TC执行过程解析
查看iproute2源码文件中的tc文件夹：
先看makefile
| TCOBJ= tc.o tc_qdisc.o tc_class.o tc_filter.o tc_util.o / m_police.o m_estimator.o m_action.o m_ematch.o / emp_ematch.yacc.o emp_ematch.lex.o
include ../Config
TCMODULES := TCMODULES += q_fifo.o TCMODULES += q_sfq.o TCMODULES += q_red.o TCMODULES += q_prio.o TCMODULES += q_tbf.o TCMODULES += q_cbq.o TCMODULES += f_rsvp.o TCMODULES += f_u32.o TCMODULES += f_route.o TCMODULES += f_fw.o TCMODULES += f_basic.o TCMODULES += q_dsmark.o TCMODULES += q_gred.o TCMODULES += f_tcindex.o TCMODULES += q_ingress.o TCMODULES += q_hfsc.o TCMODULES += q_htb.o TCMODULES += m_gact.o TCMODULES += m_mirred.o TCMODULES += m_ipt.o TCMODULES += m_pedit.o TCMODULES += p_ip.o TCMODULES += p_icmp.o TCMODULES += p_tcp.o TCMODULES += p_udp.o TCMODULES += em_nbyte.o TCMODULES += em_cmp.o TCMODULES += em_u32.o TCMODULES += em_meta.o
TCOBJ += $(TCMODULES)
TCLIB := tc_core.o TCLIB += tc_red.o TCLIB += tc_cbq.o TCLIB += tc_estimator.o
CFLAGS += -DCONFIG_GACT -DCONFIG_GACT_PROB
TCSO := TCSO += q_netem.so ifeq ($(TC_CONFIG_ATM),y) TCSO += q_atm.so endif
LDLIBS += -L. -ltc -lm -ldl
LDFLAGS += -Wl,-export-dynamic
YACC := bison LEX := flex
%.so: %.c $(CC) $(CFLAGS) -shared -fpic $< -o $@
all: libtc.a tc $(TCSO)
tc: $(TCOBJ) $(LIBNETLINK) $(LIBUTIL) $(TCLIB)
libtc.a: $(TCLIB) $(AR) rcs $@ $(TCLIB)
install: all mkdir -p $(DESTDIR)/usr/lib/tc install -m 0755 -s tc $(DESTDIR)$(SBINDIR) for i in $(TCSO); / do install -m 755 -s $$i $(DESTDIR)/usr/lib/tc; / done
clean: rm -f $(TCOBJ) $(TCLIB) libtc.a tc *.so emp_ematch.yacc.h; / rm -f emp_ematch.yacc.output
q_atm.so: q_atm.c $(CC) $(CFLAGS) -shared -fpic -o q_atm.so q_atm.c -latm
%.yacc.c: %.y $(YACC) $(YACCFLAGS) -o $@ $<
%.lex.c: %.l $(LEX) $(LEXFLAGS) -o$@ $< |
淡蓝色标记的为我们关注.
为了跟踪程序是如何执行的，以执行下面命令为例：
Code
| tc qdisc add dev eth0 root handle 1: htb default 12[cmd1] |
下面看tc.c中的main函数
| int main(int argc, char **argv) { /** *cmd1:argc=11(maybe 12) *argv:{ `tc` ` qdisc` ` add` ` dev ` `eth0` ` root ` `handle` 1:` `htb`` default` *`12`} */ int ret; int do_batching = 0; char *batchfile = NULL;
while (argc > 1) { if (argv[1][0] != '-') break; if (matches(argv[1], "-stats") == 0 || matches(argv[1], "-statistics") == 0) { ++show_stats; } else if (matches(argv[1], "-details") == 0) { ++show_details; } else if (matches(argv[1], "-raw") == 0) { ++show_raw; } else if (matches(argv[1], "-Version") == 0) { printf("tc utility, iproute2-ss%s/n", SNAPSHOT); return 0; } else if (matches(argv[1], "-iec") == 0) { ++use_iec; } else if (matches(argv[1], "-help") == 0) { usage(); return 0; } else if (matches(argv[1], "-force") == 0) { ++force; } else if (matches(argv[1], "-batch") == 0) { do_batching = 1; if (argc > 2) batchfile = argv[2]; argc--; argv++; } else { fprintf(stderr, "Option /"%s/" is unknown, try /"tc -help/"./n", argv[1]); return -1; } argc--; argv++; }
if (do_batching) return batch(batchfile);
if (argc <= 1) { usage(); return 0; }
tc_core_init();//从系统中读出一些设置参数 if (rtnl_open(&rth, 0) < 0) { /** *用户层和系统内核通信采用的是rtnetlink机制， *所以需要rtnetlink的支持 */ fprintf(stderr, "Cannot open rtnetlink/n"); exit(1); } /** *tc的命令形式：tc …… *所以传递给do_cmd的参数中，去掉了tc这个关键字 *这就是为什么argc-1和argv+1 *argv:{ ` qdisc` ` add` ` dev ` `eth0` ` root ` `handle` 1:` default` *`12`} */ ret = do_cmd(argc-1, argv+1); rtnl_close(&rth); return ret; } |
