From patchwork Tue Jan 29 03:22:10 2013 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Liu Bo X-Patchwork-Id: 2058921 Return-Path: X-Original-To: patchwork-linux-btrfs@patchwork.kernel.org Delivered-To: patchwork-process-083081@patchwork2.kernel.org Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by patchwork2.kernel.org (Postfix) with ESMTP id 1E7DADF2EC for ; Tue, 29 Jan 2013 03:25:42 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1752773Ab3A2DZj (ORCPT ); Mon, 28 Jan 2013 22:25:39 -0500 Received: from aserp1040.oracle.com ([141.146.126.69]:29397 "EHLO aserp1040.oracle.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751801Ab3A2DZi (ORCPT ); Mon, 28 Jan 2013 22:25:38 -0500 Received: from acsinet22.oracle.com (acsinet22.oracle.com [141.146.126.238]) by aserp1040.oracle.com (Sentrion-MTA-4.3.1/Sentrion-MTA-4.3.1) with ESMTP id r0T3OrLk029009 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK); Tue, 29 Jan 2013 03:24:54 GMT Received: from acsmt358.oracle.com (acsmt358.oracle.com [141.146.40.158]) by acsinet22.oracle.com (8.14.4+Sun/8.14.4) with ESMTP id r0T3OqDT018065 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NO); Tue, 29 Jan 2013 03:24:53 GMT Received: from abhmt119.oracle.com (abhmt119.oracle.com [141.146.116.71]) by acsmt358.oracle.com (8.12.11.20060308/8.12.11) with ESMTP id r0T3Oq4n007218; Mon, 28 Jan 2013 21:24:52 -0600 Received: from liubo.jp.oracle.com (/10.191.5.43) by default (Oracle Beehive Gateway v4.0) with ESMTP ; Mon, 28 Jan 2013 19:24:52 -0800 From: Liu Bo To: linux-btrfs@vger.kernel.org Cc: mitch.harder@sabayonlinux.org, dsterba@suse.cz Subject: [PATCH v2] Btrfs: fix race between snapshot deletion and getting inode Date: Tue, 29 Jan 2013 11:22:10 +0800 Message-Id: <1359429730-16729-1-git-send-email-bo.li.liu@oracle.com> X-Mailer: git-send-email 1.7.7.6 X-Source-IP: acsinet22.oracle.com [141.146.126.238] Sender: linux-btrfs-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: linux-btrfs@vger.kernel.org While running snapshot testscript created by Mitch and David, the race between autodefrag and snapshot deletion can lead to corruption of dead_root list so that we can get crash on btrfs_clean_old_snapshots(). And besides autodefrag, scrub also does the same thing, ie. read root first and get inode. Here is the story(take autodefrag as an example): (1) when we delete a snapshot or subvolume, it will set its root's refs to zero and do a iput() on its own inode, and if this inode happens to be the only active in-meory one in root's inode rbtree, it will add itself to the global dead_roots list for later cleanup. (2) after (1), the autodefrag thread may read another inode for defrag and the inode is just in the deleted snapshot/subvolume, but all of these are without checking if the root is still valid(refs > 0). So the end up result is adding the deleted snapshot/subvolume's root to the global dead_roots list AGAIN. Fortunately, we already have a srcu lock to avoid the race, ie. subvol_srcu. So all we need to do is to take the lock to protect 'read root and get inode', since we synchronize to wait for the rcu grace period before adding something to the global dead_roots list. Reported-by: Mitch Harder Signed-off-by: Liu Bo --- v1->v2: Address nice cleanup comments from David Sterba. fs/btrfs/file.c | 22 ++++++++++++++++++---- fs/btrfs/scrub.c | 25 ++++++++++++++++++++----- 2 files changed, 38 insertions(+), 9 deletions(-) diff --git a/fs/btrfs/file.c b/fs/btrfs/file.c index 841cfe3..38ffd28 100644 --- a/fs/btrfs/file.c +++ b/fs/btrfs/file.c @@ -293,15 +293,24 @@ static int __btrfs_run_defrag_inode(struct btrfs_fs_info *fs_info, struct btrfs_key key; struct btrfs_ioctl_defrag_range_args range; int num_defrag; + int index; + int ret; /* get the inode */ key.objectid = defrag->root; btrfs_set_key_type(&key, BTRFS_ROOT_ITEM_KEY); key.offset = (u64)-1; + + index = srcu_read_lock(&fs_info->subvol_srcu); + inode_root = btrfs_read_fs_root_no_name(fs_info, &key); if (IS_ERR(inode_root)) { - kmem_cache_free(btrfs_inode_defrag_cachep, defrag); - return PTR_ERR(inode_root); + ret = PTR_ERR(inode_root); + goto cleanup; + } + if (btrfs_root_refs(&inode_root->root_item) == 0) { + ret = -ENOENT; + goto cleanup; } key.objectid = defrag->ino; @@ -309,9 +318,10 @@ static int __btrfs_run_defrag_inode(struct btrfs_fs_info *fs_info, key.offset = 0; inode = btrfs_iget(fs_info->sb, &key, inode_root, NULL); if (IS_ERR(inode)) { - kmem_cache_free(btrfs_inode_defrag_cachep, defrag); - return PTR_ERR(inode); + ret = PTR_ERR(inode); + goto cleanup; } + srcu_read_unlock(&fs_info->subvol_srcu, index); /* do a chunk of defrag */ clear_bit(BTRFS_INODE_IN_DEFRAG, &BTRFS_I(inode)->runtime_flags); @@ -346,6 +356,10 @@ static int __btrfs_run_defrag_inode(struct btrfs_fs_info *fs_info, iput(inode); return 0; +cleanup: + srcu_read_unlock(&fs_info->subvol_srcu, index); + kmem_cache_free(btrfs_inode_defrag_cachep, defrag); + return ret; } /* diff --git a/fs/btrfs/scrub.c b/fs/btrfs/scrub.c index bdbb94f..67783e0 100644 --- a/fs/btrfs/scrub.c +++ b/fs/btrfs/scrub.c @@ -580,20 +580,29 @@ static int scrub_fixup_readpage(u64 inum, u64 offset, u64 root, void *fixup_ctx) int corrected = 0; struct btrfs_key key; struct inode *inode = NULL; + struct btrfs_fs_info *fs_info; u64 end = offset + PAGE_SIZE - 1; struct btrfs_root *local_root; + int srcu_index; key.objectid = root; key.type = BTRFS_ROOT_ITEM_KEY; key.offset = (u64)-1; - local_root = btrfs_read_fs_root_no_name(fixup->root->fs_info, &key); - if (IS_ERR(local_root)) + + fs_info = fixup->root->fs_info; + srcu_index = srcu_read_lock(&fs_info->subvol_srcu); + + local_root = btrfs_read_fs_root_no_name(fs_info, &key); + if (IS_ERR(local_root)) { + srcu_read_unlock(&fs_info->subvol_srcu, srcu_index); return PTR_ERR(local_root); + } key.type = BTRFS_INODE_ITEM_KEY; key.objectid = inum; key.offset = 0; - inode = btrfs_iget(fixup->root->fs_info->sb, &key, local_root, NULL); + inode = btrfs_iget(fs_info->sb, &key, local_root, NULL); + srcu_read_unlock(&fs_info->subvol_srcu, srcu_index); if (IS_ERR(inode)) return PTR_ERR(inode); @@ -606,7 +615,6 @@ static int scrub_fixup_readpage(u64 inum, u64 offset, u64 root, void *fixup_ctx) } if (PageUptodate(page)) { - struct btrfs_fs_info *fs_info; if (PageDirty(page)) { /* * we need to write the data to the defect sector. the @@ -3180,18 +3188,25 @@ static int copy_nocow_pages_for_inode(u64 inum, u64 offset, u64 root, void *ctx) u64 physical_for_dev_replace; u64 len; struct btrfs_fs_info *fs_info = nocow_ctx->sctx->dev_root->fs_info; + int srcu_index; key.objectid = root; key.type = BTRFS_ROOT_ITEM_KEY; key.offset = (u64)-1; + + srcu_index = srcu_read_lock(&fs_info->subvol_srcu); + local_root = btrfs_read_fs_root_no_name(fs_info, &key); - if (IS_ERR(local_root)) + if (IS_ERR(local_root)) { + srcu_read_unlock(&fs_info->subvol_srcu, srcu_index); return PTR_ERR(local_root); + } key.type = BTRFS_INODE_ITEM_KEY; key.objectid = inum; key.offset = 0; inode = btrfs_iget(fs_info->sb, &key, local_root, NULL); + srcu_read_unlock(&fs_info->subvol_srcu, srcu_index); if (IS_ERR(inode)) return PTR_ERR(inode);