[PATCH] powerpc: fsl_pci: Fix PCI/PCI-X regression

[Scott Wood]

On Fri, 2014-08-22 at 14:00 -0500, Aaron Sierra wrote:
> ----- Original Message -----
> > From: "Scott Wood" <scottwood at freescale.com>
> > To: "Aaron Sierra" <asierra at xes-inc.com>
> > Cc: linuxppc-dev at lists.ozlabs.org, "Minghuan Lian" <Minghuan.Lian at freescale.com>
> > Sent: Friday, August 22, 2014 1:36:31 PM
> > Subject: Re: [PATCH] powerpc: fsl_pci: Fix PCI/PCI-X regression
> > 
> > On Fri, 2014-08-22 at 12:54 -0500, Aaron Sierra wrote:
> > > ----- Original Message -----
> > > > From: "Scott Wood" <scottwood at freescale.com>
> > > > Sent: Thursday, August 21, 2014 5:01:46 PM
> > > > 
> > > > On Thu, 2014-08-21 at 16:54 -0500, Aaron Sierra wrote:
> > > > > ----- Original Message -----
> > > > > > From: "Scott Wood" <scottwood at freescale.com>
> > > > > > Sent: Thursday, August 21, 2014 4:19:56 PM
> > > > > > 
> > > > > > Why wouldn't a normal PCI agent be able to bus master?
> > > > > > 
> > > > > > -Scott
> > > > > > 
> > > > > 
> > > > > Short answer:
> > > > > 
> > > > > Simply because the hardware strapping for Host/Agent determines the
> > > > > default state of the Bus Master bit in the Command register. Without
> > > > > that bit being set, an Agent won't be able to send the PCI cycles
> > > > > necessary to enumerate the bus.
> > > > 
> > > > But what if the host has already set that bit before Linux boots?
> > > 
> > > That's a very good point. I think that concern can be addressed by looking
> > > for another telltale sign of enumeration, whether an address has been
> > > assigned to the bridge's BAR 0 (PCSRBAR).
> > 
> > I don't see how that's any different.  The host may or may not have
> > assigned an address.
> 
> I don't agree with that. If the host has enabled bus mastering for the
> device, then it also surely would also have assigned an address to the
> always on PCSRBAR during enumeration.
> 
> In what sort of environment would a host have enabled bus mastering for
> a peripheral device _before_ assigning addresses to its BARs?

It's unusual, but not illegal and doing so shouldn't cause a PCI device
to try to take over the system.

Plus, zero is (arguably) a valid BAR address.

> > > > I understand why you need to do this -- I just don't think this is a
> > > > reliable way of detecting that you're in that situation.  How about a
> > > > kernel command line setting?
> > > 
> > > I'd like to avoid requiring a kernel command-line option for this.
> > 
> > It's hardware description, so you could use a device tree property.
> > 
> 
> I like this better than a command-line option, but what are you
> suggesting? Would you want a device-tree property to gate the logic that
> I've proposed?

I don't think you need the logic you've proposed if you have a device
tree property -- the property would assert that you are the owner of the
bus despite being an agent.

-Scott