On Thu, Feb 28, 2013 at 10:45:29PM +0400, Dmitry Monakhov wrote:
> On Fri, 1 Mar 2013 02:26:44 +0800, Zheng Liu <gnehzuil.liu@xxxxxxxxx> wrote:
> > From: Zheng Liu <wenqing.lz@xxxxxxxxxx>
> >
> > When we try to split an extent, this extent could be zeroed out and mark
> > as initialized. But we don't know this in ext4_map_blocks because it
> > only returns a allocated length of extent. Meanwhile we will mark this
> > extent as uninitialized because we only check m_flags.
> >
> > This commit update extent status tree when we try to split an unwritten
> > extent. We don't need to worry about the status of this extent because
> > we always mark it as initialized.
> Your solution seems too complex, why not just update es_cache inside
> ext4_ext_zeroout() like this:
Honestly this is my first solution. But I don't think we can update
es_cache in ext4_ext_zeroout(). The reason is that when
ext4_ext_zerouout is called ext4_ext_dirty hasn't been called. That
means that extent tree on disk is not changed. We can't update es_cache
before metadata is changed. Otherwise the assertation in
ext4_es_insert_extent_check() will fail. Am I miss something?
Thanks
- Zheng
>
> From e0565336c9ea540b0297781196453b358c3c3313 Mon Sep 17 00:00:00 2001
> From: Dmitry Monakhov <dmonakhov@xxxxxxxxxx>
> Date: Thu, 28 Feb 2013 22:40:14 +0400
> Subject: [PATCH] ext4: update status tree on extent zeroout
>
>
> Signed-off-by: Dmitry Monakhov <dmonakhov@xxxxxxxxxx>
> ---
> fs/ext4/extents.c | 4 ++++
> include/trace/events/ext4.h | 27 +++++++++++++++++++++++++++
> 2 files changed, 31 insertions(+), 0 deletions(-)
>
> diff --git a/fs/ext4/extents.c b/fs/ext4/extents.c
> index c89e850..15cc324 100644
> --- a/fs/ext4/extents.c
> +++ b/fs/ext4/extents.c
> @@ -2885,12 +2885,16 @@ static int ext4_ext_zeroout(struct inode *inode, struct ext4_extent *ex)
> unsigned int ee_len;
> int ret;
>
> + trace_ext4_ext_zeroout(inode, ex);
> ee_len = ext4_ext_get_actual_len(ex);
> ee_pblock = ext4_ext_pblock(ex);
>
> ret = sb_issue_zeroout(inode->i_sb, ee_pblock, ee_len, GFP_NOFS);
> if (ret > 0)
> ret = 0;
> + if (!ret)
> + ret = ext4_es_insert_extent(inode, le32_to_cpu(ex->ee_block),
> + ee_len, ee_pblock, EXTENT_STATUS_WRITTEN);
>
> return ret;
> }
> diff --git a/include/trace/events/ext4.h b/include/trace/events/ext4.h
> index c0457c0..cd750a6 100644
> --- a/include/trace/events/ext4.h
> +++ b/include/trace/events/ext4.h
> @@ -2092,6 +2092,33 @@ TRACE_EVENT(ext4_ext_remove_space_done,
> (unsigned short) __entry->eh_entries)
> );
>
> +TRACE_EVENT(ext4_ext_zeroout,
> + TP_PROTO(struct inode *inode, struct ext4_extent *ex),
> +
> + TP_ARGS(inode, ex),
> +
> + TP_STRUCT__entry(
> + __field( dev_t, dev )
> + __field( ino_t, ino )
> + __field( ext4_lblk_t, u_lblk )
> + __field( unsigned, u_len )
> + __field( ext4_fsblk_t, u_pblk )
> + ),
> +
> + TP_fast_assign(
> + __entry->dev = inode->i_sb->s_dev;
> + __entry->ino = inode->i_ino;
> + __entry->u_lblk = le32_to_cpu(ex->ee_block);
> + __entry->u_len = ext4_ext_get_actual_len(ex);
> + __entry->u_pblk = ext4_ext_pblock(ex);
> + ),
> +
> + TP_printk("dev %d,%d ino %lu ee_lblk %u ee_len %u ee_pblk %llu",
> + MAJOR(__entry->dev), MINOR(__entry->dev),
> + (unsigned long) __entry->ino,
> + __entry->u_lblk, __entry->u_len, __entry->u_pblk)
> +);
> +
> TRACE_EVENT(ext4_es_insert_extent,
> TP_PROTO(struct inode *inode, struct extent_status *es),
>
> --
> 1.7.1
>
>
--
To unsubscribe from this list: send the line "unsubscribe linux-ext4" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
