This documentation is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation; either version 2 of the License, or (at your option) any later version. This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details. You should have received a copy of the GNU General Public License along with this program; if not, write to the Free Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA For more details see the file COPYING in the source distribution of Linux. Kernel Hackers Manual October 2009 enum sock_type 9 2.6.32-rc5 enum sock_type Socket types enum sock_type { SOCK_STREAM, SOCK_DGRAM, SOCK_RAW, SOCK_RDM, SOCK_SEQPACKET, SOCK_DCCP, SOCK_PACKET }; SOCK_STREAM stream (connection) socket SOCK_DGRAM datagram (conn.less) socket SOCK_RAW raw socket SOCK_RDM reliably-delivered message SOCK_SEQPACKET sequential packet socket SOCK_DCCP Datagram Congestion Control Protocol socket SOCK_PACKET linux specific way of getting packets at the dev level. For writing rarp and other similar things on the user level. When adding some new socket type please grep ARCH_HAS_SOCKET_TYPE include/asm-* /socket.h, at least MIPS overrides this enum for binary compat reasons. Kernel Hackers Manual October 2009 struct socket 9 2.6.32-rc5 struct socket general BSD socket struct socket { socket_state state; short type; unsigned long flags; struct fasync_struct * fasync_list; wait_queue_head_t wait; struct file * file; struct sock * sk; const struct proto_ops * ops; }; state socket state (SS_CONNECTED, etc) type socket type (SOCK_STREAM, etc) flags socket flags (SOCK_ASYNC_NOSPACE, etc) fasync_list Asynchronous wake up list wait wait queue for several uses file File back pointer for gc sk internal networking protocol agnostic socket representation ops protocol specific socket operations Kernel Hackers Manual October 2009 struct skb_shared_hwtstamps 9 2.6.32-rc5 struct skb_shared_hwtstamps hardware time stamps struct skb_shared_hwtstamps { ktime_t hwtstamp; ktime_t syststamp; }; hwtstamp hardware time stamp transformed into duration since arbitrary point in time syststamp hwtstamp transformed to system time base Software time stamps generated by ktime_get_real are stored in skb->tstamp. The relation between the different kinds of time syststamp and tstamp can be compared against each other in arbitrary combinations. The accuracy of a syststamp/tstamp/syststamp from other device comparison is limited by the accuracy of the transformation into system time base. This depends on the device driver and its underlying hardware. hwtstamps can only be compared against other hwtstamps from the same device. This structure is attached to packets as part of the skb_shared_info. Use skb_hwtstamps to get a pointer. Kernel Hackers Manual October 2009 struct skb_shared_tx 9 2.6.32-rc5 struct skb_shared_tx instructions for time stamping of outgoing packets struct skb_shared_tx { struct {unnamed_struct}; __u8 flags; }; {unnamed_struct} anonymous flags all shared_tx flags These flags are attached to packets as part of the skb_shared_info. Use skb_tx to get a pointer. Kernel Hackers Manual October 2009 struct sk_buff 9 2.6.32-rc5 struct sk_buff socket buffer struct sk_buff { struct sk_buff * next; struct sk_buff * prev; struct sock * sk; ktime_t tstamp; struct net_device * dev; unsigned long _skb_dst; #ifdef CONFIG_XFRM struct sec_path * sp; #endif char cb[48]; unsigned int len; unsigned int data_len; __u16 mac_len; __u16 hdr_len; union {unnamed_union}; __u32 priority; __u8 local_df:1; __u8 cloned:1; __u8 ip_summed:2; __u8 nohdr:1; __u8 nfctinfo:3; __u8 pkt_type:3; __u8 fclone:2; __u8 ipvs_property:1; __u8 peeked:1; __u8 nf_trace:1; __be16 protocol; void (* destructor) (struct sk_buff *skb); #if defined(CONFIG_NF_CONNTRACK) || defined(CONFIG_NF_CONNTRACK_MODULE) struct nf_conntrack * nfct; struct sk_buff * nfct_reasm; #endif #ifdef CONFIG_BRIDGE_NETFILTER struct nf_bridge_info * nf_bridge; #endif int iif; __u16 queue_mapping; #ifdef CONFIG_NET_SCHED __u16 tc_index; #ifdef CONFIG_NET_CLS_ACT __u16 tc_verd; #endif #endif #ifdef CONFIG_IPV6_NDISC_NODETYPE __u8 ndisc_nodetype:2; #endif #ifdef CONFIG_NET_DMA dma_cookie_t dma_cookie; #endif #ifdef CONFIG_NETWORK_SECMARK __u32 secmark; #endif __u32 mark; __u16 vlan_tci; sk_buff_data_t transport_header; sk_buff_data_t network_header; sk_buff_data_t mac_header; sk_buff_data_t tail; sk_buff_data_t end; unsigned char * head; unsigned char * data; unsigned int truesize; atomic_t users; }; next Next buffer in list prev Previous buffer in list sk Socket we are owned by tstamp Time we arrived dev Device we arrived on/are leaving by _skb_dst destination entry sp the security path, used for xfrm cb[48] Control buffer. Free for use by every layer. Put private vars here len Length of actual data data_len Data length mac_len Length of link layer header hdr_len writable header length of cloned skb {unnamed_union} anonymous priority Packet queueing priority local_df allow local fragmentation cloned Head may be cloned (check refcnt to be sure) ip_summed Driver fed us an IP checksum nohdr Payload reference only, must not modify header nfctinfo Relationship of this skb to the connection pkt_type Packet class fclone skbuff clone status ipvs_property skbuff is owned by ipvs peeked this packet has been seen already, so stats have been done for it, don't do them again nf_trace netfilter packet trace flag protocol Packet protocol from driver destructor Destruct function nfct Associated connection, if any nfct_reasm netfilter conntrack re-assembly pointer nf_bridge Saved data about a bridged frame - see br_netfilter.c iif ifindex of device we arrived on queue_mapping Queue mapping for multiqueue devices tc_index Traffic control index tc_verd traffic control verdict ndisc_nodetype router type (from link layer) dma_cookie a cookie to one of several possible DMA operations done by skb DMA functions secmark security marking mark Generic packet mark vlan_tci vlan tag control information transport_header Transport layer header network_header Network layer header mac_header Link layer header tail Tail pointer end End pointer head Head of buffer data Data head pointer truesize Buffer size users User count - see {datagram,tcp}.c Kernel Hackers Manual October 2009 skb_queue_empty 9 2.6.32-rc5 skb_queue_empty check if a queue is empty int skb_queue_empty const struct sk_buff_head * list list queue head Returns true if the queue is empty, false otherwise. Kernel Hackers Manual October 2009 skb_queue_is_last 9 2.6.32-rc5 skb_queue_is_last check if skb is the last entry in the queue bool skb_queue_is_last const struct sk_buff_head * list const struct sk_buff * skb list queue head skb buffer Returns true if skb is the last buffer on the list. Kernel Hackers Manual October 2009 skb_queue_is_first 9 2.6.32-rc5 skb_queue_is_first check if skb is the first entry in the queue bool skb_queue_is_first const struct sk_buff_head * list const struct sk_buff * skb list queue head skb buffer Returns true if skb is the first buffer on the list. Kernel Hackers Manual October 2009 skb_queue_next 9 2.6.32-rc5 skb_queue_next return the next packet in the queue struct sk_buff * skb_queue_next const struct sk_buff_head * list const struct sk_buff * skb list queue head skb current buffer Return the next packet in list after skb. It is only valid to call this if skb_queue_is_last evaluates to false. Kernel Hackers Manual October 2009 skb_queue_prev 9 2.6.32-rc5 skb_queue_prev return the prev packet in the queue struct sk_buff * skb_queue_prev const struct sk_buff_head * list const struct sk_buff * skb list queue head skb current buffer Return the prev packet in list before skb. It is only valid to call this if skb_queue_is_first evaluates to false. Kernel Hackers Manual October 2009 skb_get 9 2.6.32-rc5 skb_get reference buffer struct sk_buff * skb_get struct sk_buff * skb skb buffer to reference Makes another reference to a socket buffer and returns a pointer to the buffer. Kernel Hackers Manual October 2009 skb_cloned 9 2.6.32-rc5 skb_cloned is the buffer a clone int skb_cloned const struct sk_buff * skb skb buffer to check Returns true if the buffer was generated with skb_clone and is one of multiple shared copies of the buffer. Cloned buffers are shared data so must not be written to under normal circumstances. Kernel Hackers Manual October 2009 skb_header_cloned 9 2.6.32-rc5 skb_header_cloned is the header a clone int skb_header_cloned const struct sk_buff * skb skb buffer to check Returns true if modifying the header part of the buffer requires the data to be copied. Kernel Hackers Manual October 2009 skb_header_release 9 2.6.32-rc5 skb_header_release release reference to header void skb_header_release struct sk_buff * skb skb buffer to operate on Drop a reference to the header part of the buffer. This is done by acquiring a payload reference. You must not read from the header part of skb->data after this. Kernel Hackers Manual October 2009 skb_shared 9 2.6.32-rc5 skb_shared is the buffer shared int skb_shared const struct sk_buff * skb skb buffer to check Returns true if more than one person has a reference to this buffer. Kernel Hackers Manual October 2009 skb_share_check 9 2.6.32-rc5 skb_share_check check if buffer is shared and if so clone it struct sk_buff * skb_share_check struct sk_buff * skb gfp_t pri skb buffer to check pri priority for memory allocation If the buffer is shared the buffer is cloned and the old copy drops a reference. A new clone with a single reference is returned. If the buffer is not shared the original buffer is returned. When being called from interrupt status or with spinlocks held pri must be GFP_ATOMIC. NULL is returned on a memory allocation failure. Kernel Hackers Manual October 2009 skb_unshare 9 2.6.32-rc5 skb_unshare make a copy of a shared buffer struct sk_buff * skb_unshare struct sk_buff * skb gfp_t pri skb buffer to check pri priority for memory allocation If the socket buffer is a clone then this function creates a new copy of the data, drops a reference count on the old copy and returns the new copy with the reference count at 1. If the buffer is not a clone the original buffer is returned. When called with a spinlock held or from interrupt state pri must be GFP_ATOMIC NULL is returned on a memory allocation failure. Kernel Hackers Manual October 2009 skb_peek 9 2.6.32-rc5 skb_peek struct sk_buff * skb_peek struct sk_buff_head * list_ list_ list to peek at Peek an sk_buff. Unlike most other operations you _MUST_ be careful with this one. A peek leaves the buffer on the list and someone else may run off with it. You must hold the appropriate locks or have a private queue to do this. Returns NULL for an empty list or a pointer to the head element. The reference count is not incremented and the reference is therefore volatile. Use with caution. Kernel Hackers Manual October 2009 skb_peek_tail 9 2.6.32-rc5 skb_peek_tail struct sk_buff * skb_peek_tail struct sk_buff_head * list_ list_ list to peek at Peek an sk_buff. Unlike most other operations you _MUST_ be careful with this one. A peek leaves the buffer on the list and someone else may run off with it. You must hold the appropriate locks or have a private queue to do this. Returns NULL for an empty list or a pointer to the tail element. The reference count is not incremented and the reference is therefore volatile. Use with caution. Kernel Hackers Manual October 2009 skb_queue_len 9 2.6.32-rc5 skb_queue_len get queue length __u32 skb_queue_len const struct sk_buff_head * list_ list_ list to measure Return the length of an sk_buff queue. Kernel Hackers Manual October 2009 __skb_queue_head_init 9 2.6.32-rc5 __skb_queue_head_init initialize non-spinlock portions of sk_buff_head void __skb_queue_head_init struct sk_buff_head * list list queue to initialize This initializes only the list and queue length aspects of an sk_buff_head object. This allows to initialize the list aspects of an sk_buff_head without reinitializing things like the spinlock. It can also be used for on-stack sk_buff_head objects where the spinlock is known to not be used. Kernel Hackers Manual October 2009 skb_queue_splice 9 2.6.32-rc5 skb_queue_splice join two skb lists, this is designed for stacks void skb_queue_splice const struct sk_buff_head * list struct sk_buff_head * head list the new list to add head the place to add it in the first list Kernel Hackers Manual October 2009 skb_queue_splice_init 9 2.6.32-rc5 skb_queue_splice_init join two skb lists and reinitialise the emptied list void skb_queue_splice_init struct sk_buff_head * list struct sk_buff_head * head list the new list to add head the place to add it in the first list The list at list is reinitialised Kernel Hackers Manual October 2009 skb_queue_splice_tail 9 2.6.32-rc5 skb_queue_splice_tail join two skb lists, each list being a queue void skb_queue_splice_tail const struct sk_buff_head * list struct sk_buff_head * head list the new list to add head the place to add it in the first list Kernel Hackers Manual October 2009 skb_queue_splice_tail_init 9 2.6.32-rc5 skb_queue_splice_tail_init join two skb lists and reinitialise the emptied list void skb_queue_splice_tail_init struct sk_buff_head * list struct sk_buff_head * head list the new list to add head the place to add it in the first list Each of the lists is a queue. The list at list is reinitialised Kernel Hackers Manual October 2009 __skb_queue_after 9 2.6.32-rc5 __skb_queue_after queue a buffer at the list head void __skb_queue_after struct sk_buff_head * list struct sk_buff * prev struct sk_buff * newsk list list to use prev place after this buffer newsk buffer to queue Queue a buffer int the middle of a list. This function takes no locks and you must therefore hold required locks before calling it. A buffer cannot be placed on two lists at the same time. Kernel Hackers Manual October 2009 skb_headroom 9 2.6.32-rc5 skb_headroom bytes at buffer head unsigned int skb_headroom const struct sk_buff * skb skb buffer to check Return the number of bytes of free space at the head of an sk_buff. Kernel Hackers Manual October 2009 skb_tailroom 9 2.6.32-rc5 skb_tailroom bytes at buffer end int skb_tailroom const struct sk_buff * skb skb buffer to check Return the number of bytes of free space at the tail of an sk_buff Kernel Hackers Manual October 2009 skb_reserve 9 2.6.32-rc5 skb_reserve adjust headroom void skb_reserve struct sk_buff * skb int len skb buffer to alter len bytes to move Increase the headroom of an empty sk_buff by reducing the tail room. This is only allowed for an empty buffer. Kernel Hackers Manual October 2009 pskb_trim_unique 9 2.6.32-rc5 pskb_trim_unique remove end from a paged unique (not cloned) buffer void pskb_trim_unique struct sk_buff * skb unsigned int len skb buffer to alter len new length This is identical to pskb_trim except that the caller knows that the skb is not cloned so we should never get an error due to out- of-memory. Kernel Hackers Manual October 2009 skb_orphan 9 2.6.32-rc5 skb_orphan orphan a buffer void skb_orphan struct sk_buff * skb skb buffer to orphan If a buffer currently has an owner then we call the owner's destructor function and make the skb unowned. The buffer continues to exist but is no longer charged to its former owner. Kernel Hackers Manual October 2009 __dev_alloc_skb 9 2.6.32-rc5 __dev_alloc_skb allocate an skbuff for receiving struct sk_buff * __dev_alloc_skb unsigned int length gfp_t gfp_mask length length to allocate gfp_mask get_free_pages mask, passed to alloc_skb Allocate a new sk_buff and assign it a usage count of one. The buffer has unspecified headroom built in. Users should allocate the headroom they think they need without accounting for the built in space. The built in space is used for optimisations. NULL is returned if there is no free memory. Kernel Hackers Manual October 2009 netdev_alloc_skb 9 2.6.32-rc5 netdev_alloc_skb allocate an skbuff for rx on a specific device struct sk_buff * netdev_alloc_skb struct net_device * dev unsigned int length dev network device to receive on length length to allocate Allocate a new sk_buff and assign it a usage count of one. The buffer has unspecified headroom built in. Users should allocate the headroom they think they need without accounting for the built in space. The built in space is used for optimisations. NULL is returned if there is no free memory. Although this function allocates memory it can be called from an interrupt. Kernel Hackers Manual October 2009 netdev_alloc_page 9 2.6.32-rc5 netdev_alloc_page allocate a page for ps-rx on a specific device struct page * netdev_alloc_page struct net_device * dev dev network device to receive on Allocate a new page node local to the specified device. NULL is returned if there is no free memory. Kernel Hackers Manual October 2009 skb_clone_writable 9 2.6.32-rc5 skb_clone_writable is the header of a clone writable int skb_clone_writable struct sk_buff * skb unsigned int len skb buffer to check len length up to which to write Returns true if modifying the header part of the cloned buffer does not requires the data to be copied. Kernel Hackers Manual October 2009 skb_cow 9 2.6.32-rc5 skb_cow copy header of skb when it is required int skb_cow struct sk_buff * skb unsigned int headroom skb buffer to cow headroom needed headroom If the skb passed lacks sufficient headroom or its data part is shared, data is reallocated. If reallocation fails, an error is returned and original skb is not changed. The result is skb with writable area skb->head...skb->tail and at least headroom of space at head. Kernel Hackers Manual October 2009 skb_cow_head 9 2.6.32-rc5 skb_cow_head skb_cow but only making the head writable int skb_cow_head struct sk_buff * skb unsigned int headroom skb buffer to cow headroom needed headroom This function is identical to skb_cow except that we replace the skb_cloned check by skb_header_cloned. It should be used when you only need to push on some header and do not need to modify the data. Kernel Hackers Manual October 2009 skb_padto 9 2.6.32-rc5 skb_padto pad an skbuff up to a minimal size int skb_padto struct sk_buff * skb unsigned int len skb buffer to pad len minimal length Pads up a buffer to ensure the trailing bytes exist and are blanked. If the buffer already contains sufficient data it is untouched. Otherwise it is extended. Returns zero on success. The skb is freed on error. Kernel Hackers Manual October 2009 skb_linearize 9 2.6.32-rc5 skb_linearize convert paged skb to linear one int skb_linearize struct sk_buff * skb skb buffer to linarize If there is no free memory -ENOMEM is returned, otherwise zero is returned and the old skb data released. Kernel Hackers Manual October 2009 skb_linearize_cow 9 2.6.32-rc5 skb_linearize_cow make sure skb is linear and writable int skb_linearize_cow struct sk_buff * skb skb buffer to process If there is no free memory -ENOMEM is returned, otherwise zero is returned and the old skb data released. Kernel Hackers Manual October 2009 skb_postpull_rcsum 9 2.6.32-rc5 skb_postpull_rcsum update checksum for received skb after pull void skb_postpull_rcsum struct sk_buff * skb const void * start unsigned int len skb buffer to update start start of data before pull len length of data pulled After doing a pull on a received packet, you need to call this to update the CHECKSUM_COMPLETE checksum, or set ip_summed to CHECKSUM_NONE so that it can be recomputed