do_cmd会根据不同的类型，调用不同的函数来处理，例如，如果是qdisc，那么函数do_qdisc会被调用，当然传递给do_qdisc函数的参数，关键字qdisc也是会被去处的。那么[cmd1]在do_qdisc为：
| argv:{ ` add` ` dev ` `eth0` ` root ` `handle` 1:` `htb` ` default` *`12`} |
Code::do_cmd
| static int do_cmd(int argc, char **argv) { /** *cssalp&csalp： *do argc-1 and argv+1 is to escape the key word, i.e. *do_qdisc must escape 'qdisc' */ if (matches(*argv, "qdisc") == 0) return do_qdisc(argc-1, argv+1);//这是要跟踪的函数
if (matches(*argv, "class") == 0) return do_class(argc-1, argv+1);
if (matches(*argv, "filter") == 0) return do_filter(argc-1, argv+1);
if (matches(*argv, "actions") == 0) return do_action(argc-1, argv+1);
if (matches(*argv, "help") == 0) { usage(); return 0; }
fprintf(stderr, "Object /"%s/" is unknown, try /"tc help/"./n", *argv); return -1; } |
下面看看do_qdisc的
| int do_qdisc(int argc, char **argv) { if (argc < 1) return tc_qdisc_list(0, NULL);//列出系统默认的qdisc和filter等信息 if (matches(*argv, "add") == 0) return tc_qdisc_modify(RTM_NEWQDISC, NLM_F_EXCL|NLM_F_CREATE, argc-1, argv+1); if (matches(*argv, "change") == 0) return tc_qdisc_modify(RTM_NEWQDISC, 0, argc-1, argv+1); if (matches(*argv, "replace") == 0) return tc_qdisc_modify(RTM_NEWQDISC, NLM_F_CREATE|NLM_F_REPLACE, argc-1, argv+1); if (matches(*argv, "link") == 0) return tc_qdisc_modify(RTM_NEWQDISC, NLM_F_REPLACE, argc-1, argv+1); if (matches(*argv, "delete") == 0) return tc_qdisc_modify(RTM_DELQDISC, 0, argc-1, argv+1); #if 0 if (matches(*argv, "get") == 0) return tc_qdisc_get(RTM_GETQDISC, 0, argc-1, argv+1); #endif if (matches(*argv, "list") == 0 || matches(*argv, "show") == 0 || matches(*argv, "lst") == 0) return tc_qdisc_list(argc-1, argv+1); if (matches(*argv, "help") == 0) usage(); fprintf(stderr, "Command /"%s/" is unknown, try /"tc qdisc help/"./n", *argv); return -1; } |
tc_qdisc_modify函数
传递给tc_qdisc_modify函数的argv为：
| argv:{ ` dev ` `eth0` ` root ` `handle` 1:` `htb`` default` *`12`} |
tc_qdisc_modify(RTM_NEWQDISC, NLM_F_EXCL|NLM_F_CREATE, argc-1, argv+1);
对tc_qdisc_modify函数参数的说明：
1、定义rtnetlink（netlink）消息类型：RTM_NEWQDISC，即新建一个qdisc
2、设置标志flag，flag的有如下几种：
Ø NLM_F_REPLACE(ox100)覆盖以存在的
Ø NLM_F_EXCL(ox200)如果存在就不创建
Ø NLM_F_CREATE(ox400)不存在就创建
Ø NLM_F_APPEND(ox800)添加到末尾
Code::flag
| #define NLM_F_REPLACE 0x100 /* Override existing */ #define NLM_F_EXCL 0x200 /* Do not touch, if it exists */ #define NLM_F_CREATE 0x400 /* Create, if it does not exist */ #define NLM_F_APPEND 0x800 /* Add to end of list */ /*在linux系统里面有相同的定义*/ |
3、设置命令参数的个数和命令参数
Code：：tc_qdisc_modify
| int tc_qdisc_modify(int cmd, unsigned flags, int argc, char **argv) { /** * argv:{ ` dev ` `eth0` ` root ` `handle` 1:` `htb`` default` *`12`} */ struct qdisc_util *q = NULL; struct tc_estimator est; char d[16]; char k[16]; struct { /** * nlmsghdr为和系统内核通信的消息结构的头 *tcmsg为tc消息结构 */ struct nlmsghdr n; struct tcmsg t; char buf[TCA_BUF_MAX]; } req;
memset(&req, 0, sizeof(req)); memset(&est, 0, sizeof(est)); memset(&d, 0, sizeof(d)); memset(&k, 0, sizeof(k));
