qemu/util/aio-posix.c

790 lines
23 KiB
C
Raw Normal View History

/*
* QEMU aio implementation
*
* Copyright IBM, Corp. 2008
*
* 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.
*
* Contributions after 2012-01-13 are licensed under the terms of the
* GNU GPL, version 2 or (at your option) any later version.
*/
#include "qemu/osdep.h"
#include "block/block.h"
#include "block/thread-pool.h"
#include "qemu/main-loop.h"
#include "qemu/rcu.h"
#include "qemu/rcu_queue.h"
#include "qemu/sockets.h"
aio: add polling mode to AioContext The AioContext event loop uses ppoll(2) or epoll_wait(2) to monitor file descriptors or until a timer expires. In cases like virtqueues, Linux AIO, and ThreadPool it is technically possible to wait for events via polling (i.e. continuously checking for events without blocking). Polling can be faster than blocking syscalls because file descriptors, the process scheduler, and system calls are bypassed. The main disadvantage to polling is that it increases CPU utilization. In classic polling configuration a full host CPU thread might run at 100% to respond to events as quickly as possible. This patch implements a timeout so we fall back to blocking syscalls if polling detects no activity. After the timeout no CPU cycles are wasted on polling until the next event loop iteration. The run_poll_handlers_begin() and run_poll_handlers_end() trace events are added to aid performance analysis and troubleshooting. If you need to know whether polling mode is being used, trace these events to find out. Note that the AioContext is now re-acquired before disabling notify_me in the non-polling case. This makes the code cleaner since notify_me was enabled outside the non-polling AioContext release region. This change is correct since it's safe to keep notify_me enabled longer (disabling is an optimization) but potentially causes unnecessary event_notifer_set() calls. I think the chance of performance regression is small here. Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> Reviewed-by: Paolo Bonzini <pbonzini@redhat.com> Message-id: 20161201192652.9509-4-stefanha@redhat.com Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
2016-12-01 22:26:42 +03:00
#include "qemu/cutils.h"
#include "trace.h"
#include "aio-posix.h"
/* Stop userspace polling on a handler if it isn't active for some time */
#define POLL_IDLE_INTERVAL_NS (7 * NANOSECONDS_PER_SECOND)
bool aio_poll_disabled(AioContext *ctx)
{
return qatomic_read(&ctx->poll_disable_cnt);
}
void aio_add_ready_handler(AioHandlerList *ready_list,
AioHandler *node,
int revents)
{
QLIST_SAFE_REMOVE(node, node_ready); /* remove from nested parent's list */
node->pfd.revents = revents;
QLIST_INSERT_HEAD(ready_list, node, node_ready);
}
static void aio_add_poll_ready_handler(AioHandlerList *ready_list,
AioHandler *node)
{
QLIST_SAFE_REMOVE(node, node_ready); /* remove from nested parent's list */
node->poll_ready = true;
QLIST_INSERT_HEAD(ready_list, node, node_ready);
}
static AioHandler *find_aio_handler(AioContext *ctx, int fd)
{
AioHandler *node;
QLIST_FOREACH(node, &ctx->aio_handlers, node) {
if (node->pfd.fd == fd) {
if (!QLIST_IS_INSERTED(node, node_deleted)) {
return node;
}
}
}
return NULL;
}
static bool aio_remove_fd_handler(AioContext *ctx, AioHandler *node)
{
/* If the GSource is in the process of being destroyed then
* g_source_remove_poll() causes an assertion failure. Skip
* removal in that case, because glib cleans up its state during
* destruction anyway.
*/
if (!g_source_is_destroyed(&ctx->source)) {
g_source_remove_poll(&ctx->source, &node->pfd);
}
node->pfd.revents = 0;
node->poll_ready = false;
/* If the fd monitor has already marked it deleted, leave it alone */
if (QLIST_IS_INSERTED(node, node_deleted)) {
return false;
}
/* If a read is in progress, just mark the node as deleted */
if (qemu_lockcnt_count(&ctx->list_lock)) {
QLIST_INSERT_HEAD_RCU(&ctx->deleted_aio_handlers, node, node_deleted);
return false;
}
/* Otherwise, delete it for real. We can't just mark it as
* deleted because deleted nodes are only cleaned up while
* no one is walking the handlers list.
*/
QLIST_SAFE_REMOVE(node, node_poll);
QLIST_REMOVE(node, node);
return true;
}
void aio_set_fd_handler(AioContext *ctx,
int fd,
bool is_external,
IOHandler *io_read,
IOHandler *io_write,
AioPollFn *io_poll,
aio-posix: split poll check from ready handler Adaptive polling measures the execution time of the polling check plus handlers called when a polled event becomes ready. Handlers can take a significant amount of time, making it look like polling was running for a long time when in fact the event handler was running for a long time. For example, on Linux the io_submit(2) syscall invoked when a virtio-blk device's virtqueue becomes ready can take 10s of microseconds. This can exceed the default polling interval (32 microseconds) and cause adaptive polling to stop polling. By excluding the handler's execution time from the polling check we make the adaptive polling calculation more accurate. As a result, the event loop now stays in polling mode where previously it would have fallen back to file descriptor monitoring. The following data was collected with virtio-blk num-queues=2 event_idx=off using an IOThread. Before: 168k IOPS, IOThread syscalls: 9837.115 ( 0.020 ms): IO iothread1/620155 io_submit(ctx_id: 140512552468480, nr: 16, iocbpp: 0x7fcb9f937db0) = 16 9837.158 ( 0.002 ms): IO iothread1/620155 write(fd: 103, buf: 0x556a2ef71b88, count: 8) = 8 9837.161 ( 0.001 ms): IO iothread1/620155 write(fd: 104, buf: 0x556a2ef71b88, count: 8) = 8 9837.163 ( 0.001 ms): IO iothread1/620155 ppoll(ufds: 0x7fcb90002800, nfds: 4, tsp: 0x7fcb9f1342d0, sigsetsize: 8) = 3 9837.164 ( 0.001 ms): IO iothread1/620155 read(fd: 107, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.174 ( 0.001 ms): IO iothread1/620155 read(fd: 105, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.176 ( 0.001 ms): IO iothread1/620155 read(fd: 106, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.209 ( 0.035 ms): IO iothread1/620155 io_submit(ctx_id: 140512552468480, nr: 32, iocbpp: 0x7fca7d0cebe0) = 32 174k IOPS (+3.6%), IOThread syscalls: 9809.566 ( 0.036 ms): IO iothread1/623061 io_submit(ctx_id: 140539805028352, nr: 32, iocbpp: 0x7fd0cdd62be0) = 32 9809.625 ( 0.001 ms): IO iothread1/623061 write(fd: 103, buf: 0x5647cfba5f58, count: 8) = 8 9809.627 ( 0.002 ms): IO iothread1/623061 write(fd: 104, buf: 0x5647cfba5f58, count: 8) = 8 9809.663 ( 0.036 ms): IO iothread1/623061 io_submit(ctx_id: 140539805028352, nr: 32, iocbpp: 0x7fd0d0388b50) = 32 Notice that ppoll(2) and eventfd read(2) syscalls are eliminated because the IOThread stays in polling mode instead of falling back to file descriptor monitoring. As usual, polling is not implemented on Windows so this patch ignores the new io_poll_read() callback in aio-win32.c. Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> Reviewed-by: Stefano Garzarella <sgarzare@redhat.com> Message-id: 20211207132336.36627-2-stefanha@redhat.com [Fixed up aio_set_event_notifier() calls in tests/unit/test-fdmon-epoll.c added after this series was queued. --Stefan] Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
2021-12-07 16:23:31 +03:00
IOHandler *io_poll_ready,
void *opaque)
{
AioHandler *node;
AioHandler *new_node = NULL;
bool is_new = false;
bool deleted = false;
int poll_disable_change;
aio-posix: split poll check from ready handler Adaptive polling measures the execution time of the polling check plus handlers called when a polled event becomes ready. Handlers can take a significant amount of time, making it look like polling was running for a long time when in fact the event handler was running for a long time. For example, on Linux the io_submit(2) syscall invoked when a virtio-blk device's virtqueue becomes ready can take 10s of microseconds. This can exceed the default polling interval (32 microseconds) and cause adaptive polling to stop polling. By excluding the handler's execution time from the polling check we make the adaptive polling calculation more accurate. As a result, the event loop now stays in polling mode where previously it would have fallen back to file descriptor monitoring. The following data was collected with virtio-blk num-queues=2 event_idx=off using an IOThread. Before: 168k IOPS, IOThread syscalls: 9837.115 ( 0.020 ms): IO iothread1/620155 io_submit(ctx_id: 140512552468480, nr: 16, iocbpp: 0x7fcb9f937db0) = 16 9837.158 ( 0.002 ms): IO iothread1/620155 write(fd: 103, buf: 0x556a2ef71b88, count: 8) = 8 9837.161 ( 0.001 ms): IO iothread1/620155 write(fd: 104, buf: 0x556a2ef71b88, count: 8) = 8 9837.163 ( 0.001 ms): IO iothread1/620155 ppoll(ufds: 0x7fcb90002800, nfds: 4, tsp: 0x7fcb9f1342d0, sigsetsize: 8) = 3 9837.164 ( 0.001 ms): IO iothread1/620155 read(fd: 107, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.174 ( 0.001 ms): IO iothread1/620155 read(fd: 105, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.176 ( 0.001 ms): IO iothread1/620155 read(fd: 106, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.209 ( 0.035 ms): IO iothread1/620155 io_submit(ctx_id: 140512552468480, nr: 32, iocbpp: 0x7fca7d0cebe0) = 32 174k IOPS (+3.6%), IOThread syscalls: 9809.566 ( 0.036 ms): IO iothread1/623061 io_submit(ctx_id: 140539805028352, nr: 32, iocbpp: 0x7fd0cdd62be0) = 32 9809.625 ( 0.001 ms): IO iothread1/623061 write(fd: 103, buf: 0x5647cfba5f58, count: 8) = 8 9809.627 ( 0.002 ms): IO iothread1/623061 write(fd: 104, buf: 0x5647cfba5f58, count: 8) = 8 9809.663 ( 0.036 ms): IO iothread1/623061 io_submit(ctx_id: 140539805028352, nr: 32, iocbpp: 0x7fd0d0388b50) = 32 Notice that ppoll(2) and eventfd read(2) syscalls are eliminated because the IOThread stays in polling mode instead of falling back to file descriptor monitoring. As usual, polling is not implemented on Windows so this patch ignores the new io_poll_read() callback in aio-win32.c. Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> Reviewed-by: Stefano Garzarella <sgarzare@redhat.com> Message-id: 20211207132336.36627-2-stefanha@redhat.com [Fixed up aio_set_event_notifier() calls in tests/unit/test-fdmon-epoll.c added after this series was queued. --Stefan] Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
2021-12-07 16:23:31 +03:00
if (io_poll && !io_poll_ready) {
io_poll = NULL; /* polling only makes sense if there is a handler */
}
qemu_lockcnt_lock(&ctx->list_lock);
node = find_aio_handler(ctx, fd);
/* Are we deleting the fd handler? */
aio: add polling mode to AioContext The AioContext event loop uses ppoll(2) or epoll_wait(2) to monitor file descriptors or until a timer expires. In cases like virtqueues, Linux AIO, and ThreadPool it is technically possible to wait for events via polling (i.e. continuously checking for events without blocking). Polling can be faster than blocking syscalls because file descriptors, the process scheduler, and system calls are bypassed. The main disadvantage to polling is that it increases CPU utilization. In classic polling configuration a full host CPU thread might run at 100% to respond to events as quickly as possible. This patch implements a timeout so we fall back to blocking syscalls if polling detects no activity. After the timeout no CPU cycles are wasted on polling until the next event loop iteration. The run_poll_handlers_begin() and run_poll_handlers_end() trace events are added to aid performance analysis and troubleshooting. If you need to know whether polling mode is being used, trace these events to find out. Note that the AioContext is now re-acquired before disabling notify_me in the non-polling case. This makes the code cleaner since notify_me was enabled outside the non-polling AioContext release region. This change is correct since it's safe to keep notify_me enabled longer (disabling is an optimization) but potentially causes unnecessary event_notifer_set() calls. I think the chance of performance regression is small here. Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> Reviewed-by: Paolo Bonzini <pbonzini@redhat.com> Message-id: 20161201192652.9509-4-stefanha@redhat.com Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
2016-12-01 22:26:42 +03:00
if (!io_read && !io_write && !io_poll) {
if (node == NULL) {
qemu_lockcnt_unlock(&ctx->list_lock);
return;
}
/* Clean events in order to unregister fd from the ctx epoll. */
node->pfd.events = 0;
poll_disable_change = -!node->io_poll;
} else {
poll_disable_change = !io_poll - (node && !node->io_poll);
if (node == NULL) {
is_new = true;
}
/* Alloc and insert if it's not already there */
new_node = g_new0(AioHandler, 1);
aio: add polling mode to AioContext The AioContext event loop uses ppoll(2) or epoll_wait(2) to monitor file descriptors or until a timer expires. In cases like virtqueues, Linux AIO, and ThreadPool it is technically possible to wait for events via polling (i.e. continuously checking for events without blocking). Polling can be faster than blocking syscalls because file descriptors, the process scheduler, and system calls are bypassed. The main disadvantage to polling is that it increases CPU utilization. In classic polling configuration a full host CPU thread might run at 100% to respond to events as quickly as possible. This patch implements a timeout so we fall back to blocking syscalls if polling detects no activity. After the timeout no CPU cycles are wasted on polling until the next event loop iteration. The run_poll_handlers_begin() and run_poll_handlers_end() trace events are added to aid performance analysis and troubleshooting. If you need to know whether polling mode is being used, trace these events to find out. Note that the AioContext is now re-acquired before disabling notify_me in the non-polling case. This makes the code cleaner since notify_me was enabled outside the non-polling AioContext release region. This change is correct since it's safe to keep notify_me enabled longer (disabling is an optimization) but potentially causes unnecessary event_notifer_set() calls. I think the chance of performance regression is small here. Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> Reviewed-by: Paolo Bonzini <pbonzini@redhat.com> Message-id: 20161201192652.9509-4-stefanha@redhat.com Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
2016-12-01 22:26:42 +03:00
/* Update handler with latest information */
new_node->io_read = io_read;
new_node->io_write = io_write;
new_node->io_poll = io_poll;
aio-posix: split poll check from ready handler Adaptive polling measures the execution time of the polling check plus handlers called when a polled event becomes ready. Handlers can take a significant amount of time, making it look like polling was running for a long time when in fact the event handler was running for a long time. For example, on Linux the io_submit(2) syscall invoked when a virtio-blk device's virtqueue becomes ready can take 10s of microseconds. This can exceed the default polling interval (32 microseconds) and cause adaptive polling to stop polling. By excluding the handler's execution time from the polling check we make the adaptive polling calculation more accurate. As a result, the event loop now stays in polling mode where previously it would have fallen back to file descriptor monitoring. The following data was collected with virtio-blk num-queues=2 event_idx=off using an IOThread. Before: 168k IOPS, IOThread syscalls: 9837.115 ( 0.020 ms): IO iothread1/620155 io_submit(ctx_id: 140512552468480, nr: 16, iocbpp: 0x7fcb9f937db0) = 16 9837.158 ( 0.002 ms): IO iothread1/620155 write(fd: 103, buf: 0x556a2ef71b88, count: 8) = 8 9837.161 ( 0.001 ms): IO iothread1/620155 write(fd: 104, buf: 0x556a2ef71b88, count: 8) = 8 9837.163 ( 0.001 ms): IO iothread1/620155 ppoll(ufds: 0x7fcb90002800, nfds: 4, tsp: 0x7fcb9f1342d0, sigsetsize: 8) = 3 9837.164 ( 0.001 ms): IO iothread1/620155 read(fd: 107, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.174 ( 0.001 ms): IO iothread1/620155 read(fd: 105, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.176 ( 0.001 ms): IO iothread1/620155 read(fd: 106, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.209 ( 0.035 ms): IO iothread1/620155 io_submit(ctx_id: 140512552468480, nr: 32, iocbpp: 0x7fca7d0cebe0) = 32 174k IOPS (+3.6%), IOThread syscalls: 9809.566 ( 0.036 ms): IO iothread1/623061 io_submit(ctx_id: 140539805028352, nr: 32, iocbpp: 0x7fd0cdd62be0) = 32 9809.625 ( 0.001 ms): IO iothread1/623061 write(fd: 103, buf: 0x5647cfba5f58, count: 8) = 8 9809.627 ( 0.002 ms): IO iothread1/623061 write(fd: 104, buf: 0x5647cfba5f58, count: 8) = 8 9809.663 ( 0.036 ms): IO iothread1/623061 io_submit(ctx_id: 140539805028352, nr: 32, iocbpp: 0x7fd0d0388b50) = 32 Notice that ppoll(2) and eventfd read(2) syscalls are eliminated because the IOThread stays in polling mode instead of falling back to file descriptor monitoring. As usual, polling is not implemented on Windows so this patch ignores the new io_poll_read() callback in aio-win32.c. Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> Reviewed-by: Stefano Garzarella <sgarzare@redhat.com> Message-id: 20211207132336.36627-2-stefanha@redhat.com [Fixed up aio_set_event_notifier() calls in tests/unit/test-fdmon-epoll.c added after this series was queued. --Stefan] Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
2021-12-07 16:23:31 +03:00
new_node->io_poll_ready = io_poll_ready;
new_node->opaque = opaque;
new_node->is_external = is_external;
if (is_new) {
new_node->pfd.fd = fd;
} else {
new_node->pfd = node->pfd;
}
g_source_add_poll(&ctx->source, &new_node->pfd);
new_node->pfd.events = (io_read ? G_IO_IN | G_IO_HUP | G_IO_ERR : 0);
new_node->pfd.events |= (io_write ? G_IO_OUT | G_IO_ERR : 0);
QLIST_INSERT_HEAD_RCU(&ctx->aio_handlers, new_node, node);
}
/* No need to order poll_disable_cnt writes against other updates;
* the counter is only used to avoid wasting time and latency on
* iterated polling when the system call will be ultimately necessary.
* Changing handlers is a rare event, and a little wasted polling until
* the aio_notify below is not an issue.
