Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
From:   Ryusuke Konishi <konishi.ryusuke@gmail.com>
To:     Andrew Morton <akpm@linux-foundation.org>
Cc:     linux-nilfs@vger.kernel.org,
        syzbot <syzbot+78d4495558999f55d1da@syzkaller.appspotmail.com>,
        syzkaller-bugs@googlegroups.com, linux-kernel@vger.kernel.org
Subject: [PATCH] nilfs2: fix use-after-free bug of nilfs_root in nilfs_evict_inode()
Date:   Wed, 10 May 2023 00:29:56 +0900
Message-Id: <20230509152956.8313-1-konishi.ryusuke@gmail.com>
In-Reply-To: <00000000000099e5ac05fb1c3b85@google.com>
References: <00000000000099e5ac05fb1c3b85@google.com>
MIME-Version: 1.0
Content-Transfer-Encoding: 8bit
Precedence: bulk

During unmount process of nilfs2, nothing holds nilfs_root structure after
nilfs2 detaches its writer in nilfs_detach_log_writer().  However, since
nilfs_evict_inode() uses nilfs_root for some cleanup operations, it may
cause use-after-free read if inodes are left in "garbage_list" and
released by nilfs_dispose_list() at the end of nilfs_detach_log_writer().

Fix this issue by modifying nilfs_evict_inode() to only clear inode
without additional metadata changes that use nilfs_root if the file system
is degraded to read-only or the writer is detached.

Signed-off-by: Ryusuke Konishi <konishi.ryusuke@gmail.com>
Reported-by: syzbot+78d4495558999f55d1da@syzkaller.appspotmail.com
Closes: https://lkml.kernel.org/r/00000000000099e5ac05fb1c3b85@google.com
Tested-by: Ryusuke Konishi <konishi.ryusuke@gmail.com>
Cc: stable@vger.kernel.org
---
 fs/nilfs2/inode.c | 18 ++++++++++++++++++
 1 file changed, 18 insertions(+)

diff --git a/fs/nilfs2/inode.c b/fs/nilfs2/inode.c
index 1310d2d5feb3..a8ce522ac747 100644
--- a/fs/nilfs2/inode.c
+++ b/fs/nilfs2/inode.c
@@ -917,6 +917,7 @@ void nilfs_evict_inode(struct inode *inode)
 	struct nilfs_transaction_info ti;
 	struct super_block *sb = inode->i_sb;
 	struct nilfs_inode_info *ii = NILFS_I(inode);
+	struct the_nilfs *nilfs;
 	int ret;
 
 	if (inode->i_nlink || !ii->i_root || unlikely(is_bad_inode(inode))) {
@@ -929,6 +930,23 @@ void nilfs_evict_inode(struct inode *inode)
 
 	truncate_inode_pages_final(&inode->i_data);
 
+	nilfs = sb->s_fs_info;
+	if (unlikely(sb_rdonly(sb) || !nilfs->ns_writer)) {
+		/*
+		 * If this inode is about to be disposed after the file system
+		 * has been degraded to read-only due to file system corruption
+		 * or after the writer has been detached, do not make any
+		 * changes that cause writes, just clear it.
+		 * Do this check after read-locking ns_segctor_sem by
+		 * nilfs_transaction_begin() in order to avoid a race with
+		 * the writer detach operation.
+		 */
+		clear_inode(inode);
+		nilfs_clear_inode(inode);
+		nilfs_transaction_abort(sb);
+		return;
+	}
+
 	/* TODO: some of the following operations may fail.  */
 	nilfs_truncate_bmap(ii, 0);
 	nilfs_mark_inode_dirty(inode);
-- 
2.34.1