Messages in this thread | | | Subject | RE: [PATCH] 2.6.5- es7000 subarch update | Date | Wed, 14 Apr 2004 11:45:28 -0500 | From | "Protasevich, Natalie" <> |
| |
Hi Len,
>please test it on your ES7000. I'd love to see >the full dmesg from an ES7000 if you can grow your >CONFIG_LOG_BUF_SHIFT to capture it all.
The patch worked great, my system came up beautifully. Len, it was so well done, and it was fast, too - thanks! I forgot to increase this parameter, but still collected a full trace with some extra IO-APIC snapshots. For now, I am attaching a serial console trace, and later will provide you full dmesg.
>I expect that if the Linux IRQ convention was working >on the ES7000 before, this patch will maintain >compatibility with that. However, if things were >not working as expected before, I just wanted to >point out that another convention is possible -- >particularly with a sub-architecture.
Yes, the IRQ schema you described was used on the ES7000 previously on 2.6. For example, the way interrupts looked with the old interrupt code (with irq_balance off, and my tweak for the IDE): 0: 1566205 0 0 0 IO-APIC-edge timer 1: 12 0 0 0 IO-APIC-edge i8042 2: 0 0 0 0 XT-PIC cascade 4: 20 0 0 0 IO-APIC-edge serial 8: 3 0 0 0 IO-APIC-edge rtc 12: 399 0 0 0 IO-APIC-edge i8042 15: 37 0 0 0 IO-APIC-edge ide1 17: 8083 0 0 0 IO-APIC-level megaraid 20: 20141 0 0 0 IO-APIC-level eth0 23: 0 0 0 0 IO-APIC-level acpi NMI: 0 0 0 0 LOC: 1562738 1562605 1562608 1562607 ERR: 0 MIS: 0
With your patch, it looks like this:
0: 21266 79032 0 0 IO-APIC-edge timer 1: 14 0 0 0 IO-APIC-edge i8042 2: 0 0 0 0 XT-PIC cascade 4: 19 0 0 0 IO-APIC-edge serial 8: 2 0 0 0 IO-APIC-edge rtc 12: 503 0 0 0 IO-APIC-edge i8042 15: 39 0 0 1 IO-APIC-edge ide1 17: 1762 0 0 0 IO-APIC-level megaraid 20: 868 0 0 0 IO-APIC-level eth0 23: 0 0 0 0 IO-APIC-level acpi NMI: 0 0 0 0 LOC: 94472 94484 94483 94482 ERR: 0 MIS: 0
There were other strange schemas that I've used before in 2.4, mostly due to ongoing inconsistencies in the BIOS and ACPI. Those were eventually cleaned up and now I guess a general convention can be applied, even though BIOS-clean IRQ schema still looks pretty exotic, as you noticed... With this patch, it looks like everything's taken care off.
Let me know if you need particular debug done with this patch. I'll send you a dmesg you've requested with increased CONFIG_LOG_BUF_SHIFT shortly (as soon as the system becomes available today).
Thanks, --Natalie
| |