Re: [PATCH v4 04/10] s390/pci: Add architecture specific resource/bus address translation

[Niklas Schnelle <schnelle@linux.ibm.com>]

On Thu, 2025-10-02 at 12:00 -0500, Bjorn Helgaas wrote:
> On Thu, Oct 02, 2025 at 02:58:45PM +0200, Niklas Schnelle wrote:
> > On Wed, 2025-09-24 at 10:16 -0700, Farhan Ali wrote:
> > > On s390 today we overwrite the PCI BAR resource address to either an
> > > artificial cookie address or MIO address. However this address is different
> > > from the bus address of the BARs programmed by firmware. The artificial
> > > cookie address was created to index into an array of function handles
> > > (zpci_iomap_start). The MIO (mapped I/O) addresses are provided by firmware
> > > but maybe different from the bus address. This creates an issue when trying
> > > to convert the BAR resource address to bus address using the generic
> > > pcibios_resource_to_bus().
> > > 
> > > Implement an architecture specific pcibios_resource_to_bus() function to
> > > correctly translate PCI BAR resource addresses to bus addresses for s390.
> > > Similarly add architecture specific pcibios_bus_to_resource function to do
> > > the reverse translation.
> > > 
> > > Signed-off-by: Farhan Ali <alifm@linux.ibm.com>
> > > ---
> > >  arch/s390/pci/pci.c       | 74 +++++++++++++++++++++++++++++++++++++++
> > >  drivers/pci/host-bridge.c |  4 +--
> > >  2 files changed, 76 insertions(+), 2 deletions(-)
> > > 
> > 
> > @Bjorn, interesting new development. This actually fixes a current
> > linux-next breakage for us. In linux-next commit 06b77d5647a4 ("PCI:
> > Mark resources IORESOURCE_UNSET when outside bridge windows") from Ilpo
> > (added) breaks PCI on s390 because the check he added in
> > __pci_read_base() doesn't find the resource because the BAR address
> > does not match our MIO / address cookie addresses. With this patch
> > added however the pcibios_bus_to_resource() in __pci_read_base()
> > converts  the region correctly and then Ilpo's check works. I was
> > looking at this code quite intensely today wondering about Benjamin's
> > comment if we do need to check for containment rather than exact match.
> > I concluded that I think it is fine as is and was about to give my R-b
> > before Gerd had tracked down the linux-next issue and I found that this
> > fixes it.
> > 
> > So now I wonder if we might want to pick this one already to fix the
> > linux-next regression? Either way I'd like to add my:
> > 
> > Reviewed-by: Niklas Schnelle <schnelle@linux.ibm.com>
> 
> Hmmm, thanks for the report.  I'm about ready to send the pull
> request, and I hate to include something that is known to break s390
> and would require a fix before v6.18.  At the same time, I hate to add
> non-trivial code, including more weak functions, this late in the
> window.
> 
> 06b77d5647a4 ("PCI: Mark resources IORESOURCE_UNSET when outside
> bridge windows") fixes some bogus messages, but I'm not sure that it's
> actually a functional change.  So maybe the simplest at this point
> would be to defer that commit until we can do it and the s390 change
> together.
> 
> Bjorn

Yeah that makes a lot of sense. I agree this change is not completely
trivial. Might have been a little enthusiastic with it reviving PCI
support from the dead on linux-next. If the other patch is not an
important fix and Ilpo seems to agree, then simply reverting it is the
safe solution.

Thanks,
Niklas