[Bug 33958] New: Wine vulnerability with (USB) devices

http://bugs.winehq.org/show_bug.cgi?id=33958

Bug #: 33958
           Summary: Wine vulnerability with (USB) devices
           Product: Wine
           Version: unspecified
          Platform: x86-64
        OS/Version: Mac OS X
            Status: UNCONFIRMED
          Severity: normal
          Priority: P2
         Component: -unknown
        AssignedTo: [e-mail]
        ReportedBy: [e-mail]
    Classification: Unclassified

I have a original XBOX controller and use a third-party driver (which is
officially discontinued since OS X Tiger, but was updated some time ago to be
64bit compatible. It is the only driver out there for XBOX 1 controller for OS
X so I have to use it.
This driver seems to mess up the UNIX device manager (or similar, as I don't
have any consolidated knowledge), resulting in crashing any wine process
accessing the controller as a input device.

I get the following error message (repetitive) from the console whenever the
controller is connected:

sudo[4755]:   [CURRENT_USER] : TTY=unknown ; PWD=/ ; USER=root ;
COMMAND=/usr/bin/renice [#] -p [#]

Although this issue does not seem to be related to wine itself it reveals some
vulnerability (I think).
It may be worth fixing, as this might occur this other devices, too.

Best,
Rastafabi

[Bug 33958] Wine vulnerability with (USB) devices

http://bugs.winehq.org/show_bug.cgi?id=33958

--- Comment #1 from Nikolay Sivov  2013-07-05 03:55:30 CDT ---
I don't see how a driver malfunction is related to Wine. Do you have any logs
or any idea how exactly does it crash?

[Bug 33958] Wine vulnerability with (USB) devices

http://bugs.winehq.org/show_bug.cgi?id=33958

--- Comment #2 from [e-mail] 2013-07-05 04:07:40 CDT ---
Well, I just experienced that under exactly the same conditions (expect for the
controller being connected) any wine variant I'm using expect for the
CiderEngines (e.g. Wineskin/macports) and any wine version crashes giving
different error messages (look here:
http://portingteam.com/topic/9416-wineskin-crashes-when-gaming-controller-co...).
As the device (actually not being HID compliant) needs to be added to the OS
somehow this is done through the driver (obviously). This driver seems to
provide the device in a buggy way somehow, there is the chance that an
application that tries to access it "get's confused" or, like in this case
crashes.
Older versions of wine, which do not offer controller support within OS X are
not affected. Up to know wine also is the only "application" having trouble.
the driver works fine in any OS X native or cider ported game, as well as with
any remapping solution I've tried.

[Bug 33958] Wine vulnerability with (USB) devices

http://bugs.winehq.org/show_bug.cgi?id=33958

--- Comment #3 from Rastafabi  2013-07-05 04:19:45 CDT ---
Created attachment 45109
  --> http://bugs.winehq.org/attachment.cgi?id=45109
Wine log files

[Bug 33958] Wine vulnerability with (USB) devices

http://bugs.winehq.org/show_bug.cgi?id=33958

--- Comment #4 from Rastafabi  2013-08-12 02:56:13 CDT ---
Hello there,
I just played around with this issue again and finally I could get some more
output. 
I think now it's finally somehow clear where to search for the issue.

I just post it here as it's quiet short:

fixme:d3d:init_driver_info Unable to find a driver/device info for
vendor_id=0x1002 device_id=0x71c5 for driver_model=2
fixme:win:EnumDisplayDevicesW ((null),0,0x32ef98,0x00000000), stub!
fixme:d3d:context_check_fbo_status FBO status GL_FRAMEBUFFER_UNDEFINED (0x8219)
err:d3d:context_check_fbo_status FBO 0 is incomplete, driver bug?
err:d3d:device_clear_render_targets >>>>>>>>>>>>>>>>>
GL_INVALID_FRAMEBUFFER_OPERATION (0x506) from glClear @ device.c / 672
fixme:d3d:init_driver_info Unable to find a driver/device info for
vendor_id=0x1002 device_id=0x71c5 for driver_model=2
fixme:win:EnumDisplayDevicesW ((null),0,0x32f218,0x00000000), stub!
fixme:dinput:IDirectInputDevice2WImpl_EnumEffects
(this=0x18517c,0x8aa9b0,0x149a7c8,0x00000003): stub!
fixme:dinput:IDirectInputDevice2WImpl_EnumEffects
(this=0x1b531c,0x8aa9b0,0x149a890,0x00000003): stub!
fixme:dinput:get_osx_device_elements Unhandled type 129
fixme:dinput:get_osx_device_elements Unhandled type 129
fixme:dinput:get_osx_device_elements Unhandled usage 1
fixme:dinput:get_osx_device_elements Unhandled usage 2
fixme:dinput:get_osx_device_elements Unhandled usage 3
fixme:dinput:get_osx_device_elements Unhandled usage 4
fixme:dinput:get_osx_device_elements Unhandled usage 5
fixme:dinput:get_osx_device_elements Unhandled usage 6

