Skip site navigation (1)Skip section navigation (2)
Date:      Fri, 05 Dec 2014 01:34:50 +0100
From:      Jan Beich <jbeich@vfemail.net>
To:        =?utf-8?Q?Jean-S=C3=A9bastien_P=C3=A9dron?= <jean-sebastien.pedron@dumbbell.fr>
Cc:        freebsd-x11@freebsd.org
Subject:   Re: [Call for testers] Update to xorg-server 1.14
Message-ID:  <oarj-kk7p-wny@vfemail.net>
References:  <547E4B03.2090508@FreeBSD.org> <lhmo-8tv3-wny@vfemail.net> <5480D3BA.8090104@dumbbell.fr>

next in thread | previous in thread | raw e-mail | index | archive | help
Jean-S=C3=A9bastien P=C3=A9dron <jean-sebastien.pedron@dumbbell.fr> writes:

>> DEVD doesn't seem to support devices provided by multimedia/webcamd.
>> For one, xf86-input-wacom uses HAL to simplify configuration of
>> multi-interface tablets: more than one /dev/input/eventN that may
>> change order during reboot or replug.
>
> What prevents you from putting the configuration in standard X
> configuration file, using the InputClass section and Match* directives?

webcamd doesn't pass enough data via devctl(4) to associate input/eventN
with any X11 driver or even any ugenX.Y. Not to mention '+' events never
happen for ugen(4) devices. Here's an example for CTH-470:

  !system=3DUSB subsystem=3DINTERFACE type=3DATTACH ugen=3Dugen2.3 cdev=3Du=
gen2.3 vendor=3D0x056a product=3D0x00de devclass=3D0x00 devsubclass=3D0x00 =
sernum=3D"" release=3D0x0100 mode=3Dhost interface=3D0 endpoints=3D1 intcla=
ss=3D0x03 intsubclass=3D0x01 intprotocol=3D0x02
  !system=3DUSB subsystem=3DINTERFACE type=3DATTACH ugen=3Dugen2.3 cdev=3Du=
gen2.3 vendor=3D0x056a product=3D0x00de devclass=3D0x00 devsubclass=3D0x00 =
sernum=3D"" release=3D0x0100 mode=3Dhost interface=3D1 endpoints=3D1 intcla=
ss=3D0x03 intsubclass=3D0x00 intprotocol=3D0x00
  [...]
  !system=3DDEVFS subsystem=3DCDEV type=3DCREATE cdev=3Dinput/event0
  !system=3DDEVFS subsystem=3DCDEV type=3DCREATE cdev=3Dinput/event1

MatchDevicePath "/dev/input/event*" may still work except it doesn't.
config/devd only sees fake ums(4) and uhid(4) devices[1] exposed by the
tablet and while trying to attach to them fails to produce correct
Identifier thanks to a comma.

  [117006.123] (II) config/devd: adding input device Co. (/dev/uhid1)
  [117006.123] (II) Using input driver 'mouse' for 'Co.'
  [117006.123] (**) Co.: always reports core events

from

  dev.uhid.1.%desc: Wacom Co.,Ltd. CTH-470, class 0/0, rev 2.00/1.00, addr 3

  iManufacturer =3D 0x0001  <Wacom Co.,Ltd.>
  iProduct =3D 0x0002  <CTH-470>

WRKSRC/conf/50-wacom.conf of xf86-input-wacom is an example config.
I've tried to remove everything but the first InputClass rule
and drop MatchUSBID with no luck.

>
>> At least ports/183478 implementation could be extended with
>> etc/devd/*.conf if there're out-of-tree kernel or cuse(3) drivers.
>
> Points 1 to 4, and point 6 of PR 183478 are addressed in this update.
>
> Missing USB ID parsing is a known issue and on the TODO. Point 5 about
> incorrectly parsed vendor and product properties need to be addressed.

Would that help with hw_types whitelist of allowed drivers?

--
[1] they can be disabled via UQ_UMS_IGNORE and UQ_HID_IGNORE

-------------------------------------------------

VFEmail.net - http://www.vfemail.net
ONLY AT VFEmail! - Use our Metadata Mitigator to keep your email out of the NSA's hands!
$24.95 ONETIME Lifetime accounts with Privacy Features!  
15GB disk! No bandwidth quotas!
Commercial and Bulk Mail Options!  



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?oarj-kk7p-wny>