Re: [PATCH tip/core/rcu 1/4] srcu: Fix __call_srcu()/process_srcu() datarace

[Peter Zijlstra]

On Fri, Feb 14, 2020 at 04:29:29PM -0800, paulmck@xxxxxxxxxx wrote:
> From: "Paul E. McKenney" <paulmck@xxxxxxxxxx>
> 
> The srcu_node structure's ->srcu_gp_seq_needed_exp field is accessed
> locklessly, so updates must use WRITE_ONCE().  This commit therefore
> adds the needed WRITE_ONCE() invocations.
> 
> This data race was reported by KCSAN.  Not appropriate for backporting
> due to failure being unlikely.

This does indeed look like there can be a failure; but this Changelog
fails to describe an actual problem.

> Signed-off-by: Paul E. McKenney <paulmck@xxxxxxxxxx>
> ---
>  kernel/rcu/srcutree.c | 4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/kernel/rcu/srcutree.c b/kernel/rcu/srcutree.c
> index 657e6a7..b1edac9 100644
> --- a/kernel/rcu/srcutree.c
> +++ b/kernel/rcu/srcutree.c
> @@ -550,7 +550,7 @@ static void srcu_gp_end(struct srcu_struct *ssp)
>  		snp->srcu_have_cbs[idx] = gpseq;
>  		rcu_seq_set_state(&snp->srcu_have_cbs[idx], 1);
>  		if (ULONG_CMP_LT(snp->srcu_gp_seq_needed_exp, gpseq))
> -			snp->srcu_gp_seq_needed_exp = gpseq;
> +			WRITE_ONCE(snp->srcu_gp_seq_needed_exp, gpseq);
>  		mask = snp->srcu_data_have_cbs[idx];
>  		snp->srcu_data_have_cbs[idx] = 0;
>  		spin_unlock_irq_rcu_node(snp);
> @@ -660,7 +660,7 @@ static void srcu_funnel_gp_start(struct srcu_struct *ssp, struct srcu_data *sdp,
>  		if (snp == sdp->mynode)
>  			snp->srcu_data_have_cbs[idx] |= sdp->grpmask;
>  		if (!do_norm && ULONG_CMP_LT(snp->srcu_gp_seq_needed_exp, s))
> -			snp->srcu_gp_seq_needed_exp = s;
> +			WRITE_ONCE(snp->srcu_gp_seq_needed_exp, s);
>  		spin_unlock_irqrestore_rcu_node(snp, flags);
>  	}
>  
> -- 
> 2.9.5
>