[Skiboot] [PATCH 1/8] hdata/iohub: Look for IOVPD on P9

Oliver oohall at gmail.com
Fri Mar 29 15:46:30 AEDT 2019

On Fri, Mar 29, 2019 at 3:32 PM Stewart Smith <stewart at linux.ibm.com> wrote:
> Vasant Hegde <hegdevasant at linux.vnet.ibm.com> writes:
> > On 03/18/2019 11:40 AM, Oliver O'Halloran wrote:
> >> P8 and P9 use the same IO VPD setup, so we need to load the IOHUB VPD on
> >> P9 systems too.
> >
> > I have tested this patchset and it working fine.
> > It fixed issue introduced by 43f11d51 (disk detection issue in petitboot prompt).
> I'm almost afraid to ask :)

It wasn't that bad. The ZZ Vasant was using seems to have a
mis-configured PCIe switch which disabled power on some ports at boot.
Skiboot does enable slot power, but the initial power state detection
was broken.

> I've taken this series as of 84aa25921a09c7c6b703cf56a9396e2d55d49223.
> I did a few cleanups on the first patch for the hdat_to_dt test, and I
> *think* we're all okay with the base location code being a prefix on the
> location codes we spit out for PCI slots on boot (they contain the full
> information anyway).
> At least this way we're consistent? Only slightly nervous as there's
> location codes involved.

I'd rather we didn't include the base location code on OpenPower
boxes. It sort of makes sense on FSP systems because the big systems
require identifying the chassis in addition to the slot. As far as I
know we don't have any multi-drawer openpower systems so prefixing the
location code with the base location code is largely pointless. It's
pointless on the scale out ZZs as well, but whatever...

> Pavaman: could you double check that everything looks correct on
> Boston/Witherspoon with this merged (i.e. skiboot v6.3-rc1).
> --
> Stewart Smith
> OPAL Architect, IBM.

More information about the Skiboot mailing list