[Skiboot] [PATCH v11 20/23] core/pci: Power off empty hotpluggable slot

Stewart Smith stewart at linux.vnet.ibm.com
Thu Jun 9 18:48:03 AEST 2016

Gavin Shan <gwshan at linux.vnet.ibm.com> writes:
>>I'd prefer we use a callback rather than busy looping (and there's other
>>places in PCI code that does this too, but this appears to be new, so it
>>gets unfairly picked on :)
> I'm afraid we cannot relinquish the CPU for other tasks. This function is
> called when skiboot is booting, meaning it's executed in a job context.
> I guess we don't support stop/resume job in current implementation. Even
> with stop/resume job supported, more information needed to track the states
> even that's supported - a bit complexity than the polling mechanism. Lets
> have the polling mechanism for now and I will think about it in future.

Ahh... I was wondering about that.

I probably need to have a think about how I'd like it to look, and what
we care about at boot time versus at run time.

Stewart Smith
OPAL Architect, IBM.

More information about the Skiboot mailing list