Thanks Jesse, I have intel kernel modesetting working in some way now. It works on the initial (first) text-mode switch.

But the bug still stands where if it's in gfx mode, THEN i switch to a text console, I just get vertical lines.

I realize now that now that the text is actually there, it's just the wrong monitor frequency or something, because i see the text flash up very quickly and slantendly/magic-eye-style-repeatedly, and then just dotted vertical lines.

This is on a Fujitsu Lifebook S6510 notebook, on its built in LCD.

Does that behaviour sound like the gfx card sending the wrong LCD refresh/sync frequency to you?

tom.

On Mon, Nov 17, 2008 at 3:59 AM, Jesse Barnes <jbarnes@virtuousgeek.org> wrote:
I just pushed some repos to kernel.org:
git://git.kernel.org/pub/scm/linux/kernel/git/jbarnes/drm-2.6.git
 drm-modesetting branch contains the latest bits, rebased to drm-intel-next
git://git.kernel.org/pub/scm/linux/kernel/git/jbarnes/libdrm.git
 modesetting-gem branch contains the latest headers & libdrm (and a wip test
 program, modetest)

I'm not sure about the problem you're seeing; I'm pretty sure the interrupt
stuff still needs some work when modeset=1 (and possibly when modeset=1), but
my main test machine is a 965GM (Thinkpad t61) and I'm able to modprobe the
driver with modeset=1 ok.

One thing you might run into if you build the driver in is the load order wrt
the AGP driver; the latter must load first for everything to work, and in the
builtin case I'm not sure if that happens.

Jesse

On Saturday, November 15, 2008 8:19 pm tom jacobs wrote:
> Thanks Dan for the quick work on this one!
> They patched in fine and compiled, but when running that kernel,
> trying to switch to a text-mode console resulted in just a flashing
> vertical stripe motif screen, and every second boot fails with just
> jibberish colours flashing on the screen, so i'm unable to try out the
> modeset=1 flag. Although I did get a nice small-text text-mode mode
> switch on bootup before gnome started.
>
> My card is "Intel Corporation Mobile GM965/GL960 Integrated Graphics
> Controller (rev 03)", do you know if this works with the 915
> modesetting patches?
>
> Can either of you take a guess at what's going on with this? is it a
> vblank issue?
>
> tom.
>
> On Nov 16, 9:10 am, Dan Ranger <dranger...@gmail.com> wrote:
> > Ok, I updated jbarnes kms-7 patch set to work with the latest commit
> > from anholt's repo.
> > Patches are available here ->http://sites.google.com/site/dranger003/Home
> >
> > At least now I have KMS working and can move on.
> >
> > Dan
> >
> > On Nov 15, 10:48 am, Dan Ranger <dranger...@gmail.com> wrote:
> > > Tom, there has been new commits to anholt's drm-intel-next repo and
> > > jbarnes kms-7 patches are out of sync now.
> > > You will need commit 2db46b5668dfed59acacd75dfc49ecf947ea5aeb to apply
> > > those patches as I mentionned earlier.
> > >
> > > Otherwise, the latest commits from jbarnes to anholt's drm-intel-next
> > > added GEM mmap support and GEM GTT mapping support (which refers to
> > > the first 2 patches from jbarnes kms-7 patch set I believe).
> > >
> > > jbarnes remaining patches 3 to 5 from kms-7 currently don't apply
> > > without modifications (I suspect minor modifications) on last drm-
> > > intel-next commit.
> > >
> > > Dan
> > >
> > > On Nov 15, 3:30 am, "tom jacobs" <hatsma...@gmail.com> wrote:
> > > > Hi Jesse - I'm looking to use your fantastic modesetting kernel --
> > > > but I can't get a hold of it.
> > > > I tried patching using your klm-7 directory of patches onto
> > > > drm-intel-next but they're out of sync.
> > > >
> > > > Can you please push a git repo of it to freedesktop so people can
> > > > pull it down easily?
> > > >
> > > > thanks
> > > > Tom.

--
Jesse Barnes, Intel Open Source Technology Center




--
Tom.