Re: 2.6.28-rc9: USB Mouse unhappy

Re: 2.6.28-rc9: USB Mouse unhappy

[Andrew Morton]

(cc's added)

On Sat, 20 Dec 2008 22:11:17 -0500 "Parag Warudkar" <parag.lkml-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> wrote:

> Just installed today's git on my workstation and my Logitech USB Mouse
> only wants to move vertically around the centre of the monitor.
> 
> I have no clue where to look - all  I know is that it works with the
> 2.6.27 and various 2.6.28-rcs prior to this one.
> 
> Parag
> 
> [parag@xw6600 ~]$ diff Xorg.0.log.not.working /var/log/Xorg.0.log |grep Logi
> < (**) Logitech USB Receiver: Device: "/dev/input/event5"
> > (**) Logitech USB Receiver: Device: "/dev/input/event6"
> < (II) config/hal: Adding input device Logitech USB Receiver
> < (**) Logitech USB Receiver: always reports core events
> < (**) Logitech USB Receiver: Device: "/dev/input/event3"
> < (II) Logitech USB Receiver: Found x and y relative axes
> < (II) Logitech USB Receiver: Found mouse buttons
> < (II) Logitech USB Receiver: Configuring as mouse
> < (II) XINPUT: Adding extended input device "Logitech USB Receiver"
> (type: MOUSE)
> > (II) config/hal: Adding input device Logitech USB Receiver
> > (**) Logitech USB Receiver: always reports core events
> > (**) Logitech USB Receiver: Device: "/dev/input/event5"
> > (II) Logitech USB Receiver: Found x and y relative axes
> > (II) Logitech USB Receiver: Found mouse buttons
> > (II) Logitech USB Receiver: Configuring as mouse
> > (II) XINPUT: Adding extended input device "Logitech USB Receiver" (type: MOUSE)

Could be input, could be HID, could even be USB.

You're sure that 2.6.28-rc8 was OK?

2.6.28-rc9 had only a small number of changes.  Nothing I can see in
input or HID.  A few USB changes, but pretty specific ones -
unusual_devs entries and such.

Maybe your mouse has died?  Did you retest with an older kernel?
--
To unsubscribe from this list: send the line "unsubscribe linux-usb" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: 2.6.28-rc9: USB Mouse unhappy

[Parag Warudkar]

On Sun, Dec 21, 2008 at 1:12 AM, Andrew Morton
<akpm@linux-foundation.org> wrote:
> You're sure that 2.6.28-rc8 was OK?

Nope - not sure about 28-rc8. I meant to say some previous 28-rc's
that I tested worked fine - can't recall which ones. it is likely I
missed testing .28-rc8.

>
> 2.6.28-rc9 had only a small number of changes.  Nothing I can see in
> input or HID.  A few USB changes, but pretty specific ones -
> unusual_devs entries and such.
>
> Maybe your mouse has died?  Did you retest with an older kernel?
>
I did - rebooting into 2.6.27 (original distro kernel) makes it work again.

Thanks

Parag

Re: 2.6.28-rc9: USB Mouse unhappy

[Anssi Hannula]

Parag Warudkar wrote:
> On Sun, Dec 21, 2008 at 1:12 AM, Andrew Morton
> <akpm@linux-foundation.org> wrote:
>> You're sure that 2.6.28-rc8 was OK?
> 
> Nope - not sure about 28-rc8. I meant to say some previous 28-rc's
> that I tested worked fine - can't recall which ones. it is likely I
> missed testing .28-rc8.
> 
>> 2.6.28-rc9 had only a small number of changes.  Nothing I can see in
>> input or HID.  A few USB changes, but pretty specific ones -
>> unusual_devs entries and such.
>>
>> Maybe your mouse has died?  Did you retest with an older kernel?
>>
> I did - rebooting into 2.6.27 (original distro kernel) makes it work again.

This sounds to me like 'usbmouse' gets used instead of 'usbhid' on
2.6.28-rc9. Many users have reported only vertical movement with
'usbmouse'. Better check the kernel configuration, if you had usbmouse
disabled in 2.6.27, then disable it with 2.6.28-rc9 as well.

-- 
Anssi Hannula

Re: 2.6.28-rc9: USB Mouse unhappy

[Parag Warudkar]

On Sun, Dec 21, 2008 at 5:56 AM, Anssi Hannula <anssi.hannula@gmail.com> wrote:
>
> This sounds to me like 'usbmouse' gets used instead of 'usbhid' on
> 2.6.28-rc9. Many users have reported only vertical movement with
> 'usbmouse'. Better check the kernel configuration, if you had usbmouse
> disabled in 2.6.27, then disable it with 2.6.28-rc9 as well.

