lwip-users
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [lwip-users] PPPoS: ppp session disconnect if hit bad fcs for LCP pa


From: Axel Lin
Subject: Re: [lwip-users] PPPoS: ppp session disconnect if hit bad fcs for LCP packet
Date: Sat, 1 Jul 2017 12:23:17 +0800

2017-07-01 0:18 GMT+08:00 Sylvain Rochet <address@hidden>:
> Hi Axel,
>
> On Fri, Jun 30, 2017 at 11:18:52PM +0800, Axel Lin wrote:
>>
>> BTW, my device always keep sending data to LTE through pppos.
>> I'm wondering why the bad fcs usually happened on LCP packet.
>
> What I can say is, since the PPP link is only between your MCU and your
> LTE modem, which are probably a few centimers away on the same board,
> having bad FCS should *NEVER* happen! Checksums are close to being
> totally useless in this case, they are just here for protocol conformity
> regarding HDLC.
>
> Having bad FCS in this case means a software issue here, either on the
> lwIP side or, most likely, on the LTE modem side.
Hi Sylvain,
Thanks for your comments.

I tried test it on 2 different LTE modules (different model, using for
different carrier).

One device always got bad fcs on "VJ compressed TCP/IP".
I got 10 times bad fcs:
fcs  protocol
6c1f 2d
aa7c 2d
59df 2d
2c11 2d
f698 2d
50e1 2d
68ab 2d
e9a6 2d
2890 2d
245c 2d

>From the stats display I found 20 link.drop, 10 link.chkerr (bad fcs).
The drop count is always double of chkerr count.
The additional 10 drop is in below code path:
  /*
   * Toss all non-LCP packets unless LCP is OPEN.
   */
  if (protocol != PPP_LCP && pcb->lcp_fsm.state != PPP_FSM_OPENED) {
    ppp_dbglog("Discarded non-LCP packet when LCP not open");
    goto drop;
  }

On the other device I got checksum as below:
fcs  protocol
6f32 d          (Unknown protocol?)
eae4 c021       (LCP)

And it disconnect after got the LCP checksum error.

>From above, I guess the problem is not at LTE side because I tried
different LTE from different vender.

Regards,
Axel



reply via email to

[Prev in Thread] Current Thread [Next in Thread]