Linux内核网络子系统代码架构详解

需积分: 9 62 浏览量更新于2024-07-26 收藏 320KB PDF 举报

" Linux 网络子系统架构与实现" Linux 网络子系统是 Linux 操作系统中负责处理网络通信的组件，它是 Linux 内核的一个重要组件。 Linux 网络子系统的实现是基于 TCP/IP 协议栈的，它定义了如何将数据从一个设备传输到另一个设备。 Linux 网络子系统的架构可以分为三个层次：链路层、网络层和传输层。链路层负责将数据从一个网络设备传输到另一个网络设备，网络层负责将数据从一个设备传输到另一个设备，而传输层负责将数据从一个进程传输到另一个进程。在 Linux 网络子系统中，链路层由网络接口卡（NIC）和驱动程序组成。网络接口卡是指连接到网络的硬件设备，而驱动程序是指负责控制网络接口卡的软件。 Linux 内核提供了多种网络接口卡驱动程序，例如 eth0、eth1 等。网络层是 Linux 网络子系统的核心组件之一，它负责将数据从一个设备传输到另一个设备。网络层主要由路由表和路由协议组成。路由表是指存储在内核中的路由信息，而路由协议是指用于更新路由表的协议。传输层是 Linux 网络子系统的最高层次，它负责将数据从一个进程传输到另一个进程。传输层主要由套接字（Socket）和协议栈组成。套接字是指用于进程间通信的接口，而协议栈是指用于实现数据传输的协议。 Linux 网络子系统的实现是基于 TCP/IP 协议栈的，TCP/IP 协议栈是指一组用于实现网络通信的协议。TCP/IP 协议栈包括四个层次：链路层、网络层、传输层和应用层。链路层负责将数据从一个网络设备传输到另一个网络设备，网络层负责将数据从一个设备传输到另一个设备，传输层负责将数据从一个进程传输到另一个进程，而应用层负责将数据从一个应用程序传输到另一个应用程序。在 Linux 网络子系统中，TCP/IP 协议栈的实现是通过内核模块来实现的。内核模块是指加载到内核中的小型程序，它们负责实现特定的功能。 Linux 内核提供了多种内核模块，例如 TCP 模块、UDP 模块和 IP 模块等。 Linux 网络子系统是 Linux 操作系统中负责处理网络通信的组件，它的架构可以分为三个层次：链路层、网络层和传输层。 Linux 网络子系统的实现是基于 TCP/IP 协议栈的，它定义了如何将数据从一个设备传输到另一个设备。

char cb[48];

2 unsigned int len; /

Length of actual data

unsigned int data_len ;

4 unsigned int csum; /

Checksum

unsigned char __unused, /

Dead ﬁeld , may be reused

6 cloned, /

head may be cloned (check refcnt to be sure)

pkt_type, /

Packet class

8 ip_summed; /

Driver fed us an IP checksum

__u32 priority ; /

Packet queueing priority

10 atomic_t users ; /

User count − see datagram.c,tcp .c

unsigned short protocol ; /

Packet protocol from driver

12 unsigned short security ; /

Security level of packet

unsigned int truesize ; /

Buﬀer size

14 unsigned char

head; /

Head of buﬀer

unsigned char

data; /

Data head pointer

16 unsigned char

tail ; /

Tail pointer

unsigned char

end; /

End pointer

18 };

3.2 sock

The sock data structure keeps data about a speciﬁc TCP connection (e.g. TCP state) or virtual UDP connection.

Whenever a socket is created in user space, a sock structure is allocated. The ﬁrst ﬁelds contain the source and

destination addresses and ports of the socket pair.

struct sock {

2 /

Socket demultiplex comparisons on incoming packets .

__u32 daddr; /

Foreign IPv4 addr

4 __u32 rcv_saddr; /

Bound local IPv4 addr

__u16 dport ; /

Destination port

6 unsigned short num; /

Local port

int bound_dev_if; /

Bound device index if != 0

Among many other ﬁelds, the sock structure contains protocol-speciﬁc information. These ﬁelds contain state

information about each layer.

