[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[Qemu-devel] [PULL v3 29/38] coroutine: rewrite pool to avoid mutex
From: |
Stefan Hajnoczi |
Subject: |
[Qemu-devel] [PULL v3 29/38] coroutine: rewrite pool to avoid mutex |
Date: |
Tue, 13 Jan 2015 13:48:07 +0000 |
From: Paolo Bonzini <address@hidden>
This patch removes the mutex by using fancy lock-free manipulation of
the pool. Lock-free stacks and queues are not hard, but they can suffer
from the ABA problem so they are better avoided unless you have some
deferred reclamation scheme like RCU. Otherwise you have to stick
with adding to a list, and emptying it completely. This is what this
patch does, by coupling a lock-free global list of available coroutines
with per-CPU lists that are actually used on coroutine creation.
Whenever the destruction pool is big enough, the next thread that runs
out of coroutines will steal the whole destruction pool. This is positive
in two ways:
1) the allocation does not have to do any atomic operation in the fast
path, it's entirely using thread-local storage. Once every POOL_BATCH_SIZE
allocations it will do a single atomic_xchg. Release does an atomic_cmpxchg
loop, that hopefully doesn't cause any starvation, and an atomic_inc.
A later patch will also remove atomic operations from the release path,
and try to avoid the atomic_xchg altogether---succeeding in doing so if
all devices either use ioeventfd or are not submitting requests actively.
2) in theory this should be completely adaptive. The number of coroutines
around should be a little more than POOL_BATCH_SIZE * number of allocating
threads; so this also empties qemu_coroutine_adjust_pool_size. (The previous
pool size was POOL_BATCH_SIZE * number of block backends, so it was a bit
more generous. But if you actually have many high-iodepth disks, it's better
to put them in different iothreads, which will also use separate thread
pools and aio=native file descriptors).
This speeds up perf/cost (in tests/test-coroutine) by a factor of ~1.33.
No matter if we end with some kind of coroutine bypass scheme or not,
it cannot hurt to optimize hot code.
Signed-off-by: Paolo Bonzini <address@hidden>
Reviewed-by: Fam Zheng <address@hidden>
Message-id: address@hidden
Signed-off-by: Stefan Hajnoczi <address@hidden>
---
qemu-coroutine.c | 92 +++++++++++++++++++++++++-------------------------------
1 file changed, 41 insertions(+), 51 deletions(-)
diff --git a/qemu-coroutine.c b/qemu-coroutine.c
index bd574aa..93fddc7 100644
--- a/qemu-coroutine.c
+++ b/qemu-coroutine.c
@@ -15,31 +15,57 @@
#include "trace.h"
#include "qemu-common.h"
#include "qemu/thread.h"
+#include "qemu/atomic.h"
#include "block/coroutine.h"
#include "block/coroutine_int.h"
enum {
- POOL_DEFAULT_SIZE = 64,
+ POOL_BATCH_SIZE = 64,
};
/** Free list to speed up creation */
-static QemuMutex pool_lock;
-static QSLIST_HEAD(, Coroutine) pool = QSLIST_HEAD_INITIALIZER(pool);
-static unsigned int pool_size;
-static unsigned int pool_max_size = POOL_DEFAULT_SIZE;
+static QSLIST_HEAD(, Coroutine) release_pool = QSLIST_HEAD_INITIALIZER(pool);
+static unsigned int release_pool_size;
+static __thread QSLIST_HEAD(, Coroutine) alloc_pool =
QSLIST_HEAD_INITIALIZER(pool);
+static __thread Notifier coroutine_pool_cleanup_notifier;
+
+static void coroutine_pool_cleanup(Notifier *n, void *value)
+{
+ Coroutine *co;
+ Coroutine *tmp;
+
+ QSLIST_FOREACH_SAFE(co, &alloc_pool, pool_next, tmp) {
+ QSLIST_REMOVE_HEAD(&alloc_pool, pool_next);
+ qemu_coroutine_delete(co);
+ }
+}
Coroutine *qemu_coroutine_create(CoroutineEntry *entry)
{
Coroutine *co = NULL;
if (CONFIG_COROUTINE_POOL) {
- qemu_mutex_lock(&pool_lock);
- co = QSLIST_FIRST(&pool);
+ co = QSLIST_FIRST(&alloc_pool);
+ if (!co) {
+ if (release_pool_size > POOL_BATCH_SIZE) {
+ /* Slow path; a good place to register the destructor, too. */
+ if (!coroutine_pool_cleanup_notifier.notify) {
+ coroutine_pool_cleanup_notifier.notify =
coroutine_pool_cleanup;
+ qemu_thread_atexit_add(&coroutine_pool_cleanup_notifier);
+ }
+
+ /* This is not exact; there could be a little skew between
+ * release_pool_size and the actual size of release_pool. But
+ * it is just a heuristic, it does not need to be perfect.
