"Martin Velek" wrote: > Maybe there is a problem with your mbox mechanism, because usually the > socket functions are waiting for confirmation from tcp/ip thread via > semaphore or message queue. I do not know how is handled the > non-blocking socket exactly.
Non-blocking sockets are handled the same way, the only difference is they do not wait for free buffer space if the current send buffer is not big enough. But while data is copied from application buffer to stack-internal buffers, the application thread calling send() is still blocked. As such, this looks like a bug in the mbox/semaphore implementation. Simon _______________________________________________ lwip-users mailing list [email protected] https://lists.nongnu.org/mailman/listinfo/lwip-users