*/
qatomic_set(&ctx->poll_disable_cnt,
qatomic_read(&ctx->poll_disable_cnt) + poll_disable_change);
ctx->fdmon_ops->update(ctx, node, new_node);
if (node) {
deleted = aio_remove_fd_handler(ctx, node);
}
qemu_lockcnt_unlock(&ctx->list_lock);
aio_notify(ctx);
aio: add polling mode to AioContext The AioContext event loop uses ppoll(2) or epoll_wait(2) to monitor file descriptors or until a timer expires. In cases like virtqueues, Linux AIO, and ThreadPool it is technically possible to wait for events via polling (i.e. continuously checking for events without blocking). Polling can be faster than blocking syscalls because file descriptors, the process scheduler, and system calls are bypassed. The main disadvantage to polling is that it increases CPU utilization. In classic polling configuration a full host CPU thread might run at 100% to respond to events as quickly as possible. This patch implements a timeout so we fall back to blocking syscalls if polling detects no activity. After the timeout no CPU cycles are wasted on polling until the next event loop iteration. The run_poll_handlers_begin() and run_poll_handlers_end() trace events are added to aid performance analysis and troubleshooting. If you need to know whether polling mode is being used, trace these events to find out. Note that the AioContext is now re-acquired before disabling notify_me in the non-polling case. This makes the code cleaner since notify_me was enabled outside the non-polling AioContext release region. This change is correct since it's safe to keep notify_me enabled longer (disabling is an optimization) but potentially causes unnecessary event_notifer_set() calls. I think the chance of performance regression is small here. Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> Reviewed-by: Paolo Bonzini <pbonzini@redhat.com> Message-id: 20161201192652.9509-4-stefanha@redhat.com Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
2016-12-01 22:26:42 +03:00
if (deleted) {
g_free(node);
}
}
void aio_set_fd_poll(AioContext *ctx, int fd,
IOHandler *io_poll_begin,
IOHandler *io_poll_end)
{
AioHandler *node = find_aio_handler(ctx, fd);
if (!node) {
return;
}
node->io_poll_begin = io_poll_begin;
node->io_poll_end = io_poll_end;
}
void aio_set_event_notifier(AioContext *ctx,
EventNotifier *notifier,
bool is_external,
EventNotifierHandler *io_read,
aio-posix: split poll check from ready handler Adaptive polling measures the execution time of the polling check plus handlers called when a polled event becomes ready. Handlers can take a significant amount of time, making it look like polling was running for a long time when in fact the event handler was running for a long time. For example, on Linux the io_submit(2) syscall invoked when a virtio-blk device's virtqueue becomes ready can take 10s of microseconds. This can exceed the default polling interval (32 microseconds) and cause adaptive polling to stop polling. By excluding the handler's execution time from the polling check we make the adaptive polling calculation more accurate. As a result, the event loop now stays in polling mode where previously it would have fallen back to file descriptor monitoring. The following data was collected with virtio-blk num-queues=2 event_idx=off using an IOThread. Before: 168k IOPS, IOThread syscalls: 9837.115 ( 0.020 ms): IO iothread1/620155 io_submit(ctx_id: 140512552468480, nr: 16, iocbpp: 0x7fcb9f937db0) = 16 9837.158 ( 0.002 ms): IO iothread1/620155 write(fd: 103, buf: 0x556a2ef71b88, count: 8) = 8 9837.161 ( 0.001 ms): IO iothread1/620155 write(fd: 104, buf: 0x556a2ef71b88, count: 8) = 8 9837.163 ( 0.001 ms): IO iothread1/620155 ppoll(ufds: 0x7fcb90002800, nfds: 4, tsp: 0x7fcb9f1342d0, sigsetsize: 8) = 3 9837.164 ( 0.001 ms): IO iothread1/620155 read(fd: 107, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.174 ( 0.001 ms): IO iothread1/620155 read(fd: 105, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.176 ( 0.001 ms): IO iothread1/620155 read(fd: 106, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.209 ( 0.035 ms): IO iothread1/620155 io_submit(ctx_id: 140512552468480, nr: 32, iocbpp: 0x7fca7d0cebe0) = 32 174k IOPS (+3.6%), IOThread syscalls: 9809.566 ( 0.036 ms): IO iothread1/623061 io_submit(ctx_id: 140539805028352, nr: 32, iocbpp: 0x7fd0cdd62be0) = 32 9809.625 ( 0.001 ms): IO iothread1/623061 write(fd: 103, buf: 0x5647cfba5f58, count: 8) = 8 9809.627 ( 0.002 ms): IO iothread1/623061 write(fd: 104, buf: 0x5647cfba5f58, count: 8) = 8 9809.663 ( 0.036 ms): IO iothread1/623061 io_submit(ctx_id: 140539805028352, nr: 32, iocbpp: 0x7fd0d0388b50) = 32 Notice that ppoll(2) and eventfd read(2) syscalls are eliminated because the IOThread stays in polling mode instead of falling back to file descriptor monitoring. As usual, polling is not implemented on Windows so this patch ignores the new io_poll_read() callback in aio-win32.c. Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> Reviewed-by: Stefano Garzarella <sgarzare@redhat.com> Message-id: 20211207132336.36627-2-stefanha@redhat.com [Fixed up aio_set_event_notifier() calls in tests/unit/test-fdmon-epoll.c added after this series was queued. --Stefan] Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
2021-12-07 16:23:31 +03:00
AioPollFn *io_poll,
EventNotifierHandler *io_poll_ready)
{
aio_set_fd_handler(ctx, event_notifier_get_fd(notifier), is_external,
aio-posix: split poll check from ready handler Adaptive polling measures the execution time of the polling check plus handlers called when a polled event becomes ready. Handlers can take a significant amount of time, making it look like polling was running for a long time when in fact the event handler was running for a long time. For example, on Linux the io_submit(2) syscall invoked when a virtio-blk device's virtqueue becomes ready can take 10s of microseconds. This can exceed the default polling interval (32 microseconds) and cause adaptive polling to stop polling. By excluding the handler's execution time from the polling check we make the adaptive polling calculation more accurate. As a result, the event loop now stays in polling mode where previously it would have fallen back to file descriptor monitoring. The following data was collected with virtio-blk num-queues=2 event_idx=off using an IOThread. Before: 168k IOPS, IOThread syscalls: 9837.115 ( 0.020 ms): IO iothread1/620155 io_submit(ctx_id: 140512552468480, nr: 16, iocbpp: 0x7fcb9f937db0) = 16 9837.158 ( 0.002 ms): IO iothread1/620155 write(fd: 103, buf: 0x556a2ef71b88, count: 8) = 8 9837.161 ( 0.001 ms): IO iothread1/620155 write(fd: 104, buf: 0x556a2ef71b88, count: 8) = 8 9837.163 ( 0.001 ms): IO iothread1/620155 ppoll(ufds: 0x7fcb90002800, nfds: 4, tsp: 0x7fcb9f1342d0, sigsetsize: 8) = 3 9837.164 ( 0.001 ms): IO iothread1/620155 read(fd: 107, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.174 ( 0.001 ms): IO iothread1/620155 read(fd: 105, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.176 ( 0.001 ms): IO iothread1/620155 read(fd: 106, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.209 ( 0.035 ms): IO iothread1/620155 io_submit(ctx_id: 140512552468480, nr: 32, iocbpp: 0x7fca7d0cebe0) = 32 174k IOPS (+3.6%), IOThread syscalls: 9809.566 ( 0.036 ms): IO iothread1/623061 io_submit(ctx_id: 140539805028352, nr: 32, iocbpp: 0x7fd0cdd62be0) = 32 9809.625 ( 0.001 ms): IO iothread1/623061 write(fd: 103, buf: 0x5647cfba5f58, count: 8) = 8 9809.627 ( 0.002 ms): IO iothread1/623061 write(fd: 104, buf: 0x5647cfba5f58, count: 8) = 8 9809.663 ( 0.036 ms): IO iothread1/623061 io_submit(ctx_id: 140539805028352, nr: 32, iocbpp: 0x7fd0d0388b50) = 32 Notice that ppoll(2) and eventfd read(2) syscalls are eliminated because the IOThread stays in polling mode instead of falling back to file descriptor monitoring. As usual, polling is not implemented on Windows so this patch ignores the new io_poll_read() callback in aio-win32.c. Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> Reviewed-by: Stefano Garzarella <sgarzare@redhat.com> Message-id: 20211207132336.36627-2-stefanha@redhat.com [Fixed up aio_set_event_notifier() calls in tests/unit/test-fdmon-epoll.c added after this series was queued. --Stefan] Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
2021-12-07 16:23:31 +03:00
(IOHandler *)io_read, NULL, io_poll,
(IOHandler *)io_poll_ready, notifier);
}
void aio_set_event_notifier_poll(AioContext *ctx,
EventNotifier *notifier,
EventNotifierHandler *io_poll_begin,
EventNotifierHandler *io_poll_end)
{
aio_set_fd_poll(ctx, event_notifier_get_fd(notifier),
(IOHandler *)io_poll_begin,
(IOHandler *)io_poll_end);
}
aio-posix: split poll check from ready handler Adaptive polling measures the execution time of the polling check plus handlers called when a polled event becomes ready. Handlers can take a significant amount of time, making it look like polling was running for a long time when in fact the event handler was running for a long time. For example, on Linux the io_submit(2) syscall invoked when a virtio-blk device's virtqueue becomes ready can take 10s of microseconds. This can exceed the default polling interval (32 microseconds) and cause adaptive polling to stop polling. By excluding the handler's execution time from the polling check we make the adaptive polling calculation more accurate. As a result, the event loop now stays in polling mode where previously it would have fallen back to file descriptor monitoring. The following data was collected with virtio-blk num-queues=2 event_idx=off using an IOThread. Before: 168k IOPS, IOThread syscalls: 9837.115 ( 0.020 ms): IO iothread1/620155 io_submit(ctx_id: 140512552468480, nr: 16, iocbpp: 0x7fcb9f937db0) = 16 9837.158 ( 0.002 ms): IO iothread1/620155 write(fd: 103, buf: 0x556a2ef71b88, count: 8) = 8 9837.161 ( 0.001 ms): IO iothread1/620155 write(fd: 104, buf: 0x556a2ef71b88, count: 8) = 8 9837.163 ( 0.001 ms): IO iothread1/620155 ppoll(ufds: 0x7fcb90002800, nfds: 4, tsp: 0x7fcb9f1342d0, sigsetsize: 8) = 3 9837.164 ( 0.001 ms): IO iothread1/620155 read(fd: 107, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.174 ( 0.001 ms): IO iothread1/620155 read(fd: 105, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.176 ( 0.001 ms): IO iothread1/620155 read(fd: 106, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.209 ( 0.035 ms): IO iothread1/620155 io_submit(ctx_id: 140512552468480, nr: 32, iocbpp: 0x7fca7d0cebe0) = 32 174k IOPS (+3.6%), IOThread syscalls: 9809.566 ( 0.036 ms): IO iothread1/623061 io_submit(ctx_id: 140539805028352, nr: 32, iocbpp: 0x7fd0cdd62be0) = 32 9809.625 ( 0.001 ms): IO iothread1/623061 write(fd: 103, buf: 0x5647cfba5f58, count: 8) = 8 9809.627 ( 0.002 ms): IO iothread1/623061 write(fd: 104, buf: 0x5647cfba5f58, count: 8) = 8 9809.663 ( 0.036 ms): IO iothread1/623061 io_submit(ctx_id: 140539805028352, nr: 32, iocbpp: 0x7fd0d0388b50) = 32 Notice that ppoll(2) and eventfd read(2) syscalls are eliminated because the IOThread stays in polling mode instead of falling back to file descriptor monitoring. As usual, polling is not implemented on Windows so this patch ignores the new io_poll_read() callback in aio-win32.c. Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> Reviewed-by: Stefano Garzarella <sgarzare@redhat.com> Message-id: 20211207132336.36627-2-stefanha@redhat.com [Fixed up aio_set_event_notifier() calls in tests/unit/test-fdmon-epoll.c added after this series was queued. --Stefan] Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
2021-12-07 16:23:31 +03:00
static bool poll_set_started(AioContext *ctx, AioHandlerList *ready_list,
bool started)
{
AioHandler *node;
bool progress = false;
if (started == ctx->poll_started) {
return false;
}
ctx->poll_started = started;
qemu_lockcnt_inc(&ctx->list_lock);
QLIST_FOREACH(node, &ctx->poll_aio_handlers, node_poll) {
IOHandler *fn;
if (QLIST_IS_INSERTED(node, node_deleted)) {
continue;
}
if (started) {
fn = node->io_poll_begin;
} else {
fn = node->io_poll_end;
}
if (fn) {
fn(node->opaque);
}
/* Poll one last time in case ->io_poll_end() raced with the event */
aio-posix: split poll check from ready handler Adaptive polling measures the execution time of the polling check plus handlers called when a polled event becomes ready. Handlers can take a significant amount of time, making it look like polling was running for a long time when in fact the event handler was running for a long time. For example, on Linux the io_submit(2) syscall invoked when a virtio-blk device's virtqueue becomes ready can take 10s of microseconds. This can exceed the default polling interval (32 microseconds) and cause adaptive polling to stop polling. By excluding the handler's execution time from the polling check we make the adaptive polling calculation more accurate. As a result, the event loop now stays in polling mode where previously it would have fallen back to file descriptor monitoring. The following data was collected with virtio-blk num-queues=2 event_idx=off using an IOThread. Before: 168k IOPS, IOThread syscalls: 9837.115 ( 0.020 ms): IO iothread1/620155 io_submit(ctx_id: 140512552468480, nr: 16, iocbpp: 0x7fcb9f937db0) = 16 9837.158 ( 0.002 ms): IO iothread1/620155 write(fd: 103, buf: 0x556a2ef71b88, count: 8) = 8 9837.161 ( 0.001 ms): IO iothread1/620155 write(fd: 104, buf: 0x556a2ef71b88, count: 8) = 8 9837.163 ( 0.001 ms): IO iothread1/620155 ppoll(ufds: 0x7fcb90002800, nfds: 4, tsp: 0x7fcb9f1342d0, sigsetsize: 8) = 3 9837.164 ( 0.001 ms): IO iothread1/620155 read(fd: 107, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.174 ( 0.001 ms): IO iothread1/620155 read(fd: 105, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.176 ( 0.001 ms): IO iothread1/620155 read(fd: 106, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.209 ( 0.035 ms): IO iothread1/620155 io_submit(ctx_id: 140512552468480, nr: 32, iocbpp: 0x7fca7d0cebe0) = 32 174k IOPS (+3.6%), IOThread syscalls: 9809.566 ( 0.036 ms): IO iothread1/623061 io_submit(ctx_id: 140539805028352, nr: 32, iocbpp: 0x7fd0cdd62be0) = 32 9809.625 ( 0.001 ms): IO iothread1/623061 write(fd: 103, buf: 0x5647cfba5f58, count: 8) = 8 9809.627 ( 0.002 ms): IO iothread1/623061 write(fd: 104, buf: 0x5647cfba5f58, count: 8) = 8 9809.663 ( 0.036 ms): IO iothread1/623061 io_submit(ctx_id: 140539805028352, nr: 32, iocbpp: 0x7fd0d0388b50) = 32 Notice that ppoll(2) and eventfd read(2) syscalls are eliminated because the IOThread stays in polling mode instead of falling back to file descriptor monitoring. As usual, polling is not implemented on Windows so this patch ignores the new io_poll_read() callback in aio-win32.c. Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> Reviewed-by: Stefano Garzarella <sgarzare@redhat.com> Message-id: 20211207132336.36627-2-stefanha@redhat.com [Fixed up aio_set_event_notifier() calls in tests/unit/test-fdmon-epoll.c added after this series was queued. --Stefan] Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
2021-12-07 16:23:31 +03:00
if (!started && node->io_poll(node->opaque)) {
aio_add_poll_ready_handler(ready_list, node);
aio-posix: split poll check from ready handler Adaptive polling measures the execution time of the polling check plus handlers called when a polled event becomes ready. Handlers can take a significant amount of time, making it look like polling was running for a long time when in fact the event handler was running for a long time. For example, on Linux the io_submit(2) syscall invoked when a virtio-blk device's virtqueue becomes ready can take 10s of microseconds. This can exceed the default polling interval (32 microseconds) and cause adaptive polling to stop polling. By excluding the handler's execution time from the polling check we make the adaptive polling calculation more accurate. As a result, the event loop now stays in polling mode where previously it would have fallen back to file descriptor monitoring. The following data was collected with virtio-blk num-queues=2 event_idx=off using an IOThread. Before: 168k IOPS, IOThread syscalls: 9837.115 ( 0.020 ms): IO iothread1/620155 io_submit(ctx_id: 140512552468480, nr: 16, iocbpp: 0x7fcb9f937db0) = 16 9837.158 ( 0.002 ms): IO iothread1/620155 write(fd: 103, buf: 0x556a2ef71b88, count: 8) = 8 9837.161 ( 0.001 ms): IO iothread1/620155 write(fd: 104, buf: 0x556a2ef71b88, count: 8) = 8 9837.163 ( 0.001 ms): IO iothread1/620155 ppoll(ufds: 0x7fcb90002800, nfds: 4, tsp: 0x7fcb9f1342d0, sigsetsize: 8) = 3 9837.164 ( 0.001 ms): IO iothread1/620155 read(fd: 107, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.174 ( 0.001 ms): IO iothread1/620155 read(fd: 105, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.176 ( 0.001 ms): IO iothread1/620155 read(fd: 106, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.209 ( 0.035 ms): IO iothread1/620155 io_submit(ctx_id: 140512552468480, nr: 32, iocbpp: 0x7fca7d0cebe0) = 32 174k IOPS (+3.6%), IOThread syscalls: 9809.566 ( 0.036 ms): IO iothread1/623061 io_submit(ctx_id: 140539805028352, nr: 32, iocbpp: 0x7fd0cdd62be0) = 32 9809.625 ( 0.001 ms): IO iothread1/623061 write(fd: 103, buf: 0x5647cfba5f58, count: 8) = 8 9809.627 ( 0.002 ms): IO iothread1/623061 write(fd: 104, buf: 0x5647cfba5f58, count: 8) = 8 9809.663 ( 0.036 ms): IO iothread1/623061 io_submit(ctx_id: 140539805028352, nr: 32, iocbpp: 0x7fd0d0388b50) = 32 Notice that ppoll(2) and eventfd read(2) syscalls are eliminated because the IOThread stays in polling mode instead of falling back to file descriptor monitoring. As usual, polling is not implemented on Windows so this patch ignores the new io_poll_read() callback in aio-win32.c. Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> Reviewed-by: Stefano Garzarella <sgarzare@redhat.com> Message-id: 20211207132336.36627-2-stefanha@redhat.com [Fixed up aio_set_event_notifier() calls in tests/unit/test-fdmon-epoll.c added after this series was queued. --Stefan] Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
2021-12-07 16:23:31 +03:00
progress = true;
}
}
qemu_lockcnt_dec(&ctx->list_lock);
return progress;
}
bool aio_prepare(AioContext *ctx)
{
aio-posix: split poll check from ready handler Adaptive polling measures the execution time of the polling check plus handlers called when a polled event becomes ready. Handlers can take a significant amount of time, making it look like polling was running for a long time when in fact the event handler was running for a long time. For example, on Linux the io_submit(2) syscall invoked when a virtio-blk device's virtqueue becomes ready can take 10s of microseconds. This can exceed the default polling interval (32 microseconds) and cause adaptive polling to stop polling. By excluding the handler's execution time from the polling check we make the adaptive polling calculation more accurate. As a result, the event loop now stays in polling mode where previously it would have fallen back to file descriptor monitoring. The following data was collected with virtio-blk num-queues=2 event_idx=off using an IOThread. Before: 168k IOPS, IOThread syscalls: 9837.115 ( 0.020 ms): IO iothread1/620155 io_submit(ctx_id: 140512552468480, nr: 16, iocbpp: 0x7fcb9f937db0) = 16 9837.158 ( 0.002 ms): IO iothread1/620155 write(fd: 103, buf: 0x556a2ef71b88, count: 8) = 8 9837.161 ( 0.001 ms): IO iothread1/620155 write(fd: 104, buf: 0x556a2ef71b88, count: 8) = 8 9837.163 ( 0.001 ms): IO iothread1/620155 ppoll(ufds: 0x7fcb90002800, nfds: 4, tsp: 0x7fcb9f1342d0, sigsetsize: 8) = 3 9837.164 ( 0.001 ms): IO iothread1/620155 read(fd: 107, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.174 ( 0.001 ms): IO iothread1/620155 read(fd: 105, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.176 ( 0.001 ms): IO iothread1/620155 read(fd: 106, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.209 ( 0.035 ms): IO iothread1/620155 io_submit(ctx_id: 140512552468480, nr: 32, iocbpp: 0x7fca7d0cebe0) = 32 174k IOPS (+3.6%), IOThread syscalls: 9809.566 ( 0.036 ms): IO iothread1/623061 io_submit(ctx_id: 140539805028352, nr: 32, iocbpp: 0x7fd0cdd62be0) = 32 9809.625 ( 0.001 ms): IO iothread1/623061 write(fd: 103, buf: 0x5647cfba5f58, count: 8) = 8 9809.627 ( 0.002 ms): IO iothread1/623061 write(fd: 104, buf: 0x5647cfba5f58, count: 8) = 8 9809.663 ( 0.036 ms): IO iothread1/623061 io_submit(ctx_id: 140539805028352, nr: 32, iocbpp: 0x7fd0d0388b50) = 32 Notice that ppoll(2) and eventfd read(2) syscalls are eliminated because the IOThread stays in polling mode instead of falling back to file descriptor monitoring. As usual, polling is not implemented on Windows so this patch ignores the new io_poll_read() callback in aio-win32.c. Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> Reviewed-by: Stefano Garzarella <sgarzare@redhat.com> Message-id: 20211207132336.36627-2-stefanha@redhat.com [Fixed up aio_set_event_notifier() calls in tests/unit/test-fdmon-epoll.c added after this series was queued. --Stefan] Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
2021-12-07 16:23:31 +03:00
AioHandlerList ready_list = QLIST_HEAD_INITIALIZER(ready_list);
/* Poll mode cannot be used with glib's event loop, disable it. */
aio-posix: split poll check from ready handler Adaptive polling measures the execution time of the polling check plus handlers called when a polled event becomes ready. Handlers can take a significant amount of time, making it look like polling was running for a long time when in fact the event handler was running for a long time. For example, on Linux the io_submit(2) syscall invoked when a virtio-blk device's virtqueue becomes ready can take 10s of microseconds. This can exceed the default polling interval (32 microseconds) and cause adaptive polling to stop polling. By excluding the handler's execution time from the polling check we make the adaptive polling calculation more accurate. As a result, the event loop now stays in polling mode where previously it would have fallen back to file descriptor monitoring. The following data was collected with virtio-blk num-queues=2 event_idx=off using an IOThread. Before: 168k IOPS, IOThread syscalls: 9837.115 ( 0.020 ms): IO iothread1/620155 io_submit(ctx_id: 140512552468480, nr: 16, iocbpp: 0x7fcb9f937db0) = 16 9837.158 ( 0.002 ms): IO iothread1/620155 write(fd: 103, buf: 0x556a2ef71b88, count: 8) = 8 9837.161 ( 0.001 ms): IO iothread1/620155 write(fd: 104, buf: 0x556a2ef71b88, count: 8) = 8 9837.163 ( 0.001 ms): IO iothread1/620155 ppoll(ufds: 0x7fcb90002800, nfds: 4, tsp: 0x7fcb9f1342d0, sigsetsize: 8) = 3 9837.164 ( 0.001 ms): IO iothread1/620155 read(fd: 107, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.174 ( 0.001 ms): IO iothread1/620155 read(fd: 105, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.176 ( 0.001 ms): IO iothread1/620155 read(fd: 106, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.209 ( 0.035 ms): IO iothread1/620155 io_submit(ctx_id: 140512552468480, nr: 32, iocbpp: 0x7fca7d0cebe0) = 32 174k IOPS (+3.6%), IOThread syscalls: 9809.566 ( 0.036 ms): IO iothread1/623061 io_submit(ctx_id: 140539805028352, nr: 32, iocbpp: 0x7fd0cdd62be0) = 32 9809.625 ( 0.001 ms): IO iothread1/623061 write(fd: 103, buf: 0x5647cfba5f58, count: 8) = 8 9809.627 ( 0.002 ms): IO iothread1/623061 write(fd: 104, buf: 0x5647cfba5f58, count: 8) = 8 9809.663 ( 0.036 ms): IO iothread1/623061 io_submit(ctx_id: 140539805028352, nr: 32, iocbpp: 0x7fd0d0388b50) = 32 Notice that ppoll(2) and eventfd read(2) syscalls are eliminated because the IOThread stays in polling mode instead of falling back to file descriptor monitoring. As usual, polling is not implemented on Windows so this patch ignores the new io_poll_read() callback in aio-win32.c. Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> Reviewed-by: Stefano Garzarella <sgarzare@redhat.com> Message-id: 20211207132336.36627-2-stefanha@redhat.com [Fixed up aio_set_event_notifier() calls in tests/unit/test-fdmon-epoll.c added after this series was queued. --Stefan] Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
2021-12-07 16:23:31 +03:00
poll_set_started(ctx, &ready_list, false);
/* TODO what to do with this list? */
return false;
}
bool aio_pending(AioContext *ctx)
{
AioHandler *node;
bool result = false;
/*
* We have to walk very carefully in case aio_set_fd_handler is
* called while we're walking.
