cvs commit: src/sys/dev/acpica acpi_perf.c

[M. Warner Losh]

In message: <4241E29C.7010407 at root.org>
            Nate Lawson <nate at root.org> writes:
: Warner Losh wrote:
: >>>Ok, I've committed a fix:  acpi.c rev 1.207.  Please cvsup and test.
: >>
: >>It does not help, still same panic
: > 
: > 
: > It doesn't help me either, and an audit of the code suggests that it
: > will have no effect.  There's a resource that's marked as being
: > active, which is why the panic is happening....  I've always hated
: > these panics, but have appreciated them in hindsight once the bug has
: > been fixed because it is a very effective enforcement of proper
: > operations.
: > 
: > Warner
: 
: I'm trying to figure out what's going on.  The probe routine should not 
: be called multiple times since it returns 0 and now with my commit it 
: should completely clean up after itself.  Can you put this in 
: acpi_perf_identify, probe, attach and let me know the full dmesg up to 
: the panic?

I'll give it a try.  I'll reiterate that probe routines should do
nothing more than 'I'm here' and all other processing should be done
in attach.  It looks like that might be a little hard in this case,
but I've not looked into it in detail.

Warner