[lwip-users] lwip 1.3.1 MEMP_NUM_TCP_SEG overflow.

[Misaka10032]

Hi all, I needed your help. I use LwIP without RTOS. In the process, I use
the RAW API. Stack almost debugged, with the exception of trouble, which I
fight a few days. During the exchange at high speed (about 30-40 packets per
second, the size of the packet - 10-20 bytes) after a large (0.5 - 1 million
packets) occurs gradually overflow TCP segments (number of segments in a
short time begins to rise and eventually overflow occurs). The number of
segments taken with a large margin. All kinds of memory protection I have
included sys_lightweight_prot defined as 1, the protection function is
realized, although I have nothing interrupts from lwIP API is not used.
Suspicion fell on the functions tcp_write() and tcp_output(), because I
changed the location tcp_enqueue pbufs on PBUF_POOL. I if before the memory
leak was in the stack memory (area MEM_SIZE), now with the segments begins
to leak memory PBUF_POOL.

Thank you for your help.
Best regards,
Dmitriy.



--
View this message in context: 
http://lwip.100.n7.nabble.com/lwip-1-3-1-MEMP-NUM-TCP-SEG-overflow-tp26980.html
Sent from the lwip-users mailing list archive at Nabble.com.