Re: [lwip-users] sys_timeout

[Peter Graf]

Curt McDowell wrote:

> The way sys_arch is implemented, timeouts can only fire when the main process 
> is sleeping on a semaphore or mailbox using either
> sys_sem_wait() or sys_mbox_fetch() in src/core/sys.c.
>
> I think it's an architectural deficiency that any of sys_arch is implemented in 
> the lwip core.

I found it a benefit.

> The whole idea of sys_arch is to provide a system abstraction, but currently 
> it's difficult to replace
> sys_timeout() and other sys_arch functionality with a native O/S 
implementation.

I don't see a point in replacing sys_timeout(). Adam had moved this and 
other stuff which previously required architure specific implementation 
into the core, and thereby reached the goal of easier portability, in my 
case and many others. I don't see a proper multitasking O/S where the 
few system functions are hard to implement the way things are.

> Worse, any application that wants to use lwip is forced to be based on lwip's 
> sys_arch primitives.

If lwIP is intialized and the main+tcp+interface threads are active 
(this can be set up outside the application), the application just has 
to link itself into lwIP's thread list once. Afterwards it is free to 
call the lwIP networking API(s) or not, but it's not required to use the 
sys_arch primitives.

All the best
Peter