Re: SMP broken on Dell PowerEdge 4100/200 under 2.6.0-testxx?

From: Ethan Weinstein
Date: Sat Dec 06 2003 - 15:06:38 EST


Colin Coe wrote:
Sorry about the delay.

Booted with noirqbalance.

[root@host root]# cat /proc/interrupts
CPU0 CPU1
0: 7411777 5971987 IO-APIC-edge timer
1: 7 4 IO-APIC-edge i8042
2: 0 0 XT-PIC cascade
4: 16 42 IO-APIC-edge serial
5: 4915 4820 IO-APIC-level eth1
10: 67 69 IO-APIC-level aic7xxx
11: 325 266 IO-APIC-level eth0
12: 47 109 IO-APIC-edge i8042
14: 0 0 IO-APIC-level CS46XX
15: 6398 6401 IO-APIC-level megaraid
NMI: 0 0
LOC: 13383659 13383658
ERR: 0
MIS: 0

That looks a lot better...

Thanks!

--
"Obnoxious frog..." Spike, 2071AD

William Lee Irwin III said:

On Sat, Dec 06, 2003 at 01:48:54PM +1100, Nick Piggin wrote:

Although in this case Colin has 2 PPro 200s.
Colin - process load should be evenly distributed between CPUs, and this
is generally the most important thing. Big networking loads (most
commonly)
can put a lot of time into processing interrupts though.

That is rather busted, then.

Colin, could you try booting with noirqbalance on the kernel command
line?


-- wli



I'll throw my hat in here as well. This is an old compaq proliant I have at the office- dual 400 PII, booted with "noirqbalance" 2.6.0-test11:

CPU0 CPU1
0: 2580383 1920931 IO-APIC-edge timer
1: 6 3 IO-APIC-edge i8042
2: 0 0 XT-PIC cascade
5: 467 423 IO-APIC-level TLAN
8: 1 0 IO-APIC-edge rtc
9: 15 15 IO-APIC-level sym53c8xx
10: 17 17 IO-APIC-level sym53c8xx
11: 1602 1593 IO-APIC-level ida0
14: 8 2 IO-APIC-edge ide0
NMI: 0 0
LOC: 4501366 4501354
ERR: 0
MIS: 0

witout "noirqbalance" we interrupt on CPU0 solely.

-E
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/