Hi,

Having received the replacement express card (197b:2363) I can confirm 
that the card in question appears to have been faulty.

The new express card performs flawlessly, I can cold boot / warm boot to 
Linux or windows without receiving any error message, and after several 
days of testing external disks I can honestly say that everything went 
without a hitch.

So, I would be happy if the bug report was closed with reason being due 
to faulty hardware.

regards

Orson

On 16/12/11 04:55, Show lu wrote:

Dear Sirs,
Good day,
I am JMicron/Show and take over JMB36X FAE activity
Please help to shows the PCIE CFG registers(of JMB36X) at first.


Thanks
Sincerely
Show



-----Original Message-----
From: Steven Chen [mailto:stevenchen@jmicron.com]
Sent: Friday, December 16, 2011 10:52 AM
To: 'Jonathan Nieder'; linux-ide@vger.kernel.org; 'Show lu'
Cc: 'Tejun Heo'; 'Orson'; 'Aries Lee'
Subject: RE: JMicron Technology JMB362 JMB363 SError: { DevExch }
ExpressCard, sometimes fails to boot

Add Show in this mail loop.

-----Original Message-----
From: Jonathan Nieder [mailto:jrnieder@gmail.com]
Sent: Wednesday, December 14, 2011 6:45 AM
To: linux-ide@vger.kernel.org
Cc: 'Tejun Heo'; 'Orson'; Aries Lee
Subject: Re: JMicron Technology JMB362 JMB363 SError: { DevExch }
ExpressCard, sometimes fails to boot

(Aries: sorry for the duplicate message)
Hi,

Aries Lee wrote:

quoted

Is it possible to install windows OS on the same hardware,
If it still get the same error, that we can conclude it as a hardware

issue.

Orson wrote:

quoted

I installed WinXP SP3 as a dual boot option onto my HP NX6325.
I booted into windows and my dual port express card was found immediately,

I

quoted

installed the latest driver from J Micron and everything seemed to be OK.
I looked through the winxp system messages and all was OK, I looked

through

quoted

device manager and it said everything was OK, all resources were OK.
I attached an external esata drive and it was detected instantly and was
available in my list of drives, I copied some files over, all went OK, I

ran

quoted

scandisk to check the drive and all went OK, the drive did not seem to
re-attach itself or give any errors.  So it seems to work fine in windows.

All right; looks like the hardware works.

quoted

I then shut down my laptop and removed the esata drive and booted into

Debian

quoted

wheezy, same error appeared during boot (without any drive attached)
irq_stat 0x00000040, connection status changed
ata1: SError: { PHYInt CommWake DevExch }
ata1: hard resetting link
ata1: COMRESET failed (errno=-32)
ata1: reset failed (errno=-32), retrying in 8 secs
ata1: limiting SATA link speed to 1.5 Gbps
ata1: hard resetting link
ata1: SATA link down (SStatus 0 SControl 310)
ata1: EH complete

etc, etc.

Right.

quoted

I booted back to windows and all was OK, but, here is what I have found,

if I

quoted

boot into Linux from windows without shutting down (warm boot) then the
express card is initialised in Debian without problems and works as

normal, if

quoted

I shut down and cold boot straight to Debian, the error appears again.

Interesting.  So it sounds like Windows is doing some initialization
that Linux does not.

quoted

Perhaps the card is faulty or (I suspect) there is a problem with the bios

on

quoted

the NX6325, the latest bios is dated 2007 and HP seem to have stopped
supporting this model, even though mine was purchased new in 2007.

Please send "acpidump" output as an attachment for reference.

quoted

Anyway, because the express card is so cheap I have just ordered a

replacement

quoted

of the same type, so this will tell me if it's the card that is at fault.

Ok.  Thanks for your work.

linux-ide folks: any ideas for debugging this further?  For example,
is there a way to find the difference between the state after a cold
boot and a warm boot from windows?

Ciao,
Jonathan