你不好奇 Linux 網絡發包過程嗎？

{"type":"doc","content":[{"type":"heading","attrs":{"align":null,"level":2},"content":[{"type":"text","text":"前言","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"這次，就圍繞一個問題來說。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"strong","attrs":{}}],"text":"Linux 系統是如何收發網絡包的？","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"horizontalrule","attrs":{}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"heading","attrs":{"align":null,"level":2},"content":[{"type":"text","text":"正文","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"heading","attrs":{"align":null,"level":3},"content":[{"type":"text","text":"網絡模型","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"爲了使得多種設備能通過網絡相互通信，和爲了解決各種不同設備在網絡互聯中的兼容性問題，國際標標準化組織制定了開放式系統互聯通信參考模型（","attrs":{}},{"type":"text","marks":[{"type":"italic","attrs":{}}],"text":"pen System Interconnection Reference Model","attrs":{}},{"type":"text","text":"），也就是 OSI 網絡模型，該模型主要有 7 層，分別是應用層、表示層、會話層、傳輸層、網絡層、數據鏈路層以及物理層。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"每一層負責的職能都不同，如下：","attrs":{}}]},{"type":"bulletedlist","content":[{"type":"listitem","content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"應用層，負責給應用程序提供統一的接口；","attrs":{}}]}],"attrs":{}},{"type":"listitem","content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"表示層，負責把數據轉換成兼容另一個系統能識別的格式；","attrs":{}}]}],"attrs":{}},{"type":"listitem","content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"會話層，負責建立、管理和終止表示層實體之間的通信會話；","attrs":{}}]}],"attrs":{}},{"type":"listitem","content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"傳輸層，負責端到端的數據傳輸；","attrs":{}}]}],"attrs":{}},{"type":"listitem","content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"網絡層，負責數據的路由、轉發、分片；","attrs":{}}]}],"attrs":{}},{"type":"listitem","content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"數據鏈路層，負責數據的封幀和差錯檢測，以及 MAC 尋址；","attrs":{}}]}],"attrs":{}},{"type":"listitem","content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"物理層，負責在物理網絡中傳輸數據幀；","attrs":{}}]}],"attrs":{}}],"attrs":{}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"由於 OSI 模型實在太複雜，提出的也只是概念理論上的分層，並沒有提供具體的實現方案。事實上，我們比較常見，也比較實用的是四層模型，即 TCP/IP 網絡模型，Linux 系統正是按照這套網絡模型來實現網絡協議棧的。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"TCP/IP 網絡模型共有 4 層，分別是應用層、傳輸層、網絡層和網絡接口層，每一層負責的職能如下：","attrs":{}}]},{"type":"bulletedlist","content":[{"type":"listitem","content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"應用層，負責向用戶提供一組應用程序，比如 HTTP、DNS、FTP 等;","attrs":{}}]}],"attrs":{}},{"type":"listitem","content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"傳輸層，負責端到端的通信，比如 TCP、UDP 等；","attrs":{}}]}],"attrs":{}},{"type":"listitem","content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"網絡層，負責網絡包的封裝、分片、路由、轉發，比如 IP、ICMP 等；","attrs":{}}]}],"attrs":{}},{"type":"listitem","content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"網絡接口層，負責網絡包在物理網絡中的傳輸，比如網絡包的封幀、 MAC 尋址、差錯檢測，以及通過網卡傳輸網絡幀等；","attrs":{}}]}],"attrs":{}}],"attrs":{}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"TCP/IP 網絡模型相比 OSI 網絡模型簡化了不少，也更加易記，它們之間的關係如下圖：","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"image","attrs":{"src":"https://static001.geekbang.org/infoq/ac/accdc75a60926191d4ad8d8343e50301.png","alt":null,"title":null,"style":[{"key":"width","value":"75%"},{"key":"bordertype","value":"none"}],"href":null,"fromPaste":true,"pastePass":true}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"不過，我們常說的七層和四層負載均衡，是用 OSI 網絡模型來描述的，七層對應的是應用層，四層對應的是傳輸層。