[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Bug#785606: linux-image-4.0.0-1-amd64: DragonRise joystick support broken in 4.0.2-1



retitle 785606 HID support broken for some devices (incl. DragonRise USB joystick)
tags 785606 upstream
notfound 785606 3.16.7-ckt9-3
thanks

On Mon, 18 May 2015 11:29:32 +0200 Florent Rougon <f.rougon@free.fr> wrote:

> The DragonRise joystick used to work fine in linux-image-3.16.0-4-amd64
> version 3.16.7-ckt9-3 and earlier versions, however its support is broken in
> linux-image-4.0.0-1-amd64 version 4.0.2-1. The problem is with the right
> stick:
>   - moving it horizontally is reflected on axis 2 (instead of 3 in previous
>     versions), but the values are mostly unusable, for instance 0 both in
>     the center and leftmost positions;
>   - moving it vertically changes axes 2 and 3 at the same time!
>
> (in previous versions, axis 2 was already bogus, but axes 0, 1 and 3-6 were
> all usable and reflected all possible movements; there is one less axis now
> [0-5], and the movements of the right stick are not reflected properly)

To put this in a simpler way, axis 3 of this joystick is just missing in
recent kernels; and it corresponds to horizontal movement of the right
stick. But I have now found which commit introduced the bug (thanks to
'git bisect'), see below.

> What puzzles me is that 'git diff v3.16..v4.0 drivers/hid/hid-dr.c' (in a
> clone of the upstream linux repo) gives no output. The change must be
> elsewhere?..

Yup, the change causing the bug belongs to drivers/hid/hid-input.c, and
was introduced in commit 79346d620e9de87912de73337f6df8b7f9a46888 of
<git://git.kernel.org/pub/scm/linux/kernel/git/stable/linux-stable.git>,
aka v3.17-rc1-94-g79346d6, dating from August 2014. Applying the reverse
patch (obtained with 'git show -R v3.17-rc1-94-g79346d6') to the current
linux-image-4.0.0-1-amd64 package in sid (version 4.0.2-1) fixes the
problem for me. Of course, it probably causes a regression for the
people who introduced this change.

So, how do we proceed from here? Should I contact the commit author
myself? TIA

-- 
Florent


Reply to: