1
  2Network Devices, the Kernel, and You!
  3
  4
  5Introduction
  6============
  7The following is a random collection of documentation regarding
  8network devices.
  9
 10struct net_device allocation rules
 11==================================
 12Network device structures need to persist even after module is unloaded and
 13must be allocated with kmalloc.  If device has registered successfully,
 14it will be freed on last use by free_netdev.  This is required to handle the
 15pathologic case cleanly (example: rmmod mydriver </sys/class/net/myeth/mtu )
 16
 17There are routines in net_init.c to handle the common cases of
 18alloc_etherdev, alloc_netdev.  These reserve extra space for driver
 19private data which gets freed when the network device is freed. If
 20separately allocated data is attached to the network device
 21(netdev_priv(dev)) then it is up to the module exit handler to free that.
 22
 23MTU
 24===
 25Each network device has a Maximum Transfer Unit. The MTU does not
 26include any link layer protocol overhead. Upper layer protocols must
 27not pass a socket buffer (skb) to a device to transmit with more data
 28than the mtu. The MTU does not include link layer header overhead, so
 29for example on Ethernet if the standard MTU is 1500 bytes used, the
 30actual skb will contain up to 1514 bytes because of the Ethernet
 31header. Devices should allow for the 4 byte VLAN header as well.
 32
 33Segmentation Offload (GSO, TSO) is an exception to this rule.  The
 34upper layer protocol may pass a large socket buffer to the device
 35transmit routine, and the device will break that up into separate
 36packets based on the current MTU.
 37
 38MTU is symmetrical and applies both to receive and transmit. A device
 39must be able to receive at least the maximum size packet allowed by
 40the MTU. A network device may use the MTU as mechanism to size receive
 41buffers, but the device should allow packets with VLAN header. With
 42standard Ethernet mtu of 1500 bytes, the device should allow up to
 431518 byte packets (1500 + 14 header + 4 tag).  The device may either:
 44drop, truncate, or pass up oversize packets, but dropping oversize
 45packets is preferred.
 46
 47
 48struct net_device synchronization rules
 49=======================================
 50dev->open:
 51	Synchronization: rtnl_lock() semaphore.
 52	Context: process
 53
 54dev->stop:
 55	Synchronization: rtnl_lock() semaphore.
 56	Context: process
 57	Note1: netif_running() is guaranteed false
 58	Note2: dev->poll() is guaranteed to be stopped
 59
 60dev->do_ioctl:
 61	Synchronization: rtnl_lock() semaphore.
 62	Context: process
 63
 64dev->get_stats:
 65	Synchronization: dev_base_lock rwlock.
 66	Context: nominally process, but don't sleep inside an rwlock
 67
 68dev->hard_start_xmit:
 69	Synchronization: netif_tx_lock spinlock.
 70
 71	When the driver sets NETIF_F_LLTX in dev->features this will be
 72	called without holding netif_tx_lock. In this case the driver
 73	has to lock by itself when needed. It is recommended to use a try lock
 74	for this and return NETDEV_TX_LOCKED when the spin lock fails.
 75	The locking there should also properly protect against 
 76	set_multicast_list. Note that the use of NETIF_F_LLTX is deprecated.
 77	Don't use it for new drivers.
 78
 79	Context: Process with BHs disabled or BH (timer),
 80	         will be called with interrupts disabled by netconsole.
 81
 82	Return codes: 
 83	o NETDEV_TX_OK everything ok. 
 84	o NETDEV_TX_BUSY Cannot transmit packet, try later 
 85	  Usually a bug, means queue start/stop flow control is broken in
 86	  the driver. Note: the driver must NOT put the skb in its DMA ring.
 87	o NETDEV_TX_LOCKED Locking failed, please retry quickly.
 88	  Only valid when NETIF_F_LLTX is set.
 89
 90dev->tx_timeout:
 91	Synchronization: netif_tx_lock spinlock.
 92	Context: BHs disabled
 93	Notes: netif_queue_stopped() is guaranteed true
 94
 95dev->set_multicast_list:
 96	Synchronization: netif_tx_lock spinlock.
 97	Context: BHs disabled
 98
 99struct napi_struct synchronization rules
100========================================
101napi->poll:
102	Synchronization: NAPI_STATE_SCHED bit in napi->state.  Device
103		driver's dev->close method will invoke napi_disable() on
104		all NAPI instances which will do a sleeping poll on the
105		NAPI_STATE_SCHED napi->state bit, waiting for all pending
106		NAPI activity to cease.
107	Context: softirq
108	         will be called with interrupts disabled by netconsole.