lwip-users
[Top][All Lists]
Advanced

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

Re: [lwip-users] Dropped UDP packages


From: Andrew Foster
Subject: Re: [lwip-users] Dropped UDP packages
Date: Thu, 18 Aug 2011 11:21:03 -0400
User-agent: Mozilla/5.0 (X11; U; Linux x86_64; en-US; rv:1.9.2.18) Gecko/20110621 Fedora/3.1.11-1.fc14 Lightning/1.0b3pre Thunderbird/3.1.11

To be fair, I haven't looked at the latest SDK from TI for the Tempest class (9xxx). TI may have corrected the issues that I experienced. The E2E forum would probably be a better place to get updates on TI's lwip/FreeRTOS example. You may also try searching the lwip-users archive for others who have came across similar issues.

When I did my development it was based on the 6965 eth SDK. This implementation called into lwIP from the interrupt level. Generally, that is a bad idea because it can lead to issues resulting from concurrent access on the pbuf chain. Without an entire redesign of the Stellaris model, you are pretty much left with accepting the vulnerability or disabling interrupts while calling into lwIP.

That being said, I would still work to prove/disprove the vulnerability. When I was working through my issue, the application failed pretty spectacularly. When you dork up the pbuf chain the tcp_ip_thread typically freezes or writes to an erroneous memory location. Your issue could still be caused by a number of other things.

Andrew

On 08/18/2011 10:26 AM, Emil Ohlsson wrote:
Actually I'm using a TI Stellaris Cortex M3 (LM3S9B96). What kind of flaws 
where you thinking of?

Thanks
Emil Ohlsson

-----Original Message-----
From: address@hidden [mailto:address@hidden On Behalf Of Andrew Foster
Sent: den 18 augusti 2011 15:11
To: Mailing list for lwIP users
Subject: Re: [lwip-users] Dropped UDP packages

Emil,

I've had to debug a similar event on my target and found it helpful to get 
familiar with stats.c. The various compiler switches can be found in 
lwipopts.h(sounds like you are already aware of UDP_STATS.) More specifically, 
look in to MEM, MEMP, PBUF and SYS stats. Verify you don't have any errors in 
these paths before getting down and dirty with your ISR. I know the Luminary/TI 
ISR has some flaws but hopefully your target platform's ISR has been 
implemented correctly.

Hope this helps,

Andrew



On 08/18/2011 08:59 AM, Emil Ohlsson wrote:
Hi!

I'm writing and application that listens for RTP-packages via UDP multicast. 
But for some reason there is a small package drop, say 1 per 1000 packets.

I've activated UDP stats and it seems like lwIP doesn't think there are any 
dropped packets, but from my application I can see that there are missing 
sequence numbers. And I can see that those packages are being sent using 
wireshark.

Does lwIP discard packages if the application isn't fast enough to
fetch them? Could there be any other thing I've missed? Could there be
some memory issue? Any comments or ideas are very welcome

Best regards
Emil Ohlsson

_______________________________________________
lwip-users mailing list
address@hidden
https://lists.nongnu.org/mailman/listinfo/lwip-users
_______________________________________________
lwip-users mailing list
address@hidden
https://lists.nongnu.org/mailman/listinfo/lwip-users

_______________________________________________
lwip-users mailing list
address@hidden
https://lists.nongnu.org/mailman/listinfo/lwip-users



reply via email to

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