Bug 7695

Summary:	Supermicro X7DB8+ Boot failure
Product:	ACPI	Reporter:	Len Brown (lenb)
Component:	Config-Interrupts	Assignee:	Zhang Rui (rui.zhang)
Status:	CLOSED CODE_FIX
Severity:	normal	CC:	acpi-bugzilla
Priority:	P2
Hardware:	i386
OS:	Linux
Kernel Version:	2.6.20-rc1+sysfs patch	Subsystem:
Regression:	---	Bisected commit-id:
Attachments:	console.log.2.6.20-rc1 console.log.2.6.20-rc+sysfs-patch acpidump Rui's debug patch for 2.6.20-rc1 2.6.20-rc1 dmest with above pach included patch-hadle-_ADR-Based-devices patch against 2.6.20-rc1-mm dmesg

Description Len Brown 2006-12-16 22:21:52 UTC

A Supermicro X7DB8+ does not boot -- SCSI driver dies,
apparently due to lack of interrupts.  First interesting
difference in dmest without and with the sysfs patch below:


< ACPI: PCI Interrupt Routing Table [\_SB_.PCI0.P0P2._PRT]
< ACPI: PCI Interrupt Routing Table [\_SB_.PCI0.P0P2.BMD0._PRT]
< ACPI: PCI Interrupt Routing Table [\_SB_.PCI0.P0P2.BMD0.BPD0.PXH0._PRT]
< ACPI: PCI Interrupt Routing Table [\_SB_.PCI0.P0P2.BMD0.BPD0.PXH1._PRT]
< ACPI: PCI Interrupt Routing Table [\_SB_.PCI0.P0P2.BMD0.BPD2._PRT]
< ACPI: PCI Interrupt Routing Table [\_SB_.PCI0.P0P2.BMF3._PRT]
< ACPI: PCI Interrupt Routing Table [\_SB_.PCI0.P0P4._PRT]
< ACPI: PCI Interrupt Routing Table [\_SB_.PCI0.P0P6._PRT]
< ACPI: PCI Interrupt Routing Table [\_SB_.PCI0.PEX0._PRT]
< ACPI: PCI Interrupt Routing Table [\_SB_.PCI0.PCIB._PRT]
268a318,327
> ACPI: PCI Interrupt Routing Table [\_SB_.PCI0.P0P2._PRT]
> ACPI: PCI Interrupt Routing Table [\_SB_.PCI0.P0P2.BMD0._PRT]
> ACPI Exception (pci_bind-0153): AE_NOT_FOUND, Invalid ACPI-PCI context for 
parent device BPD0 [20061109]
> ACPI Exception (pci_bind-0153): AE_NOT_FOUND, Invalid ACPI-PCI context for 
parent device BPD0 [20061109]
> ACPI: PCI Interrupt Routing Table [\_SB_.PCI0.P0P2.BMD0.BPD2._PRT]
> ACPI: PCI Interrupt Routing Table [\_SB_.PCI0.P0P2.BMF3._PRT]
> ACPI: PCI Interrupt Routing Table [\_SB_.PCI0.P0P4._PRT]
> ACPI: PCI Interrupt Routing Table [\_SB_.PCI0.P0P6._PRT]
> ACPI: PCI Interrupt Routing Table [\_SB_.PCI0.PEX0._PRT]
> ACPI: PCI Interrupt Routing Table [\_SB_.PCI0.PCIB._PRT]

Comment 1 Len Brown 2006-12-16 22:23:38 UTC

Created attachment 9834 [details]
console.log.2.6.20-rc1

Comment 2 Len Brown 2006-12-16 22:24:37 UTC

Created attachment 9835 [details]
console.log.2.6.20-rc+sysfs-patch

Comment 3 Len Brown 2006-12-16 22:25:45 UTC

Created attachment 9836 [details]
acpidump

Comment 4 Len Brown 2006-12-16 22:32:21 UTC

Created attachment 9837 [details]
Rui's debug patch for 2.6.20-rc1

Comment 5 Len Brown 2006-12-16 22:42:47 UTC

Created attachment 9838 [details]
2.6.20-rc1 dmest with above pach included

Comment 6 Zhang Rui 2006-12-18 03:43:34 UTC

Created attachment 9862 [details]
patch-hadle-_ADR-Based-devices

Comment 7 Zhang Rui 2006-12-18 03:51:59 UTC

According to the ACPIDUMP above, some _ADR-Based deivces may not have _PRT, 
while we suppose they should.
So we didn't bind this kind of devices in the former patches.
This patch will fix this problem.

Len, can you try this patch based on the sysfs branch?
And, please don't remove the sysfs patches from acpi-test. It's never gonna be 
stable without tests.
Thanks

Comment 8 Zhang Rui 2006-12-19 00:58:15 UTC

Created attachment 9873 [details]
patch against 2.6.20-rc1-mm

The last patch can only be applied to -rc1+sysfs
This one is against -mm.

Comment 9 Zhang Rui 2006-12-20 23:21:39 UTC

Hi, Len
I won't generate a new patch as the patch in comment #8 seems to be the most 
reasonable way to fix this bug.
Can you have a try? :)

Comment 10 Len Brown 2006-12-20 23:44:23 UTC

Created attachment 9909 [details]
dmesg

with patch from comment #4 the supermicro box boots.
patch applied to sysfs branch.

Comment 11 Len Brown 2007-01-03 18:08:52 UTC

This bug was against 2.6.20-rc1-mm1
and the fix shipped in 2.6.20-rc2-mm1