英文:
Understanding Linux Kernel Code: net/ipv4/igmp.c - Only on a v3 query the function timer_expire() is called but nothing similar on a v1 query?
问题
在net/ipv4/igmp.c文件中,您在IGMP v1查询的情况下找不到负责响应IGMP v1查询的代码行。您已经在igmp.c的每个函数的开头写了printk。当您发送一个IGMP v3查询时,您可以通过printk消息直接观察到函数调用顺序如下:
igmp_heard_query() -> igmp_gq_start_timer() -> igmp_gq_timer_expire() -> igmpv3_send_report()
这是您预期的结果,也是您正在寻找的,但只有在v3查询时才能观察到这一点。
当您发送IGMP v1查询时,您只能观察到以下调用:
igmp_rcv() -> igmp_heard_query()
这对我来说毫无意义,因为一个函数必须发送报告(我假设是igmp_send_report())。
您的设置似乎没有问题,您可以在Wireshark中看到v1查询/报告,并验证它们已发送到您自己编译的Linux内核。您没有意外地将查询发送到自己的笔记本电脑。
console_loglevel设置为8:
# cat /proc/sys/kernel/printk
8 4 1 7
在igmp_heard_query()函数中,我没有看到与v1相关的以下内容:
- 启动定时器
- 触发定时器到期
- 调用igmp_send_report()的内容
唯一的假设是更改in_dev->mr_v1_seen将自动触发in_dev->timer吗?但即使如此,仍然不清楚为什么我在printk中没有看到igmp_timer_expire()的调用。
英文:
I have trouble to find the code line in net/ipv4/igmp.c, which responds to a IGMP v1 query with a IGMP report.
I wrote printk in the beginning of each function in igmp.c. When I send a IGMP v3 query, I can observe via the printk messages straight forward the function calls:
igmp_heard_query() -> igmp_gq_start_timer() -> igmp_gq_timer_expire() -> igmpv3_send_report()
This is what I expected and what I'm looking for, except that I only can observe this on a v3 query.
When I send a IGMP v1 query I can only observe the calls
igmp_rcv() -> igmp_heard_query()
That makes absolutely no sense to me, because one function have to send the report. (I assume igmp_send_report())
My setup is fine, I can see the v1 queries/reports in wireshark and verified, that they are sent to my self compiled linux kernel. I'm not sending accidentally the query to my notebook.
console_loglevel is set to 8:
# cat /proc/sys/kernel/printk
8 4 1 7
igmp_heard_query()
https://github.com/torvalds/linux/blob/8032bf1233a74627ce69b803608e650f3f35971c/net/ipv4/igmp.c#L940
In this whole function, I don't see anything related for v1 to
- starting a timer
- trigger a timer to expire
- something which then calls igmp_send_report()
/* return true if packet was dropped */
static bool igmp_heard_query(struct in_device *in_dev, struct sk_buff *skb,
int len)
{
struct igmphdr *ih = igmp_hdr(skb);
struct igmpv3_query *ih3 = igmpv3_query_hdr(skb);
struct ip_mc_list *im;
__be32 group = ih->group;
int max_delay;
int mark = 0;
struct net *net = dev_net(in_dev->dev);
if (len == 8) {
if (ih->code == 0) {
/* Alas, old v1 router presents here. */
max_delay = IGMP_QUERY_RESPONSE_INTERVAL;
in_dev->mr_v1_seen = jiffies +
(in_dev->mr_qrv * in_dev->mr_qi) +
in_dev->mr_qri;
group = 0;
} else {
/* v2 router present */
max_delay = ih->code*(HZ/IGMP_TIMER_SCALE);
in_dev->mr_v2_seen = jiffies +
(in_dev->mr_qrv * in_dev->mr_qi) +
in_dev->mr_qri;
}
/* cancel the interface change timer */
WRITE_ONCE(in_dev->mr_ifc_count, 0);
if (del_timer(&in_dev->mr_ifc_timer))
__in_dev_put(in_dev);
/* clear deleted report items */
igmpv3_clear_delrec(in_dev);
} else if (len < 12) {
return true; /* ignore bogus packet; freed by caller */
} else if (IGMP_V1_SEEN(in_dev)) {
/* This is a v3 query with v1 queriers present */
max_delay = IGMP_QUERY_RESPONSE_INTERVAL;
group = 0;
} else if (IGMP_V2_SEEN(in_dev)) {
/* this is a v3 query with v2 queriers present;
* Interpretation of the max_delay code is problematic here.
