ata_piix: port disabled. ignoring.

ata_piix: port disabled. ignoring.

[Geert Uytterhoeven]

	Hi Jeff,

I'm trying to install Linux on a SATA disk in a Dell PowerEdge 750, which has
an Intel 82875P chipset with an Intel 6300ESB SATA Storage Controller.

I managed to do the actual installation by booting from a Knoppix 3.4 CD (using
kernel 2.4.26 or 2.6.5 (apparently SATA drivers were not included), doesn't
matter much) in an external USB CD-ROM driver, and installing Debian testing
using FAI (Fully Automatic Installation). It took a while because the SATA
drive was used in backwards compatiblity mode without DMA, but it succeeded.

The kernel I installed was kernel-image-2.6.5-1-686-smp from Debian, which has
SATA support included. But when I want to boot it, I get (manual copy):

| ata1: PATA max UDMA/33 cmd 0x1F0 ctl 0x3F6 bmdma 0xFEA0 irq 14
| ata1: port disabled. ignoring.

I don't care, no PATA devices attached.

| ata2: SATA max UDMA/133 cmd 0x170 ctl 0x376 bmdma 0xFEA8 irq 15
| ata2: port disabled. ignoring.

Bad, the Maxtor 6Y080M0 I want to boot from is there :-(

And after that it tries the normal IDE driver, which of course fails with:

| ide0: I/O resource 0x1F0-0x1F7 not free

I looked at ata_piix.c, and apparently the driver decides whether a port is
disabled by checking a bit in PCI config space, so this looks like a BIOS setup
problem to me. But the BIOS has the first SATA port enabled (`AUTO', and it
does see a 80 GB disk there), while the PATA and second SATA ports are marked
`OFF'.

Just to be sure, I also tried acpi=off and pci=noacpi, but no avail.

Do you have a clue? Would it help to try 2.6.6?

Thanks for all suggestions!

Gr{oetje,eeting}s,

						Geert

--
Geert Uytterhoeven -- Sony Network and Software Technology Center Europe (NSCE)
Geert.Uytterhoeven@sonycom.com ------- The Corporate Village, Da Vincilaan 7-D1
Voice +32-2-7008453 Fax +32-2-7008622 ---------------- B-1935 Zaventem, Belgium

Re: ata_piix: port disabled. ignoring.

[Matt Domsch]

On Fri, May 14, 2004 at 03:11:00PM +0200, Geert Uytterhoeven wrote:
> I'm trying to install Linux on a SATA disk in a Dell PowerEdge 750, which has
> an Intel 82875P chipset with an Intel 6300ESB SATA Storage Controller.

[snip]
 
> I looked at ata_piix.c, and apparently the driver decides whether a port is
> disabled by checking a bit in PCI config space, so this looks like a BIOS setup
> problem to me. But the BIOS has the first SATA port enabled (`AUTO', and it
> does see a 80 GB disk there), while the PATA and second SATA ports are marked
> `OFF'.

Right.  At present you need to enable the second SATA  port in the
BIOS.  Disabling it also disables the first port from being
recognized by the driver.  It's a figment of the "compatability mode"
that the PE750 runs in.  I believe Stuart Hayes sent Jeff a patch to
address this, but I can't find it handy...

-- 
Matt Domsch
Sr. Software Engineer, Lead Engineer
Dell Linux Solutions linux.dell.com & www.dell.com/linux
Linux on Dell mailing lists @ http://lists.us.dell.com

Re: ata_piix: port disabled. ignoring.

[Geert Uytterhoeven]

On Fri, 14 May 2004, Matt Domsch wrote:
> On Fri, May 14, 2004 at 03:11:00PM +0200, Geert Uytterhoeven wrote:
> > I'm trying to install Linux on a SATA disk in a Dell PowerEdge 750, which has
> > an Intel 82875P chipset with an Intel 6300ESB SATA Storage Controller.
>
> [snip]
>
> > I looked at ata_piix.c, and apparently the driver decides whether a port is
> > disabled by checking a bit in PCI config space, so this looks like a BIOS setup
> > problem to me. But the BIOS has the first SATA port enabled (`AUTO', and it
> > does see a 80 GB disk there), while the PATA and second SATA ports are marked
> > `OFF'.
>
> Right.  At present you need to enable the second SATA  port in the
> BIOS.  Disabling it also disables the first port from being
> recognized by the driver.  It's a figment of the "compatability mode"
> that the PE750 runs in.  I believe Stuart Hayes sent Jeff a patch to
> address this, but I can't find it handy...

