Alan Stern wrote: > What makes you think the phone should run at high speed? Does it run > at high speed when you plug it into any other computers?
Well, I had thought that devices claiming 2.0 compliance would run at high speed instead of full. For a mass storage device, I would expect my ultra-mega-hyper-fast-and-expensive memory card so be accessed at something more than 12 Mb/s without needing to extract the card from the phone, plugging it to a reader, etc. Anyway it's a nokia implementation issue. It works on my laptop at full speed too, bound on the uhci driver. > You need to understand that there are USB 2.0 devices which don't run > at high speed. USB 2.0 is backwards compatible with USB 1.1; a USB 2.0 > device is allowed to run at any speed it wants. Okidoki. Too bad I thought that usb 2.0 conformance would require at least mass-storage devices to set up at high speed... >> Where does the ehci/companion driver selection occur, and >> what is the logic behind that ? > > Speed selection occurs in the hardware. The phone has to send a > special "chirp" signal to connect at high speed; if it doesn't send > that signal then it connects at full speed. > > Following that, EHCI/companion selection occurs in the driver. If the > connection was made at high speed then the EHCI controller is used, > otherwise the port is switched over to the companion controller. So no way to prevent that switch. That's good, because it would have been a bit more mess for the users ! >> - Did someone test such devices in mass storage mode with >> ohci drivers ? May this be an endianness issue on the ohci driver ? > > The OHCI drivers work just fine on PowerPC machines, which are > big-endian. Have you tried using a 2.6.23 kernel? Not yet, this is a very specific embedded hardware prototype, and I don't want to brick it in the next few days. Furthermore it probably requires specific kernel patches to run properly and I am not sure to have the exact source tree... > It sounds like the phone works perfectly on the x86 system. The mipseb > system has some problem, but it's hard to tell what. Have you tried > plugging other USB devices into that computer? Yes. ehci mass storage devices seem to work fine. Many other 1.1 mass storage (at least 3-4) fail. Thank you for the explanation, I'll feel less dumb tonight. For now I'll have a look at the diff between the usb tree of my device and a recent kernel one, to see if somethings looks plausibly fixing the bug. Maybe then I'll recompile just the modules. -- Eric Estievenart ------------------------------------------------------------------------- SF.Net email is sponsored by: Check out the new SourceForge.net Marketplace. It's the best place to buy or sell services for just about anything Open Source. http://sourceforge.net/services/buy/index.php _______________________________________________ Linux-usb-users@lists.sourceforge.net To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-users
