> > I have had no trouble from my Tyan board, I got a S1668 ATX Dual PPro
...
> It looks like their quality of hardware is great but their BIOSes suck.
> Unfortunately, BIOS is used for initial startup, before Linux is loaded,
> and their BIOS bugs limit Linux features. For example their MP tables
> report IRQs in a way that prevents PCI IRQs from being mapped to
> IRQ[19:16]. As a result you lose four ISA interrupts.
...
If you stick with BIOS 3.03 like I did, you can map PCI IRQs to APIC lines
19:16. I currently have:
CPU0 CPU1
0: 2258788 2147634 IO-APIC-edge timer
1: 14292 13622 IO-APIC-edge keyboard
2: 0 0 XT-PIC cascade
3: 58755 58235 IO-APIC-edge serial
4: 333871 322707 IO-APIC-edge serial
5: 0 1 IO-APIC-edge soundblaster
8: 1 1 IO-APIC-edge rtc
12: 139962 144027 IO-APIC-edge PS/2 Mouse
13: 1 0 XT-PIC fpu
16: 28 34 IO-APIC-level BusLogic BT-948
17: 106076 104386 IO-APIC-level BusLogic BT-958
18: 39404 39308 IO-APIC-level eth0
One time I had a four-port serial card that used seperate IRQs for all four
lines on IRQs 9, 10, 11, and 15. I was doin' interrupts all the time!!
(Not really--just how many interrupts do you think a 9600 baud terminal
generates? :)
Anyway, I noticed (with the S1668D) that the "new" BIOS (5.0? 5.1?)
greatly sucks in the IRQ-IOAPIC mapping and that's why I went back to
3.03. Had Tyan not made us lucky with the 3.03, we would have been
screwed ('cause it's unlikely they would have fixed it). I can't
understand a company that breaks stuff in subsequent releases and
neither acknowledges it nor fixes it.
Brendan
-
Linux SMP list: FIRST see FAQ at http://www.irisa.fr/prive/dmentre/smp-howto/
To Unsubscribe: send "unsubscribe linux-smp" to [EMAIL PROTECTED]