Thanks for your suggestion!

Unfortunately enabling the second SATA port doesn't make a difference. I also
tried enabling the PATA port as well, but no change :-(

Gr{oetje,eeting}s,

						Geert

--
Geert Uytterhoeven -- Sony Network and Software Technology Center Europe (NSCE)
Geert.Uytterhoeven@sonycom.com ------- The Corporate Village, Da Vincilaan 7-D1
Voice +32-2-7008453 Fax +32-2-7008622 ---------------- B-1935 Zaventem, Belgium

Re: ata_piix: port disabled. ignoring.

[Jeff Garzik]

Matt Domsch wrote:
> On Fri, May 14, 2004 at 03:11:00PM +0200, Geert Uytterhoeven wrote:
> 
>>I'm trying to install Linux on a SATA disk in a Dell PowerEdge 750, which has
>>an Intel 82875P chipset with an Intel 6300ESB SATA Storage Controller.
> 
> 
> [snip]
>  
> 
>>I looked at ata_piix.c, and apparently the driver decides whether a port is
>>disabled by checking a bit in PCI config space, so this looks like a BIOS setup
>>problem to me. But the BIOS has the first SATA port enabled (`AUTO', and it
>>does see a 80 GB disk there), while the PATA and second SATA ports are marked
>>`OFF'.
> 
> 
> Right.  At present you need to enable the second SATA  port in the
> BIOS.  Disabling it also disables the first port from being
> recognized by the driver.  It's a figment of the "compatability mode"
> that the PE750 runs in.  I believe Stuart Hayes sent Jeff a patch to
> address this, but I can't find it handy...


That _should_ have been fixed in the last update to ata_piix.  If you 
could get Stuart to re-test and re-submit that patch if necessary, that 
would be useful.

Geert, two things to try:
1) Try the latest kernel
2) Fiddle with BIOS settings until your PATA and SATA devices appear as 
_separate_ devices on the PCI bus.

Re: ata_piix: port disabled. ignoring.

[Geert Uytterhoeven]

On Fri, 14 May 2004, Jeff Garzik wrote:
> Matt Domsch wrote:
> > On Fri, May 14, 2004 at 03:11:00PM +0200, Geert Uytterhoeven wrote:
> >
> >>I'm trying to install Linux on a SATA disk in a Dell PowerEdge 750, which has
> >>an Intel 82875P chipset with an Intel 6300ESB SATA Storage Controller.
> >
> >
> > [snip]
> >
> >
> >>I looked at ata_piix.c, and apparently the driver decides whether a port is
> >>disabled by checking a bit in PCI config space, so this looks like a BIOS setup
> >>problem to me. But the BIOS has the first SATA port enabled (`AUTO', and it
> >>does see a 80 GB disk there), while the PATA and second SATA ports are marked
> >>`OFF'.
> >
> >
> > Right.  At present you need to enable the second SATA  port in the
> > BIOS.  Disabling it also disables the first port from being
> > recognized by the driver.  It's a figment of the "compatability mode"
> > that the PE750 runs in.  I believe Stuart Hayes sent Jeff a patch to
> > address this, but I can't find it handy...
>
>
> That _should_ have been fixed in the last update to ata_piix.  If you
> could get Stuart to re-test and re-submit that patch if necessary, that
> would be useful.
>
> Geert, two things to try:
> 1) Try the latest kernel

Thanks! I'll try that one...

> 2) Fiddle with BIOS settings until your PATA and SATA devices appear as
> _separate_ devices on the PCI bus.

There's no option in the BIOS for that (at least I didn't find it), only to
disable IDE completely, which upon selection makes the Intel 6300ESB SATA
Storage Controller disappear from the lspci listing.

Gr{oetje,eeting}s,

						Geert

--
Geert Uytterhoeven -- Sony Network and Software Technology Center Europe (NSCE)
Geert.Uytterhoeven@sonycom.com ------- The Corporate Village, Da Vincilaan 7-D1
Voice +32-2-7008453 Fax +32-2-7008622 ---------------- B-1935 Zaventem, Belgium

Re: ata_piix: port disabled. ignoring.

