Re: [PATCH v2 4/9] jbd2: jump to new copy_done tag when b_frozen_data is created concurrently

From: Jan Kara
Date: Tue May 14 2024 - 19:43:00 EST


On Tue 14-05-24 19:24:33, Kemeng Shi wrote:
> If b_frozen_data is created concurrently, we can update new_folio and
> new_offset with b_frozen_data and then move forward
>
> Signed-off-by: Kemeng Shi <shikemeng@xxxxxxxxxxxxxxx>

Looks good. Feel free to add:

Reviewed-by: Jan Kara <jack@xxxxxxx>

Honza

> ---
> fs/jbd2/journal.c | 13 ++++++-------
> 1 file changed, 6 insertions(+), 7 deletions(-)
>
> diff --git a/fs/jbd2/journal.c b/fs/jbd2/journal.c
> index 714e2ef0115a..5fb5062cf7ae 100644
> --- a/fs/jbd2/journal.c
> +++ b/fs/jbd2/journal.c
> @@ -351,7 +351,6 @@ int jbd2_journal_write_metadata_buffer(transaction_t *transaction,
> atomic_set(&new_bh->b_count, 1);
>
> spin_lock(&jh_in->b_state_lock);
> -repeat:
> /*
> * If a new transaction has already done a buffer copy-out, then
> * we use that version of the data for the commit.
> @@ -399,22 +398,22 @@ int jbd2_journal_write_metadata_buffer(transaction_t *transaction,
> spin_lock(&jh_in->b_state_lock);
> if (jh_in->b_frozen_data) {
> jbd2_free(tmp, bh_in->b_size);
> - goto repeat;
> + goto copy_done;
> }
>
> jh_in->b_frozen_data = tmp;
> memcpy_from_folio(tmp, new_folio, new_offset, bh_in->b_size);
> -
> - new_folio = virt_to_folio(tmp);
> - new_offset = offset_in_folio(new_folio, tmp);
> - done_copy_out = 1;
> -
> /*
> * This isn't strictly necessary, as we're using frozen
> * data for the escaping, but it keeps consistency with
> * b_frozen_data usage.
> */
> jh_in->b_frozen_triggers = jh_in->b_triggers;
> +
> +copy_done:
> + new_folio = virt_to_folio(jh_in->b_frozen_data);
> + new_offset = offset_in_folio(new_folio, jh_in->b_frozen_data);
> + done_copy_out = 1;
> }
>
> /*
> --
> 2.30.0
>
--
Jan Kara <jack@xxxxxxxx>
SUSE Labs, CR