From nobody Wed Nov 27 00:39:13 2024 Delivered-To: importer@patchew.org Authentication-Results: mx.zohomail.com; dkim=pass; spf=pass (zohomail.com: domain of gnu.org designates 209.51.188.17 as permitted sender) smtp.mailfrom=qemu-devel-bounces+importer=patchew.org@nongnu.org; dmarc=pass(p=none dis=none) header.from=redhat.com ARC-Seal: i=1; a=rsa-sha256; t=1703186775; cv=none; d=zohomail.com; s=zohoarc; b=XnTCgBj7VyDbiOGVKl4Ya0TXiHVjH4pfJ6gjHwyWquKp11e2BPTo/yskPNwN1GD4bbeUiWUuv6uqyvCXZrZetMs9gtICE+BwSzlV6IB2C7WnaUhp49BkDkEiM479HkRvO4BHR+coBXk9CUTPzMwzgKjmyATJ+tgv01YeCf2bnBE= ARC-Message-Signature: i=1; a=rsa-sha256; c=relaxed/relaxed; d=zohomail.com; s=zohoarc; t=1703186775; h=Content-Transfer-Encoding:Cc:Cc:Date:Date:From:From:In-Reply-To:List-Subscribe:List-Post:List-Id:List-Archive:List-Help:List-Unsubscribe:MIME-Version:Message-ID:References:Sender:Subject:Subject:To:To:Message-Id:Reply-To; bh=wy0gnLiaJ8egKg9TrWO6YxGfb6MtQIx7kCPRsGcYHFM=; b=Qq6t7s/BMrNteJNRpkRQzBeR2sQL3ttmVgAVPxaR3yn/Wag6xpZRp7pnkptp0v1hYDaCkYwtwgZvhO1ddCYwhKMrcu3mZdcXzkBGlB7ocQUxjd1oLxLXtA8ii8dY4vbiA40cnr9iftM3kPizrJpj+zDdBW4JLkv2r/qZeynQ5eo= ARC-Authentication-Results: i=1; mx.zohomail.com; dkim=pass; spf=pass (zohomail.com: domain of gnu.org designates 209.51.188.17 as permitted sender) smtp.mailfrom=qemu-devel-bounces+importer=patchew.org@nongnu.org; dmarc=pass header.from= (p=none dis=none) Return-Path: Received: from lists.gnu.org (lists.gnu.org [209.51.188.17]) by mx.zohomail.com with SMTPS id 170318677549974.52881628062698; Thu, 21 Dec 2023 11:26:15 -0800 (PST) Received: from localhost ([::1] helo=lists1p.gnu.org) by lists.gnu.org with esmtp (Exim 4.90_1) (envelope-from ) id 1rGOfZ-0001TG-HH; Thu, 21 Dec 2023 14:25:21 -0500 Received: from eggs.gnu.org ([2001:470:142:3::10]) by lists.gnu.org with esmtps (TLS1.2:ECDHE_RSA_AES_256_GCM_SHA384:256) (Exim 4.90_1) (envelope-from ) id 1rGOfX-0001Rq-47 for qemu-devel@nongnu.org; Thu, 21 Dec 2023 14:25:19 -0500 Received: from us-smtp-delivery-124.mimecast.com ([170.10.129.124]) by eggs.gnu.org with esmtps (TLS1.2:ECDHE_RSA_AES_256_GCM_SHA384:256) (Exim 4.90_1) (envelope-from ) id 1rGOfP-00023R-0t for qemu-devel@nongnu.org; Thu, 21 Dec 2023 14:25:18 -0500 Received: from mimecast-mx02.redhat.com (mx-ext.redhat.com [66.187.233.73]) by relay.mimecast.com with ESMTP with STARTTLS (version=TLSv1.3, cipher=TLS_AES_256_GCM_SHA384) id us-mta-647-Vs4iSQtOPXWfa5GYXvhDWg-1; Thu, 21 Dec 2023 14:25:08 -0500 Received: from smtp.corp.redhat.com (int-mx05.intmail.prod.int.rdu2.redhat.com [10.11.54.5]) (using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits) key-exchange X25519 server-signature RSA-PSS (2048 bits) server-digest SHA256) (No client certificate requested) by mimecast-mx02.redhat.com (Postfix) with ESMTPS id 11B40380008B; Thu, 21 Dec 2023 19:25:08 +0000 (UTC) Received: from localhost (unknown [10.39.194.145]) by smtp.corp.redhat.com (Postfix) with ESMTP id 1F25251D5; Thu, 21 Dec 2023 19:25:06 