+ */
+ release_pool_size = 0;
+ QSLIST_MOVE_ATOMIC(&alloc_pool, &release_pool);
+ co = QSLIST_FIRST(&alloc_pool);
+ }
+ }
if (co) {
- QSLIST_REMOVE_HEAD(&pool, pool_next);
- pool_size--;
+ QSLIST_REMOVE_HEAD(&alloc_pool, pool_next);
}
- qemu_mutex_unlock(&pool_lock);
}
if (!co) {
@@ -53,39 +79,19 @@ Coroutine *qemu_coroutine_create(CoroutineEntry *entry)
static void coroutine_delete(Coroutine *co)
{
+ co->caller = NULL;
+
if (CONFIG_COROUTINE_POOL) {
- qemu_mutex_lock(&pool_lock);
- if (pool_size < pool_max_size) {
- QSLIST_INSERT_HEAD(&pool, co, pool_next);
- co->caller = NULL;
- pool_size++;
- qemu_mutex_unlock(&pool_lock);
+ if (release_pool_size < POOL_BATCH_SIZE * 2) {
+ QSLIST_INSERT_HEAD_ATOMIC(&release_pool, co, pool_next);
+ atomic_inc(&release_pool_size);
return;
}
- qemu_mutex_unlock(&pool_lock);
}
qemu_coroutine_delete(co);
}
-static void __attribute__((constructor)) coroutine_pool_init(void)
-{
- qemu_mutex_init(&pool_lock);
-}
-
-static void __attribute__((destructor)) coroutine_pool_cleanup(void)
-{
- Coroutine *co;
- Coroutine *tmp;
-
- QSLIST_FOREACH_SAFE(co, &pool, pool_next, tmp) {
- QSLIST_REMOVE_HEAD(&pool, pool_next);
- qemu_coroutine_delete(co);
- }
-
- qemu_mutex_destroy(&pool_lock);
-}
-
static void coroutine_swap(Coroutine *from, Coroutine *to)
{
CoroutineAction ret;
@@ -140,20 +146,4 @@ void coroutine_fn qemu_coroutine_yield(void)
void qemu_coroutine_adjust_pool_size(int n)
{
- qemu_mutex_lock(&pool_lock);
-
- pool_max_size += n;
-
- /* Callers should never take away more than they added */
- assert(pool_max_size >= POOL_DEFAULT_SIZE);
-
- /* Trim oversized pool down to new max */
- while (pool_size > pool_max_size) {
- Coroutine *co = QSLIST_FIRST(&pool);
- QSLIST_REMOVE_HEAD(&pool, pool_next);
- pool_size--;
- qemu_coroutine_delete(co);
- }
-
- qemu_mutex_unlock(&pool_lock);
}
--
2.1.0
- [Qemu-devel] [PULL v3 17/38] qemu-iotests: Test blockdev-backup in 055, (continued)
- [Qemu-devel] [PULL v3 17/38] qemu-iotests: Test blockdev-backup in 055, Stefan Hajnoczi, 2015/01/13
- [Qemu-devel] [PULL v3 18/38] iotests: Filter out "I/O thread spun..." warning, Stefan Hajnoczi, 2015/01/13
- [Qemu-devel] [PULL v3 21/38] .gitignore: Ignore generated "common.env", Stefan Hajnoczi, 2015/01/13
- [Qemu-devel] [PULL v3 19/38] migration/block: fix pending() return value, Stefan Hajnoczi, 2015/01/13
- [Qemu-devel] [PULL v3 22/38] qemu-iotests: Replace "/bin/true" with "true", Stefan Hajnoczi, 2015/01/13
- [Qemu-devel] [PULL v3 23/38] qemu-iotests: Add "_supported_os Linux" to 058, Stefan Hajnoczi, 2015/01/13
- [Qemu-devel] [PULL v3 20/38] libqos: Convert malloc-pc allocator to a generic allocator, Stefan Hajnoczi, 2015/01/13
- [Qemu-devel] [PULL v3 25/38] coroutine-ucontext: use __thread, Stefan Hajnoczi, 2015/01/13
- [Qemu-devel] [PULL v3 24/38] qemu-iotests: Add supported os parameter for python tests, Stefan Hajnoczi, 2015/01/13
- [Qemu-devel] [PULL v3 26/38] qemu-thread: add per-thread atexit functions, Stefan Hajnoczi, 2015/01/13
- [Qemu-devel] [PULL v3 29/38] coroutine: rewrite pool to avoid mutex,
Stefan Hajnoczi <=
- [Qemu-devel] [PULL v3 30/38] coroutine: drop qemu_coroutine_adjust_pool_size, Stefan Hajnoczi, 2015/01/13
- [Qemu-devel] [PULL v3 27/38] test-coroutine: avoid overflow on 32-bit systems, Stefan Hajnoczi, 2015/01/13
- [Qemu-devel] [PULL v3 28/38] QSLIST: add lock-free operations, Stefan Hajnoczi, 2015/01/13
- [Qemu-devel] [PULL v3 32/38] block: limited request size in write zeroes unsupported path, Stefan Hajnoczi, 2015/01/13
- [Qemu-devel] [PULL v3 33/38] block: Split BLOCK_OP_TYPE_COMMIT to BLOCK_OP_TYPE_COMMIT_{SOURCE, TARGET}, Stefan Hajnoczi, 2015/01/13
- [Qemu-devel] [PULL v3 31/38] coroutine: try harder not to delete coroutines, Stefan Hajnoczi, 2015/01/13
- [Qemu-devel] [PULL v3 34/38] ide: Implement VPD response for ATAPI, Stefan Hajnoczi, 2015/01/13
- [Qemu-devel] [PULL v3 35/38] nvme: Fix get/set number of queues feature, Stefan Hajnoczi, 2015/01/13
- [Qemu-devel] [PULL v3 38/38] NVMe: Set correct VS Value for 1.1 Compliant Controllers, Stefan Hajnoczi, 2015/01/13
- [Qemu-devel] [PULL v3 36/38] MAINTAINERS: Update email addresses for Chrysostomos Nanakos, Stefan Hajnoczi, 2015/01/13