hid-generic hangs on keyboard plug: usb_submit_urb(ctrl) failed

[pytony]

Hi everybody,

 

I'm currently working on improving my kernel and I figured out (taking a look at dmesg) that hid-generic keeps hanging on 10 secondes because of my keyboard (which causes by system to boot up within 15s instead of 5 :( )

 

I made lots of tests, plug, unplug, switch USB port, try with another computer. I always get the same error and the 10s timeout.

[  172.067760] usb 1-1.2: new low-speed USB device number 4 using ehci-pci
[  172.171664] usb 1-1.2: New USB device found, idVendor=1a2c, idProduct=0021
[  172.171674] usb 1-1.2: New USB device strings: Mfr=1, Product=2, SerialNumber=0
[  172.171679] usb 1-1.2: Product: USB Keykoard
[  172.171683] usb 1-1.2: Manufacturer: USB
[  172.175537] input: USB USB Keykoard as /devices/pci0000:00/0000:00:1a.0/usb1/1-1/1-1.2/1-1.2:1.0/0003:1A2C:0021.0003/input/input20
[  172.177877] hid-generic 0003:1A2C:0021.0003: input,hidraw2: USB HID v1.10 Keyboard [USB USB Keykoard] on usb-0000:00:1a.0-1.2/input0
[  182.176726] hid-generic 0003:1A2C:0021.0004: usb_submit_urb(ctrl) failed: -1
[  182.176770] hid-generic 0003:1A2C:0021.0004: timeout initializing reports
[  182.177265] input: USB USB Keykoard as /devices/pci0000:00/0000:00:1a.0/usb1/1-1/1-1.2/1-1.2:1.1/0003:1A2C:0021.0004/input/input21
[  182.177800] hid-generic 0003:1A2C:0021.0004: input,hidraw3: USB HID v1.10 Device [USB USB Keykoard] on usb-0000:00:1a.0-1.2/input1

It sounds like two keyboards are detected when I plug it on. The first screws up and times out, and the second works fine. And indeed, I have to keyboards in /dev/input/by-id/

 

I googled for usb_submit_urb(ctrl) but only found old issues on 2.6 kernel.

 

I found one workaround : https://bbs.archlinux.org/viewtopic.php?pid=1500574#p1500574 but I am a bit reticent about this since I don't even understand what it means nor what it is supposed to do actually.

 

Any idea on what I could do on the kernel or on udev to fix this?

 

Thanks

[sputnik]

Hehehe.  15 seconds instead of 5  :P  What would the neighbors think if they knew?

Here in the ghetto we wouldn't worry about such, but I do comprehend the workaround (I think).

Staubsauger is using:

GRUB_CMDLINE_LINUX_DEFAULT="resume=UUID=376b974a-9cc7-4a48-9e6a-de19471107f2 usbhid.quirks=0x1B1C:0x1B12:0x20000000 quiet splash"

The only part of that relevant to your issue is the usbhid.quirks statement.  So in boot.conf you would add:

params += usbhid.quirks=0x1A2C:0x0021:0x20000000

To any other parameters you have.  Take a deep breath, buckle your seatbelt & give 'er a boot after doing boot-update!  Hold on to your hat.  I don't know about the 0x20000000 option, you'll have to dig into the kernel documentation on usbhid quirks, I see also some people use 0x04 (ignore) and others.  Worst case scenario: it doesn't help, so you've got nothing to lose.

I did find though that this fellow came up with what may be a more elegant solution for a similar situation:

http://ubuntuforums.org/showthread.php?t=1161574&p=8512485#post8512485

Using udev rules instead.  Just be sure to note that he posted that 6 years ago, the udev rule syntax has changed since, in particular using SYSFS, don't.  You'll have to come up with that on your own, udev makes my head hurt.

[pytony]

Yeah, it worked with that usbhid.quirks thing. :) I won't have to buy another keyboard to get my computer boot faster. :P

 

However I'm more interested in the udev solution. I'll dig into it.

 

Thanks!

[sputnik]

If I had your computer, even if there was nobody around to impress I would say "Vroooom vrooooom!" every time it started up.

yw

[pytony]

My hardware is far from being F1 stuff, but I enjoy tweaking my kernel. :P I don't have any initramfs, desktop manager or such things either. XD

 

Actually this is 5s until mounting /home. Then it's a bit slower:

[    4.956638] XFS (sdb1): Mounting V4 Filesystem
[    5.072782] XFS (sdb1): Ending clean mount
[    5.364787] ip (591) used greatest stack depth: 12888 bytes left
[    5.959298] mount (666) used greatest stack depth: 12760 bytes left
[    6.050238] Adding 2097148k swap on /dev/sda2.  Priority:-1 extents:1 across:2097148k SS
[    6.698498] touch (945) used greatest stack depth: 12712 bytes left
[    9.724836] random: nonblocking pool is initialized
[   11.304092] r8169 0000:04:00.0 eth0: link down
[   11.304117] r8169 0000:04:00.0 eth0: link down
[   11.304133] IPv6: ADDRCONF(NETDEV_UP): eth0: link is not ready
[   16.382966] r8169 0000:04:00.0 eth0: link up
[   16.382977] IPv6: ADDRCONF(NETDEV_CHANGE): eth0: link becomes ready

But I'm sure I can reach 5s until login prompt. I don't really need swap and I could run some rc services (such as net) in background after the login prompt is displayed on the tty. I don't know if such a runlevel "after default, in background" is scriptable though...