Possible Bug: 2.4.14 USB Keyboard

Possible Bug: 2.4.14 USB Keyboard

[Martin McWhorter]

Hello,

I wanted to try the new preemptive kernel patch so I downloaded the 
2.4.14 sources from kernel.org and the preemptive patch for 2.4.14, 
configured the kernel, built it, mkinitrded and liloed.

Everything came up fine, though my usb keyboard does not respond. The 
currios thing is that when the kernel is boooting and sysinit starts the 
USB devices they all report that they start ok:

Nov 14 09:42:01 m_mcwhorter rc.sysinit: Mounting USB filesystem:  succeeded
Nov 14 09:42:01 m_mcwhorter rc.sysinit: Initializing USB controller 
(usb-uhci):  succeeded
Nov 14 09:42:01 m_mcwhorter rc.sysinit: Initializing USB HID interface: 
  succeeded
Nov 14 09:42:01 m_mcwhorter rc.sysinit: Initializing USB keyboard: 
succeeded
Nov 14 09:42:01 m_mcwhorter rc.sysinit: Initializing USB mouse:  succeeded

But the USB keyboard does not repond, though the mouse connected through 
the keyboard hub works.

Is this a possible bug?

Linux 2.4.14 with preemptive patch
Redhat 7.1
HP Pavillion 9690C with supplied USB keyboard USB mouse
Intel Pentium III 800

Thanks,
Martin

Re: Possible Bug: 2.4.14 USB Keyboard

[Greg KH]

On Wed, Nov 14, 2001 at 03:18:55PM -0600, Martin McWhorter wrote:
> 
> But the USB keyboard does not repond, though the mouse connected through 
> the keyboard hub works.
> 
> Is this a possible bug?

Do you have the same problem without the preemptive patch?

If so, can you send the results of /proc/bus/usb/devices and
/proc/bus/usb/drivers?

thanks,

greg k-h

Re: Possible Bug: 2.4.14 USB Keyboard

[Martin McWhorter]

Greg

Do you have the same problem without the preemptive patch?


Yes. I appoligize for not being clear. I get the same results with plain 
2.4.14 w/out the preemptive patch.


If so, can you send the results of /proc/bus/usb/devices andT:  Bus=01 
Lev=00 Prnt=00 Port=00 Cnt=00 Dev#=  1 Spd=12  MxCh= 2


