[lwip-devel] TCP Client system fault

[Vinicius Bernardi]

Hi Everyone,

I`m facing a issue that is very complex to understand how it's happening.
The issue happens like this : 

I have a PPP connection + TCP Client, I keep just 1 client connection open.
This connection can be used for many hours without any problem, sending and 
receiving packages.
But if I decide to send a huge buffer (1400 bytes) with very short interval, as 
I receive the TCP ACK, I send a new one, the code is generating a system fault.
I'm using a RTOS uC/OS-II port, and as I said it's working very well, even with 
Huge buffers in case that the interval it's bigger.
I already increase STACK size, buffer's sizes, but as far as I don't know when 
the problem will happens, and sometimes it happnes in 50 seconds, sometimes 
with 300 seconds after I had start it, it's vey complex to understand why the 
system fault it`s cause by.
The it`s always in the same state, and the data is 0x00000000.
I'm using a Cortex M4.
Does anybody have some clue, or additional information how can I debbug it, and 
get more information?
I already have all the trace information enabled, and I didn't find a route 
cause, or a different sittuation inside the log.

My best regards,
Vinicius