*/
qemu_lockcnt_inc(&ctx->list_lock);
QLIST_FOREACH_RCU(node, &ctx->aio_handlers, node) {
int revents;
/* TODO should this check poll ready? */
revents = node->pfd.revents & node->pfd.events;
if (revents & (G_IO_IN | G_IO_HUP | G_IO_ERR) && node->io_read &&
aio_node_check(ctx, node->is_external)) {
result = true;
break;
}
if (revents & (G_IO_OUT | G_IO_ERR) && node->io_write &&
aio_node_check(ctx, node->is_external)) {
result = true;
break;
}
}
qemu_lockcnt_dec(&ctx->list_lock);
return result;
}
static void aio_free_deleted_handlers(AioContext *ctx)
{
AioHandler *node;
if (QLIST_EMPTY_RCU(&ctx->deleted_aio_handlers)) {
return;
}
if (!qemu_lockcnt_dec_if_lock(&ctx->list_lock)) {
return; /* we are nested, let the parent do the freeing */
}
while ((node = QLIST_FIRST_RCU(&ctx->deleted_aio_handlers))) {
QLIST_REMOVE(node, node);
QLIST_REMOVE(node, node_deleted);
QLIST_SAFE_REMOVE(node, node_poll);
g_free(node);
}
qemu_lockcnt_inc_and_unlock(&ctx->list_lock);
}
static bool aio_dispatch_handler(AioContext *ctx, AioHandler *node)
{
bool progress = false;
bool poll_ready;
int revents;
revents = node->pfd.revents & node->pfd.events;
node->pfd.revents = 0;
poll_ready = node->poll_ready;
node->poll_ready = false;
/*
* Start polling AioHandlers when they become ready because activity is
* likely to continue. Note that starvation is theoretically possible when
* fdmon_supports_polling(), but only until the fd fires for the first
* time.
*/
if (!QLIST_IS_INSERTED(node, node_deleted) &&
!QLIST_IS_INSERTED(node, node_poll) &&
node->io_poll) {
trace_poll_add(ctx, node, node->pfd.fd, revents);
if (ctx->poll_started && node->io_poll_begin) {
node->io_poll_begin(node->opaque);
}
QLIST_INSERT_HEAD(&ctx->poll_aio_handlers, node, node_poll);
}
aio-posix: split poll check from ready handler Adaptive polling measures the execution time of the polling check plus handlers called when a polled event becomes ready. Handlers can take a significant amount of time, making it look like polling was running for a long time when in fact the event handler was running for a long time. For example, on Linux the io_submit(2) syscall invoked when a virtio-blk device's virtqueue becomes ready can take 10s of microseconds. This can exceed the default polling interval (32 microseconds) and cause adaptive polling to stop polling. By excluding the handler's execution time from the polling check we make the adaptive polling calculation more accurate. As a result, the event loop now stays in polling mode where previously it would have fallen back to file descriptor monitoring. The following data was collected with virtio-blk num-queues=2 event_idx=off using an IOThread. Before: 168k IOPS, IOThread syscalls: 9837.115 ( 0.020 ms): IO iothread1/620155 io_submit(ctx_id: 140512552468480, nr: 16, iocbpp: 0x7fcb9f937db0) = 16 9837.158 ( 0.002 ms): IO iothread1/620155 write(fd: 103, buf: 0x556a2ef71b88, count: 8) = 8 9837.161 ( 0.001 ms): IO iothread1/620155 write(fd: 104, buf: 0x556a2ef71b88, count: 8) = 8 9837.163 ( 0.001 ms): IO iothread1/620155 ppoll(ufds: 0x7fcb90002800, nfds: 4, tsp: 0x7fcb9f1342d0, sigsetsize: 8) = 3 9837.164 ( 0.001 ms): IO iothread1/620155 read(fd: 107, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.174 ( 0.001 ms): IO iothread1/620155 read(fd: 105, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.176 ( 0.001 ms): IO iothread1/620155 read(fd: 106, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.209 ( 0.035 ms): IO iothread1/620155 io_submit(ctx_id: 140512552468480, nr: 32, iocbpp: 0x7fca7d0cebe0) = 32 174k IOPS (+3.6%), IOThread syscalls: 9809.566 ( 0.036 ms): IO iothread1/623061 io_submit(ctx_id: 140539805028352, nr: 32, iocbpp: 0x7fd0cdd62be0) = 32 9809.625 ( 0.001 ms): IO iothread1/623061 write(fd: 103, buf: 0x5647cfba5f58, count: 8) = 8 9809.627 ( 0.002 ms): IO iothread1/623061 write(fd: 104, buf: 0x5647cfba5f58, count: 8) = 8 9809.663 ( 0.036 ms): IO iothread1/623061 io_submit(ctx_id: 140539805028352, nr: 32, iocbpp: 0x7fd0d0388b50) = 32 Notice that ppoll(2) and eventfd read(2) syscalls are eliminated because the IOThread stays in polling mode instead of falling back to file descriptor monitoring. As usual, polling is not implemented on Windows so this patch ignores the new io_poll_read() callback in aio-win32.c. Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> Reviewed-by: Stefano Garzarella <sgarzare@redhat.com> Message-id: 20211207132336.36627-2-stefanha@redhat.com [Fixed up aio_set_event_notifier() calls in tests/unit/test-fdmon-epoll.c added after this series was queued. --Stefan] Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
2021-12-07 16:23:31 +03:00
if (!QLIST_IS_INSERTED(node, node_deleted) &&
poll_ready && revents == 0 &&
aio-posix: split poll check from ready handler Adaptive polling measures the execution time of the polling check plus handlers called when a polled event becomes ready. Handlers can take a significant amount of time, making it look like polling was running for a long time when in fact the event handler was running for a long time. For example, on Linux the io_submit(2) syscall invoked when a virtio-blk device's virtqueue becomes ready can take 10s of microseconds. This can exceed the default polling interval (32 microseconds) and cause adaptive polling to stop polling. By excluding the handler's execution time from the polling check we make the adaptive polling calculation more accurate. As a result, the event loop now stays in polling mode where previously it would have fallen back to file descriptor monitoring. The following data was collected with virtio-blk num-queues=2 event_idx=off using an IOThread. Before: 168k IOPS, IOThread syscalls: 9837.115 ( 0.020 ms): IO iothread1/620155 io_submit(ctx_id: 140512552468480, nr: 16, iocbpp: 0x7fcb9f937db0) = 16 9837.158 ( 0.002 ms): IO iothread1/620155 write(fd: 103, buf: 0x556a2ef71b88, count: 8) = 8 9837.161 ( 0.001 ms): IO iothread1/620155 write(fd: 104, buf: 0x556a2ef71b88, count: 8) = 8 9837.163 ( 0.001 ms): IO iothread1/620155 ppoll(ufds: 0x7fcb90002800, nfds: 4, tsp: 0x7fcb9f1342d0, sigsetsize: 8) = 3 9837.164 ( 0.001 ms): IO iothread1/620155 read(fd: 107, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.174 ( 0.001 ms): IO iothread1/620155 read(fd: 105, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.176 ( 0.001 ms): IO iothread1/620155 read(fd: 106, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.209 ( 0.035 ms): IO iothread1/620155 io_submit(ctx_id: 140512552468480, nr: 32, iocbpp: 0x7fca7d0cebe0) = 32 174k IOPS (+3.6%), IOThread syscalls: 9809.566 ( 0.036 ms): IO iothread1/623061 io_submit(ctx_id: 140539805028352, nr: 32, iocbpp: 0x7fd0cdd62be0) = 32 9809.625 ( 0.001 ms): IO iothread1/623061 write(fd: 103, buf: 0x5647cfba5f58, count: 8) = 8 9809.627 ( 0.002 ms): IO iothread1/623061 write(fd: 104, buf: 0x5647cfba5f58, count: 8) = 8 9809.663 ( 0.036 ms): IO iothread1/623061 io_submit(ctx_id: 140539805028352, nr: 32, iocbpp: 0x7fd0d0388b50) = 32 Notice that ppoll(2) and eventfd read(2) syscalls are eliminated because the IOThread stays in polling mode instead of falling back to file descriptor monitoring. As usual, polling is not implemented on Windows so this patch ignores the new io_poll_read() callback in aio-win32.c. Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> Reviewed-by: Stefano Garzarella <sgarzare@redhat.com> Message-id: 20211207132336.36627-2-stefanha@redhat.com [Fixed up aio_set_event_notifier() calls in tests/unit/test-fdmon-epoll.c added after this series was queued. --Stefan] Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
2021-12-07 16:23:31 +03:00
aio_node_check(ctx, node->is_external) &&
node->io_poll_ready) {
node->io_poll_ready(node->opaque);
/*
* Return early since revents was zero. aio_notify() does not count as
* progress.
*/
return node->opaque != &ctx->notifier;
}
if (!QLIST_IS_INSERTED(node, node_deleted) &&
(revents & (G_IO_IN | G_IO_HUP | G_IO_ERR)) &&
aio_node_check(ctx, node->is_external) &&
node->io_read) {
node->io_read(node->opaque);
/* aio_notify() does not count as progress */
if (node->opaque != &ctx->notifier) {
progress = true;
}
}
if (!QLIST_IS_INSERTED(node, node_deleted) &&
(revents & (G_IO_OUT | G_IO_ERR)) &&
aio_node_check(ctx, node->is_external) &&
node->io_write) {
node->io_write(node->opaque);
progress = true;
}
return progress;
}
/*
* If we have a list of ready handlers then this is more efficient than
* scanning all handlers with aio_dispatch_handlers().
*/
static bool aio_dispatch_ready_handlers(AioContext *ctx,
AioHandlerList *ready_list)
{
bool progress = false;
AioHandler *node;
while ((node = QLIST_FIRST(ready_list))) {
QLIST_REMOVE(node, node_ready);
progress = aio_dispatch_handler(ctx, node) || progress;
}
return progress;
}
/* Slower than aio_dispatch_ready_handlers() but only used via glib */
static bool aio_dispatch_handlers(AioContext *ctx)
{
AioHandler *node, *tmp;
bool progress = false;
QLIST_FOREACH_SAFE_RCU(node, &ctx->aio_handlers, node, tmp) {
progress = aio_dispatch_handler(ctx, node) || progress;
}
return progress;
}
void aio_dispatch(AioContext *ctx)
{
qemu_lockcnt_inc(&ctx->list_lock);
aio_bh_poll(ctx);
aio_dispatch_handlers(ctx);
aio_free_deleted_handlers(ctx);
qemu_lockcnt_dec(&ctx->list_lock);
timerlistgroup_run_timers(&ctx->tlg);
}
static bool run_poll_handlers_once(AioContext *ctx,
aio-posix: split poll check from ready handler Adaptive polling measures the execution time of the polling check plus handlers called when a polled event becomes ready. Handlers can take a significant amount of time, making it look like polling was running for a long time when in fact the event handler was running for a long time. For example, on Linux the io_submit(2) syscall invoked when a virtio-blk device's virtqueue becomes ready can take 10s of microseconds. This can exceed the default polling interval (32 microseconds) and cause adaptive polling to stop polling. By excluding the handler's execution time from the polling check we make the adaptive polling calculation more accurate. As a result, the event loop now stays in polling mode where previously it would have fallen back to file descriptor monitoring. The following data was collected with virtio-blk num-queues=2 event_idx=off using an IOThread. Before: 168k IOPS, IOThread syscalls: 9837.115 ( 0.020 ms): IO iothread1/620155 io_submit(ctx_id: 140512552468480, nr: 16, iocbpp: 0x7fcb9f937db0) = 16 9837.158 ( 0.002 ms): IO iothread1/620155 write(fd: 103, buf: 0x556a2ef71b88, count: 8) = 8 9837.161 ( 0.001 ms): IO iothread1/620155 write(fd: 104, buf: 0x556a2ef71b88, count: 8) = 8 9837.163 ( 0.001 ms): IO iothread1/620155 ppoll(ufds: 0x7fcb90002800, nfds: 4, tsp: 0x7fcb9f1342d0, sigsetsize: 8) = 3 9837.164 ( 0.001 ms): IO iothread1/620155 read(fd: 107, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.174 ( 0.001 ms): IO iothread1/620155 read(fd: 105, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.176 ( 0.001 ms): IO iothread1/620155 read(fd: 106, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.209 ( 0.035 ms): IO iothread1/620155 io_submit(ctx_id: 140512552468480, nr: 32, iocbpp: 0x7fca7d0cebe0) = 32 174k IOPS (+3.6%), IOThread syscalls: 9809.566 ( 0.036 ms): IO iothread1/623061 io_submit(ctx_id: 140539805028352, nr: 32, iocbpp: 0x7fd0cdd62be0) = 32 9809.625 ( 0.001 ms): IO iothread1/623061 write(fd: 103, buf: 0x5647cfba5f58, count: 8) = 8 9809.627 ( 0.002 ms): IO iothread1/623061 write(fd: 104, buf: 0x5647cfba5f58, count: 8) = 8 9809.663 ( 0.036 ms): IO iothread1/623061 io_submit(ctx_id: 140539805028352, nr: 32, iocbpp: 0x7fd0d0388b50) = 32 Notice that ppoll(2) and eventfd read(2) syscalls are eliminated because the IOThread stays in polling mode instead of falling back to file descriptor monitoring. As usual, polling is not implemented on Windows so this patch ignores the new io_poll_read() callback in aio-win32.c. Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> Reviewed-by: Stefano Garzarella <sgarzare@redhat.com> Message-id: 20211207132336.36627-2-stefanha@redhat.com [Fixed up aio_set_event_notifier() calls in tests/unit/test-fdmon-epoll.c added after this series was queued. --Stefan] Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
2021-12-07 16:23:31 +03:00
AioHandlerList *ready_list,
int64_t now,
int64_t *timeout)
{
bool progress = false;
AioHandler *node;
AioHandler *tmp;
QLIST_FOREACH_SAFE(node, &ctx->poll_aio_handlers, node_poll, tmp) {
if (aio_node_check(ctx, node->is_external) &&
node->io_poll(node->opaque)) {
aio_add_poll_ready_handler(ready_list, node);
aio-posix: split poll check from ready handler Adaptive polling measures the execution time of the polling check plus handlers called when a polled event becomes ready. Handlers can take a significant amount of time, making it look like polling was running for a long time when in fact the event handler was running for a long time. For example, on Linux the io_submit(2) syscall invoked when a virtio-blk device's virtqueue becomes ready can take 10s of microseconds. This can exceed the default polling interval (32 microseconds) and cause adaptive polling to stop polling. By excluding the handler's execution time from the polling check we make the adaptive polling calculation more accurate. As a result, the event loop now stays in polling mode where previously it would have fallen back to file descriptor monitoring. The following data was collected with virtio-blk num-queues=2 event_idx=off using an IOThread. Before: 168k IOPS, IOThread syscalls: 9837.115 ( 0.020 ms): IO iothread1/620155 io_submit(ctx_id: 140512552468480, nr: 16, iocbpp: 0x7fcb9f937db0) = 16 9837.158 ( 0.002 ms): IO iothread1/620155 write(fd: 103, buf: 0x556a2ef71b88, count: 8) = 8 9837.161 ( 0.001 ms): IO iothread1/620155 write(fd: 104, buf: 0x556a2ef71b88, count: 8) = 8 9837.163 ( 0.001 ms): IO iothread1/620155 ppoll(ufds: 0x7fcb90002800, nfds: 4, tsp: 0x7fcb9f1342d0, sigsetsize: 8) = 3 9837.164 ( 0.001 ms): IO iothread1/620155 read(fd: 107, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.174 ( 0.001 ms): IO iothread1/620155 read(fd: 105, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.176 ( 0.001 ms): IO iothread1/620155 read(fd: 106, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.209 ( 0.035 ms): IO iothread1/620155 io_submit(ctx_id: 140512552468480, nr: 32, iocbpp: 0x7fca7d0cebe0) = 32 174k IOPS (+3.6%), IOThread syscalls: 9809.566 ( 0.036 ms): IO iothread1/623061 io_submit(ctx_id: 140539805028352, nr: 32, iocbpp: 0x7fd0cdd62be0) = 32 9809.625 ( 0.001 ms): IO iothread1/623061 write(fd: 103, buf: 0x5647cfba5f58, count: 8) = 8 9809.627 ( 0.002 ms): IO iothread1/623061 write(fd: 104, buf: 0x5647cfba5f58, count: 8) = 8 9809.663 ( 0.036 ms): IO iothread1/623061 io_submit(ctx_id: 140539805028352, nr: 32, iocbpp: 0x7fd0d0388b50) = 32 Notice that ppoll(2) and eventfd read(2) syscalls are eliminated because the IOThread stays in polling mode instead of falling back to file descriptor monitoring. As usual, polling is not implemented on Windows so this patch ignores the new io_poll_read() callback in aio-win32.c. Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> Reviewed-by: Stefano Garzarella <sgarzare@redhat.com> Message-id: 20211207132336.36627-2-stefanha@redhat.com [Fixed up aio_set_event_notifier() calls in tests/unit/test-fdmon-epoll.c added after this series was queued. --Stefan] Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
2021-12-07 16:23:31 +03:00
node->poll_idle_timeout = now + POLL_IDLE_INTERVAL_NS;
/*
* Polling was successful, exit try_poll_mode immediately
* to adjust the next polling time.
