drive side 80-wire cable detection failed

drive side 80-wire cable detection failed

[Tobias Hoffmann]

Hi!

On my NV3 board with a Samsung SP1634N Harddisk I wrongly got
"drive side 80-wire cable detection failed" with the current kernel.

A possible fix is attached.

Please CC any replies as I'm not on the list.

  Tobias


diff --git a/drivers/ide/ide-iops.c b/drivers/ide/ide-iops.c
index dcda0f1..e3cfd6c 100644
--- a/drivers/ide/ide-iops.c
+++ b/drivers/ide/ide-iops.c
@@ -588,6 +588,7 @@ EXPORT_SYMBOL_GPL(ide_in_drive_list);
 static const struct drive_list_entry ivb_list[] = {
        { "QUANTUM FIREBALLlct10 05"    , "A03.0900"    },
        { "TSSTcorp CDDVDW SH-S202J"    , "SB00"        },
+       { "SAMSUNG SP1634N"             , "UZ100-03"    },
        { NULL                          , NULL          }
 };

Re: drive side 80-wire cable detection failed

[Alan Cox]

On Sun, 04 Nov 2007 00:09:49 +0100
Tobias Hoffmann <lk-list@the-axe-effect.de> wrote:

> Hi!
> 
> On my NV3 board with a Samsung SP1634N Harddisk I wrongly got
> "drive side 80-wire cable detection failed" with the current kernel.

Does the drive cable detect correctly on a non Nvidia chipset ?

Re: drive side 80-wire cable detection failed

[Tobias Hoffmann]

Alan Cox wrote:
>> On my NV3 board with a Samsung SP1634N Harddisk I wrongly got
>> "drive side 80-wire cable detection failed" with the current kernel.
> 
> Does the drive cable detect correctly on a non Nvidia chipset ?
> 
I have no box here to test this - my laptop won't help.

But I did another test:

I replaced the 80-wire cable with a 40-wire cable. Then the BIOS warns 
about "no 80-wire cable connected". The kernel boot and seems to enable 
UDMA100 which leads to some BadCRC errors. After that the drive is in 
udma3-mode(hdparm) IIRC but not in udma5, as when using the right cable.

Maybe it's not that clever to add the drive to the ivb_list?

As far as I understood it, pata_acpi should fix the problem, but I'm not 
sure on the required configuration. Do I have to disable the normal 
IDE-support? How 'experimental' is the PATA support?


   Tobias

Re: drive side 80-wire cable detection failed

[Alan Cox]

> As far as I understood it, pata_acpi should fix the problem, but I'm not 
> sure on the required configuration. Do I have to disable the normal 
> IDE-support? How 'experimental' is the PATA support?

The pata_amd (AMD/Nvidia) driver is pretty robust, but I would expect the
same results from both in your case at the moment. Tejun is still fixing
ACPI bits so after that it might be worth trying.

Alan