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"horizontalrule","attrs":{}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"heading","attrs":{"align":null,"level":3},"content":[{"type":"text","text":"Linux 網絡協議棧","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"我們可以把自己的身體比作應用層中的數據，打底衣服比作傳輸層中的 TCP 頭，外套比作網絡層中 IP 頭，帽子和鞋子分別比作網絡接口層的幀頭和幀尾。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"在冬天這個季節，當我們要從家裏出去玩的時候，自然要先穿個打底衣服，再套上保暖外套，最後穿上帽子和鞋子纔出門，這個過程就好像我們把 TCP 協議通信的網絡包發出去的時候，會把應用層的數據按照網絡協議棧層層封裝和處理。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"你從下面這張圖可以看到，應用層數據在每一層的封裝格式。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"image","attrs":{"src":"https://static001.geekbang.org/infoq/4d/4daed9e15b85f1691497bf911642b229.png","alt":null,"title":null,"style":[{"key":"width","value":"75%"},{"key":"bordertype","value":"none"}],"href":null,"fromPaste":true,"pastePass":true}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"其中：","attrs":{}}]},{"type":"bulletedlist","content":[{"type":"listitem","content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"傳輸層，給應用數據前面增加了 TCP 頭；","attrs":{}}]}],"attrs":{}},{"type":"listitem","content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"網絡層，給 TCP 數據包前面增加了 IP 頭；","attrs":{}}]}],"attrs":{}},{"type":"listitem","content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"網絡接口層，給 IP 數據包前後分別增加了幀頭和幀尾；","attrs":{}}]}],"attrs":{}}],"attrs":{}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"這些新增和頭部和尾部，都有各自的作用，也都是按照特定的協議格式填充，這每一層都增加了各自的協議頭，那自然網絡包的大小就增大了，但物理鏈路並不能傳輸任意大小的數據包，所以在以太網中，規定了最大傳輸單元（MTU）是 ","attrs":{}},{"type":"codeinline","content":[{"type":"text","text":"1500","attrs":{}}],"attrs":{}},{"type":"text","text":" 字節，也就是規定了單次傳輸的最大 IP 包大小。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"當網絡包超過 MTU 的大小，就會在網絡層分片，以確保分片後的 IP 包不會超過 MTU 大小，如果 MTU 越小，需要的分包就越多，那麼網絡吞吐能力就越差，相反的，如果 MTU 越大，需要的分包就越小，那麼網絡吞吐能力就越好。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"知道了 TCP/IP 網絡模型，以及網絡包的封裝原理後，那麼 Linux 網絡協議棧的樣子，你想必猜到了大概，它其實就類似於 TCP/IP 的四層結構：","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"image","attrs":{"src":"https://static001.geekbang.org/infoq/36/361e77d512918f5cd16a340b3bd351af.png","alt":null,"title":null,"style":[{"key":"width","value":"75%"},{"key":"bordertype","value":"none"}],"href":null,"fromPaste":true,"pastePass":true}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"從上圖的的網絡協議棧，你可以看到：","attrs":{}}]},{"type":"bulletedlist","content":[{"type":"listitem","content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"應用程序需要通過系統調用，來跟 Socket 層進行數據交互；","attrs":{}}]}],"attrs":{}},{"type":"listitem","content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"Socket 層的下面就是傳輸層、網絡層和網絡接口層；","attrs":{}}]}],"attrs":{}},{"type":"listitem","content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"最下面的一層，則是網卡驅動程序和硬件網卡設備；","attrs":{}}]}],"attrs":{}}],"attrs":{}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"horizontalrule","attrs":{}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"heading","attrs":{"align":null,"level":3},"content":[{"type":"text","text":"Linux 接收網絡包的流程","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"網卡是計算機裏的一個硬件，專門負責接收和發送網絡包，當網卡接收到一個網絡包後，會通過 DMA 技術，將網絡包放入到 Ring Buffer，這個是一個環形緩衝區。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"那接收到網絡包後，應該怎麼告訴操作系統這個網絡包已經到達了呢？","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"最簡單的一種方式就是觸發中斷，也就是每當網卡收到一個網絡包，就觸發一箇中斷告訴操作系統。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"但是，這存在一個問題，在高性能網絡場景下，網絡包的數量會非常多，那麼就會觸發非常多的中斷，要知道當 CPU 收到了中斷，就會停下手裏的事情，而去處理這些網絡包，處理完畢後，纔會回去繼續其他事情，那麼頻繁地觸發中斷，則會導致 CPU 一直沒玩沒了的處理中斷，而導致其他任務可能無法繼續前進，從而影響系統的整體效率。