Sure enough - rmmod usbhid; modprobe usbmouse reproduces the exact
same problem on a working kernel.

But the thing is I use the same base configuration across all kernels
I build (sans of course the newly added config options) - so something
changed somewhere and now usbmouse gets used instead of usbhid.

What's the deal with usbmouse though - is it supposed to be working or
is it considered lost cause in favor of usbhid? May be if people don't
need to use it and it is not actively looked into, we can
disable/remove it?

Thanks

Parag

Re: 2.6.28-rc9: USB Mouse unhappy

[Jiri Kosina]

On Sun, 21 Dec 2008, Parag Warudkar wrote:

> > This sounds to me like 'usbmouse' gets used instead of 'usbhid' on
> > 2.6.28-rc9. Many users have reported only vertical movement with
> > 'usbmouse'. Better check the kernel configuration, if you had usbmouse
> > disabled in 2.6.27, then disable it with 2.6.28-rc9 as well.
> Sure enough - rmmod usbhid; modprobe usbmouse reproduces the exact same 
> problem on a working kernel.
> But the thing is I use the same base configuration across all kernels
> I build (sans of course the newly added config options) - so something
> changed somewhere and now usbmouse gets used instead of usbhid.

Could you please

- make sure whether usbhid is used on 2.6.27 to drive the mouse and 
  usbmouse on 2.6.28-rc with the exact same mouse? (either check 
  /sys/bus/usb/devices or even lsusb shouldn't ideally show usbmouse 
  loaded at all)
- if so, make sure that you really need to build usbmouse (please read the 
  help text in Kconfig for explanation)

> What's the deal with usbmouse though - is it supposed to be working or 
> is it considered lost cause in favor of usbhid? May be if people don't 
> need to use it and it is not actively looked into, we can disable/remove 
> it?

Really, please read the help text.

Maybe we should make it dependent on EMBEDDED.

Thanks,

-- 
Jiri Kosina
SUSE Labs

Re: 2.6.28-rc9: USB Mouse unhappy

[Jiri Kosina]

On Sun, 21 Dec 2008, Jiri Kosina wrote:

>   usbmouse on 2.6.28-rc with the exact same mouse? (either check 
>   /sys/bus/usb/devices or even lsusb shouldn't ideally show usbmouse 
>   loaded at all)

Sorry for the above typo, 'lsusb' should of course be 'lsmod' here.

-- 
Jiri Kosina
SUSE Labs

--
To unsubscribe from this list: send the line "unsubscribe linux-usb" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

[PATCH] USBHID: Make Boot Protocol drivers depend on EMBEDDED

[Parag Warudkar]

The usbmouse and usbkbd modules are not supposed to be used with regular 
USB mice and keyboards. Make them depend on EMBEDDED to prevent them from 
being built and loaded on non-EMBEDDED configs.

Signed-off-by: Parag Warudkar <parag.lkml-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>

diff --git a/drivers/hid/usbhid/Kconfig b/drivers/hid/usbhid/Kconfig
index 5d9aa95..4edb3be 100644
--- a/drivers/hid/usbhid/Kconfig
+++ b/drivers/hid/usbhid/Kconfig
@@ -45,7 +45,7 @@ config USB_HIDDEV
 	  If unsure, say Y.
 
 menu "USB HID Boot Protocol drivers"
-	depends on USB!=n && USB_HID!=y
+	depends on USB!=n && USB_HID!=y && EMBEDDED
 
 config USB_KBD
 	tristate "USB HIDBP Keyboard (simple Boot) support"
--
To unsubscribe from this list: send the line "unsubscribe linux-usb" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH] USBHID: Make Boot Protocol drivers depend on EMBEDDED

[Jiri Kosina]

On Sun, 21 Dec 2008, Parag Warudkar wrote:

