From patchwork Mon Mar 10 03:11:45 2014 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: "Yan, Zheng" X-Patchwork-Id: 3801091 Return-Path: X-Original-To: patchwork-ceph-devel@patchwork.kernel.org Delivered-To: patchwork-parsemail@patchwork1.web.kernel.org Received: from mail.kernel.org (mail.kernel.org [198.145.19.201]) by patchwork1.web.kernel.org (Postfix) with ESMTP id 020D29F369 for ; Mon, 10 Mar 2014 03:12:09 +0000 (UTC) Received: from mail.kernel.org (localhost [127.0.0.1]) by mail.kernel.org (Postfix) with ESMTP id ECB9E20383 for ; Mon, 10 Mar 2014 03:12:07 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by mail.kernel.org (Postfix) with ESMTP id 9423B20131 for ; Mon, 10 Mar 2014 03:12:06 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1752898AbaCJDME (ORCPT ); Sun, 9 Mar 2014 23:12:04 -0400 Received: from mga11.intel.com ([192.55.52.93]:51965 "EHLO mga11.intel.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752403AbaCJDMC (ORCPT ); Sun, 9 Mar 2014 23:12:02 -0400 Received: from fmsmga001.fm.intel.com ([10.253.24.23]) by fmsmga102.fm.intel.com with ESMTP; 09 Mar 2014 20:11:49 -0700 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="4.97,621,1389772800"; d="scan'208";a="488925158" Received: from zyan5-mobl.sh.intel.com ([10.239.13.4]) by fmsmga001.fm.intel.com with ESMTP; 09 Mar 2014 20:11:48 -0700 From: "Yan, Zheng" To: ceph-devel@vger.kernel.org, sage@inktank.com Cc: "Yan, Zheng" Subject: [PATCH] ceph: use fl->fl_file as owner identifier of flock and posix lock Date: Mon, 10 Mar 2014 11:11:45 +0800 Message-Id: <1394421105-7215-1-git-send-email-zheng.z.yan@intel.com> X-Mailer: git-send-email 1.8.5.3 Sender: ceph-devel-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: ceph-devel@vger.kernel.org X-Spam-Status: No, score=-6.9 required=5.0 tests=BAYES_00, RCVD_IN_DNSWL_HI, T_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 flock and posix lock should use fl->fl_file instead of process ID as owner identifier. (posix lock uses fl->fl_owner, fl->fl_owner is usually equal to fl->fl_file in most cases). The process ID of who holds the lock is just for F_GETLK fcntl(2). The fix is rename the 'pid' fields of struct ceph_mds_request_args and struct ceph_filelock to 'owner', rename 'pid_namespace' fields to 'pid'. Assign fl->fl_file to the 'owner' field of lock messages. The MDS counterpart of this patch modifies the flock code to not take the 'pid_namespace' into consideration when checking conflict locks. Signed-off-by: Yan, Zheng --- fs/ceph/locks.c | 44 ++++++++++++++++++++++++-------------------- include/linux/ceph/ceph_fs.h | 4 ++-- 2 files changed, 26 insertions(+), 22 deletions(-) diff --git a/fs/ceph/locks.c b/fs/ceph/locks.c index f91a569a..4857007 100644 --- a/fs/ceph/locks.c +++ b/fs/ceph/locks.c @@ -14,9 +14,9 @@ static int ceph_lock_message(u8 lock_type, u16 operation, struct file *file, int cmd, u8 wait, struct file_lock *fl) { struct inode *inode = file_inode(file); - struct ceph_mds_client *mdsc = - ceph_sb_to_client(inode->i_sb)->mdsc; + struct ceph_mds_client *mdsc = ceph_sb_to_client(inode->i_sb)->mdsc; struct ceph_mds_request *req; + void *owner; int err; u64 length = 0; @@ -32,25 +32,28 @@ static int ceph_lock_message(u8 lock_type, u16 operation, struct file *file, else length = fl->fl_end - fl->fl_start + 1; - dout("ceph_lock_message: rule: %d, op: %d, pid: %llu, start: %llu, " - "length: %llu, wait: %d, type: %d", (int)lock_type, - (int)operation, (u64)fl->fl_pid, fl->fl_start, - length, wait, fl->fl_type); + if (lock_type == CEPH_LOCK_FCNTL) + owner = fl->fl_owner; + else + owner = fl->fl_file; + + dout("ceph_lock_message: rule: %d, op: %d, owner: %p, pid: %llu, " + "start: %llu, length: %llu, wait: %d, type: %d", (int)lock_type, + (int)operation, owner, (u64)fl->fl_pid, fl->fl_start, length, + wait, fl->fl_type); req->r_args.filelock_change.rule = lock_type; req->r_args.filelock_change.type = cmd; + req->r_args.filelock_change.owner = + cpu_to_le64((u64)(unsigned long)owner); req->r_args.filelock_change.pid = cpu_to_le64((u64)fl->fl_pid); - /* This should be adjusted, but I'm not sure if - namespaces actually get id numbers*/ - req->r_args.filelock_change.pid_namespace = - cpu_to_le64((u64)(unsigned long)fl->fl_nspid); req->r_args.filelock_change.start = cpu_to_le64(fl->fl_start); req->r_args.filelock_change.length = cpu_to_le64(length); req->r_args.filelock_change.wait = wait; err = ceph_mdsc_do_request(mdsc, inode, req); - if ( operation == CEPH_MDS_OP_GETFILELOCK){ + if (operation == CEPH_MDS_OP_GETFILELOCK) { fl->fl_pid = le64_to_cpu(req->r_reply_info.filelock_reply->pid); if (CEPH_LOCK_SHARED == req->r_reply_info.filelock_reply->type) fl->fl_type = F_RDLCK; @@ -93,8 +96,7 @@ int ceph_lock(struct file *file, int cmd, struct file_lock *fl) if (__mandatory_lock(file->f_mapping->host) && fl->fl_type != F_UNLCK) return -ENOLCK; - fl->fl_nspid = get_pid(task_tgid(current)); - dout("ceph_lock, fl_pid:%d", fl->fl_pid); + dout("ceph_lock, fl_owner: %p", fl->fl_owner); /* set wait bit as appropriate, then make command as Ceph expects it*/ if (IS_GETLK(cmd)) @@ -111,7 +113,7 @@ int ceph_lock(struct file *file, int cmd, struct file_lock *fl) err = ceph_lock_message(CEPH_LOCK_FCNTL, op, file, lock_cmd, wait, fl); if (!err) { - if ( op != CEPH_MDS_OP_GETFILELOCK ){ + if (op != CEPH_MDS_OP_GETFILELOCK) { dout("mds locked, locking locally"); err = posix_lock_file(file, fl, NULL); if (err && (CEPH_MDS_OP_SETFILELOCK == op)) { @@ -145,8 +147,7 @@ int ceph_flock(struct file *file, int cmd, struct file_lock *fl) if (__mandatory_lock(file->f_mapping->host) && fl->fl_type != F_UNLCK) return -ENOLCK; - fl->fl_nspid = get_pid(task_tgid(current)); - dout("ceph_flock, fl_pid:%d", fl->fl_pid); + dout("ceph_flock, fl_file: %p", fl->fl_file); if (IS_SETLKW(cmd)) wait = 1; @@ -289,13 +290,16 @@ int lock_to_ceph_filelock(struct file_lock *lock, struct ceph_filelock *cephlock) { int err = 0; - cephlock->start = cpu_to_le64(lock->fl_start); cephlock->length = cpu_to_le64(lock->fl_end - lock->fl_start + 1); cephlock->client = cpu_to_le64(0); - cephlock->pid = cpu_to_le64(lock->fl_pid); - cephlock->pid_namespace = - cpu_to_le64((u64)(unsigned long)lock->fl_nspid); + cephlock->pid = cpu_to_le64((u64)lock->fl_pid); + if (lock->fl_flags & FL_POSIX) + cephlock->owner = + cpu_to_le64((u64)(unsigned long)lock->fl_owner); + else + cephlock->owner = + cpu_to_le64((u64)(unsigned long)lock->fl_file); switch (lock->fl_type) { case F_RDLCK: diff --git a/include/linux/ceph/ceph_fs.h b/include/linux/ceph/ceph_fs.h index 35f345f..5f6db18 100644 --- a/include/linux/ceph/ceph_fs.h +++ b/include/linux/ceph/ceph_fs.h @@ -421,8 +421,8 @@ union ceph_mds_request_args { struct { __u8 rule; /* currently fcntl or flock */ __u8 type; /* shared, exclusive, remove*/ + __le64 owner; /* owner of the lock */ __le64 pid; /* process id requesting the lock */ - __le64 pid_namespace; __le64 start; /* initial location to lock */ __le64 length; /* num bytes to lock from start */ __u8 wait; /* will caller wait for lock to become available? */ @@ -533,8 +533,8 @@ struct ceph_filelock { __le64 start;/* file offset to start lock at */ __le64 length; /* num bytes to lock; 0 for all following start */ __le64 client; /* which client holds the lock */ + __le64 owner; /* owner the lock */ __le64 pid; /* process id holding the lock on the client */ - __le64 pid_namespace; __u8 type; /* shared lock, exclusive lock, or unlock */ } __attribute__ ((packed));