I think you are somewhat right about that. Right now my last hope is to make 2.4.22-rc2 compile with the latest acpi patch.
2.4.22-rc2 has been running for four days down without a problem. I had to remove most of the device (logical remove via rmmod) because every device that uses an interrupt that is loaded in the acpi space (irq 16 and above) generates billions of spurious interrupts every day. Last I looked my system had trillions of interrupts generated within the last few days. Anyway each device loaded in the acpi space causes the sytem to run slower and slower. Right now only two devices loaded in that area, and it runs about 1/3 the speed that I think it should be running.
If this latest acpi patch fixes this problem without reintroducing my lockup problem, I will be in business.
It is strange that 2.4.22-rc2 does't lockup, but if I run it with pci=noapic or acpi=off, my lockup problem comes back. Most of the people I know with this problem say that interrupts are somehow causing the problem. I have to agree with that from what I see now. Everybody that I have found with this problem have the nforce2 chipset revision 162.