*/
*timeout = 0;
if (node->opaque != &ctx->notifier) {
progress = true;
}
}
/* Caller handles freeing deleted nodes. Don't do it here. */
}
return progress;
}
static bool fdmon_supports_polling(AioContext *ctx)
{
return ctx->fdmon_ops->need_wait != aio_poll_disabled;
}
aio-posix: split poll check from ready handler Adaptive polling measures the execution time of the polling check plus handlers called when a polled event becomes ready. Handlers can take a significant amount of time, making it look like polling was running for a long time when in fact the event handler was running for a long time. For example, on Linux the io_submit(2) syscall invoked when a virtio-blk device's virtqueue becomes ready can take 10s of microseconds. This can exceed the default polling interval (32 microseconds) and cause adaptive polling to stop polling. By excluding the handler's execution time from the polling check we make the adaptive polling calculation more accurate. As a result, the event loop now stays in polling mode where previously it would have fallen back to file descriptor monitoring. The following data was collected with virtio-blk num-queues=2 event_idx=off using an IOThread. Before: 168k IOPS, IOThread syscalls: 9837.115 ( 0.020 ms): IO iothread1/620155 io_submit(ctx_id: 140512552468480, nr: 16, iocbpp: 0x7fcb9f937db0) = 16 9837.158 ( 0.002 ms): IO iothread1/620155 write(fd: 103, buf: 0x556a2ef71b88, count: 8) = 8 9837.161 ( 0.001 ms): IO iothread1/620155 write(fd: 104, buf: 0x556a2ef71b88, count: 8) = 8 9837.163 ( 0.001 ms): IO iothread1/620155 ppoll(ufds: 0x7fcb90002800, nfds: 4, tsp: 0x7fcb9f1342d0, sigsetsize: 8) = 3 9837.164 ( 0.001 ms): IO iothread1/620155 read(fd: 107, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.174 ( 0.001 ms): IO iothread1/620155 read(fd: 105, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.176 ( 0.001 ms): IO iothread1/620155 read(fd: 106, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.209 ( 0.035 ms): IO iothread1/620155 io_submit(ctx_id: 140512552468480, nr: 32, iocbpp: 0x7fca7d0cebe0) = 32 174k IOPS (+3.6%), IOThread syscalls: 9809.566 ( 0.036 ms): IO iothread1/623061 io_submit(ctx_id: 140539805028352, nr: 32, iocbpp: 0x7fd0cdd62be0) = 32 9809.625 ( 0.001 ms): IO iothread1/623061 write(fd: 103, buf: 0x5647cfba5f58, count: 8) = 8 9809.627 ( 0.002 ms): IO iothread1/623061 write(fd: 104, buf: 0x5647cfba5f58, count: 8) = 8 9809.663 ( 0.036 ms): IO iothread1/623061 io_submit(ctx_id: 140539805028352, nr: 32, iocbpp: 0x7fd0d0388b50) = 32 Notice that ppoll(2) and eventfd read(2) syscalls are eliminated because the IOThread stays in polling mode instead of falling back to file descriptor monitoring. As usual, polling is not implemented on Windows so this patch ignores the new io_poll_read() callback in aio-win32.c. Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> Reviewed-by: Stefano Garzarella <sgarzare@redhat.com> Message-id: 20211207132336.36627-2-stefanha@redhat.com [Fixed up aio_set_event_notifier() calls in tests/unit/test-fdmon-epoll.c added after this series was queued. --Stefan] Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
2021-12-07 16:23:31 +03:00
static bool remove_idle_poll_handlers(AioContext *ctx,
AioHandlerList *ready_list,
int64_t now)
{
AioHandler *node;
AioHandler *tmp;
bool progress = false;
/*
* File descriptor monitoring implementations without userspace polling
* support suffer from starvation when a subset of handlers is polled
* because fds will not be processed in a timely fashion. Don't remove
* idle poll handlers.
*/
if (!fdmon_supports_polling(ctx)) {
return false;
}
QLIST_FOREACH_SAFE(node, &ctx->poll_aio_handlers, node_poll, tmp) {
if (node->poll_idle_timeout == 0LL) {
node->poll_idle_timeout = now + POLL_IDLE_INTERVAL_NS;
} else if (now >= node->poll_idle_timeout) {
trace_poll_remove(ctx, node, node->pfd.fd);
node->poll_idle_timeout = 0LL;
QLIST_SAFE_REMOVE(node, node_poll);
if (ctx->poll_started && node->io_poll_end) {
node->io_poll_end(node->opaque);
/*
* Final poll in case ->io_poll_end() races with an event.
* Nevermind about re-adding the handler in the rare case where
* this causes progress.
*/
aio-posix: split poll check from ready handler Adaptive polling measures the execution time of the polling check plus handlers called when a polled event becomes ready. Handlers can take a significant amount of time, making it look like polling was running for a long time when in fact the event handler was running for a long time. For example, on Linux the io_submit(2) syscall invoked when a virtio-blk device's virtqueue becomes ready can take 10s of microseconds. This can exceed the default polling interval (32 microseconds) and cause adaptive polling to stop polling. By excluding the handler's execution time from the polling check we make the adaptive polling calculation more accurate. As a result, the event loop now stays in polling mode where previously it would have fallen back to file descriptor monitoring. The following data was collected with virtio-blk num-queues=2 event_idx=off using an IOThread. Before: 168k IOPS, IOThread syscalls: 9837.115 ( 0.020 ms): IO iothread1/620155 io_submit(ctx_id: 140512552468480, nr: 16, iocbpp: 0x7fcb9f937db0) = 16 9837.158 ( 0.002 ms): IO iothread1/620155 write(fd: 103, buf: 0x556a2ef71b88, count: 8) = 8 9837.161 ( 0.001 ms): IO iothread1/620155 write(fd: 104, buf: 0x556a2ef71b88, count: 8) = 8 9837.163 ( 0.001 ms): IO iothread1/620155 ppoll(ufds: 0x7fcb90002800, nfds: 4, tsp: 0x7fcb9f1342d0, sigsetsize: 8) = 3 9837.164 ( 0.001 ms): IO iothread1/620155 read(fd: 107, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.174 ( 0.001 ms): IO iothread1/620155 read(fd: 105, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.176 ( 0.001 ms): IO iothread1/620155 read(fd: 106, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.209 ( 0.035 ms): IO iothread1/620155 io_submit(ctx_id: 140512552468480, nr: 32, iocbpp: 0x7fca7d0cebe0) = 32 174k IOPS (+3.6%), IOThread syscalls: 9809.566 ( 0.036 ms): IO iothread1/623061 io_submit(ctx_id: 140539805028352, nr: 32, iocbpp: 0x7fd0cdd62be0) = 32 9809.625 ( 0.001 ms): IO iothread1/623061 write(fd: 103, buf: 0x5647cfba5f58, count: 8) = 8 9809.627 ( 0.002 ms): IO iothread1/623061 write(fd: 104, buf: 0x5647cfba5f58, count: 8) = 8 9809.663 ( 0.036 ms): IO iothread1/623061 io_submit(ctx_id: 140539805028352, nr: 32, iocbpp: 0x7fd0d0388b50) = 32 Notice that ppoll(2) and eventfd read(2) syscalls are eliminated because the IOThread stays in polling mode instead of falling back to file descriptor monitoring. As usual, polling is not implemented on Windows so this patch ignores the new io_poll_read() callback in aio-win32.c. Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> Reviewed-by: Stefano Garzarella <sgarzare@redhat.com> Message-id: 20211207132336.36627-2-stefanha@redhat.com [Fixed up aio_set_event_notifier() calls in tests/unit/test-fdmon-epoll.c added after this series was queued. --Stefan] Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
2021-12-07 16:23:31 +03:00
if (node->io_poll(node->opaque)) {
aio_add_poll_ready_handler(ready_list, node);
aio-posix: split poll check from ready handler Adaptive polling measures the execution time of the polling check plus handlers called when a polled event becomes ready. Handlers can take a significant amount of time, making it look like polling was running for a long time when in fact the event handler was running for a long time. For example, on Linux the io_submit(2) syscall invoked when a virtio-blk device's virtqueue becomes ready can take 10s of microseconds. This can exceed the default polling interval (32 microseconds) and cause adaptive polling to stop polling. By excluding the handler's execution time from the polling check we make the adaptive polling calculation more accurate. As a result, the event loop now stays in polling mode where previously it would have fallen back to file descriptor monitoring. The following data was collected with virtio-blk num-queues=2 event_idx=off using an IOThread. Before: 168k IOPS, IOThread syscalls: 9837.115 ( 0.020 ms): IO iothread1/620155 io_submit(ctx_id: 140512552468480, nr: 16, iocbpp: 0x7fcb9f937db0) = 16 9837.158 ( 0.002 ms): IO iothread1/620155 write(fd: 103, buf: 0x556a2ef71b88, count: 8) = 8 9837.161 ( 0.001 ms): IO iothread1/620155 write(fd: 104, buf: 0x556a2ef71b88, count: 8) = 8 9837.163 ( 0.001 ms): IO iothread1/620155 ppoll(ufds: 0x7fcb90002800, nfds: 4, tsp: 0x7fcb9f1342d0, sigsetsize: 8) = 3 9837.164 ( 0.001 ms): IO iothread1/620155 read(fd: 107, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.174 ( 0.001 ms): IO iothread1/620155 read(fd: 105, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.176 ( 0.001 ms): IO iothread1/620155 read(fd: 106, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.209 ( 0.035 ms): IO iothread1/620155 io_submit(ctx_id: 140512552468480, nr: 32, iocbpp: 0x7fca7d0cebe0) = 32 174k IOPS (+3.6%), IOThread syscalls: 9809.566 ( 0.036 ms): IO iothread1/623061 io_submit(ctx_id: 140539805028352, nr: 32, iocbpp: 0x7fd0cdd62be0) = 32 9809.625 ( 0.001 ms): IO iothread1/623061 write(fd: 103, buf: 0x5647cfba5f58, count: 8) = 8 9809.627 ( 0.002 ms): IO iothread1/623061 write(fd: 104, buf: 0x5647cfba5f58, count: 8) = 8 9809.663 ( 0.036 ms): IO iothread1/623061 io_submit(ctx_id: 140539805028352, nr: 32, iocbpp: 0x7fd0d0388b50) = 32 Notice that ppoll(2) and eventfd read(2) syscalls are eliminated because the IOThread stays in polling mode instead of falling back to file descriptor monitoring. As usual, polling is not implemented on Windows so this patch ignores the new io_poll_read() callback in aio-win32.c. Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> Reviewed-by: Stefano Garzarella <sgarzare@redhat.com> Message-id: 20211207132336.36627-2-stefanha@redhat.com [Fixed up aio_set_event_notifier() calls in tests/unit/test-fdmon-epoll.c added after this series was queued. --Stefan] Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
2021-12-07 16:23:31 +03:00
progress = true;
}
}
}
}
return progress;
}
aio: add polling mode to AioContext The AioContext event loop uses ppoll(2) or epoll_wait(2) to monitor file descriptors or until a timer expires. In cases like virtqueues, Linux AIO, and ThreadPool it is technically possible to wait for events via polling (i.e. continuously checking for events without blocking). Polling can be faster than blocking syscalls because file descriptors, the process scheduler, and system calls are bypassed. The main disadvantage to polling is that it increases CPU utilization. In classic polling configuration a full host CPU thread might run at 100% to respond to events as quickly as possible. This patch implements a timeout so we fall back to blocking syscalls if polling detects no activity. After the timeout no CPU cycles are wasted on polling until the next event loop iteration. The run_poll_handlers_begin() and run_poll_handlers_end() trace events are added to aid performance analysis and troubleshooting. If you need to know whether polling mode is being used, trace these events to find out. Note that the AioContext is now re-acquired before disabling notify_me in the non-polling case. This makes the code cleaner since notify_me was enabled outside the non-polling AioContext release region. This change is correct since it's safe to keep notify_me enabled longer (disabling is an optimization) but potentially causes unnecessary event_notifer_set() calls. I think the chance of performance regression is small here. Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> Reviewed-by: Paolo Bonzini <pbonzini@redhat.com> Message-id: 20161201192652.9509-4-stefanha@redhat.com Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
2016-12-01 22:26:42 +03:00
/* run_poll_handlers:
* @ctx: the AioContext
aio-posix: split poll check from ready handler Adaptive polling measures the execution time of the polling check plus handlers called when a polled event becomes ready. Handlers can take a significant amount of time, making it look like polling was running for a long time when in fact the event handler was running for a long time. For example, on Linux the io_submit(2) syscall invoked when a virtio-blk device's virtqueue becomes ready can take 10s of microseconds. This can exceed the default polling interval (32 microseconds) and cause adaptive polling to stop polling. By excluding the handler's execution time from the polling check we make the adaptive polling calculation more accurate. As a result, the event loop now stays in polling mode where previously it would have fallen back to file descriptor monitoring. The following data was collected with virtio-blk num-queues=2 event_idx=off using an IOThread. Before: 168k IOPS, IOThread syscalls: 9837.115 ( 0.020 ms): IO iothread1/620155 io_submit(ctx_id: 140512552468480, nr: 16, iocbpp: 0x7fcb9f937db0) = 16 9837.158 ( 0.002 ms): IO iothread1/620155 write(fd: 103, buf: 0x556a2ef71b88, count: 8) = 8 9837.161 ( 0.001 ms): IO iothread1/620155 write(fd: 104, buf: 0x556a2ef71b88, count: 8) = 8 9837.163 ( 0.001 ms): IO iothread1/620155 ppoll(ufds: 0x7fcb90002800, nfds: 4, tsp: 0x7fcb9f1342d0, sigsetsize: 8) = 3 9837.164 ( 0.001 ms): IO iothread1/620155 read(fd: 107, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.174 ( 0.001 ms): IO iothread1/620155 read(fd: 105, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.176 ( 0.001 ms): IO iothread1/620155 read(fd: 106, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.209 ( 0.035 ms): IO iothread1/620155 io_submit(ctx_id: 140512552468480, nr: 32, iocbpp: 0x7fca7d0cebe0) = 32 174k IOPS (+3.6%), IOThread syscalls: 9809.566 ( 0.036 ms): IO iothread1/623061 io_submit(ctx_id: 140539805028352, nr: 32, iocbpp: 0x7fd0cdd62be0) = 32 9809.625 ( 0.001 ms): IO iothread1/623061 write(fd: 103, buf: 0x5647cfba5f58, count: 8) = 8 9809.627 ( 0.002 ms): IO iothread1/623061 write(fd: 104, buf: 0x5647cfba5f58, count: 8) = 8 9809.663 ( 0.036 ms): IO iothread1/623061 io_submit(ctx_id: 140539805028352, nr: 32, iocbpp: 0x7fd0d0388b50) = 32 Notice that ppoll(2) and eventfd read(2) syscalls are eliminated because the IOThread stays in polling mode instead of falling back to file descriptor monitoring. As usual, polling is not implemented on Windows so this patch ignores the new io_poll_read() callback in aio-win32.c. Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> Reviewed-by: Stefano Garzarella <sgarzare@redhat.com> Message-id: 20211207132336.36627-2-stefanha@redhat.com [Fixed up aio_set_event_notifier() calls in tests/unit/test-fdmon-epoll.c added after this series was queued. --Stefan] Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
2021-12-07 16:23:31 +03:00
* @ready_list: the list to place ready handlers on
aio: add polling mode to AioContext The AioContext event loop uses ppoll(2) or epoll_wait(2) to monitor file descriptors or until a timer expires. In cases like virtqueues, Linux AIO, and ThreadPool it is technically possible to wait for events via polling (i.e. continuously checking for events without blocking). Polling can be faster than blocking syscalls because file descriptors, the process scheduler, and system calls are bypassed. The main disadvantage to polling is that it increases CPU utilization. In classic polling configuration a full host CPU thread might run at 100% to respond to events as quickly as possible. This patch implements a timeout so we fall back to blocking syscalls if polling detects no activity. After the timeout no CPU cycles are wasted on polling until the next event loop iteration. The run_poll_handlers_begin() and run_poll_handlers_end() trace events are added to aid performance analysis and troubleshooting. If you need to know whether polling mode is being used, trace these events to find out. Note that the AioContext is now re-acquired before disabling notify_me in the non-polling case. This makes the code cleaner since notify_me was enabled outside the non-polling AioContext release region. This change is correct since it's safe to keep notify_me enabled longer (disabling is an optimization) but potentially causes unnecessary event_notifer_set() calls. I think the chance of performance regression is small here. Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> Reviewed-by: Paolo Bonzini <pbonzini@redhat.com> Message-id: 20161201192652.9509-4-stefanha@redhat.com Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
2016-12-01 22:26:42 +03:00
* @max_ns: maximum time to poll for, in nanoseconds
*
* Polls for a given time.
