On Tue, Aug 01, 2006 at 12:27:43PM +1000, Brendon Higgins wrote:
> Package: xserver-xorg-input-synaptics
> Version: 0.14.6-1
> Severity: normal
>
> This presented itself to me originally as a problem with pbbuttonsd (see bug
> #377447), as that daemon seemed to not notice when I moved the trackpad, so it
> would suspend the machine even while I was using it. The maintainer of
> pbbuttonsd claims this is a known bug in the synaptics driver.
>
> The cause (apparently) is that thhe synaptics driver opens the event device
> exclusively, so pbbuttonsd can't read any events from it. This would certainly
> cause the problem. Apparently there's a patch floating around, too (no idea
> where, though. I wouldn't think opening an event device exclusively would be
> standard behaviour.
it's not because of an exclusive open, instead because of this (eventcomm.c:46):
/* Try to grab the event device so that data don't leak to /dev/input/mice
*/
int ret;
SYSCALL(ret = ioctl(local->fd, EVIOCGRAB, (pointer)1));
and this is necessary to avoid doubling mouse events to /dev/input/mice[1] which
essentially would cause also a doubled reaction to such events (clicks,
movements).
The patch that is floating around is probably here[2] but this includes
adding a new ioctl to the kernel (read: patch and rebuild the kernel)
and this is the reason the synaptics driver doesn't use it (and won't
until that ioctl is available in stock kernels).
Maybe an alternative solution could be polling some X status parameter?
Well I'm not an X guru so I could just be saying something stupid...
thanks
[1]: /dev/input/mouse reports events from _any_ input device and this
works great until you have only one of such devices.
[2]: http://www.artha.org/shammash/ppc/adb_syn/
--
mattia
:wq!
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
