Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1754724AbcDSUrd (ORCPT ); Tue, 19 Apr 2016 16:47:33 -0400 Received: from mail-wm0-f50.google.com ([74.125.82.50]:37921 "EHLO mail-wm0-f50.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752703AbcDSUrc (ORCPT ); Tue, 19 Apr 2016 16:47:32 -0400 Message-ID: <1461098921.2921.3.camel@margaine.com> Subject: [PATCH] fs: reintroduce freezing nesting From: Florian Margaine To: Alexander Viro , Jeff Layton , "J. Bruce Fields" , linux-fsdevel@vger.kernel.org, "linux-kernel@vger.kernel.org" Date: Tue, 19 Apr 2016 22:48:41 +0200 Content-Type: text/plain; charset="UTF-8" X-Mailer: Evolution 3.20.1 Mime-Version: 1.0 Content-Transfer-Encoding: 8bit Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 2195 Lines: 82 The behavior was removed in 18e9e5104fcd9a973ffe3eed3816c87f2a1b6cd2 noting that this was a better idea than using a counter. However, this behavior is actually wanted if multiple applications want to freeze concurrently while remaining non-racy. This patch reintroduces this feature by using a counter. ---  fs/super.c         | 15 +++++++++++----  include/linux/fs.h |  1 +  2 files changed, 12 insertions(+), 4 deletions(-) diff --git a/fs/super.c b/fs/super.c index 74914b1..9fa8ca1 100644 --- a/fs/super.c +++ b/fs/super.c @@ -231,6 +231,7 @@ static struct super_block *alloc_super(struct file_system_type *type, int flags)    */   down_write_nested(&s->s_umount, SINGLE_DEPTH_NESTING);   s->s_count = 1; + s->s_freezers = 0;   atomic_set(&s->s_active, 1);   mutex_init(&s->s_vfs_rename_mutex);   lockdep_set_class(&s->s_vfs_rename_mutex, &type- >s_vfs_rename_key); @@ -1275,12 +1276,12 @@ int freeze_super(struct super_block *sb)  {   int ret;   + sb->s_freezers++; + if (sb->s_writers.frozen != SB_UNFROZEN) + return 0; +   atomic_inc(&sb->s_active);   down_write(&sb->s_umount); - if (sb->s_writers.frozen != SB_UNFROZEN) { - deactivate_locked_super(sb); - return -EBUSY; - }     if (!(sb->s_flags & MS_BORN)) {   up_write(&sb->s_umount); @@ -1338,14 +1339,20 @@ EXPORT_SYMBOL(freeze_super);   * @sb: the super to thaw   *   * Unlocks the filesystem and marks it writeable again after freeze_super(). + * Since nesting freezes is allowed, only the last freeze actually unlocks.   */  int thaw_super(struct super_block *sb)  {   int error;   + sb->s_freezers--; + if (sb->s_freezers > 0) + return 0; +   down_write(&sb->s_umount);   if (sb->s_writers.frozen == SB_UNFROZEN) {   up_write(&sb->s_umount); + sb->s_freezers++;   return -EINVAL;   }   diff --git a/include/linux/fs.h b/include/linux/fs.h index e514f76..c045e2a 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -1333,6 +1333,7 @@ struct super_block {   struct quota_info s_dquot; /* Diskquota specific options */     struct sb_writers s_writers; + int s_freezers;     char s_id[32]; /* Informational name */   u8 s_uuid[16]; /* UUID */ --  2.8.0