Re: [PATCH] md/raid5: fix possible null-pointer dereferences in raid5_store_group_thread_cnt()

Hi,

在 2025/12/10 15:41, Tuo Li 写道:
> The variable mddev->private is first assigned to conf and then checked:
>
>    conf = mddev->private;
>    if (!conf) ...
>
> If conf is NULL, then mddev->private is also NULL. However, the function
> does not return at this point, and raid5_quiesce() is later called with
> mddev as the argument. Inside raid5_quiesce(), mddev->private is again
> assigned to conf, which is then dereferenced in multiple places, for
> example:
>
>    conf->quiesce = 0;
>    wake_up(&conf->wait_for_quiescent);
>    ...
>
> This can lead to several null-pointer dereferences.
>
> To fix these issues, the function should unlock mddev and return early when
> conf is NULL, following the pattern in raid5_change_consistency_policy().
>
> Signed-off-by: Tuo Li <[email protected]>
> ---
>   drivers/md/raid5.c | 7 ++++---
>   1 file changed, 4 insertions(+), 3 deletions(-)
>
> diff --git a/drivers/md/raid5.c b/drivers/md/raid5.c
> index e57ce3295292..be3f9a127212 100644
> --- a/drivers/md/raid5.c
> +++ b/drivers/md/raid5.c
> @@ -7190,9 +7190,10 @@ raid5_store_group_thread_cnt(struct mddev *mddev, const char *page, size_t len)
>   	raid5_quiesce(mddev, true);
>   
>   	conf = mddev->private;
> -	if (!conf)
> -		err = -ENODEV;
> -	else if (new != conf->worker_cnt_per_group) {
> +	if (!conf) {
> +		mddev_unlock_and_resume(mddev);
> +		return -ENODEV;

+CC Xiao

This is still wrong, please add the NULL check and return early before raid5_quise().
And also add a fix tag:

fa1944bbe622 md/raid5: Wait sync io to finish before changing group cnt

> +	} else if (new != conf->worker_cnt_per_group) {
>   		old_groups = conf->worker_groups;
>   		if (old_groups)
>   			flush_workqueue(raid5_wq);

-- 
Thansk,
Kuai