tjh.dev
Linux kernel mirror (for testing)
git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git
kernel
os
linux
doc, net: Update netdev operation names
Commits d314774cf2cd5dfeb39a00d37deee65d4c627927 ('netdev: network
device operations infrastructure') and
008298231abbeb91bc7be9e8b078607b816d1a4a ('netdev: add more functions
to netdevice ops') moved and renamed net device operation pointers.
Signed-off-by: Ben Hutchings <bhutchings@solarflare.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
authored by
Ben Hutchings
and committed by
David S. Miller
b3cf6545
04fd3d35
+14
-14
+6
-6
Documentation/networking/driver.txt
+6
-6
Documentation/networking/driver.txt
···
2
2
3
3
Transmit path guidelines:
4
4
5
-
1) The hard_start_xmit method must never return '1' under any
5
+
1) The ndo_start_xmit method must never return '1' under any
6
6
normal circumstances. It is considered a hard error unless
7
7
there is no way your device can tell ahead of time when it's
8
8
transmit function will become busy.
···
61
61
2) Do not forget to update netdev->trans_start to jiffies after
62
62
each new tx packet is given to the hardware.
63
63
64
-
3) A hard_start_xmit method must not modify the shared parts of a
64
+
3) An ndo_start_xmit method must not modify the shared parts of a
65
65
cloned SKB.
66
66
67
-
4) Do not forget that once you return 0 from your hard_start_xmit
67
+
4) Do not forget that once you return 0 from your ndo_start_xmit
68
68
method, it is your driver's responsibility to free up the SKB
69
69
and in some finite amount of time.
70
70
···
74
74
This error can deadlock sockets waiting for send buffer room
75
75
to be freed up.
76
76
77
-
If you return 1 from the hard_start_xmit method, you must not keep
77
+
If you return 1 from the ndo_start_xmit method, you must not keep
78
78
any reference to that SKB and you must not attempt to free it up.
79
79
80
80
Probing guidelines:
···
85
85
86
86
Close/stop guidelines:
87
87
88
-
1) After the dev->stop routine has been called, the hardware must
88
+
1) After the ndo_stop routine has been called, the hardware must
89
89
not receive or transmit any data. All in flight packets must
90
90
be aborted. If necessary, poll or wait for completion of
91
91
any reset commands.
92
92
93
-
2) The dev->stop routine will be called by unregister_netdevice
93
+
2) The ndo_stop routine will be called by unregister_netdevice
94
94
if device is still UP.
+8
-8
Documentation/networking/netdevices.txt
+8
-8
Documentation/networking/netdevices.txt
···
47
47
48
48
struct net_device synchronization rules
49
49
=======================================
50
-
dev->open:
50
+
ndo_open:
51
51
Synchronization: rtnl_lock() semaphore.
52
52
Context: process
53
53
54
-
dev->stop:
54
+
ndo_stop:
55
55
Synchronization: rtnl_lock() semaphore.
56
56
Context: process
57
57
Note: netif_running() is guaranteed false
58
58
59
-
dev->do_ioctl:
59
+
ndo_do_ioctl:
60
60
Synchronization: rtnl_lock() semaphore.
61
61
Context: process
62
62
63
-
dev->get_stats:
63
+
ndo_get_stats:
64
64
Synchronization: dev_base_lock rwlock.
65
65
Context: nominally process, but don't sleep inside an rwlock
66
66
67
-
dev->hard_start_xmit:
67
+
ndo_start_xmit:
68
68
Synchronization: __netif_tx_lock spinlock.
69
69
70
70
When the driver sets NETIF_F_LLTX in dev->features this will be
···
86
86
o NETDEV_TX_LOCKED Locking failed, please retry quickly.
87
87
Only valid when NETIF_F_LLTX is set.
88
88
89
-
dev->tx_timeout:
89
+
ndo_tx_timeout:
90
90
Synchronization: netif_tx_lock spinlock; all TX queues frozen.
91
91
Context: BHs disabled
92
92
Notes: netif_queue_stopped() is guaranteed true
93
93
94
-
dev->set_rx_mode:
94
+
ndo_set_rx_mode:
95
95
Synchronization: netif_addr_lock spinlock.
96
96
Context: BHs disabled
97
97
···
99
99
========================================
100
100
napi->poll:
101
101
Synchronization: NAPI_STATE_SCHED bit in napi->state. Device
102
-
driver's dev->close method will invoke napi_disable() on
102
+
driver's ndo_stop method will invoke napi_disable() on
103
103
all NAPI instances which will do a sleeping poll on the
104
104
NAPI_STATE_SCHED napi->state bit, waiting for all pending
105
105
NAPI activity to cease.