Bug 7699 - irq9: nobody cared - regression 2.6.17 - VIA KT600 - PIC-mode
irq9: nobody cared - regression 2.6.17 - VIA KT600 - PIC-mode
Status: REJECTED INSUFFICIENT_DATA
Product: ACPI
Classification: Unclassified
Component: Config-Interrupts
i386 Linux
: P2 normal
Assigned To: Len Brown
:
Depends on:
Blocks:
  Show dependency treegraph
 
Reported: 2006-12-17 10:05 UTC by Emmanuel Chantr
Modified: 2007-09-05 15:11 UTC (History)
2 users (show)

See Also:
Kernel Version: 2.6.19.1
Tree: Mainline
Regression: ---


Attachments
lspci with 2.6.16 (1.48 KB, text/plain)
2006-12-17 10:08 UTC, Emmanuel Chantr
Details
cat /proc/interrupts with 2.6.16 (603 bytes, text/plain)
2006-12-17 10:09 UTC, Emmanuel Chantr
Details
screenshot 1 or acpi infos, before the fatal error message (10.81 KB, image/png)
2006-12-17 10:46 UTC, Emmanuel Chantr
Details
screenshot 2, the error message (31.56 KB, image/png)
2006-12-17 10:51 UTC, Emmanuel Chantr
Details
screenshot 3, just after the error message (28.62 KB, image/png)
2006-12-17 10:52 UTC, Emmanuel Chantr
Details
after, the kernel loop on this message (33.39 KB, image/png)
2006-12-17 10:57 UTC, Emmanuel Chantr
Details
acpidump (58.19 KB, text/plain)
2006-12-17 10:59 UTC, Emmanuel Chantr
Details
acpidump with 2.6.16 (58.19 KB, text/plain)
2006-12-17 10:59 UTC, Emmanuel Chantr
Details
config of the 2.6.19.1 kernel (43.02 KB, text/plain)
2006-12-17 11:00 UTC, Emmanuel Chantr
Details
2.6.19.1 success on KT400 box - dmesg (16.18 KB, text/plain)
2006-12-19 21:51 UTC, Len Brown
Details
dmessage for kernel 2.6.17 which have the problem (45.80 KB, text/plain)
2006-12-20 02:02 UTC, Emmanuel Chantr
Details
The kernel config file of the 2.6.17 (41.39 KB, text/plain)
2006-12-20 02:03 UTC, Emmanuel Chantr
Details
2.6.21-rc5 patch to remove irq compression (5.95 KB, patch)
2007-03-30 18:55 UTC, Len Brown
Details | Diff

Description Emmanuel Chantr 2006-12-17 10:05:04 UTC
Most recent kernel where this bug did *NOT* occur: 2.6.16 (not tested between
2.6.16 and 2.6.19)


Distribution: Debian but it's my own kernel not from Debian.

Hardware Environment:
-motherboard: A7V600-X-EAYKZ
-CPU: AMD Duron 750Mhz

Software Environment:
Same problem with 2.6.19 and 2.6.19.1

Problem Description: On boot, this message is printed:
> irq9: nobody cared (try booting with the "irqpoll" option).
I have no problem with the 2.6.16 and I have not changed the ACPI options.
With 2.6.19 I have tried the irqpoll option and the bug appeared also but I
don't remember if it's later or not; because I also tryed to revert the patch as
described in #7601, and in one case the "irqn: nobody cared" appaered later.
Comment 1 Emmanuel Chantr 2006-12-17 10:08:20 UTC
Created attachment 9845 [details]
lspci with 2.6.16

lspci output with 2.6.16
Comment 2 Emmanuel Chantr 2006-12-17 10:09:48 UTC
Created attachment 9846 [details]
cat /proc/interrupts with 2.6.16
Comment 3 Emmanuel Chantr 2006-12-17 10:46:08 UTC
Created attachment 9847 [details]
screenshot 1 or acpi infos, before the fatal error message
Comment 4 Emmanuel Chantr 2006-12-17 10:51:11 UTC
Created attachment 9848 [details]
screenshot 2, the error message
Comment 5 Emmanuel Chantr 2006-12-17 10:52:58 UTC
Created attachment 9849 [details]
screenshot 3, just after the error message
Comment 6 Emmanuel Chantr 2006-12-17 10:57:21 UTC
Created attachment 9850 [details]
after, the kernel loop on this message
Comment 7 Emmanuel Chantr 2006-12-17 10:59:00 UTC
Created attachment 9851 [details]
acpidump
Comment 8 Emmanuel Chantr 2006-12-17 10:59:17 UTC
Created attachment 9852 [details]
acpidump with 2.6.16
Comment 9 Emmanuel Chantr 2006-12-17 11:00:46 UTC
Created attachment 9853 [details]
config of the 2.6.19.1 kernel
Comment 10 Len Brown 2006-12-17 18:46:36 UTC
Please attach the full dmesg -s64000 (or serial console log
with "debug" enabled)and /proc/interrupts of the latest working kernel,
and the first failing kernel (failing kernel will need irqpoll to boot)

