Hi,
I've been working on a Cypress SX2 controller driver and things are starting to fall into place :) Pings with a small packet size work (<= max packet size).
Cool -- another high speed controller!
Presumably you didn't need many changes to make "ether.c" behave. Please send me the patch to teach it the endpoint mappings when you're ready. The current "ether.c" adds a non-CDC mode, since it turned out the PXA-2xx UDC can't really support any more than one interface with a single altsetting. (What a waste, with that many endpoints!)
I don't (yet) see how usb packets are gathered to form a ethernet frame. I remember with the usbnet driver, packets are gatheres until a < max packet size (usb) packet arrives and indicates an end of ethernet frame. I'll dig deeper into the CDCEther stuff, but maybe somebody knows...
Same framing: one ethernet packet maps into N USB packets, of maxpacketsize (usb) except for the last one. If the last one is "naturally" short, fine; otherwise the drivers add an extra (short) packet on xmit. ("usbnet" adds a few more framing schemes, which are worth avoiding, but that's the most widely used one.)
Note that there were some bugs in the CDCEther driver on 2.4 relating to that framing, and to formatting the status events.
I know that 2.4.20 had both types of bugs. It may be worth updating to the 2.4.22-pre version of CDCEther.c, it ought to be a drop-in replacement for your 2.4.19 kernel.
- Dave
------------------------------------------------------- This SF.Net email sponsored by: Free pre-built ASP.NET sites including Data Reports, E-commerce, Portals, and Forums are available now. Download today and enter to win an XBOX or Visual Studio .NET. http://aspnet.click-url.com/go/psa00100006ave/direct;at.asp_061203_01/01 _______________________________________________ [EMAIL PROTECTED] To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
