[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-Id: <20150925.124848.1041664902815491015.davem@davemloft.net>
Date: Fri, 25 Sep 2015 12:48:48 -0700 (PDT)
From: David Miller <davem@...emloft.net>
To: madalin.bucur@...escale.com
Cc: netdev@...r.kernel.org, linuxppc-dev@...ts.ozlabs.org,
linux-kernel@...r.kernel.org, scottwood@...escale.com,
igal.liberman@...escale.com, roy.pledge@...escale.com,
ppc@...dchasers.com, joe@...ches.com, pebolle@...cali.nl,
joakim.tjernlund@...nsmode.se, f.fainelli@...il.com, stsp@...t.ru
Subject: Re: [v3 2/8] dpaa_eth: add support for DPAA Ethernet
From: Madalin Bucur <madalin.bucur@...escale.com>
Date: Thu, 24 Sep 2015 18:00:13 +0300
> +#define DPA_NAPI_WEIGHT 64
This is just the default, so simply use "NAPI_POLL_WEIGHT" rather
than defining your own macro unnecessarily for this.
> +static int dpa_eth_priv_stop(struct net_device *net_dev)
> +{
> + int err;
> + struct dpa_priv_s *priv;
> +
> + err = dpa_stop(net_dev);
> + /* Allow NAPI to consume any frame still in the Rx/TxConfirm
> + * ingress queues. This is to avoid a race between the current
> + * context and ksoftirqd which could leave NAPI disabled while
> + * in fact there's still Rx traffic to be processed.
> + */
> + usleep_range(5000, 10000);
A wait never fixes a race. Instead you must use appropriate mutual
exclusion mechanisms to ensure that NAPI is really completed it's
run, and we definitely have a means by which to do that properly.
> + net_dev->hw_features |= (NETIF_F_IP_CSUM | NETIF_F_IPV6_CSUM |
> + NETIF_F_LLTX);
This second line should start exactly at the first column after the
openning parenthesis of the first line.
--
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