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"所以爲了解決頻繁中斷帶來的性能開銷，Linux 內核在 2.6 版本中引入了 ","attrs":{}},{"type":"text","marks":[{"type":"strong","attrs":{}}],"text":"NAPI 機制","attrs":{}},{"type":"text","text":"，它是混合「中斷和輪詢」的方式來接收網絡包，它的核心概念就是","attrs":{}},{"type":"text","marks":[{"type":"strong","attrs":{}}],"text":"不採用中斷的方式讀取數據","attrs":{}},{"type":"text","text":"，而是首先採用中斷喚醒數據接收的服務程序，然後 ","attrs":{}},{"type":"codeinline","content":[{"type":"text","text":"poll","attrs":{}}],"attrs":{}},{"type":"text","text":" 的方法來輪詢數據。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"比如，當有網絡包到達時，網卡發起硬件中斷，於是會執行網卡硬件中斷處理函數，","attrs":{}},{"type":"text","marks":[{"type":"strong","attrs":{}}],"text":"中斷處理函數處理完需要「暫時屏蔽中斷」，然後喚醒「軟中斷」來輪詢處理數據，直到沒有新數據時才恢復中斷，這樣一次中斷處理多個網絡包","attrs":{}},{"type":"text","text":"，於是就可以降低網卡中斷帶來的性能開銷。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"那軟中斷是怎麼處理網絡包的呢？它會從 Ring Buffer 中拷貝數據到內核 struct sk_buff 緩衝區中，從而可以作爲一個網絡包交給網絡協議棧進行逐層處理。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"首先，會先進入到網絡接口層，在這一層會檢查報文的合法性，如果不合法則丟棄，合法則會找出該網絡包的上層協議的類型，比如是 IPv4，還是 IPv6，接着再去掉幀頭和幀尾，然後交給網絡層。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"到了網絡層，則取出 IP 包，判斷網絡包下一步的走向，比如是交給上層處理還是轉發出去。當確認這個網絡包要發送給本機後，就會從 IP 頭裏看看上一層協議的類型是 TCP 還是 UDP，接着去掉 IP 頭，然後交給傳輸層。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"傳輸層取出 TCP 頭或 UDP 頭，根據四元組「源 IP、源端口、目的 IP、目的端口」 作爲標識，找出對應的 Socket，並把數據拷貝到 Socket 的接收緩衝區。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"最後，應用層程序調用 Socket 接口，從內核的 Socket 接收緩衝區讀取新到來的數據到應用層。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"至此，一個網絡包的接收過程就已經結束了，你也可以從下圖左邊部分看到網絡包接收的流程，右邊部分剛好反過來，它是網絡包發送的流程。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"image","attrs":{"src":"https://static001.geekbang.org/infoq/42/427db83b285da4c3e6b28090f73f43da.png","alt":null,"title":null,"style":[{"key":"width","value":"75%"},{"key":"bordertype","value":"none"}],"href":null,"fromPaste":true,"pastePass":true}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"heading","attrs":{"align":null,"level":3},"content":[{"type":"text","text":"Linux 發送網絡包的流程","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"如上圖的有半部分，發送網絡包的流程正好和接收流程相反。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"首先，應用程序會調用 Socket 發送數據包的接口，由於這個是系統調用，所以會從用戶態陷入到內核態中的 Socket 層，Socket 層會將應用層數據拷貝到 Socket 發送緩衝區中。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"接下來，網絡協議棧從 Socket 發送緩衝區中取出數據包，並按照 TCP/IP 協議棧從上到下逐層處理。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"如果使用的是 TCP 傳輸協議發送數據，那麼會在傳輸層增加 TCP 包頭，然後交給網絡層，網絡層會給數據包增加 IP 包，然後通過查詢路由表確認下一跳的 IP，並按照 MTU 大小進行分片。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"分片後的網絡包，就會被送到網絡接口層，在這裏會通過 ARP 協議獲得下一跳的 MAC 地址，然後增加幀頭和幀尾，放到發包隊列中。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"這一些準備好後，會觸發軟中斷告訴網卡驅動程序，這裏有新的網絡包需要發送，最後驅動程序通過 DMA，從發包隊列中讀取網絡包，將其放入到硬件網卡的隊列中，隨後物理網卡再將它發送出去。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"----","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"heading","attrs":{"align":null,"level":3},"content":[{"type":"text","text":"總結","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"電腦與電腦之間通常都是通話網卡、交換機、路由器等網絡設備連接到一起，那由於網絡設備的異構性，國際標準化組織定義了一個七層的 OSI 網絡模型，但是這個模型由於比較複雜，實際應用中並沒有採用，而是採用了更爲簡化的 TCP/IP 模型，Linux 網絡協議棧就是按照了該模型來實現的。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"TCP/IP 模型主要分爲應用層、傳輸層、網絡層、網絡接口層四層，每一層負責的職責都不同，這也是 Linux 網絡協議棧主要構成部分。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"當應用程序通過 Socket 接口發送數據包，數據包會被網絡協議棧從上到下進行逐層處理後，纔會被送到網卡隊列中，隨後由網卡將網絡包發送出去。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"而在接收網絡包時，同樣也要先經過網絡協議棧從下到上的逐層處理，最後纔會被送到應用程序。","attrs":{}}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}}]}