Re: [lwip-devel] core/timers.c with NO_SYS?

[Simon Goldschmidt]

Leon Woestenberg <address@hidden> wrote:
> I am not following the criticism. Who talked about interrupt locks? I
> certainly did not.

I was talking about interrupt locks since its currently a common implementation 
to protect the memp and pbuf code.

> Why would pbuf, mem and memp be interrupt disabled locks? That makes
> no sense to me. They are not in any way related to an interrupt
> (handler).

Of course they are: RX interrupts from the MAC may call pbuf_alloc(PBUF_POOL) 
to allocate a pbuf to copy received data to.

You may implement the MAC driver in a different way, but on some platforms you 
are a) limited in RAM so cannot preallocate too many pbufs and/or b) must pull 
packets from the MAC as fast as you can, so cannot defer this to the main loop 
(without losing packets due to MAC buffer overrun).


Simon
-- 
NEU: FreePhone - kostenlos mobil telefonieren!                  
Jetzt informieren: http://www.gmx.net/de/go/freephone