Subject: Re: ESP SCSI controller errors?
To: None <he@netbsd.org, port-sparc@netbsd.org>
From: None <eeh@netbsd.org>
List: port-sparc
Date: 01/27/2002 19:31:28
|  o I am unsure whether the error messages of the form
|
|    esp0: !TC on DATA XFER [intr 10, stat 83, step 4] prevphase 1, resid=
|  c00
|
|    is indicative of "drive rot" or "driver rot".  I've considered
|    installing another drive in the machine, but if the driver is at
|    fault that effort will be wasted.  Comments appreciated.

This is different from some of the other errors reported.  We really
need to handle different errors separately.

In this case what's happened is that the driver got a bus service
interrupt and the status says the transfer is complete but the data 
count has not decremented to 0.  The phase is actually a status phase
so it appears the disk switched from a data in phase to a status phase
before transferring all the data.  

Given that, it sounds like this is "drive rot" rather than "driver rot",
but it would be much more interesting to learn what status the drive
is returning in this circumstance.  It may be possible to work around
it in the driver.

So exactly how long ago did this problem start?  And what's the date
of the kernel sources?

Eduardo