Confused about the comment regarding reverting patch in bug #7601 --
as that patch is already reverted in 2.6.19.1 -- it appeared only
in 2.6.19.

acpidump shows that this machine has an IOAPIC:

ACPI: APIC (v001 ASUS   A7V600-X 0x42302e31 MSFT 0x31313031) @ 0x(nil)
ACPI: IOAPIC (id[0x02] address[0xfec00000] global_irq_base[0x0])
ACPI: INT_SRC_OVR (bus 0 bus_irq 0 global_irq 2 dfl edge)
ACPI: INT_SRC_OVR (bus 0 bus_irq 9 global_irq 9 low level)
ACPI: LAPIC_NMI (acpi_id[0x00] high edge lint[0x1])
ACPI: LAPIC (acpi_id[0x00] lapic_id[0x00] enabled)
Length 90 OK
Checksum OK

however, the 2.6.16 /proc/interrupts shows PIC mode
is being used, and the config for 2.6.19.1
suggests that it isn't being used there either.

While this appears to be a PIC-mode regression, and worthy of
being debugged, it would also be interesting to try IOAPIC
mode on the old and new kernel as well -- simple to enble
this with CONFIG_SMP.  As IOAPIC mode is preferable if it works.
The MADT shows that the APIC-mode entries are all hard-coded
so from an ACPI point of view, it should be very simple
and reliable -- and the only obstacle might be the dreaded VIA quirk.
Comment 11 Len Brown 2006-12-19 21:51:00 UTC
Created attachment 9892 [details]
2.6.19.1 success on KT400 box - dmesg

I dug up a KT400 box to see if I could reproduce this failure,
but it works for me.

00:00.0 Host bridge: VIA Technologies, Inc. VT8377 [KT400/KT600 AGP] Host
Bridge
00:01.0 PCI bridge: VIA Technologies, Inc. VT8235 PCI Bridge
00:10.0 USB Controller: VIA Technologies, Inc. VT82xxxxx UHCI USB 1.1
Controller (rev 80)
00:10.1 USB Controller: VIA Technologies, Inc. VT82xxxxx UHCI USB 1.1
Controller (rev 80)
00:10.2 USB Controller: VIA Technologies, Inc. VT82xxxxx UHCI USB 1.1
Controller (rev 80)
00:10.3 USB Controller: VIA Technologies, Inc. USB 2.0 (rev 82)
00:11.0 ISA bridge: VIA Technologies, Inc. VT8235 ISA Bridge
00:11.1 IDE interface: VIA Technologies, Inc.
VT82C586A/B/VT82C686/A/B/VT823x/A/C PIPC Bus Master IDE (rev 06)
00:11.5 Multimedia audio controller: VIA Technologies, Inc. VT8233/A/8235/8237
AC97 Audio Controller (rev 50)
00:13.0 Ethernet controller: Realtek Semiconductor Co., Ltd.
RTL-8139/8139C/8139C+ (rev 10)
01:00.0 VGA compatible controller: ATI Technologies Inc Rage 128 Pro Ultra TF
	   CPU0
  0:	  49914    XT-PIC-XT	    timer
  1:	     10    XT-PIC-XT	    i8042
  2:	      0    XT-PIC-XT	    cascade
  4:	    736    XT-PIC-XT	    serial
  6:	      5    XT-PIC-XT	    floppy
  7:	    121    XT-PIC-XT	    parport0
  8:	      0    XT-PIC-XT	    rtc
  9:	      0    XT-PIC-XT	    acpi
 10:	      0    XT-PIC-XT	    uhci_hcd:usb1, uhci_hcd:usb2
 11:	   3419    XT-PIC-XT	    uhci_hcd:usb3, VIA8233, ehci_hcd:usb4, eth0

