irq 3: nobody cared! with Intel 31244 SATA.... Advice??

irq 3: nobody cared! with Intel 31244 SATA.... Advice??

Post by dps.. » Tue, 25 Jan 2005 05:57:31



Hi folks,

I have a Dell Powervault 745N appliance that I am trying to install
Slackware Linux on....  I have temporarily installed a dual port IDE
card in the expansion slot so I can have the benefit of a CD-ROM and
IDE PATA HD.... Slackware 9.1 installs straightforwardly to the IDE
PATA HD and it boots up fine....

I first tried to get the SATA disk talking with the 2.4.29 kernel...
This kernel boots but hangs when it tries to load module sata_vsc
either from rc.modules or manually after startup using "modprobe
sata_vsc"....

I then downloaded and built kernel 2.6.10 which boots up fine without
the
sata_vsc module.... If you then load the sata_vsc module manually using
"modprobe sata_vsc" it will cause the following error once for each
attached disk drive:

Jan 23 09:08:21 linux kernel: Disabling IRQ #3
Jan 23 09:08:23 linux kernel: irq 3: nobody cared!
Jan 23 09:08:23 linux kernel:  [<c0127a62>] __report_bad_irq+0x22/0x90
Jan 23 09:08:23 linux kernel:  [<c0127b58>] note_interrupt+0x58/0x90
Jan 23 09:08:23 linux kernel:  [<c01276e8>] __do_IRQ+0xd8/0xe0
Jan 23 09:08:23 linux kernel:  [<c0103a6a>] do_IRQ+0x1a/0x30
Jan 23 09:08:23 linux kernel:  [<c010254a>] common_interrupt+0x1a/0x20
Jan 23 09:08:23 linux kernel:  [<c0114880>] __do_softirq+0x30/0x90
Jan 23 09:08:23 linux kernel:  [<c0114915>] do_softirq+0x35/0x40
Jan 23 09:08:23 linux kernel:  [<c0103a6f>] do_IRQ+0x1f/0x30
Jan 23 09:08:23 linux kernel:  [<c010254a>] common_interrupt+0x1a/0x20
Jan 23 09:08:23 linux kernel:  [<c0100590>] default_idle+0x0/0x40
Jan 23 09:08:23 linux kernel:  [<c01005b4>] default_idle+0x24/0x40
Jan 23 09:08:23 linux kernel:  [<c010063e>] cpu_idle+0x2e/0x40
Jan 23 09:08:23 linux kernel:  [<c03d077b>] start_kernel+0x15b/0x190
Jan 23 09:08:23 linux kernel: handlers:
Jan 23 09:08:23 linux kernel: [<c0245a30>] (ide_intr+0x0/0x120)
Jan 23 09:08:23 linux kernel: [<c0245a30>] (ide_intr+0x0/0x120)
Jan 23 09:08:23 linux kernel: [<e08ef250>] (vsc_sata_interrupt+0x0/0xa0
[sata_vsc])

but it will then load and work ok and allow me to fdisk, format and
mount
the disk normally.....

If I run 'modprobe sata_vsc' from rc.modules, then it hangs while
trying
to talk to the first disk just like kernel 2.4.29 did.... I have
repetitively seen this behavior both with a module and with the module
builtin to the kernel.... I have also built the kernel with and without
ACPI as some of the traffic here indicated that could be a problem...
This
caused no change in behavior....

Since my ambition is to get this machine to boot from the SATA disk I
am
stuck at this point.... If anyone can provide a little advice or
direction
I would be indebted... I have fiddled with the PCI BIOS IRQ settings
without joy... In the BIOS you can change which IRQ is used by the sata
controller but then the error message just changes to that IRQ (i.e.,
irq 7: nobody cared!).... It may be of interest to note that the SATA
controller (Intel
31244) and the dual port IDE card in the expansion slot seem to be
forced
in the BIOS to share the IRQ that is assigned to either one...

Again, any advice would be very welcome at this point....
TIA,

Dave Sims
Houston, Texas
USA

 
 
 

irq 3: nobody cared! with Intel 31244 SATA.... Advice??