req.n.nlmsg_len = NLMSG_LENGTH(sizeof(struct tcmsg)); req.n.nlmsg_flags = NLM_F_REQUEST|flags; req.n.nlmsg_type = cmd;//对应[cmd1],type=rtm_newqdisc req.t.tcm_family = AF_UNSPEC;//family
while (argc > 0) { if (strcmp(*argv, "dev") == 0) { NEXT_ARG();//point to nextà ethi(i is number) if (d[0]) duparg("dev", *argv); strncpy(d, *argv, sizeof(d)-1);//eth0 } else if (strcmp(*argv, "handle") == 0) { __u32 handle; if (req.t.tcm_handle) duparg("handle", *argv); NEXT_ARG(); if (get_qdisc_handle(&handle, *argv))//handle=1 invarg(*argv, "invalid qdisc ID"); req.t.tcm_handle = handle; } else if (strcmp(*argv, "root") == 0) { if (req.t.tcm_parent) { /** *如果存在parent， 那么就不能为root */ fprintf(stderr, "Error: /"root/" is duplicate parent ID/n"); return -1; } req.t.tcm_parent = TC_H_ROOT; #ifdef TC_H_INGRESS } else if (strcmp(*argv, "ingress") == 0) { if (req.t.tcm_parent) { fprintf(stderr, "Error: /"ingress/" is a duplicate parent ID/n"); return -1; } req.t.tcm_parent = TC_H_INGRESS; strncpy(k, "ingress", sizeof(k)-1); q = get_qdisc_kind(k); req.t.tcm_handle = 0xffff0000;
argc--; argv++; break; #endif } else if (strcmp(*argv, "parent") == 0) { __u32 handle; NEXT_ARG(); if (req.t.tcm_parent) duparg("parent", *argv); if (get_tc_classid(&handle, *argv)) invarg(*argv, "invalid parent ID"); req.t.tcm_parent = handle; } else if (matches(*argv, "estimator") == 0) { if (parse_estimator(&argc, &argv, &est)) return -1; } else if (matches(*argv, "help") == 0) { usage(); } else { strncpy(k, *argv, sizeof(k)-1);//k=`htb` /** *get_qdisc_kind *1、首先查看，与要添加的qdisc相关的qdisc_util是否已经存在了 * 如果存在，那么返回 *2、如果qdisc_util 还存在，那么查找系统的动态链接库，看是否存在 *通过函数dlopen打开动态链接库，然后通过函数q = dlsym(dlh, buf) *得到相应的qdisc_util *3、如果系统的动态链接也不存在,[cmd1]就是这种能够情况 *那么，就创建一个新的(q = malloc(sizeof(*q));)，如果成功 *然后设置一下参数{ * q->id = strcpy(malloc(strlen(str)+1), str); * q->parse_qopt = parse_noqopt; * q->print_qopt = print_noqopt; } *4、最后将找到的qdisc_util,挂载到qdisc_list的上（链表）,并返回q *5、如果不成功，那么返回null(无能为力了) */ q = get_qdisc_kind(k); argc--; argv++; break; } argc--; argv++; } /** *addattr函数设置nlmsghdr中的一些参数，为和内核通信传递参数做准备 */ if (k[0]) addattr_l(&req.n, sizeof(req), TCA_KIND, k, strlen(k)+1); if (est.ewma_log) addattr_l(&req.n, sizeof(req), TCA_RATE, &est, sizeof(est));
if (q) { if (!q->parse_qopt) { fprintf(stderr, "qdisc '%s' does not support option parsing/n", k); return -1; } /** *在这个例子中，parse_qopt=parse_noqopt, do nothing and return 0 */ if (q->parse_qopt(q, argc, argv, &req.n)) return 1; } else { if (argc) { if (matches(*argv, "help") == 0) usage();
fprintf(stderr, "Garbage instead of arguments /"%s .../". Try /"tc qdisc help/"./n", *argv); return -1; } } /** *根据设备名，找到设备，并设置消息，传递设备信息到内核 */ if (d[0]) { int idx;
ll_init_map(&rth);
if ((idx = ll_name_to_index(d)) == 0) { fprintf(stderr, "Cannot find device /"%s/"/n", d); return 1; } req.t.tcm_ifindex = idx; } /** *rtnl_talk函数，实现和内核通信 *在rtnl_talk函数中，会调用sendmsg函数，将信息传递到内核， *在内核中，最终会调用tc_modify_qidsc函数来创建qdisc *在pktsched_init函数中(linux 内核) *rtnl_register(PF_UNSPEC, RTM_NEWQDISC, tc_modify_qdisc, NULL); *而消息设定信息： * req.n.nlmsg_type = cmd;//对应[cmd1],type=RTM_NEWQDISC * req.t.tcm_family = AF_UNSPEC; *当然rtnl_talk会监听内核返回的消息，并做相应的处理 */ if (rtnl_talk(&rth, &req.n, 0, 0, NULL, NULL, NULL) < 0) return 2;
return 0; } |