*
* Note that the caller must have incremented ctx->list_lock.
aio: add polling mode to AioContext The AioContext event loop uses ppoll(2) or epoll_wait(2) to monitor file descriptors or until a timer expires. In cases like virtqueues, Linux AIO, and ThreadPool it is technically possible to wait for events via polling (i.e. continuously checking for events without blocking). Polling can be faster than blocking syscalls because file descriptors, the process scheduler, and system calls are bypassed. The main disadvantage to polling is that it increases CPU utilization. In classic polling configuration a full host CPU thread might run at 100% to respond to events as quickly as possible. This patch implements a timeout so we fall back to blocking syscalls if polling detects no activity. After the timeout no CPU cycles are wasted on polling until the next event loop iteration. The run_poll_handlers_begin() and run_poll_handlers_end() trace events are added to aid performance analysis and troubleshooting. If you need to know whether polling mode is being used, trace these events to find out. Note that the AioContext is now re-acquired before disabling notify_me in the non-polling case. This makes the code cleaner since notify_me was enabled outside the non-polling AioContext release region. This change is correct since it's safe to keep notify_me enabled longer (disabling is an optimization) but potentially causes unnecessary event_notifer_set() calls. I think the chance of performance regression is small here. Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> Reviewed-by: Paolo Bonzini <pbonzini@redhat.com> Message-id: 20161201192652.9509-4-stefanha@redhat.com Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
2016-12-01 22:26:42 +03:00
*
* Returns: true if progress was made, false otherwise
*/
aio-posix: split poll check from ready handler Adaptive polling measures the execution time of the polling check plus handlers called when a polled event becomes ready. Handlers can take a significant amount of time, making it look like polling was running for a long time when in fact the event handler was running for a long time. For example, on Linux the io_submit(2) syscall invoked when a virtio-blk device's virtqueue becomes ready can take 10s of microseconds. This can exceed the default polling interval (32 microseconds) and cause adaptive polling to stop polling. By excluding the handler's execution time from the polling check we make the adaptive polling calculation more accurate. As a result, the event loop now stays in polling mode where previously it would have fallen back to file descriptor monitoring. The following data was collected with virtio-blk num-queues=2 event_idx=off using an IOThread. Before: 168k IOPS, IOThread syscalls: 9837.115 ( 0.020 ms): IO iothread1/620155 io_submit(ctx_id: 140512552468480, nr: 16, iocbpp: 0x7fcb9f937db0) = 16 9837.158 ( 0.002 ms): IO iothread1/620155 write(fd: 103, buf: 0x556a2ef71b88, count: 8) = 8 9837.161 ( 0.001 ms): IO iothread1/620155 write(fd: 104, buf: 0x556a2ef71b88, count: 8) = 8 9837.163 ( 0.001 ms): IO iothread1/620155 ppoll(ufds: 0x7fcb90002800, nfds: 4, tsp: 0x7fcb9f1342d0, sigsetsize: 8) = 3 9837.164 ( 0.001 ms): IO iothread1/620155 read(fd: 107, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.174 ( 0.001 ms): IO iothread1/620155 read(fd: 105, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.176 ( 0.001 ms): IO iothread1/620155 read(fd: 106, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.209 ( 0.035 ms): IO iothread1/620155 io_submit(ctx_id: 140512552468480, nr: 32, iocbpp: 0x7fca7d0cebe0) = 32 174k IOPS (+3.6%), IOThread syscalls: 9809.566 ( 0.036 ms): IO iothread1/623061 io_submit(ctx_id: 140539805028352, nr: 32, iocbpp: 0x7fd0cdd62be0) = 32 9809.625 ( 0.001 ms): IO iothread1/623061 write(fd: 103, buf: 0x5647cfba5f58, count: 8) = 8 9809.627 ( 0.002 ms): IO iothread1/623061 write(fd: 104, buf: 0x5647cfba5f58, count: 8) = 8 9809.663 ( 0.036 ms): IO iothread1/623061 io_submit(ctx_id: 140539805028352, nr: 32, iocbpp: 0x7fd0d0388b50) = 32 Notice that ppoll(2) and eventfd read(2) syscalls are eliminated because the IOThread stays in polling mode instead of falling back to file descriptor monitoring. As usual, polling is not implemented on Windows so this patch ignores the new io_poll_read() callback in aio-win32.c. Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> Reviewed-by: Stefano Garzarella <sgarzare@redhat.com> Message-id: 20211207132336.36627-2-stefanha@redhat.com [Fixed up aio_set_event_notifier() calls in tests/unit/test-fdmon-epoll.c added after this series was queued. --Stefan] Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
2021-12-07 16:23:31 +03:00
static bool run_poll_handlers(AioContext *ctx, AioHandlerList *ready_list,
int64_t max_ns, int64_t *timeout)
aio: add polling mode to AioContext The AioContext event loop uses ppoll(2) or epoll_wait(2) to monitor file descriptors or until a timer expires. In cases like virtqueues, Linux AIO, and ThreadPool it is technically possible to wait for events via polling (i.e. continuously checking for events without blocking). Polling can be faster than blocking syscalls because file descriptors, the process scheduler, and system calls are bypassed. The main disadvantage to polling is that it increases CPU utilization. In classic polling configuration a full host CPU thread might run at 100% to respond to events as quickly as possible. This patch implements a timeout so we fall back to blocking syscalls if polling detects no activity. After the timeout no CPU cycles are wasted on polling until the next event loop iteration. The run_poll_handlers_begin() and run_poll_handlers_end() trace events are added to aid performance analysis and troubleshooting. If you need to know whether polling mode is being used, trace these events to find out. Note that the AioContext is now re-acquired before disabling notify_me in the non-polling case. This makes the code cleaner since notify_me was enabled outside the non-polling AioContext release region. This change is correct since it's safe to keep notify_me enabled longer (disabling is an optimization) but potentially causes unnecessary event_notifer_set() calls. I think the chance of performance regression is small here. Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> Reviewed-by: Paolo Bonzini <pbonzini@redhat.com> Message-id: 20161201192652.9509-4-stefanha@redhat.com Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
2016-12-01 22:26:42 +03:00
{
bool progress;
int64_t start_time, elapsed_time;
aio: add polling mode to AioContext The AioContext event loop uses ppoll(2) or epoll_wait(2) to monitor file descriptors or until a timer expires. In cases like virtqueues, Linux AIO, and ThreadPool it is technically possible to wait for events via polling (i.e. continuously checking for events without blocking). Polling can be faster than blocking syscalls because file descriptors, the process scheduler, and system calls are bypassed. The main disadvantage to polling is that it increases CPU utilization. In classic polling configuration a full host CPU thread might run at 100% to respond to events as quickly as possible. This patch implements a timeout so we fall back to blocking syscalls if polling detects no activity. After the timeout no CPU cycles are wasted on polling until the next event loop iteration. The run_poll_handlers_begin() and run_poll_handlers_end() trace events are added to aid performance analysis and troubleshooting. If you need to know whether polling mode is being used, trace these events to find out. Note that the AioContext is now re-acquired before disabling notify_me in the non-polling case. This makes the code cleaner since notify_me was enabled outside the non-polling AioContext release region. This change is correct since it's safe to keep notify_me enabled longer (disabling is an optimization) but potentially causes unnecessary event_notifer_set() calls. I think the chance of performance regression is small here. Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> Reviewed-by: Paolo Bonzini <pbonzini@redhat.com> Message-id: 20161201192652.9509-4-stefanha@redhat.com Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
2016-12-01 22:26:42 +03:00
assert(qemu_lockcnt_count(&ctx->list_lock) > 0);
aio: add polling mode to AioContext The AioContext event loop uses ppoll(2) or epoll_wait(2) to monitor file descriptors or until a timer expires. In cases like virtqueues, Linux AIO, and ThreadPool it is technically possible to wait for events via polling (i.e. continuously checking for events without blocking). Polling can be faster than blocking syscalls because file descriptors, the process scheduler, and system calls are bypassed. The main disadvantage to polling is that it increases CPU utilization. In classic polling configuration a full host CPU thread might run at 100% to respond to events as quickly as possible. This patch implements a timeout so we fall back to blocking syscalls if polling detects no activity. After the timeout no CPU cycles are wasted on polling until the next event loop iteration. The run_poll_handlers_begin() and run_poll_handlers_end() trace events are added to aid performance analysis and troubleshooting. If you need to know whether polling mode is being used, trace these events to find out. Note that the AioContext is now re-acquired before disabling notify_me in the non-polling case. This makes the code cleaner since notify_me was enabled outside the non-polling AioContext release region. This change is correct since it's safe to keep notify_me enabled longer (disabling is an optimization) but potentially causes unnecessary event_notifer_set() calls. I think the chance of performance regression is small here. Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> Reviewed-by: Paolo Bonzini <pbonzini@redhat.com> Message-id: 20161201192652.9509-4-stefanha@redhat.com Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
2016-12-01 22:26:42 +03:00
trace_run_poll_handlers_begin(ctx, max_ns, *timeout);
aio: add polling mode to AioContext The AioContext event loop uses ppoll(2) or epoll_wait(2) to monitor file descriptors or until a timer expires. In cases like virtqueues, Linux AIO, and ThreadPool it is technically possible to wait for events via polling (i.e. continuously checking for events without blocking). Polling can be faster than blocking syscalls because file descriptors, the process scheduler, and system calls are bypassed. The main disadvantage to polling is that it increases CPU utilization. In classic polling configuration a full host CPU thread might run at 100% to respond to events as quickly as possible. This patch implements a timeout so we fall back to blocking syscalls if polling detects no activity. After the timeout no CPU cycles are wasted on polling until the next event loop iteration. The run_poll_handlers_begin() and run_poll_handlers_end() trace events are added to aid performance analysis and troubleshooting. If you need to know whether polling mode is being used, trace these events to find out. Note that the AioContext is now re-acquired before disabling notify_me in the non-polling case. This makes the code cleaner since notify_me was enabled outside the non-polling AioContext release region. This change is correct since it's safe to keep notify_me enabled longer (disabling is an optimization) but potentially causes unnecessary event_notifer_set() calls. I think the chance of performance regression is small here. Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> Reviewed-by: Paolo Bonzini <pbonzini@redhat.com> Message-id: 20161201192652.9509-4-stefanha@redhat.com Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
2016-12-01 22:26:42 +03:00
/*
* Optimization: ->io_poll() handlers often contain RCU read critical
* sections and we therefore see many rcu_read_lock() -> rcu_read_unlock()
* -> rcu_read_lock() -> ... sequences with expensive memory
* synchronization primitives. Make the entire polling loop an RCU
* critical section because nested rcu_read_lock()/rcu_read_unlock() calls
* are cheap.
*/
RCU_READ_LOCK_GUARD();
start_time = qemu_clock_get_ns(QEMU_CLOCK_REALTIME);
aio: add polling mode to AioContext The AioContext event loop uses ppoll(2) or epoll_wait(2) to monitor file descriptors or until a timer expires. In cases like virtqueues, Linux AIO, and ThreadPool it is technically possible to wait for events via polling (i.e. continuously checking for events without blocking). Polling can be faster than blocking syscalls because file descriptors, the process scheduler, and system calls are bypassed. The main disadvantage to polling is that it increases CPU utilization. In classic polling configuration a full host CPU thread might run at 100% to respond to events as quickly as possible. This patch implements a timeout so we fall back to blocking syscalls if polling detects no activity. After the timeout no CPU cycles are wasted on polling until the next event loop iteration. The run_poll_handlers_begin() and run_poll_handlers_end() trace events are added to aid performance analysis and troubleshooting. If you need to know whether polling mode is being used, trace these events to find out. Note that the AioContext is now re-acquired before disabling notify_me in the non-polling case. This makes the code cleaner since notify_me was enabled outside the non-polling AioContext release region. This change is correct since it's safe to keep notify_me enabled longer (disabling is an optimization) but potentially causes unnecessary event_notifer_set() calls. I think the chance of performance regression is small here. Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> Reviewed-by: Paolo Bonzini <pbonzini@redhat.com> Message-id: 20161201192652.9509-4-stefanha@redhat.com Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
2016-12-01 22:26:42 +03:00
do {
aio-posix: split poll check from ready handler Adaptive polling measures the execution time of the polling check plus handlers called when a polled event becomes ready. Handlers can take a significant amount of time, making it look like polling was running for a long time when in fact the event handler was running for a long time. For example, on Linux the io_submit(2) syscall invoked when a virtio-blk device's virtqueue becomes ready can take 10s of microseconds. This can exceed the default polling interval (32 microseconds) and cause adaptive polling to stop polling. By excluding the handler's execution time from the polling check we make the adaptive polling calculation more accurate. As a result, the event loop now stays in polling mode where previously it would have fallen back to file descriptor monitoring. The following data was collected with virtio-blk num-queues=2 event_idx=off using an IOThread. Before: 168k IOPS, IOThread syscalls: 9837.115 ( 0.020 ms): IO iothread1/620155 io_submit(ctx_id: 140512552468480, nr: 16, iocbpp: 0x7fcb9f937db0) = 16 9837.158 ( 0.002 ms): IO iothread1/620155 write(fd: 103, buf: 0x556a2ef71b88, count: 8) = 8 9837.161 ( 0.001 ms): IO iothread1/620155 write(fd: 104, buf: 0x556a2ef71b88, count: 8) = 8 9837.163 ( 0.001 ms): IO iothread1/620155 ppoll(ufds: 0x7fcb90002800, nfds: 4, tsp: 0x7fcb9f1342d0, sigsetsize: 8) = 3 9837.164 ( 0.001 ms): IO iothread1/620155 read(fd: 107, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.174 ( 0.001 ms): IO iothread1/620155 read(fd: 105, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.176 ( 0.001 ms): IO iothread1/620155 read(fd: 106, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.209 ( 0.035 ms): IO iothread1/620155 io_submit(ctx_id: 140512552468480, nr: 32, iocbpp: 0x7fca7d0cebe0) = 32 174k IOPS (+3.6%), IOThread syscalls: 9809.566 ( 0.036 ms): IO iothread1/623061 io_submit(ctx_id: 140539805028352, nr: 32, iocbpp: 0x7fd0cdd62be0) = 32 9809.625 ( 0.001 ms): IO iothread1/623061 write(fd: 103, buf: 0x5647cfba5f58, count: 8) = 8 9809.627 ( 0.002 ms): IO iothread1/623061 write(fd: 104, buf: 0x5647cfba5f58, count: 8) = 8 9809.663 ( 0.036 ms): IO iothread1/623061 io_submit(ctx_id: 140539805028352, nr: 32, iocbpp: 0x7fd0d0388b50) = 32 Notice that ppoll(2) and eventfd read(2) syscalls are eliminated because the IOThread stays in polling mode instead of falling back to file descriptor monitoring. As usual, polling is not implemented on Windows so this patch ignores the new io_poll_read() callback in aio-win32.c. Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> Reviewed-by: Stefano Garzarella <sgarzare@redhat.com> Message-id: 20211207132336.36627-2-stefanha@redhat.com [Fixed up aio_set_event_notifier() calls in tests/unit/test-fdmon-epoll.c added after this series was queued. --Stefan] Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
2021-12-07 16:23:31 +03:00
progress = run_poll_handlers_once(ctx, ready_list,
start_time, timeout);
elapsed_time = qemu_clock_get_ns(QEMU_CLOCK_REALTIME) - start_time;
max_ns = qemu_soonest_timeout(*timeout, max_ns);
assert(!(max_ns && progress));
} while (elapsed_time < max_ns && !ctx->fdmon_ops->need_wait(ctx));
aio: add polling mode to AioContext The AioContext event loop uses ppoll(2) or epoll_wait(2) to monitor file descriptors or until a timer expires. In cases like virtqueues, Linux AIO, and ThreadPool it is technically possible to wait for events via polling (i.e. continuously checking for events without blocking). Polling can be faster than blocking syscalls because file descriptors, the process scheduler, and system calls are bypassed. The main disadvantage to polling is that it increases CPU utilization. In classic polling configuration a full host CPU thread might run at 100% to respond to events as quickly as possible. This patch implements a timeout so we fall back to blocking syscalls if polling detects no activity. After the timeout no CPU cycles are wasted on polling until the next event loop iteration. The run_poll_handlers_begin() and run_poll_handlers_end() trace events are added to aid performance analysis and troubleshooting. If you need to know whether polling mode is being used, trace these events to find out. Note that the AioContext is now re-acquired before disabling notify_me in the non-polling case. This makes the code cleaner since notify_me was enabled outside the non-polling AioContext release region. This change is correct since it's safe to keep notify_me enabled longer (disabling is an optimization) but potentially causes unnecessary event_notifer_set() calls. I think the chance of performance regression is small here. Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> Reviewed-by: Paolo Bonzini <pbonzini@redhat.com> Message-id: 20161201192652.9509-4-stefanha@redhat.com Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
2016-12-01 22:26:42 +03:00
aio-posix: split poll check from ready handler Adaptive polling measures the execution time of the polling check plus handlers called when a polled event becomes ready. Handlers can take a significant amount of time, making it look like polling was running for a long time when in fact the event handler was running for a long time. For example, on Linux the io_submit(2) syscall invoked when a virtio-blk device's virtqueue becomes ready can take 10s of microseconds. This can exceed the default polling interval (32 microseconds) and cause adaptive polling to stop polling. By excluding the handler's execution time from the polling check we make the adaptive polling calculation more accurate. As a result, the event loop now stays in polling mode where previously it would have fallen back to file descriptor monitoring. The following data was collected with virtio-blk num-queues=2 event_idx=off using an IOThread. Before: 168k IOPS, IOThread syscalls: 9837.115 ( 0.020 ms): IO iothread1/620155 io_submit(ctx_id: 140512552468480, nr: 16, iocbpp: 0x7fcb9f937db0) = 16 9837.158 ( 0.002 ms): IO iothread1/620155 write(fd: 103, buf: 0x556a2ef71b88, count: 8) = 8 9837.161 ( 0.001 ms): IO iothread1/620155 write(fd: 104, buf: 0x556a2ef71b88, count: 8) = 8 9837.163 ( 0.001 ms): IO iothread1/620155 ppoll(ufds: 0x7fcb90002800, nfds: 4, tsp: 0x7fcb9f1342d0, sigsetsize: 8) = 3 9837.164 ( 0.001 ms): IO iothread1/620155 read(fd: 107, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.174 ( 0.001 ms): IO iothread1/620155 read(fd: 105, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.176 ( 0.001 ms): IO iothread1/620155 read(fd: 106, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.209 ( 0.035 ms): IO iothread1/620155 io_submit(ctx_id: 140512552468480, nr: 32, iocbpp: 0x7fca7d0cebe0) = 32 174k IOPS (+3.6%), IOThread syscalls: 9809.566 ( 0.036 ms): IO iothread1/623061 io_submit(ctx_id: 140539805028352, nr: 32, iocbpp: 0x7fd0cdd62be0) = 32 9809.625 ( 0.001 ms): IO iothread1/623061 write(fd: 103, buf: 0x5647cfba5f58, count: 8) = 8 9809.627 ( 0.002 ms): IO iothread1/623061 write(fd: 104, buf: 0x5647cfba5f58, count: 8) = 8 9809.663 ( 0.036 ms): IO iothread1/623061 io_submit(ctx_id: 140539805028352, nr: 32, iocbpp: 0x7fd0d0388b50) = 32 Notice that ppoll(2) and eventfd read(2) syscalls are eliminated because the IOThread stays in polling mode instead of falling back to file descriptor monitoring. As usual, polling is not implemented on Windows so this patch ignores the new io_poll_read() callback in aio-win32.c. Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> Reviewed-by: Stefano Garzarella <sgarzare@redhat.com> Message-id: 20211207132336.36627-2-stefanha@redhat.com [Fixed up aio_set_event_notifier() calls in tests/unit/test-fdmon-epoll.c added after this series was queued. --Stefan] Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
2021-12-07 16:23:31 +03:00
if (remove_idle_poll_handlers(ctx, ready_list,
start_time + elapsed_time)) {
*timeout = 0;
progress = true;
}
/* If time has passed with no successful polling, adjust *timeout to
* keep the same ending time.