[Bug 33958] OS X joystick-support (partial improved)

http://bugs.winehq.org/show_bug.cgi?id=33958

Rastafabi  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |RESOLVED
          Component|-unknown                    |directx-dinput
            Version|unspecified                 |1.7.17
         Resolution|---                         |WORKSFORME
            Summary|Wine vulnerability with     |OS X joystick-support
                   |(USB) devices               |(partial improved)

--- Comment #5 from Rastafabi  ---
Finally Wine 1.7.17 solved it (partially). The issue seemingly was solely
related to the way Wine handles OS X joysticks. As recently input
handling/joystick mapping on OS X got improved, wine does not crash anymore
with the controller connected (which was the case previously - see older
posts).

Anyway some work still needs to be done, as not every button is recognized
(missing: A, B, X, Y, black, white - XBOX [original] Controller).
Furthermore wine does not handle the Z/RZ axis properly. They are treated as
independent axis while actually both represent half an axis.
- This makes acceleration (racing game) to be active while not being actually
pressed and braking being activated upon hitting "acceleration" (Z-axis). The
RZ axis is being ignored completely although it's assigned to breaking in the
system settings. This is not an issue with the game (Colin McRea Rally 2005) as
it occurs within other games, too.

I'm using my XBOX one controller with a special OS X driver
(http://xhd.sourceforge.net) in version 2.0.0
(http://xhd.cvs.sourceforge.net/viewvc/xhd/xhd/Release/) with 64bit support,
which is required by OS X Mountain Lion and greater. (I'm on 10.8.5.)
 - The driver itself does work (tested with OS X native/cider games and
confirmed with the drivers panel graphical feedback.)

[Bug 33958] OS X joystick-support (partial improved)

http://bugs.winehq.org/show_bug.cgi?id=33958

Rastafabi  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |[e-mail]

--- Comment #6 from Rastafabi  ---
*Concerning buttons I have do add that every axis basically gets recognized
(did not test using buttons as axis, yet - a XBOX [original] controller
feature) apart of the Z/RZ issue.
Start and Back work perfectly. (They are mapped as the default go
forward/backward functions).
The D-Pad works, too, but it cannot be used for navigation (or similar), as
it's only recognized as being 4 independent buttons.

[Bug 33958] OS X joystick-support (partial improved)

http://bugs.winehq.org/show_bug.cgi?id=33958

Bruno Jesus  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
         Resolution|WORKSFORME                  |FIXED

--- Comment #7 from Bruno Jesus  ---
(In reply to Rastafabi from comment #5)
> Finally Wine 1.7.17 solved it (partially). The issue seemingly was solely
> related to the way Wine handles OS X joysticks. As recently input
> handling/joystick mapping on OS X got improved, wine does not crash anymore
> with the controller connected (which was the case previously - see older
> posts).

Since this bug has a very wide subject we can split it, this one can be about
the crash that was fixed. For the remaining problems please open new bugs so
they can be better tracked.

[Bug 33958] wine programs crash with joystick connected

http://bugs.winehq.org/show_bug.cgi?id=33958

Rastafabi  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|OS X joystick-support       |wine programs crash with
                   |(partial improved)          |joystick connected

[Bug 33958] wine programs crash with joystick connected (OS X)

http://bugs.winehq.org/show_bug.cgi?id=33958

Rastafabi  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|wine programs crash with    |wine programs crash with
                   |joystick connected          |joystick connected (OS X)

[Bug 33958] wine programs crash with joystick connected (OS X)

http://bugs.winehq.org/show_bug.cgi?id=33958

--- Comment #8 from Rastafabi  ---
issue moved to bug 36035 (http://bugs.winehq.org/show_bug.cgi?id=36035)

[Bug 33958] wine programs crash with joystick connected (OS X)

http://bugs.winehq.org/show_bug.cgi?id=33958

Rastafabi  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
           See Also|                            |http://bugs.winehq.org/show
                   |                            |_bug.cgi?id=36035

[Bug 33958] wine programs crash with joystick connected (OS X)

https://bugs.winehq.org/show_bug.cgi?id=33958

Alexandre Julliard  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|RESOLVED                    |CLOSED

--- Comment #9 from Alexandre Julliard  ---
Closing bugs fixed in 1.7.18.