 11:16:58 +01:00

556 lines
18 KiB
C

/*
* QEMU monitor
*
* Copyright (c) 2003-2004 Fabrice Bellard
*
* Permission is hereby granted, free of charge, to any person obtaining a copy
* of this software and associated documentation files (the "Software"), to deal
* in the Software without restriction, including without limitation the rights
* to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
* copies of the Software, and to permit persons to whom the Software is
* furnished to do so, subject to the following conditions:
*
* The above copyright notice and this permission notice shall be included in
* all copies or substantial portions of the Software.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
* IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
* FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
* THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
* LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
* OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
* THE SOFTWARE.
*/
#include "qemu/osdep.h"
#include "chardev/char-io.h"
#include "monitor-internal.h"
#include "qapi/error.h"
#include "qapi/qapi-commands-control.h"
#include "qapi/qmp/qdict.h"
#include "qapi/qmp/qjson.h"
#include "qapi/qmp/qlist.h"
#include "trace.h"
/*
* qmp_dispatcher_co_busy is used for synchronisation between the
* monitor thread and the main thread to ensure that the dispatcher
* coroutine never gets scheduled a second time when it's already
* scheduled (scheduling the same coroutine twice is forbidden).
*
* It is true if the coroutine will process at least one more request
* before going to sleep. Either it has been kicked already, or it
* is active and processing requests. Additional requests may therefore
* be pushed onto mon->qmp_requests, and @qmp_dispatcher_co_shutdown may
* be set without further ado. @qmp_dispatcher_co must not be woken up
* in this case.
*
* If false, you have to wake up @qmp_dispatcher_co after pushing new
* requests. You also have to set @qmp_dispatcher_co_busy to true
* before waking up the coroutine.
*
* The coroutine will automatically change this variable back to false
* before it yields. Nobody else may set the variable to false.
*
* Access must be atomic for thread safety.
*/
static bool qmp_dispatcher_co_busy = true;
struct QMPRequest {
/* Owner of the request */
MonitorQMP *mon;
/*
* Request object to be handled or Error to be reported
* (exactly one of them is non-null)
*/
QObject *req;
Error *err;
};
typedef struct QMPRequest QMPRequest;
QmpCommandList qmp_commands, qmp_cap_negotiation_commands;
static bool qmp_oob_enabled(MonitorQMP *mon)
{
return mon->capab[QMP_CAPABILITY_OOB];
}
static void monitor_qmp_caps_reset(MonitorQMP *mon)
{
memset(mon->capab_offered, 0, sizeof(mon->capab_offered));
memset(mon->capab, 0, sizeof(mon->capab));
mon->capab_offered[QMP_CAPABILITY_OOB] = mon->common.use_io_thread;
}
static void qmp_request_free(QMPRequest *req)
{
qobject_unref(req->req);
error_free(req->err);
g_free(req);
}
/* Caller must hold mon->qmp.qmp_queue_lock */
static void monitor_qmp_cleanup_req_queue_locked(MonitorQMP *mon)
{
while (!g_queue_is_empty(mon->qmp_requests)) {
qmp_request_free(g_queue_pop_head(mon->qmp_requests));
}
}
static void monitor_qmp_cleanup_queue_and_resume(MonitorQMP *mon)
{
QEMU_LOCK_GUARD(&mon->qmp_queue_lock);
/*
* Same condition as in monitor_qmp_dispatcher_co(), but before
* removing an element from the queue (hence no `- 1`).
* Also, the queue should not be empty either, otherwise the
* monitor hasn't been suspended yet (or was already resumed).
*/
bool need_resume = (!qmp_oob_enabled(mon) ||
mon->qmp_requests->length == QMP_REQ_QUEUE_LEN_MAX)
&& !g_queue_is_empty(mon->qmp_requests);
monitor_qmp_cleanup_req_queue_locked(mon);
if (need_resume) {
/*
* handle_qmp_command() suspended the monitor because the
* request queue filled up, to be resumed when the queue has
* space again. We just emptied it; resume the monitor.
*
* Without this, the monitor would remain suspended forever
* when we get here while the monitor is suspended. An
* unfortunately timed CHR_EVENT_CLOSED can do the trick.
*/
monitor_resume(&mon->common);
}
}
void qmp_send_response(MonitorQMP *mon, const QDict *rsp)
{
const QObject *data = QOBJECT(rsp);
GString *json;
json = qobject_to_json_pretty(data, mon->pretty);
assert(json != NULL);
trace_monitor_qmp_respond(mon, json->str);
g_string_append_c(json, '\n');
monitor_puts(&mon->common, json->str);
g_string_free(json, true);
}
/*
* Emit QMP response @rsp to @mon.
* Null @rsp can only happen for commands with QCO_NO_SUCCESS_RESP.
* Nothing is emitted then.
*/
static void monitor_qmp_respond(MonitorQMP *mon, QDict *rsp)
{
if (rsp) {
qmp_send_response(mon, rsp);
}
}
/*
* Runs outside of coroutine context for OOB commands, but in
* coroutine context for everything else.
*/
static void monitor_qmp_dispatch(MonitorQMP *mon, QObject *req)
{
QDict *rsp;
QDict *error;
rsp = qmp_dispatch(mon->commands, req, qmp_oob_enabled(mon),
&mon->common);
if (mon->commands == &qmp_cap_negotiation_commands) {
error = qdict_get_qdict(rsp, "error");
if (error
&& !g_strcmp0(qdict_get_try_str(error, "class"),
QapiErrorClass_str(ERROR_CLASS_COMMAND_NOT_FOUND))) {
/* Provide a more useful error message */
qdict_del(error, "desc");
qdict_put_str(error, "desc", "Expecting capabilities negotiation"
" with 'qmp_capabilities'");
}
}
monitor_qmp_respond(mon, rsp);
qobject_unref(rsp);
}
/*
* Pop a QMP request from a monitor request queue.
* Return the request, or NULL all request queues are empty.
* We are using round-robin fashion to pop the request, to avoid
* processing commands only on a very busy monitor. To achieve that,
* when we process one request on a specific monitor, we put that
* monitor to the end of mon_list queue.
*
* Note: if the function returned with non-NULL, then the caller will
* be with qmp_mon->qmp_queue_lock held, and the caller is responsible
* to release it.
*/
static QMPRequest *monitor_qmp_requests_pop_any_with_lock(void)
{
QMPRequest *req_obj = NULL;
Monitor *mon;
MonitorQMP *qmp_mon;
QTAILQ_FOREACH(mon, &mon_list, entry) {
if (!monitor_is_qmp(mon)) {
continue;
}
qmp_mon = container_of(mon, MonitorQMP, common);
qemu_mutex_lock(&qmp_mon->qmp_queue_lock);
req_obj = g_queue_pop_head(qmp_mon->qmp_requests);
if (req_obj) {
/* With the lock of corresponding queue held */
break;
}
qemu_mutex_unlock(&qmp_mon->qmp_queue_lock);
}
if (req_obj) {
/*
* We found one request on the monitor. Degrade this monitor's
* priority to lowest by re-inserting it to end of queue.
*/
QTAILQ_REMOVE(&mon_list, mon, entry);
QTAILQ_INSERT_TAIL(&mon_list, mon, entry);
}
return req_obj;
}
static QMPRequest *monitor_qmp_dispatcher_pop_any(void)
{
while (true) {
/*
* To avoid double scheduling, busy is true on entry to
* monitor_qmp_dispatcher_co(), and must be set again before
* aio_co_wake()-ing it.
*/
assert(qatomic_read(&qmp_dispatcher_co_busy) == true);
/*
* Mark the dispatcher as not busy already here so that we
* don't miss any new requests coming in the middle of our
* processing.
*
* Clear qmp_dispatcher_co_busy before reading request.
*/
qatomic_set_mb(&qmp_dispatcher_co_busy, false);
WITH_QEMU_LOCK_GUARD(&monitor_lock) {
QMPRequest *req_obj;
/* On shutdown, don't take any more requests from the queue */
if (qmp_dispatcher_co_shutdown) {
return NULL;
}
req_obj = monitor_qmp_requests_pop_any_with_lock();
if (req_obj) {
return req_obj;
}
}
/*
* No more requests to process. Wait to be reentered from
* handle_qmp_command() when it pushes more requests, or
* from monitor_cleanup() when it requests shutdown.
*/
qemu_coroutine_yield();
}
}
void coroutine_fn monitor_qmp_dispatcher_co(void *data)
{
QMPRequest *req_obj;
QDict *rsp;
bool oob_enabled;
MonitorQMP *mon;
while ((req_obj = monitor_qmp_dispatcher_pop_any()) != NULL) {
trace_monitor_qmp_in_band_dequeue(req_obj,
req_obj->mon->qmp_requests->length);
/*
* @req_obj has a request, we hold req_obj->mon->qmp_queue_lock
*/
mon = req_obj->mon;
/*
* We need to resume the monitor if handle_qmp_command()
* suspended it. Two cases:
* 1. OOB enabled: mon->qmp_requests has no more space
* Resume right away, so that OOB commands can get executed while
* this request is being processed.
* 2. OOB disabled: always
* Resume only after we're done processing the request,
* We need to save qmp_oob_enabled() for later, because
* qmp_qmp_capabilities() can change it.
*/
oob_enabled = qmp_oob_enabled(mon);
if (oob_enabled
&& mon->qmp_requests->length == QMP_REQ_QUEUE_LEN_MAX - 1) {
monitor_resume(&mon->common);
}
/*
* Drop the queue mutex now, before yielding, otherwise we might
* deadlock if the main thread tries to lock it.
*/
qemu_mutex_unlock(&mon->qmp_queue_lock);
if (qatomic_xchg(&qmp_dispatcher_co_busy, true) == true) {
/*
* Someone rescheduled us (probably because a new requests
* came in), but we didn't actually yield. Do that now,
* only to be immediately reentered and removed from the
* list of scheduled coroutines.
*/
qemu_coroutine_yield();
}
/* Process request */
if (req_obj->req) {
if (trace_event_get_state(TRACE_MONITOR_QMP_CMD_IN_BAND)) {
QDict *qdict = qobject_to(QDict, req_obj->req);
QObject *id = qdict ? qdict_get(qdict, "id") : NULL;
GString *id_json;
id_json = id ? qobject_to_json(id) : g_string_new(NULL);
trace_monitor_qmp_cmd_in_band(id_json->str);
g_string_free(id_json, true);
}
monitor_qmp_dispatch(mon, req_obj->req);
} else {
assert(req_obj->err);
trace_monitor_qmp_err_in_band(error_get_pretty(req_obj->err));
rsp = qmp_error_response(req_obj->err);
req_obj->err = NULL;
monitor_qmp_respond(mon, rsp);
qobject_unref(rsp);
}
if (!oob_enabled) {
monitor_resume(&mon->common);
}
qmp_request_free(req_obj);
}
qatomic_set(&qmp_dispatcher_co, NULL);
}
void qmp_dispatcher_co_wake(void)
{
/* Write request before reading qmp_dispatcher_co_busy. */
smp_mb__before_rmw();
if (!qatomic_xchg(&qmp_dispatcher_co_busy, true)) {
aio_co_wake(qmp_dispatcher_co);
}
}
static void handle_qmp_command(void *opaque, QObject *req, Error *err)
{
MonitorQMP *mon = opaque;
QDict *qdict = qobject_to(QDict, req);
QMPRequest *req_obj;
assert(!req != !err);
if (req && trace_event_get_state_backends(TRACE_HANDLE_QMP_COMMAND)) {
GString *req_json = qobject_to_json(req);
trace_handle_qmp_command(mon, req_json->str);
g_string_free(req_json, true);
}
if (qdict && qmp_is_oob(qdict)) {
/* OOB commands are executed immediately */
if (trace_event_get_state(TRACE_MONITOR_QMP_CMD_OUT_OF_BAND)) {
QObject *id = qdict_get(qdict, "id");
GString *id_json;
id_json = id ? qobject_to_json(id) : g_string_new(NULL);
trace_monitor_qmp_cmd_out_of_band(id_json->str);
g_string_free(id_json, true);
}
monitor_qmp_dispatch(mon, req);
qobject_unref(req);
return;
}
req_obj = g_new0(QMPRequest, 1);
req_obj->mon = mon;
req_obj->req = req;
req_obj->err = err;
/* Protect qmp_requests and fetching its length. */
WITH_QEMU_LOCK_GUARD(&mon->qmp_queue_lock) {
/*
* Suspend the monitor when we can't queue more requests after
* this one. Dequeuing in monitor_qmp_dispatcher_co() or
* monitor_qmp_cleanup_queue_and_resume() will resume it.
* Note that when OOB is disabled, we queue at most one command,
* for backward compatibility.
*/
if (!qmp_oob_enabled(mon) ||
mon->qmp_requests->length == QMP_REQ_QUEUE_LEN_MAX - 1) {
monitor_suspend(&mon->common);
}
/*
* Put the request to the end of queue so that requests will be
* handled in time order. Ownership for req_obj, req,
* etc. will be delivered to the handler side.
