Problem in ‘pico_protocol_generic_loop’

[StupidZJF]

I read the source code of picotcp, and I have some questions.
First, what is the ‘pico_protocol_generic_loop’ function purpose? What is in the loop? The protocol or the layer?
Second, Whether the project is multi-threading, if so, where does the code create the new thread? if not , how does the function 'pico_stack_tick' to assign CPU time slices to all the modules previously registered

[danielinux]

Hello,
the function you mention is used to loop across all the modules (i.e. each of the protocol) input and output functions, and this happens every time pico_stack_tick is called. This allows the packets to flow in both direction at every tick. In a bare-metal, single thread, pico_stack_tick should be called regularly from the main event loop.

The support for multithreaded environments works as follows:

• PICO_SUPPORT_MUTEX must be enable, to ensure that the socket queues (which may be shared across threads) are protected, and you should provide the implementation of the mutex primitives (lock, unlock)
• The OS should spawn a separate thread which calls pico_stack_tick at regular interval
• Other threads can open sockets and call picotcp API

For examples of integration in RTOS see also:

• https://github.com/tass-belgium/picotcp-bsd (generic picotcp OSAL)
• https://github.com/insane-adding-machines/frosted (embedded GPL *nix kernel implementing POSIX socket calls using picoTCP)

[danielinux]

More details on how to port on *OS in the wiki documentation:
https://github.com/tass-belgium/picotcp/wiki/Porting-picoTCP-OS

[StupidZJF]

Thank you for your reply.
there is a question that does that means it is single thread if I do not call some additional thread function manually? So is it a event-trigger structure？ So what does upper layer do while lower layer is processing the frame. For example, when ipv4 is processing a frame which sent by udp, what dose udp doing? Looping or waiting?