Migrating suspend quirks away from hal

Victor Lowther victor.lowther at gmail.com
Fri Nov 13 18:42:20 PST 2009

On Fri, 2009-11-13 at 12:30 -0600, Victor Lowther wrote:
> On Nov 13, 2009, at 7:30 AM, Martin Pitt <martin.pitt at ubuntu.com> wrote:
> > Hello Victor, hello Richard, hello devkitters,
> >
> > in the past half year we have come quite far with the "halsectomy",
> > i.e. deprecating hal [1]. The main two remaining items are now X.org
> > input devices (where a patch is already being discussed), and the
> > suspend quirks in pm-utils.
> >
> > I would like to push this forward a little. From what I can see, the
> > are two sensible options:
> >
> > (1) Convert the hal-info suspend quirks to a simple text format and
> >     put it into/maintain it directly in pm-utils.
> >
> >     → nice encapsulation, maintained at the place where they are
> >     actually needed
> I did this as an experiment last year, and it worked out well at the  
> time. I will see if I still have that branch around somewhere to make  
> available for comment.

Attached is the code I posted to the hal mailing list last November to
migrate the video quirk database out of HAL and into something a little
easier for bash to deal with.  It did not have the full expressive power
of the hal-info XML format, but that was mostly not needed anyways --
the only place it was used was the 21-nvidia quirk for handling g80
chipsets with the nv driver, but that can just as easily be hacked into
a few lines of bash.

Hacking this into a pm-utils hook would not be that tricky -- the flow
would go something like this:

if has_kms || has_nvidia || has_fglrx || has_smart_intel; then
elif has_nvidia_g80; then

video_quirks definitly has bugs, but it did OK as a proof of concept for
-------------- next part --------------
A non-text attachment was scrubbed...
Name: video-quirks.bash
Type: application/x-shellscript
Size: 10898 bytes
Desc: not available
Url : http://lists.freedesktop.org/archives/devkit-devel/attachments/20091113/f51f9b34/attachment.bin 

More information about the devkit-devel mailing list