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=redhat.com; s=mimecast20190719; t=1703186709; h=from:from:reply-to:subject:subject:date:date:message-id:message-id: to:to:cc:cc:mime-version:mime-version: content-transfer-encoding:content-transfer-encoding: in-reply-to:in-reply-to:references:references; bh=wy0gnLiaJ8egKg9TrWO6YxGfb6MtQIx7kCPRsGcYHFM=; b=fs6K5NmBsaWYtOOV9B4S5+4mQpixTqxDisJRElARmcAeKVygH125xNjiYwap4liYjJ0xIc B/c2Xxhi9A0+Kkm4jfLdHaMke0XvyeoEuYPHLnXPkPj63pE5BrPl0fHAzGwCr6ozJ5ePRQ yIcLLVTX8a5yZcNuqNw/L2aDA0U2c4s= X-MC-Unique: Vs4iSQtOPXWfa5GYXvhDWg-1 From: Stefan Hajnoczi To: qemu-devel@nongnu.org, Kevin Wolf Cc: Juan Quintela , Vladimir Sementsov-Ogievskiy , pbonzini@redhat.com, Peter Xu , qemu-block@nongnu.org, Hanna Reitz , Fabiano Rosas , Leonardo Bras , Eric Blake , Stefan Hajnoczi , Fam Zheng Subject: [PATCH v3 6/6] nbd/server: introduce NBDClient->lock to protect fields Date: Thu, 21 Dec 2023 14:24:52 -0500 Message-ID: <20231221192452.1785567-7-stefanha@redhat.com> In-Reply-To: <20231221192452.1785567-1-stefanha@redhat.com> References: <20231221192452.1785567-1-stefanha@redhat.com> MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable X-Scanned-By: MIMEDefang 3.4.1 on 10.11.54.5 Received-SPF: pass (zohomail.com: domain of gnu.org designates 209.51.188.17 as permitted sender) client-ip=209.51.188.17; envelope-from=qemu-devel-bounces+importer=patchew.org@nongnu.org; helo=lists.gnu.org; Received-SPF: pass client-ip=170.10.129.124; envelope-from=stefanha@redhat.com; helo=us-smtp-delivery-124.mimecast.com X-Spam_score_int: -21 X-Spam_score: -2.2 X-Spam_bar: -- X-Spam_report: (-2.2 / 5.0 requ) BAYES_00=-1.9, DKIMWL_WL_HIGH=-0.061, DKIM_SIGNED=0.1, DKIM_VALID=-0.1, DKIM_VALID_AU=-0.1, DKIM_VALID_EF=-0.1, RCVD_IN_DNSWL_NONE=-0.0001, RCVD_IN_MSPIKE_H4=0.001, RCVD_IN_MSPIKE_WL=0.001, SPF_HELO_NONE=0.001, SPF_PASS=-0.001, T_SCC_BODY_TEXT_LINE=-0.01 autolearn=ham autolearn_force=no X-Spam_action: no action X-BeenThere: qemu-devel@nongnu.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: qemu-devel-bounces+importer=patchew.org@nongnu.org Sender: qemu-devel-bounces+importer=patchew.org@nongnu.org X-ZohoMail-DKIM: pass (identity @redhat.com) X-ZM-MESSAGEID: 1703186776891100003 Content-Type: text/plain; charset="utf-8" NBDClient has a number of fields that are accessed by both the export AioContext and the main loop thread. When the AioContext lock is removed these fields will need another form of protection. Add NBDClient->lock and protect fields that are accessed by both threads. Also add assertions where possible and otherwise add doc comments stating assumptions about which thread and lock holding. Note this patch moves the client->recv_coroutine assertion from nbd_co_receive_request() to nbd_trip() where client->lock is held. Signed-off-by: Stefan Hajnoczi --- nbd/server.c | 144 +++++++++++++++++++++++++++++++++++++++------------ 1 file changed, 111 insertions(+), 33 deletions(-) diff --git a/nbd/server.c