Re: [git pull request] ACPI patches for 2.6.34-rc6

[Linus Torvalds <torvalds@linux-foundation.org>]

On Sat, 8 May 2010, Rafael J. Wysocki wrote:

> On Saturday 08 May 2010, Linus Torvalds wrote:
> > 
> > On Sat, 8 May 2010, Rafael J. Wysocki wrote:
> > 
> > > On Friday 07 May 2010, Linus Torvalds wrote:
> > > > 
> > > > Btw, why don't we just force the SCI_EN by hand?
> > > 
> > > Because there are boxes that are broken right away by this.
> > > 
> > > I have one on my desk in front of me. :-)
> > 
> > Also if we check whether it was already set before?
> 
> Yes.

I don't think you understood the question.

If we check whether it was already set before, we wouldn't _do_ it by 
hand.

> > The thing is, if that bit is clear (_after_ we've done the regular ACPI 
> > setup), the box can't work correctly as far as I know.
> 
> That's correct.  But on that particular box acpi_enable() is necessary to
> make things work.  Setting SCI_EN by hand makes it hang solid.

Let me explain what I think we should do with actual code:

	acpi_enable();
	if (test_if_that_f_cking_bit_still_isnt_set)
		set_it_by_hand();

wouldn't that work on your box? The acpi_enable() seems to work for you, 
so it wouldn't actually ever set it by hand. 

But the problem is that on a number of boxes, acpi_enable() apparently 
doesn't do what it should do. Because the stupid BIOS "knows" it is 
already in ACPI mode, but it forgot to actually set the bit!

			Linus