from scratch. Kernel Hackers Manual October 2009 pskb_trim_rcsum 9 2.6.32-rc5 pskb_trim_rcsum trim received skb and update checksum int pskb_trim_rcsum struct sk_buff * skb unsigned int len skb buffer to trim len new length This is exactly the same as pskb_trim except that it ensures the checksum of received packets are still valid after the operation. Kernel Hackers Manual October 2009 skb_get_timestamp 9 2.6.32-rc5 skb_get_timestamp get timestamp from a skb void skb_get_timestamp const struct sk_buff * skb struct timeval * stamp skb skb to get stamp from stamp pointer to struct timeval to store stamp in Timestamps are stored in the skb as offsets to a base timestamp. This function converts the offset back to a struct timeval and stores it in stamp. Kernel Hackers Manual October 2009 skb_checksum_complete 9 2.6.32-rc5 skb_checksum_complete Calculate checksum of an entire packet __sum16 skb_checksum_complete struct sk_buff * skb skb packet to process This function calculates the checksum over the entire packet plus the value of skb->csum. The latter can be used to supply the checksum of a pseudo header as used by TCP/UDP. It returns the checksum. For protocols that contain complete checksums such as ICMP/TCP/UDP, this function can be used to verify that checksum on received packets. In that case the function should return zero if the checksum is correct. In particular, this function will return zero if skb->ip_summed is CHECKSUM_UNNECESSARY which indicates that the hardware has already verified the correctness of the checksum. Kernel Hackers Manual October 2009 struct sock_common 9 2.6.32-rc5 struct sock_common minimal network layer representation of sockets struct sock_common { union {unnamed_union}; atomic_t skc_refcnt; unsigned int skc_hash; unsigned short skc_family; volatile unsigned char skc_state; unsigned char skc_reuse; int skc_bound_dev_if; struct hlist_node skc_bind_node; struct proto * skc_prot; #ifdef CONFIG_NET_NS struct net * skc_net; #endif }; {unnamed_union} anonymous skc_refcnt reference count skc_hash hash value used with various protocol lookup tables skc_family network address family skc_state Connection state skc_reuse SO_REUSEADDR setting skc_bound_dev_if bound device index if != 0 skc_bind_node bind hash linkage for various protocol lookup tables skc_prot protocol handlers inside a network family skc_net reference to the network namespace of this socket This is the minimal network layer representation of sockets, the header for struct sock and struct inet_timewait_sock. Kernel Hackers Manual October 2009 struct sock 9 2.6.32-rc5 struct sock network layer representation of sockets struct sock { struct sock_common __sk_common; #define sk_node __sk_common.skc_node #define sk_nulls_node __sk_common.skc_nulls_node #define sk_refcnt __sk_common.skc_refcnt #define sk_copy_start __sk_common.skc_hash #define sk_hash __sk_common.skc_hash #define sk_family __sk_common.skc_family #define sk_state __sk_common.skc_state #define sk_reuse __sk_common.skc_reuse #define sk_bound_dev_if __sk_common.skc_bound_dev_if #define sk_bind_node __sk_common.skc_bind_node #define sk_prot __sk_common.skc_prot #define sk_net __sk_common.skc_net unsigned int sk_shutdown:2; unsigned int sk_no_check:2; unsigned int sk_userlocks:4; unsigned int sk_protocol:8; unsigned int sk_type:16; int sk_rcvbuf; socket_lock_t sk_lock; struct sk_backlog; wait_queue_head_t * sk_sleep; struct dst_entry * sk_dst_cache; #ifdef CONFIG_XFRM struct xfrm_policy * sk_policy[2]; #endif rwlock_t sk_dst_lock; atomic_t sk_rmem_alloc; atomic_t sk_wmem_alloc; atomic_t sk_omem_alloc; int sk_sndbuf; struct sk_buff_head sk_receive_queue; struct sk_buff_head sk_write_queue; #ifdef CONFIG_NET_DMA struct sk_buff_head sk_async_wait_queue; #endif int sk_wmem_queued; int sk_forward_alloc; gfp_t sk_allocation; int sk_route_caps; int sk_gso_type; unsigned int sk_gso_max_size; int sk_rcvlowat; unsigned long sk_flags; unsigned long sk_lingertime; struct sk_buff_head sk_error_queue; struct proto * sk_prot_creator; rwlock_t sk_callback_lock; int sk_err; int sk_err_soft; atomic_t sk_drops; unsigned short sk_ack_backlog; unsigned short sk_max_ack_backlog; __u32 sk_priority; struct ucred sk_peercred; long sk_rcvtimeo; long sk_sndtimeo; struct sk_filter * sk_filter; void * sk_protinfo; struct timer_list sk_timer; ktime_t sk_stamp; struct socket * sk_socket; void * sk_user_data; struct page * sk_sndmsg_page; struct sk_buff * sk_send_head; __u32 sk_sndmsg_off; int sk_write_pending; #ifdef CONFIG_SECURITY void * sk_security; #endif __u32 sk_mark; void (* sk_state_change) (struct sock *sk); void (* sk_data_ready) (struct sock *sk, int bytes); void (* sk_write_space) (struct sock *sk); void (* sk_error_report) (struct sock *sk); int (* sk_backlog_rcv) (struct sock *sk,struct sk_buff *skb); void (* sk_destruct) (struct sock *sk); }; __sk_common shared layout with inet_timewait_sock sk_shutdown mask of SEND_SHUTDOWN and/or RCV_SHUTDOWN sk_no_check SO_NO_CHECK setting, wether or not checkup packets sk_userlocks SO_SNDBUF and SO_RCVBUF settings sk_protocol which protocol this socket belongs in this network family sk_type socket type (SOCK_STREAM, etc) sk_rcvbuf size of receive buffer in bytes sk_lock synchronizer sk_backlog always used with the per-socket spinlock held sk_sleep sock wait queue sk_dst_cache destination cache sk_policy[2] flow policy sk_dst_lock destination cache lock sk_rmem_alloc receive queue bytes committed sk_wmem_alloc transmit queue bytes committed sk_omem_alloc "o is option or other" sk_sndbuf size of send buffer in bytes sk_receive_queue incoming packets sk_write_queue Packet sending queue sk_async_wait_queue DMA copied packets sk_wmem_queued persistent queue size sk_forward_alloc space allocated forward sk_allocation allocation mode sk_route_caps route capabilities (e.g. NETIF_F_TSO) sk_gso_type GSO type (e.g. SKB_GSO_TCPV4) sk_gso_max_size Maximum GSO segment size to build sk_rcvlowat SO_RCVLOWAT setting sk_flags SO_LINGER (l_onoff), SO_BROADCAST, SO_KEEPALIVE, SO_OOBINLINE settings, SO_TIMESTAMPING settings sk_lingertime SO_LINGER l_linger setting sk_error_queue rarely used sk_prot_creator sk_prot of original sock creator (see ipv6_setsockopt, IPV6_ADDRFORM for instance) sk_callback_lock used with the callbacks in the end of this struct sk_err last error sk_err_soft errors that don't cause failure but are the cause of a persistent failure not just 'timed out' sk_drops raw/udp drops counter sk_ack_backlog current listen backlog sk_max_ack_backlog listen backlog set in listen sk_priority SO_PRIORITY setting sk_peercred SO_PEERCRED setting sk_rcvtimeo SO_RCVTIMEO setting sk_sndtimeo SO_SNDTIMEO setting sk_filter socket filtering instructions sk_protinfo private area, net family specific, when not using slab sk_timer sock cleanup timer sk_stamp time stamp of last packet received sk_socket Identd and reporting IO signals sk_user_data RPC layer private data sk_sndmsg_page cached page for sendmsg sk_send_head front of stuff to transmit sk_sndmsg_off cached offset for sendmsg sk_write_pending a write to stream socket waits to start sk_security used by security modules sk_mark generic packet mark sk_state_change callback to indicate change in the state of the sock sk_data_ready callback to indicate there is data to be processed sk_write_space callback to indicate there is bf sending space available sk_error_report callback to indicate errors (e.g. MSG_ERRQUEUE) sk_backlog_rcv callback to process the backlog sk_destruct called at sock freeing time, i.e. when all refcnt == 0 Kernel Hackers Manual October 2009 sk_filter_release 9 2.6.32-rc5 sk_filter_release void sk_filter_release struct sk_filter * fp fp filter to remove Remove a filter from a socket and release its resources. Kernel Hackers Manual October 2009 sk_wmem_alloc_get 9 2.6.32-rc5 sk_wmem_alloc_get returns write allocations int sk_wmem_alloc_get const struct sock * sk sk socket Returns sk_wmem_alloc minus initial offset of one Kernel Hackers Manual October 2009 sk_rmem_alloc_get 9 2.6.32-rc5 sk_rmem_alloc_get returns read allocations int sk_rmem_alloc_get const struct sock * sk sk socket Returns sk_rmem_alloc Kernel Hackers Manual October 2009 sk_has_allocations 9 2.6.32-rc5 sk_has_allocations check if allocations are outstanding int sk_has_allocations const struct sock * sk sk socket Returns true if socket has write or read allocations Kernel Hackers Manual October 2009 sk_has_sleeper 9 2.6.32-rc5 sk_has_sleeper check if there are any waiting processes int sk_has_sleeper struct sock * sk sk socket Returns true if socket has waiting processes The purpose of the sk_has_sleeper and sock_poll_wait is to wrap the memory barrier call. They were added due to the race found within the tcp code. CPU1 CPU2 sys_select receive packet ... ... __add_wait_queue update tp->rcv_nxt ... ... tp->rcv_nxt check sock_def_readable ... { schedule ... if (sk->sk_sleep && waitqueue_active(sk->sk_sleep)) wake_up_interruptible(sk->sk_sleep) ... } The race for tcp fires when the __add_wait_queue changes done by CPU1 stay in its cache, and so does the tp->rcv_nxt update on CPU2 side. The CPU1 could then endup calling schedule and sleep forever if there are no more data on the socket. The sk_has_sleeper is always called right after a call to read_lock, so we can use smp_mb__after_lock barrier. Kernel Hackers Manual October 2009 sock_poll_wait 9 2.6.32-rc5 sock_poll_wait place memory barrier behind the poll_wait call. void sock_poll_wait struct file * filp wait_queue_head_t * wait_address poll_table * p filp file wait_address socket wait queue p poll_table See the comments in the sk_has_sleeper function. Kernel Hackers Manual October 2009 sk_eat_skb 9 2.6.32-rc5 sk_eat_skb Release a skb if it is no longer needed void sk_eat_skb struct sock * sk struct sk_buff * skb int copied_early sk socket to eat this skb from skb socket buffer to eat copied_early flag indicating whether DMA operations copied this data early This routine must be called with interrupts disabled or with the socket locked so that the sk_buff queue operation is ok. Kernel Hackers Manual October 2009 sockfd_lookup 9 2.6.32-rc5 sockfd_lookup Go from a file number to its socket slot struct socket * sockfd_lookup int fd int * err fd file handle err pointer to an error code return The file handle passed in is locked and the socket it is bound too is returned. If an error occurs the err pointer is overwritten with a negative errno code and NULL is returned. The function checks for both invalid handles and passing a handle which is not a socket. On a success the socket object pointer is returned. Kernel Hackers Manual October 2009 sock_release 9 2.6.32-rc5 sock_release close a socket void sock_release struct socket * sock sock socket to close The socket is released from the protocol stack if it has a release callback, and the inode is then released if the socket is bound to an inode not a file. Kernel Hackers Manual October 2009 sock_register 9 2.6.32-rc5 sock_register add a socket protocol handler int sock_register const struct net_proto_family * ops ops description of protocol This function is called by a protocol handler that wants to advertise its address family, and have it linked into the socket interface. The value ops->family coresponds to the socket system call protocol family. Kernel Hackers Manual October 2009 sock_unregister 9 2.6.32-rc5 sock_unregister remove a protocol handler void sock_unregister int family family protocol family to remove This function is called by a protocol handler that wants to remove its address family, and have it unlinked from the new socket creation. If protocol handler is a module, then it can use module reference counts to protect against new references. If protocol handler is not a module then it needs to provide its own protection in the ops->create routine. Kernel Hackers Manual October 2009 skb_over_panic 9 2.6.32-rc5 skb_over_panic private function void skb_over_panic struct sk_buff * skb int sz void * here skb buffer sz size here address Out of line support code for skb_put. Not user callable. Kernel Hackers Manual October 2009 skb_under_panic 9 2.6.32-rc5 skb_under_panic private function void skb_under_panic struct sk_buff * skb int sz void * here skb buffer sz size here address Out of line support code for skb_push. Not user callable. Kernel Hackers Manual October 2009 __alloc_skb 9 2.6.32-rc5 __alloc_skb allocate a network buffer struct sk_buff * __alloc_skb unsigned int size gfp_t gfp_mask int fclone int node size size to allocate gfp_mask allocation mask fclone allocate from fclone cache instead of head cache and allocate a cloned (child) skb node numa node to allocate memory on Allocate a new sk_buff. The returned buffer has no headroom and a tail room of size bytes. The object has a reference count of one. The return is the buffer. On a failure the return is NULL. Buffers may only be allocated from interrupts using a gfp_mask of GFP_ATOMIC. Kernel Hackers Manual October 2009 __netdev_alloc_skb 9 2.6.32-rc5 __netdev_alloc_skb allocate an skbuff for rx on a specific device struct sk_buff * __netdev_alloc_skb struct net_device * dev unsigned int length gfp_t gfp_mask dev network device to receive on length length to allocate gfp_mask get_free_pages mask, passed to alloc_skb Allocate a new sk_buff and assign it a usage count of one. The buffer has unspecified headroom built in. Users should allocate the headroom they think they need without accounting for the built in space. The built in space is used for optimisations. NULL is returned if there is no free memory. Kernel Hackers Manual October 2009 dev_alloc_skb 9 2.6.32-rc5 dev_alloc_skb allocate an skbuff for receiving struct sk_buff * dev_alloc_skb unsigned int length length length to allocate Allocate a new sk_buff and assign it a usage count of one. The buffer has unspecified headroom built in. Users should allocate the headroom they think they need without accounting for the built in space. The built in space is used for optimisations. NULL is returned if there is no free memory. Although this function allocates memory it can be called from an interrupt. Kernel Hackers Manual October 2009 __kfree_skb 9 2.6.32-rc5 __kfree_skb private function void __kfree_skb struct sk_buff * skb skb buffer Free an sk_buff. Release anything attached to the buffer. Clean the state. This is an internal helper function. Users should always call kfree_skb Kernel Hackers Manual October 2009 kfree_skb 9 2.6.32-rc5 kfree_skb free an sk_buff void kfree_skb struct sk_buff * skb skb buffer to free Drop a reference to the buffer and free it if the usage count has hit zero. Kernel Hackers Manual October 2009 consume_skb 9 2.6.32-rc5 consume_skb free an skbuff void consume_skb struct sk_buff * skb skb buffer to free Drop a ref to the buffer and free it if the usage count has hit zero Functions identically to kfree_skb, but kfree_skb assumes that the frame is being dropped after a failure and notes that Kernel Hackers Manual October 2009 skb_recycle_check 9 2.6.32-rc5 skb_recycle_check check if skb can be reused for receive int skb_recycle_check struct sk_buff * skb int skb_size skb buffer skb_size minimum receive buffer size Checks that the skb passed in is not shared or cloned, and that it is linear and its head portion at least as large as skb_size so that it can be recycled as a receive buffer. If these conditions are met, this function does any necessary reference count dropping and cleans up the skbuff as if it just came from __alloc_skb. Kernel Hackers Manual October 2009 skb_morph 9 2.6.32-rc5 skb_morph morph one skb into another struct sk_buff * skb_morph struct sk_buff * dst struct sk_buff * src dst the skb to receive the contents src the skb to supply the contents This is identical to skb_clone except that the target skb is supplied by the user. The target skb is returned upon exit. Kernel Hackers Manual October 2009 skb_clone 9 2.6.32-rc5 skb_clone duplicate an sk_buff struct sk_buff * skb_clone struct sk_buff * skb gfp_t gfp_mask skb buffer to clone gfp_mask allocation priority Duplicate an sk_buff. The new one is not owned by a socket. Both copies share the same packet data but not structure. The new buffer has a reference count of 1. If the allocation fails the function returns NULL otherwise the new buffer is returned. If this function is called from an interrupt gfp_mask must be GFP_ATOMIC. Kernel Hackers Manual October 2009 skb_copy 9 2.6.32-rc5 skb_copy create private copy of an sk_buff struct sk_buff * skb_copy const struct sk_buff * skb gfp_t gfp_mask skb buffer to copy gfp_mask allocation priority Make a copy of both an sk_buff and its data. This is used when the caller wishes to modify the data and needs a private copy of the data to alter. Returns NULL on failure or the pointer to the buffer on success. The returned buffer has a reference count of 1. As by-product this function converts non-linear sk_buff to linear one, so that sk_buff becomes completely private and caller is allowed to modify all the data of returned buffer. This means that this function is not recommended for use in circumstances when only header is going to be modified. Use pskb_copy instead. Kernel Hackers Manual October 2009 pskb_copy 9 2.6.32-rc5 pskb_copy create copy of an sk_buff with private head. struct sk_buff * pskb_copy struct sk_buff * skb gfp_t gfp_mask skb buffer to copy gfp_mask allocation priority Make a copy of both an sk_buff and part of its data, located in header. Fragmented data remain shared. This is used when the caller wishes to modify only header of sk_buff and needs private copy of the header to alter. Returns NULL on failure or the pointer to the buffer on success. The returned buffer has a reference count of 1. Kernel Hackers Manual October 2009 pskb_expand_head 9 2.6.32-rc5 pskb_expand_head reallocate header of sk_buff int pskb_expand_head struct sk_buff * skb int nhead int ntail gfp_t gfp_mask skb buffer to reallocate nhead room to add at head ntail room to add at tail gfp_mask allocation priority Expands (or creates identical copy, if nhead and ntail are zero) header of skb. sk_buff itself is not changed. sk_buff MUST have reference count of 1. Returns zero in the case of success or error, if expansion failed. In the last case, sk_buff is not changed. All the pointers pointing into skb header may change and must be reloaded after call to this function. Kernel Hackers Manual October 2009 skb_copy_expand 9 2.6.32-rc5 skb_copy_expand copy and expand sk_buff struct sk_buff * skb_copy_expand const struct sk_buff * skb int newheadroom int newtailroom gfp_t gfp_mask skb buffer to copy newheadroom new free bytes at head newtailroom new free bytes at tail gfp_mask allocation priority Make a copy of both an sk_buff and its data and while doing so allocate additional space. This is used when the caller wishes to modify the data and needs a private copy of the data to alter as well as more space for new fields. Returns NULL on failure or the pointer to the buffer on success. The returned buffer has a reference count of 1. You must pass GFP_ATOMIC as the allocation priority if this function is called from an interrupt. Kernel Hackers Manual October 2009 skb_pad 9 2.6.32-rc5 skb_pad zero pad the tail of an skb int skb_pad struct sk_buff * skb int pad skb buffer to pad pad space to pad Ensure that a buffer is followed by a padding area that is zero