Patch for Promise PDC20276

Patch for Promise PDC20276

[Michel Bouissou]

Hi there,

The patch that Arne Brutschy posted here (see 
http://www.cs.helsinki.fi/linux/linux-kernel/2003-22/0386.html) almost saved 
my life as it allowed me to understand why my 2.4.22 kernel didn't activate 
my Promise PDC20276 IDE controller properly.

(For the record, my kernel was displaying the following message at boot:
PDC20276: neither IDE port enabled (BIOS)
)

But this patch was only halfway satisfactory, as it enabled only one of my 2 
controller channels.

So I did a quick-and-dirty hack that gives the following patch, that allows 
activation of both channels of my PDC20276.

This might help others encountering the same problem...

========== CUT HERE ==========
--- drivers/ide/setup-pci.c.orig	2003-08-25 13:44:41.000000000 +0200
+++ drivers/ide/setup-pci.c	2003-10-26 20:35:49.000000000 +0100
@@ -577,7 +577,7 @@
 	ata_index_t index;
 	u8 tmp = 0;
 	ide_hwif_t *hwif, *mate = NULL;
-	static int secondpdc = 0;
+/* MiB	static int secondpdc = 0; */
 
 	index.all = 0xf0f0;
 
@@ -637,10 +637,10 @@
 		 * by the bios for raid purposes. 
 		 * Skip the normal "is it enabled" test for those.
 		 */
-		if (((d->vendor == PCI_VENDOR_ID_PROMISE) &&
+		if ((d->vendor == PCI_VENDOR_ID_PROMISE) &&
 		     ((d->device == PCI_DEVICE_ID_PROMISE_20262) ||
-		      (d->device == PCI_DEVICE_ID_PROMISE_20265))) &&
-		    (secondpdc++==1) && (port==1))
+		      (d->device == PCI_DEVICE_ID_PROMISE_20265) ||
+		      (d->device == PCI_DEVICE_ID_PROMISE_20276)))
 			goto controller_ok;
 			
 		if (e->reg && (pci_read_config_byte(dev, e->reg, &tmp) ||
========== CUT HERE ==========

Regards

-- 
Michel Bouissou <michel@bouissou.net> OpenPGP ID 0xDDE8AC6E

Re: Patch for Promise PDC20276

[Michel Bouissou]

Le Lundi 27 Octobre 2003 10:49, Bartlomiej Zolnierkiewicz a écrit :
>
> This was discussed few times before.
> Just enable "Special FastTrak feature" (overriding BIOS) config option.

Uh. I may have misunderstood, but I understood that using this option would 
activate the controller's hardware RAID feature, which I don't want.

I need to use the controller as a normal, non-RAID, IDE controller (and if 
ever if was setup as hardware RAID, I'm afraid it would destroy my data..)

Cheers.

-- 
Michel Bouissou <michel@bouissou.net> OpenPGP ID 0xDDE8AC6E

Re: Patch for Promise PDC20276

[Bartlomiej Zolnierkiewicz]

This was discussed few times before.
Just enable "Special FastTrak feature" (overriding BIOS) config option.

On Monday 27 of October 2003 10:09, Michel Bouissou wrote:
> Hi there,
>
> The patch that Arne Brutschy posted here (see
> http://www.cs.helsinki.fi/linux/linux-kernel/2003-22/0386.html) almost
> saved my life as it allowed me to understand why my 2.4.22 kernel didn't
> activate my Promise PDC20276 IDE controller properly.
>
> (For the record, my kernel was displaying the following message at boot:
> PDC20276: neither IDE port enabled (BIOS)
> )
>
> But this patch was only halfway satisfactory, as it enabled only one of my
> 2 controller channels.
>
> So I did a quick-and-dirty hack that gives the following patch, that allows
> activation of both channels of my PDC20276.
>
> This might help others encountering the same problem...
>
> ========== CUT HERE ==========
> --- drivers/ide/setup-pci.c.orig	2003-08-25 13:44:41.000000000 +0200
> +++ drivers/ide/setup-pci.c	2003-10-26 20:35:49.000000000 +0100
> @@ -577,7 +577,7 @@
>  	ata_index_t index;
>  	u8 tmp = 0;
>  	ide_hwif_t *hwif, *mate = NULL;
> -	static int secondpdc = 0;
> +/* MiB	static int secondpdc = 0; */
>
>  	index.all = 0xf0f0;
>
> @@ -637,10 +637,10 @@
>  		 * by the bios for raid purposes.
>  		 * Skip the normal "is it enabled" test for those.
>  		 */
> -		if (((d->vendor == PCI_VENDOR_ID_PROMISE) &&
> +		if ((d->vendor == PCI_VENDOR_ID_PROMISE) &&
>  		     ((d->device == PCI_DEVICE_ID_PROMISE_20262) ||
> -		      (d->device == PCI_DEVICE_ID_PROMISE_20265))) &&
> -		    (secondpdc++==1) && (port==1))
> +		      (d->device == PCI_DEVICE_ID_PROMISE_20265) ||
> +		      (d->device == PCI_DEVICE_ID_PROMISE_20276)))
>  			goto controller_ok;
>
>  		if (e->reg && (pci_read_config_byte(dev, e->reg, &tmp) ||
> ========== CUT HERE ==========
>
> Regards

Re: Patch for Promise PDC20276

[Bartlomiej Zolnierkiewicz]

On Monday 27 of October 2003 10:49, Michel Bouissou wrote:
> Le Lundi 27 Octobre 2003 10:49, Bartlomiej Zolnierkiewicz a écrit :
> > This was discussed few times before.
> > Just enable "Special FastTrak feature" (overriding BIOS) config option.
>
> Uh. I may have misunderstood, but I understood that using this option would
> activate the controller's hardware RAID feature, which I don't want.

Quite opposite, it makes Linux use controller even if it was marked by BIOS
as disabled (for RAID purposes).

> I need to use the controller as a normal, non-RAID, IDE controller (and if
> ever if was setup as hardware RAID, I'm afraid it would destroy my data..)

It is software RAID (in BIOS/drivers) with propertiary format.

cheers,
--bartlomiej

Re: Patch for Promise PDC20276

[Michel Bouissou]

Le Lundi 27 Octobre 2003 11:02, Bartlomiej Zolnierkiewicz a écrit :
>
> > Uh. I may have misunderstood, but I understood that using this option
> > would activate the controller's hardware RAID feature, which I don't
> > want.
>
> Quite opposite, it makes Linux use controller even if it was marked by BIOS
> as disabled (for RAID purposes).

So the Configure.help help text associated with this option is definitely 
unclear (I would say: misleading) because it barely says (in kernel 2.4.22):

<<<<<
Special FastTrak Feature
CONFIG_PDC202XX_FORCE
  For FastTrak enable overriding BIOS.
>>>>>

And I understood this as meaning <<enable the "FastTrak" RAID feature 
regardless of what BIOS says>>

Hmmmm...

-- 
Michel Bouissou <michel@bouissou.net> OpenPGP ID 0xDDE8AC6E