[Geert Uytterhoeven]

On Mon, 17 May 2004, Geert Uytterhoeven wrote:
> On Fri, 14 May 2004, Jeff Garzik wrote:
> > Matt Domsch wrote:
> > > On Fri, May 14, 2004 at 03:11:00PM +0200, Geert Uytterhoeven wrote:
> > >
> > >>I'm trying to install Linux on a SATA disk in a Dell PowerEdge 750, which has
> > >>an Intel 82875P chipset with an Intel 6300ESB SATA Storage Controller.
> > >
> > >
> > > [snip]
> > >
> > >
> > >>I looked at ata_piix.c, and apparently the driver decides whether a port is
> > >>disabled by checking a bit in PCI config space, so this looks like a BIOS setup
> > >>problem to me. But the BIOS has the first SATA port enabled (`AUTO', and it
> > >>does see a 80 GB disk there), while the PATA and second SATA ports are marked
> > >>`OFF'.
> > >
> > >
> > > Right.  At present you need to enable the second SATA  port in the
> > > BIOS.  Disabling it also disables the first port from being
> > > recognized by the driver.  It's a figment of the "compatability mode"
> > > that the PE750 runs in.  I believe Stuart Hayes sent Jeff a patch to
> > > address this, but I can't find it handy...
> >
> >
> > That _should_ have been fixed in the last update to ata_piix.  If you
> > could get Stuart to re-test and re-submit that patch if necessary, that
> > would be useful.
> >
> > Geert, two things to try:
> > 1) Try the latest kernel
>
> Thanks! I'll try that one...

And so I did. But 2.6.6-bk4 still gives:

| Uniform Multi-Platform E-IDE driver Revision: 7.00alpha2
| ide: Assuming 33MHz system bus speed for PIO modes; override with idebus=xx
| ide1: I/O resource 0x170-0x177 not free.
| ide1: ports already in use, skipping probe
| ata_piix: combined mode detected
| ata1: PATA max UDMA/33 cmd 0x1F0 ctl 0x3F6 bmdma 0xFEA0 irq 14
| ata1: port disabled. ignoring.
| scsi0 : ata_piix
| ata2: SATA max UDMA/133 cmd 0x170 ctl 0x376 bmdma 0xFEA8 irq 15
| ata2: port disabled. ignoring.
| scsi1 : ata_piix

If I disable CONFIG_SCSI_SATA, IDE works, but very slow (no DMA).

If I try to work around the problem by applying this patch:

--- linux-2.6.6-bk4/drivers/scsi/ata_piix.c.orig	2004-05-17 20:02:25.000000000 +0200
+++ linux-2.6.6-bk4/drivers/scsi/ata_piix.c	2004-05-17 20:59:15.000000000 +0200
@@ -330,8 +330,8 @@
 	if (!pci_test_config_bits(ap->host_set->pdev,
 				  &piix_enable_bits[ap->port_no])) {
 		ata_port_disable(ap);
-		printk(KERN_INFO "ata%u: port disabled. ignoring.\n", ap->id);
-		return;
+		//printk(KERN_INFO "ata%u: port disabled. ignoring.\n", ap->id);
+		printk(KERN_INFO "ata%u: port disabled. NOT IGNORING!\n", ap->id);
 	}

 	if (!piix_sata_probe(ap)) {

everything works fine, even if I disable the second SATA port in the BIOS:

| Uniform Multi-Platform E-IDE driver Revision: 7.00alpha2
| ide: Assuming 33MHz system bus speed for PIO modes; override with idebus=xx
| ide1: I/O resource 0x170-0x177 not free.
| ide1: ports already in use, skipping probe
| ata_piix: combined mode detected
| ata1: PATA max UDMA/33 cmd 0x1F0 ctl 0x3F6 bmdma 0xFEA0 irq 14
| ata1: port disabled. ignoring.
| scsi0 : ata_piix
| ata2: SATA max UDMA/133 cmd 0x170 ctl 0x376 bmdma 0xFEA8 irq 15
| ata2: port disabled. NOT IGNORING!
| ata2: dev 0 ATA, max UDMA/133, 156250000 sectors:
| ata2: dev 0 configured for UDMA/133
| scsi1 : ata_piix
|   Vendor: ATA       Model: Maxtor 6Y080M0    Rev: 1.02
|   Type:   Direct-Access                      ANSI SCSI revision: 05
| SCSI device sda: 156250000 512-byte hdwr sectors (80000 MB)
| SCSI device sda: drive cache: write through
|  sda: sda1 sda2 sda3 < sda5 sda6 sda7 sda8 sda9 sda10 >
| Attached scsi disk sda at scsi1, channel 0, id 0, lun 0

