Hello,
I was thinking about some kind of hardware incompatibility that may cause
my SMP problem.
How are NMI and SMI (system management interrupt) managed by regular linux
and rtlinux / rtai? I would guess that NMI should normally be masked in
hardware and SMI ignored by dummy handler. As far as I understand, IOAPIC
may relay SMI to both CPUs and they have to decide somehow who is handling
it. Could be potential problem there?
The SMI signal is tied to INTIN23 input of IOAPIC. Depending on MASK bit
of IOAPIC's Reg 23, IOAPIC can deliver SMI as a normal interrupt over APIC
bus just like all other interrupts, or via SMI OUT which is tied in to both
CPUs (I wonder if this is what happens and creates the problem).
Intel B440FX motherboard has extensive system management features, and SMI
interrupt is very likely to occur (e.g. I use regular CPU fans without rpm
sensors, chassis intrusion sensor etc.). I have tried to disable all system
management events in BIOS. SMI can also be disabled in DISMIC control
registers (in ISA IO space) but I do not have documentation for that.
I would appreciate any feedback from people with better understanding of
SMI interrupt handling.
Surya, does your system have any extensive system management features?
Thanks,
Sergey.
-- [rtl] ---
To unsubscribe:
echo "unsubscribe rtl" | mail [EMAIL PROTECTED] OR
echo "unsubscribe rtl <Your_email>" | mail [EMAIL PROTECTED]
---
For more information on Real-Time Linux see:
http://www.rtlinux.org/rtlinux/
