From: Theodore Tso Subject: Re: [PATCH 2/2] ext4: journal superblock modifications in ext4_statfs() Date: Sat, 14 Nov 2009 22:29:41 -0500 Message-ID: <20091115032941.GB4323@mit.edu> References: <4AF4A429.7090507@redhat.com> <6BDA2C94-6FA5-48EE-9E68-56BDFC4B558A@sun.com> <20091108214804.GC7592@mit.edu> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: Eric Sandeen , ext4 development To: Andreas Dilger Return-path: Received: from thunk.org ([69.25.196.29]:45115 "EHLO thunker.thunk.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751787AbZKOD3l (ORCPT ); Sat, 14 Nov 2009 22:29:41 -0500 Content-Disposition: inline In-Reply-To: Sender: linux-ext4-owner@vger.kernel.org List-ID: On Sun, Nov 08, 2009 at 09:41:28PM -0700, Andreas Dilger wrote: > > The reason we added this was for running a read-only e2fsck on a > filesystem without getting spurious errors just because the superblock > summaries were incorrect. The other alternative is to change e2fsck > so that it doesn't consider just a block/inode summary an error. > We've been doing this for a while --- e2fsprogs 1.34, since April 2003. In e2fsck/super.c:check_super_block(): /* * Update the global counts from the block group counts. This * is needed for an experimental patch which eliminates * locking the entire filesystem when allocating blocks or * inodes; if the filesystem is not unmounted cleanly, the * global counts may not be accurate. */ if ((free_blocks != sb->s_free_blocks_count) || (free_inodes != sb->s_free_inodes_count)) { if (ctx->options & E2F_OPT_READONLY) ext2fs_unmark_valid(fs); else { sb->s_free_blocks_count = free_blocks; sb->s_free_inodes_count = free_inodes; ext2fs_mark_super_dirty(fs); } } I suppose I should edit out the bit about "experimental patch", which might have been true in 2003, but it's certainly not true any more. Hence, I think it's safe to eliminate the updates in ext4_statfs() altogether. - Ted commit 48efef97e03811e4492607ed7a7aefb35e2c0c00 Author: Theodore Ts'o Date: Sat Nov 14 21:37:52 2009 -0500 ext4: Don't update the superblock in ext4_statfs() commit a71ce8c6c9bf269b192f352ea555217815cf027e updated ext4_statfs() to update the on-disk superblock counters, but modified this buffer directly without any journaling of the change. This is one of the accesses that was causing the crc errors in journal replay as seen in kernel.org bugzilla #14354. Signed-off-by: "Theodore Ts'o" diff --git a/fs/ext4/super.c b/fs/ext4/super.c index 8662b2e..3cd519b 100644 --- a/fs/ext4/super.c +++ b/fs/ext4/super.c @@ -3668,13 +3668,11 @@ static int ext4_statfs(struct dentry *dentry, struct kstatfs *buf) buf->f_blocks = ext4_blocks_count(es) - sbi->s_overhead_last; buf->f_bfree = percpu_counter_sum_positive(&sbi->s_freeblocks_counter) - percpu_counter_sum_positive(&sbi->s_dirtyblocks_counter); - ext4_free_blocks_count_set(es, buf->f_bfree); buf->f_bavail = buf->f_bfree - ext4_r_blocks_count(es); if (buf->f_bfree < ext4_r_blocks_count(es)) buf->f_bavail = 0; buf->f_files = le32_to_cpu(es->s_inodes_count); buf->f_ffree = percpu_counter_sum_positive(&sbi->s_freeinodes_counter); - es->s_free_inodes_count = cpu_to_le32(buf->f_ffree); buf->f_namelen = EXT4_NAME_LEN; fsid = le64_to_cpup((void *)es->s_uuid) ^ le64_to_cpup((void *)es->s_uuid + sizeof(u64));