2003-06-02 18:38:07

by Arne Brutschy

[permalink] [raw]
Subject: [PATCH] ide driver 2.4.21-rc6

Using the Promise 20276 IDE controller without raid function
doesn't work with 2.4.21*.

If you want to use the IDE controller just as plain controller without
Promise software raid (i.e. if you prefer to trust the linux software
raid), the kernel always reports this:

PDC20276: IDE controller at PCI slot 02:02.0
PDC20276: chipset revision 1
PDC20276: not 100% native mode: will probe irqs later
PDC20276: neither IDE port enabled (BIOS)

Afterwards, the kernel disables the controller. This has been reported
by serveral other users. This small patch solves the problem.

Arne


--- linux-2.4.21-rc6/drivers/ide/setup-pci.c.orig 2003-06-01 11:38:23.000000000 +0200
+++ linux-2.4.21-rc6/drivers/ide/setup-pci.c 2003-06-01 11:40:12.000000000 +0200
@@ -640,7 +640,8 @@
*/
if (((d->vendor == PCI_VENDOR_ID_PROMISE) &&
((d->device == PCI_DEVICE_ID_PROMISE_20262) ||
- (d->device == PCI_DEVICE_ID_PROMISE_20265))) &&
+ (d->device == PCI_DEVICE_ID_PROMISE_20265) ||
+ (d->device == PCI_DEVICE_ID_PROMISE_20276))) &&
(secondpdc++==1) && (port==1))
goto controller_ok;


Subject: Re: [PATCH] ide driver 2.4.21-rc6


What about turning on "Special FastTrak Feature" instead...
--
Bartlomiej

On Mon, 2 Jun 2003, Arne Brutschy wrote:

> Using the Promise 20276 IDE controller without raid function
> doesn't work with 2.4.21*.
>
> If you want to use the IDE controller just as plain controller without
> Promise software raid (i.e. if you prefer to trust the linux software
> raid), the kernel always reports this:
>
> PDC20276: IDE controller at PCI slot 02:02.0
> PDC20276: chipset revision 1
> PDC20276: not 100% native mode: will probe irqs later
> PDC20276: neither IDE port enabled (BIOS)
>
> Afterwards, the kernel disables the controller. This has been reported
> by serveral other users. This small patch solves the problem.
>
> Arne
>
>
> --- linux-2.4.21-rc6/drivers/ide/setup-pci.c.orig 2003-06-01 11:38:23.000000000 +0200
> +++ linux-2.4.21-rc6/drivers/ide/setup-pci.c 2003-06-01 11:40:12.000000000 +0200
> @@ -640,7 +640,8 @@
> */
> if (((d->vendor == PCI_VENDOR_ID_PROMISE) &&
> ((d->device == PCI_DEVICE_ID_PROMISE_20262) ||
> - (d->device == PCI_DEVICE_ID_PROMISE_20265))) &&
> + (d->device == PCI_DEVICE_ID_PROMISE_20265) ||
> + (d->device == PCI_DEVICE_ID_PROMISE_20276))) &&
> (secondpdc++==1) && (port==1))
> goto controller_ok;

2003-06-03 13:08:40

by Arne Brutschy

[permalink] [raw]
Subject: Re[2]: [PATCH] ide driver 2.4.21-rc6

Bartlomiej Zolnierkiewicz wrote:
BZ> What about turning on "Special FastTrak Feature" instead...
Didn't work for me, don't know why.

Arne

Subject: Re: Re[2]: [PATCH] ide driver 2.4.21-rc6


It MUST work.

Clean your kernel source with 'make mrproper' (just in case) and compile
again with Special FastTrak Feature enabled (CONFIG_PDC202XX_FORCE=y).

If it really won't work contact me.
--
Bartlomiej

On Tue, 3 Jun 2003, Arne Brutschy wrote:

> Bartlomiej Zolnierkiewicz wrote:
> BZ> What about turning on "Special FastTrak Feature" instead...
> Didn't work for me, don't know why.
>
> Arne



2003-06-26 21:46:32

by Arne Brutschy

[permalink] [raw]
Subject: Re[4]: [PATCH] ide driver 2.4.21-rc6

Hi,

as I promised, I'm back at home and investigated the problem with the
CONFIG_PDC202XX_FORCE setting further.

The special cas with my setup is, that I have an ICH4 onboard, an
PDC20276 onboard and an PCI SCSI controller. The root filesystem is on
the SCSI disks. I can not boot from the SCSI disks directly, as the
BIOS enables the PDC first. So I have to boot from the IDE disks
connected to the PDC. This didn't work, so I replaced the simple
Promise RAID bios with an normal UDMA bios thats allows you to use the
controller as normal IDE controller. The result is, I can boot from
the IDE disks and I can use the IDE disks as normal disks (used it in
times the CONFIG_PDC202XX_FORCE wasn't in the linux kernel).

I do not know if the PDC 20276 is being sold with such a normal bios.
I think so, as there are serveral users reporting the same error.

The problem is, that the driver thinks the bios/drives didn't get enabled
by the bios end disables the whole controller. I patched the driver to
skip this step. I still have to use the CONFIG_PDC202XX_FORCE setting-
without it the second channel gets disabled.

This is the reason for my patch. I'm not sure, but I this it shouldn't
be a problem for the driver...

Regards,
Arne

PS: please write me if you see any problems related to that patch

Bartlomiej Zolnierkiewicz wrote:
BZ> It MUST work.

BZ> Clean your kernel source with 'make mrproper' (just in case) and compile
BZ> again with Special FastTrak Feature enabled (CONFIG_PDC202XX_FORCE=y).

BZ> If it really won't work contact me.
BZ> --
BZ> Bartlomiej

BZ> On Tue, 3 Jun 2003, Arne Brutschy wrote:

>> Bartlomiej Zolnierkiewicz wrote:
>> BZ> What about turning on "Special FastTrak Feature" instead...
>> Didn't work for me, don't know why.
>>
>> Arne