Re: SATA PMP support in 2.6.24-rc3?

[Tejun Heo <htejun@gmail.com>]

Gaston, Jason D wrote:
> For reference, here is the dmesg output from the 2.6.24-rc7 kernel W/O
> the patch applied.

Man, this is weird.  Do you happen to have a addonics device too?  I
have one 4726 and two 3726s here and none shows such problem.  Also,
there are quite a few people confirming that detection is okay on 3726
and 4726s.

The failure you're seeing is very similar to what happens on my 4726 if
the first fan-out port is not occupied.  4726 behaves very strangely
when its first port is empty.  Even BIOS has to time out on 4726 if the
first port is empty, so I've been thinking it's the peculiarity of the
chip (and considering the shitty ATA emulation of the pseudo config
device I'm not surprised).  Also, marvell PMP and new generation of SIMG
PMPs (5723 and friends) work fine with the current code.

I've never used windows or the binary only Linux driver.  Do they work
fine when the first port is unoccupied?  Also, is your PMP from
addonics?  In most reports, they're saying the array is addonics.  I'm
not sure whether it's because that vendor is so popular that there
aren't many alternatives or their hardware is somehow quirky.

-- 
tejun