To avoid jumps, the original version of the handler clears
the scroll flags. The method works well if a scroll gesture
continues, but it can preclude "coasting". The new version
adjusts the coordinates that determine the speed and direction
of scrolling.
ok shadchin@, mpi@
As of 1.17.2, the xserver is more picky about the order in which
headers are included. We need to account for this in order for the
compile time exa.h check to succeed.
Patch from Matthew Rezny upstreams.
As of 1.17.2, the xserver is more picky about the order in which
headers are included. We need to account for this in order for the
compile time exa.h check to succeed.
As of 1.17.2, the xserver is more picky about the order in which
headers are included. We need to account for this in order for the
compile time exa.h check to succeed.
As of 1.17.2, the xserver is more picky about the order in which
headers are included. We need to account for this in order for the
compile time exa.h check to succeed.
Problem noticed by P. Graham, diagnostic by kettenis@, fix from
Connor Behan in upstreams driver
With the changes for multitouch input, new event types have been
introduced in wsmouse. There are separate types for contact count
and width values, and RESET events for controlling/stopping
pointer motion.
ok mpi@, shadchin@
restore something properly after resume which makes the render ring get stuck.
By defaulting to the "blt" acceleration method we avoid the use of this ring
and the machine remains usable after resume.
People who really want full 2D acceleration can still override the acceleration
method in their xorg.conf.
ok jsg@
From Chris Wilson in
7a9bdadd71730adc5266bc6758982abec5917b93
Fixes rendering problems on Broadwell GT3 (Iris 6100/Iris Pro 6200)
Tested by Thomas Weinbrenner on bugs@
Before:
Both drivers currently do one read(2) per event, but since we use a
"sync" event we have a least two of them for ws(4) but more commonly
3 for ws(4) and 4 for synaptics(4).
most of the code and ok mpi@
by kettenis@ that are still needed:
Fix a problem introduced by the libpciaccess conversion where the chip ID
would be set wrong and the proper resources wouldn't be claimed.
