Re: [PATCH] lpc_ich: Fix a 3.5 kernel regression for iTCO_wdt driver

[Samuel Ortiz <sameo@linux.intel.com>]

Hi Feng,

On Thu, Aug 23, 2012 at 01:08:14PM +0800, Feng Tang wrote:
> On Wed, 22 Aug 2012 22:55:43 +0100
> Matthew Garrett <mjg59@srcf.ucam.org> wrote:
> 
> > On Wed, Aug 22, 2012 at 09:55:12PM +0200, Wim Van Sebroeck wrote:
> > 
> > > Any idea why the acpi_check_resource_conflict() check gives a conflict?
> > 
> > Because the resource range is declared in ACPI and we assume that that 
> > means the firmware wants to scribble on it. We'd need the output of 
> > acpidump to work out whether that's safe or not.
> 
> Good point, I checked the conflict for iTCO_wdt, the conflict exists on
> almost all the machines I have.
> 
> According to ICH (7/8/9 etc)spec, the TCO watchdog has a 32 bytes long IO 
> space resource, and the bit 9 of TCO1_STS register is "DMISCI_STS", which
>  indicates whether a SCI happens, and will be cleared by writing 1
> to it. Most of DSDT table will claim a TCO op region only for one bit:
> "DMISCI_STS" , as some method may need to access that bit. 
> 
> I think there is some risk, but it's quite safe as the DMISCI_STS bit has
> nothing to do with TCO driver itself, and TCO driver never access it, also
> this TCO driver has been there for years, and this resource conflict also
> exists for many generations hardware. 
Makes sense to me.
I'm queueing this one to my for-linus branch, I'll send a pull request soon.

Cheers,
Samuel.

-- 
Intel Open Source Technology Centre
http://oss.intel.com/