rx_process for isoc

[Duncan Sands]

On Friday 18 November 2005 17:04, Stanislaw Gruszka wrote:
> > that would work too, but in the case when data is coming in fast,
> > so all iso packets are full, you have to do a partial copy to
> > cell_buf for *every* iso packet (at least with the speedtouch)
> > because the packet length is not divisible by the stride (for
> > the speedtouch, the stride is 53, and the packet size is 640
> > or 960).  Maybe it should be profiled.
> AFIAK when all iso packets are full then urb->actual_length 
> == packet_size*num_packets == rx_channel.buf_size. 

Yes, but the same situation arises partially whenever large network
packets are coming in (large = bigger than the iso packet size).  For
example, a standard maximum ethernet packet amounts to about 2 full
iso packets followed by a partial packet (assuming an iso packet size
of 640 bytes).  It seems a pity to do two pointless copies to cell_buf.
I can't deny that it adds complexity.  I will try to do some profiling
and see if it actually gains anything (probably not).  If not, then it
is easy to go back to the original, simpler version.

> if (usb_pipeisoc(urb->pipe)) {
> 	if (urb->actual_length == instance->rx_channel.buf_size)
> 		urbatm_extract_cells(instance, urb->transfer_buffer, urb->actual_lenght)	
> 	else
>  		for (i = 0; i < urb->number_of_packets; i++) {
> 
> The only problem here is if assumption 
> 
> len = 0
> for (i = 0; i < urb->number_of_packets; i++)
> 	len += urb->iso_fame_desc[i].actual_length
> ASSERT(urb->actual_length == len);
> 
> is true and actual_length == 0 when error occurs. 
> USB guys should know this.

Yes - want to ask them?

Ciao,

Duncan.