[speedtouch] Re: The size of the send buffer

2003-02-03 Thread Edouard Gomez
Duncan Sands ([EMAIL PROTECTED]) wrote: > Receive buffer: 64 * 53 is the least common multiple of 64 (packet > size of the bulk endpoint) and 53 (ATM cell size). That way there is > no wasted USB bandwidth. But you can use other sizes too. > > Sending: the send buffer can be up to 64k. > No b

[speedtouch] Re: The size of the send buffer

2003-02-03 Thread Duncan Sands
Hi Edouard, I found out about the send and receive buffer sizes. The person who told me would not like it to be publically known that he told me. Receive buffer: 64 * 53 is the least common multiple of 64 (packet size of the bulk endpoint) and 53 (ATM cell size). That way there is no wasted USB

[speedtouch] Re: The size of the send buffer

2003-01-31 Thread Duncan Sands
> > PS: I guess in practice pppd never sends very big packets, > > so probably the actual size of the usb buffer sent to 0x07 > > never exceeds about 3k. > > It's even worse. pppd+pppoa3 are sending single atm cells most of the > time or when real payload is a bit longer it sends 2 cells because o

[speedtouch] Re: The size of the send buffer

2003-01-31 Thread Edouard Gomez
Duncan Sands ([EMAIL PROTECTED]) wrote: > Hi, I have a technical question: when reading from the > bulk (0x87) endpoint, the buffer size is 64*53, i.e. 64 ATM > cells. Does anyone know why this size limitation exists? I entered the project after the pusb lib was ready, so i don't know precisely