[RFC] [PATCH] usbatm.[ch]: multiple changes

[matthieu castet]

Hi Roman,

Roman Kagan wrote:
> On Wed, Mar 30, 2005 at 07:26:37PM +0200, matthieu castet wrote:
> 
>>Roman Kagan wrote:
	/* don't expect iso out endpoints */
>>
>>But you could add interrupt out endpoint ;)
> 
> 
> Interrupt transfers seem to be even less adequate for ATM :)...  The
> comment actually refers to the next line
It was a little joke because the sagem seem to support it [1]...


>>>+			iso_size -= iso_size % trx->channel->stride;	/* 
>>>alignment */
>>>+			BUG_ON(!iso_size);
>>>+			iso_packets = (trx->channel->buf_size - 1) / 
>>>iso_size + 1;
>>
>>using default buf_size is quite low : for iso_size 1000, we have only 3 
>>or 4 iso_packets per urb instead of 16.
> 
> 
> Is 16 the recommended number?  Anyway, feel free to suggest better
It was the value used in the official driver, I don't know what's the 
best value...

>>>+		}
>>>+
>>>+		trx->urb = usb_alloc_urb(iso_packets, GFP_KERNEL);
>>
>>                 **********************
> 
> 
> Sorry, what is wrong here?
Nothing it was refering with the pb earlier (calling before allocating)

Matthieu

[1]
I:  If#= 1 Alt= 0 #EPs= 2 Cls=0a(data ) Sub=00 Prot=00 Driver=ueagle-atm
E:  Ad=04(O) Atr=02(Bulk) MxPS=  64 Ivl=0ms <--firmware
E:  Ad=02(O) Atr=02(Bulk) MxPS=  64 Ivl=0ms <-- data
I:  If#= 1 Alt= 1 #EPs= 2 Cls=0a(data ) Sub=00 Prot=00 Driver=ueagle-atm
E:  Ad=04(O) Atr=02(Bulk) MxPS=  64 Ivl=0ms
E:  Ad=02(O) Atr=03(Int.) MxPS=  64 Ivl=1ms