mbox series

[0/5] btrfs: zoned: remove extent_buffer redirtying

Message ID 20231121-josef-generic-163-v1-0-049e37185841@wdc.com (mailing list archive)
Headers show
Series btrfs: zoned: remove extent_buffer redirtying | expand

Message

Johannes Thumshirn Nov. 21, 2023, 4:32 p.m. UTC
Since the beginning of zoned mode, I've promised Josef to get rid of the
extent_buffer redirtying, but never actually got around to doing so.

Then 2 weeks ago our CI has hit an ASSERT() in this area and I started to look
into it again. After some discussion with Christoph we came to the conclusion
to finally take the time and get rid of the extent_buffer redirtying once and
for all.

Patch one renames EXTENT_BUFFER_NO_CHECK into EXTENT_BUFFER_CANCELLED, because
this fits the new model somewhat better.

Number two sets the cancel bit instead of clearing the dirty bit from a zoned
extent_buffer.

Number three removes the last remaining bits of btrfs_redirty_list_add().

The last two patches in this series are just trivial cleanups I came across
while looking at the code.

---
Johannes Thumshirn (5):
      btrfs: rename EXTENT_BUFFER_NO_CHECK to EXTENT_BUFFER_CANCELLED
      btrfs: zoned: don't clear dirty flag of extent buffer
      btrfs: remove now unneeded btrfs_redirty_list_add
      btrfs: use memset_page instead of opencoding it
      btrfs: reflow btrfs_free_tree_block

 fs/btrfs/disk-io.c     |   4 +-
 fs/btrfs/extent-tree.c | 102 ++++++++++++++++++++++++-------------------------
 fs/btrfs/extent_io.c   |   9 +++--
 fs/btrfs/extent_io.h   |   3 +-
 fs/btrfs/tree-log.c    |   1 -
 fs/btrfs/zoned.c       |  16 --------
 fs/btrfs/zoned.h       |   5 ---
 7 files changed, 60 insertions(+), 80 deletions(-)
---
base-commit: 592afe8e8b7ceee58107757fd29ff3290e6539e3
change-id: 20231120-josef-generic-163-f4df4eab2c98

Best regards,

Comments

Josef Bacik Nov. 21, 2023, 9:01 p.m. UTC | #1
On Tue, Nov 21, 2023 at 08:32:29AM -0800, Johannes Thumshirn wrote:
> Since the beginning of zoned mode, I've promised Josef to get rid of the
> extent_buffer redirtying, but never actually got around to doing so.
> 
> Then 2 weeks ago our CI has hit an ASSERT() in this area and I started to look
> into it again. After some discussion with Christoph we came to the conclusion
> to finally take the time and get rid of the extent_buffer redirtying once and
> for all.
> 
> Patch one renames EXTENT_BUFFER_NO_CHECK into EXTENT_BUFFER_CANCELLED, because
> this fits the new model somewhat better.
> 
> Number two sets the cancel bit instead of clearing the dirty bit from a zoned
> extent_buffer.
> 
> Number three removes the last remaining bits of btrfs_redirty_list_add().
> 
> The last two patches in this series are just trivial cleanups I came across
> while looking at the code.
>

Reviewed-by: Josef Bacik <josef@toxicpanda.com>

Thanks,

Josef