b/nbd/server.c index e91e2e0903..941832f178 100644 --- a/nbd/server.c +++ b/nbd/server.c @@ -125,23 +125,25 @@ struct NBDClient { int refcount; /* atomic */ void (*close_fn)(NBDClient *client, bool negotiated); =20 + QemuMutex lock; + NBDExport *exp; QCryptoTLSCreds *tlscreds; char *tlsauthz; QIOChannelSocket *sioc; /* The underlying data channel */ QIOChannel *ioc; /* The current I/O channel which may differ (eg TLS) = */ =20 - Coroutine *recv_coroutine; + Coroutine *recv_coroutine; /* protected by lock */ =20 CoMutex send_lock; Coroutine *send_coroutine; =20 - bool read_yielding; - bool quiescing; + bool read_yielding; /* protected by lock */ + bool quiescing; /* protected by lock */ =20 QTAILQ_ENTRY(NBDClient) next; - int nb_requests; - bool closing; + int nb_requests; /* protected by lock */ + bool closing; /* protected by lock */ =20 uint32_t check_align; /* If non-zero, check for aligned client request= s */ =20 @@ -1415,11 +1417,18 @@ nbd_read_eof(NBDClient *client, void *buffer, size_= t size, Error **errp) =20 len =3D qio_channel_readv(client->ioc, &iov, 1, errp); if (len =3D=3D QIO_CHANNEL_ERR_BLOCK) { - client->read_yielding =3D true; + WITH_QEMU_LOCK_GUARD(&client->lock) { + client->read_yielding =3D true; + + /* Prompt main loop thread to re-run nbd_drained_poll() */ + aio_wait_kick(); + } qio_channel_yield(client->ioc, G_IO_IN); - client->read_yielding =3D false; - if (client->quiescing) { - return -EAGAIN; + WITH_QEMU_LOCK_GUARD(&client->lock) { + client->read_yielding =3D false; + if (client->quiescing) { + return -EAGAIN; + } } continue; } else if (len < 0) { @@ -1528,6 +1537,7 @@ void nbd_client_put(NBDClient *client) blk_exp_unref(&client->exp->common); } g_free(client->contexts.bitmaps); + qemu_mutex_destroy(&client->lock); g_free(client); } } @@ -1561,11 +1571,13 @@ static void client_close(NBDClient *client, bool ne= gotiated) { assert(qemu_in_main_thread()); =20 - if (client->closing) { - return; - } + WITH_QEMU_LOCK_GUARD(&client->lock) { + if (client->closing) { + return; + } =20 - client->closing =3D true; + client->closing =3D true; + } =20 /* Force requests to finish. They will drop their own references, * then we'll close the socket and free the NBDClient. @@ -1579,6 +1591,7 @@ static void client_close(NBDClient *client, bool nego= tiated) } } =20 +/* Runs in export AioContext with client->lock held */ static NBDRequestData *nbd_request_get(NBDClient *client) { NBDRequestData *req; @@ -1591,6 +1604,7 @@ static NBDRequestData *nbd_request_get(NBDClient *cli= ent) return req; } =20 +/* Runs in export AioContext with client->lock held */ static void nbd_request_put(NBDRequestData *req) { NBDClient *client =3D req->client; @@ -1614,14 +1628,18 @@ static void blk_aio_attached(AioContext *ctx, void = *opaque) NBDExport *exp =3D opaque; NBDClient *client; =20 + assert(qemu_in_main_thread()); + trace_nbd_blk_aio_attached(exp->name, ctx); =20 exp->common.ctx =3D ctx; =20 