B:  Alloc=159/900 us (18%), #Int=  4, #Iso=  0
D:  Ver= 1.00 Cls=09(hub  ) Sub=00 Prot=00 MxPS= 8 #Cfgs=  1
P:  Vendor=0000 ProdID=0000 Rev= 0.00
S:  Product=USB UHCI Root Hub
S:  SerialNumber=d400
C:* #Ifs= 1 Cfg#= 1 Atr=40 MxPwr=  0mA
I:  If#= 0 Alt= 0 #EPs= 1 Cls=09(hub  ) Sub=00 Prot=00 Driver=hub
E:  Ad=81(I) Atr=03(Int.) MxPS=   8 Ivl=255ms
T:  Bus=01 Lev=01 Prnt=01 Port=00 Cnt=01 Dev#=  2 Spd=12  MxCh= 3
D:  Ver= 1.10 Cls=09(hub  ) Sub=00 Prot=00 MxPS= 8 #Cfgs=  1
P:  Vendor=03f0 ProdID=010c Rev= 0.01
S:  Manufacturer=HP
S:  Product=Multimedia Keyboard Hub
C:* #Ifs= 1 Cfg#= 1 Atr=e0 MxPwr=100mA
I:  If#= 0 Alt= 0 #EPs= 1 Cls=09(hub  ) Sub=00 Prot=00 Driver=hub
E:  Ad=81(I) Atr=03(Int.) MxPS=   1 Ivl=255ms
T:  Bus=01 Lev=02 Prnt=02 Port=00 Cnt=01 Dev#=  3 Spd=12  MxCh= 0
D:  Ver= 1.10 Cls=00(>ifc ) Sub=00 Prot=00 MxPS= 8 #Cfgs=  1
P:  Vendor=03f0 ProdID=020c Rev= 0.01
S:  Manufacturer=HP
S:  Product=Multimedia Keyboard Hub
C:* #Ifs= 2 Cfg#= 1 Atr=e0 MxPwr=  0mA
I:  If#= 0 Alt= 0 #EPs= 1 Cls=03(HID  ) Sub=01 Prot=01 Driver=hid
E:  Ad=81(I) Atr=03(Int.) MxPS=   8 Ivl= 10ms
I:  If#= 1 Alt= 0 #EPs= 1 Cls=03(HID  ) Sub=00 Prot=00 Driver=hid
E:  Ad=82(I) Atr=03(Int.) MxPS=   4 Ivl=255ms
T:  Bus=01 Lev=02 Prnt=02 Port=02 Cnt=02 Dev#=  4 Spd=1.5 MxCh= 0
D:  Ver= 1.00 Cls=00(>ifc ) Sub=00 Prot=00 MxPS= 8 #Cfgs=  1
P:  Vendor=046d ProdID=c001 Rev=11.01
S:  Manufacturer=Logitech
C:* #Ifs= 1 Cfg#= 1 Atr=a0 MxPwr= 26mA
I:  If#= 0 Alt= 0 #EPs= 1 Cls=03(HID  ) Sub=01 Prot=02 Driver=hid
E:  Ad=81(I) Atr=03(Int.) MxPS=   8 Ivl= 10ms


 > /proc/bus/usb/drivers?

          usbdevfs
          hub
          hid
          keyboard

Thanks,
Martin

Re: Possible Bug: 2.4.14 USB Keyboard

[Greg KH]

On Thu, Nov 15, 2001 at 08:24:41AM -0600, Martin McWhorter wrote:
> 
> Yes. I appoligize for not being clear. I get the same results with plain 
> 2.4.14 w/out the preemptive patch.

Thanks for the files.  It looks like the USB drivers are being bound to
your devices properly.

What does the kernel log say when you plug in your keyboard?

thanks,

greg k-h

Re: Possible Bug: 2.4.14 USB Keyboard

[Martin McWhorter]

Greg,

Sorry I had a think-o. I sent you the info on the wrong kernel.