*/
if (*timeout != -1) {
*timeout -= MIN(*timeout, elapsed_time);
}
aio: add polling mode to AioContext The AioContext event loop uses ppoll(2) or epoll_wait(2) to monitor file descriptors or until a timer expires. In cases like virtqueues, Linux AIO, and ThreadPool it is technically possible to wait for events via polling (i.e. continuously checking for events without blocking). Polling can be faster than blocking syscalls because file descriptors, the process scheduler, and system calls are bypassed. The main disadvantage to polling is that it increases CPU utilization. In classic polling configuration a full host CPU thread might run at 100% to respond to events as quickly as possible. This patch implements a timeout so we fall back to blocking syscalls if polling detects no activity. After the timeout no CPU cycles are wasted on polling until the next event loop iteration. The run_poll_handlers_begin() and run_poll_handlers_end() trace events are added to aid performance analysis and troubleshooting. If you need to know whether polling mode is being used, trace these events to find out. Note that the AioContext is now re-acquired before disabling notify_me in the non-polling case. This makes the code cleaner since notify_me was enabled outside the non-polling AioContext release region. This change is correct since it's safe to keep notify_me enabled longer (disabling is an optimization) but potentially causes unnecessary event_notifer_set() calls. I think the chance of performance regression is small here. Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> Reviewed-by: Paolo Bonzini <pbonzini@redhat.com> Message-id: 20161201192652.9509-4-stefanha@redhat.com Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
2016-12-01 22:26:42 +03:00
trace_run_poll_handlers_end(ctx, progress, *timeout);
aio: add polling mode to AioContext The AioContext event loop uses ppoll(2) or epoll_wait(2) to monitor file descriptors or until a timer expires. In cases like virtqueues, Linux AIO, and ThreadPool it is technically possible to wait for events via polling (i.e. continuously checking for events without blocking). Polling can be faster than blocking syscalls because file descriptors, the process scheduler, and system calls are bypassed. The main disadvantage to polling is that it increases CPU utilization. In classic polling configuration a full host CPU thread might run at 100% to respond to events as quickly as possible. This patch implements a timeout so we fall back to blocking syscalls if polling detects no activity. After the timeout no CPU cycles are wasted on polling until the next event loop iteration. The run_poll_handlers_begin() and run_poll_handlers_end() trace events are added to aid performance analysis and troubleshooting. If you need to know whether polling mode is being used, trace these events to find out. Note that the AioContext is now re-acquired before disabling notify_me in the non-polling case. This makes the code cleaner since notify_me was enabled outside the non-polling AioContext release region. This change is correct since it's safe to keep notify_me enabled longer (disabling is an optimization) but potentially causes unnecessary event_notifer_set() calls. I think the chance of performance regression is small here. Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> Reviewed-by: Paolo Bonzini <pbonzini@redhat.com> Message-id: 20161201192652.9509-4-stefanha@redhat.com Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
2016-12-01 22:26:42 +03:00
return progress;
}
/* try_poll_mode:
* @ctx: the AioContext
aio-posix: split poll check from ready handler Adaptive polling measures the execution time of the polling check plus handlers called when a polled event becomes ready. Handlers can take a significant amount of time, making it look like polling was running for a long time when in fact the event handler was running for a long time. For example, on Linux the io_submit(2) syscall invoked when a virtio-blk device's virtqueue becomes ready can take 10s of microseconds. This can exceed the default polling interval (32 microseconds) and cause adaptive polling to stop polling. By excluding the handler's execution time from the polling check we make the adaptive polling calculation more accurate. As a result, the event loop now stays in polling mode where previously it would have fallen back to file descriptor monitoring. The following data was collected with virtio-blk num-queues=2 event_idx=off using an IOThread. Before: 168k IOPS, IOThread syscalls: 9837.115 ( 0.020 ms): IO iothread1/620155 io_submit(ctx_id: 140512552468480, nr: 16, iocbpp: 0x7fcb9f937db0) = 16 9837.158 ( 0.002 ms): IO iothread1/620155 write(fd: 103, buf: 0x556a2ef71b88, count: 8) = 8 9837.161 ( 0.001 ms): IO iothread1/620155 write(fd: 104, buf: 0x556a2ef71b88, count: 8) = 8 9837.163 ( 0.001 ms): IO iothread1/620155 ppoll(ufds: 0x7fcb90002800, nfds: 4, tsp: 0x7fcb9f1342d0, sigsetsize: 8) = 3 9837.164 ( 0.001 ms): IO iothread1/620155 read(fd: 107, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.174 ( 0.001 ms): IO iothread1/620155 read(fd: 105, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.176 ( 0.001 ms): IO iothread1/620155 read(fd: 106, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.209 ( 0.035 ms): IO iothread1/620155 io_submit(ctx_id: 140512552468480, nr: 32, iocbpp: 0x7fca7d0cebe0) = 32 174k IOPS (+3.6%), IOThread syscalls: 9809.566 ( 0.036 ms): IO iothread1/623061 io_submit(ctx_id: 140539805028352, nr: 32, iocbpp: 0x7fd0cdd62be0) = 32 9809.625 ( 0.001 ms): IO iothread1/623061 write(fd: 103, buf: 0x5647cfba5f58, count: 8) = 8 9809.627 ( 0.002 ms): IO iothread1/623061 write(fd: 104, buf: 0x5647cfba5f58, count: 8) = 8 9809.663 ( 0.036 ms): IO iothread1/623061 io_submit(ctx_id: 140539805028352, nr: 32, iocbpp: 0x7fd0d0388b50) = 32 Notice that ppoll(2) and eventfd read(2) syscalls are eliminated because the IOThread stays in polling mode instead of falling back to file descriptor monitoring. As usual, polling is not implemented on Windows so this patch ignores the new io_poll_read() callback in aio-win32.c. Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> Reviewed-by: Stefano Garzarella <sgarzare@redhat.com> Message-id: 20211207132336.36627-2-stefanha@redhat.com [Fixed up aio_set_event_notifier() calls in tests/unit/test-fdmon-epoll.c added after this series was queued. --Stefan] Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
2021-12-07 16:23:31 +03:00
* @ready_list: list to add handlers that need to be run
* @timeout: timeout for blocking wait, computed by the caller and updated if
* polling succeeds.
aio: add polling mode to AioContext The AioContext event loop uses ppoll(2) or epoll_wait(2) to monitor file descriptors or until a timer expires. In cases like virtqueues, Linux AIO, and ThreadPool it is technically possible to wait for events via polling (i.e. continuously checking for events without blocking). Polling can be faster than blocking syscalls because file descriptors, the process scheduler, and system calls are bypassed. The main disadvantage to polling is that it increases CPU utilization. In classic polling configuration a full host CPU thread might run at 100% to respond to events as quickly as possible. This patch implements a timeout so we fall back to blocking syscalls if polling detects no activity. After the timeout no CPU cycles are wasted on polling until the next event loop iteration. The run_poll_handlers_begin() and run_poll_handlers_end() trace events are added to aid performance analysis and troubleshooting. If you need to know whether polling mode is being used, trace these events to find out. Note that the AioContext is now re-acquired before disabling notify_me in the non-polling case. This makes the code cleaner since notify_me was enabled outside the non-polling AioContext release region. This change is correct since it's safe to keep notify_me enabled longer (disabling is an optimization) but potentially causes unnecessary event_notifer_set() calls. I think the chance of performance regression is small here. Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> Reviewed-by: Paolo Bonzini <pbonzini@redhat.com> Message-id: 20161201192652.9509-4-stefanha@redhat.com Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
2016-12-01 22:26:42 +03:00
*
* Note that the caller must have incremented ctx->list_lock.
aio: add polling mode to AioContext The AioContext event loop uses ppoll(2) or epoll_wait(2) to monitor file descriptors or until a timer expires. In cases like virtqueues, Linux AIO, and ThreadPool it is technically possible to wait for events via polling (i.e. continuously checking for events without blocking). Polling can be faster than blocking syscalls because file descriptors, the process scheduler, and system calls are bypassed. The main disadvantage to polling is that it increases CPU utilization. In classic polling configuration a full host CPU thread might run at 100% to respond to events as quickly as possible. This patch implements a timeout so we fall back to blocking syscalls if polling detects no activity. After the timeout no CPU cycles are wasted on polling until the next event loop iteration. The run_poll_handlers_begin() and run_poll_handlers_end() trace events are added to aid performance analysis and troubleshooting. If you need to know whether polling mode is being used, trace these events to find out. Note that the AioContext is now re-acquired before disabling notify_me in the non-polling case. This makes the code cleaner since notify_me was enabled outside the non-polling AioContext release region. This change is correct since it's safe to keep notify_me enabled longer (disabling is an optimization) but potentially causes unnecessary event_notifer_set() calls. I think the chance of performance regression is small here. Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> Reviewed-by: Paolo Bonzini <pbonzini@redhat.com> Message-id: 20161201192652.9509-4-stefanha@redhat.com Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
2016-12-01 22:26:42 +03:00
*
* Returns: true if progress was made, false otherwise
*/
aio-posix: split poll check from ready handler Adaptive polling measures the execution time of the polling check plus handlers called when a polled event becomes ready. Handlers can take a significant amount of time, making it look like polling was running for a long time when in fact the event handler was running for a long time. For example, on Linux the io_submit(2) syscall invoked when a virtio-blk device's virtqueue becomes ready can take 10s of microseconds. This can exceed the default polling interval (32 microseconds) and cause adaptive polling to stop polling. By excluding the handler's execution time from the polling check we make the adaptive polling calculation more accurate. As a result, the event loop now stays in polling mode where previously it would have fallen back to file descriptor monitoring. The following data was collected with virtio-blk num-queues=2 event_idx=off using an IOThread. Before: 168k IOPS, IOThread syscalls: 9837.115 ( 0.020 ms): IO iothread1/620155 io_submit(ctx_id: 140512552468480, nr: 16, iocbpp: 0x7fcb9f937db0) = 16 9837.158 ( 0.002 ms): IO iothread1/620155 write(fd: 103, buf: 0x556a2ef71b88, count: 8) = 8 9837.161 ( 0.001 ms): IO iothread1/620155 write(fd: 104, buf: 0x556a2ef71b88, count: 8) = 8 9837.163 ( 0.001 ms): IO iothread1/620155 ppoll(ufds: 0x7fcb90002800, nfds: 4, tsp: 0x7fcb9f1342d0, sigsetsize: 8) = 3 9837.164 ( 0.001 ms): IO iothread1/620155 read(fd: 107, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.174 ( 0.001 ms): IO iothread1/620155 read(fd: 105, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.176 ( 0.001 ms): IO iothread1/620155 read(fd: 106, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.209 ( 0.035 ms): IO iothread1/620155 io_submit(ctx_id: 140512552468480, nr: 32, iocbpp: 0x7fca7d0cebe0) = 32 174k IOPS (+3.6%), IOThread syscalls: 9809.566 ( 0.036 ms): IO iothread1/623061 io_submit(ctx_id: 140539805028352, nr: 32, iocbpp: 0x7fd0cdd62be0) = 32 9809.625 ( 0.001 ms): IO iothread1/623061 write(fd: 103, buf: 0x5647cfba5f58, count: 8) = 8 9809.627 ( 0.002 ms): IO iothread1/623061 write(fd: 104, buf: 0x5647cfba5f58, count: 8) = 8 9809.663 ( 0.036 ms): IO iothread1/623061 io_submit(ctx_id: 140539805028352, nr: 32, iocbpp: 0x7fd0d0388b50) = 32 Notice that ppoll(2) and eventfd read(2) syscalls are eliminated because the IOThread stays in polling mode instead of falling back to file descriptor monitoring. As usual, polling is not implemented on Windows so this patch ignores the new io_poll_read() callback in aio-win32.c. Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> Reviewed-by: Stefano Garzarella <sgarzare@redhat.com> Message-id: 20211207132336.36627-2-stefanha@redhat.com [Fixed up aio_set_event_notifier() calls in tests/unit/test-fdmon-epoll.c added after this series was queued. --Stefan] Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
2021-12-07 16:23:31 +03:00
static bool try_poll_mode(AioContext *ctx, AioHandlerList *ready_list,
int64_t *timeout)
aio: add polling mode to AioContext The AioContext event loop uses ppoll(2) or epoll_wait(2) to monitor file descriptors or until a timer expires. In cases like virtqueues, Linux AIO, and ThreadPool it is technically possible to wait for events via polling (i.e. continuously checking for events without blocking). Polling can be faster than blocking syscalls because file descriptors, the process scheduler, and system calls are bypassed. The main disadvantage to polling is that it increases CPU utilization. In classic polling configuration a full host CPU thread might run at 100% to respond to events as quickly as possible. This patch implements a timeout so we fall back to blocking syscalls if polling detects no activity. After the timeout no CPU cycles are wasted on polling until the next event loop iteration. The run_poll_handlers_begin() and run_poll_handlers_end() trace events are added to aid performance analysis and troubleshooting. If you need to know whether polling mode is being used, trace these events to find out. Note that the AioContext is now re-acquired before disabling notify_me in the non-polling case. This makes the code cleaner since notify_me was enabled outside the non-polling AioContext release region. This change is correct since it's safe to keep notify_me enabled longer (disabling is an optimization) but potentially causes unnecessary event_notifer_set() calls. I think the chance of performance regression is small here. Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> Reviewed-by: Paolo Bonzini <pbonzini@redhat.com> Message-id: 20161201192652.9509-4-stefanha@redhat.com Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
2016-12-01 22:26:42 +03:00
{
int64_t max_ns;
if (QLIST_EMPTY_RCU(&ctx->poll_aio_handlers)) {
return false;
}
aio: add polling mode to AioContext The AioContext event loop uses ppoll(2) or epoll_wait(2) to monitor file descriptors or until a timer expires. In cases like virtqueues, Linux AIO, and ThreadPool it is technically possible to wait for events via polling (i.e. continuously checking for events without blocking). Polling can be faster than blocking syscalls because file descriptors, the process scheduler, and system calls are bypassed. The main disadvantage to polling is that it increases CPU utilization. In classic polling configuration a full host CPU thread might run at 100% to respond to events as quickly as possible. This patch implements a timeout so we fall back to blocking syscalls if polling detects no activity. After the timeout no CPU cycles are wasted on polling until the next event loop iteration. The run_poll_handlers_begin() and run_poll_handlers_end() trace events are added to aid performance analysis and troubleshooting. If you need to know whether polling mode is being used, trace these events to find out. Note that the AioContext is now re-acquired before disabling notify_me in the non-polling case. This makes the code cleaner since notify_me was enabled outside the non-polling AioContext release region. This change is correct since it's safe to keep notify_me enabled longer (disabling is an optimization) but potentially causes unnecessary event_notifer_set() calls. I think the chance of performance regression is small here. Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> Reviewed-by: Paolo Bonzini <pbonzini@redhat.com> Message-id: 20161201192652.9509-4-stefanha@redhat.com Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
2016-12-01 22:26:42 +03:00
max_ns = qemu_soonest_timeout(*timeout, ctx->poll_ns);
if (max_ns && !ctx->fdmon_ops->need_wait(ctx)) {
/*
* Enable poll mode. It pairs with the poll_set_started() in
* aio_poll() which disables poll mode.
*/
aio-posix: split poll check from ready handler Adaptive polling measures the execution time of the polling check plus handlers called when a polled event becomes ready. Handlers can take a significant amount of time, making it look like polling was running for a long time when in fact the event handler was running for a long time. For example, on Linux the io_submit(2) syscall invoked when a virtio-blk device's virtqueue becomes ready can take 10s of microseconds. This can exceed the default polling interval (32 microseconds) and cause adaptive polling to stop polling. By excluding the handler's execution time from the polling check we make the adaptive polling calculation more accurate. As a result, the event loop now stays in polling mode where previously it would have fallen back to file descriptor monitoring. The following data was collected with virtio-blk num-queues=2 event_idx=off using an IOThread. Before: 168k IOPS, IOThread syscalls: 9837.115 ( 0.020 ms): IO iothread1/620155 io_submit(ctx_id: 140512552468480, nr: 16, iocbpp: 0x7fcb9f937db0) = 16 9837.158 ( 0.002 ms): IO iothread1/620155 write(fd: 103, buf: 0x556a2ef71b88, count: 8) = 8 9837.161 ( 0.001 ms): IO iothread1/620155 write(fd: 104, buf: 0x556a2ef71b88, count: 8) = 8 9837.163 ( 0.001 ms): IO iothread1/620155 ppoll(ufds: 0x7fcb90002800, nfds: 4, tsp: 0x7fcb9f1342d0, sigsetsize: 8) = 3 9837.164 ( 0.001 ms): IO iothread1/620155 read(fd: 107, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.174 ( 0.001 ms): IO iothread1/620155 read(fd: 105, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.176 ( 0.001 ms): IO iothread1/620155 read(fd: 106, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.209 ( 0.035 ms): IO iothread1/620155 io_submit(ctx_id: 140512552468480, nr: 32, iocbpp: 0x7fca7d0cebe0) = 32 174k IOPS (+3.6%), IOThread syscalls: 9809.566 ( 0.036 ms): IO iothread1/623061 io_submit(ctx_id: 140539805028352, nr: 32, iocbpp: 0x7fd0cdd62be0) = 32 9809.625 ( 0.001 ms): IO iothread1/623061 write(fd: 103, buf: 0x5647cfba5f58, count: 8) = 8 9809.627 ( 0.002 ms): IO iothread1/623061 write(fd: 104, buf: 0x5647cfba5f58, count: 8) = 8 9809.663 ( 0.036 ms): IO iothread1/623061 io_submit(ctx_id: 140539805028352, nr: 32, iocbpp: 0x7fd0d0388b50) = 32 Notice that ppoll(2) and eventfd read(2) syscalls are eliminated because the IOThread stays in polling mode instead of falling back to file descriptor monitoring. As usual, polling is not implemented on Windows so this patch ignores the new io_poll_read() callback in aio-win32.c. Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> Reviewed-by: Stefano Garzarella <sgarzare@redhat.com> Message-id: 20211207132336.36627-2-stefanha@redhat.com [Fixed up aio_set_event_notifier() calls in tests/unit/test-fdmon-epoll.c added after this series was queued. --Stefan] Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
2021-12-07 16:23:31 +03:00
poll_set_started(ctx, ready_list, true);
aio-posix: split poll check from ready handler Adaptive polling measures the execution time of the polling check plus handlers called when a polled event becomes ready. Handlers can take a significant amount of time, making it look like polling was running for a long time when in fact the event handler was running for a long time. For example, on Linux the io_submit(2) syscall invoked when a virtio-blk device's virtqueue becomes ready can take 10s of microseconds. This can exceed the default polling interval (32 microseconds) and cause adaptive polling to stop polling. By excluding the handler's execution time from the polling check we make the adaptive polling calculation more accurate. As a result, the event loop now stays in polling mode where previously it would have fallen back to file descriptor monitoring. The following data was collected with virtio-blk num-queues=2 event_idx=off using an IOThread. Before: 168k IOPS, IOThread syscalls: 9837.115 ( 0.020 ms): IO iothread1/620155 io_submit(ctx_id: 140512552468480, nr: 16, iocbpp: 0x7fcb9f937db0) = 16 9837.158 ( 0.002 ms): IO iothread1/620155 write(fd: 103, buf: 0x556a2ef71b88, count: 8) = 8 9837.161 ( 0.001 ms): IO iothread1/620155 write(fd: 104, buf: 0x556a2ef71b88, count: 8) = 8 9837.163 ( 0.001 ms): IO iothread1/620155 ppoll(ufds: 0x7fcb90002800, nfds: 4, tsp: 0x7fcb9f1342d0, sigsetsize: 8) = 3 9837.164 ( 0.001 ms): IO iothread1/620155 read(fd: 107, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.174 ( 0.001 ms): IO iothread1/620155 read(fd: 105, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.176 ( 0.001 ms): IO iothread1/620155 read(fd: 106, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.209 ( 0.035 ms): IO iothread1/620155 io_submit(ctx_id: 140512552468480, nr: 32, iocbpp: 0x7fca7d0cebe0) = 32 174k IOPS (+3.6%), IOThread syscalls: 9809.566 ( 0.036 ms): IO iothread1/623061 io_submit(ctx_id: 140539805028352, nr: 32, iocbpp: 0x7fd0cdd62be0) = 32 9809.625 ( 0.001 ms): IO iothread1/623061 write(fd: 103, buf: 0x5647cfba5f58, count: 8) = 8 9809.627 ( 0.002 ms): IO iothread1/623061 write(fd: 104, buf: 0x5647cfba5f58, count: 8) = 8 9809.663 ( 0.036 ms): IO iothread1/623061 io_submit(ctx_id: 140539805028352, nr: 32, iocbpp: 0x7fd0d0388b50) = 32 Notice that ppoll(2) and eventfd read(2) syscalls are eliminated because the IOThread stays in polling mode instead of falling back to file descriptor monitoring. As usual, polling is not implemented on Windows so this patch ignores the new io_poll_read() callback in aio-win32.c. Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> Reviewed-by: Stefano Garzarella <sgarzare@redhat.com> Message-id: 20211207132336.36627-2-stefanha@redhat.com [Fixed up aio_set_event_notifier() calls in tests/unit/test-fdmon-epoll.c added after this series was queued. --Stefan] Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
2021-12-07 16:23:31 +03:00
if (run_poll_handlers(ctx, ready_list, max_ns, timeout)) {
return true;
aio: add polling mode to AioContext The AioContext event loop uses ppoll(2) or epoll_wait(2) to monitor file descriptors or until a timer expires. In cases like virtqueues, Linux AIO, and ThreadPool it is technically possible to wait for events via polling (i.e. continuously checking for events without blocking). Polling can be faster than blocking syscalls because file descriptors, the process scheduler, and system calls are bypassed. The main disadvantage to polling is that it increases CPU utilization. In classic polling configuration a full host CPU thread might run at 100% to respond to events as quickly as possible. This patch implements a timeout so we fall back to blocking syscalls if polling detects no activity. After the timeout no CPU cycles are wasted on polling until the next event loop iteration. The run_poll_handlers_begin() and run_poll_handlers_end() trace events are added to aid performance analysis and troubleshooting. If you need to know whether polling mode is being used, trace these events to find out. Note that the AioContext is now re-acquired before disabling notify_me in the non-polling case. This makes the code cleaner since notify_me was enabled outside the non-polling AioContext release region. This change is correct since it's safe to keep notify_me enabled longer (disabling is an optimization) but potentially causes unnecessary event_notifer_set() calls. I think the chance of performance regression is small here. Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> Reviewed-by: Paolo Bonzini <pbonzini@redhat.com> Message-id: 20161201192652.9509-4-stefanha@redhat.com Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
2016-12-01 22:26:42 +03:00
}
}
return false;
aio: add polling mode to AioContext The AioContext event loop uses ppoll(2) or epoll_wait(2) to monitor file descriptors or until a timer expires. In cases like virtqueues, Linux AIO, and ThreadPool it is technically possible to wait for events via polling (i.e. continuously checking for events without blocking). Polling can be faster than blocking syscalls because file descriptors, the process scheduler, and system calls are bypassed. The main disadvantage to polling is that it increases CPU utilization. In classic polling configuration a full host CPU thread might run at 100% to respond to events as quickly as possible. This patch implements a timeout so we fall back to blocking syscalls if polling detects no activity. After the timeout no CPU cycles are wasted on polling until the next event loop iteration. The run_poll_handlers_begin() and run_poll_handlers_end() trace events are added to aid performance analysis and troubleshooting. If you need to know whether polling mode is being used, trace these events to find out. Note that the AioContext is now re-acquired before disabling notify_me in the non-polling case. This makes the code cleaner since notify_me was enabled outside the non-polling AioContext release region. This change is correct since it's safe to keep notify_me enabled longer (disabling is an optimization) but potentially causes unnecessary event_notifer_set() calls. I think the chance of performance regression is small here. Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> Reviewed-by: Paolo Bonzini <pbonzini@redhat.com> Message-id: 20161201192652.9509-4-stefanha@redhat.com Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
2016-12-01 22:26:42 +03:00
}
bool aio_poll(AioContext *ctx, bool blocking)
{
AioHandlerList ready_list = QLIST_HEAD_INITIALIZER(ready_list);
aio: stop using .io_flush() Now that aio_poll() users check their termination condition themselves, it is no longer necessary to call .io_flush() handlers. The behavior of aio_poll() changes as follows: 1. .io_flush() is no longer invoked and file descriptors are *always* monitored. Previously returning 0 from .io_flush() would skip this file descriptor. Due to this change it is essential to check that requests are pending before calling qemu_aio_wait(). Failure to do so means we block, for example, waiting for an idle iSCSI socket to become readable when there are no requests. Currently all qemu_aio_wait()/aio_poll() callers check before calling. 2. aio_poll() now returns true if progress was made (BH or fd handlers executed) and false otherwise. Previously it would return true whenever 'busy', which means that .io_flush() returned true. The 'busy' concept no longer exists so just progress is returned. Due to this change we need to update tests/test-aio.c which asserts aio_poll() return values. Note that QEMU doesn't actually rely on these return values so only tests/test-aio.c cares. Note that ctx->notifier, the EventNotifier fd used for aio_notify(), is now handled as a special case. This is a little ugly but maintains aio_poll() semantics, i.e. aio_notify() does not count as 'progress' and aio_poll() avoids blocking when the user has not set any fd handlers yet. Patches after this remove .io_flush() handler code until we can finally drop the io_flush arguments to aio_set_fd_handler() and friends. Reviewed-by: Paolo Bonzini <pbonzini@redhat.com> Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
2013-04-11 18:56:50 +04:00
bool progress;
bool use_notify_me;
int64_t timeout;
int64_t start = 0;
/*
* There cannot be two concurrent aio_poll calls for the same AioContext (or
* an aio_poll concurrent with a GSource prepare/check/dispatch callback).
