[00/10] A bunch of misc cleanups

Message

Nikolay Borisov July 2, 2020, 1:46 p.m. UTC

Here's an assortment of little quality-of-life patches that I created while
looking into the raid56 code. They should bear no functional changes and have
tested them with xfstest and nothing fell over so should be rather low risk.

Patch 1 moves code in __btrfs_map_block, essentially assigning tgtdev_map/raid_map
closet to where space for them is allocated. This also neccesiated moving the
call to sort_parity_stripes. The end result is (hopefully) slightly easier to
follow __btrfs_map_block.

Next 5 patches cleanup minor things in raid56.c such as removing redundant checks,
making code interacting with bio_list more in line with what the rest of the
kernel is doing. Finally it's using some macros/functions instead of open-coding
them. Really just a bunch of low hanging fruit.

Final 4 patches gradually remove all labels in btrfs_submit_compressed_read.
Current failures can be handled "inline" so to speak, without the need for
extra labels. This likely will change once the BUG_ONs are removed but we are
not there yet.

Nikolay Borisov (10):
  btrfs: Always initialize btrfs_bio::tgtdev_map/raid_map pointers
  btrfs: raid56: Remove redundant check in rbio_add_io_page
  btrfs: raid56: Assign bio in while()
  btrfs: raid56: Remove out label in __raid56_parity_recover
  btrfs: raid56: Use in_range where applicable
  btrfs: raid56: Don't opencode swap()
  btrfs: Remove fail label in check_compressed_csum
  btrfs: Remove fail1 label in btrfs_submit_compressed_read
  btrfs: Remove fail2 label from btrfs_submit_compressed_read
  btrfs: Remove out label in btrfs_submit_compressed_read

 fs/btrfs/compression.c | 48 +++++++++++++-------------------
 fs/btrfs/raid56.c      | 63 ++++++++++--------------------------------
 fs/btrfs/volumes.c     | 34 ++++++++++-------------
 3 files changed, 49 insertions(+), 96 deletions(-)

--
2.17.1

Comments

David Sterba July 3, 2020, 4:21 p.m. UTC | #1

On Thu, Jul 02, 2020 at 04:46:40PM +0300, Nikolay Borisov wrote:
> Here's an assortment of little quality-of-life patches that I created while
> looking into the raid56 code. They should bear no functional changes and have
> tested them with xfstest and nothing fell over so should be rather low risk.
> 
> Patch 1 moves code in __btrfs_map_block, essentially assigning tgtdev_map/raid_map
> closet to where space for them is allocated. This also neccesiated moving the
> call to sort_parity_stripes. The end result is (hopefully) slightly easier to
> follow __btrfs_map_block.
> 
> Next 5 patches cleanup minor things in raid56.c such as removing redundant checks,
> making code interacting with bio_list more in line with what the rest of the
> kernel is doing. Finally it's using some macros/functions instead of open-coding
> them. Really just a bunch of low hanging fruit.
> 
> Final 4 patches gradually remove all labels in btrfs_submit_compressed_read.
> Current failures can be handled "inline" so to speak, without the need for
> extra labels. This likely will change once the BUG_ONs are removed but we are
> not there yet.
> 
> Nikolay Borisov (10):
>   btrfs: Always initialize btrfs_bio::tgtdev_map/raid_map pointers
>   btrfs: raid56: Remove redundant check in rbio_add_io_page
>   btrfs: raid56: Assign bio in while()
>   btrfs: raid56: Remove out label in __raid56_parity_recover
>   btrfs: raid56: Use in_range where applicable
>   btrfs: raid56: Don't opencode swap()
>   btrfs: Remove fail label in check_compressed_csum
>   btrfs: Remove fail1 label in btrfs_submit_compressed_read
>   btrfs: Remove fail2 label from btrfs_submit_compressed_read
>   btrfs: Remove out label in btrfs_submit_compressed_read

Except patches 4, 8, 9, 10, series merged to misc-next. We can have a
look at the label/return cleanups next week, the other cleanups are good
so I don't want to stall this patchset. Thanks.