* A real v2 host would use ih_code directly, while v3 has a
* different encoding. We use the v3 encoding as more likely
* to be intended in a v3 query.
*/
max_delay = IGMPV3_MRC(ih3->code)*(HZ/IGMP_TIMER_SCALE);
if (!max_delay)
max_delay = 1; /* can't mod w/ 0 */
} else { /* v3 */
if (!pskb_may_pull(skb, sizeof(struct igmpv3_query)))
return true;
ih3 = igmpv3_query_hdr(skb);
if (ih3->nsrcs) {
if (!pskb_may_pull(skb, sizeof(struct igmpv3_query)
+ ntohs(ih3->nsrcs)*sizeof(__be32)))
return true;
ih3 = igmpv3_query_hdr(skb);
}
max_delay = IGMPV3_MRC(ih3->code)*(HZ/IGMP_TIMER_SCALE);
if (!max_delay)
max_delay = 1; /* can't mod w/ 0 */
in_dev->mr_maxdelay = max_delay;
/* RFC3376, 4.1.6. QRV and 4.1.7. QQIC, when the most recently
* received value was zero, use the default or statically
* configured value.
*/
in_dev->mr_qrv = ih3->qrv ?: READ_ONCE(net->ipv4.sysctl_igmp_qrv);
in_dev->mr_qi = IGMPV3_QQIC(ih3->qqic)*HZ ?: IGMP_QUERY_INTERVAL;
/* RFC3376, 8.3. Query Response Interval:
* The number of seconds represented by the [Query Response
* Interval] must be less than the [Query Interval].
*/
if (in_dev->mr_qri >= in_dev->mr_qi)
in_dev->mr_qri = (in_dev->mr_qi/HZ - 1)*HZ;
if (!group) { /* general query */
if (ih3->nsrcs)
return true; /* no sources allowed */
igmp_gq_start_timer(in_dev);
return false;
}
/* mark sources to include, if group & source-specific */
mark = ih3->nsrcs != 0;
}
/*
* - Start the timers in all of our membership records
* that the query applies to for the interface on
* which the query arrived excl. those that belong
* to a "local" group (224.0.0.X)
* - For timers already running check if they need to
* be reset.
* - Use the igmp->igmp_code field as the maximum
* delay possible
*/
rcu_read_lock();
for_each_pmc_rcu(in_dev, im) {
int changed;
if (group && group != im->multiaddr)
continue;
if (im->multiaddr == IGMP_ALL_HOSTS)
continue;
if (ipv4_is_local_multicast(im->multiaddr) &&
!READ_ONCE(net->ipv4.sysctl_igmp_llm_reports))
continue;
spin_lock_bh(&im->lock);
if (im->tm_running)
im->gsquery = im->gsquery && mark;
else
im->gsquery = mark;
changed = !im->gsquery ||
igmp_marksources(im, ntohs(ih3->nsrcs), ih3->srcs);
spin_unlock_bh(&im->lock);
if (changed)
igmp_mod_timer(im, max_delay);
}
rcu_read_unlock();
return false;
}
The only assumption I have, is that changing in_dev->mr_v1_seen will automatically trigger the in_dev->timer? But then still it is unclear, why I don't see the call of igmp_timer_expire() with my printk.
/*
* A socket has joined a multicast group on device dev.
*/
static void ____ip_mc_inc_group(struct in_device *in_dev, __be32 addr,
unsigned int mode, gfp_t gfp)
...
timer_setup(&im->timer, igmp_timer_expire, 0);
答案1
得分: 0
我找到了解决方案:
这就是行为,如果我没有加入一个多播组。
英文:
I found the solution:
That's the behavior, if I didn't joined a multicast group.
通过集体智慧和协作来改善编程学习和解决问题的方式。致力于成为全球开发者共同参与的知识库,让每个人都能够通过互相帮助和分享经验来进步。
评论