From patchwork Fri Nov 20 08:23:16 2015 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Ryan Ding X-Patchwork-Id: 7665511 Return-Path: X-Original-To: patchwork-ocfs2-devel@patchwork.kernel.org Delivered-To: patchwork-parsemail@patchwork1.web.kernel.org Received: from mail.kernel.org (mail.kernel.org [198.145.29.136]) by patchwork1.web.kernel.org (Postfix) with ESMTP id E53159F392 for ; Fri, 20 Nov 2015 08:24:44 +0000 (UTC) Received: from mail.kernel.org (localhost [127.0.0.1]) by mail.kernel.org (Postfix) with ESMTP id D54A4203AE for ; Fri, 20 Nov 2015 08:24:43 +0000 (UTC) Received: from aserp1040.oracle.com (aserp1040.oracle.com [141.146.126.69]) (using TLSv1.2 with cipher DHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by mail.kernel.org (Postfix) with ESMTPS id 8FB122035E for ; Fri, 20 Nov 2015 08:24:42 +0000 (UTC) Received: from userv0021.oracle.com (userv0021.oracle.com [156.151.31.71]) by aserp1040.oracle.com (Sentrion-MTA-4.3.2/Sentrion-MTA-4.3.2) with ESMTP id tAK8Nrx5013225 (version=TLSv1 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Fri, 20 Nov 2015 08:23:54 GMT Received: from oss.oracle.com (oss-old-reserved.oracle.com [137.254.22.2]) by userv0021.oracle.com (8.13.8/8.13.8) with ESMTP id tAK8NpwP005351 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NO); Fri, 20 Nov 2015 08:23:51 GMT Received: from localhost ([127.0.0.1] helo=lb-oss.oracle.com) by oss.oracle.com with esmtp (Exim 4.63) (envelope-from ) id 1Zzgz1-0007kt-9x; Fri, 20 Nov 2015 00:23:51 -0800 Received: from userv0021.oracle.com ([156.151.31.71]) by oss.oracle.com with esmtp (Exim 4.63) (envelope-from ) id 1Zzgyd-0007jt-Bz for ocfs2-devel@oss.oracle.com; Fri, 20 Nov 2015 00:23:27 -0800 Received: from userv0122.oracle.com (userv0122.oracle.com [156.151.31.75]) by userv0021.oracle.com (8.13.8/8.13.8) with ESMTP id tAK8NQrm003877 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=FAIL); Fri, 20 Nov 2015 08:23:26 GMT Received: from abhmp0012.oracle.com (abhmp0012.oracle.com [141.146.116.18]) by userv0122.oracle.com (8.13.8/8.13.8) with ESMTP id tAK8NQbp002240; Fri, 20 Nov 2015 08:23:26 GMT Received: from dhcp-cn-10-182-70-63.cn.oracle.com (/10.182.37.238) by default (Oracle Beehive Gateway v4.0) with ESMTP ; Fri, 20 Nov 2015 00:23:26 -0800 From: Ryan Ding To: ocfs2-devel@oss.oracle.com Date: Fri, 20 Nov 2015 16:23:16 +0800 Message-Id: <1448007799-10914-1-git-send-email-ryan.ding@oracle.com> X-Mailer: git-send-email 1.7.1 Cc: mfasheh@suse.de Subject: [Ocfs2-devel] [PATCH 1/4] ocfs2: fix ip_unaligned_aio deadlock with dio work queue X-BeenThere: ocfs2-devel@oss.oracle.com X-Mailman-Version: 2.1.9 Precedence: list List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , MIME-Version: 1.0 Sender: ocfs2-devel-bounces@oss.oracle.com Errors-To: ocfs2-devel-bounces@oss.oracle.com X-Source-IP: userv0021.oracle.com [156.151.31.71] X-Spam-Status: No, score=-4.8 required=5.0 tests=BAYES_00, RCVD_IN_DNSWL_MED, RP_MATCHES_RCVD, UNPARSEABLE_RELAY autolearn=unavailable version=3.3.1 X-Spam-Checker-Version: SpamAssassin 3.3.1 (2010-03-16) on mail.kernel.org X-Virus-Scanned: ClamAV using ClamSMTP In the current implementation of unaligned aio+dio, lock order behave as follow: in user process context: -> call io_submit() -> get i_mutex <== window1 -> get ip_unaligned_aio -> submit direct io to block device -> release i_mutex -> io_submit() return in dio work queue context(the work queue is created in __blockdev_direct_IO): -> release ip_unaligned_aio <== window2 -> get i_mutex -> clear unwritten flag & change i_size -> release i_mutex There is a limitation to the thread number of dio work queue. 