Re: [PATCH] btrfs: extract the parity scrub code into a helper

[Qu Wenruo <wqu@suse.com>]

在 2025/11/6 19:32, David Sterba 写道:
> On Thu, Nov 06, 2025 at 01:32:05PM +1030, Qu Wenruo wrote:
>> The function scrub_raid56_parity_stripe() is handling the partity stripe
>> by the following steps:
>>
>> - Scrub each data stripes
>>    And make sure everything is fine in each data stripe
>>
>> - Cache the data stripe into the raid bio
>>
>> - Use the cached raid bio to scrub the target parity stripe
>>
>> Extract the last two steps into a new helper,
>> scrub_radi56_cached_parity(), as a cleanup and make the error handling
>> more straightforward.
>>
>> With the following minor cleanups:
>>
>> - Use on-stack bio structure
>>    The bio is always empty thus we do not need any bio vector nor the
>>    block device. Thus there is no need to allocate a bio, the on-stack
>>    one is more than enough to cut it.
>>
>> - Remove the unnecessary btrfs_put_bioc() call if btrfs_map_block()
>>    failed
>>    If btrfs_map_block() is failed, @bioc_ret will not be touched thus
>>    there is no need to call btrfs_put_bioc() in this case.
>>
>> - Use a proper out: tag to do the cleanup
>>    Now the error cleanup is much shorter and simpler, just
>>    btrfs_bio_counter_dec() and bio_uninit().
>>
>> Signed-off-by: Qu Wenruo <wqu@suse.com>
>> ---
>>   fs/btrfs/scrub.c | 90 ++++++++++++++++++++++++++++--------------------
>>   1 file changed, 52 insertions(+), 38 deletions(-)
>>
>> diff --git a/fs/btrfs/scrub.c b/fs/btrfs/scrub.c
>> index e3612202ba55..8c360d941bd5 100644
>> --- a/fs/btrfs/scrub.c
>> +++ b/fs/btrfs/scrub.c
>> @@ -2113,6 +2113,56 @@ static int should_cancel_scrub(const struct scrub_ctx *sctx)
>>   	return 0;
>>   }
>>   
>> +static int scrub_raid56_cached_parity(struct scrub_ctx *sctx,
>> +				      struct btrfs_device *scrub_dev,
>> +				      struct btrfs_chunk_map *map,
>> +				      u64 full_stripe_start,
>> +				      unsigned long *extent_bitmap)
>> +{
>> +	DECLARE_COMPLETION_ONSTACK(io_done);
>> +	struct btrfs_fs_info *fs_info = sctx->fs_info;
>> +	struct btrfs_io_context *bioc = NULL;
>> +	struct btrfs_raid_bio *rbio;
>> +	struct bio bio;
>> +	const int data_stripes = nr_data_stripes(map);
>> +	u64 length = btrfs_stripe_nr_to_offset(data_stripes);
>> +	int ret;
>> +
>> +	bio_init(&bio, NULL, NULL, 0, REQ_OP_READ);
>> +	bio.bi_iter.bi_sector = full_stripe_start >> SECTOR_SHIFT;
>> +	bio.bi_private = &io_done;
>> +	bio.bi_end_io = raid56_scrub_wait_endio;
>> +
>> +	btrfs_bio_counter_inc_blocked(fs_info);
>> +	ret = btrfs_map_block(fs_info, BTRFS_MAP_WRITE, full_stripe_start,
>> +			      &length, &bioc, NULL, NULL);
>> +	if (ret < 0)
>> +		goto out;
>> +	/* For RAID56 write there must be an @bioc allocated. */
>> +	ASSERT(bioc);
>> +	rbio = raid56_parity_alloc_scrub_rbio(&bio, bioc, scrub_dev, extent_bitmap,
>> +				BTRFS_STRIPE_LEN >> fs_info->sectorsize_bits);
>> +	btrfs_put_bioc(bioc);
>> +	if (!rbio) {
>> +		ret = -ENOMEM;
>> +		goto out;
>> +	}
>> +	/* Use the recovered stripes as cache to avoid read them from disk again. */
>> +	for (int i = 0; i < data_stripes; i++) {
>> +		struct scrub_stripe *stripe = &sctx->raid56_data_stripes[i];
>> +
>> +		raid56_parity_cache_data_folios(rbio, stripe->folios,
>> +				full_stripe_start + (i << BTRFS_STRIPE_LEN_SHIFT));
>> +	}
>> +	raid56_parity_submit_scrub_rbio(rbio);
>> +	wait_for_completion_io(&io_done);
>> +	ret = blk_status_to_errno(bio.bi_status);
>> +out:
>> +	btrfs_bio_counter_dec(fs_info);
>> +	bio_uninit(&bio);
>> +	return ret;
>> +}
>> +
>>   static int scrub_raid56_parity_stripe(struct scrub_ctx *sctx,
>>   				      struct btrfs_device *scrub_dev,
>>   				      struct btrfs_block_group *bg,
>> @@ -2121,16 +2171,12 @@ static int scrub_raid56_parity_stripe(struct scrub_ctx *sctx,
>>   {
>>   	DECLARE_COMPLETION_ONSTACK(io_done);
> 
> This should be deleted as well, as it's in scrub_raid56_cached_parity()

Right, will sent out a v2 just in case.
As I mostly rely on b4 to handle the tags, only updating it locally 
won't make it persistent here.

Thanks,
Qu

> 
> The stack meter says that the new function adds 240 bytes (and has
> dynamic stack size) while scrub_raid56_parity_stripe() shrinks only by
> 24 bytes. So this basically adds 240 - 24 = 216 bytes to the stack.
> 
> With the completion removed is another -32 bytes it's down to 184. The
> on-stack bio is 112 bytes from that, 184 - 112 = 72 for remaining
> variables.