On 8/6/20 2:55 AM, Felix Fietkau wrote:
> For some drivers (especially 802.11 drivers), doing a lot of work in the NAPI
> poll function does not perform well. Since NAPI poll is bound to the CPU it
> was scheduled from, we can easily end up with a few very busy CPUs spending
> most of their time in softirq/ksoftirqd and some idle ones.
>
> Introduce threaded NAPI for such drivers based on a workqueue. The API is the
> same except for using netif_threaded_napi_add instead of netif_napi_add.
>
> In my tests with mt76 on MT7621 using threaded NAPI + a thread for tx
> scheduling
> improves LAN->WLAN bridging throughput by 10-50%. Throughput without threaded
> NAPI is wildly inconsistent, depending on the CPU that runs the tx scheduling
> thread.
>
> With threaded NAPI, throughput seems stable and consistent (and higher than
> the best results I got without it).
>
> Based on a patch by Hillf Danton
>
> Cc: Hillf Danton <hdan...@sina.com>
> Signed-off-by: Felix Fietkau <n...@nbd.name>
...
> index e353b822bb15..99233e86f4c5 100644
> --- a/net/core/net-sysfs.c
> +++ b/net/core/net-sysfs.c
> @@ -471,6 +471,47 @@ static ssize_t proto_down_store(struct device *dev,
> }
> NETDEVICE_SHOW_RW(proto_down, fmt_dec);
>
This belongs to a separate patch, with correct attribution.
> +static int change_napi_threaded(struct net_device *dev, unsigned long val)
> +{
> + struct napi_struct *napi;
> +
> + if (list_empty(&dev->napi_list))
> + return -EOPNOTSUPP;
> + list_for_each_entry(napi, &dev->napi_list, dev_list) {
> + if (val)
> + set_bit(NAPI_STATE_THREADED, &napi->state);
> + else
> + clear_bit(NAPI_STATE_THREADED, &napi->state);
> + }
> +
> + return 0;
> +}
> +
> +static ssize_t napi_threaded_store(struct device *dev,
> + struct device_attribute *attr,
> + const char *buf, size_t len)
> +{
> + return netdev_store(dev, attr, buf, len, change_napi_threaded);
> +}
> +
> +static ssize_t napi_threaded_show(struct device *dev,
> + struct device_attribute *attr,
> + char *buf)
> +{
> + struct net_device *netdev = to_net_dev(dev);
> + struct napi_struct *napi;
> + bool enabled = false;
> +
You probably want to use RTNL protection, list could change under us otherwise.
The write side part is protected already in netdev_store()
> + list_for_each_entry(napi, &netdev->napi_list, dev_list) {
> + if (test_bit(NAPI_STATE_THREADED, &napi->state))
> + enabled = true;
> + }
> +
> + return sprintf(buf, fmt_dec, enabled);
> +}
> +DEVICE_ATTR_RW(napi_threaded);
> +
> static ssize_t phys_port_id_show(struct device *dev,
> struct device_attribute *attr, char *buf)
> {
> @@ -563,6 +604,7 @@ static struct attribute *net_class_attrs[]
> __ro_after_init = {
> &dev_attr_tx_queue_len.attr,
> &dev_attr_gro_flush_timeout.attr,
> &dev_attr_napi_defer_hard_irqs.attr,
> + &dev_attr_napi_threaded.attr,
> &dev_attr_phys_port_id.attr,
> &dev_attr_phys_port_name.attr,
> &dev_attr_phys_switch_id.attr,
>
