Hi,
On Sat, Oct 07, 2023 at 11:52:33AM +0800, Wang Jinchao wrote:
> Signed-off-by: Wang Jinchao <wangjinchao@xxxxxxxxxxx>
The Fixes: tag should be near your Signed-off-by:, and the subject of
the patch should be something descriptive like
padata: Fix refcnt handling in padata_free_shell()
Here's some documentation about this:
https://docs.kernel.org/process/5.Posting.html#patch-formatting-and-changelogs
> Acked-by: Daniel Jordan <daniel.m.jordan@xxxxxxxxxx>
> ---
> V3:
> Include Daniel's ack
>
> V2: https://lore.kernel.org/all/ZRTLHY5A+VqIKhA2@fedora/
> To satisfy Sparse, use rcu_dereference_protected.
> Reported-by: kernel test robot <lkp@xxxxxxxxx>
> Closes: https://lore.kernel.org/oe-kbuild-all/202309270829.xHgTOMKw-lkp@xxxxxxxxx/
These two tags can also go near your SoB.
> V1: https://lore.kernel.org/all/ZRE4XvOOhz4HSOgR@fedora/
>
> kernel/padata.c | 6 +++++-
> 1 file changed, 5 insertions(+), 1 deletion(-)
>
> diff --git a/kernel/padata.c b/kernel/padata.c
> index 222d60195de6..79d04a97ded6 100644
> --- a/kernel/padata.c
> +++ b/kernel/padata.c
> @@ -1102,12 +1102,16 @@ EXPORT_SYMBOL(padata_alloc_shell);
> */
> void padata_free_shell(struct padata_shell *ps)
> {
> + struct parallel_data *pd;
> +
> if (!ps)
> return;
>
> mutex_lock(&ps->pinst->lock);
> list_del(&ps->list);
> - padata_free_pd(rcu_dereference_protected(ps->pd, 1));
> + pd = rcu_dereference_protected(ps->pd, 1);
> + if (refcount_dec_and_test(&pd->refcnt))
> + padata_free_pd(rcu_dereference_protected(ps->pd, 1));
As Herbert points out, this version changes the code by
rcu-dereferencing pd twice. Usually previous acks are dropped when
introducing new changes, but you can have it back by only
rcu-dereferencing once and addressing the other comments above.
> mutex_unlock(&ps->pinst->lock);
>
> kfree(ps);
> --
> 2.40.0
>
