From: Greg KH Subject: [patch 26/39] ext4: Fix race between read_block_bitmap() and mark_diskspace_used() Date: Wed, 18 Feb 2009 13:33:03 -0800 Message-ID: <20090218213303.GA19814@kroah.com> References: <20090218212144.965748151@mini.kroah.org> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: Justin Forbes , Zwane Mwaikambo , Theodore Ts'o , Randy Dunlap , Dave Jones , Chuck Wolber , Chris Wedgwood , Michael Krufky , Chuck Ebbert , Domenico Andreoli , Willy Tarreau , Rodrigo Rubira Branco , Jake Edge , Eugene Teo , torvalds@linux-foundation.org, akpm@linux-foundation.org, alan@lxorguk.ukuu.org.uk, linux-ext4@vger.kernel.org, "Aneesh Kumar K.V" To: linux-kernel@vger.kernel.org, stable@kernel.org Return-path: Received: from kroah.org ([198.145.64.141]:56323 "EHLO coco.kroah.org" rhost-flags-OK-OK-OK-FAIL) by vger.kernel.org with ESMTP id S1757217AbZBRVfw (ORCPT ); Wed, 18 Feb 2009 16:35:52 -0500 Content-Disposition: inline; filename="ext4-fix-race-between-read_block_bitmap-and-mark_diskspace_used.patch" In-Reply-To: <20090218213021.GA19814@kroah.com> Sender: linux-ext4-owner@vger.kernel.org List-ID: 2.6.28-stable review patch. If anyone has any objections, please let us know. ------------------ From: Aneesh Kumar K.V (cherry picked from commit e8134b27e351e813414da3b95aa8eac6d3908088) We need to make sure we update the block bitmap and clear EXT4_BG_BLOCK_UNINIT flag with sb_bgl_lock held, since ext4_read_block_bitmap() looks at EXT4_BG_BLOCK_UNINIT to decide whether to initialize the block bitmap each time it is called (introduced by commit c806e68f), and this can race with block allocations in ext4_mb_mark_diskspace_used(). ext4_read_block_bitmap does: spin_lock(sb_bgl_lock(EXT4_SB(sb), block_group)); if (desc->bg_flags & cpu_to_le16(EXT4_BG_BLOCK_UNINIT)) { ext4_init_block_bitmap(sb, bh, block_group, desc); Now on the block allocation side we do mb_set_bits(sb_bgl_lock(sbi, ac->ac_b_ex.fe_group), bitmap_bh->b_data, ac->ac_b_ex.fe_start, ac->ac_b_ex.fe_len); .... spin_lock(sb_bgl_lock(sbi, ac->ac_b_ex.fe_group)); if (gdp->bg_flags & cpu_to_le16(EXT4_BG_BLOCK_UNINIT)) { gdp->bg_flags &= cpu_to_le16(~EXT4_BG_BLOCK_UNINIT); ie on allocation we update the bitmap then we take the sb_bgl_lock and clear the EXT4_BG_BLOCK_UNINIT flag. What can happen is a parallel ext4_read_block_bitmap can zero out the bitmap in between the above mb_set_bits and spin_lock(sb_bg_lock..) The race results in below user visible errors EXT4-fs error (device sdb1): ext4_mb_release_inode_pa: free 100, pa_free 105 EXT4-fs error (device sdb1): mb_free_blocks: double-free of inode 0's block .. Signed-off-by: Aneesh Kumar K.V Signed-off-by: "Theodore Ts'o" Signed-off-by: Greg Kroah-Hartman --- fs/ext4/mballoc.c | 15 ++++++++++----- 1 file changed, 10 insertions(+), 5 deletions(-) --- a/fs/ext4/mballoc.c +++ b/fs/ext4/mballoc.c @@ -1070,7 +1070,10 @@ static void mb_clear_bits(spinlock_t *lo cur += 32; continue; } - mb_clear_bit_atomic(lock, cur, bm); + if (lock) + mb_clear_bit_atomic(lock, cur, bm); + else + mb_clear_bit(cur, bm); cur++; } } @@ -1088,7 +1091,10 @@ static void mb_set_bits(spinlock_t *lock cur += 32; continue; } - mb_set_bit_atomic(lock, cur, bm); + if (lock) + mb_set_bit_atomic(lock, cur, bm); + else + mb_set_bit(cur, bm); cur++; } } @@ -3033,10 +3039,9 @@ ext4_mb_mark_diskspace_used(struct ext4_ } } #endif - mb_set_bits(sb_bgl_lock(sbi, ac->ac_b_ex.fe_group), bitmap_bh->b_data, - ac->ac_b_ex.fe_start, ac->ac_b_ex.fe_len);