[Libdlo] randr.c RRFirstOutput() crash when using "displaylink" driver in combination with "nvidia" driver

[Mulchman]

On 3/12/2010 12:34, Bernie Thompson wrote:
>
> http://www.x.org/wiki/Projects/XRandR is a starting point, but
> slightly out of date.
> Here's the place where the latest formal spec lives:
> http://cgit.freedesktop.org/xorg/proto/randrproto/tree/randrproto.txt
>   

Hey Bernie, thanks for the links.

> I think all the majors have some level of randr support, but what
> version (1.2, 1.3 ..) varies, etc. Here's the Intel implementation,
> that relies on all the new DRI/KMS Intel support.
>
> http://cgit.freedesktop.org/xorg/driver/xf86-video-intel/tree/src/drmmode_display.c
>
>   

I believe that's one of the ones I've been looking at currently.

One random side note I wanted to mention about adding the RRScreeInit
call and adding some function pointer stubs is that the edit to
/etc/gdm/Init/Default is not needed any more.  This is the snippet I'm
referring to:

XRANDR=`gdmwhich xrandr`
if [ "x$XRANDR" != "x" ]; then
  $XRANDR -o 0
fi

The attached file (which is quite worthless at the moment) is a diff
against src/displaylink.c in the displaylink driver git repo here:
http://git.plugable.com/webdav/xf-video-udlfb/ and it shows the
RRScreenInit & stub function additions which make the snippet above not
necessary.  Of course, it's far from complete and I'm not even sure if
it's the correct way to do things in the first place (as indicated by
the comments).  And, it probably leaks a bit as there's no shutdown code
- only initialization (that I'm not even sure of is correct, also).

Just wanted to throw that out there.
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: displaylink.c.diff
URL: <http://lists.freedesktop.org/archives/libdlo/attachments/20100315/2bb56565/attachment.asc>