> The usbmouse and usbkbd modules are not supposed to be used with regular 
> USB mice and keyboards. Make them depend on EMBEDDED to prevent them from 
> being built and loaded on non-EMBEDDED configs.
> 
> Signed-off-by: Parag Warudkar <parag.lkml-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
> 
> diff --git a/drivers/hid/usbhid/Kconfig b/drivers/hid/usbhid/Kconfig
> index 5d9aa95..4edb3be 100644
> --- a/drivers/hid/usbhid/Kconfig
> +++ b/drivers/hid/usbhid/Kconfig
> @@ -45,7 +45,7 @@ config USB_HIDDEV
>  	  If unsure, say Y.
>  
>  menu "USB HID Boot Protocol drivers"
> -	depends on USB!=n && USB_HID!=y
> +	depends on USB!=n && USB_HID!=y && EMBEDDED
>  
>  config USB_KBD
>  	tristate "USB HIDBP Keyboard (simple Boot) support"

So do I get it right that the only reason of the problem you were seeing 
was that in the 2.6.27 kernel you have compiled usbhid module, which was 
then used for driving your mouse,but in 2.6.28-rcX you compiled and used 
'usbmouse' instead, and therefore this is not a regression?

Thanks,

-- 
Jiri Kosina
SUSE Labs
--
To unsubscribe from this list: send the line "unsubscribe linux-usb" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH] USBHID: Make Boot Protocol drivers depend on EMBEDDED

[Andrew Morton]

On Sun, 21 Dec 2008 22:48:21 +0100 (CET) Jiri Kosina <jkosina-AlSwsSmVLrQ@public.gmane.org> wrote:

> On Sun, 21 Dec 2008, Parag Warudkar wrote:
> 
> > The usbmouse and usbkbd modules are not supposed to be used with regular 
> > USB mice and keyboards. Make them depend on EMBEDDED to prevent them from 
> > being built and loaded on non-EMBEDDED configs.
> > 
> > Signed-off-by: Parag Warudkar <parag.lkml-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
> > 
> > diff --git a/drivers/hid/usbhid/Kconfig b/drivers/hid/usbhid/Kconfig
> > index 5d9aa95..4edb3be 100644
> > --- a/drivers/hid/usbhid/Kconfig
> > +++ b/drivers/hid/usbhid/Kconfig
> > @@ -45,7 +45,7 @@ config USB_HIDDEV
> >  	  If unsure, say Y.
> >  
> >  menu "USB HID Boot Protocol drivers"
> > -	depends on USB!=n && USB_HID!=y
> > +	depends on USB!=n && USB_HID!=y && EMBEDDED
> >  
> >  config USB_KBD
> >  	tristate "USB HIDBP Keyboard (simple Boot) support"
> 
> So do I get it right that the only reason of the problem you were seeing 
> was that in the 2.6.27 kernel you have compiled usbhid module, which was 
> then used for driving your mouse,but in 2.6.28-rcX you compiled and used 
> 'usbmouse' instead, and therefore this is not a regression?
> 

Shouldn't usbmouse support two-dimensional mice as well?
--
To unsubscribe from this list: send the line "unsubscribe linux-usb" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH] USBHID: Make Boot Protocol drivers depend on EMBEDDED

[Jiri Kosina]

On Sun, 21 Dec 2008, Andrew Morton wrote:

> > So do I get it right that the only reason of the problem you were 
> > seeing was that in the 2.6.27 kernel you have compiled usbhid module, 
> > which was then used for driving your mouse,but in 2.6.28-rcX you 
> > compiled and used 'usbmouse' instead, and therefore this is not a 
> > regression?
> Shouldn't usbmouse support two-dimensional mice as well?

Yup, but that's a separate issue on my TODO. I first want to know whether 
this is a regression or not.

Thanks,

-- 
Jiri Kosina
SUSE Labs

Re: [PATCH] USBHID: Make Boot Protocol drivers depend on EMBEDDED

[Parag Warudkar]

On Sun, 21 Dec 2008, Jiri Kosina wrote:

> On Sun, 21 Dec 2008, Andrew Morton wrote:
> 
> > > So do I get it right that the only reason of the problem you were 
> > > seeing was that in the 2.6.27 kernel you have compiled usbhid module, 
> > > which was then used for driving your mouse,but in 2.6.28-rcX you 
> > > compiled and used 'usbmouse' instead, and therefore this is not a 
> > > regression?
> > Shouldn't usbmouse support two-dimensional mice as well?
> 
> Yup, but that's a separate issue on my TODO. I first want to know whether 
> this is a regression or not.
> 

I have always compiled usbmouse as a module - the problem or regression 
here was that it was loaded instead of usbhid when both were compiled as 
modules - but I don't think this is a kernel regression, here is why -

The problem happened with Fedora 10 - when I booted into Ubuntu 8.10 with 
the same kernel compiled with same config, usbhid was correctly loaded and 
mouse worked. So I would forget this as a Fedora specific issue.