/proc/bus/usb/devices for 2.4.14 VANILLA
T:  Bus=01 Lev=00 Prnt=00 Port=00 Cnt=00 Dev#=  1 Spd=12  MxCh= 2
B:  Alloc=146/900 us (16%), #Int=  3, #Iso=  0
D:  Ver= 1.00 Cls=09(hub  ) Sub=00 Prot=00 MxPS= 8 #Cfgs=  1
P:  Vendor=0000 ProdID=0000 Rev= 0.00
S:  Product=USB UHCI Root Hub
S:  SerialNumber=d400
C:* #Ifs= 1 Cfg#= 1 Atr=40 MxPwr=  0mA
I:  If#= 0 Alt= 0 #EPs= 1 Cls=09(hub  ) Sub=00 Prot=00 Driver=hub
E:  Ad=81(I) Atr=03(Int.) MxPS=   8 Ivl=255ms
T:  Bus=01 Lev=01 Prnt=01 Port=00 Cnt=01 Dev#=  2 Spd=12  MxCh= 3
D:  Ver= 1.10 Cls=09(hub  ) Sub=00 Prot=00 MxPS= 8 #Cfgs=  1
P:  Vendor=03f0 ProdID=010c Rev= 0.01
S:  Manufacturer=HP
S:  Product=Multimedia Keyboard Hub
C:* #Ifs= 1 Cfg#= 1 Atr=e0 MxPwr=100mA
I:  If#= 0 Alt= 0 #EPs= 1 Cls=09(hub  ) Sub=00 Prot=00 Driver=hub
E:  Ad=81(I) Atr=03(Int.) MxPS=   1 Ivl=255ms
T:  Bus=01 Lev=02 Prnt=02 Port=00 Cnt=01 Dev#=  3 Spd=12  MxCh= 0
D:  Ver= 1.10 Cls=00(>ifc ) Sub=00 Prot=00 MxPS= 8 #Cfgs=  1
P:  Vendor=03f0 ProdID=020c Rev= 0.01
S:  Manufacturer=HP
S:  Product=Multimedia Keyboard Hub
C:* #Ifs= 2 Cfg#= 1 Atr=e0 MxPwr=  0mA
I:  If#= 0 Alt= 0 #EPs= 1 Cls=03(HID  ) Sub=01 Prot=01 Driver=hid
E:  Ad=81(I) Atr=03(Int.) MxPS=   8 Ivl= 10ms
I:  If#= 1 Alt= 0 #EPs= 1 Cls=03(HID  ) Sub=00 Prot=00 Driver=hid
E:  Ad=82(I) Atr=03(Int.) MxPS=   4 Ivl=255ms
T:  Bus=01 Lev=02 Prnt=02 Port=02 Cnt=02 Dev#=  4 Spd=1.5 MxCh= 0
D:  Ver= 1.00 Cls=00(>ifc ) Sub=00 Prot=00 MxPS= 8 #Cfgs=  1
P:  Vendor=046d ProdID=c001 Rev=11.01
S:  Manufacturer=Logitech
C:* #Ifs= 1 Cfg#= 1 Atr=a0 MxPwr= 26mA
I:  If#= 0 Alt= 0 #EPs= 1 Cls=03(HID  ) Sub=01 Prot=02 Driver=hid
E:  Ad=81(I) Atr=03(Int.) MxPS=   8 Ivl= 10ms

/proc/bus/usb/drivers
          usbdevfs
          hub
          hid
          keyboard


> 
> What does the kernel log say when you plug in your keyboard?
> 

Nov 15 12:38:10 m_mcwhorter kernel: usb.c: USB disconnect on device 2
Nov 15 12:38:10 m_mcwhorter kernel: usb.c: USB disconnect on device 3
Nov 15 12:38:10 m_mcwhorter kernel: usb-uhci.c: interrupt, status 3, 
frame# 1572
Nov 15 12:38:10 m_mcwhorter kernel: usb.c: USB disconnect on device 4
Nov 15 12:38:12 m_mcwhorter kernel: hub.c: USB new device connect on 
bus1/1, assigned device number 5
Nov 15 12:38:12 m_mcwhorter kernel: hub.c: USB hub found
Nov 15 12:38:12 m_mcwhorter kernel: hub.c: 3 ports detected
Nov 15 12:38:13 m_mcwhorter kernel: hub.c: USB new device connect on 
bus1/1/1, assigned device number 6
Nov 15 12:38:13 m_mcwhorter kernel: input0: USB HID v1.00 Keyboard [HP 
Multimedia Keyboard Hub] on usb1:6.0
Nov 15 12:38:13 m_mcwhorter kernel: : USB HID v1.00 Device [HP 
Multimedia Keyboard Hub] on usb1:6.1
Nov 15 12:38:13 m_mcwhorter /etc/hotplug/usb.agent: ... no drivers for 
USB product 3f0/10c/1
Nov 15 12:38:13 m_mcwhorter kernel: hub.c: USB new device connect on 
bus1/1/3, assigned device number 7
Nov 15 12:38:13 m_mcwhorter kernel: input1: USB HID v1.00 Mouse 
[Logitech] on usb1:7.0


Thanks again,
Martin

Re: Possible Bug: 2.4.14 USB Keyboard

[Vojtech Pavlik]

On Thu, Nov 15, 2001 at 12:40:03PM -0600, Martin McWhorter wrote:
> Greg,
> 
> Sorry I had a think-o. I sent you the info on the wrong kernel.