filled. Used by network drivers which may DMA or transfer data beyond the buffer end onto the wire. May return error in out of memory cases. The skb is freed on error. Kernel Hackers Manual October 2009 skb_put 9 2.6.32-rc5 skb_put add data to a buffer unsigned char * skb_put struct sk_buff * skb unsigned int len skb buffer to use len amount of data to add This function extends the used data area of the buffer. If this would exceed the total buffer size the kernel will panic. A pointer to the first byte of the extra data is returned. Kernel Hackers Manual October 2009 skb_push 9 2.6.32-rc5 skb_push add data to the start of a buffer unsigned char * skb_push struct sk_buff * skb unsigned int len skb buffer to use len amount of data to add This function extends the used data area of the buffer at the buffer start. If this would exceed the total buffer headroom the kernel will panic. A pointer to the first byte of the extra data is returned. Kernel Hackers Manual October 2009 skb_pull 9 2.6.32-rc5 skb_pull remove data from the start of a buffer unsigned char * skb_pull struct sk_buff * skb unsigned int len skb buffer to use len amount of data to remove This function removes data from the start of a buffer, returning the memory to the headroom. A pointer to the next data in the buffer is returned. Once the data has been pulled future pushes will overwrite the old data. Kernel Hackers Manual October 2009 skb_trim 9 2.6.32-rc5 skb_trim remove end from a buffer void skb_trim struct sk_buff * skb unsigned int len skb buffer to alter len new length Cut the length of a buffer down by removing data from the tail. If the buffer is already under the length specified it is not modified. The skb must be linear. Kernel Hackers Manual October 2009 __pskb_pull_tail 9 2.6.32-rc5 __pskb_pull_tail advance tail of skb header unsigned char * __pskb_pull_tail struct sk_buff * skb int delta skb buffer to reallocate delta number of bytes to advance tail The function makes a sense only on a fragmented sk_buff, it expands header moving its tail forward and copying necessary data from fragmented part. sk_buff MUST have reference count of 1. Returns NULL (and sk_buff does not change) if pull failed or value of new tail of skb in the case of success. All the pointers pointing into skb header may change and must be reloaded after call to this function. Kernel Hackers Manual October 2009 skb_store_bits 9 2.6.32-rc5 skb_store_bits store bits from kernel buffer to skb int skb_store_bits struct sk_buff * skb int offset const void * from int len skb destination buffer offset offset in destination from source buffer len number of bytes to copy Copy the specified number of bytes from the source buffer to the destination skb. This function handles all the messy bits of traversing fragment lists and such. Kernel Hackers Manual October 2009 skb_dequeue 9 2.6.32-rc5 skb_dequeue remove from the head of the queue struct sk_buff * skb_dequeue struct sk_buff_head * list list list to dequeue from Remove the head of the list. The list lock is taken so the function may be used safely with other locking list functions. The head item is returned or NULL if the list is empty. Kernel Hackers Manual October 2009 skb_dequeue_tail 9 2.6.32-rc5 skb_dequeue_tail remove from the tail of the queue struct sk_buff * skb_dequeue_tail struct sk_buff_head * list list list to dequeue from Remove the tail of the list. The list lock is taken so the function may be used safely with other locking list functions. The tail item is returned or NULL if the list is empty. Kernel Hackers Manual October 2009 skb_queue_purge 9 2.6.32-rc5 skb_queue_purge empty a list void skb_queue_purge struct sk_buff_head * list list list to empty Delete all buffers on an sk_buff list. Each buffer is removed from the list and one reference dropped. This function takes the list lock and is atomic with respect to other list locking functions. Kernel Hackers Manual October 2009 skb_queue_head 9 2.6.32-rc5 skb_queue_head queue a buffer at the list head void skb_queue_head struct sk_buff_head * list struct sk_buff * newsk list list to use newsk buffer to queue Queue a buffer at the start of the list. This function takes the list lock and can be used safely with other locking sk_buff functions safely. A buffer cannot be placed on two lists at the same time. Kernel Hackers Manual October 2009 skb_queue_tail 9 2.6.32-rc5 skb_queue_tail queue a buffer at the list tail void skb_queue_tail struct sk_buff_head * list struct sk_buff * newsk list list to use newsk buffer to queue Queue a buffer at the tail of the list. This function takes the list lock and can be used safely with other locking sk_buff functions safely. A buffer cannot be placed on two lists at the same time. Kernel Hackers Manual October 2009 skb_unlink 9 2.6.32-rc5 skb_unlink remove a buffer from a list void skb_unlink struct sk_buff * skb struct sk_buff_head * list skb buffer to remove list list to use Remove a packet from a list. The list locks are taken and this function is atomic with respect to other list locked calls You must know what list the SKB is on. Kernel Hackers Manual October 2009 skb_append 9 2.6.32-rc5 skb_append append a buffer void skb_append struct sk_buff * old struct sk_buff * newsk struct sk_buff_head * list old buffer to insert after newsk buffer to insert list list to use Place a packet after a given packet in a list. The list locks are taken and this function is atomic with respect to other list locked calls. A buffer cannot be placed on two lists at the same time. Kernel Hackers Manual October 2009 skb_insert 9 2.6.32-rc5 skb_insert insert a buffer void skb_insert struct sk_buff * old struct sk_buff * newsk struct sk_buff_head * list old buffer to insert before newsk buffer to insert list list to use Place a packet before a given packet in a list. The list locks are taken and this function is atomic with respect to other list locked calls. A buffer cannot be placed on two lists at the same time. Kernel Hackers Manual October 2009 skb_split 9 2.6.32-rc5 skb_split Split fragmented skb to two parts at length len. void skb_split struct sk_buff * skb struct sk_buff * skb1 const u32 len skb the buffer to split skb1 the buffer to receive the second part len new length for skb Kernel Hackers Manual October 2009 skb_prepare_seq_read 9 2.6.32-rc5 skb_prepare_seq_read Prepare a sequential read of skb data void skb_prepare_seq_read struct sk_buff * skb unsigned int from unsigned int to struct skb_seq_state * st skb the buffer to read from lower offset of data to be read to upper offset of data to be read st state variable Initializes the specified state variable. Must be called before invoking skb_seq_read for the first time. Kernel Hackers Manual October 2009 skb_seq_read 9 2.6.32-rc5 skb_seq_read Sequentially read skb data unsigned int skb_seq_read unsigned int consumed const u8 ** data struct skb_seq_state * st consumed number of bytes consumed by the caller so far data destination pointer for data to be returned st state variable Reads a block of skb data at consumed relative to the lower offset specified to skb_prepare_seq_read. Assigns the head of the data block to data and returns the length of the block or 0 if the end of the skb data or the upper offset has been reached. The caller is not required to consume all of the data returned, i.e. consumed is typically set to the number of bytes already consumed and the next call to skb_seq_read will return the remaining part of the block. The size of each block of data returned can be arbitary, this limitation is the cost for zerocopy seqeuental reads of potentially non linear data. Fragment lists within fragments are not implemented at the moment, state->root_skb could be replaced with a stack for this purpose. Kernel Hackers Manual October 2009 skb_abort_seq_read 9 2.6.32-rc5 skb_abort_seq_read Abort a sequential read of skb data void skb_abort_seq_read struct skb_seq_state * st st state variable Must be called if skb_seq_read was not called until it returned 0. Kernel Hackers Manual October 2009 skb_find_text 9 2.6.32-rc5 skb_find_text Find a text pattern in skb data unsigned int skb_find_text struct sk_buff * skb unsigned int from unsigned int to struct ts_config * config struct ts_state * state skb the buffer to look in from search offset to search limit config textsearch configuration state uninitialized textsearch state variable Finds a pattern in the skb data according to the specified textsearch configuration. Use textsearch_next to retrieve subsequent occurrences of the pattern. Returns the offset to the first occurrence or UINT_MAX if no match was found. Kernel Hackers Manual October 2009 skb_append_datato_frags 9 2.6.32-rc5 skb_append_datato_frags append the user data to a skb int skb_append_datato_frags struct sock * sk struct sk_buff * skb int (*getfrag) void *from, char *to, int offset, int len, int odd, struct sk_buff *skb void * from int length sk sock structure skb skb structure to be appened with user data. getfrag call back function to be used for getting the user data from pointer to user message iov length length of the iov message This procedure append the user data in the fragment part of the skb if any page alloc fails user this procedure returns -ENOMEM Kernel Hackers Manual October 2009 skb_pull_rcsum 9 2.6.32-rc5 skb_pull_rcsum pull skb and update receive checksum unsigned char * skb_pull_rcsum struct sk_buff * skb unsigned int len skb buffer to update len length of data pulled This function performs an skb_pull on the packet and updates the CHECKSUM_COMPLETE checksum. It should be used on receive path processing instead of skb_pull unless you know that the checksum difference is zero (e.g., a valid IP header) or you are setting ip_summed to CHECKSUM_NONE. Kernel Hackers Manual October 2009 skb_segment 9 2.6.32-rc5 skb_segment Perform protocol segmentation on skb. struct sk_buff * skb_segment struct sk_buff * skb int features skb buffer to segment features features for the output path (see dev->features) This function performs segmentation on the given skb. It returns a pointer to the first in a list of new skbs for the segments. In case of error it returns ERR_PTR(err). Kernel Hackers Manual October 2009 skb_cow_data 9 2.6.32-rc5 skb_cow_data Check that a socket buffer's data buffers are writable int skb_cow_data struct sk_buff * skb int tailbits struct sk_buff ** trailer skb The socket buffer to check. tailbits Amount of trailing space to be added trailer Returned pointer to the skb where the tailbits space begins Make sure that the data buffers attached to a socket buffer are writable. If they are not, private copies are made of the data buffers and the socket buffer is set to use these instead. If tailbits is given, make sure that there is space to write tailbits bytes of data beyond current end of socket buffer. trailer will be set to point to the skb in which this space begins. The number of scatterlist elements required to completely map the COW'd and extended socket buffer will be returned. Kernel Hackers Manual October 2009 skb_partial_csum_set 9 2.6.32-rc5 skb_partial_csum_set set up and verify partial csum values for packet bool skb_partial_csum_set struct sk_buff * skb u16 start u16 off skb the skb to set start the number of bytes after skb->data to start checksumming. off the offset from start to place the checksum. For untrusted partially-checksummed packets, we need to make sure the values for skb->csum_start and skb->csum_offset are valid so we don't oops. This function checks and sets those values and skb->ip_summed: if this returns false you should drop the packet. Kernel Hackers Manual October 2009 sk_alloc 9 2.6.32-rc5 sk_alloc All socket objects are allocated here struct sock * sk_alloc struct net * net int family gfp_t priority struct proto * prot net the applicable net namespace family protocol family priority for allocation (GFP_KERNEL, GFP_ATOMIC, etc) prot struct proto associated with this new sock instance Kernel Hackers Manual October 2009 sk_wait_data 9 2.6.32-rc5 sk_wait_data wait for data to arrive at sk_receive_queue int sk_wait_data struct sock * sk long * timeo sk sock to wait on timeo for how long Now socket state including sk->sk_err is changed only under lock, hence we may omit checks after joining wait queue. We check receive queue before schedule only as optimization; it is very likely that release_sock added new data. Kernel Hackers Manual October 2009 __sk_mem_schedule 9 2.6.32-rc5 __sk_mem_schedule increase sk_forward_alloc and memory_allocated int __sk_mem_schedule struct sock * sk int size int kind sk socket size memory size to allocate kind allocation type If kind is SK_MEM_SEND, it means wmem allocation. Otherwise it means rmem allocation. This function assumes that protocols which have memory_pressure use sk_wmem_queued as write buffer accounting. Kernel Hackers Manual October 2009 __sk_mem_reclaim 9 2.6.32-rc5 __sk_mem_reclaim reclaim memory_allocated void __sk_mem_reclaim struct sock * sk sk socket Kernel Hackers Manual October 2009 __skb_recv_datagram 9 2.6.32-rc5 __skb_recv_datagram Receive a datagram skbuff struct sk_buff * __skb_recv_datagram struct sock * sk unsigned flags int * peeked int * err sk socket flags MSG_ flags peeked returns non-zero if this packet has been seen before err error code returned Get a datagram skbuff, understands the peeking, nonblocking wakeups and possible races. This replaces identical code in packet, raw and udp, as well as the IPX AX.25 and Appletalk. It also finally fixes the long standing peek and read race for datagram sockets. If you alter this routine remember it must be re-entrant. This function will lock the socket if a skb is returned, so the caller needs to unlock the socket in that case (usually by calling skb_free_datagram) * It does not lock socket since today. This function is * free of race conditions. This measure should/can improve * significantly datagram socket latencies at high loads, * when data copying to user space takes lots of time. * (BTW I've just killed the last cli in IP/IPv6/core/netlink/packet * 8) Great win.) * --ANK (980729) The order of the tests when we find no data waiting are specified quite explicitly by POSIX 1003.1g, don't change them without having the standard around please. Kernel Hackers Manual October 2009 skb_kill_datagram 9 2.6.32-rc5 skb_kill_datagram Free a datagram skbuff forcibly int skb_kill_datagram struct sock * sk struct sk_buff * skb unsigned int flags sk socket skb datagram skbuff flags MSG_ flags This function frees a datagram skbuff that was received by skb_recv_datagram. The flags argument must match the one used for skb_recv_datagram. If the MSG_PEEK flag is set, and the packet is still on the receive queue of the socket, it will be taken off the queue before it is freed. This function currently only disables BH when acquiring the sk_receive_queue lock. Therefore it must not be used in a context where that lock is acquired in an IRQ context. It returns 0 if the packet was removed by us. Kernel Hackers Manual October 2009 skb_copy_datagram_iovec 9 2.6.32-rc5 skb_copy_datagram_iovec Copy a datagram to an iovec. int skb_copy_datagram_iovec const struct sk_buff * skb int offset struct iovec * to int len skb buffer to copy offset offset in the buffer to start copying from to io vector to copy to len amount of data to copy from buffer to iovec the iovec is modified during the copy. Kernel Hackers Manual October 2009 skb_copy_datagram_const_iovec 9 2.6.32-rc5 skb_copy_datagram_const_iovec Copy a datagram to an iovec. int skb_copy_datagram_const_iovec const struct sk_buff * skb int offset const struct iovec * to int to_offset int len skb buffer to copy offset offset in the buffer to start copying from to io vector to copy to to_offset offset in the io vector to start copying to len amount of data to copy from buffer to iovec Returns 0 or -EFAULT. the iovec is not modified during the copy. Kernel Hackers Manual October 2009 skb_copy_datagram_from_iovec 9 2.6.32-rc5 skb_copy_datagram_from_iovec Copy a datagram from an iovec. int skb_copy_datagram_from_iovec struct sk_buff * skb int offset const struct iovec * from int from_offset int len skb buffer to copy offset offset in the buffer to start copying to from io vector to copy to from_offset offset in the io vector to start copying from len amount of data to copy to buffer from iovec Returns 0 or -EFAULT. the iovec is not modified during the copy. Kernel Hackers Manual October 2009 skb_copy_and_csum_datagram_iovec 9 2.6.32-rc5 skb_copy_and_csum_datagram_iovec Copy and checkum skb to user iovec. int skb_copy_and_csum_datagram_iovec struct sk_buff * skb int hlen struct iovec * iov skb skbuff hlen hardware length iov io vector Caller _must_ check that skb will fit to this iovec. 