Thanks for the pointers.

Parag
--
To unsubscribe from this list: send the line "unsubscribe linux-usb" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH] USBHID: Make Boot Protocol drivers depend on EMBEDDED

[Anssi Hannula]

Parag Warudkar wrote:
> 
> On Sun, 21 Dec 2008, Jiri Kosina wrote:
> 
>> On Sun, 21 Dec 2008, Andrew Morton wrote:
>>
>>>> So do I get it right that the only reason of the problem you were 
>>>> seeing was that in the 2.6.27 kernel you have compiled usbhid module, 
>>>> which was then used for driving your mouse,but in 2.6.28-rcX you 
>>>> compiled and used 'usbmouse' instead, and therefore this is not a 
>>>> regression?
>>> Shouldn't usbmouse support two-dimensional mice as well?
>> Yup, but that's a separate issue on my TODO. I first want to know whether 
>> this is a regression or not.
>>
> 
> I have always compiled usbmouse as a module - the problem or regression 
> here was that it was loaded instead of usbhid when both were compiled as 
> modules - but I don't think this is a kernel regression, here is why -
> 
> The problem happened with Fedora 10 - when I booted into Ubuntu 8.10 with 
> the same kernel compiled with same config, usbhid was correctly loaded and 
> mouse worked. So I would forget this as a Fedora specific issue.

It may well be that Ubuntu has usbmouse blacklisted or employs udev
rules that force usbhid to be loaded before usbmouse.

On Mandriva it was handled in a similar way, before usbmouse was
disabled from the kernel (no reason to have it in a "normal" kernel).

> Thanks for the pointers.

-- 
Anssi Hannula

Re: [PATCH] USBHID: Make Boot Protocol drivers depend on EMBEDDED

[Parag Warudkar]

On Mon, 22 Dec 2008, Anssi Hannula wrote:

> It may well be that Ubuntu has usbmouse blacklisted or employs udev
> rules that force usbhid to be loaded before usbmouse.
> 
Right - Ubuntu blacklists it and it works. Fedora doesn't blacklist it but 
it is not a problem with the distro shipped kernel - somehow. The problem 
happens only when using vanilla kernel on Fedora - not sure what's going 
on there. 

> On Mandriva it was handled in a similar way, before usbmouse was
> disabled from the kernel (no reason to have it in a "normal" kernel).
> 

Yep, and so the patch to make it dependent on EMBEDDED still makes sense.

Parag

Re: [PATCH] USBHID: Make Boot Protocol drivers depend on EMBEDDED

[Jiri Kosina]

On Sun, 21 Dec 2008, Parag Warudkar wrote:

> > It may well be that Ubuntu has usbmouse blacklisted or employs udev 
> > rules that force usbhid to be loaded before usbmouse.
> Right - Ubuntu blacklists it and it works. Fedora doesn't blacklist it but 
> it is not a problem with the distro shipped kernel - somehow. The problem 
> happens only when using vanilla kernel on Fedora - not sure what's going 
> on there. 

Do they even compile usbmouse and usbkbd?

-- 
Jiri Kosina
SUSE Labs

Re: [PATCH] USBHID: Make Boot Protocol drivers depend on EMBEDDED

[Parag Warudkar]

On Sun, 21 Dec 2008, Jiri Kosina wrote:

> Do they even compile usbmouse and usbkbd?

Turns out they don't (and I thought all distros ship make allmodconfig 
kernels for some reason) - now I will be puzzled for ever as to how did those pre rc-9 
kernels worked without usbmouse being blacklisted. Might be config, might 
be timing, dunno.

Parag

Re: [PATCH] USBHID: Make Boot Protocol drivers depend on EMBEDDED

[Jiri Kosina]

On Sun, 21 Dec 2008, Parag Warudkar wrote:

> The usbmouse and usbkbd modules are not supposed to be used with regular 
> USB mice and keyboards. Make them depend on EMBEDDED to prevent them from 
> being built and loaded on non-EMBEDDED configs.
> Signed-off-by: Parag Warudkar <parag.lkml-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
> 
> diff --git a/drivers/hid/usbhid/Kconfig b/drivers/hid/usbhid/Kconfig
> index 5d9aa95..4edb3be 100644
> --- a/drivers/hid/usbhid/Kconfig
> +++ b/drivers/hid/usbhid/Kconfig
> @@ -45,7 +45,7 @@ config USB_HIDDEV
>  	  If unsure, say Y.
>  
>  menu "USB HID Boot Protocol drivers"
> -	depends on USB!=n && USB_HID!=y
> +	depends on USB!=n && USB_HID!=y && EMBEDDED
>  
>  config USB_KBD
>  	tristate "USB HIDBP Keyboard (simple Boot) support"

Applied, thanks Parag.