... and I'm a happy man (for now ;-)

Gr{oetje,eeting}s,

						Geert

--
Geert Uytterhoeven -- Sony Network and Software Technology Center Europe (NSCE)
Geert.Uytterhoeven@sonycom.com ------- The Corporate Village, Da Vincilaan 7-D1
Voice +32-2-7008453 Fax +32-2-7008622 ---------------- B-1935 Zaventem, Belgium

Re: ata_piix: port disabled. ignoring.

[Bartlomiej Zolnierkiewicz]

On Monday 17 of May 2004 16:19, Geert Uytterhoeven wrote:

> If I disable CONFIG_SCSI_SATA, IDE works, but very slow (no DMA).

due to CONFIG_BLK_DEV_PIIX=n but ata_piix.c is prefferred for SATA

Re: ata_piix: port disabled. ignoring.

[Geert Uytterhoeven]

On Tue, 18 May 2004, Bartlomiej Zolnierkiewicz wrote:
> On Monday 17 of May 2004 16:19, Geert Uytterhoeven wrote:
> > If I disable CONFIG_SCSI_SATA, IDE works, but very slow (no DMA).
>
> due to CONFIG_BLK_DEV_PIIX=n but ata_piix.c is prefferred for SATA

No, I had

    CONFIG_BLK_DEV_PIIX=y
    CONFIG_BLK_DEV_IDEDMA_PCI=y

but I get an error when trying to enable DMA using hdparm.

Gr{oetje,eeting}s,

						Geert

--
Geert Uytterhoeven -- Sony Network and Software Technology Center Europe (NSCE)
Geert.Uytterhoeven@sonycom.com ------- The Corporate Village, Da Vincilaan 7-D1
Voice +32-2-7008453 Fax +32-2-7008622 ---------------- B-1935 Zaventem, Belgium

Re: ata_piix: port disabled. ignoring.

[Bartlomiej Zolnierkiewicz]

On Wednesday 19 of May 2004 09:48, Geert Uytterhoeven wrote:
> On Tue, 18 May 2004, Bartlomiej Zolnierkiewicz wrote:
> > On Monday 17 of May 2004 16:19, Geert Uytterhoeven wrote:
> > > If I disable CONFIG_SCSI_SATA, IDE works, but very slow (no DMA).
> >
> > due to CONFIG_BLK_DEV_PIIX=n but ata_piix.c is prefferred for SATA
>
> No, I had
>
>     CONFIG_BLK_DEV_PIIX=y
>     CONFIG_BLK_DEV_IDEDMA_PCI=y
>
> but I get an error when trying to enable DMA using hdparm.

I suppose that you're using chipset which ID is not in piix.c.
If not than this is a bug and I would like to know more about it.

> Gr{oetje,eeting}s,
>
> 						Geert
>
> --
> Geert Uytterhoeven -- Sony Network and Software Technology Center Europe
> (NSCE) Geert.Uytterhoeven@sonycom.com ------- The Corporate Village, Da
> Vincilaan 7-D1 Voice +32-2-7008453 Fax +32-2-7008622 ----------------
> B-1935 Zaventem, Belgium

Re: ata_piix: port disabled. ignoring.

[Geert Uytterhoeven]

On Wed, 19 May 2004, Bartlomiej Zolnierkiewicz wrote:
> On Wednesday 19 of May 2004 09:48, Geert Uytterhoeven wrote:
> > On Tue, 18 May 2004, Bartlomiej Zolnierkiewicz wrote:
> > > On Monday 17 of May 2004 16:19, Geert Uytterhoeven wrote:
> > > > If I disable CONFIG_SCSI_SATA, IDE works, but very slow (no DMA).
> > >
> > > due to CONFIG_BLK_DEV_PIIX=n but ata_piix.c is prefferred for SATA
> >
> > No, I had
> >
> >     CONFIG_BLK_DEV_PIIX=y
> >     CONFIG_BLK_DEV_IDEDMA_PCI=y
> >
> > but I get an error when trying to enable DMA using hdparm.
>
> I suppose that you're using chipset which ID is not in piix.c.
> If not than this is a bug and I would like to know more about it.