*/
trace_monitor_qmp_in_band_enqueue(req_obj, mon,
mon->qmp_requests->length);
assert(mon->qmp_requests->length < QMP_REQ_QUEUE_LEN_MAX);
g_queue_push_tail(mon->qmp_requests, req_obj);
}
/* Kick the dispatcher routine */
qmp_dispatcher_co_wake();
}
static void monitor_qmp_read(void *opaque, const uint8_t *buf, int size)
{
MonitorQMP *mon = opaque;
json_message_parser_feed(&mon->parser, (const char *) buf, size);
}
static QDict *qmp_greeting(MonitorQMP *mon)
{
QList *cap_list = qlist_new();
QObject *ver = NULL;
QDict *args;
QMPCapability cap;
args = qdict_new();
qmp_marshal_query_version(args, &ver, NULL);
qobject_unref(args);
for (cap = 0; cap < QMP_CAPABILITY__MAX; cap++) {
if (mon->capab_offered[cap]) {
qlist_append_str(cap_list, QMPCapability_str(cap));
}
}
return qdict_from_jsonf_nofail(
"{'QMP': {'version': %p, 'capabilities': %p}}",
ver, cap_list);
}
static void monitor_qmp_event(void *opaque, QEMUChrEvent event)
{
QDict *data;
MonitorQMP *mon = opaque;
switch (event) {
case CHR_EVENT_OPENED:
mon->commands = &qmp_cap_negotiation_commands;
monitor_qmp_caps_reset(mon);
data = qmp_greeting(mon);
qmp_send_response(mon, data);
qobject_unref(data);
mon_refcount++;
break;
case CHR_EVENT_CLOSED:
/*
* Note: this is only useful when the output of the chardev
* backend is still open. For example, when the backend is
* stdio, it's possible that stdout is still open when stdin
* is closed.
*/
monitor_qmp_cleanup_queue_and_resume(mon);
json_message_parser_destroy(&mon->parser);
json_message_parser_init(&mon->parser, handle_qmp_command,
mon, NULL);
mon_refcount--;
monitor_fdsets_cleanup();
break;
case CHR_EVENT_BREAK:
case CHR_EVENT_MUX_IN:
case CHR_EVENT_MUX_OUT:
/* Ignore */
break;
}
}
void monitor_data_destroy_qmp(MonitorQMP *mon)
{
json_message_parser_destroy(&mon->parser);
qemu_mutex_destroy(&mon->qmp_queue_lock);
monitor_qmp_cleanup_req_queue_locked(mon);
g_queue_free(mon->qmp_requests);
}
static void monitor_qmp_setup_handlers_bh(void *opaque)
{
MonitorQMP *mon = opaque;
GMainContext *context;
assert(mon->common.use_io_thread);
context = iothread_get_g_main_context(mon_iothread);
assert(context);
qemu_chr_fe_set_handlers(&mon->common.chr, monitor_can_read,
monitor_qmp_read, monitor_qmp_event,
NULL, &mon->common, context, true);
monitor_list_append(&mon->common);
}
void monitor_init_qmp(Chardev *chr, bool pretty, Error **errp)
{
MonitorQMP *mon = g_new0(MonitorQMP, 1);
if (!qemu_chr_fe_init(&mon->common.chr, chr, errp)) {
g_free(mon);
return;
}
qemu_chr_fe_set_echo(&mon->common.chr, true);
/* Note: we run QMP monitor in I/O thread when @chr supports that */
monitor_data_init(&mon->common, true, false,
qemu_chr_has_feature(chr, QEMU_CHAR_FEATURE_GCONTEXT));
mon->pretty = pretty;
qemu_mutex_init(&mon->qmp_queue_lock);
mon->qmp_requests = g_queue_new();
json_message_parser_init(&mon->parser, handle_qmp_command, mon, NULL);
if (mon->common.use_io_thread) {
/*
* Make sure the old iowatch is gone. It's possible when
* e.g. the chardev is in client mode, with wait=on.
*/
remove_fd_in_watch(chr);
/*
* We can't call qemu_chr_fe_set_handlers() directly here
* since chardev might be running in the monitor I/O
* thread. Schedule a bottom half.
*/
aio_bh_schedule_oneshot(iothread_get_aio_context(mon_iothread),
monitor_qmp_setup_handlers_bh, mon);
/* The bottom half will add @mon to @mon_list */
} else {
qemu_chr_fe_set_handlers(&mon->common.chr, monitor_can_read,
monitor_qmp_read, monitor_qmp_event,
NULL, &mon->common, NULL, true);
monitor_list_append(&mon->common);
}
}