* We rely on this below to avoid slow locked accesses to ctx->notify_me.
*
* aio_poll() may only be called in the AioContext's thread. iohandler_ctx
* is special in that it runs in the main thread, but that thread's context
* is qemu_aio_context.
*/
assert(in_aio_context_home_thread(ctx == iohandler_get_aio_context() ?
qemu_get_aio_context() : ctx));
qemu_lockcnt_inc(&ctx->list_lock);
if (ctx->poll_max_ns) {
start = qemu_clock_get_ns(QEMU_CLOCK_REALTIME);
}
timeout = blocking ? aio_compute_timeout(ctx) : 0;
aio-posix: split poll check from ready handler Adaptive polling measures the execution time of the polling check plus handlers called when a polled event becomes ready. Handlers can take a significant amount of time, making it look like polling was running for a long time when in fact the event handler was running for a long time. For example, on Linux the io_submit(2) syscall invoked when a virtio-blk device's virtqueue becomes ready can take 10s of microseconds. This can exceed the default polling interval (32 microseconds) and cause adaptive polling to stop polling. By excluding the handler's execution time from the polling check we make the adaptive polling calculation more accurate. As a result, the event loop now stays in polling mode where previously it would have fallen back to file descriptor monitoring. The following data was collected with virtio-blk num-queues=2 event_idx=off using an IOThread. Before: 168k IOPS, IOThread syscalls: 9837.115 ( 0.020 ms): IO iothread1/620155 io_submit(ctx_id: 140512552468480, nr: 16, iocbpp: 0x7fcb9f937db0) = 16 9837.158 ( 0.002 ms): IO iothread1/620155 write(fd: 103, buf: 0x556a2ef71b88, count: 8) = 8 9837.161 ( 0.001 ms): IO iothread1/620155 write(fd: 104, buf: 0x556a2ef71b88, count: 8) = 8 9837.163 ( 0.001 ms): IO iothread1/620155 ppoll(ufds: 0x7fcb90002800, nfds: 4, tsp: 0x7fcb9f1342d0, sigsetsize: 8) = 3 9837.164 ( 0.001 ms): IO iothread1/620155 read(fd: 107, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.174 ( 0.001 ms): IO iothread1/620155 read(fd: 105, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.176 ( 0.001 ms): IO iothread1/620155 read(fd: 106, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.209 ( 0.035 ms): IO iothread1/620155 io_submit(ctx_id: 140512552468480, nr: 32, iocbpp: 0x7fca7d0cebe0) = 32 174k IOPS (+3.6%), IOThread syscalls: 9809.566 ( 0.036 ms): IO iothread1/623061 io_submit(ctx_id: 140539805028352, nr: 32, iocbpp: 0x7fd0cdd62be0) = 32 9809.625 ( 0.001 ms): IO iothread1/623061 write(fd: 103, buf: 0x5647cfba5f58, count: 8) = 8 9809.627 ( 0.002 ms): IO iothread1/623061 write(fd: 104, buf: 0x5647cfba5f58, count: 8) = 8 9809.663 ( 0.036 ms): IO iothread1/623061 io_submit(ctx_id: 140539805028352, nr: 32, iocbpp: 0x7fd0d0388b50) = 32 Notice that ppoll(2) and eventfd read(2) syscalls are eliminated because the IOThread stays in polling mode instead of falling back to file descriptor monitoring. As usual, polling is not implemented on Windows so this patch ignores the new io_poll_read() callback in aio-win32.c. Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> Reviewed-by: Stefano Garzarella <sgarzare@redhat.com> Message-id: 20211207132336.36627-2-stefanha@redhat.com [Fixed up aio_set_event_notifier() calls in tests/unit/test-fdmon-epoll.c added after this series was queued. --Stefan] Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
2021-12-07 16:23:31 +03:00
progress = try_poll_mode(ctx, &ready_list, &timeout);
assert(!(timeout && progress));
/*
* aio_notify can avoid the expensive event_notifier_set if
* everything (file descriptors, bottom halves, timers) will
* be re-evaluated before the next blocking poll(). This is
* already true when aio_poll is called with blocking == false;
AioContext: fix broken ctx->dispatching optimization This patch rewrites the ctx->dispatching optimization, which was the cause of some mysterious hangs that could be reproduced on aarch64 KVM only. The hangs were indirectly caused by aio_poll() and in particular by flash memory updates's call to blk_write(), which invokes aio_poll(). Fun stuff: they had an extremely short race window, so much that adding all kind of tracing to either the kernel or QEMU made it go away (a single printf made it half as reproducible). On the plus side, the failure mode (a hang until the next keypress) made it very easy to examine the state of the process with a debugger. And there was a very nice reproducer from Laszlo, which failed pretty often (more than half of the time) on any version of QEMU with a non-debug kernel; it also failed fast, while still in the firmware. So, it could have been worse. For some unknown reason they happened only with virtio-scsi, but that's not important. It's more interesting that they disappeared with io=native, making thread-pool.c a likely suspect for where the bug arose. thread-pool.c is also one of the few places which use bottom halves across threads, by the way. I hope that no other similar bugs exist, but just in case :) I am going to describe how the successful debugging went... Since the likely culprit was the ctx->dispatching optimization, which mostly affects bottom halves, the first observation was that there are two qemu_bh_schedule() invocations in the thread pool: the one in the aio worker and the one in thread_pool_completion_bh. The latter always causes the optimization to trigger, the former may or may not. In order to restrict the possibilities, I introduced new functions qemu_bh_schedule_slow() and qemu_bh_schedule_fast(): /* qemu_bh_schedule_slow: */ ctx = bh->ctx; bh->idle = 0; if (atomic_xchg(&bh->scheduled, 1) == 0) { event_notifier_set(&ctx->notifier); } /* qemu_bh_schedule_fast: */ ctx = bh->ctx; bh->idle = 0; assert(ctx->dispatching); atomic_xchg(&bh->scheduled, 1); Notice how the atomic_xchg is still in qemu_bh_schedule_slow(). This was already debated a few months ago, so I assumed it to be correct. In retrospect this was a very good idea, as you'll see later. Changing thread_pool_completion_bh() to qemu_bh_schedule_fast() didn't trigger the assertion (as expected). Changing the worker's invocation to qemu_bh_schedule_slow() didn't hide the bug (another assumption which luckily held). This already limited heavily the amount of interaction between the threads, hinting that the problematic events must have triggered around thread_pool_completion_bh(). As mentioned early, invoking a debugger to examine the state of a hung process was pretty easy; the iothread was always waiting on a poll(..., -1) system call. Infinite timeouts are much rarer on x86, and this could be the reason why the bug was never observed there. With the buggy sequence more or less resolved to an interaction between thread_pool_completion_bh() and poll(..., -1), my "tracing" strategy was to just add a few qemu_clock_get_ns(QEMU_CLOCK_REALTIME) calls, hoping that the ordering of aio_ctx_prepare(), aio_ctx_dispatch, poll() and qemu_bh_schedule_fast() would provide some hint. The output was: (gdb) p last_prepare $3 = 103885451 (gdb) p last_dispatch $4 = 103876492 (gdb) p last_poll $5 = 115909333 (gdb) p last_schedule $6 = 115925212 Notice how the last call to qemu_poll_ns() came after aio_ctx_dispatch(). This makes little sense unless there is an aio_poll() call involved, and indeed with a slightly different instrumentation you can see that there is one: (gdb) p last_prepare $3 = 107569679 (gdb) p last_dispatch $4 = 107561600 (gdb) p last_aio_poll $5 = 110671400 (gdb) p last_schedule $6 = 110698917 So the scenario becomes clearer: iothread VCPU thread -------------------------------------------------------------------------- aio_ctx_prepare aio_ctx_check qemu_poll_ns(timeout=-1) aio_poll aio_dispatch thread_pool_completion_bh qemu_bh_schedule() At this point bh->scheduled = 1 and the iothread has not been woken up. The solution must be close, but this alone should not be a problem, because the bottom half is only rescheduled to account for rare situations (see commit 3c80ca1, thread-pool: avoid deadlock in nested aio_poll() calls, 2014-07-15). Introducing a third thread---a thread pool worker thread, which also does qemu_bh_schedule()---does bring out the problematic case. The third thread must be awakened *after* the callback is complete and thread_pool_completion_bh has redone the whole loop, explaining the short race window. And then this is what happens: thread pool worker -------------------------------------------------------------------------- <I/O completes> qemu_bh_schedule() Tada, bh->scheduled is already 1, so qemu_bh_schedule() does nothing and the iothread is never woken up. This is where the bh->scheduled optimization comes into play---it is correct, but removing it would have masked the bug. So, what is the bug? Well, the question asked by the ctx->dispatching optimization ("is any active aio_poll dispatching?") was wrong. The right question to ask instead is "is any active aio_poll *not* dispatching", i.e. in the prepare or poll phases? In that case, the aio_poll is sleeping or might go to sleep anytime soon, and the EventNotifier must be invoked to wake it up. In any other case (including if there is *no* active aio_poll at all!) we can just wait for the next prepare phase to pick up the event (e.g. a bottom half); the prepare phase will avoid the blocking and service the bottom half. Expressing the invariant with a logic formula, the broken one looked like: !(exists(thread): in_dispatching(thread)) => !optimize or equivalently: !(exists(thread): in_aio_poll(thread) && in_dispatching(thread)) => !optimize In the correct one, the negation is in a slightly different place: (exists(thread): in_aio_poll(thread) && !in_dispatching(thread)) => !optimize or equivalently: (exists(thread): in_prepare_or_poll(thread)) => !optimize Even if the difference boils down to moving an exclamation mark :) the implementation is quite different. However, I think the new one is simpler to understand. In the old implementation, the "exists" was implemented with a boolean value. This didn't really support well the case of multiple concurrent event loops, but I thought that this was okay: aio_poll holds the AioContext lock so there cannot be concurrent aio_poll invocations, and I was just considering nested event loops. However, aio_poll _could_ indeed be concurrent with the GSource. This is why I came up with the wrong invariant. In the new implementation, "exists" is computed simply by counting how many threads are in the prepare or poll phases. There are some interesting points to consider, but the gist of the idea remains: 1) AioContext can be used through GSource as well; as mentioned in the patch, bit 0 of the counter is reserved for the GSource. 2) the counter need not be updated for a non-blocking aio_poll, because it won't sleep forever anyway. This is just a matter of checking the "blocking" variable. This requires some changes to the win32 implementation, but is otherwise not too complicated. 3) as mentioned above, the new implementation will not call aio_notify when there is *no* active aio_poll at all. The tests have to be adjusted for this change. The calls to aio_notify in async.c are fine; they only want to kick aio_poll out of a blocking wait, but need not do anything if aio_poll is not running. 4) nested aio_poll: these just work with the new implementation; when a nested event loop is invoked, the outer event loop is never in the prepare or poll phases. The outer event loop thus has already decremented the counter. Reported-by: Richard W. M. Jones <rjones@redhat.com> Reported-by: Laszlo Ersek <lersek@redhat.com> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com> Reviewed-by: Fam Zheng <famz@redhat.com> Tested-by: Richard W.M. Jones <rjones@redhat.com> Message-id: 1437487673-23740-5-git-send-email-pbonzini@redhat.com Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
2015-07-21 17:07:51 +03:00
* if blocking == true, it is only true after poll() returns,
* so disable the optimization now.
*/
use_notify_me = timeout != 0;
if (use_notify_me) {
qatomic_set(&ctx->notify_me, qatomic_read(&ctx->notify_me) + 2);
/*
* Write ctx->notify_me before reading ctx->notified. Pairs with
* smp_mb in aio_notify().
*/
smp_mb();
/* Don't block if aio_notify() was called */
if (qatomic_read(&ctx->notified)) {
timeout = 0;
}
}
/* If polling is allowed, non-blocking aio_poll does not need the
* system call---a single round of run_poll_handlers_once suffices.
*/
if (timeout || ctx->fdmon_ops->need_wait(ctx)) {
/*
* Disable poll mode. poll mode should be disabled before the call
* of ctx->fdmon_ops->wait() so that guest's notification can wake
* up IO threads when some work becomes pending. It is essential to
* avoid hangs or unnecessary latency.