256 at default. If all 256 thread are in the above 'window2' stage, and there is a user process in the 'window1' stage, the system will became deadlock. Since the user process hold i_mutex to wait ip_unaligned_aio lock, while there is a direct bio hold ip_unaligned_aio mutex who is waiting for a dio work queue thread to be schedule. But all the dio work queue thread is waiting for i_mutex lock in 'window2'. This case only happened in a test which send a large number(more than 256) of aio at one io_submit() call. My design is to remove ip_unaligned_aio lock. Change it to a sync io instead. Just like ip_unaligned_aio lock, serialize the unaligned aio dio. Signed-off-by: Ryan Ding Reviewed-by: Junxiao Bi --- fs/ocfs2/aops.c | 6 ------ fs/ocfs2/aops.h | 7 ------- fs/ocfs2/file.c | 27 +++++++++------------------ fs/ocfs2/inode.h | 3 --- fs/ocfs2/super.c | 1 - 5 files changed, 9 insertions(+), 35 deletions(-) diff --git a/fs/ocfs2/aops.c b/fs/ocfs2/aops.c index 4bb9921..cc604a2 100644 --- a/fs/ocfs2/aops.c +++ b/fs/ocfs2/aops.c @@ -2388,12 +2388,6 @@ static void ocfs2_dio_end_io(struct kiocb *iocb, /* this io's submitter should not have unlocked this before we could */ BUG_ON(!ocfs2_iocb_is_rw_locked(iocb)); - if (ocfs2_iocb_is_unaligned_aio(iocb)) { - ocfs2_iocb_clear_unaligned_aio(iocb); - - mutex_unlock(&OCFS2_I(inode)->ip_unaligned_aio); - } - if (private) ocfs2_dio_end_io_write(inode, private, offset, bytes); diff --git a/fs/ocfs2/aops.h b/fs/ocfs2/aops.h index d06b80f..7fc1189 100644 --- a/fs/ocfs2/aops.h +++ b/fs/ocfs2/aops.h @@ -93,11 +93,4 @@ enum ocfs2_iocb_lock_bits { #define ocfs2_iocb_rw_locked_level(iocb) \ test_bit(OCFS2_IOCB_RW_LOCK_LEVEL, (unsigned long *)&iocb->private) -#define ocfs2_iocb_set_unaligned_aio(iocb) \ - set_bit(OCFS2_IOCB_UNALIGNED_IO, (unsigned long *)&iocb->private) -#define ocfs2_iocb_clear_unaligned_aio(iocb) \ - clear_bit(OCFS2_IOCB_UNALIGNED_IO, (unsigned long *)&iocb->private) -#define ocfs2_iocb_is_unaligned_aio(iocb) \ - test_bit(OCFS2_IOCB_UNALIGNED_IO, (unsigned long *)&iocb->private) - #endif /* OCFS2_FILE_H */ diff --git a/fs/ocfs2/file.c b/fs/ocfs2/file.c index 05346fb..6ab91cd 100644 --- a/fs/ocfs2/file.c +++ b/fs/ocfs2/file.c @@ -2170,7 +2170,7 @@ static ssize_t ocfs2_file_write_iter(struct kiocb *iocb, struct ocfs2_super *osb = OCFS2_SB(inode->i_sb); int full_coherency = !(osb->s_mount_opt & OCFS2_MOUNT_COHERENCY_BUFFERED); - int unaligned_dio = 0; + void *saved_ki_complete = NULL; int append_write = ((iocb->ki_pos + count) >= i_size_read(inode) ? 1 : 0); @@ -2233,17 +2233,12 @@ static ssize_t ocfs2_file_write_iter(struct kiocb *iocb, goto out; } - if (direct_io && !is_sync_kiocb(iocb)) - unaligned_dio = ocfs2_is_io_unaligned(inode, count, iocb->ki_pos); - - if (unaligned_dio) { + if (direct_io && !is_sync_kiocb(iocb) && + ocfs2_is_io_unaligned(inode, count, iocb->ki_pos)) { /* - * Wait on previous unaligned aio to complete before - * proceeding. + * Make it a sync io if it's an unaligned aio. */ - mutex_lock(&OCFS2_I(inode)->ip_unaligned_aio); - /* Mark the iocb as needing an unlock in ocfs2_dio_end_io */ - ocfs2_iocb_set_unaligned_aio(iocb); + saved_ki_complete = xchg(&iocb->ki_complete, NULL); } /* communicate with ocfs2_dio_end_io */ @@ -2264,11 +2259,10 @@ static ssize_t ocfs2_file_write_iter(struct kiocb *iocb, */ if ((written == -EIOCBQUEUED) || (!ocfs2_iocb_is_rw_locked(iocb))) { rw_level = -1; - unaligned_dio = 0; } if (unlikely(written <= 0)) - goto no_sync; + goto out; if (((file->f_flags & O_DSYNC) && !direct_io) || IS_SYNC(inode)) { @@ -2290,13 +2284,10 @@ static ssize_t ocfs2_file_write_iter(struct kiocb *iocb, iocb->ki_pos - 1); } -no_sync: - if (unaligned_dio && ocfs2_iocb_is_unaligned_aio(iocb)) { - ocfs2_iocb_clear_unaligned_aio(iocb); - mutex_unlock(&OCFS2_I(inode)->ip_unaligned_aio); - } - out: + if (saved_ki_complete) + xchg(&iocb->ki_complete, saved_ki_complete); + if (rw_level != -1) ocfs2_rw_unlock(inode, rw_level); diff --git a/fs/ocfs2/inode.h b/fs/ocfs2/inode.h index 0c22ddd..6e0633e 100644 --- a/fs/ocfs2/inode.h +++ b/fs/ocfs2/inode.h @@ -43,9 +43,6 @@ struct ocfs2_inode_info /* protects extended attribute changes on this inode */ struct rw_semaphore ip_xattr_sem; - /* Number of outstanding AIO's which are not page aligned */ - struct mutex ip_unaligned_aio; - /* These fields are protected by ip_lock */ spinlock_t ip_lock; u32 ip_open_count; diff --git a/fs/ocfs2/super.c b/fs/ocfs2/super.c index c70a80b..3fd171f 100644 --- a/fs/ocfs2/super.c +++ b/fs/ocfs2/super.c @@ -1723,7 +1723,6 @@ static void ocfs2_inode_init_once(void *data) INIT_LIST_HEAD(&oi->ip_io_markers); INIT_LIST_HEAD(&oi->ip_unwritten_list); oi->ip_dir_start_lookup = 0; - mutex_init(&oi->ip_unaligned_aio); init_rwsem(&oi->ip_alloc_sem); init_rwsem(&oi->ip_xattr_sem); mutex_init(&oi->ip_io_mutex);