Re: [lwip-devel] Re: [task #7040] Work on tcp_enqueue

[address@hidden]

Alain M. wrote:
> Yes, and a bit more... The problem gets worse on 32 bit arch like arm7 
> that only operate in 32 bit aligned chunks. If independent write blocks 
> are not a multiple of 4 bytes, they will have to be concatenated later 
> on. So it might be just as well to do that early. And it is no use 
> atempting zero copy in that case too, for that same reason.
>   

Yes, I had the same problems. But I managed to write my own version of 
memcpy that showed no significant slowdown when copying to/from 
unaligned pointers by loading into a register and shifting in order to 
always have 32-bit accesses. Of course that only works if you have a 
1-cycle-shift operation.

As to zero-copy: pbufs should be allocated on a boundary that is OK for 
DMA on your platform, but of course, if the first pbuf has the wrong 
length, the target of the second pbuf will be unaligned. That would also 
be solved with the one-pbuf-per-segment approach.

Simon