Re: 2.6.17-rc5-mm1-lockdep: a rather strange oops

[Barry K. Nathan]

On 5/31/06, Andrew Morton <akpm@xxxxxxxx> wrote:
> The original oops was a jump-to-null.  I had a few of those when getting
> the latest git-libata-all tree working, due to missing
> ata_port_operations.data_xfer vectors.  But it appears that both sata_sil.c
> and sata_promise.c do have those filled in.

Ah, but pata_pdc2027x.c doesn't. (Oh, by the way, neither does sata_sil24.c.)

I tried filling it in, with the following patch, but booting that gave
me lots of weird output before the kernel finally failed to boot from
the root device. "Lots" meaning, enough that I think I'll need a
serial console to get anything meaningful. I didn't see any oopses;
rather, it seemed like the driver was misbehaving. I don't know
whether the fault is in my patch, or elsewhere in the pdc2027x driver.
I don't have time tonight (or probably this week, for that matter) to
look into this further.

As a reminder (in case anyone else jumps into this thread in the
future), 2.6.17-rc4-mm3 works perfectly...
--
-Barry K. Nathan <barryn@xxxxxxxxx>