svn commit: r194540 - head/sys/dev/drm
rnoland at FreeBSD.org
Sat Jun 20 19:24:56 UTC 2009
On Sat, 2009-06-20 at 12:56 -0400, Daniel Eischen wrote:
> On Sat, 20 Jun 2009, Robert Noland wrote:
> > Author: rnoland
> > Date: Sat Jun 20 16:45:14 2009
> > New Revision: 194540
> > URL: http://svn.freebsd.org/changeset/base/194540
> > Log:
> > The G45 docs indicate that all G4X chips use the new framecount register.
> > Intel agrees with my reading of the docs, make it so for all G4X chips.
> > The new register also has a 32 bit width as opposed to 24 bits. Fix
> > things up so that the counters roll over properly.
> Could this possibly fix the problem I'm seeing with the screen being
> garbage after a logout from KDE 3.x (using kdm)? Everything works
> fine after logging in, but when you log out, the screen is left with
> garbage/lots of vertical striping. This only happened after upgrading
> my system (and all ports) to Xorg 7.4, worked just fine before that.
> agp0: <Intel 82945GM (945GM GMCH) SVGA controller> on vgapci0
> agp0: detected 7932k stolen memory
> agp0: aperture size is 256M
> vgapci1: <VGA-compatible display> mem 0xdff80000-0xdfffffff at device 2.1 on pci0
No, this only change only effects g45 chips, which would have been
incorrectly handling the hardware vblank counter. It will have no
impact on 945. I'm trying to pull changes that are definite wins from
my Intel branch right now. I don't think that I have seen the symptoms
that you describe on my 945, except when it is resuming from suspend. I
do only run gnome on it though, it is far too slow to do kde as well...
I'm trying to fix up the interrupt handling in a less invasive way than
I did previously, but even that should only effect 965+ with msi
Robert Noland <rnoland at FreeBSD.org>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Size: 196 bytes
Desc: This is a digitally signed message part
Url : http://lists.freebsd.org/pipermail/svn-src-head/attachments/20090620/c53dd2a9/attachment.pgp
More information about the svn-src-head