Re: IORESOURCE_WINDOW for PCI-to-PCI bridges

[Bjorn Helgaas <helgaas@kernel.org>]

On Fri, Nov 11, 2022 at 10:48:16PM +0100, Pali Rohár wrote:
> On Friday 11 November 2022 15:05:55 Bjorn Helgaas wrote:
> > On Fri, Nov 11, 2022 at 09:09:45PM +0100, Pali Rohár wrote:
> > > On Thursday 20 January 2022 15:02:12 Bjorn Helgaas wrote:
> > > > On Thu, Jan 20, 2022 at 09:45:05PM +0100, Pali Rohár wrote:
> > 
> > [trimmed material; beginning of thread is at
> > https://lore.kernel.org/r/20211220155448.1233-3-pali@kernel.org]
> > 
> > > > > Meanwhile I found out that in linux/ioport.h file is IORESOURCE_WINDOW
> > > > > constant with comment /* forwarded by bridge */
> > > > > https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/include/linux/ioport.h?h=v5.15#n56
> > > > > 
> > > > > But apparently it is not set for resources behind PCI bridges and
> > > > > therefore it is not available in column of "resources" sysfs file.
> > > > > 
> > > > > So maybe instead of adding new sysfs files, it would be better way to
> > > > > implement this flag and export it in flags column of "resources" file
> > > > > for every row which belongs to resources behind bridges?
> > > > 
> > > > I looked at that, too.  Today we only set IORESOURCE_WINDOW for host
> > > > bridge windows.  Maybe it could be set for PCI-to-PCI bridge windows,
> > > > too.  Would have to audit users to make sure it wouldn't break
> > > > anything.
> > > 
> > > Hello Bjorn, I would like to remind this older issue. Did you have a time
> > > to audit usage of IORESOURCE_WINDOW? Some flag for resource forwarding
> > > windows in PCI-to-PCI bridges would really help userspace application to
> > > distinguish between IO/MEM BARs an IO/MEM forwarding windows.
> > 
> > I had forgotten all about this issue.  IIUC, the ultimate goal here
> > is to help lspci distinguish between an I/O window that's disabled and
> > one that's enabled at [io 0x0000-0x0fff].
> > 
> > I have not done the research to see whether it would be safe to set
> > IORESOURCE_WINDOW for PCI-to-PCI bridge windows.  I'm sorry if I left
> > the impression that I intended to do that.  I would welcome your help
> > to do that.
> 
> Ok, do you have some resources or other information at which I should
> look? I just do not know where to start or what to check for that
> research.
> 
> I looked into kernel sources and the only places where is code checking
> for IORESOURCE_WINDOW is ACPI related: arch/arm64/kernel/pci.c and
> drivers/pnp/resource.c. And I do not fully understand how is ACPI
> connected with PCI resources at this level. Other places which check
> (lib/vsprintf.c and drivers/pnp/interface.c) just use it for
> printf-formats.

Yeah, that's the kind of thing I have in mind.  I can't remember if I
had any specific concern.

Bjorn