0 - success. -EINVAL - checksum failure. -EFAULT - fault during copy. Beware, in this case iovec can be modified! Kernel Hackers Manual October 2009 datagram_poll 9 2.6.32-rc5 datagram_poll generic datagram poll unsigned int datagram_poll struct file * file struct socket * sock poll_table * wait file file struct sock socket wait poll table Again totally generic. This also handles sequenced packet sockets providing the socket receive queue is only ever holding data ready to receive. when you _don't_ use this routine for this protocol, and you use a different write policy from sock_writeable then please supply your own write_space callback. Kernel Hackers Manual October 2009 sk_stream_write_space 9 2.6.32-rc5 sk_stream_write_space stream socket write_space callback. void sk_stream_write_space struct sock * sk sk socket write proper description Kernel Hackers Manual October 2009 sk_stream_wait_connect 9 2.6.32-rc5 sk_stream_wait_connect Wait for a socket to get into the connected state int sk_stream_wait_connect struct sock * sk long * timeo_p sk sock to wait on timeo_p for how long to wait Must be called with the socket locked. Kernel Hackers Manual October 2009 sk_stream_wait_memory 9 2.6.32-rc5 sk_stream_wait_memory Wait for more memory for a socket int sk_stream_wait_memory struct sock * sk long * timeo_p sk socket to wait for memory timeo_p for how long Kernel Hackers Manual October 2009 sk_filter 9 2.6.32-rc5 sk_filter run a packet through a socket filter int sk_filter struct sock * sk struct sk_buff * skb sk sock associated with sk_buff skb buffer to filter Run the filter code and then cut skb->data to correct size returned by sk_run_filter. If pkt_len is 0 we toss packet. If skb->len is smaller than pkt_len we keep whole skb->data. This is the socket level wrapper to sk_run_filter. It returns 0 if the packet should be accepted or -EPERM if the packet should be tossed. Kernel Hackers Manual October 2009 sk_run_filter 9 2.6.32-rc5 sk_run_filter run a filter on a socket unsigned int sk_run_filter struct sk_buff * skb struct sock_filter * filter int flen skb buffer to run the filter on filter filter to apply flen length of filter Decode and apply filter instructions to the skb->data. Return length to keep, 0 for none. skb is the data we are filtering, filter is the array of filter instructions, and len is the number of filter blocks in the array. Kernel Hackers Manual October 2009 sk_chk_filter 9 2.6.32-rc5 sk_chk_filter verify socket filter code int sk_chk_filter struct sock_filter * filter int flen filter filter to verify flen length of filter Check the user's filter code. If we let some ugly filter code slip through kaboom! The filter must contain no references or jumps that are out of range, no illegal instructions, and must end with a RET instruction. All jumps are forward as they are not signed. Returns 0 if the rule set is legal or -EINVAL if not. Kernel Hackers Manual October 2009 struct gnet_stats_basic 9 2.6.32-rc5 struct gnet_stats_basic byte/packet throughput statistics struct gnet_stats_basic { __u64 bytes; __u32 packets; }; bytes number of seen bytes packets number of seen packets Kernel Hackers Manual October 2009 struct gnet_stats_rate_est 9 2.6.32-rc5 struct gnet_stats_rate_est rate estimator struct gnet_stats_rate_est { __u32 bps; __u32 pps; }; bps current byte rate pps current packet rate Kernel Hackers Manual October 2009 struct gnet_stats_queue 9 2.6.32-rc5 struct gnet_stats_queue queuing statistics struct gnet_stats_queue { __u32 qlen; __u32 backlog; __u32 drops; __u32 requeues; __u32 overlimits; }; qlen queue length backlog backlog size of queue drops number of dropped packets requeues number of requeues overlimits number of enqueues over the limit Kernel Hackers Manual October 2009 struct gnet_estimator 9 2.6.32-rc5 struct gnet_estimator rate estimator configuration struct gnet_estimator { signed char interval; unsigned char ewma_log; }; interval sampling period ewma_log the log of measurement window weight Kernel Hackers Manual October 2009 gnet_stats_start_copy_compat 9 2.6.32-rc5 gnet_stats_start_copy_compat start dumping procedure in compatibility mode int gnet_stats_start_copy_compat struct sk_buff * skb int type int tc_stats_type int xstats_type spinlock_t * lock struct gnet_dump * d skb socket buffer to put statistics TLVs into type TLV type for top level statistic TLV tc_stats_type TLV type for backward compatibility struct tc_stats TLV xstats_type TLV type for backward compatibility xstats TLV lock statistics lock d dumping handle Initializes the dumping handle, grabs the statistic lock and appends an empty TLV header to the socket buffer for use a container for all other statistic TLVS. The dumping handle is marked to be in backward compatibility mode telling all gnet_stats_copy_XXX functions to fill a local copy of struct tc_stats. Returns 0 on success or -1 if the room in the socket buffer was not sufficient. Kernel Hackers Manual October 2009 gnet_stats_start_copy 9 2.6.32-rc5 gnet_stats_start_copy start dumping procedure in compatibility mode int gnet_stats_start_copy struct sk_buff * skb int type spinlock_t * lock struct gnet_dump * d skb socket buffer to put statistics TLVs into type TLV type for top level statistic TLV lock statistics lock d dumping handle Initializes the dumping handle, grabs the statistic lock and appends an empty TLV header to the socket buffer for use a container for all other statistic TLVS. Returns 0 on success or -1 if the room in the socket buffer was not sufficient. Kernel Hackers Manual October 2009 gnet_stats_copy_basic 9 2.6.32-rc5 gnet_stats_copy_basic copy basic statistics into statistic TLV int gnet_stats_copy_basic struct gnet_dump * d struct gnet_stats_basic_packed * b d dumping handle b basic statistics Appends the basic statistics to the top level TLV created by gnet_stats_start_copy. Returns 0 on success or -1 with the statistic lock released if the room in the socket buffer was not sufficient. Kernel Hackers Manual October 2009 gnet_stats_copy_rate_est 9 2.6.32-rc5 gnet_stats_copy_rate_est copy rate estimator statistics into statistics TLV int gnet_stats_copy_rate_est struct gnet_dump * d struct gnet_stats_rate_est * r d dumping handle r rate estimator statistics Appends the rate estimator statistics to the top level TLV created by gnet_stats_start_copy. Returns 0 on success or -1 with the statistic lock released if the room in the socket buffer was not sufficient. Kernel Hackers Manual October 2009 gnet_stats_copy_queue 9 2.6.32-rc5 gnet_stats_copy_queue copy queue statistics into statistics TLV int gnet_stats_copy_queue struct gnet_dump * d struct gnet_stats_queue * q d dumping handle q queue statistics Appends the queue statistics to the top level TLV created by gnet_stats_start_copy. Returns 0 on success or -1 with the statistic lock released if the room in the socket buffer was not sufficient. Kernel Hackers Manual October 2009 gnet_stats_copy_app 9 2.6.32-rc5 gnet_stats_copy_app copy application specific statistics into statistics TLV int gnet_stats_copy_app struct gnet_dump * d void * st int len d dumping handle st application specific statistics data len length of data Appends the application sepecific statistics to the top level TLV created by gnet_stats_start_copy and remembers the data for XSTATS if the dumping handle is in backward compatibility mode. Returns 0 on success or -1 with the statistic lock released if the room in the socket buffer was not sufficient. Kernel Hackers Manual October 2009 gnet_stats_finish_copy 9 2.6.32-rc5 gnet_stats_finish_copy finish dumping procedure int gnet_stats_finish_copy struct gnet_dump * d d dumping handle Corrects the length of the top level TLV to include all TLVs added by gnet_stats_copy_XXX calls. Adds the backward compatibility TLVs if gnet_stats_start_copy_compat was used and releases the statistics lock. Returns 0 on success or -1 with the statistic lock released if the room in the socket buffer was not sufficient. Kernel Hackers Manual October 2009 gen_new_estimator 9 2.6.32-rc5 gen_new_estimator create a new rate estimator int gen_new_estimator struct gnet_stats_basic_packed * bstats struct gnet_stats_rate_est * rate_est spinlock_t * stats_lock struct nlattr * opt bstats basic statistics rate_est rate estimator statistics stats_lock statistics lock opt rate estimator configuration TLV Creates a new rate estimator with bstats as source and rate_est as destination. A new timer with the interval specified in the configuration TLV is created. Upon each interval, the latest statistics will be read from bstats and the estimated rate will be stored in rate_est with the statistics lock grabed during this period. Returns 0 on success or a negative error code. Called under rtnl_mutex Kernel Hackers Manual October 2009 gen_kill_estimator 9 2.6.32-rc5 gen_kill_estimator remove a rate estimator void gen_kill_estimator struct gnet_stats_basic_packed * bstats struct gnet_stats_rate_est * rate_est bstats basic statistics rate_est rate estimator statistics Removes the rate estimator specified by bstats and rate_est. Called under rtnl_mutex Kernel Hackers Manual October 2009 gen_replace_estimator 9 2.6.32-rc5 gen_replace_estimator replace rate estimator configuration int gen_replace_estimator struct gnet_stats_basic_packed * bstats struct gnet_stats_rate_est * rate_est spinlock_t * stats_lock struct nlattr * opt bstats basic statistics rate_est rate estimator statistics stats_lock statistics lock opt rate estimator configuration TLV Replaces the configuration of a rate estimator by calling gen_kill_estimator and gen_new_estimator. Returns 0 on success or a negative error code. Kernel Hackers Manual October 2009 gen_estimator_active 9 2.6.32-rc5 gen_estimator_active test if estimator is currently in use bool gen_estimator_active const struct gnet_stats_basic_packed * bstats const struct gnet_stats_rate_est * rate_est bstats basic statistics rate_est rate estimator statistics Returns true if estimator is active, and false if not. Kernel Hackers Manual October 2009 xdr_encode_opaque_fixed 9 2.6.32-rc5 xdr_encode_opaque_fixed Encode fixed length opaque data __be32 * xdr_encode_opaque_fixed __be32 * p const void * ptr unsigned int nbytes p pointer to current position in XDR buffer. ptr pointer to data to encode (or NULL) nbytes size of data. Copy the array of data of length nbytes at ptr to the XDR buffer at position p, then align to the next 32-bit boundary by padding with zero bytes (see RFC1832). if ptr is NULL, only the padding is performed. Returns the updated current XDR buffer position Kernel Hackers Manual October 2009 xdr_encode_opaque 9 2.6.32-rc5 xdr_encode_opaque Encode variable length opaque data __be32 * xdr_encode_opaque __be32 * p const void * ptr unsigned int nbytes p pointer to current position in XDR buffer. ptr pointer to data to encode (or NULL) nbytes size of data. Returns the updated current XDR buffer position Kernel Hackers Manual October 2009 xdr_init_encode 9 2.6.32-rc5 xdr_init_encode Initialize a struct xdr_stream for sending data. void xdr_init_encode struct xdr_stream * xdr struct xdr_buf * buf __be32 * p xdr pointer to xdr_stream struct buf pointer to XDR buffer in which to encode data p current pointer inside XDR buffer at the moment the RPC client only passes the length of our scratch buffer in the xdr_buf's header kvec. Previously this meant we needed to call xdr_adjust_iovec after encoding the data. With the new scheme, the xdr_stream manages the details of the buffer length, and takes care of adjusting the kvec length for us. Kernel Hackers Manual October 2009 xdr_reserve_space 9 2.6.32-rc5 xdr_reserve_space Reserve buffer space for sending __be32 * xdr_reserve_space struct xdr_stream * xdr size_t nbytes xdr pointer to xdr_stream nbytes number of bytes to reserve Checks that we have enough buffer space to encode 'nbytes' more bytes of data. If so, update the total xdr_buf length, and adjust the length of the current kvec. Kernel Hackers Manual October 2009 xdr_write_pages 9 2.6.32-rc5 xdr_write_pages Insert a list of pages into an XDR buffer for sending void xdr_write_pages struct xdr_stream * xdr struct page ** pages unsigned int base unsigned int len xdr pointer to xdr_stream pages list of pages base offset of first byte len length of data in bytes Kernel Hackers Manual October 2009 xdr_init_decode 9 2.6.32-rc5 xdr_init_decode Initialize an xdr_stream for decoding data. void xdr_init_decode struct xdr_stream * xdr struct xdr_buf * buf __be32 * p xdr pointer to xdr_stream struct buf pointer to XDR buffer from which to decode data p current pointer inside XDR buffer Kernel Hackers Manual October 2009 xdr_inline_decode 9 2.6.32-rc5 xdr_inline_decode Retrieve non-page XDR data to decode __be32 * xdr_inline_decode struct xdr_stream * xdr size_t nbytes xdr pointer to xdr_stream struct nbytes number of bytes of data to decode Check if the input buffer is long enough to enable us to decode 'nbytes' more bytes of data starting at the current position. If so return the current pointer, then update the current pointer position. Kernel Hackers Manual October 2009 xdr_read_pages 9 2.6.32-rc5 xdr_read_pages Ensure page-based XDR data to decode is aligned at current pointer position void xdr_read_pages struct xdr_stream * xdr unsigned int len xdr pointer to xdr_stream struct len number of bytes of page data Moves data beyond the current pointer position from the XDR head[] buffer into the page list. Any data that lies beyond current position + len bytes is moved into the XDR tail[]. Kernel Hackers Manual October 2009 xdr_enter_page 9 2.6.32-rc5 xdr_enter_page decode data from the XDR page void xdr_enter_page struct xdr_stream * xdr unsigned int len xdr pointer to xdr_stream struct len number of bytes of page data Moves data beyond the current pointer position from the XDR head[] buffer into the page list. Any data that lies beyond current position + len bytes is moved into the XDR tail[]. The current pointer is then repositioned at the beginning of the first XDR page. Kernel Hackers Manual October 2009 svc_print_addr 9 2.6.32-rc5 svc_print_addr Format rq_addr field for printing char * svc_print_addr struct svc_rqst * rqstp char * buf size_t len rqstp svc_rqst struct containing address to print buf target buffer for formatted address len length of target buffer Kernel Hackers Manual October 2009 svc_reserve 9 2.6.32-rc5 svc_reserve change the space reserved for the reply to a request. void svc_reserve struct svc_rqst * rqstp int space rqstp The request in question space new max space to reserve Each request reserves some space on the output queue of the transport to make sure the reply fits. This function reduces that reserved space to be the amount of space used already, plus space. Kernel Hackers Manual October 2009 svc_find_xprt 9 2.6.32-rc5 svc_find_xprt find an RPC transport instance struct svc_xprt * svc_find_xprt struct svc_serv * serv const char * xcl_name const sa_family_t af const unsigned short port serv pointer to svc_serv to search xcl_name C string containing transport's class name af Address family of transport's local address port transport's IP port number Return the transport instance pointer for the endpoint accepting connections/peer traffic from the specified transport class, address family and port. Specifying 0 for the address family or port is effectively a wild-card, and will result in matching the first transport in the service's list that has a matching class name. Kernel Hackers Manual October 2009 svc_xprt_names 9 2.6.32-rc5 svc_xprt_names format a buffer with a list of transport names int svc_xprt_names struct svc_serv * serv char * buf const int buflen serv pointer to an RPC service buf pointer to a buffer to be filled in buflen length of buffer to be filled in Fills in buf with a string containing a list of transport names, each name terminated with '\n'. Returns positive length of the filled-in string on success; otherwise a negative errno value is returned if an error occurs. Kernel Hackers Manual October 2009 xprt_register_transport 9 2.6.32-rc5 xprt_register_transport register a transport implementation int xprt_register_transport struct xprt_class * transport transport transport to register If a transport implementation is loaded as a kernel module, it can call this interface to make itself known to the RPC client. transport successfully registered -EEXIST: transport already registered -EINVAL: transport module being unloaded Kernel Hackers Manual October 2009 xprt_unregister_transport 9 2.6.32-rc5 xprt_unregister_transport unregister a transport implementation int xprt_unregister_transport struct xprt_class * transport transport transport to unregister transport successfully unregistered -ENOENT: transport never registered Kernel Hackers Manual October 2009 xprt_load_transport 9 2.6.32-rc5 xprt_load_transport load a transport implementation int xprt_load_transport const char * transport_name transport_name transport to load transport successfully loaded -ENOENT: transport module not available Kernel Hackers Manual October 2009 xprt_reserve_xprt 9 2.6.32-rc5 xprt_reserve_xprt serialize write access to transports int xprt_reserve_xprt struct rpc_task * task task task that is requesting access to the transport This prevents mixing the payload of separate requests, and prevents transport connects from colliding with writes. No congestion control is provided. Kernel Hackers Manual October 2009 xprt_release_xprt 9 2.6.32-rc5 xprt_release_xprt allow other requests to use a transport void xprt_release_xprt struct rpc_xprt * xprt struct rpc_task * task xprt transport with other tasks potentially waiting task task that is releasing access to the transport Note that task can be NULL. No congestion control is provided. Kernel Hackers Manual October 2009 xprt_release_xprt_cong 9 2.6.32-rc5 xprt_release_xprt_cong allow other requests to use a transport void xprt_release_xprt_cong struct rpc_xprt * xprt struct rpc_task * task xprt transport with other tasks potentially waiting task task that is releasing access to the transport Note that task can be NULL. Another task is awoken to use the transport if the transport's congestion window allows it. Kernel Hackers Manual October 2009 xprt_release_rqst_cong 9 2.6.32-rc5 xprt_release_rqst_cong housekeeping when request is complete void xprt_release_rqst_cong struct rpc_task * task task RPC request that recently completed Useful for transports that require congestion control. Kernel Hackers Manual October 2009 xprt_adjust_cwnd 9 2.6.32-rc5 xprt_adjust_cwnd adjust transport congestion window void xprt_adjust_cwnd struct rpc_task * task int result task recently completed RPC request used to adjust window result result code of completed RPC request We use a time-smoothed congestion estimator to avoid heavy oscillation. Kernel Hackers Manual October 2009 xprt_wake_pending_tasks 9 2.6.32-rc5 xprt_wake_pending_tasks wake all tasks on a transport's pending queue void xprt_wake_pending_tasks struct rpc_xprt * xprt int status xprt transport with waiting tasks status result code to plant in each task before waking it Kernel Hackers Manual October 2009 xprt_wait_for_buffer_space 9 2.6.32-rc5 xprt_wait_for_buffer_space wait for transport output buffer to clear void xprt_wait_for_buffer_space struct rpc_task * task rpc_action action task task to be put to sleep action function pointer to be executed after wait Kernel Hackers Manual October 2009 xprt_write_space 9 2.6.32-rc5 xprt_write_space wake the task waiting for transport output buffer space void xprt_write_space struct rpc_xprt * xprt xprt transport with waiting tasks Can be called in a soft IRQ context, so xprt_write_space never sleeps. Kernel Hackers Manual October 2009 xprt_set_retrans_timeout_def 9 2.6.32-rc5 xprt_set_retrans_timeout_def set a request's retransmit timeout void xprt_set_retrans_timeout_def struct rpc_task * task task task whose timeout is to be set Set a request's retransmit timeout based on the transport's default timeout parameters. Used by transports that don't adjust the retransmit timeout based on round-trip time estimation. Kernel Hackers Manual October 2009 xprt_disconnect_done 9 2.6.32-rc5 xprt_disconnect_done mark a transport as disconnected void xprt_disconnect_done struct rpc_xprt * xprt xprt transport to flag for disconnect Kernel Hackers Manual October 2009 xprt_lookup_rqst 9 2.6.32-rc5 xprt_lookup_rqst find an RPC request corresponding to an XID struct rpc_rqst * xprt_lookup_rqst struct rpc_xprt * xprt __be32 xid xprt transport on which the original request was transmitted xid RPC XID of incoming reply Kernel Hackers Manual October 2009 xprt_update_rtt 9 2.6.32-rc5 xprt_update_rtt update an RPC client's RTT state after receiving a reply void xprt_update_rtt struct rpc_task * task task RPC request that recently completed Kernel Hackers Manual October 2009 xprt_complete_rqst 9 2.6.32-rc5 xprt_complete_rqst called when reply processing is complete void xprt_complete_rqst struct rpc_task * task int copied task RPC request that recently completed copied actual number of bytes received from the transport Caller holds transport lock. Kernel Hackers Manual October 2009 rpc_wake_up 9 2.6.32-rc5 rpc_wake_up wake up all rpc_tasks void rpc_wake_up struct rpc_wait_queue * queue queue rpc_wait_queue on which the tasks are sleeping Grabs queue->lock Kernel Hackers Manual October 2009 rpc_wake_up_status 9 2.6.32-rc5 rpc_wake_up_status wake up all rpc_tasks and set their status value. void rpc_wake_up_status struct rpc_wait_queue * queue int status queue rpc_wait_queue on which the tasks are sleeping status status value to set Grabs queue->lock Kernel Hackers Manual October 2009 rpc_malloc 9 2.6.32-rc5 rpc_malloc allocate an RPC buffer void * rpc_malloc struct rpc_task * task size_t size task RPC task that will use this buffer size requested byte size To prevent rpciod from hanging, this allocator never sleeps, returning NULL if the request cannot be serviced