[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <4877CE6B.2050805@cosmosbay.com>
Date: Fri, 11 Jul 2008 23:19:39 +0200
From: Eric Dumazet <dada1@...mosbay.com>
To: David Miller <davem@...emloft.net>
Cc: netdev@...r.kernel.org
Subject: Re: [PATCH 1/13]: netdev: Allocate multiple queues for TX.
David Miller a écrit :
> alloc_netdev_mq() now allocates an array of netdev_queue
> structures for TX, based upon the queue_count argument.
>
> Furthermore, all accesses to the TX queues are now vectored
> through the netdev_get_tx_queue() and netdev_for_each_tx_queue()
> interfaces. This makes it easy to grep the tree for all
> things that want to get to a TX queue of a net device.
>
> Problem spots which are not really multiqueue aware yet, and
> only work with one queue, can easily be spotted by grepping
> for all netdev_get_tx_queue() calls that pass in a zero index.
>
> Signed-off-by: David S. Miller <davem@...emloft.net>
> ---
> }
>
> + tx = kzalloc(sizeof(struct netdev_queue) * queue_count, GFP_KERNEL);
> + if (!tx) {
> + printk(KERN_ERR "alloc_netdev: Unable to allocate "
> + "tx qdiscs.\n");
> + kfree(p);
> + return NULL;
> + }
> +
On i386 for example, sizeof(struct netdev_queue) = 0x2c
Shouldnt we make sure each queue sits on its own cache line on CONFIG_SMP ?
Also, dev->_tx and dev->num_tx_queues might share a cache line with dev->rx_queue, was it intended ?
> --- a/include/linux/netdevice.h
> +++ b/include/linux/netdevice.h
> @@ -640,7 +640,9 @@ struct net_device
> unsigned char broadcast[MAX_ADDR_LEN]; /* hw bcast add */
>
> struct netdev_queue rx_queue;
> - struct netdev_queue tx_queue ____cacheline_aligned_in_smp;
> +
> + struct netdev_queue *_tx;
> + unsigned int num_tx_queues;
> unsigned long tx_queue_len; /* Max frames per queue allowed */
--
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