[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Wed, 1 Jan 2014 16:34:30 +0000
From: Ben Hutchings <bhutchings@...arflare.com>
To: Amir Vadai <amirv@...lanox.com>
CC: "David S. Miller" <davem@...emloft.net>, <netdev@...r.kernel.org>,
"Or Gerlitz" <ogerlitz@...lanox.com>,
Yuval Atias <yuvala@...lanox.com>
Subject: Re: [PATCH net-next 1/2] net/mlx4_en: Use affinity hint
On Wed, 2014-01-01 at 15:05 +0200, Amir Vadai wrote:
> From: Yuval Atias <yuvala@...lanox.com>
>
> The “affinity hint” mechanism is used by the user space
> daemon, irqbalancer, to indicate a preferred CPU mask for irqs.
> Irqbalancer can use this hint to balance the irqs between the
> cpus indicated by the mask.
>
> We wish the HCA to preferentially map the IRQs it uses to numa cores
> close to it.
> To accomplish this, we use affinity hint: first we map IRQs to “close”
> numa cores.
> If these are exhausted, the remaining IRQs are mapped to “far” numa
> cores.
>
> Signed-off-by: Yuval Atias <yuvala@...lanox.com>
[...]
> --- a/drivers/net/ethernet/mellanox/mlx4/en_netdev.c
> +++ b/drivers/net/ethernet/mellanox/mlx4/en_netdev.c
> @@ -1532,6 +1532,45 @@ static void mlx4_en_linkstate(struct work_struct *work)
> mutex_unlock(&mdev->state_lock);
> }
>
> +static int mlx4_en_rings_affinity_hint(struct mlx4_en_priv *priv ,
> + int *affinity_hint_mapping)
> +{
> + int cpu;
> + int cores_arr_index = 0;
> + cpumask_t *numa_cores_mask;
> + cpumask_t *non_numa_cores_mask;
The types of these masks must be cpumask_var_t.
> + if (!zalloc_cpumask_var(&non_numa_cores_mask, GFP_KERNEL)) {
> + en_err(priv, "Failed allocating cores mask\n");
> + return -EINVAL;
> + }
> + numa_cores_mask = (cpumask_t *)cpumask_of_node(
> + priv->mdev->dev->numa_node);
> + if (!cpumask_and(numa_cores_mask, cpu_online_mask, numa_cores_mask))
> + en_warn(priv, "Failed to find online cores for local numa\n");
This changes the CPU mask of your device's NUMA node. Don't do that.
You need to copy it to numa_cores_mask and then start changing it.
[...]
> @@ -1557,6 +1596,9 @@ int mlx4_en_start_port(struct net_device *dev)
> memset(&priv->ethtool_rules[0], 0,
> sizeof(struct ethtool_flow_id) * MAX_NUM_OF_FS_RULES);
>
> + affinity_hint_mapping = kzalloc(num_online_cpus() * sizeof(int),
> + GFP_KERNEL);
[...]
What makes you think that the online CPU IDs are all contiguous starting
at 0? The correct array size is nr_cpu_ids, if I remember correctly.
Ben.
--
Ben Hutchings, Staff Engineer, Solarflare
Not speaking for my employer; that's the marketing department's job.
They asked us to note that Solarflare product names are trademarked.
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Powered by blists - more mailing lists