immediately. The caller can arrange to sleep in a way that is safe for rpciod. Most requests are 'small' (under 2KiB) and can be serviced from a mempool, ensuring that NFS reads and writes can always proceed, and that there is good locality of reference for these buffers. In order to avoid memory starvation triggering more writebacks of NFS requests, we avoid using GFP_KERNEL. Kernel Hackers Manual October 2009 rpc_free 9 2.6.32-rc5 rpc_free free buffer allocated via rpc_malloc void rpc_free void * buffer buffer buffer to free Kernel Hackers Manual October 2009 xdr_skb_read_bits 9 2.6.32-rc5 xdr_skb_read_bits copy some data bits from skb to internal buffer size_t xdr_skb_read_bits struct xdr_skb_reader * desc void * to size_t len desc sk_buff copy helper to copy destination len number of bytes to copy Possibly called several times to iterate over an sk_buff and copy data out of it. Kernel Hackers Manual October 2009 xdr_partial_copy_from_skb 9 2.6.32-rc5 xdr_partial_copy_from_skb copy data out of an skb ssize_t xdr_partial_copy_from_skb struct xdr_buf * xdr unsigned int base struct xdr_skb_reader * desc xdr_skb_read_actor copy_actor xdr target XDR buffer base starting offset desc sk_buff copy helper copy_actor virtual method for copying data Kernel Hackers Manual October 2009 csum_partial_copy_to_xdr 9 2.6.32-rc5 csum_partial_copy_to_xdr checksum and copy data int csum_partial_copy_to_xdr struct xdr_buf * xdr struct sk_buff * skb xdr target XDR buffer skb source skb We have set things up such that we perform the checksum of the UDP packet in parallel with the copies into the RPC client iovec. -DaveM Kernel Hackers Manual October 2009 rpc_alloc_iostats 9 2.6.32-rc5 rpc_alloc_iostats allocate an rpc_iostats structure struct rpc_iostats * rpc_alloc_iostats struct rpc_clnt * clnt clnt RPC program, version, and xprt Kernel Hackers Manual October 2009 rpc_free_iostats 9 2.6.32-rc5 rpc_free_iostats release an rpc_iostats structure void rpc_free_iostats struct rpc_iostats * stats stats doomed rpc_iostats structure Kernel Hackers Manual October 2009 rpc_queue_upcall 9 2.6.32-rc5 rpc_queue_upcall int rpc_queue_upcall struct inode * inode struct rpc_pipe_msg * msg inode inode of upcall pipe on which to queue given message msg message to queue Call with an inode created by rpc_mkpipe to queue an upcall. A userspace process may then later read the upcall by performing a read on an open file for this inode. It is up to the caller to initialize the fields of msg (other than msg->list) appropriately. Kernel Hackers Manual October 2009 rpc_mkpipe 9 2.6.32-rc5 rpc_mkpipe make an rpc_pipefs file for kernel<->userspace communication struct dentry * rpc_mkpipe struct dentry * parent const char * name void * private const struct rpc_pipe_ops * ops int flags parent dentry of directory to create new pipe in name name of pipe private private data to associate with the pipe, for the caller's use ops operations defining the behavior of the pipe: upcall, downcall, release_pipe, open_pipe, and destroy_msg. flags rpc_inode flags Data is made available for userspace to read by calls to rpc_queue_upcall. The actual reads will result in calls to ops->upcall, which will be called with the file pointer, message, and userspace buffer to copy to. Writes can come at any time, and do not necessarily have to be responses to upcalls. They will result in calls to msg->downcall. The private argument passed here will be available to all these methods from the file pointer, via RPC_I(file->f_dentry->d_inode)->private. Kernel Hackers Manual October 2009 rpc_unlink 9 2.6.32-rc5 rpc_unlink remove a pipe int rpc_unlink struct dentry * dentry dentry dentry for the pipe, as returned from rpc_mkpipe After this call, lookups will no longer find the pipe, and any attempts to read or write using preexisting opens of the pipe will return -EPIPE. Kernel Hackers Manual October 2009 rpcb_getport_sync 9 2.6.32-rc5 rpcb_getport_sync obtain the port for an RPC service on a given host int rpcb_getport_sync struct sockaddr_in * sin u32 prog u32 vers int prot sin address of remote peer prog RPC program number to bind vers RPC version number to bind prot transport protocol to use to make this request Return value is the requested advertised port number, or a negative errno value. Called from outside the RPC client in a synchronous task context. Uses default timeout parameters specified by underlying transport. Needs to support IPv6 Kernel Hackers Manual October 2009 rpcb_getport_async 9 2.6.32-rc5 rpcb_getport_async obtain the port for a given RPC service on a given host void rpcb_getport_async struct rpc_task * task task task that is waiting for portmapper request This one can be called for an ongoing RPC request, and can be used in an async (rpciod) context. Kernel Hackers Manual October 2009 rpc_bind_new_program 9 2.6.32-rc5 rpc_bind_new_program bind a new RPC program to an existing client struct rpc_clnt * rpc_bind_new_program struct rpc_clnt * old struct rpc_program * program u32 vers old old rpc_client program rpc program to set vers rpc program version Clones the rpc client and sets up a new RPC program. This is mainly of use for enabling different RPC programs to share the same transport. The Sun NFSv2/v3 ACL protocol can do this. Kernel Hackers Manual October 2009 rpc_run_task 9 2.6.32-rc5 rpc_run_task Allocate a new RPC task, then run rpc_execute against it struct rpc_task * rpc_run_task const struct rpc_task_setup * task_setup_data task_setup_data pointer to task initialisation data Kernel Hackers Manual October 2009 rpc_call_sync 9 2.6.32-rc5 rpc_call_sync Perform a synchronous RPC call int rpc_call_sync struct rpc_clnt * clnt const struct rpc_message * msg int flags clnt pointer to RPC client msg RPC call parameters flags RPC call flags Kernel Hackers Manual October 2009 rpc_call_async 9 2.6.32-rc5 rpc_call_async Perform an asynchronous RPC call int rpc_call_async struct rpc_clnt * clnt const struct rpc_message * msg int flags const struct rpc_call_ops * tk_ops void * data clnt pointer to RPC client msg RPC call parameters flags RPC call flags tk_ops RPC call ops data user call data Kernel Hackers Manual October 2009 rpc_peeraddr 9 2.6.32-rc5 rpc_peeraddr extract remote peer address from clnt's xprt size_t rpc_peeraddr struct rpc_clnt * clnt struct sockaddr * buf size_t bufsize clnt RPC client structure buf target buffer bufsize length of target buffer Returns the number of bytes that are actually in the stored address. Kernel Hackers Manual October 2009 rpc_peeraddr2str 9 2.6.32-rc5 rpc_peeraddr2str return remote peer address in printable format const char * rpc_peeraddr2str struct rpc_clnt * clnt enum rpc_display_format_t format clnt RPC client structure format address format Kernel Hackers Manual October 2009 rpc_force_rebind 9 2.6.32-rc5 rpc_force_rebind force transport to check that remote port is unchanged void rpc_force_rebind struct rpc_clnt * clnt clnt client to rebind Kernel Hackers Manual October 2009 wimax_msg_alloc 9 2.6.32-rc5 wimax_msg_alloc Create a new skb for sending a message to userspace struct sk_buff * wimax_msg_alloc struct wimax_dev * wimax_dev const char * pipe_name const void * msg size_t size gfp_t gfp_flags wimax_dev WiMAX device descriptor pipe_name "named pipe" the message will be sent to msg pointer to the message data to send size size of the message to send (in bytes), including the header. gfp_flags flags for memory allocation. 0 if ok, negative errno code on error Allocates an skb that will contain the message to send to user space over the messaging pipe and initializes it, copying the payload. Once this call is done, you can deliver it with wimax_msg_send. Don't use skb_push/skb_pull/skb_reserve on the skb, as wimax_msg_send depends on skb->data being placed at the beginning of the user message. Unlike other WiMAX stack calls, this call can be used way early, even before wimax_dev_add is called, as long as the wimax_dev->net_dev pointer is set to point to a proper net_dev. This is so that drivers can use it early in case they need to send stuff around or communicate with user space. Kernel Hackers Manual October 2009 wimax_msg_data_len 9 2.6.32-rc5 wimax_msg_data_len Return a pointer and size of a message's payload const void * wimax_msg_data_len struct sk_buff * msg size_t * size msg Pointer to a message created with wimax_msg_alloc size Pointer to where to store the message's size Returns the pointer to the message data. Kernel Hackers Manual October 2009 wimax_msg_data 9 2.6.32-rc5 wimax_msg_data Return a pointer to a message's payload const void * wimax_msg_data struct sk_buff * msg msg Pointer to a message created with wimax_msg_alloc Kernel Hackers Manual October 2009 wimax_msg_len 9 2.6.32-rc5 wimax_msg_len Return a message's payload length ssize_t wimax_msg_len struct sk_buff * msg msg Pointer to a message created with wimax_msg_alloc Kernel Hackers Manual October 2009 wimax_msg_send 9 2.6.32-rc5 wimax_msg_send Send a pre-allocated message to user space int wimax_msg_send struct wimax_dev * wimax_dev struct sk_buff * skb wimax_dev WiMAX device descriptor skb struct sk_buff returned by wimax_msg_alloc. Note the ownership of skb is transferred to this function. 0 if ok, < 0 errno code on error Sends a free-form message that was preallocated with wimax_msg_alloc and filled up. Assumes that once you pass an skb to this function for sending, it owns it and will release it when done (on success). Don't use skb_push/skb_pull/skb_reserve on the skb, as wimax_msg_send depends on skb->data being placed at the beginning of the user message. Unlike other WiMAX stack calls, this call can be used way early, even before wimax_dev_add is called, as long as the wimax_dev->net_dev pointer is set to point to a proper net_dev. This is so that drivers can use it early in case they need to send stuff around or communicate with user space. Kernel Hackers Manual October 2009 wimax_msg 9 2.6.32-rc5 wimax_msg Send a message to user space int wimax_msg struct wimax_dev * wimax_dev const char * pipe_name const void * buf size_t size gfp_t gfp_flags wimax_dev WiMAX device descriptor (properly referenced) pipe_name "named pipe" the message will be sent to buf pointer to the message to send. size size of the buffer pointed to by buf (in bytes). gfp_flags flags for memory allocation. 0 if ok, negative errno code on error. Sends a free-form message to user space on the device wimax_dev. Once the skb is given to this function, who will own it and will release it when done (unless it returns error). Kernel Hackers Manual October 2009 wimax_reset 9 2.6.32-rc5 wimax_reset Reset a WiMAX device int wimax_reset struct wimax_dev * wimax_dev wimax_dev WiMAX device descriptor 0 if ok and a warm reset was done (the device still exists in the system). -ENODEV if a cold/bus reset had to be done (device has disconnected and reconnected, so current handle is not valid any more). -EINVAL if the device is not even registered. Any other negative error code shall be considered as non-recoverable. Called when wanting to reset the device for any reason. Device is taken back to power on status. This call blocks; on succesful return, the device has completed the reset process and is ready to operate. Kernel Hackers Manual October 2009 wimax_report_rfkill_hw 9 2.6.32-rc5 wimax_report_rfkill_hw Reports changes in the hardware RF switch void wimax_report_rfkill_hw struct wimax_dev * wimax_dev enum wimax_rf_state state wimax_dev WiMAX device descriptor state New state of the RF Kill switch. WIMAX_RF_ON radio on, WIMAX_RF_OFF radio off. When the device detects a change in the state of thehardware RF switch, it must call this function to let the WiMAX kernel stack know that the state has changed so it can be properly propagated. The WiMAX stack caches the state (the driver doesn't need to). As well, as the change is propagated it will come back as a request to change the software state to mirror the hardware state. If the device doesn't have a hardware kill switch, just report it on initialization as always on (WIMAX_RF_ON, radio on). Kernel Hackers Manual October 2009 wimax_report_rfkill_sw 9 2.6.32-rc5 wimax_report_rfkill_sw Reports changes in the software RF switch void wimax_report_rfkill_sw struct wimax_dev * wimax_dev enum wimax_rf_state state wimax_dev WiMAX device descriptor state New state of the RF kill switch. WIMAX_RF_ON radio on, WIMAX_RF_OFF radio off. Reports changes in the software RF switch state to the the WiMAX stack. The main use is during initialization, so the driver can query the device for its current software radio kill switch state and feed it to the system. On the side, the device does not change the software state by itself. In practice, this can happen, as the device might decide to switch (in software) the radio off for different reasons. Kernel Hackers Manual October 2009 wimax_rfkill 9 2.6.32-rc5 wimax_rfkill Set the software RF switch state for a WiMAX device int wimax_rfkill struct wimax_dev * wimax_dev enum wimax_rf_state state wimax_dev WiMAX device descriptor state New RF state. >= 0 toggle state if ok, < 0 errno code on error. The toggle state is returned as a bitmap, bit 0 being the hardware RF state, bit 1 the software RF state. 0 means disabled (WIMAX_RF_ON, radio on), 1 means enabled radio off (WIMAX_RF_OFF). Called by the user when he wants to request the WiMAX radio to be switched on (WIMAX_RF_ON) or off (WIMAX_RF_OFF). With WIMAX_RF_QUERY, just the current state is returned. This call will block until the operation is complete. Kernel Hackers Manual October 2009 wimax_state_change 9 2.6.32-rc5 wimax_state_change Set the current state of a WiMAX device void wimax_state_change struct wimax_dev * wimax_dev enum wimax_st new_state wimax_dev WiMAX device descriptor (properly referenced) new_state New state to switch to This implements the state changes for the wimax devices. It will - verify that the state transition is legal (for now it'll just print a warning if not) according to the table in linux/wimax.h's documentation for 'enum wimax_st'. - perform the actions needed for leaving the current state and whichever are needed for entering the new state. - issue a report to user space indicating the new state (and an optional payload with information about the new state). wimax_dev must be locked Kernel Hackers Manual October 2009 wimax_state_get 9 2.6.32-rc5 wimax_state_get Return the current state of a WiMAX device enum wimax_st wimax_state_get struct wimax_dev * wimax_dev wimax_dev WiMAX device descriptor Current state of the device according to its driver. Kernel Hackers Manual October 2009 wimax_dev_init 9 2.6.32-rc5 wimax_dev_init initialize a newly allocated instance void wimax_dev_init struct wimax_dev * wimax_dev wimax_dev WiMAX device descriptor to initialize. Initializes fields of a freshly allocated wimax_dev instance. This function assumes that after allocation, the memory occupied by wimax_dev was zeroed. Kernel Hackers Manual October 2009 wimax_dev_add 9 2.6.32-rc5 wimax_dev_add Register a new WiMAX device int wimax_dev_add struct wimax_dev * wimax_dev struct net_device * net_dev wimax_dev WiMAX device descriptor (as embedded in your net_dev's priv data). You must have called wimax_dev_init on it before. net_dev net device the wimax_dev is associated with. The function expects SET_NETDEV_DEV and register_netdev were already called on it. Registers the new WiMAX device, sets up the user-kernel control interface (generic netlink) and common WiMAX infrastructure. Note that the parts that will allow interaction with user space are setup at the very end, when the rest is in place, as once that happens, the driver might get user space control requests via netlink or from debugfs that might translate into calls into wimax_dev->op_*(). Kernel Hackers Manual October 2009 wimax_dev_rm 9 2.6.32-rc5 wimax_dev_rm Unregister an existing WiMAX device void wimax_dev_rm struct wimax_dev * wimax_dev wimax_dev WiMAX device descriptor Unregisters a WiMAX device previously registered for use with wimax_add_rm. IMPORTANT! Must call before calling unregister_netdev. After this function returns, you will not get any more user space control requests (via netlink or debugfs) and thus to wimax_dev->ops. Reentrancy control is ensured by setting the state to __WIMAX_ST_QUIESCING. rfkill operations coming through wimax_*rfkill*() will be stopped by the quiescing state; ops coming from the rfkill subsystem will be stopped by the support being removed by wimax_rfkill_rm. Kernel Hackers Manual October 2009 struct wimax_dev 9 2.6.32-rc5 struct wimax_dev Generic WiMAX device struct wimax_dev { struct net_device * net_dev; struct list_head id_table_node; struct mutex mutex; struct mutex mutex_reset; enum wimax_st state; int (* op_msg_from_user) (struct wimax_dev *wimax_dev,const char *,const void *, size_t,const struct genl_info *info); int (* op_rfkill_sw_toggle) (struct wimax_dev *wimax_dev,enum wimax_rf_state); int (* op_reset) (struct wimax_dev *wimax_dev); struct rfkill * rfkill; unsigned rf_hw; unsigned rf_sw; char name[32]; struct dentry * debugfs_dentry; }; net_dev [fill] Pointer to the struct net_device this WiMAX device implements. id_table_node [private] link to the list of wimax devices kept by id-table.c. Protected by it's own spinlock. mutex [private] Serializes all concurrent access and execution of operations. mutex_reset [private] Serializes reset operations. Needs to be a different mutex because as part of the reset operation, the driver has to call back into the stack to do things such as state change, that require wimax_dev->mutex. state [private] Current state of the WiMAX device. op_msg_from_user [fill] Driver-specific operation to handle a raw message from user space to the driver. The driver can send messages to user space using with wimax_msg_to_user. op_rfkill_sw_toggle [fill] Driver-specific operation to act on userspace (or any other agent) requesting the WiMAX device to change the RF Kill software switch (WIMAX_RF_ON or WIMAX_RF_OFF). If such hardware support is not present, it is assumed the radio cannot be switched off and it is always on (and the stack will error out when trying to switch it off). In such case, this function pointer can be left as NULL. op_reset [fill] Driver specific operation to reset the device. This operation should always attempt first a warm reset that does not disconnect the device from the bus and return 0. If that fails, it should resort to some sort of cold or bus reset (even if it implies a bus disconnection and device dissapearance). In that case, -ENODEV should be returned to indicate the device is gone. This operation has to be synchronous, and return only when the reset is complete. In case of having had to resort to bus/cold reset implying a device disconnection, the call is allowed to return inmediately. rfkill [private] integration into the RF-Kill infrastructure. rf_hw [private] State of the hardware radio switch (OFF/ON) rf_sw [private] State of the software radio switch (OFF/ON) name[32] [fill] A way to identify this device. We need to register a name with many subsystems (rfkill, workqueue creation, etc). We can't use the network device name as that might change and in some instances we don't know it yet (until we don't call register_netdev). So we generate an unique one using the driver name and device bus id, place it here and use it across the board. Recommended naming: DRIVERNAME-BUSNAME:BUSID (dev->bus->name, dev->bus_id). debugfs_dentry [private] Used to hook up a debugfs entry. This shows up in the debugfs root as wimax\:DEVICENAME. wimax_dev->mutex is NOT locked when this op is being called; however, wimax_dev->mutex_reset IS locked to ensure serialization of calls to wimax_reset. See wimax_reset's documentation. This structure defines a common interface to access all WiMAX devices from different vendors and provides a common API as well as a free-form device-specific messaging channel. 