On Fri, May 27, 2011 at 03:37:37PM +0800, Greg KH wrote: > On Thu, May 26, 2011 at 10:45:10PM -0700, Sarah Sharp wrote: > Thanks for taking the kernel boot parameter out, much nicer. > > Just one tiny nit: > > > + if (!ports_available) > > + dev_warn(&xhci_pdev->dev, > > + "WARNING: BIOS reports no USB 3.0 ports? " > > + "Email linux-usb@xxxxxxxxxxxxxxx\n"); > > + > > Do you really want to get emails about this? > We've been burned in the past, but I know we still have messages for > this for the usb-storage stuff, and it's useful there. > > What would a user be able to do, or you be able to do, if someone does > email you based on this, and the other warning message like this? How > are we going to be able to resolve it, by fixing the kernel code, or > their BIOS? If we find that BIOSes don't set the port routing mask registers correctly, we can just add a quirk for that platform/BIOS version that writes all ones to the port switchover registers. I really have no hope for fixing BIOSes. :) Originally, when there was some internal FUD about whether all USB 2.0 devices would work under xHCI, I offered the port mask as a compromise. If OEMs found that some internal USB 2.0 device didn't work under xHCI (and they were using the xHCI ports internally instead of EHCI for whatever reason), then they could tell their BIOS writers to disable that port in the port mask. So there could be a practical reason why the port mask register doesn't get set. In practice, BIOS writers being what they are, I think it will either never get set (since Windows doesn't use this particular mask register, and we all know if a BIOS works with Windows it gets shipped), or will always get set to the suggested value in the EDS documentation, which is exactly the number of ports that the xHCI host supports. We could just write all ones to the switchover ports, which avoids BIOS writers all together and assumes the non-working USB 2.0 device under xHCI is a rare case, which it should be. Sarah Sharp -- To unsubscribe from this list: send the line "unsubscribe linux-usb" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
