[Skiboot] [RFC/PATCH] affinity: Use fabric node ID on most systems

Stewart Smith stewart at linux.vnet.ibm.com
Fri Nov 10 17:20:52 AEDT 2017

Michael Ellerman <mpe at ellerman.id.au> writes:
> Benjamin Herrenschmidt <benh at kernel.crashing.org> writes:
>> Except Murano where we stick to the module ID from HDAT
>> Signed-off-by: Benjamin Herrenschmidt <benh at kernel.crashing.org>
>> ---
>>  core/affinity.c | 42 ++++++++++++++++++++++--------------------
>>  1 file changed, 22 insertions(+), 20 deletions(-)
> Ping?
> It seems this patch is being used in the wild, so it would be good to
> either get it merged or nack'ed.
> Changing the numbering on existing P8 systems seems like it could
> definitely cause breakage in the kernel or apps, but on P9 maybe we can
> get away with it.

Yeah, I'm thinking we should be able to get away with it on P9...

Early next week I'll take a stab at running this through a bunch of
what-I-can-think-of tests, probably hacking something into op-test, and
generally trying to ensure that we're not too batshit insane to do it.

Or, I could merge it right now, git push and just FRIDAY YOLO the whole
thing (but I won't :)

Stewart Smith
OPAL Architect, IBM.

More information about the Skiboot mailing list