http://bugzilla.suse.com/show_bug.cgi?id=954324 http://bugzilla.suse.com/show_bug.cgi?id=954324#c1 Takashi Iwai changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |kolAflash@kolAhilft.de, | |oneukum@suse.com, | |tiwai@suse.com Flags| |needinfo?(kolAflash@kolAhil | |ft.de) --- Comment #1 from Takashi Iwai --- How did you check that nousb doesn't take effect? Looking through the source code, indeed xhci-hcd doesn't check this option at probe unlike other host drivers, so it may be present. But USB device drivers won't load at the registration level, so it should be still effective. We should add the check in xhcd-hcd.c, too, but I wonder what any other is missing. -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.suse.com/show_bug.cgi?id=954324 http://bugzilla.suse.com/show_bug.cgi?id=954324#c3 --- Comment #3 from Takashi Iwai --- I see what you meant, thanks. The culprit is that the USB core is now built as a module where it was built into kernel in the earlier distro versions. You need to pass usbcore.nousb=1 instead. IMO, this is a bug in the code, though. The driver should refer to nousb no matter how it's built. Also, xhci_pci is still loaded even if usb is disabled. This is another thing to be fixed. Other host drivers refuse to load. -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.suse.com/show_bug.cgi?id=954324 http://bugzilla.suse.com/show_bug.cgi?id=954324#c5 --- Comment #5 from Oliver Neukum --- (In reply to Takashi Iwai from comment #3)

I see what you meant, thanks.

The culprit is that the USB core is now built as a module where it was built into kernel in the earlier distro versions. You need to pass usbcore.nousb=1 instead.

IMO, this is a bug in the code, though. The driver should refer to nousb no matter how it's built.

It does. usbcore.nousb always should work.

Also, xhci_pci is still loaded even if usb is disabled. This is another thing to be fixed. Other host drivers refuse to load.

Yes, that is a bug. -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.suse.com/show_bug.cgi?id=954324 http://bugzilla.suse.com/show_bug.cgi?id=954324#c7 --- Comment #7 from kolA flash --- Thank you a lot for your help! Will you report/fix this upstream? I did some more testing: The full error message lsusb usually reports when nousb is in use on openSUSE 13.2 is: unable to initialize libusb: -99 Return code is 1 (normally 0, if USB isn't disabled). Using usbcore.nousb=1 on openSUSE 13.2 or Ubuntu 15.10 seems to have no effect. So there's probably not one way to disable USB on kernels with buildin USB and kernels with USB module, except putting both nousb and usbcore.nousb=1 in the commandline (I actually did that and it worked fine). Using usbcore.nousb=1 on openSUSE 42.1 and Tumbleweed disables USB as expected. /sys/module/usbcore/parameters/nousb is Y is usbcore.nousb=1 is in use. But lsusb behaves a little different. Is exits without any message. Nevertheless this is probably not relevant (and maybe due to the different lsusb version on 42.1/current Tumbleweed relative to 13.2) Return code of lsusb is 1 as expected. -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.suse.com/show_bug.cgi?id=954324 http://bugzilla.suse.com/show_bug.cgi?id=954324#c9 Oliver Neukum changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |RESOLVED Resolution|--- |FIXED --- Comment #9 from Oliver Neukum --- Fix added to tree -- You are receiving this mail because: You are on the CC list for the bug.