*/
if (poll_set_started(ctx, &ready_list, false)) {
timeout = 0;
progress = true;
}
aio-posix: split poll check from ready handler Adaptive polling measures the execution time of the polling check plus handlers called when a polled event becomes ready. Handlers can take a significant amount of time, making it look like polling was running for a long time when in fact the event handler was running for a long time. For example, on Linux the io_submit(2) syscall invoked when a virtio-blk device's virtqueue becomes ready can take 10s of microseconds. This can exceed the default polling interval (32 microseconds) and cause adaptive polling to stop polling. By excluding the handler's execution time from the polling check we make the adaptive polling calculation more accurate. As a result, the event loop now stays in polling mode where previously it would have fallen back to file descriptor monitoring. The following data was collected with virtio-blk num-queues=2 event_idx=off using an IOThread. Before: 168k IOPS, IOThread syscalls: 9837.115 ( 0.020 ms): IO iothread1/620155 io_submit(ctx_id: 140512552468480, nr: 16, iocbpp: 0x7fcb9f937db0) = 16 9837.158 ( 0.002 ms): IO iothread1/620155 write(fd: 103, buf: 0x556a2ef71b88, count: 8) = 8 9837.161 ( 0.001 ms): IO iothread1/620155 write(fd: 104, buf: 0x556a2ef71b88, count: 8) = 8 9837.163 ( 0.001 ms): IO iothread1/620155 ppoll(ufds: 0x7fcb90002800, nfds: 4, tsp: 0x7fcb9f1342d0, sigsetsize: 8) = 3 9837.164 ( 0.001 ms): IO iothread1/620155 read(fd: 107, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.174 ( 0.001 ms): IO iothread1/620155 read(fd: 105, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.176 ( 0.001 ms): IO iothread1/620155 read(fd: 106, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.209 ( 0.035 ms): IO iothread1/620155 io_submit(ctx_id: 140512552468480, nr: 32, iocbpp: 0x7fca7d0cebe0) = 32 174k IOPS (+3.6%), IOThread syscalls: 9809.566 ( 0.036 ms): IO iothread1/623061 io_submit(ctx_id: 140539805028352, nr: 32, iocbpp: 0x7fd0cdd62be0) = 32 9809.625 ( 0.001 ms): IO iothread1/623061 write(fd: 103, buf: 0x5647cfba5f58, count: 8) = 8 9809.627 ( 0.002 ms): IO iothread1/623061 write(fd: 104, buf: 0x5647cfba5f58, count: 8) = 8 9809.663 ( 0.036 ms): IO iothread1/623061 io_submit(ctx_id: 140539805028352, nr: 32, iocbpp: 0x7fd0d0388b50) = 32 Notice that ppoll(2) and eventfd read(2) syscalls are eliminated because the IOThread stays in polling mode instead of falling back to file descriptor monitoring. As usual, polling is not implemented on Windows so this patch ignores the new io_poll_read() callback in aio-win32.c. Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> Reviewed-by: Stefano Garzarella <sgarzare@redhat.com> Message-id: 20211207132336.36627-2-stefanha@redhat.com [Fixed up aio_set_event_notifier() calls in tests/unit/test-fdmon-epoll.c added after this series was queued. --Stefan] Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
2021-12-07 16:23:31 +03:00
ctx->fdmon_ops->wait(ctx, &ready_list, timeout);
}
aio: add polling mode to AioContext The AioContext event loop uses ppoll(2) or epoll_wait(2) to monitor file descriptors or until a timer expires. In cases like virtqueues, Linux AIO, and ThreadPool it is technically possible to wait for events via polling (i.e. continuously checking for events without blocking). Polling can be faster than blocking syscalls because file descriptors, the process scheduler, and system calls are bypassed. The main disadvantage to polling is that it increases CPU utilization. In classic polling configuration a full host CPU thread might run at 100% to respond to events as quickly as possible. This patch implements a timeout so we fall back to blocking syscalls if polling detects no activity. After the timeout no CPU cycles are wasted on polling until the next event loop iteration. The run_poll_handlers_begin() and run_poll_handlers_end() trace events are added to aid performance analysis and troubleshooting. If you need to know whether polling mode is being used, trace these events to find out. Note that the AioContext is now re-acquired before disabling notify_me in the non-polling case. This makes the code cleaner since notify_me was enabled outside the non-polling AioContext release region. This change is correct since it's safe to keep notify_me enabled longer (disabling is an optimization) but potentially causes unnecessary event_notifer_set() calls. I think the chance of performance regression is small here. Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> Reviewed-by: Paolo Bonzini <pbonzini@redhat.com> Message-id: 20161201192652.9509-4-stefanha@redhat.com Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
2016-12-01 22:26:42 +03:00
if (use_notify_me) {
/* Finish the poll before clearing the flag. */
qatomic_store_release(&ctx->notify_me,
qatomic_read(&ctx->notify_me) - 2);
AioContext: fix broken ctx->dispatching optimization This patch rewrites the ctx->dispatching optimization, which was the cause of some mysterious hangs that could be reproduced on aarch64 KVM only. The hangs were indirectly caused by aio_poll() and in particular by flash memory updates's call to blk_write(), which invokes aio_poll(). Fun stuff: they had an extremely short race window, so much that adding all kind of tracing to either the kernel or QEMU made it go away (a single printf made it half as reproducible). On the plus side, the failure mode (a hang until the next keypress) made it very easy to examine the state of the process with a debugger. And there was a very nice reproducer from Laszlo, which failed pretty often (more than half of the time) on any version of QEMU with a non-debug kernel; it also failed fast, while still in the firmware. So, it could have been worse. For some unknown reason they happened only with virtio-scsi, but that's not important. It's more interesting that they disappeared with io=native, making thread-pool.c a likely suspect for where the bug arose. thread-pool.c is also one of the few places which use bottom halves across threads, by the way. I hope that no other similar bugs exist, but just in case :) I am going to describe how the successful debugging went... Since the likely culprit was the ctx->dispatching optimization, which mostly affects bottom halves, the first observation was that there are two qemu_bh_schedule() invocations in the thread pool: the one in the aio worker and the one in thread_pool_completion_bh. The latter always causes the optimization to trigger, the former may or may not. In order to restrict the possibilities, I introduced new functions qemu_bh_schedule_slow() and qemu_bh_schedule_fast(): /* qemu_bh_schedule_slow: */ ctx = bh->ctx; bh->idle = 0; if (atomic_xchg(&bh->scheduled, 1) == 0) { event_notifier_set(&ctx->notifier); } /* qemu_bh_schedule_fast: */ ctx = bh->ctx; bh->idle = 0; assert(ctx->dispatching); atomic_xchg(&bh->scheduled, 1); Notice how the atomic_xchg is still in qemu_bh_schedule_slow(). This was already debated a few months ago, so I assumed it to be correct. In retrospect this was a very good idea, as you'll see later. Changing thread_pool_completion_bh() to qemu_bh_schedule_fast() didn't trigger the assertion (as expected). Changing the worker's invocation to qemu_bh_schedule_slow() didn't hide the bug (another assumption which luckily held). This already limited heavily the amount of interaction between the threads, hinting that the problematic events must have triggered around thread_pool_completion_bh(). As mentioned early, invoking a debugger to examine the state of a hung process was pretty easy; the iothread was always waiting on a poll(..., -1) system call. Infinite timeouts are much rarer on x86, and this could be the reason why the bug was never observed there. With the buggy sequence more or less resolved to an interaction between thread_pool_completion_bh() and poll(..., -1), my "tracing" strategy was to just add a few qemu_clock_get_ns(QEMU_CLOCK_REALTIME) calls, hoping that the ordering of aio_ctx_prepare(), aio_ctx_dispatch, poll() and qemu_bh_schedule_fast() would provide some hint. The output was: (gdb) p last_prepare $3 = 103885451 (gdb) p last_dispatch $4 = 103876492 (gdb) p last_poll $5 = 115909333 (gdb) p last_schedule $6 = 115925212 Notice how the last call to qemu_poll_ns() came after aio_ctx_dispatch(). This makes little sense unless there is an aio_poll() call involved, and indeed with a slightly different instrumentation you can see that there is one: (gdb) p last_prepare $3 = 107569679 (gdb) p last_dispatch $4 = 107561600 (gdb) p last_aio_poll $5 = 110671400 (gdb) p last_schedule $6 = 110698917 So the scenario becomes clearer: iothread VCPU thread -------------------------------------------------------------------------- aio_ctx_prepare aio_ctx_check qemu_poll_ns(timeout=-1) aio_poll aio_dispatch thread_pool_completion_bh qemu_bh_schedule() At this point bh->scheduled = 1 and the iothread has not been woken up. The solution must be close, but this alone should not be a problem, because the bottom half is only rescheduled to account for rare situations (see commit 3c80ca1, thread-pool: avoid deadlock in nested aio_poll() calls, 2014-07-15). Introducing a third thread---a thread pool worker thread, which also does qemu_bh_schedule()---does bring out the problematic case. The third thread must be awakened *after* the callback is complete and thread_pool_completion_bh has redone the whole loop, explaining the short race window. And then this is what happens: thread pool worker -------------------------------------------------------------------------- <I/O completes> qemu_bh_schedule() Tada, bh->scheduled is already 1, so qemu_bh_schedule() does nothing and the iothread is never woken up. This is where the bh->scheduled optimization comes into play---it is correct, but removing it would have masked the bug. So, what is the bug? Well, the question asked by the ctx->dispatching optimization ("is any active aio_poll dispatching?") was wrong. The right question to ask instead is "is any active aio_poll *not* dispatching", i.e. in the prepare or poll phases? In that case, the aio_poll is sleeping or might go to sleep anytime soon, and the EventNotifier must be invoked to wake it up. In any other case (including if there is *no* active aio_poll at all!) we can just wait for the next prepare phase to pick up the event (e.g. a bottom half); the prepare phase will avoid the blocking and service the bottom half. Expressing the invariant with a logic formula, the broken one looked like: !(exists(thread): in_dispatching(thread)) => !optimize or equivalently: !(exists(thread): in_aio_poll(thread) && in_dispatching(thread)) => !optimize In the correct one, the negation is in a slightly different place: (exists(thread): in_aio_poll(thread) && !in_dispatching(thread)) => !optimize or equivalently: (exists(thread): in_prepare_or_poll(thread)) => !optimize Even if the difference boils down to moving an exclamation mark :) the implementation is quite different. However, I think the new one is simpler to understand. In the old implementation, the "exists" was implemented with a boolean value. This didn't really support well the case of multiple concurrent event loops, but I thought that this was okay: aio_poll holds the AioContext lock so there cannot be concurrent aio_poll invocations, and I was just considering nested event loops. However, aio_poll _could_ indeed be concurrent with the GSource. This is why I came up with the wrong invariant. In the new implementation, "exists" is computed simply by counting how many threads are in the prepare or poll phases. There are some interesting points to consider, but the gist of the idea remains: 1) AioContext can be used through GSource as well; as mentioned in the patch, bit 0 of the counter is reserved for the GSource. 2) the counter need not be updated for a non-blocking aio_poll, because it won't sleep forever anyway. This is just a matter of checking the "blocking" variable. This requires some changes to the win32 implementation, but is otherwise not too complicated. 3) as mentioned above, the new implementation will not call aio_notify when there is *no* active aio_poll at all. The tests have to be adjusted for this change. The calls to aio_notify in async.c are fine; they only want to kick aio_poll out of a blocking wait, but need not do anything if aio_poll is not running. 4) nested aio_poll: these just work with the new implementation; when a nested event loop is invoked, the outer event loop is never in the prepare or poll phases. The outer event loop thus has already decremented the counter. Reported-by: Richard W. M. Jones <rjones@redhat.com> Reported-by: Laszlo Ersek <lersek@redhat.com> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com> Reviewed-by: Fam Zheng <famz@redhat.com> Tested-by: Richard W.M. Jones <rjones@redhat.com> Message-id: 1437487673-23740-5-git-send-email-pbonzini@redhat.com Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
2015-07-21 17:07:51 +03:00
}
aio_notify_accept(ctx);
/* Adjust polling time */
if (ctx->poll_max_ns) {
int64_t block_ns = qemu_clock_get_ns(QEMU_CLOCK_REALTIME) - start;
if (block_ns <= ctx->poll_ns) {
/* This is the sweet spot, no adjustment needed */
} else if (block_ns > ctx->poll_max_ns) {
/* We'd have to poll for too long, poll less */
int64_t old = ctx->poll_ns;
if (ctx->poll_shrink) {
ctx->poll_ns /= ctx->poll_shrink;
} else {
ctx->poll_ns = 0;
}
trace_poll_shrink(ctx, old, ctx->poll_ns);
} else if (ctx->poll_ns < ctx->poll_max_ns &&
block_ns < ctx->poll_max_ns) {
/* There is room to grow, poll longer */
int64_t old = ctx->poll_ns;
int64_t grow = ctx->poll_grow;
if (grow == 0) {
grow = 2;
}
if (ctx->poll_ns) {
ctx->poll_ns *= grow;
} else {
ctx->poll_ns = 4000; /* start polling at 4 microseconds */
}
if (ctx->poll_ns > ctx->poll_max_ns) {
ctx->poll_ns = ctx->poll_max_ns;
}
trace_poll_grow(ctx, old, ctx->poll_ns);
}
}
progress |= aio_bh_poll(ctx);
aio-posix: split poll check from ready handler Adaptive polling measures the execution time of the polling check plus handlers called when a polled event becomes ready. Handlers can take a significant amount of time, making it look like polling was running for a long time when in fact the event handler was running for a long time. For example, on Linux the io_submit(2) syscall invoked when a virtio-blk device's virtqueue becomes ready can take 10s of microseconds. This can exceed the default polling interval (32 microseconds) and cause adaptive polling to stop polling. By excluding the handler's execution time from the polling check we make the adaptive polling calculation more accurate. As a result, the event loop now stays in polling mode where previously it would have fallen back to file descriptor monitoring. The following data was collected with virtio-blk num-queues=2 event_idx=off using an IOThread. Before: 168k IOPS, IOThread syscalls: 9837.115 ( 0.020 ms): IO iothread1/620155 io_submit(ctx_id: 140512552468480, nr: 16, iocbpp: 0x7fcb9f937db0) = 16 9837.158 ( 0.002 ms): IO iothread1/620155 write(fd: 103, buf: 0x556a2ef71b88, count: 8) = 8 9837.161 ( 0.001 ms): IO iothread1/620155 write(fd: 104, buf: 0x556a2ef71b88, count: 8) = 8 9837.163 ( 0.001 ms): IO iothread1/620155 ppoll(ufds: 0x7fcb90002800, nfds: 4, tsp: 0x7fcb9f1342d0, sigsetsize: 8) = 3 9837.164 ( 0.001 ms): IO iothread1/620155 read(fd: 107, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.174 ( 0.001 ms): IO iothread1/620155 read(fd: 105, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.176 ( 0.001 ms): IO iothread1/620155 read(fd: 106, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.209 ( 0.035 ms): IO iothread1/620155 io_submit(ctx_id: 140512552468480, nr: 32, iocbpp: 0x7fca7d0cebe0) = 32 174k IOPS (+3.6%), IOThread syscalls: 9809.566 ( 0.036 ms): IO iothread1/623061 io_submit(ctx_id: 140539805028352, nr: 32, iocbpp: 0x7fd0cdd62be0) = 32 9809.625 ( 0.001 ms): IO iothread1/623061 write(fd: 103, buf: 0x5647cfba5f58, count: 8) = 8 9809.627 ( 0.002 ms): IO iothread1/623061 write(fd: 104, buf: 0x5647cfba5f58, count: 8) = 8 9809.663 ( 0.036 ms): IO iothread1/623061 io_submit(ctx_id: 140539805028352, nr: 32, iocbpp: 0x7fd0d0388b50) = 32 Notice that ppoll(2) and eventfd read(2) syscalls are eliminated because the IOThread stays in polling mode instead of falling back to file descriptor monitoring. As usual, polling is not implemented on Windows so this patch ignores the new io_poll_read() callback in aio-win32.c. Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> Reviewed-by: Stefano Garzarella <sgarzare@redhat.com> Message-id: 20211207132336.36627-2-stefanha@redhat.com [Fixed up aio_set_event_notifier() calls in tests/unit/test-fdmon-epoll.c added after this series was queued. --Stefan] Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
2021-12-07 16:23:31 +03:00
progress |= aio_dispatch_ready_handlers(ctx, &ready_list);
aio_free_deleted_handlers(ctx);
qemu_lockcnt_dec(&ctx->list_lock);
progress |= timerlistgroup_run_timers(&ctx->tlg);
aio: stop using .io_flush() Now that aio_poll() users check their termination condition themselves, it is no longer necessary to call .io_flush() handlers. The behavior of aio_poll() changes as follows: 1. .io_flush() is no longer invoked and file descriptors are *always* monitored. Previously returning 0 from .io_flush() would skip this file descriptor. Due to this change it is essential to check that requests are pending before calling qemu_aio_wait(). Failure to do so means we block, for example, waiting for an idle iSCSI socket to become readable when there are no requests. Currently all qemu_aio_wait()/aio_poll() callers check before calling. 2. aio_poll() now returns true if progress was made (BH or fd handlers executed) and false otherwise. Previously it would return true whenever 'busy', which means that .io_flush() returned true. The 'busy' concept no longer exists so just progress is returned. Due to this change we need to update tests/test-aio.c which asserts aio_poll() return values. Note that QEMU doesn't actually rely on these return values so only tests/test-aio.c cares. Note that ctx->notifier, the EventNotifier fd used for aio_notify(), is now handled as a special case. This is a little ugly but maintains aio_poll() semantics, i.e. aio_notify() does not count as 'progress' and aio_poll() avoids blocking when the user has not set any fd handlers yet. Patches after this remove .io_flush() handler code until we can finally drop the io_flush arguments to aio_set_fd_handler() and friends. Reviewed-by: Paolo Bonzini <pbonzini@redhat.com> Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
2013-04-11 18:56:50 +04:00
return progress;
}
void aio_context_setup(AioContext *ctx)
{
ctx->fdmon_ops = &fdmon_poll_ops;
ctx->epollfd = -1;
/* Use the fastest fd monitoring implementation if available */
if (fdmon_io_uring_setup(ctx)) {
return;
}
fdmon_epoll_setup(ctx);
}
aio: add polling mode to AioContext The AioContext event loop uses ppoll(2) or epoll_wait(2) to monitor file descriptors or until a timer expires. In cases like virtqueues, Linux AIO, and ThreadPool it is technically possible to wait for events via polling (i.e. continuously checking for events without blocking). Polling can be faster than blocking syscalls because file descriptors, the process scheduler, and system calls are bypassed. The main disadvantage to polling is that it increases CPU utilization. In classic polling configuration a full host CPU thread might run at 100% to respond to events as quickly as possible. This patch implements a timeout so we fall back to blocking syscalls if polling detects no activity. After the timeout no CPU cycles are wasted on polling until the next event loop iteration. The run_poll_handlers_begin() and run_poll_handlers_end() trace events are added to aid performance analysis and troubleshooting. If you need to know whether polling mode is being used, trace these events to find out. Note that the AioContext is now re-acquired before disabling notify_me in the non-polling case. This makes the code cleaner since notify_me was enabled outside the non-polling AioContext release region. This change is correct since it's safe to keep notify_me enabled longer (disabling is an optimization) but potentially causes unnecessary event_notifer_set() calls. I think the chance of performance regression is small here. Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> Reviewed-by: Paolo Bonzini <pbonzini@redhat.com> Message-id: 20161201192652.9509-4-stefanha@redhat.com Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
2016-12-01 22:26:42 +03:00
void aio_context_destroy(AioContext *ctx)
{
fdmon_io_uring_destroy(ctx);
fdmon_epoll_disable(ctx);
aio_free_deleted_handlers(ctx);
}
void aio_context_use_g_source(AioContext *ctx)
{
/*
* Disable io_uring when the glib main loop is used because it doesn't
* support mixed glib/aio_poll() usage. It relies on aio_poll() being
* called regularly so that changes to the monitored file descriptors are
* submitted, otherwise a list of pending fd handlers builds up.
*/
fdmon_io_uring_destroy(ctx);
aio_free_deleted_handlers(ctx);
}
void aio_context_set_poll_params(AioContext *ctx, int64_t max_ns,
int64_t grow, int64_t shrink, Error **errp)
aio: add polling mode to AioContext The AioContext event loop uses ppoll(2) or epoll_wait(2) to monitor file descriptors or until a timer expires. In cases like virtqueues, Linux AIO, and ThreadPool it is technically possible to wait for events via polling (i.e. continuously checking for events without blocking). Polling can be faster than blocking syscalls because file descriptors, the process scheduler, and system calls are bypassed. The main disadvantage to polling is that it increases CPU utilization. In classic polling configuration a full host CPU thread might run at 100% to respond to events as quickly as possible. This patch implements a timeout so we fall back to blocking syscalls if polling detects no activity. After the timeout no CPU cycles are wasted on polling until the next event loop iteration. The run_poll_handlers_begin() and run_poll_handlers_end() trace events are added to aid performance analysis and troubleshooting. If you need to know whether polling mode is being used, trace these events to find out. Note that the AioContext is now re-acquired before disabling notify_me in the non-polling case. This makes the code cleaner since notify_me was enabled outside the non-polling AioContext release region. This change is correct since it's safe to keep notify_me enabled longer (disabling is an optimization) but potentially causes unnecessary event_notifer_set() calls. I think the chance of performance regression is small here. Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> Reviewed-by: Paolo Bonzini <pbonzini@redhat.com> Message-id: 20161201192652.9509-4-stefanha@redhat.com Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
2016-12-01 22:26:42 +03:00
{
/* No thread synchronization here, it doesn't matter if an incorrect value
* is used once.
aio: add polling mode to AioContext The AioContext event loop uses ppoll(2) or epoll_wait(2) to monitor file descriptors or until a timer expires. In cases like virtqueues, Linux AIO, and ThreadPool it is technically possible to wait for events via polling (i.e. continuously checking for events without blocking). Polling can be faster than blocking syscalls because file descriptors, the process scheduler, and system calls are bypassed. The main disadvantage to polling is that it increases CPU utilization. In classic polling configuration a full host CPU thread might run at 100% to respond to events as quickly as possible. This patch implements a timeout so we fall back to blocking syscalls if polling detects no activity. After the timeout no CPU cycles are wasted on polling until the next event loop iteration. The run_poll_handlers_begin() and run_poll_handlers_end() trace events are added to aid performance analysis and troubleshooting. If you need to know whether polling mode is being used, trace these events to find out. Note that the AioContext is now re-acquired before disabling notify_me in the non-polling case. This makes the code cleaner since notify_me was enabled outside the non-polling AioContext release region. This change is correct since it's safe to keep notify_me enabled longer (disabling is an optimization) but potentially causes unnecessary event_notifer_set() calls. I think the chance of performance regression is small here. Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> Reviewed-by: Paolo Bonzini <pbonzini@redhat.com> Message-id: 20161201192652.9509-4-stefanha@redhat.com Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
2016-12-01 22:26:42 +03:00
*/
ctx->poll_max_ns = max_ns;
ctx->poll_ns = 0;
ctx->poll_grow = grow;
ctx->poll_shrink = shrink;
aio: add polling mode to AioContext The AioContext event loop uses ppoll(2) or epoll_wait(2) to monitor file descriptors or until a timer expires. In cases like virtqueues, Linux AIO, and ThreadPool it is technically possible to wait for events via polling (i.e. continuously checking for events without blocking). Polling can be faster than blocking syscalls because file descriptors, the process scheduler, and system calls are bypassed. The main disadvantage to polling is that it increases CPU utilization. In classic polling configuration a full host CPU thread might run at 100% to respond to events as quickly as possible. This patch implements a timeout so we fall back to blocking syscalls if polling detects no activity. After the timeout no CPU cycles are wasted on polling until the next event loop iteration. The run_poll_handlers_begin() and run_poll_handlers_end() trace events are added to aid performance analysis and troubleshooting. If you need to know whether polling mode is being used, trace these events to find out. Note that the AioContext is now re-acquired before disabling notify_me in the non-polling case. This makes the code cleaner since notify_me was enabled outside the non-polling AioContext release region. This change is correct since it's safe to keep notify_me enabled longer (disabling is an optimization) but potentially causes unnecessary event_notifer_set() calls. I think the chance of performance regression is small here. Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> Reviewed-by: Paolo Bonzini <pbonzini@redhat.com> Message-id: 20161201192652.9509-4-stefanha@redhat.com Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
2016-12-01 22:26:42 +03:00
aio_notify(ctx);
}
void aio_context_set_aio_params(AioContext *ctx, int64_t max_batch,
Error **errp)
{
/*
* No thread synchronization here, it doesn't matter if an incorrect value
* is used once.
*/
ctx->aio_max_batch = max_batch;
aio_notify(ctx);
}