Bug 10692
Summary: | mainboard INTEL D201GLY2A, pci devices only available with acpi=noirq | ||
---|---|---|---|
Product: | ACPI | Reporter: | Robert Goldner (robert) |
Component: | Config-Interrupts | Assignee: | ykzhao (yakui.zhao) |
Status: | REJECTED WILL_NOT_FIX | ||
Severity: | normal | CC: | acpi_bios |
Priority: | P1 | ||
Hardware: | All | ||
OS: | Linux | ||
Kernel Version: | 2.6.25.3 | Subsystem: | |
Regression: | --- | Bisected commit-id: | |
Attachments: |
output of acpidump with boot-parameter acpi=noirq
acpidump without acpi=noirq output of lspci -vxxxx with parameter acpi=noirq output of lspci without any parameter lspci -vxxx with acpi=noirq lspci -vxxx without any parameter output of acpidump dmesg output without any parameter |
Description
Robert Goldner
2008-05-14 01:33:28 UTC
Please attach the output of acpidump. Please attach the output of "lspci -vxxx" when the system is booted with/without "acpi=noirq". Thanks. Created attachment 16140 [details]
output of acpidump with boot-parameter acpi=noirq
Created attachment 16141 [details]
acpidump without acpi=noirq
It´s exact the same as with parameter acpi=noirq. I checked this two times.
Created attachment 16142 [details]
output of lspci -vxxxx with parameter acpi=noirq
Created attachment 16143 [details]
output of lspci without any parameter
Created attachment 16157 [details]
lspci -vxxx with acpi=noirq
Intel provided a new Bios release there days, but there is no change in this behaviour.
Created attachment 16158 [details]
lspci -vxxx without any parameter
Intel provided a new Bios release there days, but there is no change in this behaviour.
Created attachment 16159 [details]
output of acpidump
Intel provided a new Bios release there days, but there is no change in this behaviour.
Created attachment 16160 [details]
dmesg output without any parameter
Intel provided a new Bios release there days, but there is no change in this behaviour.
Hi, Robert From the acpidump it seems that there exists the following error in the bios. Method (_PRT, 0, NotSerialized) { If (And (PICM, LEqual (\_SB.PCI0.SBRG.SATS, 0x01))) { Return (AR00) } Else { Return (AR01) } If (\_SB.PCI0.SBRG.SATS) { Return (PR00) } Return (PR01) } General speaking ,the object AR00/AR01 will be returned only when the PICM is 1. But in the above AML code, the AR01 object will be returned when the PICM is zero. And unfortunately there is no PCI interrupt routing entry for some PCI devices in AR01 object. For example: wifi device At the same time the system works in PIC mode(The GSI number should be less than 16 in PIC mode). But the GSI number greater than 16 is returned by some PCI devices. For example: USB host controller: 20/21/22/23. It means that PCI interrupt routing is incorrect. Will you please try to build SMP kernel and see whether the problem still exists? (Notes: please enable the Local & I/O APIC in kernel configuration). Thanks. Hi, thanks for the hint with the SMP kernel. This works fine. I think we can close this bug. Thanks a lot for your very fast and good support. All the best Robert Thanks for the test. It seems that the bug is related with BIOS. But the system can work well on the SMP kernel(local & I/O APIC is enabled). So the bug will be rejected. |