Intel i830M behind transparent AGP bridge?

Oliver Fromme olli at lurza.secnetix.de
Wed Aug 2 08:38:17 UTC 2006


John Baldwin wrote:
 > Andreas Kohn wrote:
 > > a friend of mine has a laptop which has an Intel i830M chipset.
 > > Linux recognizes the chip, and connects it to its agpgart.
 > > FreeBSD* also recognizes the chip, but see it as connected to the pci
 > > bus. This makes the drm parts fail to attach, as they require
 > > (according to the code) the i830 to be attached as AGP chip.
 > > 
 > > Is this a known problem, and even more, is there any known work-around
 > > to achieve working drm with i830M in this setup?
 > 
 > It should work fine in current, and possibly on 6.1-stable.  You need 'device 
 > agp' in your kernel, and I think the patches in -stable as with drmsub the 
 > idea is that agp0 attaches to the PCI device and creates a drmsub0 child, and 
 > then the drm driver attaches to drmsub0.

I can confirm that last week's RELENG_6 works perfectly
fine with the i915, which is similar enough to the i830
and supported by the same drivers, AFAIK.

The only problem is that X11's hardware cursor seems to
occupy the resources of the brightness-icon.  When I press
a certain hotkey combination (Fn+CursorUp/Down), I can
change the brightness of the screen backlight, and the
current brightness is indicated with a small icon in the
upper left corner.  However, when I run X11 with DRI/DRM
enabled, that brightness icon doesn't work anymore, and
instead the hardware cursor changes its shape to that
brightness icon when I use the key combination (I can
provide screen shots if that description wasn't clear
enough).

("Doesn't work anymore" means that the icon in the top
left corner is either bit garbage, or it always shows
the last brightness state before starting X11, no matter
what the current state actually is.  The chance seems
to be about 50/50 which of those two cases occur.)

There are no errors or warnings in the Xorg.log that
might be related to the problem.  Maybe the resources
of the hardware cursor need to be assigned differently,
but that's probably a bug in Xorg, not in FreeBSD, and
I'm not familiar with the bug reporting facilities of
Xorg.

Fortunately, it's not a very serious problem, just a
little bit annoying.  Not annoying enough to let me
look how to write an Xorg bug report anyway.  :-)

Best regards
   Oliver

-- 
Oliver Fromme,  secnetix GmbH & Co. KG, Marktplatz 29, 85567 Grafing
Dienstleistungen mit Schwerpunkt FreeBSD: http://www.secnetix.de/bsd
Any opinions expressed in this message may be personal to the author
and may not necessarily reflect the opinions of secnetix in any way.

"[...]  one observation we can make here is that Python makes
an excellent pseudocoding language, with the wonderful attribute
that it can actually be executed."  --  Bruce Eckel


More information about the freebsd-mobile mailing list