[linux-pm] suspend slow in 2.6.25-rc6 (was Re: Suspend and hibernation patchset against -rc6)

[Rafael J. Wysocki]

On Wednesday, 19 of March 2008, Pavel Machek wrote:
> On Wed 2008-03-19 23:35:10, Alexey Starikovskiy wrote:
> > Pavel Machek wrote:
> >> Hi!
> >>
> >>>> Ok, 7c0... is innocent. Reverting not only the mdelay, but also rest
> >>>> of drivers/acpi/ec.c patch fixed the
> >>>> "rc6-breaks-backlight-in-X-over-lid-close", good.
> >>> Can you clarify a bit?
> >>>
> >>> Do you mean that the full revert of 2c81ce4 that already got committed is 
> >>> sufficient for you, or do you need to totally undo everything we've done 
> >>> to ec.c since -rc5, and thus also to revert the _other_ revert we did (in 
> >>> 4af8e10a6c57e7292862bd1703712f0565c7e429)?
> >>
> >> I did some more testing, and realized I was wrong. We need to totally
> >> undo everything we've done to ec.c since -rc5... (that has small
> >> sideffect of brightness up/down keys no longer working -- regression
> >> since 2.6.24, but saves us from "backlight gone after
> >> closing/reopening lid" which is _NASTY_ regression from 2.6.25-rc5).
> >>
> >> ACPI sucks.
> >>
> >> I.e. this patch:
> >>
> >>
> >> Revert ec.c to 2.6.25-rc5 state:
> >>
> >> * this fixes backlight after closing/reopening the lid while in X on
> >> thinkpad x60
> >>
> >> * unfortunately it breaks fn-home/end keyboard brightness control
> > It should also break volume up/down/mute keys. Essentially all the keys on Thinkpad.
> > And it would also be regression to 2.6.24 level.
> 
> As I explained above.
> 
> Unfortunately, applying this means I loose screen as soon as I
> close/reopen lid. 2.6.24 did not behave like that. And it makes my
> machine mostly unusable -- as soon as you try to carry it away, you
> loose video. BAD.

Pavel,

Can you text the current -git (preferably without anything on top) pretty please?

Thanks,
Rafael