Re: PCI error on x86_64 2.6.13 kernel

Re: PCI error on x86_64 2.6.13 kernel

[Niklas Kallman]

Jack Howarth wrote:
>    Has anyone reported the following? For both of the 2.6.13 based
> kernels released so far on Fedora Core 4 for x86_64, we are seeing
> error messages of the form...
> 
> Oct  3 11:21:48 XXXXX  kernel:   MEM window: d0200000-d02fffff
> Oct  3 11:21:48 XXXXX  kernel:   PREFETCH window: disabled.
> Oct  3 11:21:48 XXXXX  kernel: PCI: Failed to allocate mem resource #6:20000 f0000000 for 0000:09:00.0
> 
> ..on a Sun W2100Z dual opteron. These memory allocations errors 
> don't occur under the 2.6.12 based kernels that Fedora has released 
> for FC4. So far the errors don't seem to be manifesting themselves 
> as any noticable system errors in using the machine itself.
>                     Jack
> -

I have the same problem on a Dell Inspiron 510m laptop and Fedora Core 4 
when upgrading from 2.6.12 to 2.6.13.* :

I get the following messages at 2.6.13.* startup:

PCI: Ignore bogus resource 6 [0:0] of 0000:00:02.0
PCI: Failed to allocate I/O resource #7:1000@10000 for 0000:01:01.0
PCI: Failed to allocate I/O resource #8:1000@10000 for 0000:01:01.0
PCI: Bus 2, cardbus bridge: 0000:01:01.0

I have tried "pci=routeirq", but it makes no difference.
When running 2.6.13.4 I cannot use my WLAN card (ndiswrapper).
With 2.6.12.6 there is no problem at all...

I posted lspci output and more in another mail with subject:
PCI: Failed to allocate I/O resource when upgrading to 2.6.13.*

I'm starting to suspect something in Fedora Core 4 ..

/Niklas

Re: PCI error on x86_64 2.6.13 kernel

[Rajesh Shah]

On Wed, Oct 26, 2005 at 11:06:05PM +0200, Niklas Kallman wrote:
> Jack Howarth wrote:
> >    Has anyone reported the following? For both of the 2.6.13 based
> > kernels released so far on Fedora Core 4 for x86_64, we are seeing
> > error messages of the form...
> > 
> > Oct  3 11:21:48 XXXXX  kernel:   MEM window: d0200000-d02fffff
> > Oct  3 11:21:48 XXXXX  kernel:   PREFETCH window: disabled.
> > Oct  3 11:21:48 XXXXX  kernel: PCI: Failed to allocate mem resource #6:20000 f0000000 for 0000:09:00.0
> > 

I ran into a similar problem, and posted a fix, see
http://marc.theaimsgroup.com/?l=linux-pci&m=113225006603745&w=2

Can you try it to see if this problem goes away?

thanks,
Rajesh

Re: PCI error on x86_64 2.6.13 kernel

[Voluspa]

On 2005-11-18 0:24:51 Rajesh Shah wrote:
>On Wed, Oct 26, 2005 at 11:06:05PM +0200, Niklas Kallman wrote:
>>Jack Howarth wrote:
>>> Has anyone reported the following? For both of the 2.6.13 based
>>> kernels released so far on Fedora Core 4 for x86_64, we are seeing
>>> error messages of the form...
>>> 
>>> Oct  3 11:21:48 XXXXX  kernel:   MEM window: d0200000-d02fffff
>>> Oct  3 11:21:48 XXXXX  kernel:   PREFETCH window: disabled.
>>> Oct  3 11:21:48 XXXXX  kernel: PCI: Failed to allocate mem resource #6:20000 \
>>> f0000000 for 0000:09:00.0 
>
>I ran into a similar problem, and posted a fix, see
>http://marc.theaimsgroup.com/?l=linux-pci&m=113225006603745&w=2
>
>Can you try it to see if this problem goes away?

Even though your patch touched arch/i386/pci/i386.c I tried it in my pure AMD64
environment. No luck... I remember when this PCI error turned up, but since it
was non-fatal I shrugged it off. Early 2.6.13 it was. Booting back on my
non-distro, plain kernel.org notebook I indeed see that 2.6.11.11 and 2.6.12
are fine.