Do you have the keybdev module loaded? Also, don't load the usbkbd
module, if you load hid ...

-- 
Vojtech Pavlik
SuSE Labs

Re: Possible Bug: 2.4.14 USB Keyboard

[Pete Zaitcev]

> Do you have the keybdev module loaded? Also, don't load the usbkbd
> module, if you load hid ...
> 
> -- 
> Vojtech Pavlik
> SuSE Labs

There is a small problem with this approach: users have no clue
how to control what modules are loaded, hotplug loads whatever
was built (and recorded in modules.usbmap), and some users
have keyboards that plainly refuse to work with hid, therefore
vendors have to build both modules.

See this little gem, for instance:
 https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=55878

I suspect some distributions can get away with "load the right
module" approach because their userbase is so small and technical
that they do not hit these cases often. I think something needs
fixing in hid.

-- Pete

Re: Possible Bug: 2.4.14 USB Keyboard

[Vojtech Pavlik]

On Thu, Nov 15, 2001 at 02:30:12PM -0500, Pete Zaitcev wrote:
> > Do you have the keybdev module loaded? Also, don't load the usbkbd
> > module, if you load hid ...
> > 
> > -- 
> > Vojtech Pavlik
> > SuSE Labs
> 
> There is a small problem with this approach: users have no clue
> how to control what modules are loaded, hotplug loads whatever
> was built (and recorded in modules.usbmap), and some users
> have keyboards that plainly refuse to work with hid, therefore
> vendors have to build both modules.
> 
> See this little gem, for instance:
>  https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=55878
> 
> I suspect some distributions can get away with "load the right
> module" approach because their userbase is so small and technical
> that they do not hit these cases often. I think something needs
> fixing in hid.

Interesting. Could you by any chance try to run the 'evtest' program (i
can send it to you if you don't have it handy) on the /dev/input/event
device created for this keyboard? And/or dmesg with DEBUG enabled in
hid-core.c? Also, latest kernels should make even the extra keys of this
keyboard work ...

-- 
Vojtech Pavlik
SuSE Labs

Re: Possible Bug: 2.4.14 USB Keyboard

[Martin McWhorter]

Pete,

> ok, so no usbkbd, it seems. But still, do lsmod too, please.

[root@m_mcwhorter m_mcwhorter]# /sbin/lsmod
Module                  Size  Used by
emu10k1                49584   0  (autoclean)
sr_mod                 13968   0  (autoclean)
ac97_codec              9248   0  (autoclean) [emu10k1]
soundcore               3376   4  (autoclean) [emu10k1]
usbkbd                  2944   0  (unused)
autofs                  9184   1  (autoclean)
8139too                13152   1  (autoclean)
ipchains               31232   0
ide-scsi                7712   0
scsi_mod               86256   2  [sr_mod ide-scsi]
ide-cd                 26624   0
cdrom                  27328   0  [sr_mod ide-cd]
mousedev                3936   1
keybdev                 1728   0  (unused)
hid                    12576   0  (unused)
input                   3136   0  [usbkbd mousedev keybdev hid]
usb-uhci               21344   0  (unused)
usbcore                49184   1  [usbkbd hid usb-uhci]


Thanks for the help,
Martin

Re: Possible Bug: 2.4.14 USB Keyboard

[Pete Zaitcev]

> Date: Thu, 15 Nov 2001 12:44:39 -0600
> From: Martin McWhorter <m_mcwhorter@prairiegroup.com>
> To: Pete Zaitcev <zaitcev@redhat.com>,
>    linux-kernel <linux-kernel@vger.kernel.org>
> Subject: Re: Possible Bug: 2.4.14 USB Keyboard
> 
> > ok, so no usbkbd, it seems. But still, do lsmod too, please.
> 
> [root@m_mcwhorter m_mcwhorter]# /sbin/lsmod
> Module                  Size  Used by
> emu10k1                49584   0  (autoclean)
> sr_mod                 13968   0  (autoclean)
> ac97_codec              9248   0  (autoclean) [emu10k1]
> soundcore               3376   4  (autoclean) [emu10k1]
> usbkbd                  2944   0  (unused)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

YES! Kill this bastard (the best of all is to use rm(1), then reboot).

