Maison > Article > développement back-end > 《深入理解Nginx》笔记之ngx_event_pipe_s结构体
ngx_event_pipe_s
<code><span>typedef</span><span>struct</span> ngx_event_pipe_s ngx_event_pipe_t; <span>// 处理接收自上游的包体的回调函数原型</span><span>typedef</span> ngx_int_t (*ngx_event_pipe_input_filter_pt)(ngx_event_pipe_t *p, ngx_buf_t *buf); <span>// 向下游发送响应的回调函数原型</span><span>typedef</span> ngx_int_t (*ngx_event_pipe_output_filter_pt)(<span>void</span> *data, ngx_chain_t *chain); <span>struct</span> ngx_event_pipe_s { <span>// Nginx与上游服务器之间的连接</span> ngx_connection_t *upstream; <span>// Nginx与下游客户端之间的连接</span> ngx_connection_t *downstream; <span>/* 直接接收自上游服务器的缓冲区链表,保存的是未经任何处理的数据。这个链表是逆序的,后接受的响应插在链表头处 */</span> ngx_chain_t *free_raw_bufs; <span>// 表示接收到的上游响应缓冲区,其数据是经过input_filter处理的</span> ngx_chain_t *in; <span>// 指向刚收到的一个缓冲区</span> ngx_chain_t **last_in; <span>// 保存着将要发给客户端的缓冲区链表。在写入临时文件成功时,会把in中的缓冲区添加到out中</span> ngx_chain_t *out; <span>// 等待释放的缓冲区</span> ngx_chain_t *<span>free</span>; <span>// 表示上次调用ngx_http_output_filter函数发送响应时没有发送完的缓冲区链表</span> ngx_chain_t *busy; <span>/* * the input filter i.e. that moves HTTP/1.1 chunks * from the raw bufs to an incoming chain */</span><span>// 处理接收到的、来自上游服务器的数据</span> ngx_event_pipe_input_filter_pt input_filter; <span>// 用于input_filter的的参数,一般是ngx_http_request_t的地址</span><span>void</span> *input_ctx; <span>// 向下游发送响应的函数</span> ngx_event_pipe_output_filter_pt output_filter; <span>// output_filter的参数,指向ngx_http_request_t</span><span>void</span> *output_ctx; <span>// 1:表示当前已读取到上游的响应</span><span>unsigned</span> read:<span>1</span>; <span>// 1:启用文件缓存</span><span>unsigned</span> cacheable:<span>1</span>; <span>// 1:表示接收上游响应时,一次只能接收一个ngx_buf_t缓冲区</span><span>unsigned</span> single_buf:<span>1</span>; <span>// 1:一旦不再接收上游包体,将尽可能地释放缓冲区</span><span>unsigned</span> free_bufs:<span>1</span>; <span>// 1:表示Nginx与上游交互已经结束</span><span>unsigned</span> upstream_done:<span>1</span>; <span>// 1:Nginx与上游服务器的连接出现错误</span><span>unsigned</span> upstream_error:<span>1</span>; <span>// 1:表示与上游服务器的连接已关闭</span><span>unsigned</span> upstream_eof:<span>1</span>; <span>/* 1:表示暂时阻塞读取上游响应的的流程。此时会先调用ngx_event_pipe_write_to_downstream 函数发送缓冲区中的数据给下游,从而腾出缓冲区空间,再调用ngx_event_pipe_read_upstream 函数读取上游信息 */</span><span>unsigned</span> upstream_blocked:<span>1</span>; <span>// 1:与下游的交互已结束</span><span>unsigned</span> downstream_done:<span>1</span>; <span>// 1:与下游的连接出现错误</span><span>unsigned</span> downstream_error:<span>1</span>; <span>// 1:复用临时文件。它是由ngx_http_upstream_conf_t中的同名成员赋值的</span><span>unsigned</span> cyclic_temp_file:<span>1</span>; <span>// 已分配的缓冲区数据</span> ngx_int_t allocated; <span>// 记录了接收上游响应的内存缓冲区大小,bufs.size表示每个内存缓冲区大小,bufs.num表示最多可以有num个缓冲区</span> ngx_bufs_t bufs; <span>// 用于设置、比较缓冲区链表中的ngx_buf_t结构体的tag标志位</span> ngx_buf_tag_t tag; <span>/* busy缓冲区中待发送响应长度的最大值,当到达busy_size时,必须等待busy缓冲区发送了足够的数据, 才能继续发送out和in中的内容 */</span> ssize_t busy_size; <span>// 已经接收到来自上游响应包体的长度</span> off_t read_length; off_t length; <span>// 表示临时文件的最大长度</span> off_t max_temp_file_size; <span>// 表示一次写入文件时数据的最大长度</span> ssize_t temp_file_write_size; <span>// 读取上游响应的超时时间</span> ngx_msec_t read_timeout; <span>// 向下游发送响应的超时时间</span> ngx_msec_t send_timeout; <span>// 向下游发送响应时,TCP连接中设置的send_lowat“水位”</span> ssize_t send_lowat; <span>// 连接池</span> ngx_pool_t *pool; <span>// 日志</span> ngx_log_t *<span>log</span>; <span>// 表示在接收上游服务器响应头部阶段,已经读取到响应包体</span> ngx_chain_t *preread_bufs; <span>// 表示在接收上游服务器响应头部阶段,已经读取到响应包体长度</span> size_t preread_size; <span>// 用于缓存文件</span> ngx_buf_t *buf_to_file; size_t limit_rate; time_t start_sec; <span>// 存放上游响应的临时文件</span> ngx_temp_file_t *temp_file; <span>// 已使用的ngx_buf_t缓冲区数目</span><span>/* STUB */</span><span>int</span> num; }; </code>
版权声明:Pain is just in your mind.
以上就介绍了《深入理解Nginx》笔记之ngx_event_pipe_s结构体,包括了方面的内容,希望对PHP教程有兴趣的朋友有所帮助。