Here the effected device is the builtin nvidia GPU (GeForce FX Go5700 64MB):

[...]
Boot video device is 0000:01:00.0
ACPI: PCI Interrupt Routing Table [\_SB_.PCI0._PRT]
ACPI: PCI Interrupt Link [ALKA] (IRQs 16 17 18 19 20 21 22 23) *10, disabled.
ACPI: PCI Interrupt Link [ALKB] (IRQs 16 17 18 19 20 21 22 23) *10, disabled.
ACPI: PCI Interrupt Link [ALKC] (IRQs 22) *11, disabled.
ACPI: PCI Interrupt Link [ALKD] (IRQs 21) *11, disabled.
ACPI: PCI Interrupt Link [LNKA] (IRQs 3 4 5 7 9 12 14 15) *10
ACPI: PCI Interrupt Link [LNKB] (IRQs 3 4 5 7 *10 12 14 15)
ACPI: PCI Interrupt Link [LNKC] (IRQs 3 4 5 7 *11 12 14 15)
ACPI: PCI Interrupt Link [LNKD] (IRQs 3 4 5 7 9 10 *11 12 14 15)
ACPI: Embedded Controller [EC] (gpe 11)
Linux Plug and Play Support v0.97 (c) Adam Belay
pnp: PnP ACPI init
pnp: PnP ACPI: found 10 devices
SCSI subsystem initialized
PCI: Using ACPI for IRQ routing
PCI: If a device doesn't work, try "pci=routeirq".  If it helps, post a report
pnp: 00:05: ioport range 0x600-0x60f has been reserved
pnp: 00:05: ioport range 0x1c0-0x1cf has been reserved
pnp: 00:05: ioport range 0x4d0-0x4d1 has been reserved
pnp: 00:05: ioport range 0xfe10-0xfe11 could not be reserved
PCI: Failed to allocate mem resource #6:20000@f0000000 for 0000:01:00.0
PCI: Bridge: 0000:00:01.0
  IO window: disabled.
  MEM window: c1000000-c1ffffff
  PREFETCH window: e0000000-efffffff
PCI: Bus 2, cardbus bridge: 0000:00:0b.0
  IO window: 00002000-000020ff
  IO window: 00002400-000024ff
  PREFETCH window: 88000000-89ffffff
  MEM window: 8a000000-8bffffff
PCI: Bus 6, cardbus bridge: 0000:00:0b.1
  IO window: 00002800-000028ff
  IO window: 00002c00-00002cff
  PREFETCH window: 8c000000-8dffffff
  MEM window: 8e000000-8fffffff
PCI: Setting latency timer of device 0000:00:01.0 to 64
[... asf...]

Mvh
Mats Johannesson
--

Re: PCI error on x86_64 2.6.13 kernel

[Greg KH]

On Fri, Nov 18, 2005 at 08:04:40AM +0100, Voluspa wrote:
> 
> On 2005-11-18 0:24:51 Rajesh Shah wrote:
> >On Wed, Oct 26, 2005 at 11:06:05PM +0200, Niklas Kallman wrote:
> >>Jack Howarth wrote:
> >>> Has anyone reported the following? For both of the 2.6.13 based
> >>> kernels released so far on Fedora Core 4 for x86_64, we are seeing
> >>> error messages of the form...
> >>> 
> >>> Oct  3 11:21:48 XXXXX  kernel:   MEM window: d0200000-d02fffff
> >>> Oct  3 11:21:48 XXXXX  kernel:   PREFETCH window: disabled.
> >>> Oct  3 11:21:48 XXXXX  kernel: PCI: Failed to allocate mem resource #6:20000 \
> >>> f0000000 for 0000:09:00.0 
> >
> >I ran into a similar problem, and posted a fix, see
> >http://marc.theaimsgroup.com/?l=linux-pci&m=113225006603745&w=2
> >
> >Can you try it to see if this problem goes away?
> 
> Even though your patch touched arch/i386/pci/i386.c I tried it in my pure AMD64
> environment. No luck... I remember when this PCI error turned up, but since it
> was non-fatal I shrugged it off. Early 2.6.13 it was. Booting back on my
> non-distro, plain kernel.org notebook I indeed see that 2.6.11.11 and 2.6.12
> are fine.

