Re: [PATCH] Revert commit e8aa4667baf74dfd85fbaab86861465acb811085

[Ingo Molnar <mingo@elte.hu>]

* Thomas Gleixner <tglx@linutronix.de> wrote:

> On Thu, 4 Sep 2008, Ingo Molnar wrote:
> 
> > 
> > * Andreas Herrmann <andreas.herrmann3@amd.com> wrote:
> > 
> > > This reverts commit e8aa4667baf74dfd85fbaab86861465acb811085
> > >  (x86: enable hpet=force for AMD SB400)
> > > 
> > > Since ATI/AMD decided not to support HPET on SB4xx it doesn't
> > > make sense to enable this unsupported feature.
> > > (I was not aware of this when submitting the quirk.)
> > > 
> > > If a system with SB4xx chipset provides an ACPI HPET table and does
> > > not boot, "nohpet" should be used as kernel parameter.
> > > 
> > > Signed-off-by: Andreas Herrmann <andreas.herrmann3@amd.com>
> > 
> > applied to tip/x86/urgent, thanks Andreas. I guess a system broke due to 
> > this commit?
> 
> Hmm, why do we remove something which needs to be force enabled by the 
> user anyway ?

good point, i thought the original commit caused unconditional 
force-enabling - but indeed it is only relevant if hpet=force is 
specified. (which should be rare and specific)

> Is the HPET on these systems not working at all so the force enable 
> code is useless ?

also, if a user does hpet=force and thing break he's got to keep all the 
pieces, right?

or is there any other side-effect of the commit that matters here?

	Ingo