FC9 installation sees SATA drives but not PATA

[James Wilkinson]

Stephen Soliday wrote:
> I want to install FC9 in place of my old Debian 4
> 
> I have one PATA three SATA hard drives and a SATA DVD-RW on a
> ASUS N1L64-SLI WS motherboard   with the 0505 bios release
> 
> IDE:
> /dev/hda1    ntfs      Windows vista boot
> /dev/hda2    linux-swap
> /dev/hda3    reiserfs    debian /
> 
> SATA:
> /dev/sda1  &
> /dev/sda2   configured with LVM as RAID 1 mounted as /home
> /dev/sdc1  ntfs   windows data drive
> 
> The problem is that FC9 will not see the ide drive (/dev/hda) it does not
> show up in /proc/partitions
> I tried Fedora Live 8 with the same result, the kernel sees the SATA drives
> but not the IDE
> 
> I have tried various boot parameters such as     libata.dma=0  or ide=nodma

Hi Stephen,

Did you ever fix this?

Google has only heard of the N1L64-SLI motherboard from your posts – are
you sure that you got the model name right?

It’s possible that your PATA drive is plugged into an IDE adapter
without libata support. Is the drive plugged into the motherboard, or
into a separate PCI card? Is the chipset Nvidia, or something else, like
Intel?

Normally, I’d expect a motherboard with SLI in the name to be based on
an Nvidia chipset, and use that chipset’s on-board PATA support. I can
confirm that I’ve never had problems with the on-board PATA on an Asus
nForce 4 motherboard since about FC3; I’ve never heard of particular
problems with Nvidia’s PATA support, and I doubt Nvidia have done much to
change it over the years. But it’s possible that Asus might have used a
separate PCI or PCI Express PATA adapter.

Recent Intel chipsets don’t have PATA: any motherboard support comes
from a separate PCI or PCI Express PATA adapter on the motherboard. In
that case, you might have got an unusual chip without decent support.

In any case, the output of /sbin/lspci will be instructive.

Hope this helps,

James.

-- 
E-mail:     james@ | "Security question ... What's your dog's maiden name?"
aprilcottage.co.uk |     -- Peter Gutmann on bad security designs