Re: [PATCH 2/2] acpi,backlight: MSI S270 laptop support - driver

[Lennart Poettering <mzxreary@0pointer.de>]

On Thu, 10.08.06 12:15, Brown, Len (len.brown@intel.com) wrote:

Len,

> >drivers/video/backlight/
> >  It doesn't just do backlight control.
> 
> Perhaps the backlight part should live there.

Mhmm, that would mean I'd had to split up the driver. That would mean
duplicate code to a certain degree and two separate mini drivers with
very connected features. 

> >drivers/misc/
> >  Seems to be the last resort for everything that doesn't fit it
> >  otherwise.
> >
> >Unless anyone has a better idea I will move it to drivers/misc/, then.
> 
> Yeah, there is probably a better place than misc.

Hmm. As you might have noticed I already posted an updated driver a
few minutes ago which resides in drivers/misc. Is there any need to
move it once again?

> >I cannot map the "automatic brightness control" feature to the
> >backlight class driver, that's why I duplicated the brightness stuff
> >in /proc/acpi/s270/.
> 
> Better to extend the backlight code so that it can handle the
> special features of this platform than to duplicate brightness
> control in multiple places.

Hmm. Better not. That "automatic brightness control" feature and its
behaviour are very specific to this laptop. I don't see any value in
abstracting that. That would be quite a lot of overdesigning in my
eyes. In fact the driver disables this feature on load by default,
assuming that it was probably loaded to do the control in
software. 

> >> wlan and bluetooth indicators/controls need to appear under
> >> generic places under sysfs -- not under platform specific
> >> files under /proc/acpi.
> >
> >What are those "generic" places? I cannot think of any besides a
> >"platform" device.
> 
> They should appear as properties under their associated
> devices in /sys/devices tree rather than invening a new place.

My updated driver does this now. The attributes appear in
/sys/devices/platform/s270pf/.

> >Any ideas on that ec_transaction() patch I sent earlier?
> 
> I agree 100% that ec.c needs to be whacked.  Unfortunately
> there seem to be 3 people doing it at the same time,
> so we'll have to sort that out.

Whoever works on that: please make sure to offer a generic EC access
function similar to my ec_transaction() patch! Thanks!

Is there any timeframe for this EC rework? Is there any chance to get
my patch merged in the meantime? (Hmm, I guess i am little unpatient...)

> ps. Lennart, please understand that I didn't intend to be gruff.  I
> figured that by the quality of your work -- which is better than
> most -- that I'd go for an immediate reply, even though I was still
> holding a sharp pointed stick at the end of a 6-hour bug scrub
> marathon.

Oh, that's OK. I value a quick reply quite a lot. Thank you very much
for your quick review!

Thanks,
        Lennart

-- 
Lennart Poettering; lennart [at] poettering [dot] net
ICQ# 11060553; GPG 0x1A015CC4; http://0pointer.net/lennart/