2.6.39.1 input regression

2.6.39.1 input regression

[Ortwin Glück]

Hi Greg,

This patch in 2.6.39.1 breaks magicmouse support for me completely.
a5ccaac64c17ce3cc1878a6d479a6f1b16092c24
HID: magicmouse: ignore 'ivalid report id' while switching modes

The device will no longer work at all:
magicmouse 0005:05AC:030D.0005: input,hidraw2: BLUETOOTH HID v0.84 Mouse
[gandalf mouse] on 7C:6D:62:9F:DB:2D
magicmouse 0005:05AC:030D.0005: unable to request touch data (2)

Fine in previous kernels.

Thanks.

Ortwin

Re: 2.6.39.1 input regression

[Greg KH]

On Wed, Jun 15, 2011 at 09:16:47AM +0200, Ortwin Glück wrote:
> Hi Greg,
> 
> This patch in 2.6.39.1 breaks magicmouse support for me completely.
> a5ccaac64c17ce3cc1878a6d479a6f1b16092c24
> HID: magicmouse: ignore 'ivalid report id' while switching modes
> 
> The device will no longer work at all:
> magicmouse 0005:05AC:030D.0005: input,hidraw2: BLUETOOTH HID v0.84 Mouse
> [gandalf mouse] on 7C:6D:62:9F:DB:2D
> magicmouse 0005:05AC:030D.0005: unable to request touch data (2)
> 
> Fine in previous kernels.

This has been discussed already, and it should be fixed in Linus's tree
right now, right?

If so, can someone please send stable@kernel.org the proper fix for
this?

thanks,

greg k-h

Re: 2.6.39.1 input regression

[Ortwin Glück]

On 15.06.2011 17:03, Greg KH wrote:
> On Wed, Jun 15, 2011 at 09:16:47AM +0200, Ortwin Glück wrote:
>> Hi Greg,
>>
>> This patch in 2.6.39.1 breaks magicmouse support for me completely.
>> a5ccaac64c17ce3cc1878a6d479a6f1b16092c24
>> HID: magicmouse: ignore 'ivalid report id' while switching modes
>>
>> The device will no longer work at all:
>> magicmouse 0005:05AC:030D.0005: input,hidraw2: BLUETOOTH HID v0.84 Mouse
>> [gandalf mouse] on 7C:6D:62:9F:DB:2D
>> magicmouse 0005:05AC:030D.0005: unable to request touch data (2)
>>
>> Fine in previous kernels.
> 
> This has been discussed already, and it should be fixed in Linus's tree
> right now, right?
> 
> If so, can someone please send stable@kernel.org the proper fix for
> this?
> 
> thanks,
> 
> greg k-h

If it's in Linus' tree then I can't find it. To me this issue just looks
like a silly typo anyway:

-	if (ret != sizeof(feature)) {
+	if (ret != -EIO) {
 		hid_err(hdev, "unable to request touch data (%d)\n", ret);
 		goto err_stop_hw;
 	}

Should this not be (ret == -EIO)?

Ortwin

Re: 2.6.39.1 input regression

[Jiri Kosina]

On Wed, 15 Jun 2011, Greg KH wrote:

> > This patch in 2.6.39.1 breaks magicmouse support for me completely.
> > a5ccaac64c17ce3cc1878a6d479a6f1b16092c24
> > HID: magicmouse: ignore 'ivalid report id' while switching modes
> > 
> > The device will no longer work at all:
> > magicmouse 0005:05AC:030D.0005: input,hidraw2: BLUETOOTH HID v0.84 Mouse
> > [gandalf mouse] on 7C:6D:62:9F:DB:2D
> > magicmouse 0005:05AC:030D.0005: unable to request touch data (2)
> > 
> > Fine in previous kernels.
> 
> This has been discussed already, and it should be fixed in Linus's tree
> right now, right?

It's not yet -- apparently the symptoms aer different for different 
people, probably because of differences in firmware.

Plus one of the testers is now quite busy, so the turnaround was rather 
slow.

I already have patch in my tree that should cover all the cases. Will be 
sending pull request to Linus likely today.

> If so, can someone please send stable@kernel.org the proper fix for 
> this?

Yes, I will then send the patch stable-way as well.

-- 
Jiri Kosina
SUSE Labs