From: Eric Blake <eblake@redhat.com>
To: qemu-devel@nongnu.org
Cc: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>,
"open list:Network Block Dev..." <qemu-block@nongnu.org>
Subject: [PULL 20/34] nbd/client-connection: use QEMU_LOCK_GUARD
Date: Tue, 15 Jun 2021 15:47:42 -0500 [thread overview]
Message-ID: <20210615204756.281505-21-eblake@redhat.com> (raw)
In-Reply-To: <20210615204756.281505-1-eblake@redhat.com>
From: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
We don't update connect_thread_func() to use QEMU_LOCK_GUARD, as it
will get more complex critical sections logic in further commit, where
QEMU_LOCK_GUARD doesn't help.
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Message-Id: <20210610100802.5888-19-vsementsov@virtuozzo.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
---
nbd/client-connection.c | 95 +++++++++++++++++++----------------------
1 file changed, 43 insertions(+), 52 deletions(-)
diff --git a/nbd/client-connection.c b/nbd/client-connection.c
index f3b270d38cc8..eb5cae2eaea4 100644
--- a/nbd/client-connection.c
+++ b/nbd/client-connection.c
@@ -125,14 +125,14 @@ void nbd_client_connection_release(NBDClientConnection *conn)
return;
}
- qemu_mutex_lock(&conn->mutex);
- assert(!conn->detached);
- if (conn->running) {
- conn->detached = true;
- } else {
- do_free = true;
+ WITH_QEMU_LOCK_GUARD(&conn->mutex) {
+ assert(!conn->detached);
+ if (conn->running) {
+ conn->detached = true;
+ } else {
+ do_free = true;
+ }
}
- qemu_mutex_unlock(&conn->mutex);
if (do_free) {
nbd_client_connection_do_free(conn);
@@ -150,62 +150,55 @@ void nbd_client_connection_release(NBDClientConnection *conn)
QIOChannelSocket *coroutine_fn
nbd_co_establish_connection(NBDClientConnection *conn, Error **errp)
{
- QIOChannelSocket *sioc = NULL;
QemuThread thread;
- qemu_mutex_lock(&conn->mutex);
+ WITH_QEMU_LOCK_GUARD(&conn->mutex) {
+ /*
+ * Don't call nbd_co_establish_connection() in several coroutines in
+ * parallel. Only one call at once is supported.
+ */
+ assert(!conn->wait_co);
- /*
- * Don't call nbd_co_establish_connection() in several coroutines in
- * parallel. Only one call at once is supported.
- */
- assert(!conn->wait_co);
+ if (!conn->running) {
+ if (conn->sioc) {
+ /* Previous attempt finally succeeded in background */
+ return g_steal_pointer(&conn->sioc);
+ }
- if (!conn->running) {
- if (conn->sioc) {
- /* Previous attempt finally succeeded in background */
- sioc = g_steal_pointer(&conn->sioc);
- qemu_mutex_unlock(&conn->mutex);
-
- return sioc;
+ conn->running = true;
+ error_free(conn->err);
+ conn->err = NULL;
+ qemu_thread_create(&thread, "nbd-connect",
+ connect_thread_func, conn, QEMU_THREAD_DETACHED);
}
- conn->running = true;
- error_free(conn->err);
- conn->err = NULL;
- qemu_thread_create(&thread, "nbd-connect",
- connect_thread_func, conn, QEMU_THREAD_DETACHED);
+ conn->wait_co = qemu_coroutine_self();
}
- conn->wait_co = qemu_coroutine_self();
-
- qemu_mutex_unlock(&conn->mutex);
-
/*
* We are going to wait for connect-thread finish, but
* nbd_co_establish_connection_cancel() can interrupt.
*/
qemu_coroutine_yield();
- qemu_mutex_lock(&conn->mutex);
-
- if (conn->running) {
- /*
- * The connection attempt was canceled and the coroutine resumed
- * before the connection thread finished its job. Report the
- * attempt as failed, but leave the connection thread running,
- * to reuse it for the next connection attempt.
- */
- error_setg(errp, "Connection attempt cancelled by other operation");
- } else {
- error_propagate(errp, conn->err);
- conn->err = NULL;
- sioc = g_steal_pointer(&conn->sioc);
+ WITH_QEMU_LOCK_GUARD(&conn->mutex) {
+ if (conn->running) {
+ /*
+ * The connection attempt was canceled and the coroutine resumed
+ * before the connection thread finished its job. Report the
+ * attempt as failed, but leave the connection thread running,
+ * to reuse it for the next connection attempt.
