Disabling IRQ #10 hang after compiling 2.6.10

godhugh · 01-03-2005, 02:00 PM

I was having some issues getting eth0 and my sound card detected after compiling 2.6.10 so I decided to reinstall Slackware 10 and compile 2.6.10 knowing exactly what I needed to select.

After choosing all the drivers for my ethernet card, my on-board sound, my SATA controller (SiI SATA controller), and ACPI support I compiled.

Here are my system specs:

Processor: Intel Pentium 4 2.4ghz
Motherboard: Intel D845PEBT2
Chipset: Intel i845PE
Memory: 1GB DDR3200
Video Card: NVidia Geforce FX 5900XT
Sound: Onboard 82801DB/DBM Audio Controller
Hard Drives: 120GB SATA Maxtor (1 partition with Win XP), 75GB SATA Seagate (un-partitioned, planning to put FC3 on here)
SATA Controller: Silicon Image SiI 3112 SATALink Controller
IDE drives: Lite-On DVD Player (Secondary Master), Plextor PX-708A (Secondary Slave)

After a reboot and choosing the new kernel, This shows up and the system hangs...

Code:

GDT-HA: Storage RAID Controller Driver, Version 3.04
GDT-HA: Found 0 PCI Storage RAID Controllers
scsi0: scsi_debug, version 1.74 [20040829], dev_size_mb=8 opts=0x0
        Vendor: Linux      Model: scsi_debug           Rev: 0004
        Type:  Direct Access
PCI: Found IRQ10 for device 0000:02:06.0
ata1: SATA max UDMA/100 cmd 0xF8818C80 ctl 0xF8818C8A bmdma 0xF8818C00 irq 10
ata2: SATA max UDMA/100 cmd 0xF8818CC0 ctl 0xF8818CCA bmdma 0xF8818C08 irq 10
irq 10: nobody cared!
[<c013209a>] _report_bad_irq+0x2a/0xa0
[<c0131ab0>] handle_IRQ_event+0x30/0x70
                        note_interrupt+0x70/0x60
                        __do_IRQ_0x128/0x140
                        do_IRQ+0x19/0x30
                        common_interrupt+0x1a/0x20
                        __do_soft_irq+0x2e/0x90
                        do_soft_irq+0x27/0x30
                        irq_exit+0x35/0x40
                        do_IRQ=0x1e/0x30
                        common_interrupt+0x1a/0x20
                        default_idle+0x0/0x30
                        default_idle+0x23/0x30
                        cpu_idle+0x3a/0x60
                        start_kernel+0x16b/0x1b0
                        unknown_bootoption+0x0/0x1e0
handlers:
[<co31f1a0>] (atn_interrupt=0x0/0x1d0)
Disabling IRQ #10
ata1: dev 0 ATA, max UDMA/133, 240121728 sectors

The system just hangs there.

I've tried compiling the kernel without hotplug and ACPI. I've tried booting the kernel with "acpi=off" in LILO. I've tried turning off PnP OS in the BIOS and I've tried assigning PCI IRQ's manually in the BIOS.

I'm at my wits end here and I don't know what else to do.

Anyone have any ideas?

Cedrik · 01-03-2005, 02:12 PM

I would try append="pci=routeirq" in lilo because a message says that at boot:

Quote:

PCI: Using ACPI for IRQ routing
* PCI interrupts are no longer routed automatically. If this
** driver failed to call pci_enable_device(). As a temporary
** workaround, the "pci=routeirq" argument restores the old
** please email the output of "lspci" to bjorn.helgaas@hp.com

godhugh · 01-03-2005, 02:18 PM

Okay, I just tried that and I still have the same problem

.

Cedrik · 01-03-2005, 02:26 PM

Are you sure you do the correct syntax to add kernel parameters ? Could you post your lilo.conf
part with your kernel parameter.

Sorry for asking that but I see a suspicious line in your logs :

Quote:

...
start_kernel+0x16b/0x1b0
unknown_bootoption+0x0/0x1e0

godhugh · 01-03-2005, 02:46 PM

Ok, here it is.

Remember, I'm pretty new at this so if I've totally screwed it up try not to laugh to hard

.

Code:

# Windows bootable partition config begins
other = /dev/hda1
  label = Windows
  table = /dev/hda
# Windows bootable partition config ends
# Linux bootable partition config begins
image = /boot/vmlinuz-ide-2.4.26
  root = /dev/hde2
  label = Linux-2.4.26
  read-only
# Linux bootable partition config ends

# Linux bootable partition config begins
image = /boot/vmlinuz
  root = /dev/hde2
  label = Linux-2.6.10
  read-only
  append="acpi=off pci=routeirq pci=noacpi"
# Linux bootable partition config ends

Cedrik · 01-03-2005, 03:01 PM

That seems fine for the append option
As I don't use lilo tough, I don't know if append="..." and append = "..." does matter (spaces
before and after the equal sign)...

I guess something goes wrong with your sata driver anyway, I even would try a 2.6.9 kernel
compile to see if that change something

urka58 · 01-03-2005, 04:16 PM

Assuming you configured the sata driver "built in" the kernel, you should try configuring the PCI Bus option to "BIOS" instead of the default "ANY".
This is an extract of my .config file (2.6.10)

Bus options (PCI, PCMCIA, EISA, MCA, ISA)
#
CONFIG_PCI=y
CONFIG_PCI_GOBIOS=y
# CONFIG_PCI_GOMMCONFIG is not set
# CONFIG_PCI_GODIRECT is not set
# CONFIG_PCI_GOANY is not set
CONFIG_PCI_BIOS=y
CONFIG_PCI_LEGACY_PROC=y
CONFIG_PCI_NAMES=y
# CONFIG_ISA is not set
# CONFIG_MCA is not set
# CONFIG_SCx200 is not set

I would also remove the parameters you are passing to the kernel in your /etc/lilo.conf
Hope this helps
Ciao

godhugh · 01-04-2005, 11:29 AM

I did a compile of 2.6.9 and it worked like a charm (other then a minor ALSA issue). I didn't use my old .config file this time and just went through and selected only those options I needed. For some reason that did it.

For now, I'm just happy to have a working version of a 2.6 kernel. It's taken 3 days and been kinda frustrating but I learned a hell of a lot.

Thanks for you help!