1 union {

struct ipv6_pinfo af_inet6 ;

3 } net_pinfo ;

5 union {

struct tcp_opt af_tcp ;

7 struct raw_opt tp_raw4;

struct raw6_opt tp_raw;

9 struct spx_opt af_spx;

} tp_pinfo ;

11 };

3.3 TCP options

One of the main components of the sock structure is the TCP option ﬁeld (tcp_opt). Both IP and UDP are stateless

protocols with a minimum need to store information about their connections. TCP, however, needs to store a large

set of variables. These variables are stored in the ﬁelds of the tcp_opt structure; only the most relevant ﬁelds are

shown below (comments are self-explanatory).

1 struct tcp_opt {

int tcp_header_len ; /

Bytes of tcp header to send

3 __u32 rcv_nxt; /

What we want to receive next

__u32 snd_nxt; /

Next sequence we send

5 __u32 snd_una; /

First byte we want an ack for

__u32 snd_sml; /

Last byte of the most recently transmitted small packet

7 __u32 rcv_tstamp; /

timestamp of last received ACK (for keepalives )

__u32 lsndtime; /

timestamp of last sent data packet (for restart window)

Delayed ACK control data

11 struct {

__u8 pending; /

ACK is pending

13 __u8 quick; /

Scheduled number of quick acks

__u8 pingpong; /

The session is interactive

15 __u8 blocked; /

Delayed ACK was blocked by socket lock

__u32 ato; /

Predicted tick of soft clock

17 unsigned long timeout; /

Currently scheduled timeout

__u32 lrcvtime ; /

timestamp of last received data packet

19 __u16 last_seg_size ; /

Size of last incoming segment

__u16 rcv_mss; /

MSS used for delayed ACK decisions

21 } ack;

23 /

Data for direct copy to user

struct {

25 struct sk_buﬀ_head prequeue;

struct task_struct

task;

27 struct iovec

iov;

int memory;

29 int len ;

} ucopy;

__u32 snd_wl1; /

Sequence for window update

33 __u32 snd_wnd; /

The window we expect to receive

__u32 max_window; /

Maximal window ever seen from peer

35 __u32 pmtu_cookie; /

Last pmtu seen by socket

__u16 mss_cache; /

Cached eﬀective mss, not including SACKS

37 __u16 mss_clamp; /

Maximal mss, negotiated at connection setup

__u16 ext_header_len; /

Network protocol overhead (IP/ IPv6 options )

39 __u8 ca_state ; /

State of fast− retransmit machine

__u8 retransmits ; /

Number of unrecovered RTO timeouts.

__u8 reordering ; /

Packet reordering metric .

43 __u8 queue_shrunk; /

Write queue has been shrunk recently .

__u8 defer_accept ; /

User waits for some data after accept ()

RTT measurement

47 __u8 backoﬀ ; /

backoﬀ

__u32 srtt ; /

smothed round trip time << 3

49 __u32 mdev; /

medium deviation

__u32 mdev_max; /

maximal mdev for the last rtt period

51 __u32 rttvar ; /

smoothed mdev_max

__u32 rtt_seq ; /

sequence number to update rttvar

53 __u32 rto ; /

retransmit timeout

__u32 packets_out; /

Packets which are "in ﬂight "

55 __u32 left_out ; /

Packets which leaved network

剩余31页未读，继续阅读

teamos

粉丝: 1
资源: 4

Linux内核网络子系统代码架构详解

A Map of the Networking Code in Linux Kernel 2.4

Linux_Kernel_Networking_Implementation_and_Theory

The_Linux_Networking_Architecture

2msl in linux kernel

Linux networking

linux tasklet

使用rust生成一段代码, 将 std::net::SocketAddr 转化为 windows_sys::Win32::Networking::WinSock::SOCKADDR_IN

linux kernel networking implementation and theory pdf

CONFIG_CFG80211_REG_DEBUG

Linux C code demo on how IPv4 accesses IPv6 addresses

最新资源