nginx对keepalive和pipeline请求处理分析

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

    hc = r->http_connection;

    b = r->header_in;

//一般情况下，当解析完header_in之后，pos会设置为last。也就是读取到的数据刚好是一个完整的http请求.当pos小于last，则说明可能是一个pipeline请求。

    if(b->pos last) {

        /* the pipelined request */

        if(b != c->buffer) {

            /*

             * If the large header buffers were allocated while the previous

             * request processing then we do not use c->buffer for

             * the pipelined request (see ngx_http_init_request()).

             *

             * Now we would move the large header buffers to the free list.

             */

            cscf = ngx_http_get_module_srv_conf(r, ngx_http_core_module);

//如果free为空，则新建

            if(hc->free== NULL) {

//可以看到是large_client_headers的个数

                hc->free= ngx_palloc(c->pool,

                  cscf->large_client_header_buffers.num * sizeof(ngx_buf_t *));

                if(hc->free== NULL) {

                    ngx_http_close_request(r, 0);

                    return;

                }

            }

//然后清理当前的request的busy

            for(i = 0; i nbusy - 1; i++) {

                f = hc->busy[i];

                hc->free[hc->nfree++] = f;

                f->pos = f->start;

                f->last = f->start;

            }

//保存当前的header_in buf,以便与下次给free使用。

            hc->busy[0] = b;

            hc->nbusy = 1;

        }

    }