[PATCH v6 4/4] drivers: net: Add APM X-Gene SoC ethernet driver support.

[Florian Fainelli]

2014-06-17 12:21 GMT-07:00 Iyappan Subramanian <isubramanian at apm.com>:
> On Mon, Jun 16, 2014 at 9:42 PM, David Miller <davem at davemloft.net> wrote:
>> From: Iyappan Subramanian <isubramanian at apm.com>
>> Date: Mon, 16 Jun 2014 17:18:46 -0700
>>
>>> +static netdev_tx_t xgene_enet_start_xmit(struct sk_buff *skb,
>>> +                                      struct net_device *ndev)
>>> +{
>>> +     struct xgene_enet_pdata *pdata = netdev_priv(ndev);
>>> +     struct xgene_enet_desc_ring *tx_ring = pdata->tx_ring;
>>> +     struct xgene_enet_desc_ring *cp_ring = tx_ring->cp_ring;
>>> +     u32 tx_level, cq_level;
>>> +
>>> +     tx_level = xgene_enet_ring_len(tx_ring);
>>> +     cq_level = xgene_enet_ring_len(cp_ring);
>>> +     if (unlikely(tx_level > pdata->tx_qcnt_hi ||
>>> +                  cq_level > pdata->cp_qcnt_hi)) {
>>> +             netif_stop_queue(ndev);
>>> +             return NETDEV_TX_BUSY;
>>> +     }
>>> +
>>> +     if (xgene_enet_setup_tx_desc(tx_ring, skb))
>>> +             return NETDEV_TX_OK;
>>
>> If you return NETDEV_TX_OK, it is your responsibility to deal with the
>> SKB, because you own it.
>>
>> In particular, you have to free the packet.
>
> skb freeing happens in xgene_enet_tx_completion().

I have not looked at what xgenet_enet_setup_tx_desc() does, but
David's point is that, if this function fails and has not yet been
able to queue the SKB for transmission at the HW, you should make sure
that you free the SKB. Is xgene_enet_tx_completion() called if the
driver failed to prepare a descriptor for a SKB?
-- 
Florian