fc7t4 - problems with ATA66 module

[Joshua Andrews]

My system has an onboard Highpoint ATA66 controller that has worked fine 
with all previous releases of Fedora as well as redhat 6.0 - 9. Now with 
the switch to serial drivers it no longer functions.

Below is a section of my kernel log that does not mean much to me except 
that it shows that it tried and failed to make it work.

Apr 28 18:20:05 et kernel: ACPI: PCI Interrupt 0000:00:13.0[A] -> GSI 18 
(level, low) -> IRQ 19
Apr 28 18:20:05 et kernel: pata_hpt366 0000:00:13.0: failed to iomap PCI 
BAR 2
Apr 28 18:20:05 et kernel: ACPI: PCI interrupt for device 0000:00:13.0 
disabled
Apr 28 18:20:05 et kernel: pata_hpt366: probe of 0000:00:13.0 failed 
with error -12
Apr 28 18:20:05 et kernel: ACPI: PCI Interrupt 0000:00:13.1[B] -> GSI 18 
(level, low) -> IRQ 19
Apr 28 18:20:05 et kernel: pata_hpt366 0000:00:13.1: failed to iomap PCI 
BAR 2
Apr 28 18:20:05 et kernel: ACPI: PCI interrupt for device 0000:00:13.1 
disabled
Apr 28 18:20:05 et kernel: pata_hpt366: probe of 0000:00:13.1 failed 
with error -12

-J

[Alan Cox]

On Sat, Apr 28, 2007 at 06:48:03PM -0700, Joshua Andrews wrote:
>  Apr 28 18:20:05 et kernel: ACPI: PCI Interrupt 0000:00:13.0[A] ->
> GSI 18 
>  (level, low) -> IRQ 19
>  Apr 28 18:20:05 et kernel: pata_hpt366 0000:00:13.0: failed to iomap PCI 
>  BAR 2 
Is this the only PCI BAR it complains about ?

Can you open a bug, cc me it and attach a dmesg and lspci -vvxxx from a working
FC (eg FC6). I have some ideas I'd like to check

[Joshua Andrews]

Alan Cox wrote:
>  On Sat, Apr 28, 2007 at 06:48:03PM -0700, Joshua Andrews wrote:
>    
> > Apr 28 18:20:05 et kernel: ACPI: PCI Interrupt 0000:00:13.0[A] -> GSI 18 
> > (level, low) -> IRQ 19
> > Apr 28 18:20:05 et kernel: pata_hpt366 0000:00:13.0: failed to iomap PCI 
> > BAR 2
> >     
>
>  Is this the only PCI BAR it complains about ?
>
>  Can you open a bug, cc me it and attach a dmesg and lspci -vvxxx from a working
>  FC (eg FC6). I have some ideas I'd like to check 
>
>     
I installed Dave Jones 2.6.21-1.3125.fc7 kernel and the problem still 
exists with "kernel: pata_hpt366 0000:00:13.0: failed to iomap PCI BAR 
2". There are no other complaints.

-J

[Dave Jones]

On Wed, May 02, 2007 at 10:08:28AM -0700, Joshua Andrews wrote:
>  Alan Cox wrote:
>  > On Sat, Apr 28, 2007 at 06:48:03PM -0700, Joshua Andrews wrote:
>  >   
>  >> Apr 28 18:20:05 et kernel: ACPI: PCI Interrupt 0000:00:13.0[A] -> GSI 18 
>  >> (level, low) -> IRQ 19
>  >> Apr 28 18:20:05 et kernel: pata_hpt366 0000:00:13.0: failed to iomap PCI 
>  >> BAR 2
>  >>     
>  >
>  > Is this the only PCI BAR it complains about ?
>  >
>  > Can you open a bug, cc me it and attach a dmesg and lspci -vvxxx from a working
>  > FC (eg FC6). I have some ideas I'd like to check 
>  >
>  >   
>  I installed Dave Jones 2.6.21-1.3125.fc7 kernel and the problem still 
>  exists with "kernel: pata_hpt366 0000:00:13.0: failed to iomap PCI BAR 
>  2". There are no other complaints. 
As soon as the current build system outage has passed, I'll be pushing
through a build with some patches Alan sent me, which should fix this.

