[PATCH/2.6.17-rc4 10/10] bugs fix for marvell SATA on powerp c pl atform

[Sven Luther]

On Fri, May 26, 2006 at 11:47:11AM -0400, Mark Lord wrote:
> Sven Luther wrote:
> >On Fri, May 26, 2006 at 09:19:33AM -0400, Mark Lord wrote:
> >>Sven Luther wrote:
> >>>Ok. can i use this tree with a 2.6.16 base ?
> >>Not as-is.  Here (attached) is a patch for 2.6.16.17+ that updates
> >>the sata_mv driver to the latest source.  Completely untested,
> >>but it does compile.
> >>
> >>I will hopefully test it later today, but in the meanwhile, have a go at 
> >>it.
> >
> >And here is attached my dmesg output. The last bit of mv_host_intr was 
> >when i
> >tried to access the partition table of the disk with parted.
> 
> I don't see anything particularly bad in that dmesg output,
> apart from all of the debug output --> did you enable that,
> or was it "on" by default?

Ah, yes, i enabled it on Jeff's suggestion.

> It finds one SATA drive, with no *known* partition table format.

Well, it is a blank disk indeed, so this is no suprise.

> Can you access the disk?  Eg.  hexdump -C /dev/sda

Trying to partition the disk with parted yielded the last set of debug
messages, and a : 

Error: Unable to open /dev/sda - unrecognized disk label.

The same when trying to write a partition table to it, and naturally, there is
no partition afterward.

hexdump yielded only the debugmessages and nothing else.

> Meanwhile, I just booted 2.6.17-rc5-git1 (latest kernel.org) on my Mac G3
> box here, and sata_mv seems to be behaving for me (thus far).

Mmm, this is a G3, while i have a G4. The G4 does some I/O reordering, which
we don't have with a G3, so this may be the cause. 

Do you have a mac version of the board, with a forth/OF bios in it ? Or a
normal PC card, which is thus uninitialized ? 

Friendly,

Sven Luther