Re: [PATCH -mm] clocksource init adjustments (fix bug #7426)

[john stultz <johnstul@us.ibm.com>]

On Fri, 2007-03-02 at 02:18 -0800, Andrew Morton wrote:
> On Thu, 22 Feb 2007 16:13:02 -0800 john stultz <johnstul@us.ibm.com> wrote:
> > Thus the solution here is to register clocksources earlier (ideally when
> > the hardware is being initialized), and then we enable clocksource
> > selection at fs_initcall (before device_initcall).
> > 
> > This patch should probably get some testing time in -mm, since
> > clocksource selection is one of the most important issues for correct
> > timekeeping, and I've only been able to test this on a few of my own
> > boxes.
> 
> This doornails my Nocona box early in boot:
> http://userweb.kernel.org/~akpm/s5000431.jpg
> 
> Slab isn't ready yet - time_init()->hpet_arch_init() is called before
> start_kernel() has run kmem_cache_init().

Oh! Sorry! Yea, looking at it more the ioremap isn't actually necessary,
as we can use hpet_readl() instead of re-calculating the hpet base
address pointer.

I'll fix this up (and find an HPET enabled x86_64 box to test it on) and
get a patch to you shortly.

Sorry again!
-john