From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from smtp.kernel.org (aws-us-west-2-korg-mail-1.web.codeaurora.org [10.30.226.201]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id 43CBF1772F for ; Mon, 15 May 2023 18:02:29 +0000 (UTC) Received: by smtp.kernel.org (Postfix) with ESMTPSA id 2AA60C433EF; Mon, 15 May 2023 18:02:27 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=linuxfoundation.org; s=korg; t=1684173749; bh=/psOibmtnnkcz1+tirmhsBiDgkCvd5+0ZG4x9ICNoaM=; h=From:To:Cc:Subject:Date:In-Reply-To:References:From; b=fMxH3WjntnpakiB8lCqvWWm3OFnGsdF1ikdHSJl0+h4WFL3haVTT1H/6tpgTloKUK YfY0BDXhFk4UDqqYNw/Fdxubg88LNMZ2P84Uv1YUF88gX/PMHWCec04Rb0CYfRva/9 ZkBDpZ/kAihI5W1cdTgurYuxQj3HP5iMB0wgJtqQ= From: Greg Kroah-Hartman To: stable@vger.kernel.org Cc: Greg Kroah-Hartman , patches@lists.linux.dev, Ryusuke Konishi , syzbot+221d75710bde87fa0e97@syzkaller.appspotmail.com, Andrew Morton Subject: [PATCH 5.4 194/282] nilfs2: fix infinite loop in nilfs_mdt_get_block() Date: Mon, 15 May 2023 18:29:32 +0200 Message-Id: <20230515161728.076559485@linuxfoundation.org> X-Mailer: git-send-email 2.40.1 In-Reply-To: <20230515161722.146344674@linuxfoundation.org> References: <20230515161722.146344674@linuxfoundation.org> User-Agent: quilt/0.67 Precedence: bulk X-Mailing-List: patches@lists.linux.dev List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit From: Ryusuke Konishi commit a6a491c048882e7e424d407d32cba0b52d9ef2bf upstream. If the disk image that nilfs2 mounts is corrupted and a virtual block address obtained by block lookup for a metadata file is invalid, nilfs_bmap_lookup_at_level() may return the same internal return code as -ENOENT, meaning the block does not exist in the metadata file. This duplication of return codes confuses nilfs_mdt_get_block(), causing it to read and create a metadata block indefinitely. In particular, if this happens to the inode metadata file, ifile, semaphore i_rwsem can be left held, causing task hangs in lock_mount. Fix this issue by making nilfs_bmap_lookup_at_level() treat virtual block address translation failures with -ENOENT as metadata corruption instead of returning the error code. Link: https://lkml.kernel.org/r/20230430193046.6769-1-konishi.ryusuke@gmail.com Signed-off-by: Ryusuke Konishi Tested-by: Ryusuke Konishi Reported-by: syzbot+221d75710bde87fa0e97@syzkaller.appspotmail.com Link: https://syzkaller.appspot.com/bug?extid=221d75710bde87fa0e97 Cc: Signed-off-by: Andrew Morton Signed-off-by: Greg Kroah-Hartman --- fs/nilfs2/bmap.c | 16 ++++++++++++---- 1 file changed, 12 insertions(+), 4 deletions(-) --- a/fs/nilfs2/bmap.c +++ b/fs/nilfs2/bmap.c @@ -67,20 +67,28 @@ int nilfs_bmap_lookup_at_level(struct ni down_read(&bmap->b_sem); ret = bmap->b_ops->bop_lookup(bmap, key, level, ptrp); - if (ret < 0) { - ret = nilfs_bmap_convert_error(bmap, __func__, ret); + if (ret < 0) goto out; - } + if (NILFS_BMAP_USE_VBN(bmap)) { ret = nilfs_dat_translate(nilfs_bmap_get_dat(bmap), *ptrp, &blocknr); if (!ret) *ptrp = blocknr; + else if (ret == -ENOENT) { + /* + * If there was no valid entry in DAT for the block + * address obtained by b_ops->bop_lookup, then pass + * internal code -EINVAL to nilfs_bmap_convert_error + * to treat it as metadata corruption. + */ + ret = -EINVAL; + } } out: up_read(&bmap->b_sem); - return ret; + return nilfs_bmap_convert_error(bmap, __func__, ret); } int nilfs_bmap_lookup_contig(struct nilfs_bmap *bmap, __u64 key, __u64 *ptrp,