+ */
+ error_setg(errp, "Connection attempt cancelled by other operation");
+ return NULL;
+ } else {
+ error_propagate(errp, conn->err);
+ conn->err = NULL;
+ return g_steal_pointer(&conn->sioc);
+ }
}
- qemu_mutex_unlock(&conn->mutex);
-
- return sioc;
+ abort(); /* unreachable */
}
/*
@@ -220,11 +213,9 @@ void nbd_co_establish_connection_cancel(NBDClientConnection *conn)
{
Coroutine *wait_co;
- qemu_mutex_lock(&conn->mutex);
-
- wait_co = g_steal_pointer(&conn->wait_co);
-
- qemu_mutex_unlock(&conn->mutex);
+ WITH_QEMU_LOCK_GUARD(&conn->mutex) {
+ wait_co = g_steal_pointer(&conn->wait_co);
+ }
if (wait_co) {
aio_co_wake(wait_co);
--
2.31.1
next prev parent reply other threads:[~2021-06-15 21:01 UTC|newest]
Thread overview: 39+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-06-15 20:47 [PULL 00/34] NBD patches for 2021-06-15 Eric Blake
2021-06-15 20:47 ` [PULL 01/34] async: the main AioContext is only "current" if under the BQL Eric Blake
2021-06-15 20:47 ` [PULL 02/34] tests: cover aio_co_enter from a worker thread without BQL taken Eric Blake
2021-06-15 20:47 ` [PULL 03/34] co-queue: drop extra coroutine_fn marks Eric Blake
2021-06-15 20:47 ` [PULL 04/34] block/nbd: fix channel object leak Eric Blake
2021-06-15 20:47 ` [PULL 05/34] block/nbd: fix how state is cleared on nbd_open() failure paths Eric Blake
2021-06-15 20:47 ` [PULL 06/34] block/nbd: connect_thread_func(): do qio_channel_set_delay(false) Eric Blake
2021-06-15 20:47 ` [PULL 07/34] qemu-sockets: introduce socket_address_parse_named_fd() Eric Blake
2021-06-15 20:47 ` [PULL 08/34] block/nbd: call socket_address_parse_named_fd() in advance Eric Blake
2021-06-15 20:47 ` [PULL 09/34] block/nbd: ensure ->connection_thread is always valid Eric Blake
2021-06-15 20:47 ` [PULL 10/34] block/nbd: nbd_client_handshake(): fix leak of s->ioc Eric Blake
2021-06-15 20:47 ` [PULL 11/34] block/nbd: BDRVNBDState: drop unused connect_err and connect_status Eric Blake
2021-06-15 20:47 ` [PULL 12/34] block/nbd: simplify waking of nbd_co_establish_connection() Eric Blake
2021-06-15 20:47 ` [PULL 13/34] block/nbd: drop thr->state Eric Blake
2021-06-15 20:47 ` [PULL 14/34] block/nbd: bs-independent interface for nbd_co_establish_connection() Eric Blake
2021-06-15 20:47 ` [PULL 15/34] block/nbd: make nbd_co_establish_connection_cancel() bs-independent Eric Blake
2021-06-15 20:47 ` [PULL 16/34] block/nbd: rename NBDConnectThread to NBDClientConnection Eric Blake
2021-06-15 20:47 ` [PULL 17/34] block/nbd: introduce nbd_client_connection_new() Eric Blake
2021-06-15 20:47 ` [PULL 18/34] block/nbd: introduce nbd_client_connection_release() Eric Blake
2021-06-15 20:47 ` [PULL 19/34] nbd: move connection code from block/nbd to nbd/client-connection Eric Blake
2021-06-15 20:47 ` Eric Blake [this message]
2021-06-15 20:47 ` [PULL 21/34] nbd/client-connection: add possibility of negotiation Eric Blake
2021-06-15 20:47 ` [PULL 22/34] nbd/client-connection: implement connection retry Eric Blake
2021-06-15 20:47 ` [PULL 23/34] nbd/client-connection: shutdown connection on release Eric Blake
2021-06-15 20:47 ` [PULL 24/34] block/nbd: split nbd_handle_updated_info out of nbd_client_handshake() Eric Blake
2021-06-15 20:47 ` [PULL 25/34] block/nbd: use negotiation of NBDClientConnection Eric Blake
2021-06-15 20:47 ` [PULL 26/34] block/nbd: don't touch s->sioc in nbd_teardown_connection() Eric Blake
2021-06-15 20:47 ` [PULL 27/34] block/nbd: drop BDRVNBDState::sioc Eric Blake
2021-06-15 20:47 ` [PULL 28/34] nbd/client-connection: return only one io channel Eric Blake
2021-06-17 18:32 ` Vladimir Sementsov-Ogievskiy
2021-06-18 15:55 ` Eric Blake
2021-06-15 20:47 ` [PULL 29/34] block-coroutine-wrapper: allow non bdrv_ prefix Eric Blake
2021-06-15 20:47 ` [PULL 30/34] block/nbd: split nbd_co_do_establish_connection out of nbd_reconnect_attempt Eric Blake
2021-06-15 20:47 ` [PULL 31/34] nbd/client-connection: add option for non-blocking connection attempt Eric Blake
2021-06-15 20:47 ` [PULL 32/34] block/nbd: reuse nbd_co_do_establish_connection() in nbd_open() Eric Blake
2021-06-15 20:47 ` [PULL 33/34] block/nbd: add nbd_client_connected() helper Eric Blake
2021-06-15 20:47 ` [PULL 34/34] block/nbd: safer transition to receiving request Eric Blake
2021-06-17 9:42 ` [PULL 00/34] NBD patches for 2021-06-15 Peter Maydell
2021-06-17 18:35 ` Vladimir Sementsov-Ogievskiy
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=20210615204756.281505-21-eblake@redhat.com \
--to=eblake@redhat.com \
--cc=qemu-block@nongnu.org \
--cc=qemu-devel@nongnu.org \
--cc=vsementsov@virtuozzo.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).