[PATCH] rpaphp broken in ameslab

[Paul Mackerras]

Greg KH writes:

> But that's not relevant here.  I want to know when this hack will be
> removed, and done properly, in a arch-independent manner.  If it's not
> on the shortlist of things to do, I can't live with it, and will have to
> remove it, because that means it will never get fixed :(

What I propose is that the EEH code export a notifier list for code
that wants to know about EEH events, and do a notifier_call_chain when
events such as slot isolation occur.  The rpaphp code can then
register itself.  When a slot isolation event occurs, the rpaphp code
can then do whatever is appropriate.  This seems to me to be as
fundamental as a device driver being able to do request_irq().  The
rpaphp code is part of the device driver for the pci subsystem and it
needs to be notified when its hardware changes state.

Paul.

** Sent via the linuxppc64-dev mail list. See http://lists.linuxppc.org/