Web lists-archives.com

Re: [PATCH 1/2] rcutorture: Add usage warnings for parameters gp_normal and gp_exp




On Wed, Dec 06, 2017 at 06:51:07PM +0800, lianglihao@xxxxxxxxxx wrote:
> From: Lihao Liang <lianglihao@xxxxxxxxxx>
> 
> In rcu_torture_fakewriter(), when param gp_normal is equal to
> gp_exp (even when they are false), both cur_ops->sync() and
> cur_ops->exp_sync() may be invoked. This commit checks that
> if any of them is NULL, it will print out a warning message.
> It also emits a warning if gp_normal is true but cur_ops->sync()
> is NULL.
> 
> Signed-off-by: Lihao Liang <lianglihao@xxxxxxxxxx>

Queued for review and testing, thank you!  Passes light rcutorture
testing.

							Thanx, Paul

> ---
> 
> I think the last else-if block is redundant and I choose to include it
> for completeness. In addition, it may be better to have more informative
> warning messages for the new if-else blocks. Feel free to revise them.
> 
>  kernel/rcu/rcutorture.c | 10 ++++++++++
>  1 file changed, 10 insertions(+)
> 
> diff --git a/kernel/rcu/rcutorture.c b/kernel/rcu/rcutorture.c
> index d1b64dc..d2abebd 100644
> --- a/kernel/rcu/rcutorture.c
> +++ b/kernel/rcu/rcutorture.c
> @@ -937,6 +937,16 @@ rcu_torture_writer(void *arg)
>  		synctype[nsynctypes++] = RTWS_SYNC;
>  	else if (gp_sync && !cur_ops->sync)
>  		pr_alert("%s: gp_sync without primitives.\n", __func__);
> +	if (gp_normal == gp_exp) {
> +		if (!cur_ops->sync)
> +			pr_alert("%s: gp_sync without primitives.\n", __func__);
> +		if (!cur_ops->exp_sync)
> +			pr_alert("%s: gp_exp without primitives.\n", __func__);
> +	} else if (gp_normal && !cur_ops->sync) {
> +		pr_alert("%s: gp_sync without primitives.\n", __func__);
> +	} else if (gp_exp && !cur_ops->exp_sync) {
> +		pr_alert("%s: gp_exp without primitives.\n", __func__);
> +	}
>  	if (WARN_ONCE(nsynctypes == 0,
>  		      "rcu_torture_writer: No update-side primitives.\n")) {
>  		/*
> -- 
> 2.7.4
>