Re: Linux IP stack change to obtain TTL min/max/avg

From: mpk_india
Date: Thu Jan 01 2009 - 09:42:30 EST



Hello Sam,

Can you elaborate more on how I can fetch these TTL min/max/avg values using
netdev?

Thanks

Mpk

Sam Ravnborg wrote:
>
> netdev is a much better place for this - added.
>
> Sam
>
> On Wed, Dec 31, 2008 at 03:20:53AM -0800, mpk_india wrote:
>>
>> Hello All,
>>
>> I am trying to modify the linux ip stack to get the avg/min/max of the
>> TTL
>> value (in trasport layer) from the received UDP packet. I have modifed
>> the
>> 2.6.28 kernel for this and added new setsockopt flags as IP_RECVTTL_AVG,
>> IP_RECVTLL_MIN, IP_RECVTTL_MAX.
>>
>> Please see below changes I have made to do this:
>>
>> 1. Inside "in.h" file Added following macros
>> #define IP_RECVTTL_AVG 20
>> #define IP_RECVTTL_MIN 21
>> #define IP_RECVTTL_MAX 22
>>
>> 2. Inside "inet_sock.h" in struct inet_sock {
>> added 3 variables
>> __s16 uc_ttl_avg;
>> __s16 uc_ttl_min;
>> __s16 uc_ttl_max;
>>
>> 3. In ipv4/udp.c added following peice of code
>> sk = __udp4_lib_lookup_skb(skb, uh->source, uh->dest, udptable);
>> if (sk != NULL) {
>>
>> #if 1 // Changes to get TTL stats..Start
>> struct iphdr *iph;
>> struct inet_sock *inet;
>> inet = inet_sk(sk);
>> iph = ip_hdr(skb);
>> inet->uc_ttl_avg = (iph->ttl + inet->uc_ttl_avg)>>1;
>> if (inet->uc_ttl_max < iph->ttl)
>> inet->uc_ttl_max = iph->ttl;
>> if (inet->uc_ttl_min > iph->ttl)
>> inet->uc_ttl_min = iph->ttl;
>> #endif // Changes to get TTL stats..Start
>>
>> int ret = udp_queue_rcv_skb(sk, skb);
>> sock_put(sk);
>>
>> 4. Changes inside "ip_sockglue.c" file
>>
>> static void ip_cmsg_recv_ttl_avg(struct msghdr *msg, struct sk_buff *skb)
>> {
>> struct inet_sock *inet = inet_sk(skb->sk);
>> int ttl = inet->uc_ttl_avg;
>> printk(KERN_DEBUG "***<recv ttl avg () is called -- value = %d***\n",
>> ttl);
>> put_cmsg(msg, SOL_IP, IP_TTL, sizeof(int), &ttl);
>> }
>>
>> static void ip_cmsg_recv_ttl_max(struct msghdr *msg, struct sk_buff *skb)
>> {
>> struct inet_sock *inet = inet_sk(skb->sk);
>> int ttl = inet->uc_ttl_max;
>> printk(KERN_DEBUG "***<recv ttl max () is called -- value = %d***\n",
>> ttl);
>> put_cmsg(msg, SOL_IP, IP_TTL, sizeof(int), &ttl);
>> }
>>
>> static void ip_cmsg_recv_ttl_min(struct msghdr *msg, struct sk_buff *skb)
>> {
>> struct inet_sock *inet = inet_sk(skb->sk);
>> int ttl = inet->uc_ttl_min;
>> printk(KERN_DEBUG "***<recv ttl min () is called -- value = %d***\n",
>> ttl);
>> put_cmsg(msg, SOL_IP, IP_TTL, sizeof(int), &ttl);
>> }
>>
>> void ip_cmsg_recv(struct msghdr *msg, struct sk_buff *skb)
>> {
>> .
>> .
>> .
>> .
>> #if 1 // Changes to get TTL stats..Start
>> if (flags & 1)
>> ip_cmsg_recv_ttl_avg(msg, skb);
>> if ((flags>>=1) == 0)
>> return;
>> if (flags & 1)
>> ip_cmsg_recv_ttl_min(msg, skb);
>> if ((flags>>=1) == 0)
>> return;
>> if (flags & 1)
>> ip_cmsg_recv_ttl_max(msg, skb);
>> #endif // Changes to get TTL stats..Start
>>
>> }
>>
>> inside do_ip_setsockopt() added following code
>>
>> #if 1 // Changes to get TTL stats..Start
>> case IP_RECVTTL_AVG:
>> if (val)
>> inet->cmsg_flags |= IP_CMSG_TTL_AVG;
>> else
>> inet->cmsg_flags &= ~IP_CMSG_TTL_AVG;
>> break;
>> case IP_RECVTTL_MIN:
>> if (val)
>> inet->cmsg_flags |= IP_CMSG_TTL_MIN;
>> else
>> inet->cmsg_flags &= ~IP_CMSG_TTL_MIN;
>> break;
>> case IP_RECVTTL_MAX:
>> if (val)
>> inet->cmsg_flags |= IP_CMSG_TTL_MAX;
>> else
>> inet->cmsg_flags &= ~IP_CMSG_TTL_MAX;
>> break;
>> #endif // Changes to get TTL stats..End
>>
>> =================================================================
>>
>> I have compiled this and written the udp client server to test this.
>>
>>
>> In my client application i am sending 64 packets with some dummy message.
>> In
>> server I have used the recvmsg() and setsockopt() to test these changes
>> as
>> below:
>>
>> setsockopt(serverfd, IPPROTO_IP, IP_OPTIONS, NULL, 0);
>> if (setsockopt (serverfd,IPPROTO_IP, IP_RECVTTL_AVG, &ttl,ttlSize)<0) {
>>
>> printf ("Error occured in setsockopt\n");
>> exit(1);
>> }
>>
>> now if i do the recvmsg 64 times after recvfrom in my server code it
>> gives
>> me proper ttl avg values.. but i do not want to seek this values for
>> every
>> packet i receive.. i want them once for all sent 64 udp packets.. the
>> same
>> recvmsg fails if i doit single recvmsg after 64 received packets..
>>
>>
>> Can anyone tell about what i am doing wrong in the present scenario?
>>
>> Thanks in advance
>> mpk
>> --
>> View this message in context:
>> http://www.nabble.com/Linux-IP-stack-change-to-obtain-TTL-min-max-avg-tp21230026p21230026.html
>> Sent from the linux-kernel mailing list archive at Nabble.com.
>>
>> --
>> To unsubscribe from this list: send the line "unsubscribe linux-kernel"
>> in
>> the body of a message to majordomo@xxxxxxxxxxxxxxx
>> More majordomo info at http://vger.kernel.org/majordomo-info.html
>> Please read the FAQ at http://www.tux.org/lkml/
> --
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@xxxxxxxxxxxxxxx
> More majordomo info at http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at http://www.tux.org/lkml/
>
>

--
View this message in context: http://www.nabble.com/Linux-IP-stack-change-to-obtain-TTL-min-max-avg-tp21230026p21241797.html
Sent from the linux-kernel mailing list archive at Nabble.com.

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/