-- 
Jiri Kosina
SUSE Labs
--
To unsubscribe from this list: send the line "unsubscribe linux-usb" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: 2.6.28-rc9: USB Mouse unhappy

[Rafael J. Wysocki]

Hi,

On Sunday 21 December 2008, Jiri Kosina wrote:
> On Sun, 21 Dec 2008, Jiri Kosina wrote:
> 
> >   usbmouse on 2.6.28-rc with the exact same mouse? (either check 
> >   /sys/bus/usb/devices or even lsusb shouldn't ideally show usbmouse 
> >   loaded at all)
> 
> Sorry for the above typo, 'lsusb' should of course be 'lsmod' here.

Can you tell me what the status of this issue is, please?

Rafael
--
To unsubscribe from this list: send the line "unsubscribe linux-usb" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: 2.6.28-rc9: USB Mouse unhappy

[Parag Warudkar]

On Sat, 10 Jan 2009, Rafael J. Wysocki wrote:

> Hi,
>
> On Sunday 21 December 2008, Jiri Kosina wrote:
>> On Sun, 21 Dec 2008, Jiri Kosina wrote:
>>
>>>   usbmouse on 2.6.28-rc with the exact same mouse? (either check
>>>   /sys/bus/usb/devices or even lsusb shouldn't ideally show usbmouse
>>>   loaded at all)
>>
>> Sorry for the above typo, 'lsusb' should of course be 'lsmod' here.
>
> Can you tell me what the status of this issue is, please?
>

Wasn't a regression - for some reason USB Boot Protocol drivers for 
mouse and kbd were compiled and loaded - I sent a patch to Jiri making 
those depend on EMBEDDED - so no more an issue.

Thanks
Parag

Re: 2.6.28-rc9: USB Mouse unhappy

[Jiri Kosina]

On Sat, 10 Jan 2009, Rafael J. Wysocki wrote:

> > >   usbmouse on 2.6.28-rc with the exact same mouse? (either check 
> > >   /sys/bus/usb/devices or even lsusb shouldn't ideally show usbmouse 
> > >   loaded at all)
> > Sorry for the above typo, 'lsusb' should of course be 'lsmod' here.
> Can you tell me what the status of this issue is, please?

Parag was using incorrect driver -- usbmouse -- (which is intended to be 
used only under very special circumstances) instead of standard HID 
driver.

Kconfig help text clearly states how special the usbmouse driver is, but 
many people still try to use it, probably because of its name. Therefore 
we now have

Author: Parag Warudkar <parag.lkml-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
Date:   Mon Dec 22 22:50:52 2008 +0100

    HID: make boot protocol drivers depend on EMBEDDED

which is already in Linus tree and which makes the possibility of chosing 
this special driver dependent on CONFIG_EMBEDDED.

-- 
Jiri Kosina
SUSE Labs
--
To unsubscribe from this list: send the line "unsubscribe linux-usb" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: 2.6.28-rc9: USB Mouse unhappy

[Rafael J. Wysocki]

On Saturday 10 January 2009, Jiri Kosina wrote:
> On Sat, 10 Jan 2009, Rafael J. Wysocki wrote:
> 
> > > >   usbmouse on 2.6.28-rc with the exact same mouse? (either check 
> > > >   /sys/bus/usb/devices or even lsusb shouldn't ideally show usbmouse 
> > > >   loaded at all)
> > > Sorry for the above typo, 'lsusb' should of course be 'lsmod' here.
> > Can you tell me what the status of this issue is, please?
> 
> Parag was using incorrect driver -- usbmouse -- (which is intended to be 
> used only under very special circumstances) instead of standard HID 
> driver.
> 
> Kconfig help text clearly states how special the usbmouse driver is, but 
> many people still try to use it, probably because of its name. Therefore 
> we now have
> 
> Author: Parag Warudkar <parag.lkml@gmail.com>
> Date:   Mon Dec 22 22:50:52 2008 +0100
> 
>     HID: make boot protocol drivers depend on EMBEDDED
> 
> which is already in Linus tree and which makes the possibility of chosing 
> this special driver dependent on CONFIG_EMBEDDED.

OK, thanks!

Rafael