Upon closer look, piix.c doesn't recognize the ESB_3 yet. If I add it to
piix.[ch] and treat it the same as the ICH5-SATA, the SATA disk gets 57 MiB/s
using the PATA piix driver.

Gr{oetje,eeting}s,

						Geert

--
Geert Uytterhoeven -- Sony Network and Software Technology Center Europe (NSCE)
Geert.Uytterhoeven@sonycom.com ------- The Corporate Village, Da Vincilaan 7-D1
Voice +32-2-7008453 Fax +32-2-7008622 ---------------- B-1935 Zaventem, Belgium

Re: ata_piix: port disabled. ignoring.

[Bartlomiej Zolnierkiewicz]

On Mon, 24 May 2004, Geert Uytterhoeven wrote:

> On Wed, 19 May 2004, Bartlomiej Zolnierkiewicz wrote:
> > On Wednesday 19 of May 2004 09:48, Geert Uytterhoeven wrote:
> > > On Tue, 18 May 2004, Bartlomiej Zolnierkiewicz wrote:
> > > > On Monday 17 of May 2004 16:19, Geert Uytterhoeven wrote:
> > > > > If I disable CONFIG_SCSI_SATA, IDE works, but very slow (no DMA).
> > > >
> > > > due to CONFIG_BLK_DEV_PIIX=n but ata_piix.c is prefferred for SATA
> > >
> > > No, I had
> > >
> > >     CONFIG_BLK_DEV_PIIX=y
> > >     CONFIG_BLK_DEV_IDEDMA_PCI=y
> > >
> > > but I get an error when trying to enable DMA using hdparm.
> >
> > I suppose that you're using chipset which ID is not in piix.c.
> > If not than this is a bug and I would like to know more about it.
>
> Upon closer look, piix.c doesn't recognize the ESB_3 yet. If I add it to

This is intended, new SATA chipsets belong to ata_piix.c.

> piix.[ch] and treat it the same as the ICH5-SATA, the SATA disk gets 57 MiB/s
> using the PATA piix driver.

OK, thanks.

Cheers,
Bartlomiej

Re: ata_piix: port disabled. ignoring.

[Jeff Garzik]

Geert Uytterhoeven wrote:
> If I try to work around the problem by applying this patch:
> 
> --- linux-2.6.6-bk4/drivers/scsi/ata_piix.c.orig	2004-05-17 20:02:25.000000000 +0200
> +++ linux-2.6.6-bk4/drivers/scsi/ata_piix.c	2004-05-17 20:59:15.000000000 +0200
> @@ -330,8 +330,8 @@
>  	if (!pci_test_config_bits(ap->host_set->pdev,
>  				  &piix_enable_bits[ap->port_no])) {
>  		ata_port_disable(ap);
> -		printk(KERN_INFO "ata%u: port disabled. ignoring.\n", ap->id);
> -		return;
> +		//printk(KERN_INFO "ata%u: port disabled. ignoring.\n", ap->id);
> +		printk(KERN_INFO "ata%u: port disabled. NOT IGNORING!\n", ap->id);
>  	}
> 
>  	if (!piix_sata_probe(ap)) {
> 
> everything works fine, even if I disable the second SATA port in the BIOS:


I think this check is vaguely incorrect, because it sounds like you are 
in combined mode.  That would imply that ap->port_no is incorrect, for 
this one special case.  (details: in combined aka legacy mode, port 
number is always zero because it is initialized as two separate hosts, 
not one host with two ata_ports)

However, since this is SATA, and PIIX does at least give us a "no 
device" indication, we could probably just delete the 'if' and the code 
you are commenting out as well.

Ponder, ponder...

Another thing I am pondering is detecting combined mode in 
drivers/pci/quirks.c, and reconfiguring the motherboard such that is it 
no longer in combined mode.

	Jeff