linux-toradex.git/fs/btrfs/compression.h, branch v6.7

btrfs: pass an ordered_extent to btrfs_submit_compressed_write

2023-06-19T11:59:36+00:00

btrfs_submit_compressed_write always operates on a single ordered_extent.
Make that explicit by using btrfs_alloc_ordered_extent in the callers
and passing the ordered_extent to btrfs_submit_compressed_write.  This
will help with storing and ordered_extent pointer in the btrfs_bio in
subsequent patches.

Reviewed-by: Johannes Thumshirn 
Reviewed-by: Josef Bacik 
Signed-off-by: Christoph Hellwig 
Reviewed-by: David Sterba 
Signed-off-by: David Sterba

btrfs: remove the mirror_num argument to btrfs_submit_compressed_read

2023-06-19T11:59:27+00:00

Given that read recovery for data I/O is handled in the storage layer,
the mirror_num argument to btrfs_submit_compressed_read is always 0,
so remove it.

Reviewed-by: Johannes Thumshirn 
Reviewed-by: Qu Wenruo 
Reviewed-by: Josef Bacik 
Signed-off-by: Christoph Hellwig 
Reviewed-by: David Sterba 
Signed-off-by: David Sterba

btrfs: move kthread_associate_blkcg out of btrfs_submit_compressed_write

2023-04-17T16:01:22+00:00

btrfs_submit_compressed_write should not have to care if it is called
from a helper thread or not.  Move the kthread_associate_blkcg handling
into submit_one_async_extent, as that is the one caller that needs it.
Also move the assignment of REQ_CGROUP_PUNT into cow_file_range_async,
as that is the routine that sets up the helper thread offload.

Signed-off-by: Christoph Hellwig 
Reviewed-by: David Sterba 
Signed-off-by: David Sterba

btrfs: store a pointer to the original btrfs_bio in struct compressed_bio

2023-04-17T16:01:17+00:00

The original bio must be a btrfs_bio, so store a pointer to the
btrfs_bio for better type checking.

Reviewed-by: Anand Jain 
Reviewed-by: Johannes Thumshirn 
Reviewed-by: Qu Wenruo 
Signed-off-by: Christoph Hellwig 
Signed-off-by: David Sterba

btrfs: pass a btrfs_bio to btrfs_submit_compressed_read

2023-04-17T16:01:17+00:00

btrfs_submit_compressed_read expects the bio passed to it to be embedded
into a btrfs_bio structure.  Pass the btrfs_bio directly to increase type
safety and make the code self-documenting.

Reviewed-by: Anand Jain 
Reviewed-by: Johannes Thumshirn 
Reviewed-by: Qu Wenruo 
Signed-off-by: Christoph Hellwig 
Signed-off-by: David Sterba

btrfs: embed a btrfs_bio into struct compressed_bio

2023-04-17T16:01:14+00:00

Embed a btrfs_bio into struct compressed_bio.  This avoids potential
(so far theoretical) deadlocks due to nesting of btrfs_bioset allocations
for the original read bio and the compressed bio, and avoids an extra
memory allocation in the I/O path.

Reviewed-by: Johannes Thumshirn 
Signed-off-by: Christoph Hellwig 
Reviewed-by: David Sterba 
Signed-off-by: David Sterba

btrfs: split zone append bios in btrfs_submit_bio

2023-02-15T18:38:53+00:00

The current btrfs zoned device support is a little cumbersome in the data
I/O path as it requires the callers to not issue I/O larger than the
supported ZONE_APPEND size of the underlying device.  This leads to a lot
of extra accounting.  Instead change btrfs_submit_bio so that it can take
write bios of arbitrary size and form from the upper layers, and just
split them internally to the ZONE_APPEND queue limits.  Then remove all
the upper layer warts catering to limited write sized on zoned devices,
including the extra refcount in the compressed_bio.

Reviewed-by: Josef Bacik 
Reviewed-by: Johannes Thumshirn 
Signed-off-by: Christoph Hellwig 
Signed-off-by: David Sterba

btrfs: constify input buffer parameter in compression code

2022-12-05T17:00:55+00:00

The input buffers passed down to compression must never be changed,
switch type to u8 as it's a raw byte buffer and use const.

Reviewed-by: Anand Jain 
Signed-off-by: David Sterba

btrfs: add blk_types.h include to compression.h

2022-12-05T17:00:45+00:00

When moving the printk messages into their own file I got a compiler
error because the includes grabbed compression.h, but nothing pulled in
the blk_types.h dependency that compression.h has because it uses
blkstatus_t.  Add blk_types.h to compression.h so that this sort of
thing doesn't happen in the future.

Reviewed-by: Johannes Thumshirn 
Signed-off-by: Josef Bacik 
Reviewed-by: David Sterba 
Signed-off-by: David Sterba

btrfs: make module init/exit match their sequence

2022-12-05T17:00:40+00:00

[BACKGROUND]
In theory init_btrfs_fs() and exit_btrfs_fs() should match their
sequence, thus normally they should look like this:

    init_btrfs_fs()   |   exit_btrfs_fs()
----------------------+------------------------
    init_A();         |
    init_B();         |
    init_C();         |
                      |   exit_C();
                      |   exit_B();
                      |   exit_A();

So is for the error path of init_btrfs_fs().

But it's not the case, some exit functions don't match their init
functions sequence in init_btrfs_fs().

Furthermore in init_btrfs_fs(), we need to have a new error label for
each new init function we added.  This is not really expandable,
especially recently we may add several new functions to init_btrfs_fs().

[ENHANCEMENT]
The patch will introduce the following things to enhance the situation:

- struct init_sequence
  Just a wrapper of init and exit function pointers.

  The init function must use int type as return value, thus some init
  functions need to be updated to return 0.

  The exit function can be NULL, as there are some init sequence just
  outputting a message.

- struct mod_init_seq[] array
  This is a const array, recording all the initialization we need to do
  in init_btrfs_fs(), and the order follows the old init_btrfs_fs().

- bool mod_init_result[] array
  This is a bool array, recording if we have initialized one entry in
  mod_init_seq[].

  The reason to split mod_init_seq[] and mod_init_result[] is to avoid
  section mismatch in reference.

  All init function are in .init.text, but if mod_init_seq[] records
  the @initialized member it can no longer be const, thus will be put
  into .data section, and cause modpost warning.

For init_btrfs_fs() we just call all init functions in their order in
mod_init_seq[] array, and after each call, setting corresponding
mod_init_result[] to true.

For exit_btrfs_fs() and error handling path of init_btrfs_fs(), we just
iterate mod_init_seq[] in reverse order, and skip all uninitialized
entry.

With this patch, init_btrfs_fs()/exit_btrfs_fs() will be much easier to
expand and will always follow the strict order.

Signed-off-by: Qu Wenruo 
Signed-off-by: David Sterba