Others are reporting this problem too.

Is there any way you could be able to run 'git bisect' between 2.6.12
and 2.6.13 to try to find the offending changeset?  I would really
appreciate it.

thanks,

greg k-h

Re: PCI error on x86_64 2.6.13 kernel

[Voluspa]

On Fri, 18 Nov 2005 17:18:40 -0800 Greg KH wrote:
> On Fri, Nov 18, 2005 at 08:04:40AM +0100, Voluspa wrote:
> > 
> > On 2005-11-18 0:24:51 Rajesh Shah wrote:
> > >On Wed, Oct 26, 2005 at 11:06:05PM +0200, Niklas Kallman wrote:
> > >>Jack Howarth wrote:
> > >>> Has anyone reported the following? For both of the 2.6.13 based
> > >>> kernels released so far on Fedora Core 4 for x86_64, we are seeing
> > >>> error messages of the form...
> > >>> 
> > >>> Oct  3 11:21:48 XXXXX  kernel:   MEM window: d0200000-d02fffff
> > >>> Oct  3 11:21:48 XXXXX  kernel:   PREFETCH window: disabled.
> > >>> Oct  3 11:21:48 XXXXX  kernel: PCI: Failed to allocate mem resource #6:20000 \
> > >>> f0000000 for 0000:09:00.0 
> > >
> > >I ran into a similar problem, and posted a fix, see
> > >http://marc.theaimsgroup.com/?l=linux-pci&m=113225006603745&w=2
> > >
> > >Can you try it to see if this problem goes away?
> > 
> > Even though your patch touched arch/i386/pci/i386.c I tried it in my pure AMD64
> > environment. No luck... I remember when this PCI error turned up, but since it
> > was non-fatal I shrugged it off. Early 2.6.13 it was. Booting back on my
> > non-distro, plain kernel.org notebook I indeed see that 2.6.11.11 and 2.6.12
> > are fine.
> 
> Others are reporting this problem too.
> 
> Is there any way you could be able to run 'git bisect' between 2.6.12
> and 2.6.13 to try to find the offending changeset?  I would really
> appreciate it.

I have to do all the groundwork first (getting git, some dataset, read an howto),
but I'll try to return with some answers by weekend's end or Monday morning. If I fail,
the pointer will be the oldfashioned -rcX is Ok but -rcY not.

Mvh
Mats Johannesson
--

Re: PCI error on x86_64 2.6.13 kernel

[Voluspa]

On Fri, 18 Nov 2005 17:18:40 -0800 Greg KH wrote:
> Is there any way you could be able to run 'git bisect' between 2.6.12
> and 2.6.13 to try to find the offending changeset?  I would really
> appreciate it.

I'll take one beer (or in .ru: \u043f\u0438\u0432\u043e) for each hour ;-) This is a six-pack:

299de0343c7d18448a69c635378342e9214b14af is first bad commit
diff-tree 299de0343c7d18448a69c635378342e9214b14af (from 90b54929b626c80056262d9d99b3f48522e404d0)

Marvellous tool, that git. But it took less time - for us users - when we only had to
wave in a general -rc direction...

I've not tried to back the commit out from a 2.6.14 kernel, though. Will wait for a real fix.

Mvh
Mats Johannesson
--

PCI error on x86_64 2.6.13 kernel

[Jack Howarth]

   Has anyone reported the following? For both of the 2.6.13 based
kernels released so far on Fedora Core 4 for x86_64, we are seeing
error messages of the form...

Oct  3 11:21:48 XXXXX  kernel:   MEM window: d0200000-d02fffff
Oct  3 11:21:48 XXXXX  kernel:   PREFETCH window: disabled.
Oct  3 11:21:48 XXXXX  kernel: PCI: Failed to allocate mem resource #6:20000 f0000000 for 0000:09:00.0

...on a Sun W2100Z dual opteron. These memory allocations errors 
don't occur under the 2.6.12 based kernels that Fedora has released 
for FC4. So far the errors don't seem to be manifesting themselves 
as any noticable system errors in using the machine itself.
                    Jack