On Wed 10-02-21 10:57:39, Alexander Lochmann wrote:
> Some members of transaction_t are allowed to be read without
> any lock being held if consistency doesn't matter.
> Based on LockDoc's findings, we extended the locking
> documentation of those members.
> Each one of them is marked with a short comment:
> "no lock for quick racy checks".
> 
> Signed-off-by: Alexander Lochmann <[email protected]>
> Signed-off-by: Horst Schirmeier <[email protected]>

Thanks for the patch! Some comments below...

> ---
>  include/linux/jbd2.h | 8 ++++----
>  1 file changed, 4 insertions(+), 4 deletions(-)
> 
> diff --git a/include/linux/jbd2.h b/include/linux/jbd2.h
> index 99d3cd051ac3..18f77d9b1745 100644
> --- a/include/linux/jbd2.h
> +++ b/include/linux/jbd2.h
> @@ -594,18 +594,18 @@ struct transaction_s
>        */
>       unsigned long           t_log_start;
>  
> -     /* Number of buffers on the t_buffers list [j_list_lock] */
> +     /* Number of buffers on the t_buffers list [j_list_lock, no lock for 
> quick racy checks] */
>       int                     t_nr_buffers;

So this case is actually somewhat different now that I audited the uses.
There are two types of users - commit code (fs/jbd2/commit.c) and others.
Other users properly use j_list_lock to access t_nr_buffers. Commit code
does not use any locks because committing transaction is fully in
ownership of the jbd2 thread and all other users need to check & wait for
commit to be finished before doing anything with the transaction's buffers.

>       /*
>        * Doubly-linked circular list of all buffers reserved but not yet
> -      * modified by this transaction [j_list_lock]
> +      * modified by this transaction [j_list_lock, no lock for quick racy 
> checks]
>        */
>       struct journal_head     *t_reserved_list;
>
>       /*
>        * Doubly-linked circular list of all metadata buffers owned by this
> -      * transaction [j_list_lock]
> +      * transaction [j_list_lock, no lock for quick racy checks]
>        */
>       struct journal_head     *t_buffers;
> 
> @@ -631,7 +631,7 @@ struct transaction_s
>       /*
>        * Doubly-linked circular list of metadata buffers being shadowed by log
>        * IO.  The IO buffers on the iobuf list and the shadow buffers on this
> -      * list match each other one for one at all times. [j_list_lock]
> +      * list match each other one for one at all times. [j_list_lock, no 
> lock for quick racy checks]
>        */
>       struct journal_head     *t_shadow_list;

The above three cases are the same as t_reserved_list.

                                                                Honza
-- 
Jan Kara <[email protected]>
SUSE Labs, CR

Reply via email to