 12:	    131    XT-PIC-XT	    i8042
 14:	   4431    XT-PIC-XT	    ide0
 15:	   1569    XT-PIC-XT	    ide1
NMI:	      0
LOC:	  49850
ERR:	     97
MIS:	      0

(It also works fine in IOAPIC mode)
	   CPU0
  0:	   8589   IO-APIC-edge	    timer
  1:	      8   IO-APIC-edge	    i8042
  4:	    352   IO-APIC-edge	    serial
  6:	      5   IO-APIC-edge	    floppy
  7:	      0   IO-APIC-edge	    parport0
  8:	      0   IO-APIC-edge	    rtc
  9:	      0   IO-APIC-fasteoi   acpi
 12:	    131   IO-APIC-edge	    i8042
 14:	   3140   IO-APIC-edge	    ide0
 15:	     92   IO-APIC-edge	    ide1
 16:	    272   IO-APIC-fasteoi   eth0
 18:	      0   IO-APIC-fasteoi   uhci_hcd:usb1, uhci_hcd:usb2,
uhci_hcd:usb3, ehci_hcd:usb4
 19:	      0   IO-APIC-fasteoi   VIA8233
NMI:	      0
LOC:	   8515
ERR:	      0
MIS:	      0
Comment 12 Len Brown 2006-12-19 22:03:25 UTC
> ERR:         97
This is from:
> spurious 8259A interrupt: IRQ7.
And have been present as long as I've had this board.
It is the motherboard's way of telling the OS to use IOAPIC mode
instead of PIC mode:-)

Any progress identifying the latest working and first failing kernel
on the A7V600-X-EAYKZ, per comment #10?
Comment 13 Emmanuel Chantr 2006-12-20 02:02:25 UTC
Created attachment 9894 [details]
dmessage for kernel 2.6.17 which have the problem

Compiled with debug, the config for this kernel is following. So the kernel
2.6.16 is working and not the 2.6.17. I have not check with kernel 2.6.17.rc*
or 2.6.16.*, do you want that I check these kernels too ?

>Confused about the comment regarding reverting patch in bug #7601 --
>as that patch is already reverted in 2.6.19.1 -- it appeared only
>in 2.6.19.

the reverting patch was applied on 2.6.19
Comment 14 Emmanuel Chantr 2006-12-20 02:03:34 UTC
Created attachment 9895 [details]
The kernel config file of the 2.6.17
Comment 15 Emmanuel Chantr 2006-12-23 05:01:30 UTC
Comment on attachment 9894 [details]
dmessage for kernel 2.6.17 which have the problem

2.6.36 has not the problem so 2.6.17 is the first with the problem. It's
compiled with the debug option.
Comment 16 Len Brown 2007-03-30 18:52:26 UTC
lsppci: 0000:00:0f.1 IDE interface: VIA Technologies, Inc. 
VT82C586A/B/VT82C686/A/B/VT823x/A/C PIPC Bus Master IDE (rev 06)

dmesg:
VP_IDE: IDE controller at PCI slot 0000:00:0f.1
ACPI: PCI Interrupt Link [LNKE] enabled at IRQ 9
PCI: setting IRQ 9 as level-triggered
ACPI: PCI Interrupt 0000:00:0f.1[A] -> Link [LNKE] -> GSI 9 (level, low) -> IRQ 
9
PCI: VIA IRQ fixup for 0000:00:0f.1, from 14 to 9

Please attach the output from lspci -vv (running on any release)
so I can verify that IDE at 0000:00:0f.1 is indeed
associated with LNKE.

Last time I checked, IDE was hard-coded to IRQ14, so if the hardware
is trying to route this to IRQ9, it makes perfect sense that when
IDE starts it nukes ACPI, which can't handle the spurious interrupts
on IRQ9.
Comment 17 Len Brown 2007-03-30 18:55:21 UTC
Created attachment 11002 [details]
2.6.21-rc5 patch to remove irq compression

> 2.6.17 is the first with the problem

Interesting.  That is when "irq compression" was added.
Please reproduce the issue with the latest kernel -- 2.6.21-rc5
and then apply the this patch to remove irq compression and re-test.
Comment 18 Len Brown 2007-09-05 15:11:16 UTC
please re-open if this issue is still present in 2.6.22.stable or later

Note You need to log in before you can comment on or make changes to this bug.