boot problems on pseries

Hello,
	So I do most of my work on some IBM JS20 powerpc blades. They
usually run a git kernel on a stock RHEL4 distribution. The contents of
/proc/cpuinfo is:

processor       : 0
cpu             : PPC970, altivec supported
clock           : 1600.000000MHz
revision        : 2.2 (pvr 0039 0202)

processor       : 1
cpu             : PPC970, altivec supported
clock           : 1600.000000MHz
revision        : 2.2 (pvr 0039 0202)

timebase        : 199837651
platform        : pSeries
machine         : CHRP IBM,8842-21X


Recently, I discovered that newer kernels would no longer fully boot.
There's no oops or panic message and they get through most of the kernel
initialization before simply hanging. I spent some quality time today with
git-bisect, and traced the badness down to this commit:

0ebfff1491ef85d41ddf9c633834838be144f69f is first bad commit
diff-tree 0ebfff1491ef85d41ddf9c633834838be144f69f (from
f63e115fb50db39706b955b81e3375ef6bab2268)
Author: Benjamin Herrenschmidt [off-list ref]
Date:   Mon Jul 3 21:36:01 2006 +1000

    [POWERPC] Add new interrupt mapping core and change platforms to use it

    ...


A quick look through the patch didn't yield anything interesting to my eyes.


I tried booting a kernel built from the powerpc.git repository, and can
confirm that nothing in there fixes the problem. A search of this mailing
list didn't reveal anything either. Attached is my kernel .config and a cut
and paste of the boot log. Any insight would be greatly appreciated.
	--Mark

--
Mark Fasheh
Senior Software Developer, Oracle
mark.fasheh@oracle.com

On Tue, Aug 08, 2006 at 06:36:09PM -0700, Mark Fasheh wrote:

Recently, I discovered that newer kernels would no longer fully boot.

The IDE controller got irq32 in older kernels, now it gets irq17.
...
Uniform Multi-Platform E-IDE driver Revision: 7.00alpha2
ide: Assuming 33MHz system bus speed for PIO modes; override with idebus=xx
AMD8111: IDE controller at PCI slot 0000:00:04.1
AMD8111: chipset revision 3
AMD8111: 0000:00:04.1 (rev 03) UDMA133 controller
AMD8111: 100% native mode on irq 17
    ide0: BM-DMA at 0x7c00-0x7c07, BIOS settings: hda:pio, hdb:pio
    ide1: BM-DMA at 0x7c08-0x7c0f, BIOS settings: hdc:pio, hdd:pio
Probing IDE interface ide0...
hda: TOSHIBA MK6026GAXB, ATA DISK drive
ide0 at 0x7400-0x7407,0x6c02 on irq 17
Probing IDE interface ide1...
Probing IDE interface ide1...
hda: max request size: 128KiB
hda: lost interrupt
hda: lost interrupt
...

On Tue, 2006-08-08 at 18:36 -0700, Mark Fasheh wrote:

Hello,
	So I do most of my work on some IBM JS20 powerpc blades. They
usually run a git kernel on a stock RHEL4 distribution. The contents of
/proc/cpuinfo is:

processor       : 0
cpu             : PPC970, altivec supported
clock           : 1600.000000MHz
revision        : 2.2 (pvr 0039 0202)

processor       : 1
cpu             : PPC970, altivec supported
clock           : 1600.000000MHz
revision        : 2.2 (pvr 0039 0202)

timebase        : 199837651
platform        : pSeries
machine         : CHRP IBM,8842-21X


Recently, I discovered that newer kernels would no longer fully boot.
There's no oops or panic message and they get through most of the kernel
initialization before simply hanging. I spent some quality time today with
git-bisect, and traced the badness down to this commit:

I'm just back from vacation and looking into this. 

Ben.