Re: PROBLEM: Can't boot a (HZ = 1000) kernel using an AMD Phenom-II processor

From: Bartlomiej Zolnierkiewicz
Date: Mon Jan 19 2009 - 08:00:21 EST


On Monday 19 January 2009, Mark Hounschell wrote:
> Mark Hounschell wrote:
> > I upgraded from a Phenom to a Phenom-II processor and I could no longer boot.
> >
> > I have tried kernels from 2.6.26 through 2.6.28.
> >
> > I _can_ however boot these kernels when they are configured for HZ = 250. I
> > haven't tried HZ = 300. But certainly if HZ = 1000 the boot hangs at the spot
> > indicated below in the dmesg output.
> >
> > The machine is up and stable running 2.6.26.8 with a (HZ = 1000) configured kernel.
> >
>
> The machine is up and stable running 2.6.26.8 with a (HZ = 250) configured
> kernel. Sorry.
>
> > While running a 2.6.26.8 kernel configured for HZ = 250:

[...]

> > Checking if this processor honours the WP bit even in supervisor mode...Ok.
> >
> >
> > The 2.6.26.8 kernel hangs up right here.
> >
> >
> > CPA: page pool initialized 1 of 1 pages preallocated
> > SLUB: Genslabs=12, HWalign=64, Order=0-3, MinObjects=0, CPUs=4, Nodes=1
> > hpet clockevent registered
> > calibrate_delay_direct() failed to get a good estimate for loops_per_jiffy.
> > Probably due to long platform interrupts. Consider using "lpj=" boot option.
> > Calibrating delay loop... 3956.73 BogoMIPS (lpj=7913472)

Even if it boots it may not work reliably since lpj/BogoMIPS are too small
(they should be the same as for other CPUs) which in turn could result in too
short delay loops.

> > Security Framework initialized
> > Capability LSM initialized
> > Mount-cache hash table entries: 512
> > CPU: L1 I Cache: 64K (64 bytes/line), D cache 64K (64 bytes/line)
> > CPU: L2 Cache: 512K (64 bytes/line)
> > CPU 0(4) -> Core 0
> > Checking 'hlt' instruction... OK.
> > ACPI: Core revision 20080321
> > ENABLING IO-APIC IRQs
> > ..TIMER: vector=0x31 apic1=0 pin1=2 apic2=-1 pin2=-1
> > CPU0: AMD Processor model unknown stepping 02
> > Booting processor 1/1 ip 2000
> > Initializing CPU#1
> > Calibrating delay using timer specific routine.. 6000.60 BogoMIPS (lpj=12001200)

Here are the proper values.

> > CPU: L1 I Cache: 64K (64 bytes/line), D cache 64K (64 bytes/line)
> > CPU: L2 Cache: 512K (64 bytes/line)
> > CPU 1(4) -> Core 2
> > x86 PAT enabled: cpu 1, old 0x7040600070406, new 0x7010600070106
> > CPU1: AMD Processor model unknown stepping 02
> > Booting processor 2/2 ip 2000
> > Initializing CPU#2
> > Calibrating delay using timer specific routine.. 6000.47 BogoMIPS (lpj=12000950)

also OK

> > CPU: L1 I Cache: 64K (64 bytes/line), D cache 64K (64 bytes/line)
> > CPU: L2 Cache: 512K (64 bytes/line)
> > CPU 2(4) -> Core 1
> > x86 PAT enabled: cpu 2, old 0x7040600070406, new 0x7010600070106
> > CPU2: AMD Processor model unknown stepping 02
> > Booting processor 3/3 ip 2000
> > Initializing CPU#3
> > Calibrating delay using timer specific routine.. 6000.47 BogoMIPS (lpj=12000955)

ditto

That being said I don't know why delay calibration loop fails on CPU0.

You may try booting with "lpj=12001200" kernel parameter to workaround
the issue (it should fix booting with HZ == 1000) and/or try 2.6.29-rc2
(the problem may have been already fixed).

Thanks,
Bart
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/