kern/123140: SMP boot causes slow KB, ATA drives not detected
Bob Frazier
bobf at mrp3.com
Tue Sep 2 07:40:03 UTC 2008
The following reply was made to PR kern/123140; it has been noted by GNATS.
From: Bob Frazier <bobf at mrp3.com>
To: bug-followup at FreeBSD.org
Cc: gavin at FreeBSD.org
Subject: Re: kern/123140: SMP boot causes slow KB, ATA drives not detected
Date: Tue, 02 Sep 2008 00:33:38 -0700
1. made temporary code changes to force IRQ 9 to LOW/LEVEL - no
observed changes
2. made temporary code changes in lapic_handle_intr to determine
whether or not ANY interrupt handling for the LAPIC is EVER taking place
- it appears that it is not.
Conclusion: It appears that interrupts are NOT being serviced at all
through the APIC/LAPIC. Timers, however, ARE being serviced (but at a
rate that is equal to a single CPU but for all 4 CPUs, so that 4 times
as many timer interrupts are being serviced than for a single CPU).
Other debugging effort, such as determining whether interrupts are even
being set up correctly, suggests that the code is operating as designed
and there are no apparent code design problems or code flow problems
that might be causing this problem.
Conclusion: It appears that there is an APIC/LAPIC configuration step
that is either missing or incorrect (such as an incorrect memory address
or a register that is not being assigned properly) that is the cause of
this particular problem.
(general call for assistance)
I am attempting to do what I can to help to locate and/or correct this
particular problem. I have also personally attempted to contact ASUS
through their customer support web site and the response I received was
the electronic equivalent of a door slammed in my face. They canceled
the problem request and didn't even respond by e-mail that they had done it.
As such, if anyone out there has ANY board or NVidia chipset
documentation that might be helpful, please attach it to this bug report
(or send it to me directly). Specifically I'm looking for a wiring
diagram or register map that's specific to the NVidia chipset (and not
the APIC/LAPIC as a whole - I already have that for Intel - NVidia may
NOT be following the Intel spec 100%). Also, information (such as
APIC/LAPIC register base addresses) from a working system (even one
running windows) could be very helpful if the problem is a result of
problems in the ACPI BIOS tables.
More information about the freebsd-bugs
mailing list