1. Embed a struct wimax_dev at *the beginning* the network device structure so that netdev_priv points to it. 2. memset it to zero 3. Initialize with wimax_dev_init. This will leave the WiMAX device in the __WIMAX_ST_NULL state. 4. Fill all the fields marked with [fill]; once called wimax_dev_add, those fields CANNOT be modified. 5. Call wimax_dev_add *after* registering the network device. This will leave the WiMAX device in the WIMAX_ST_DOWN state. Protect the driver's net_device->open against succeeding if the wimax device state is lower than WIMAX_ST_DOWN. 6. Select when the device is going to be turned on/initialized; for example, it could be initialized on 'ifconfig up' (when the netdev op 'open' is called on the driver). When the device is initialized (at `ifconfig up` time, or right after calling wimax_dev_add from _probe, make sure the following steps are taken a. Move the device to WIMAX_ST_UNINITIALIZED. This is needed so some API calls that shouldn't work until the device is ready can be blocked. b. Initialize the device. Make sure to turn the SW radio switch off and move the device to state WIMAX_ST_RADIO_OFF when done. When just initialized, a device should be left in RADIO OFF state until user space devices to turn it on. c. Query the device for the state of the hardware rfkill switch and call wimax_rfkill_report_hw and wimax_rfkill_report_sw as needed. See below. wimax_dev_rm undoes before unregistering the network device. Once wimax_dev_add is called, the driver can get called on the wimax_dev->op_* function pointers The stack provides a mutex for each device that will disallow API calls happening concurrently; thus, op calls into the driver through the wimax_dev->op*() function pointers will always be serialized and *never* concurrent. For locking, take wimax_dev->mutex is taken; (most) operations in the API have to check for wimax_dev_is_ready to return 0 before continuing (this is done internally). The WiMAX device is reference counted by the associated network device. The only operation that can be used to reference the device is wimax_dev_get_by_genl_info, and the reference it acquires has to be released with dev_put(wimax_dev->net_dev). At startup, both HW and SW radio switchess are assumed to be off. At initialization time [after calling wimax_dev_add], have the driver query the device for the status of the software and hardware RF kill switches and call wimax_report_rfkill_hw and wimax_rfkill_report_sw to indicate their state. If any is missing, just call it to indicate it is ON (radio always on). Whenever the driver detects a change in the state of the RF kill switches, it should call wimax_report_rfkill_hw or wimax_report_rfkill_sw to report it to the stack. Kernel Hackers Manual October 2009 enum wimax_st 9 2.6.32-rc5 enum wimax_st The different states of a WiMAX device enum wimax_st { __WIMAX_ST_NULL, WIMAX_ST_DOWN, __WIMAX_ST_QUIESCING, WIMAX_ST_UNINITIALIZED, WIMAX_ST_RADIO_OFF, WIMAX_ST_READY, WIMAX_ST_SCANNING, WIMAX_ST_CONNECTING, WIMAX_ST_CONNECTED, __WIMAX_ST_INVALID }; __WIMAX_ST_NULL The device structure has been allocated and zeroed, but still wimax_dev_add hasn't been called. There is no state. WIMAX_ST_DOWN The device has been registered with the WiMAX and networking stacks, but it is not initialized (normally that is done with 'ifconfig DEV up' [or equivalent], which can upload firmware and enable communications with the device). In this state, the device is powered down and using as less power as possible. This state is the default after a call to wimax_dev_add. It is ok to have drivers move directly to WIMAX_ST_UNINITIALIZED or WIMAX_ST_RADIO_OFF in _probe after the call to wimax_dev_add. It is recommended that the driver leaves this state when calling 'ifconfig DEV up' and enters it back on 'ifconfig DEV down'. __WIMAX_ST_QUIESCING The device is being torn down, so no API operations are allowed to proceed except the ones needed to complete the device clean up process. WIMAX_ST_UNINITIALIZED [optional] Communication with the device is setup, but the device still requires some configuration before being operational. Some WiMAX API calls might work. WIMAX_ST_RADIO_OFF The device is fully up; radio is off (wether by hardware or software switches). It is recommended to always leave the device in this state after initialization. WIMAX_ST_READY The device is fully up and radio is on. WIMAX_ST_SCANNING [optional] The device has been instructed to scan. In this state, the device cannot be actively connected to a network. WIMAX_ST_CONNECTING The device is connecting to a network. This state exists because in some devices, the connect process can include a number of negotiations between user space, kernel space and the device. User space needs to know what the device is doing. If the connect sequence in a device is atomic and fast, the device can transition directly to CONNECTED WIMAX_ST_CONNECTED The device is connected to a network. __WIMAX_ST_INVALID This is an invalid state used to mark the maximum numeric value of states. Transitions from one state to another one are atomic and can only be caused in kernel space with wimax_state_change. To read the state, use wimax_state_get. States starting with __ are internal and shall not be used or referred to by drivers or userspace. They look ugly, but that's the point -- if any use is made non-internal to the stack, it is easier to catch on review. All API operations [with well defined exceptions] will take the device mutex before starting and then check the state. If the state is __WIMAX_ST_NULL, WIMAX_ST_DOWN, WIMAX_ST_UNINITIALIZED or __WIMAX_ST_QUIESCING, it will drop the lock and quit with -EINVAL, -ENOMEDIUM, -ENOTCONN or -ESHUTDOWN. The order of the definitions is important, so we can do numerical comparisons (eg: < WIMAX_ST_RADIO_OFF means the device is not ready to operate). Kernel Hackers Manual October 2009 dev_add_pack 9 2.6.32-rc5 dev_add_pack add packet handler void dev_add_pack struct packet_type * pt pt packet type declaration Add a protocol handler to the networking stack. The passed packet_type is linked into kernel lists and may not be freed until it has been removed from the kernel lists. This call does not sleep therefore it can not guarantee all CPU's that are in middle of receiving packets will see the new packet type (until the next received packet). Kernel Hackers Manual October 2009 __dev_remove_pack 9 2.6.32-rc5 __dev_remove_pack remove packet handler void __dev_remove_pack struct packet_type * pt pt packet type declaration Remove a protocol handler that was previously added to the kernel protocol handlers by dev_add_pack. The passed packet_type is removed from the kernel lists and can be freed or reused once this function returns. The packet type might still be in use by receivers and must not be freed until after all the CPU's have gone through a quiescent state. Kernel Hackers Manual October 2009 dev_remove_pack 9 2.6.32-rc5 dev_remove_pack remove packet handler void dev_remove_pack struct packet_type * pt pt packet type declaration Remove a protocol handler that was previously added to the kernel protocol handlers by dev_add_pack. The passed packet_type is removed from the kernel lists and can be freed or reused once this function returns. This call sleeps to guarantee that no CPU is looking at the packet type after return. Kernel Hackers Manual October 2009 netdev_boot_setup_check 9 2.6.32-rc5 netdev_boot_setup_check check boot time settings int netdev_boot_setup_check struct net_device * dev dev the netdevice Check boot time settings for the device. The found settings are set for the device to be used later in the device probing. Returns 0 if no settings found, 1 if they are. Kernel Hackers Manual October 2009 __dev_get_by_name 9 2.6.32-rc5 __dev_get_by_name find a device by its name struct net_device * __dev_get_by_name struct net * net const char * name net the applicable net namespace name name to find Find an interface by name. Must be called under RTNL semaphore or dev_base_lock. If the name is found a pointer to the device is returned. If the name is not found then NULL is returned. The reference counters are not incremented so the caller must be careful with locks. Kernel Hackers Manual October 2009 dev_get_by_name 9 2.6.32-rc5 dev_get_by_name find a device by its name struct net_device * dev_get_by_name struct net * net const char * name net the applicable net namespace name name to find Find an interface by name. This can be called from any context and does its own locking. The returned handle has the usage count incremented and the caller must use dev_put to release it when it is no longer needed. NULL is returned if no matching device is found. Kernel Hackers Manual October 2009 __dev_get_by_index 9 2.6.32-rc5 __dev_get_by_index find a device by its ifindex struct net_device * __dev_get_by_index struct net * net int ifindex net the applicable net namespace ifindex index of device Search for an interface by index. Returns NULL if the device is not found or a pointer to the device. The device has not had its reference counter increased so the caller must be careful about locking. The caller must hold either the RTNL semaphore or dev_base_lock. Kernel Hackers Manual October 2009 dev_get_by_index 9 2.6.32-rc5 dev_get_by_index find a device by its ifindex struct net_device * dev_get_by_index struct net * net int ifindex net the applicable net namespace ifindex index of device Search for an interface by index. Returns NULL if the device is not found or a pointer to the device. The device returned has had a reference added and the pointer is safe until the user calls dev_put to indicate they have finished with it. Kernel Hackers Manual October 2009 dev_getbyhwaddr 9 2.6.32-rc5 dev_getbyhwaddr find a device by its hardware address struct net_device * dev_getbyhwaddr struct net * net unsigned short type char * ha net the applicable net namespace type media type of device ha hardware address Search for an interface by MAC address. Returns NULL if the device is not found or a pointer to the device. The caller must hold the rtnl semaphore. The returned device has not had its ref count increased and the caller must therefore be careful about locking If the API was consistent this would be __dev_get_by_hwaddr Kernel Hackers Manual October 2009 dev_get_by_flags 9 2.6.32-rc5 dev_get_by_flags find any device with given flags struct net_device * dev_get_by_flags struct net * net unsigned short if_flags unsigned short mask net the applicable net namespace if_flags IFF_* values mask bitmask of bits in if_flags to check Search for any interface with the given flags. Returns NULL if a device is not found or a pointer to the device. The device returned has had a reference added and the pointer is safe until the user calls dev_put to indicate they have finished with it. Kernel Hackers Manual October 2009 dev_valid_name 9 2.6.32-rc5 dev_valid_name check if name is okay for network device int dev_valid_name const char * name name name string Network device names need to be valid file names to to allow sysfs to work. We also disallow any kind of whitespace. Kernel Hackers Manual October 2009 dev_alloc_name 9 2.6.32-rc5 dev_alloc_name allocate a name for a device int dev_alloc_name struct net_device * dev const char * name dev device name name format string Passed a format string - eg ltd it will try and find a suitable id. It scans list of devices to build up a free map, then chooses the first empty slot. The caller must hold the dev_base or rtnl lock while allocating the name and adding the device in order to avoid duplicates. Limited to bits_per_byte * page size devices (ie 32K on most platforms). Returns the number of the unit assigned or a negative errno code. Kernel Hackers Manual October 2009 netdev_features_change 9 2.6.32-rc5 netdev_features_change device changes features void netdev_features_change struct net_device * dev dev device to cause notification Called to indicate a device has changed features. Kernel Hackers Manual October 2009 netdev_state_change 9 2.6.32-rc5 netdev_state_change device changes state void netdev_state_change struct net_device * dev dev device to cause notification Called to indicate a device has changed state. This function calls the notifier chains for netdev_chain and sends a NEWLINK message to the routing socket. Kernel Hackers Manual October 2009 dev_load 9 2.6.32-rc5 dev_load load a network module void dev_load struct net * net const char * name net the applicable net namespace name name of interface If a network interface is not present and the process has suitable privileges this function loads the module. If module loading is not available in this kernel then it becomes a nop. Kernel Hackers Manual October 2009 dev_open 9 2.6.32-rc5 dev_open prepare an interface for use. int dev_open struct net_device * dev dev device to open Takes a device from down to up state. The device's private open function is invoked and then the multicast lists are loaded. Finally the device is moved into the up state and a NETDEV_UP message is sent to the netdev notifier chain. Calling this function on an active interface is a nop. On a failure a negative errno code is returned. Kernel Hackers Manual October 2009 dev_close 9 2.6.32-rc5 dev_close shutdown an interface. int dev_close struct net_device * dev dev device to shutdown This function moves an active device into down state. A NETDEV_GOING_DOWN is sent to the netdev notifier chain. The device is then deactivated and finally a NETDEV_DOWN is sent to the notifier chain. Kernel Hackers Manual October 2009 dev_disable_lro 9 2.6.32-rc5 dev_disable_lro disable Large Receive Offload on a device void dev_disable_lro struct net_device * dev dev device Disable Large Receive Offload (LRO) on a net device. Must be called under RTNL. This is needed if received packets may be forwarded to another interface. Kernel Hackers Manual October 2009 register_netdevice_notifier 9 2.6.32-rc5 register_netdevice_notifier register a network notifier block int register_netdevice_notifier struct notifier_block * nb nb notifier Register a notifier to be called when network device events occur. The notifier passed is linked into the kernel structures and must not be reused until it has been unregistered. A negative errno code is returned on a failure. When registered all registration and up events are replayed to the new notifier to allow device to have a race free view of the network device list. Kernel Hackers Manual October 2009 unregister_netdevice_notifier 9 2.6.32-rc5 unregister_netdevice_notifier unregister a network notifier block int unregister_netdevice_notifier struct notifier_block * nb nb notifier Unregister a notifier previously registered by register_netdevice_notifier. The notifier is unlinked into the kernel structures and may then be reused. A negative errno code is returned on a failure. Kernel Hackers Manual October 2009 netif_device_detach 9 2.6.32-rc5 netif_device_detach mark device as removed void netif_device_detach struct net_device * dev dev network device Mark device as removed from system and therefore no longer available. Kernel Hackers Manual October 2009 netif_device_attach 9 2.6.32-rc5 netif_device_attach mark device as attached void netif_device_attach struct net_device * dev dev network device Mark device as attached from system and restart if needed. Kernel Hackers Manual October 2009 skb_gso_segment 9 2.6.32-rc5 skb_gso_segment Perform segmentation on skb. struct sk_buff * skb_gso_segment struct sk_buff * skb int features skb buffer to segment features features for the output path (see dev->features) This function segments the given skb and returns a list of segments. It may return NULL if the skb requires no segmentation. This is only possible when GSO is used for verifying header integrity. Kernel Hackers Manual October 2009 dev_queue_xmit 9 2.6.32-rc5 dev_queue_xmit transmit a buffer int dev_queue_xmit struct sk_buff * skb skb buffer to transmit Queue a buffer for transmission to a network device. The caller must have set the device and priority and built the buffer before calling this function. The function can be called from an interrupt. A negative errno code is returned on a failure. A success does not guarantee the frame will be transmitted as it may be dropped due to congestion or traffic shaping. ----------------------------------------------------------------------------------- I notice this method can also return errors from the queue disciplines, including NET_XMIT_DROP, which is a positive value. So, errors can also be positive. Regardless of the return value, the skb is consumed, so it is currently difficult to retry a send to this method. (You can bump the ref count before sending to hold a reference for retry if you are careful.) When calling this method, interrupts MUST be enabled. This is because the BH enable code must have IRQs enabled so that it will not deadlock. --BLG Kernel Hackers Manual October 2009 netif_rx 9 2.6.32-rc5 netif_rx post buffer to the network code int netif_rx struct sk_buff * skb skb buffer to post This function receives a packet from a device driver and queues it for the upper (protocol) levels to process. It always succeeds. The buffer may be dropped during processing for congestion control or by the protocol layers. NET_RX_SUCCESS (no congestion) NET_RX_DROP (packet was dropped) Kernel Hackers Manual October 2009 netif_receive_skb 9 2.6.32-rc5 netif_receive_skb process receive buffer from network int netif_receive_skb struct sk_buff * skb skb buffer to process netif_receive_skb is the main receive data processing function. It always succeeds. The buffer may be dropped during processing for congestion control or by the protocol layers. This function may only be called from softirq context and interrupts should be enabled. Return values (usually ignored): no congestion packet was dropped Kernel Hackers Manual October 2009 __napi_schedule 9 2.6.32-rc5 __napi_schedule schedule for receive void __napi_schedule struct napi_struct * n n entry to schedule The entry's receive function will be scheduled to run Kernel Hackers Manual October 2009 register_gifconf 9 2.6.32-rc5 register_gifconf register a SIOCGIF handler int register_gifconf unsigned int family gifconf_func_t * gifconf family Address family gifconf Function handler Register protocol dependent address dumping routines. The handler that is passed must not be freed or reused until it has been replaced by another handler. Kernel Hackers Manual October 2009 netdev_set_master 9 2.6.32-rc5 netdev_set_master set up master/slave pair int netdev_set_master struct net_device * slave struct net_device * master slave slave device master new master device Changes the master device of the slave. Pass NULL to break the bonding. The caller must hold the RTNL semaphore. On