Post by simonbachmann at bluewin dot c » Thu, 27 Jan 2005 06:06:39



Quote:> Hi folks,

> I have a Dell Powervault 745N appliance that I am trying to install
> Slackware Linux on....  I have temporarily installed a dual port IDE
> card in the expansion slot so I can have the benefit of a CD-ROM and
> IDE PATA HD.... Slackware 9.1 installs straightforwardly to the IDE
> PATA HD and it boots up fine....

> I first tried to get the SATA disk talking with the 2.4.29 kernel...
> This kernel boots but hangs when it tries to load module sata_vsc
> either from rc.modules or manually after startup using "modprobe
> sata_vsc"....

> I then downloaded and built kernel 2.6.10 which boots up fine without
> the
> sata_vsc module.... If you then load the sata_vsc module manually using
> "modprobe sata_vsc" it will cause the following error once for each
> attached disk drive:

> Jan 23 09:08:21 linux kernel: Disabling IRQ #3
> Jan 23 09:08:23 linux kernel: irq 3: nobody cared!
> Jan 23 09:08:23 linux kernel:  [<c0127a62>] __report_bad_irq+0x22/0x90
> Jan 23 09:08:23 linux kernel:  [<c0127b58>] note_interrupt+0x58/0x90
> Jan 23 09:08:23 linux kernel:  [<c01276e8>] __do_IRQ+0xd8/0xe0
> Jan 23 09:08:23 linux kernel:  [<c0103a6a>] do_IRQ+0x1a/0x30
> Jan 23 09:08:23 linux kernel:  [<c010254a>] common_interrupt+0x1a/0x20
> Jan 23 09:08:23 linux kernel:  [<c0114880>] __do_softirq+0x30/0x90
> Jan 23 09:08:23 linux kernel:  [<c0114915>] do_softirq+0x35/0x40
> Jan 23 09:08:23 linux kernel:  [<c0103a6f>] do_IRQ+0x1f/0x30
> Jan 23 09:08:23 linux kernel:  [<c010254a>] common_interrupt+0x1a/0x20
> Jan 23 09:08:23 linux kernel:  [<c0100590>] default_idle+0x0/0x40
> Jan 23 09:08:23 linux kernel:  [<c01005b4>] default_idle+0x24/0x40
> Jan 23 09:08:23 linux kernel:  [<c010063e>] cpu_idle+0x2e/0x40
> Jan 23 09:08:23 linux kernel:  [<c03d077b>] start_kernel+0x15b/0x190
> Jan 23 09:08:23 linux kernel: handlers:
> Jan 23 09:08:23 linux kernel: [<c0245a30>] (ide_intr+0x0/0x120)
> Jan 23 09:08:23 linux kernel: [<c0245a30>] (ide_intr+0x0/0x120)
> Jan 23 09:08:23 linux kernel: [<e08ef250>] (vsc_sata_interrupt+0x0/0xa0
> [sata_vsc])

> but it will then load and work ok and allow me to fdisk, format and
> mount
> the disk normally.....

> If I run 'modprobe sata_vsc' from rc.modules, then it hangs while
> trying
> to talk to the first disk just like kernel 2.4.29 did.... I have
> repetitively seen this behavior both with a module and with the module
> builtin to the kernel.... I have also built the kernel with and without
> ACPI as some of the traffic here indicated that could be a problem...
> This
> caused no change in behavior....

> Since my ambition is to get this machine to boot from the SATA disk I
> am
> stuck at this point.... If anyone can provide a little advice or
> direction
> I would be indebted... I have fiddled with the PCI BIOS IRQ settings
> without joy... In the BIOS you can change which IRQ is used by the sata
> controller but then the error message just changes to that IRQ (i.e.,
> irq 7: nobody cared!).... It may be of interest to note that the SATA
> controller (Intel
> 31244) and the dual port IDE card in the expansion slot seem to be
> forced
> in the BIOS to share the IRQ that is assigned to either one...

> Again, any advice would be very welcome at this point....
> TIA,

> Dave Sims
> Houston, Texas
> USA

I did not understand a lot of your problem. Just try to help you sharing
an experience I had with IRQ's even if my problem was really different.
It seems enabling /disabling the APIC / IO-APIC in the kernel changes
the IRQ mapping and some drivers don't like that....

Ciao

Simon