qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Paolo Bonzini <pbonzini@redhat.com>
To: qemu-devel@nongnu.org
Cc: stefanha@redhat.com, famz@redhat.com, kwolf@redhat.com, berto@igalia.com
Subject: [Qemu-devel] [PATCH 11/11] coroutine-lock: make CoRwlock thread-safe and fair
Date: Fri, 15 Apr 2016 13:32:06 +0200	[thread overview]
Message-ID: <1460719926-12950-12-git-send-email-pbonzini@redhat.com> (raw)
In-Reply-To: <1460719926-12950-1-git-send-email-pbonzini@redhat.com>

This adds a CoMutex around the existing CoQueue.  Because the write-side
can just take CoMutex, the old "writer" field is not necessary anymore.
Instead of removing it altogether, count the number of pending writers
during a read-side critical section and forbid further readers from
entering.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
---
 include/qemu/coroutine.h   |  3 ++-
 util/qemu-coroutine-lock.c | 29 +++++++++++++++++++++--------
 2 files changed, 23 insertions(+), 9 deletions(-)

diff --git a/include/qemu/coroutine.h b/include/qemu/coroutine.h
index 25e31a1..0dfb301 100644
--- a/include/qemu/coroutine.h
+++ b/include/qemu/coroutine.h
@@ -198,8 +198,9 @@ bool qemu_co_queue_empty(CoQueue *queue);
 
 
 typedef struct CoRwlock {
-    bool writer;
+    int pending_writer;
     int reader;
+    CoMutex mutex;
     CoQueue queue;
 } CoRwlock;
 
diff --git a/util/qemu-coroutine-lock.c b/util/qemu-coroutine-lock.c
index 828d79a..db9cf01 100644
--- a/util/qemu-coroutine-lock.c
+++ b/util/qemu-coroutine-lock.c
@@ -322,36 +322,49 @@ void qemu_co_rwlock_init(CoRwlock *lock)
 {
     memset(lock, 0, sizeof(*lock));
     qemu_co_queue_init(&lock->queue);
+    qemu_co_mutex_init(&lock->mutex);
 }
 
 void qemu_co_rwlock_rdlock(CoRwlock *lock)
 {
-    while (lock->writer) {
-        qemu_co_queue_wait(&lock->queue, NULL);
+    qemu_co_mutex_lock(&lock->mutex);
+    /* For fairness, wait if a writer is in line.  */
+    while (lock->pending_writer) {
+        qemu_co_queue_wait(&lock->queue, &lock->mutex);
     }
     lock->reader++;
+    qemu_co_mutex_unlock(&lock->mutex);
+
+    /* The rest of the read-side critical section is run without the mutex.  */
 }
 
 void qemu_co_rwlock_unlock(CoRwlock *lock)
 {
     assert(qemu_in_coroutine());
-    if (lock->writer) {
-        lock->writer = false;
+    if (!lock->reader) {
+        /* The critical section started in qemu_co_rwlock_wrlock.  */
         qemu_co_queue_restart_all(&lock->queue);
     } else {
+        qemu_co_mutex_lock(&lock->mutex);
         lock->reader--;
-        assert(lock->reader >= 0);
         /* Wakeup only one waiting writer */
         if (!lock->reader) {
             qemu_co_queue_next(&lock->queue);
         }
     }
+    qemu_co_mutex_unlock(&lock->mutex);
 }
 
 void qemu_co_rwlock_wrlock(CoRwlock *lock)
 {
-    while (lock->writer || lock->reader) {
-        qemu_co_queue_wait(&lock->queue, NULL);
+    qemu_co_mutex_lock(&lock->mutex);
+    lock->pending_writer++;
+    while (lock->reader) {
+        qemu_co_queue_wait(&lock->queue, &lock->mutex);
     }
-    lock->writer = true;
+    lock->pending_writer--;
+
+    /* The rest of the write-side critical section is run with
+     * the mutex taken, so lock->reader remains zero.
+     */
 }
-- 
2.5.5

  parent reply	other threads:[~2016-04-15 11:32 UTC|newest]

Thread overview: 27+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2016-04-15 11:31 [Qemu-devel] [RFC PATCH resend 00/11] Make CoMutex/CoQueue/CoRwlock thread-safe Paolo Bonzini
2016-04-15 11:31 ` [Qemu-devel] [PATCH 01/11] coroutine: use QSIMPLEQ instead of QTAILQ Paolo Bonzini
2016-04-19 13:45   ` Stefan Hajnoczi
2016-04-15 11:31 ` [Qemu-devel] [PATCH 02/11] throttle-groups: restart throttled requests from coroutine context Paolo Bonzini
2016-04-19 13:49   ` Stefan Hajnoczi
2016-04-15 11:31 ` [Qemu-devel] [PATCH 03/11] coroutine: delete qemu_co_enter_next Paolo Bonzini
2016-04-19 13:49   ` Stefan Hajnoczi
2016-04-15 11:31 ` [Qemu-devel] [PATCH 04/11] aio: introduce aio_co_schedule Paolo Bonzini
2016-04-19 14:31   ` Stefan Hajnoczi
2016-05-17 14:57     ` Paolo Bonzini
2016-05-26 19:19       ` Stefan Hajnoczi
2016-04-29  5:11   ` Fam Zheng
2016-05-17 14:38     ` Paolo Bonzini
2016-04-15 11:32 ` [Qemu-devel] [PATCH 05/11] coroutine-lock: reschedule coroutine on the AioContext it was running on Paolo Bonzini
2016-04-15 11:32 ` [Qemu-devel] [PATCH 06/11] coroutine-lock: make CoMutex thread-safe Paolo Bonzini
2016-04-29  6:26   ` Fam Zheng
2016-05-17 15:34     ` Paolo Bonzini
2016-04-15 11:32 ` [Qemu-devel] [PATCH 07/11] coroutine-lock: add limited spinning to CoMutex Paolo Bonzini
2016-04-15 11:32 ` [Qemu-devel] [PATCH 08/11] test-aio-multithread: add performance comparison with thread-based mutexes Paolo Bonzini
2016-04-29  6:52   ` Fam Zheng
2016-05-12 16:49     ` Paolo Bonzini
2016-04-15 11:32 ` [Qemu-devel] [PATCH 09/11] coroutine-lock: place CoMutex before CoQueue in header Paolo Bonzini
2016-04-15 11:32 ` [Qemu-devel] [PATCH 10/11] coroutine-lock: add mutex argument to CoQueue APIs Paolo Bonzini
2016-04-15 11:32 ` Paolo Bonzini [this message]
2016-04-26 10:54 ` [Qemu-devel] [RFC PATCH resend 00/11] Make CoMutex/CoQueue/CoRwlock thread-safe Stefan Hajnoczi
2016-04-27 15:42   ` Stefan Hajnoczi
2016-05-17 15:34   ` Paolo Bonzini

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=1460719926-12950-12-git-send-email-pbonzini@redhat.com \
    --to=pbonzini@redhat.com \
    --cc=berto@igalia.com \
    --cc=famz@redhat.com \
    --cc=kwolf@redhat.com \
    --cc=qemu-devel@nongnu.org \
    --cc=stefanha@redhat.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).