lwip-users
[Top][All Lists]
Advanced

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

Re: [lwip-users] lwip_read hanging and wrong sequence number


From: Martin Persich
Subject: Re: [lwip-users] lwip_read hanging and wrong sequence number
Date: Fri, 12 Feb 2010 15:14:59 +0100

Hi,
I recomend you upgrade to LwIP 1.3.2 first. There is a lot of changes (corrected bugs).
Port version for AVR32 UC3A is in Software Framework V 1.7 file http://www.atmel.no/beta_ware/
Many files is only 1 day old ...    :-)
Martin
 
----- Original Message -----
Sent: Friday, February 12, 2010 2:54 PM
Subject: R: [lwip-users] lwip_read hanging and wrong sequence number

First of all thanks and sorry for my skipping some vital info about my environment. Here they are:
- lwip version 1.3.0
- freeRTOS version V5.1.1
 
Hanging Item:
I also cannot see any other possibility and suspect my thread to be suspended in netconn_recv at sys_arch_mbox_fetch(conn->recvmbox, (void *)&p, conn->recv_timeout). The fact that no timeout occurs is in my opinion an additional hint to my thread's hanging in the blocked tasks list and never becoming ready anymore. I am using a debugger (AVR studio, also new to me) but was not able to find a way of verifying exactly where my thread is hanging. One additional info: This problem only happens if my application receives and sends data with the external application asynchronously, no problems occur if every received message is just a response to a sent one.
 
Seqno Item:
My protocol already contains a counter in the data bytes, that's why I can say that no message went actually lost.
Some explanations to the protocol:
.130 is my application, .82 is the external application, they both send messages and acknowldge the received ones by using a counter (in blue the counter of the message sent by 82 and acked by 130, in red the viceversa)
 
frame 15 130->82: 00000900f0f1f200040b032b39 (ack to e previous msg with cnt 032b)
 
frame 16 82->130: 00004b00f0f1f2004609032c00... (new msg with cnt 032c)
 
frame 17 130->82:00000900f0f1f200040a0f2942 (new msg with cnt 0f29)
 
frame 18 82->130: 00000900f0f1f200040b0f2943 (ack to 0f29)
 
frame 20 130->82:00000900f0f1f200040b032c3a (ack to frame 16)
 
frame 21 82->130: 00004b00f0f1f2004609032d00... (new msg with cnt 032d)
 
frame 22 130->82:00000900f0f1f200040a0f2a43 (new msg with cnt 0f2a)
 
frame 23 82->130:00000900f0f1f200040b0f2a44 (ack to 0f2a)
 
frame 25 130->82:00000900f0f1f200040b032d3b (ack to frame 21)
 
No possible message loss in between, but why is the seqno counter in frame 25 increased by 13 compared to frame 24???? This does not happen e.g. in frame 20 or frame 14.
 
I hope it is clearer now.
 
Cordiali saluti
Cristina Marcucci
______________________________
mail to: address@hidden
Tel. +39.02.67493347 (PicoSoft)
 
 
 

reply via email to

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