a failure a negative errno code is returned. On success the reference counts are adjusted, RTM_NEWLINK is sent to the routing socket and the function returns zero. Kernel Hackers Manual October 2009 dev_set_promiscuity 9 2.6.32-rc5 dev_set_promiscuity update promiscuity count on a device int dev_set_promiscuity struct net_device * dev int inc dev device inc modifier Add or remove promiscuity from a device. While the count in the device remains above zero the interface remains promiscuous. Once it hits zero the device reverts back to normal filtering operation. A negative inc value is used to drop promiscuity on the device. Return 0 if successful or a negative errno code on error. Kernel Hackers Manual October 2009 dev_set_allmulti 9 2.6.32-rc5 dev_set_allmulti update allmulti count on a device int dev_set_allmulti struct net_device * dev int inc dev device inc modifier Add or remove reception of all multicast frames to a device. While the count in the device remains above zero the interface remains listening to all interfaces. Once it hits zero the device reverts back to normal filtering operation. A negative inc value is used to drop the counter when releasing a resource needing all multicasts. Return 0 if successful or a negative errno code on error. Kernel Hackers Manual October 2009 dev_addr_add 9 2.6.32-rc5 dev_addr_add Add a device address int dev_addr_add struct net_device * dev unsigned char * addr unsigned char addr_type dev device addr address to add addr_type address type Add a device address to the device or increase the reference count if it already exists. The caller must hold the rtnl_mutex. Kernel Hackers Manual October 2009 dev_addr_del 9 2.6.32-rc5 dev_addr_del Release a device address. int dev_addr_del struct net_device * dev unsigned char * addr unsigned char addr_type dev device addr address to delete addr_type address type Release reference to a device address and remove it from the device if the reference count drops to zero. The caller must hold the rtnl_mutex. Kernel Hackers Manual October 2009 dev_addr_add_multiple 9 2.6.32-rc5 dev_addr_add_multiple Add device addresses from another device int dev_addr_add_multiple struct net_device * to_dev struct net_device * from_dev unsigned char addr_type to_dev device to which addresses will be added from_dev device from which addresses will be added addr_type address type - 0 means type will be used from from_dev Add device addresses of the one device to another. * The caller must hold the rtnl_mutex. Kernel Hackers Manual October 2009 dev_addr_del_multiple 9 2.6.32-rc5 dev_addr_del_multiple Delete device addresses by another device int dev_addr_del_multiple struct net_device * to_dev struct net_device * from_dev unsigned char addr_type to_dev device where the addresses will be deleted from_dev device by which addresses the addresses will be deleted addr_type address type - 0 means type will used from from_dev Deletes addresses in to device by the list of addresses in from device. The caller must hold the rtnl_mutex. Kernel Hackers Manual October 2009 dev_unicast_delete 9 2.6.32-rc5 dev_unicast_delete Release secondary unicast address. int dev_unicast_delete struct net_device * dev void * addr dev device addr address to delete Release reference to a secondary unicast address and remove it from the device if the reference count drops to zero. The caller must hold the rtnl_mutex. Kernel Hackers Manual October 2009 dev_unicast_add 9 2.6.32-rc5 dev_unicast_add add a secondary unicast address int dev_unicast_add struct net_device * dev void * addr dev device addr address to add Add a secondary unicast address to the device or increase the reference count if it already exists. The caller must hold the rtnl_mutex. Kernel Hackers Manual October 2009 dev_unicast_sync 9 2.6.32-rc5 dev_unicast_sync Synchronize device's unicast list to another device int dev_unicast_sync struct net_device * to struct net_device * from to destination device from source device Add newly added addresses to the destination device and release addresses that have no users left. The source device must be locked by netif_tx_lock_bh. This function is intended to be called from the dev->set_rx_mode function of layered software devices. Kernel Hackers Manual October 2009 dev_unicast_unsync 9 2.6.32-rc5 dev_unicast_unsync Remove synchronized addresses from the destination device void dev_unicast_unsync struct net_device * to struct net_device * from to destination device from source device Remove all addresses that were added to the destination device by dev_unicast_sync. This function is intended to be called from the dev->stop function of layered software devices. Kernel Hackers Manual October 2009 dev_get_flags 9 2.6.32-rc5 dev_get_flags get flags reported to userspace unsigned dev_get_flags const struct net_device * dev dev device Get the combination of flag bits exported through APIs to userspace. Kernel Hackers Manual October 2009 dev_change_flags 9 2.6.32-rc5 dev_change_flags change device settings int dev_change_flags struct net_device * dev unsigned flags dev device flags device state flags Change settings on device based state flags. The flags are in the userspace exported format. Kernel Hackers Manual October 2009 dev_set_mtu 9 2.6.32-rc5 dev_set_mtu Change maximum transfer unit int dev_set_mtu struct net_device * dev int new_mtu dev device new_mtu new transfer unit Change the maximum transfer size of the network device. Kernel Hackers Manual October 2009 dev_set_mac_address 9 2.6.32-rc5 dev_set_mac_address Change Media Access Control Address int dev_set_mac_address struct net_device * dev struct sockaddr * sa dev device sa new address Change the hardware (MAC) address of the device Kernel Hackers Manual October 2009 register_netdevice 9 2.6.32-rc5 register_netdevice register a network device int register_netdevice struct net_device * dev dev device to register Take a completed network device structure and add it to the kernel interfaces. A NETDEV_REGISTER message is sent to the netdev notifier chain. 0 is returned on success. A negative errno code is returned on a failure to set up the device, or if the name is a duplicate. Callers must hold the rtnl semaphore. You may want register_netdev instead of this. The locking appears insufficient to guarantee two parallel registers will not get the same name. Kernel Hackers Manual October 2009 init_dummy_netdev 9 2.6.32-rc5 init_dummy_netdev init a dummy network device for NAPI int init_dummy_netdev struct net_device * dev dev device to init This takes a network device structure and initialize the minimum amount of fields so it can be used to schedule NAPI polls without registering a full blown interface. This is to be used by drivers that need to tie several hardware interfaces to a single NAPI poll scheduler due to HW limitations. Kernel Hackers Manual October 2009 register_netdev 9 2.6.32-rc5 register_netdev register a network device int register_netdev struct net_device * dev dev device to register Take a completed network device structure and add it to the kernel interfaces. A NETDEV_REGISTER message is sent to the netdev notifier chain. 0 is returned on success. A negative errno code is returned on a failure to set up the device, or if the name is a duplicate. This is a wrapper around register_netdevice that takes the rtnl semaphore and expands the device name if you passed a format string to alloc_netdev. Kernel Hackers Manual October 2009 dev_get_stats 9 2.6.32-rc5 dev_get_stats get network device statistics const struct net_device_stats * dev_get_stats struct net_device * dev dev device to get statistics from Get network statistics from device. The device driver may provide its own method by setting dev->netdev_ops->get_stats; otherwise the internal statistics structure is used. Kernel Hackers Manual October 2009 alloc_netdev_mq 9 2.6.32-rc5 alloc_netdev_mq allocate network device struct net_device * alloc_netdev_mq int sizeof_priv const char * name void (*setup) struct net_device * unsigned int queue_count sizeof_priv size of private data to allocate space for name device name format string setup callback to initialize device queue_count the number of subqueues to allocate Allocates a struct net_device with private data area for driver use and performs basic initialization. Also allocates subquue structs for each queue on the device at the end of the netdevice. Kernel Hackers Manual October 2009 free_netdev 9 2.6.32-rc5 free_netdev free network device void free_netdev struct net_device * dev dev device This function does the last stage of destroying an allocated device interface. The reference to the device object is released. If this is the last reference then it will be freed. Kernel Hackers Manual October 2009 synchronize_net 9 2.6.32-rc5 synchronize_net Synchronize with packet receive processing void synchronize_net void void no arguments Wait for packets currently being received to be done. Does not block later packets from starting. Kernel Hackers Manual October 2009 unregister_netdevice 9 2.6.32-rc5 unregister_netdevice remove device from the kernel void unregister_netdevice struct net_device * dev dev device This function shuts down a device interface and removes it from the kernel tables. Callers must hold the rtnl semaphore. You may want unregister_netdev instead of this. Kernel Hackers Manual October 2009 unregister_netdev 9 2.6.32-rc5 unregister_netdev remove device from the kernel void unregister_netdev struct net_device * dev dev device This function shuts down a device interface and removes it from the kernel tables. This is just a wrapper for unregister_netdevice that takes the rtnl semaphore. In general you want to use this and not unregister_netdevice. Kernel Hackers Manual October 2009 dev_change_net_namespace 9 2.6.32-rc5 dev_change_net_namespace move device to different nethost namespace int dev_change_net_namespace struct net_device * dev struct net * net const char * pat dev device net network namespace pat If not NULL name pattern to try if the current device name is already taken in the destination network namespace. This function shuts down a device interface and moves it to a new network namespace. On success 0 is returned, on a failure a netagive errno code is returned. Callers must hold the rtnl semaphore. Kernel Hackers Manual October 2009 netdev_increment_features 9 2.6.32-rc5 netdev_increment_features increment feature set by one unsigned long netdev_increment_features unsigned long all unsigned long one unsigned long mask all current feature set one new feature set mask mask feature set Computes a new feature set after adding a device with feature set one to the master device with current feature set all. Will not enable anything that is off in mask. Returns the new feature set. Kernel Hackers Manual October 2009 eth_header 9 2.6.32-rc5 eth_header create the Ethernet header int eth_header struct sk_buff * skb struct net_device * dev unsigned short type const void * daddr const void * saddr unsigned len skb buffer to alter dev source device type Ethernet type field daddr destination address (NULL leave destination address) saddr source address (NULL use device source address) len packet length (<= skb->len) Set the protocol type. For a packet of type ETH_P_802_3 we put the length in here instead. It is up to the 802.2 layer to carry protocol information. Kernel Hackers Manual October 2009 eth_rebuild_header 9 2.6.32-rc5 eth_rebuild_header rebuild the Ethernet MAC header. int eth_rebuild_header struct sk_buff * skb skb socket buffer to update This is called after an ARP or IPV6 ndisc it's resolution on this sk_buff. We now let protocol (ARP) fill in the other fields. This routine CANNOT use cached dst->neigh! Really, it is used only when dst->neigh is wrong. Kernel Hackers Manual October 2009 eth_type_trans 9 2.6.32-rc5 eth_type_trans determine the packet's protocol ID. __be16 eth_type_trans struct sk_buff * skb struct net_device * dev skb received socket data dev receiving network device The rule here is that we assume 802.3 if the type field is short enough to be a length. This is normal practice and works for any 'now in use' protocol. Kernel Hackers Manual October 2009 eth_header_parse 9 2.6.32-rc5 eth_header_parse extract hardware address from packet int eth_header_parse const struct sk_buff * skb unsigned char * haddr skb packet to extract header from haddr destination buffer Kernel Hackers Manual October 2009 eth_header_cache 9 2.6.32-rc5 eth_header_cache fill cache entry from neighbour int eth_header_cache const struct neighbour * neigh struct hh_cache * hh neigh source neighbour hh destination cache entry Create an Ethernet header template from the neighbour. Kernel Hackers Manual October 2009 eth_header_cache_update 9 2.6.32-rc5 eth_header_cache_update update cache entry void eth_header_cache_update struct hh_cache * hh const struct net_device * dev const unsigned char * haddr hh destination cache entry dev network device haddr new hardware address Called by Address Resolution module to notify changes in address. Kernel Hackers Manual October 2009 eth_mac_addr 9 2.6.32-rc5 eth_mac_addr set new Ethernet hardware address int eth_mac_addr struct net_device * dev void * p dev network device p socket address Change hardware address of device. This doesn't change hardware matching, so needs to be overridden for most real devices. Kernel Hackers Manual October 2009 eth_change_mtu 9 2.6.32-rc5 eth_change_mtu set new MTU size int eth_change_mtu struct net_device * dev int new_mtu dev network device new_mtu new Maximum Transfer Unit Allow changing MTU size. Needs to be overridden for devices supporting jumbo frames. Kernel Hackers Manual October 2009 ether_setup 9 2.6.32-rc5 ether_setup setup Ethernet network device void ether_setup struct net_device * dev dev network device Fill in the fields of the device structure with Ethernet-generic values. Kernel Hackers Manual October 2009 alloc_etherdev_mq 9 2.6.32-rc5 alloc_etherdev_mq Allocates and sets up an Ethernet device struct net_device * alloc_etherdev_mq int sizeof_priv unsigned int queue_count sizeof_priv Size of additional driver-private structure to be allocated for this Ethernet device queue_count The number of queues this device has. Fill in the fields of the device structure with Ethernet-generic values. Basically does everything except registering the device. Constructs a new net device, complete with a private data area of size (sizeof_priv). A 32-byte (not bit) alignment is enforced for this private data area. Kernel Hackers Manual October 2009 netif_carrier_on 9 2.6.32-rc5 netif_carrier_on set carrier void netif_carrier_on struct net_device * dev dev network device Device has detected that carrier. Kernel Hackers Manual October 2009 netif_carrier_off 9 2.6.32-rc5 netif_carrier_off clear carrier void netif_carrier_off struct net_device * dev dev network device Device has detected loss of carrier. Kernel Hackers Manual October 2009 is_zero_ether_addr 9 2.6.32-rc5 is_zero_ether_addr Determine if give Ethernet address is all zeros. int is_zero_ether_addr const u8 * addr addr Pointer to a six-byte array containing the Ethernet address Return true if the address is all zeroes. Kernel Hackers Manual October 2009 is_multicast_ether_addr 9 2.6.32-rc5 is_multicast_ether_addr Determine if the Ethernet address is a multicast. int is_multicast_ether_addr const u8 * addr addr Pointer to a six-byte array containing the Ethernet address Return true if the address is a multicast address. By definition the broadcast address is also a multicast address. Kernel Hackers Manual October 2009 is_local_ether_addr 9 2.6.32-rc5 is_local_ether_addr Determine if the Ethernet address is locally-assigned one (IEEE 802). int is_local_ether_addr const u8 * addr addr Pointer to a six-byte array containing the Ethernet address Return true if the address is a local address. Kernel Hackers Manual October 2009 is_broadcast_ether_addr 9 2.6.32-rc5 is_broadcast_ether_addr Determine if the Ethernet address is broadcast int is_broadcast_ether_addr const u8 * addr addr Pointer to a six-byte array containing the Ethernet address Return true if the address is the broadcast address. Kernel Hackers Manual October 2009 is_valid_ether_addr 9 2.6.32-rc5 is_valid_ether_addr Determine if the given Ethernet address is valid int is_valid_ether_addr const u8 * addr addr Pointer to a six-byte array containing the Ethernet address Check that the Ethernet address (MAC) is not 00:00:00:00:00:00, is not a multicast address, and is not FF:FF:FF:FF:FF:FF. Return true if the address is valid. Kernel Hackers Manual October 2009 random_ether_addr 9 2.6.32-rc5 random_ether_addr Generate software assigned random Ethernet address void random_ether_addr u8 * addr addr Pointer to a six-byte array containing the Ethernet address Generate a random Ethernet address (MAC) that is not multicast and has the local assigned bit set. Kernel Hackers Manual October 2009 compare_ether_addr 9 2.6.32-rc5 compare_ether_addr Compare two Ethernet addresses unsigned compare_ether_addr const u8 * addr1 const u8 * addr2 addr1 Pointer to a six-byte array containing the Ethernet address addr2 Pointer other six-byte array containing the Ethernet address Compare two ethernet addresses, returns 0 if equal Kernel Hackers Manual October 2009 compare_ether_addr_64bits 9 2.6.32-rc5 compare_ether_addr_64bits Compare two Ethernet addresses unsigned compare_ether_addr_64bits const u8 addr1[6+2] const u8 addr2[6+2] addr1[6+2] Pointer to an array of 8 bytes addr2[6+2] Pointer to an other array of 8 bytes Compare two ethernet addresses, returns 0 if equal. Same result than memcmp(addr1, addr2, ETH_ALEN) but without conditional branches, and possibly long word memory accesses on CPU allowing cheap unaligned memory reads. arrays = { byte1, byte2, byte3, byte4, byte6, byte7, pad1, pad2} Please note that alignment of addr1 & addr2 is only guaranted to be 16 bits. Kernel Hackers Manual October 2009 is_etherdev_addr 9 2.6.32-rc5 is_etherdev_addr Tell if given Ethernet address belongs to the device. bool is_etherdev_addr const struct net_device * dev const u8 addr[6 + 2] dev Pointer to a device structure addr[6 + 2] Pointer to a six-byte array containing the Ethernet address Compare passed address with all addresses of the device. Return true if the address if one of the device addresses. Note that this function calls compare_ether_addr_64bits so take care of the right padding. Kernel Hackers Manual October 2009 compare_ether_header 9 2.6.32-rc5 compare_ether_header Compare two Ethernet headers int compare_ether_header const void * a const void * b a Pointer to Ethernet header b Pointer to Ethernet header Compare two ethernet headers, returns 0 if equal. This assumes that the network header (i.e., IP header) is 4-byte aligned OR the platform can handle unaligned access. This is the case for all packets coming into netif_receive_skb or similar entry points. Kernel Hackers Manual October 2009 napi_schedule_prep 9 2.6.32-rc5 napi_schedule_prep check if napi can be scheduled int napi_schedule_prep struct napi_struct * n n napi context Test if NAPI routine is already running, and if not mark it as running. This is used as a condition variable insure only one NAPI poll instance runs. We also make sure there is no pending NAPI disable. Kernel Hackers Manual October 2009 napi_schedule 9 2.6.32-rc5 