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
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
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
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
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
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); }
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
-
Alan Cox -
Dave Jones -
Joshua Andrews