QTAILQ_FOREACH(client, &exp->clients, next) { - assert(client->nb_requests =3D=3D 0); - assert(client->recv_coroutine =3D=3D NULL); - assert(client->send_coroutine =3D=3D NULL); + WITH_QEMU_LOCK_GUARD(&client->lock) { + assert(client->nb_requests =3D=3D 0); + assert(client->recv_coroutine =3D=3D NULL); + assert(client->send_coroutine =3D=3D NULL); + } } } =20 @@ -1629,6 +1647,8 @@ static void blk_aio_detach(void *opaque) { NBDExport *exp =3D opaque; =20 + assert(qemu_in_main_thread()); + trace_nbd_blk_aio_detach(exp->name, exp->common.ctx); =20 exp->common.ctx =3D NULL; @@ -1639,8 +1659,12 @@ static void nbd_drained_begin(void *opaque) NBDExport *exp =3D opaque; NBDClient *client; =20 + assert(qemu_in_main_thread()); + QTAILQ_FOREACH(client, &exp->clients, next) { - client->quiescing =3D true; + WITH_QEMU_LOCK_GUARD(&client->lock) { + client->quiescing =3D true; + } } } =20 @@ -1649,28 +1673,48 @@ static void nbd_drained_end(void *opaque) NBDExport *exp =3D opaque; NBDClient *client; =20 + assert(qemu_in_main_thread()); + QTAILQ_FOREACH(client, &exp->clients, next) { - client->quiescing =3D false; - nbd_client_receive_next_request(client); + WITH_QEMU_LOCK_GUARD(&client->lock) { + client->quiescing =3D false; + nbd_client_receive_next_request(client); + } } } =20 +/* Runs in export AioContext */ +static void nbd_wake_read_bh(void *opaque) +{ + NBDClient *client =3D opaque; + qio_channel_wake_read(client->ioc); +} + static bool nbd_drained_poll(void *opaque) { NBDExport *exp =3D opaque; NBDClient *client; =20 + assert(qemu_in_main_thread()); + QTAILQ_FOREACH(client, &exp->clients, next) { - if (client->nb_requests !=3D 0) { - /* - * If there's a coroutine waiting for a request on nbd_read_eo= f() - * enter it here so we don't depend on the client to wake it u= p. - */ - if (client->recv_coroutine !=3D NULL && client->read_yielding)= { - qio_channel_wake_read(client->ioc); + WITH_QEMU_LOCK_GUARD(&client->lock) { + if (client->nb_requests !=3D 0) { + /* + * If there's a coroutine waiting for a request on nbd_rea= d_eof() + * enter it here so we don't depend on the client to wake = it up. + * + * Schedule a BH in the export AioContext to avoid missing= the + * wake up due to the race between qio_channel_wake_read()= and + * qio_channel_yield(). + */ + if (client->recv_coroutine !=3D NULL && client->read_yield= ing) { + aio_bh_schedule_oneshot(nbd_export_aio_context(client-= >exp), + nbd_wake_read_bh, client); + } + + return true; } - - return true; } } =20 @@ -1681,6 +1725,8 @@ static void nbd_eject_notifier(Notifier *n, void *dat= a) { NBDExport *exp =3D container_of(n, NBDExport, eject_notifier); =20 + assert(qemu_in_main_thread()); + blk_exp_request_shutdown(&exp->common); } =20 @@ -2566,7 +2612,6 @@ static int coroutine_fn nbd_co_receive_request(NBDReq= uestData *req, int ret; =20 g_assert(qemu_in_coroutine()); - assert(client->recv_coroutine =3D=3D qemu_coroutine_self()); ret =3D nbd_receive_request(client, request, errp); if (ret < 0) { return ret; @@ -2975,6 +3020,9 @@ static coroutine_fn void nbd_trip(void *opaque) */ =20 trace_nbd_trip(); + + qemu_mutex_lock(&client->lock); + if (client->closing) { goto done; } @@ -2990,7 +3038,21 @@ static coroutine_fn void nbd_trip(void *opaque) } =20 req =3D nbd_request_get(client); - ret =3D nbd_co_receive_request(req, &request, &local_err); + + /* + * nbd_co_receive_request() returns -EAGAIN when nbd_drained_begin() h= as + * set client->quiescing but by the time we get back nbd_drained_end()= may + * have already cleared client->quiescing. In that case we try again + * because nothing else will spawn an nbd_trip() coroutine until we set + * client->recv_coroutine =3D NULL further down. + */ + do { + assert(client->recv_coroutine =3D=3D qemu_coroutine_self()); + qemu_mutex_unlock(&client->lock); + ret =3D nbd_co_receive_request(req, &request, &local_err); + qemu_mutex_lock(&client->lock); + } while (ret =3D=3D -EAGAIN && !client->quiescing); + client->recv_coroutine =3D NULL; =20 if (client->closing) { @@ -3002,15 +3064,16 @@ static coroutine_fn void nbd_trip(void *opaque) } =20 if (ret =3D=3D -EAGAIN) { - assert(client->quiescing); goto done; } =20 nbd_client_receive_next_request(client); + if (ret =3D=3D -EIO) { goto disconnect; } =20 + qemu_mutex_unlock(&client->lock); qio_channel_set_cork(client->ioc, true); =20 if (ret < 0) { @@ -3030,6 +3093,10 @@ static coroutine_fn void nbd_trip(void *opaque) g_free(request.contexts->bitmaps); g_free(request.contexts); } + + qio_channel_set_cork(client->ioc, false); + qemu_mutex_lock(&client->lock); + if (ret < 0) { error_prepend(&local_err, "Failed to send reply: "); goto disconnect; @@ -3044,11 +3111,13 @@ static coroutine_fn void nbd_trip(void *opaque) goto disconnect; } =20 - qio_channel_set_cork(client->ioc, false); done: if (req) { nbd_request_put(req); } + + qemu_mutex_unlock(&client->lock); + if (!nbd_client_put_nonzero(client)) { aio_co_reschedule_self(qemu_get_aio_context()); nbd_client_put(client); @@ -3059,13 +3128,19 @@ disconnect: if (local_err) { error_reportf_err(local_err, "Disconnect client, due to: "); } + nbd_request_put(req); + qemu_mutex_unlock(&client->lock); =20 aio_co_reschedule_self(qemu_get_aio_context()); client_close(client, true); nbd_client_put(client); } =20 +/* + * Runs in export AioContext and main loop thread. Caller must hold + * client->lock. + */ static void nbd_client_receive_next_request(NBDClient *client) { if (!client->recv_coroutine && client->nb_requests < MAX_NBD_REQUESTS = && @@ -3091,7 +3166,9 @@ static coroutine_fn void nbd_co_client_start(void *op= aque) return; } =20 - nbd_client_receive_next_request(client); + WITH_QEMU_LOCK_GUARD(&client->lock) { + nbd_client_receive_next_request(client); + } } =20 /* @@ -3108,6 +3185,7 @@ void nbd_client_new(QIOChannelSocket *sioc, Coroutine *co; =20 client =3D g_new0(NBDClient, 1); + qemu_mutex_init(&client->lock); client->refcount =3D 1; client->tlscreds =3D tlscreds; if (tlscreds) { --=20 2.43.0