Improving the suspend quirks guessworking

[Victor Lowther]

On Wed, 2008-03-26 at 21:23 +0200, Daniel Stone wrote:
> On Tue, Mar 25, 2008 at 10:33:16PM -0500, ext Victor Lowther wrote:
> > On Tue, 2008-03-25 at 01:16 +0100, Danny Kukawka wrote:
> > > If you take a look at already existing quirks, I wouldn't say this is a common 
> > > combination to fix things generally or often.
> > 
> > It is getting to be more and more common.  Intel, AMD, and nVidia have
> > won the graphics wars for now.  Intel now has a sane in-kernel driver
> > that does not require quirks,
> 
> If you're running an absolute bleeding-edge kernel.

... or if you will be running a distro that has the driver installed
natively sometime in the next 6 months.

> > the newly-opened specs for the ATI gpus should really help the ATI
> > drivers improve,
> 
> Except that it will take quite a while to get this working.

Indeed.  I have no illusions that the open-source radeon driver will
magically acquire flawless suspend/resume just because better docs on
the chips exist. However, it is better to have the HAL functionality in
place to handle things properly when we have a shiny new kernel driver
written basedon the docs that does have decent suspend/resume support.

> > and the latest nVidia binary
> > drivers are really quite good at this whole power management thing.
> > 
> > I know it seems to be popular to slag the binary video drivers, but for
> > some of us they are the only reasonable option.  
> 
> There's a difference between berating you for your choices, and not
> wanting to optimise for same.

It is not so much "optimization" as "not actively ignoring them".
Currently, I see HAL as doing the latter on Linux. The freebsd code
seems to be doing a better job there.

> Cheers,
> Daniel
-- 
Victor Lowther
Ubuntu Certified Professional