Re: [bisected][resend] pnp: Huge number of "io resource overlap" messages

[Bjorn Helgaas <bjorn.helgaas@hp.com>]

On Saturday 20 September 2008 05:49:05 pm Frans Pop wrote:
> On Thursday 18 September 2008, Bjorn Helgaas wrote:
> > On Tuesday 09 September 2008 12:31:09 pm Rene Herman wrote:
> > > On 09-09-08 19:40, Bjorn Helgaas wrote:
> > > > If the PCI device isn't fully initialized, it doesn't seem right to
> > > > check it for resource conflicts.  But I don't know how to tell
> > > > that.
> > >
> > > His pci_resource_start() values are 0. How about just checking for
> > > that?
> >
> > Frans, can you test Rene's patch?  I think it will solve the problem
> > you're seeing, and if so, we should put it in for 2.6.27.  But I'd like
> > to have your "Tested-by" first.
> 
> Tested against current git (v2.6.27-rc6-158-g9824b8f) and looks good.
> Attached the patch with Rene's Signed-off and my Tested-by for 
> convenience.
> 
> I had not tested earlier as you said you wanted to better understand the 
> cause first. Did you get anything more about why things happen as they do 
> from the info I sent?

Thanks for testing this.

We're looking at some other issues in the same area, or at least,
where the fix might be in the same area:
  http://bugzilla.kernel.org/show_bug.cgi?id=10231
  http://bugzilla.kernel.org/show_bug.cgi?id=9904

I am still not 100% comfortable with this because I think we really
want to know whether the BAR value is zero, not whether the CPU
address is zero, and pci_resource_start() gives us the CPU address.

Bus and CPU addresses are currently identical on x86, but I expect
that will change someday.  They're already different on ia64 and
some other architectures.

Bottom line, I think we should tweak the patch to check the BAR
address before we put it in.

Bjorn