[ofa-general] Re: [PATCH 03/12 -Rev2] dev.c changes.

[Evgeniy Polyakov]

Hi Krishna.

On Sun, Jul 22, 2007 at 02:35:25PM +0530, Krishna Kumar (krkumar2 at in.ibm.com) wrote:
> diff -ruNp org/net/core/dev.c rev2/net/core/dev.c
> --- org/net/core/dev.c	2007-07-20 07:49:28.000000000 +0530
> +++ rev2/net/core/dev.c	2007-07-21 23:08:33.000000000 +0530
> @@ -875,6 +875,48 @@ void netdev_state_change(struct net_devi
>  	}
>  }
>  
> +/*
> + * dev_change_tx_batching - Enable or disable batching for a driver that
> + * supports batching.
> + */
> +int dev_change_tx_batching(struct net_device *dev, unsigned long new_batch_skb)
> +{
> +	int ret;
> +
> +	if (!dev->hard_start_xmit_batch) {
> +		/* Driver doesn't support skb batching */
> +		ret = -ENOTSUPP;
> +		goto out;
> +	}
> +
> +	/* Handle invalid argument */
> +	if (new_batch_skb < 0) {
> +		ret = -EINVAL;
> +		goto out;
> +	}
> +
> +	ret = 0;
> +
> +	/* Check if new value is same as the current */
> +	if (!!(dev->features & NETIF_F_BATCH_ON) == !!new_batch_skb)
> +		goto out;

o_O

Scratched head for too long before understood what it means :)

> +	spin_lock(&dev->queue_lock);
> +	if (new_batch_skb) {
> +		dev->features |= NETIF_F_BATCH_ON;
> +		dev->tx_queue_len >>= 1;
> +	} else {
> +		if (!skb_queue_empty(&dev->skb_blist))
> +			skb_queue_purge(&dev->skb_blist);
> +		dev->features &= ~NETIF_F_BATCH_ON;
> +		dev->tx_queue_len <<= 1;
> +	}
> +	spin_unlock(&dev->queue_lock);

Hmm, should this also stop interrupts?

-- 
	Evgeniy Polyakov