	Dave

-- 
http://www.codemonkey.org.uk

[Alan Cox]

On Wed, May 02, 2007 at 10:08:28AM -0700, Joshua Andrews wrote:
>  I installed Dave Jones 2.6.21-1.3125.fc7 kernel and the problem still
>
>  exists with "kernel: pata_hpt366 0000:00:13.0: failed to iomap PCI BAR 
>  2". There are no other complaints. 
Fix sent upstream

[Joshua Andrews]

Alan Cox wrote:
>  On Wed, May 02, 2007 at 10:08:28AM -0700, Joshua Andrews wrote:
>    
> > I installed Dave Jones 2.6.21-1.3125.fc7 kernel and the problem still 
> > exists with "kernel: pata_hpt366 0000:00:13.0: failed to iomap PCI BAR 
> > 2". There are no other complaints.
> >     
>
>  Fix sent upstream
>
>     
After installing kernel 2.6.21-1.3149.fc7 the controller still does not 
work.
The error message has changed.

ACPI: PCI Interrupt 0000:00:13.0[A] -> GSI 18 (level, low) -> IRQ 19
ata3: PATA max UDMA/66 cmd 0x0001b400 ctl 0x0001b802 bmdma 0x0001bc00 irq 19
scsi2 : pata_hpt366
ata3: port disabled. ignoring.
piix4_smbus 0000:00:07.3: Found 0000:00:07.3 device
ACPI: PCI Interrupt 0000:00:13.1[B] -> GSI 18 (level, low) -> IRQ 19
ata4: PATA max UDMA/66 cmd 0x0001c000 ctl 0x0001c402 bmdma 0x0001c800 irq 19
scsi3 : pata_hpt366
ata4: port disabled. ignoring.


-J

Patch for HPT366 (Re: fc7t4 - problems with ATA66 module)

[Alan Cox]

On Sat, May 12, 2007 at 03:58:17PM -0700, Joshua Andrews wrote:
>  After installing kernel 2.6.21-1.3149.fc7 the controller still does
> not 
>  work.
>  The error message has changed.
>  
>  ACPI: PCI Interrupt 0000:00:13.0[A] -> GSI 18 (level, low) -> IRQ 19
>  ata3: PATA max UDMA/66 cmd 0x0001b400 ctl 0x0001b802 bmdma 0x0001bc00 irq 19
>  scsi2 : pata_hpt366
>  ata3: port disabled. ignoring.
>  piix4_smbus 0000:00:07.3: Found 0000:00:07.3 device
>  ACPI: PCI Interrupt 0000:00:13.1[B] -> GSI 18 (level, low) -> IRQ 19
>  ata4: PATA max UDMA/66 cmd 0x0001c000 ctl 0x0001c402 bmdma 0x0001c800 irq 19
>  scsi3 : pata_hpt366
>  ata4: port disabled. ignoring. 
And the PCI dump agrees with this - your HPT366 controller is disabled. The
following might help but from all the code/docs I have something has disabled
that port. 

There is a matching upstream report of this (but with the old IDE driver which
was corrected to test these bits) so it doesn't appear unique but a problem
with a few boxes.

Alan



--- drivers/ata/pata_hpt366.c~	2007-05-13 00:14:51.971506056 +0100
+++ drivers/ata/pata_hpt366.c	2007-05-13 00:14:51.972505904 +0100
@@ -227,10 +227,10 @@
 		{ 0x54, 1, 0x04, 0x04 }
 	};
 	struct pci_dev *pdev = to_pci_dev(ap->host->dev);
-
+#if 0
 	if (!pci_test_config_bits(pdev, &hpt36x_enable_bits[ap->port_no]))
 		return -ENOENT;
-
+#endif
 	return ata_std_prereset(ap, deadline);
 }

Patch for HPT366 (Re: fc7t4 - problems with ATA66 module)

[Joshua Andrews]

Alan Cox wrote:
>  And the PCI dump agrees with this - your HPT366 controller is
> disabled. The
>  following might help but from all the code/docs I have something has disabled
>  that port. 
>
>  There is a matching upstream report of this (but with the old IDE driver which
>  was corrected to test these bits) so it doesn't appear unique but a problem
>  with a few boxes.
>
>  Alan
>
>  --- drivers/ata/pata_hpt366.c~	2007-05-13 00:14:51.971506056 +0100
>  +++ drivers/ata/pata_hpt366.c	2007-05-13 00:14:51.972505904 +0100
>  @@ -227,10 +227,10 @@
>   		{ 0x54, 1, 0x04, 0x04 }
>   	};
>   	struct pci_dev *pdev = to_pci_dev(ap->host->dev);
>  -
>  +#if 0
>   	if (!pci_test_config_bits(pdev, &hpt36x_enable_bits[ap->port_no]))
>   		return -ENOENT;
>  -
>  +#endif
>   	return ata_std_prereset(ap, deadline);
>   }
>     
No luck with the patch, it failed with "Hunk #1 FAILED at 227".

I ran it on "prepped" source from 2.6.21-1.3149.fc7.src.

Maybe it's a different pata_hpt366.c or maybe I don't know what I'm doing.

-J