napi_schedule schedule NAPI poll void napi_schedule struct napi_struct * n n napi context Schedule NAPI poll routine to be called if it is not already running. Kernel Hackers Manual October 2009 napi_disable 9 2.6.32-rc5 napi_disable prevent NAPI from scheduling void napi_disable struct napi_struct * n n napi context Stop NAPI from being scheduled on this context. Waits till any outstanding processing completes. Kernel Hackers Manual October 2009 napi_enable 9 2.6.32-rc5 napi_enable enable NAPI scheduling void napi_enable struct napi_struct * n n napi context Resume NAPI from being scheduled on this context. Must be paired with napi_disable. Kernel Hackers Manual October 2009 napi_synchronize 9 2.6.32-rc5 napi_synchronize wait until NAPI is not running void napi_synchronize const struct napi_struct * n n napi context Wait until NAPI is done being scheduled on this context. Waits till any outstanding processing completes but does not disable future activations. Kernel Hackers Manual October 2009 netdev_priv 9 2.6.32-rc5 netdev_priv access network device private data void * netdev_priv const struct net_device * dev dev network device Get network device private data Kernel Hackers Manual October 2009 netif_start_queue 9 2.6.32-rc5 netif_start_queue allow transmit void netif_start_queue struct net_device * dev dev network device Allow upper layers to call the device hard_start_xmit routine. Kernel Hackers Manual October 2009 netif_wake_queue 9 2.6.32-rc5 netif_wake_queue restart transmit void netif_wake_queue struct net_device * dev dev network device Allow upper layers to call the device hard_start_xmit routine. Used for flow control when transmit resources are available. Kernel Hackers Manual October 2009 netif_stop_queue 9 2.6.32-rc5 netif_stop_queue stop transmitted packets void netif_stop_queue struct net_device * dev dev network device Stop upper layers calling the device hard_start_xmit routine. Used for flow control when transmit resources are unavailable. Kernel Hackers Manual October 2009 netif_queue_stopped 9 2.6.32-rc5 netif_queue_stopped test if transmit queue is flowblocked int netif_queue_stopped const struct net_device * dev dev network device Test if transmit queue on device is currently unable to send. Kernel Hackers Manual October 2009 netif_running 9 2.6.32-rc5 netif_running test if up int netif_running const struct net_device * dev dev network device Test if the device has been brought up. Kernel Hackers Manual October 2009 netif_start_subqueue 9 2.6.32-rc5 netif_start_subqueue allow sending packets on subqueue void netif_start_subqueue struct net_device * dev u16 queue_index dev network device queue_index sub queue index Start individual transmit queue of a device with multiple transmit queues. Kernel Hackers Manual October 2009 netif_stop_subqueue 9 2.6.32-rc5 netif_stop_subqueue stop sending packets on subqueue void netif_stop_subqueue struct net_device * dev u16 queue_index dev network device queue_index sub queue index Stop individual transmit queue of a device with multiple transmit queues. Kernel Hackers Manual October 2009 __netif_subqueue_stopped 9 2.6.32-rc5 __netif_subqueue_stopped test status of subqueue int __netif_subqueue_stopped const struct net_device * dev u16 queue_index dev network device queue_index sub queue index Check individual transmit queue of a device with multiple transmit queues. Kernel Hackers Manual October 2009 netif_wake_subqueue 9 2.6.32-rc5 netif_wake_subqueue allow sending packets on subqueue void netif_wake_subqueue struct net_device * dev u16 queue_index dev network device queue_index sub queue index Resume individual transmit queue of a device with multiple transmit queues. Kernel Hackers Manual October 2009 netif_is_multiqueue 9 2.6.32-rc5 netif_is_multiqueue test if device has multiple transmit queues int netif_is_multiqueue const struct net_device * dev dev network device Check if device has multiple transmit queues Kernel Hackers Manual October 2009 dev_put 9 2.6.32-rc5 dev_put release reference to device void dev_put struct net_device * dev dev network device Release reference to device to allow it to be freed. Kernel Hackers Manual October 2009 dev_hold 9 2.6.32-rc5 dev_hold get reference to device void dev_hold struct net_device * dev dev network device Hold reference to device to keep it from being freed. Kernel Hackers Manual October 2009 netif_carrier_ok 9 2.6.32-rc5 netif_carrier_ok test if carrier present int netif_carrier_ok const struct net_device * dev dev network device Check if carrier is present on device Kernel Hackers Manual October 2009 netif_dormant_on 9 2.6.32-rc5 netif_dormant_on mark device as dormant. void netif_dormant_on struct net_device * dev dev network device Mark device as dormant (as per RFC2863). The dormant state indicates that the relevant interface is not actually in a condition to pass packets (i.e., it is not 'up') but is in a pending state, waiting for some external event. For on- demand interfaces, this new state identifies the situation where the interface is waiting for events to place it in the up state. Kernel Hackers Manual October 2009 netif_dormant_off 9 2.6.32-rc5 netif_dormant_off set device as not dormant. void netif_dormant_off struct net_device * dev dev network device Device is not in dormant state. Kernel Hackers Manual October 2009 netif_dormant 9 2.6.32-rc5 netif_dormant test if carrier present int netif_dormant const struct net_device * dev dev network device Check if carrier is present on device Kernel Hackers Manual October 2009 netif_oper_up 9 2.6.32-rc5 netif_oper_up test if device is operational int netif_oper_up const struct net_device * dev dev network device Check if carrier is operational Kernel Hackers Manual October 2009 netif_device_present 9 2.6.32-rc5 netif_device_present is device available or removed int netif_device_present struct net_device * dev dev network device Check if device has not been removed from system. Kernel Hackers Manual October 2009 netif_tx_lock 9 2.6.32-rc5 netif_tx_lock grab network device transmit lock void netif_tx_lock struct net_device * dev dev network device Get network device transmit lock Kernel Hackers Manual October 2009 phy_print_status 9 2.6.32-rc5 phy_print_status Convenience function to print out the current phy status void phy_print_status struct phy_device * phydev phydev the phy_device struct Kernel Hackers Manual October 2009 phy_sanitize_settings 9 2.6.32-rc5 phy_sanitize_settings make sure the PHY is set to supported speed and duplex void phy_sanitize_settings struct phy_device * phydev phydev the target phy_device struct Make sure the PHY is set to supported speeds and duplexes. Drop down by one in this order: 1000/FULL, 1000/HALF, 100/FULL, 100/HALF, 10/FULL, 10/HALF. Kernel Hackers Manual October 2009 phy_ethtool_sset 9 2.6.32-rc5 phy_ethtool_sset generic ethtool sset function, handles all the details int phy_ethtool_sset struct phy_device * phydev struct ethtool_cmd * cmd phydev target phy_device struct cmd ethtool_cmd - We don't set port or transceiver, so we don't care what they were set to. - phy_start_aneg will make sure forced settings are sane, and choose the next best ones from the ones selected, so we don't care if ethtool tries to give us bad values. Kernel Hackers Manual October 2009 phy_mii_ioctl 9 2.6.32-rc5 phy_mii_ioctl generic PHY MII ioctl interface int phy_mii_ioctl struct phy_device * phydev struct mii_ioctl_data * mii_data int cmd phydev the phy_device struct mii_data MII ioctl data cmd ioctl cmd to execute Note that this function is currently incompatible with the PHYCONTROL layer. It changes registers without regard to current state. Use at own risk. Kernel Hackers Manual October 2009 phy_start_aneg 9 2.6.32-rc5 phy_start_aneg start auto-negotiation for this PHY device int phy_start_aneg struct phy_device * phydev phydev the phy_device struct Sanitizes the settings (if we're not autonegotiating them), and then calls the driver's config_aneg function. If the PHYCONTROL Layer is operating, we change the state to reflect the beginning of Auto-negotiation or forcing. Kernel Hackers Manual October 2009 phy_enable_interrupts 9 2.6.32-rc5 phy_enable_interrupts Enable the interrupts from the PHY side int phy_enable_interrupts struct phy_device * phydev phydev target phy_device struct Kernel Hackers Manual October 2009 phy_disable_interrupts 9 2.6.32-rc5 phy_disable_interrupts Disable the PHY interrupts from the PHY side int phy_disable_interrupts struct phy_device * phydev phydev target phy_device struct Kernel Hackers Manual October 2009 phy_start_interrupts 9 2.6.32-rc5 phy_start_interrupts request and enable interrupts for a PHY device int phy_start_interrupts struct phy_device * phydev phydev target phy_device struct Request the interrupt for the given PHY. If this fails, then we set irq to PHY_POLL. Otherwise, we enable the interrupts in the PHY. This should only be called with a valid IRQ number. Returns 0 on success or < 0 on error. Kernel Hackers Manual October 2009 phy_stop_interrupts 9 2.6.32-rc5 phy_stop_interrupts disable interrupts from a PHY device int phy_stop_interrupts struct phy_device * phydev phydev target phy_device struct Kernel Hackers Manual October 2009 phy_stop 9 2.6.32-rc5 phy_stop Bring down the PHY link, and stop checking the status void phy_stop struct phy_device * phydev phydev target phy_device struct Kernel Hackers Manual October 2009 phy_start 9 2.6.32-rc5 phy_start start or restart a PHY device void phy_start struct phy_device * phydev phydev target phy_device struct Indicates the attached device's readiness to handle PHY-related work. Used during startup to start the PHY, and after a call to phy_stop to resume operation. Also used to indicate the MDIO bus has cleared an error condition. Kernel Hackers Manual October 2009 phy_clear_interrupt 9 2.6.32-rc5 phy_clear_interrupt Ack the phy device's interrupt int phy_clear_interrupt struct phy_device * phydev phydev the phy_device struct If the phydev driver has an ack_interrupt function, call it to ack and clear the phy device's interrupt. Returns 0 on success on < 0 on error. Kernel Hackers Manual October 2009 phy_config_interrupt 9 2.6.32-rc5 phy_config_interrupt configure the PHY device for the requested interrupts int phy_config_interrupt struct phy_device * phydev u32 interrupts phydev the phy_device struct interrupts interrupt flags to configure for this phydev Returns 0 on success on < 0 on error. Kernel Hackers Manual October 2009 phy_aneg_done 9 2.6.32-rc5 phy_aneg_done return auto-negotiation status int phy_aneg_done struct phy_device * phydev phydev target phy_device struct Reads the status register and returns 0 either if auto-negotiation is incomplete, or if there was an error. Returns BMSR_ANEGCOMPLETE if auto-negotiation is done. Kernel Hackers Manual October 2009 phy_find_setting 9 2.6.32-rc5 phy_find_setting find a PHY settings array entry that matches speed & duplex int phy_find_setting int speed int duplex speed speed to match duplex duplex to match Searches the settings array for the setting which matches the desired speed and duplex, and returns the index of that setting. Returns the index of the last setting if none of the others match. Kernel Hackers Manual October 2009 phy_find_valid 9 2.6.32-rc5 phy_find_valid find a PHY setting that matches the requested features mask int phy_find_valid int idx u32 features idx The first index in settings[] to search features A mask of the valid settings Returns the index of the first valid setting less than or equal to the one pointed to by idx, as determined by the mask in features. Returns the index of the last setting if nothing else matches. Kernel Hackers Manual October 2009 phy_start_machine 9 2.6.32-rc5 phy_start_machine start PHY state machine tracking void phy_start_machine struct phy_device * phydev void (*handler) struct net_device * phydev the phy_device struct handler callback function for state change notifications The PHY infrastructure can run a state machine which tracks whether the PHY is starting up, negotiating, etc. This function starts the timer which tracks the state of the PHY. If you want to be notified when the state changes, pass in the callback handler, otherwise, pass NULL. If you want to maintain your own state machine, do not call this function. Kernel Hackers Manual October 2009 phy_stop_machine 9 2.6.32-rc5 phy_stop_machine stop the PHY state machine tracking void phy_stop_machine struct phy_device * phydev phydev target phy_device struct Stops the state machine timer, sets the state to UP (unless it wasn't up yet). This function must be called BEFORE phy_detach. Kernel Hackers Manual October 2009 phy_force_reduction 9 2.6.32-rc5 phy_force_reduction reduce PHY speed/duplex settings by one step void phy_force_reduction struct phy_device * phydev phydev target phy_device struct Reduces the speed/duplex settings by one notch, in this order-- 1000/FULL, 1000/HALF, 100/FULL, 100/HALF, 10/FULL, 10/HALF. The function bottoms out at 10/HALF. Kernel Hackers Manual October 2009 phy_error 9 2.6.32-rc5 phy_error enter HALTED state for this PHY device void phy_error struct phy_device * phydev phydev target phy_device struct Moves the PHY to the HALTED state in response to a read or write error, and tells the controller the link is down. Must not be called from interrupt context, or while the phydev->lock is held. Kernel Hackers Manual October 2009 phy_interrupt 9 2.6.32-rc5 phy_interrupt PHY interrupt handler irqreturn_t phy_interrupt int irq void * phy_dat irq interrupt line phy_dat phy_device pointer When a PHY interrupt occurs, the handler disables interrupts, and schedules a work task to clear the interrupt. Kernel Hackers Manual October 2009 phy_change 9 2.6.32-rc5 phy_change Scheduled by the phy_interrupt/timer to handle PHY changes void phy_change struct work_struct * work work work_struct that describes the work to be done Kernel Hackers Manual October 2009 phy_state_machine 9 2.6.32-rc5 phy_state_machine Handle the state machine void phy_state_machine struct work_struct * work work work_struct that describes the work to be done Kernel Hackers Manual October 2009 get_phy_id 9 2.6.32-rc5 get_phy_id reads the specified addr for its ID. int get_phy_id struct mii_bus * bus int addr u32 * phy_id bus the target MII bus addr PHY address on the MII bus phy_id where to store the ID retrieved. Reads the ID registers of the PHY at addr on the bus, stores it in phy_id and returns zero on success. Kernel Hackers Manual October 2009 get_phy_device 9 2.6.32-rc5 get_phy_device reads the specified PHY device and returns its phy_device struct struct phy_device * get_phy_device struct mii_bus * bus int addr bus the target MII bus addr PHY address on the MII bus Reads the ID registers of the PHY at addr on the bus, then allocates and returns the phy_device to represent it. Kernel Hackers Manual October 2009 phy_device_register 9 2.6.32-rc5 phy_device_register Register the phy device on the MDIO bus int phy_device_register struct phy_device * phydev phydev phy_device structure to be added to the MDIO bus Kernel Hackers Manual October 2009 phy_connect_direct 9 2.6.32-rc5 phy_connect_direct connect an ethernet device to a specific phy_device int phy_connect_direct struct net_device * dev struct phy_device * phydev void (*handler) struct net_device * u32 flags phy_interface_t interface dev the network device to connect phydev the pointer to the phy device handler callback function for state change notifications flags PHY device's dev_flags interface PHY device's interface Kernel Hackers Manual October 2009 phy_connect 9 2.6.32-rc5 phy_connect connect an ethernet device to a PHY device struct phy_device * phy_connect struct net_device * dev const char * bus_id void (*handler) struct net_device * u32 flags phy_interface_t interface dev the network device to connect bus_id the id string of the PHY device to connect handler callback function for state change notifications flags PHY device's dev_flags interface PHY device's interface Convenience function for connecting ethernet devices to PHY devices. The default behavior is for the PHY infrastructure to handle everything, and only notify the connected driver when the link status changes. If you don't want, or can't use the provided functionality, you may choose to call only the subset of functions which provide the desired functionality. Kernel Hackers Manual October 2009 phy_disconnect 9 2.6.32-rc5 phy_disconnect disable interrupts, stop state machine, and detach a PHY device void phy_disconnect struct phy_device * phydev phydev target phy_device struct Kernel Hackers Manual October 2009 phy_attach_direct 9 2.6.32-rc5 phy_attach_direct attach a network device to a given PHY device pointer int phy_attach_direct struct net_device * dev struct phy_device * phydev u32 flags phy_interface_t interface dev network device to attach phydev Pointer to phy_device to attach flags PHY device's dev_flags interface PHY device's interface Called by drivers to attach to a particular PHY device. The phy_device is found, and properly hooked up to the phy_driver. If no driver is attached, then the genphy_driver is used. The phy_device is given a ptr to the attaching device, and given a callback for link status change. The phy_device is returned to the attaching driver. Kernel Hackers Manual October 2009 phy_attach 9 2.6.32-rc5 phy_attach attach a network device to a particular PHY device struct phy_device * phy_attach struct net_device * dev const char * bus_id u32 flags phy_interface_t interface dev network device to attach bus_id Bus ID of PHY device to attach flags PHY device's dev_flags interface PHY device's interface Same as phy_attach_direct except that a PHY bus_id string is passed instead of a pointer to a struct phy_device. Kernel Hackers Manual October 2009 phy_detach 9 2.6.32-rc5 phy_detach detach a PHY device from its network device void phy_detach struct phy_device * phydev phydev target phy_device struct Kernel Hackers Manual October 2009 genphy_config_advert 9 2.6.32-rc5 genphy_config_advert sanitize and advertise auto-negotation parameters int genphy_config_advert struct phy_device * phydev phydev target phy_device struct Writes MII_ADVERTISE with the appropriate values, after sanitizing the values to make sure we only advertise what is supported. Returns < 0 on error, 0 if the PHY's advertisement hasn't changed, and > 0 if it has changed. Kernel Hackers Manual October 2009 genphy_restart_aneg 9 2.6.32-rc5 genphy_restart_aneg Enable and Restart Autonegotiation int genphy_restart_aneg struct phy_device * phydev phydev target phy_device struct Kernel Hackers Manual October 2009 genphy_config_aneg 9 2.6.32-rc5 genphy_config_aneg restart auto-negotiation or write BMCR int genphy_config_aneg struct phy_device * phydev phydev target phy_device struct If auto-negotiation is enabled, we configure the advertising, and then restart auto-negotiation. If it is not enabled, then we write the BMCR. Kernel Hackers Manual October 2009 genphy_update_link 9 2.6.32-rc5 genphy_update_link update link status in phydev int genphy_update_link struct phy_device * phydev phydev target phy_device struct Update the value in phydev->link to reflect the current link value. In order to do this, we need to read the status register twice, keeping the second value. Kernel Hackers Manual October 2009 genphy_read_status 9 2.6.32-rc5