On Sat, 23 Aug 2014 15:08:35 -0400, "Theodore Ts'o" <tytso@xxxxxxx> wrote: > On Fri, Aug 22, 2014 at 03:32:25PM +0400, Dmitry Monakhov wrote: > > - handle = ext4_journal_start(inode, EXT4_HT_MISC, 4); > > + /* In worst case we have to writeout two nonadjacent unwritten blocks */ > > + credits = ext4_chunk_trans_blocks(inode, 1) * 2 - > > + EXT4_META_TRANS_BLOCKS(inode->i_sb); > > This looks like it would be a massive over-estimate, since it includes > the block group allocation bitmaps, which we wouldn't need to update, > no? Oh. You right. groups metadata/quota should not be affected. > > Wouldn't > > credts = ext4_index_trans_blocks(inode, 1, 1) * 2; Yes, but ext4_index_trans_blocks is statically defined in inode.c, but since zero_range works only for extent based ext4_ext_index_trans_blocks is sufficient. Also we must reserve 2 block for data. So final calculation looks like follows: credits = ext4_ext_index_trans_blocks(inode, 2) + 2
>From ec7e1926e8bb99ccd0eb525efc6666cb8ea36a5e Mon Sep 17 00:00:00 2001 From: Dmitry Monakhov <dmonakhov@xxxxxxxxxx> Date: Mon, 25 Aug 2014 11:12:35 +0400 Subject: [PATCH] ext4_zero_range: fix incorect journal credits reservation v2 Currently we reserve only 4 blocks but in worst case scenario ext4_zero_partial_blocks() may want to zeroout and convert two non adjacent blocks. Signed-off-by: Dmitry Monakhov <dmonakhov@xxxxxxxxxx> --- fs/ext4/extents.c | 6 ++++-- 1 files changed, 4 insertions(+), 2 deletions(-) diff --git a/fs/ext4/extents.c b/fs/ext4/extents.c index 5bf718b..f09dbe7 100644 --- a/fs/ext4/extents.c +++ b/fs/ext4/extents.c @@ -4740,6 +4740,7 @@ static long ext4_zero_range(struct file *file, loff_t offset, loff_t new_size = 0; int ret = 0; int flags; + int credits; int partial; loff_t start, end; ext4_lblk_t lblk; @@ -4839,8 +4840,9 @@ static long ext4_zero_range(struct file *file, loff_t offset, if (ret) goto out_dio; } - - handle = ext4_journal_start(inode, EXT4_HT_MISC, 4); + /* In worst case we have to writeout two nonadjacent unwritten blocks */ + credits = ext4_ext_index_trans_blocks(inode, 2) + 2; + handle = ext4_journal_start(inode, EXT4_HT_MISC, credits); if (IS_ERR(handle)) { ret = PTR_ERR(handle); ext4_std_error(inode->i_sb, ret); -- 1.7.1