Looking at ethtool GDRVINFO support in the USB network drivers, some in 2.4 use usb_make_path() for their bus_info field, and some use usb<busnum>:<devicenum>. Everything in 2.5 uses usb_make_path().
It'd be worth IMO making all 2.4 drivers use usb_make_path(), since those IDs are stable.
In 2.5, identifying devices with usb_make_path() is fairly simple; it's all exported via sysfs. Is there a similar function in 2.4 to, given a usb device, determine its usb_make_path() ID without walking the entire USB and PCI trees?
Actually, I'd say the usb_make_path() physical IDs are not really available through sysfs. Though maybe some folk would argue that you can (usually) find the components of that physical ID in the sysfs path somewhere, so that might be enough. (Bus numbering is the sticky point. Changing driver load ordering should never be able to make a device's physical path ID change.)
How exactly would you be "given a USB device"? If you're given a network interface name, GDRVINFO is the answer. HID devices have a different ioctl to expose this. Not all other driver frameworks have any notion of a device path.
- 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
