Re: "device error via D2H FIS" w/ Samsung drive and Si3726

On Sat, Jul 30, 2011 at 5:55 AM, Tejun Heo [off-list ref] wrote:

On Tue, Jul 26, 2011 at 08:59:39PM -0700, Ryan Castellucci wrote:

quoted

I've got a eight bay RAID enclosure with two Si3726 based port
multipliers. I have some Seagate ST31500341AS drives which work fine,
but my Samsung HD204UI do not like this.  I've tried swapping the
cable.  I've tried both with a on-board JMB362 AHCI and a Si3132 and
get similar errors.  On the Si3132 (with latest non-raid firmware):

I have eight of these Samsung drives and they all do the same thing,
but work fine when direct attached.

Here's what spews into the kernel log:

Can you please attach full dmesg output?  The log you posted is too
incomplete to tell what's going on.

Thanks for responding... I seem to not be getting emails from the list
- maybe majordomo doesn't like me? (i tried to subscribe...)

I eventually figured out that I was having pretty much the exact same
problem as the guy in this thread:

http://www.spinics.net/lists/linux-ide/msg38500.html

The unlock command[1] there worked once the disks settled... and my
claim of them working fine direct attached was based on an identical
disk that hadn't been molested by the 3ware controller.

Having the kernel detect a 'drive is locked' condition and give an
error about that (as suggested in the other thread) would have saved
me a few hours of troubleshooting. It looks like some work was done on
a patch to implement that functionality[2] but I'm not sure what
became of it - this particular box has 2.6.38 on it which hasn't
implemented that.
If you would still like the full dmesg output let me know and I'll dig
up my logs and attach them.

1. hdparm --user-master u --security-disable "3wareUserPassword
       " /dev/sdX
2. http://www.spinics.net/lists/linux-ide/msg38566.html
-- 
Ryan Castellucci http://ryanc.org/