>[...]
> mousedev                3936   1
> keybdev                 1728   0  (unused)
> hid                    12576   0  (unused)
> input                   3136   0  [usbkbd mousedev keybdev hid]
> usb-uhci               21344   0  (unused)
> usbcore                49184   1  [usbkbd hid usb-uhci]

Dunno how, but usbkbd often manages to disable hid.
Never let it get built. We do have reports of people who
have keyboards not working with hid, very infrequently.
Of course, those guys prefer just to build usbkbd than to
help figuring why hid does not work. If not for that,
usbkbd would be removed from kernel sources long ago.

-- Pete

Re: Possible Bug: 2.4.14 USB Keyboard

[Martin McWhorter]

Pete,


>>usbkbd                  2944   0  (unused)
>>
> ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> 
> YES! Kill this bastard (the best of all is to use rm(1), then reboot).
> 


rm /lib/modules/2.4.14/kernel/drivers/usb/usbkbd.o

reboot

HP usb-keyboard still does not work. Now the kernel that it does work 
with is 2.4.2 from redhat's tree. I have not tried it with a vanilla 2.4.2

Now when I do an
[root@m_mcwhorter m_mcwhorter]# /sbin/lsmod
Module                  Size  Used by
emu10k1                49584   0  (autoclean)
sr_mod                 13968   0  (autoclean)
ac97_codec              9248   0  (autoclean) [emu10k1]
soundcore               3376   4  (autoclean) [emu10k1]
autofs                  9184   1  (autoclean)
8139too                13152   1  (autoclean)
ipchains               31232   0
ide-scsi                7712   0
scsi_mod               86256   2  [sr_mod ide-scsi]
ide-cd                 26624   0
cdrom                  27328   0  [sr_mod ide-cd]
mousedev                3936   1
keybdev                 1728   0  (unused)
hid                    12576   0  (unused)
input                   3136   0  [mousedev keybdev hid]
usb-uhci               21344   0  (unused)
usbcore                49184   1  [hid usb-uhci]

Martin

Re: Possible Bug: 2.4.14 USB Keyboard

[Martin McWhorter]

Kinda sorta partial success


> "make modules_install", but this time rename hid.o into hid.hidden.
> Let me know if this fixes the problem.

I went ahead and did this. Booted up and got the expected error message 
when HID tried to load. Keyboard works but USB mouse ofcourse does not. 
So I went and renamed hid.hidden to hid.o and loaded it. Now my USB 
keyboard and USB mouse are working.

I suspect it has something to do with the order the modules are loaded? 
hid.o must be last? I am just guessing.

I suppose I can edit the rc.sysinit to load hid AFTER usbkbd is loaded.

Martin

Re: Possible Bug: 2.4.14 USB Keyboard

[Martin McWhorter]

Pete,



> Try to rename /sbin/hotplug into something else _temporarily_,

Gave this a try. No luck. Interestingly nothing apperently stopped working.

But I did find something that did. Remove the loading of hid in 
rc.sysinit and add it the the end of rc.local.

Thanks alot Pete,
Martin

Re: Possible Bug: 2.4.14 USB Keyboard

[Greg KH]

On Thu, Nov 15, 2001 at 04:40:04PM -0600, Martin McWhorter wrote:
> 
> >Try to rename /sbin/hotplug into something else _temporarily_,

Easier to just do:
	echo "/bin/true" > /proc/sys/kernel/hotplug

Then change it back later to /sbin/hotplug when you are done testing, or
just reboot :)

thanks,

greg k-h