Blame view
Documentation/networking/driver.txt
2.66 KB
98766fbe6 [PATCH] kernel Do... |
1 |
Document about softnet driver issues |
1da177e4c Linux-2.6.12-rc2 |
2 3 |
Transmit path guidelines: |
e34fac1c2 doc, net: Update ... |
4 5 |
1) The ndo_start_xmit method must not return NETDEV_TX_BUSY under any normal circumstances. It is considered a hard error unless |
1da177e4c Linux-2.6.12-rc2 |
6 7 8 9 10 |
there is no way your device can tell ahead of time when it's transmit function will become busy. Instead it must maintain the queue properly. For example, for a driver implementing scatter-gather this means: |
e34fac1c2 doc, net: Update ... |
11 12 |
static netdev_tx_t drv_hard_start_xmit(struct sk_buff *skb, struct net_device *dev) |
1da177e4c Linux-2.6.12-rc2 |
13 |
{ |
b74ca3a89 netdevice: Kill n... |
14 |
struct drv *dp = netdev_priv(dev); |
1da177e4c Linux-2.6.12-rc2 |
15 16 17 18 19 20 21 22 23 24 |
lock_tx(dp); ... /* This is a hard error log it. */ if (TX_BUFFS_AVAIL(dp) <= (skb_shinfo(skb)->nr_frags + 1)) { netif_stop_queue(dev); unlock_tx(dp); printk(KERN_ERR PFX "%s: BUG! Tx Ring full when queue awake! ", dev->name); |
e34fac1c2 doc, net: Update ... |
25 |
return NETDEV_TX_BUSY; |
1da177e4c Linux-2.6.12-rc2 |
26 27 28 29 30 31 32 33 34 35 36 |
} ... queue packet to card ... ... update tx consumer index ... if (TX_BUFFS_AVAIL(dp) <= (MAX_SKB_FRAGS + 1)) netif_stop_queue(dev); ... unlock_tx(dp); ... |
e34fac1c2 doc, net: Update ... |
37 |
return NETDEV_TX_OK; |
1da177e4c Linux-2.6.12-rc2 |
38 |
} |
d6bc8ac9e Fix typos in Docu... |
39 |
And then at the end of your TX reclamation event handling: |
1da177e4c Linux-2.6.12-rc2 |
40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 |
if (netif_queue_stopped(dp->dev) && TX_BUFFS_AVAIL(dp) > (MAX_SKB_FRAGS + 1)) netif_wake_queue(dp->dev); For a non-scatter-gather supporting card, the three tests simply become: /* This is a hard error log it. */ if (TX_BUFFS_AVAIL(dp) <= 0) and: if (TX_BUFFS_AVAIL(dp) == 0) and: if (netif_queue_stopped(dp->dev) && TX_BUFFS_AVAIL(dp) > 0) netif_wake_queue(dp->dev); |
de7aca16f doc, net: Remove ... |
59 |
2) An ndo_start_xmit method must not modify the shared parts of a |
ce3ba1399 Documentation: ad... |
60 |
cloned SKB. |
e34fac1c2 doc, net: Update ... |
61 62 63 |
3) Do not forget that once you return NETDEV_TX_OK from your ndo_start_xmit method, it is your driver's responsibility to free up the SKB and in some finite amount of time. |
1da177e4c Linux-2.6.12-rc2 |
64 65 66 67 68 69 |
For example, this means that it is not allowed for your TX mitigation scheme to let TX packets "hang out" in the TX ring unreclaimed forever if no new TX packets are sent. This error can deadlock sockets waiting for send buffer room to be freed up. |
e34fac1c2 doc, net: Update ... |
70 71 72 |
If you return NETDEV_TX_BUSY from the ndo_start_xmit method, you must not keep any reference to that SKB and you must not attempt to free it up. |
1da177e4c Linux-2.6.12-rc2 |
73 74 75 76 77 78 79 80 |
Probing guidelines: 1) Any hardware layer address you obtain for your device should be verified. For example, for ethernet check it with linux/etherdevice.h:is_valid_ether_addr() Close/stop guidelines: |
b3cf65457 doc, net: Update ... |
81 |
1) After the ndo_stop routine has been called, the hardware must |
1da177e4c Linux-2.6.12-rc2 |
82 83 84 |
not receive or transmit any data. All in flight packets must be aborted. If necessary, poll or wait for completion of any reset commands. |
b3cf65457 doc, net: Update ... |
85 |
2) The ndo_stop routine will be called by unregister_netdevice |
1da177e4c Linux-2.6.12-rc2 |
86 |
if device is still UP. |