[PATCH,RFC] PCI: deal with device incorrectly reporting I/O decoding being enabled

[matthew@wil.cx (Matthew Wilcox)]

On Sun, Nov 08, 2009 at 09:04:51AM +0000, Russell King - ARM Linux wrote:
> On Sat, Nov 07, 2009 at 08:03:05AM -0700, Matthew Wilcox wrote:
> > I'm OK with this ... it might provoke bugs in some _other_ piece of hardware,
> > but that seems pretty unlikely.
> 
> Please remember that some ARM hardware only has the capability to read
> and write full 32-bit quantities of configuration space, which means
> writing the control register can result in the status register being
> written as well.
> 
> That shouldn't be a problem, but is merely something else to consider
> which isn't obvious.

Wow, that's horrid.  You must have to special case each config register
that's accessed in a 16-bit way, since some bits are 'preserve' and others
are 'write 1 to clear'.  What a nightmare.

-- 
Matthew Wilcox				Intel Open Source Technology Centre
"Bill, look, we understand that you're interested in selling us this
operating system, but compare it to ours.  We can't possibly take such
a retrograde step."