Re: apic-to-drivermodel conversion

Pavel Machek writes:
 > > Try this patch instead. This is my merge of my and Pavel's earlier
 > > patches, so it differs from Pavel's in some details.
 > 
 > I tried it... Oops at kobject_add, called from device_add,
 > device_initialize, ..., init. This is on machine that does have
 > lapic. I'll try to do clean build to see if it goes away.
 > 
 > No, it still dies after ds: no socket drivers.

When does it die? At initial boot or at resume?

I don't claim that every driver that's hooked into the driver
model actually handles suspend/resume correctly, but this patch
does migrate APM/apic.c/nmi.c from the old PM system to the
driver model. I've tested it on plain P3 & P4 desktop machines.

Try a minimal .config.

/Mikael