Re: Scheduling while atomic errors on swsusp resume

2005-03-04 Thread Pavel Machek
Hi! > > Well, those are warnings, so it still works, right? Aha, "exited with > > preempt count 1" seems very wrong. Yes, please try this with > > vanilla. I'm running 2.6.11 with > > Yeah, the resume script crashes, which is a bit of a problem. I'll get > the user to try with a vanilla kernel,

Re: Scheduling while atomic errors on swsusp resume

2005-03-04 Thread Matthew Garrett
On Fri, 2005-03-04 at 12:26 +0100, Pavel Machek wrote: > Well, those are warnings, so it still works, right? Aha, "exited with > preempt count 1" seems very wrong. Yes, please try this with > vanilla. I'm running 2.6.11 with Yeah, the resume script crashes, which is a bit of a problem. I'll get

Re: Scheduling while atomic errors on swsusp resume

2005-03-04 Thread Pavel Machek
Hi! > Using the current Ubuntu development kernel (2.6.10 with acpi and swsusp > stuff backported from 2.6.11), a user is getting the following trace on > resume. Passing noapic nolapic removes the APIC error, but the rest of > the trace is identical. This is reproducible, but only seems to happen

Scheduling while atomic errors on swsusp resume

2005-03-02 Thread Matthew Garrett
Using the current Ubuntu development kernel (2.6.10 with acpi and swsusp stuff backported from 2.6.11), a user is getting the following trace on resume. Passing noapic nolapic removes the APIC error, but the rest of the trace is identical. This is reproducible, but only seems to happen on this mach