Date: Sun, 1 Dec 2002 21:23:05 -0500 (EST) From: "Michael Richards" <michael@fastmail.ca> To: freebsd-smp@freebsd.org Subject: Intel SE7500CW2 narrowed down... Message-ID: <3DEAC409.000009.36383@ns.interchange.ca>
next in thread | raw e-mail | index | archive | help
--------------Boundary-00=_HYZG0DJXFQQMYJ0CCJD0
Content-Type: Text/Plain
Content-Transfer-Encoding: 7bit
Ok, I've been hacking up /usr/src/sys/i386/i386/mp_machdep.c to spit
out some more debugging stuff. Here is what I have now.
Programming 24 pins in IOAPIC #0
IOAPIC #0 intpin 2 -> irq 0
Programming 24 pins in IOAPIC #1
Programming 24 pins in IOAPIC #2
Starting application processors...Initializing APIC...Done.
Installing AP 1st level boot code...Done.
Setting up P->V mapping...Done.
Starting AP 1 of 1
calling start_ap(1,9a000)
This is physical CPU 6 at 0x9a
Timed out waiting for processor to start!
fake processor starting y/n? [n] Done.
FreeBSD/SMP: Multiprocessor motherboard
cpu0 (BSP): apic id: 0, version: 0x00050014, at 0xfee00000
cpu1 (AP): apic id: 6, version: 0x00000000, at 0xfee00000
io0 (APIC): apic id: 2, version: 0x00178020, at 0xfec00000
io1 (APIC): apic id: 3, version: 0x00178020, at 0xfec80000
io2 (APIC): apic id: 4, version: 0x00178020, at 0xfec80400
Interesting notes:
a) the version for the AP CPU seems bogus and the address shows up in
the list the same as cpu0. This happens both in the original kernel
and my hacked up one.
b) The failure happens in start_ap and it happens at the end of the
function where mp_ncpus doesn't get incremented. For kicks I tried
hacking it up to give you the option of returning a 1 so the system
would continue. Had no effect.
Since I'm getting in over my head here I'm assuming that the start_ap
function sends the CPU off to increment the mp_ncpus and fill in the
entry giving its version and address. This is a very broad
assumption. If I'm remotely correct then our second CPU isn't doing
anything and as a result even trying to trick it the SMP never comes
on.
If anyone can direct me or take over, or provide any other debugging
type patch it would be greatly appreciated. I've been looking around
for technical docs don't really know what I'm looking for. Any ideas?
-Michael
_________________________________________________________________
http://fastmail.ca/ - Fast Secure Web Email for Canadians
--------------Boundary-00=_HYZG0DJXFQQMYJ0CCJD0--
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-smp" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?3DEAC409.000009.36383>