kern/91408: ata(4) failure: SETFEATURES SET TRANSFER MODE semaphore timeout !! DANGER Will Robinson !!

Thede Loder thede at loder.com
Fri Jan 27 15:28:00 PST 2006


>Number:         91410
>Category:       kern
>Synopsis:       Promise SATA controller + striping : "SETFEATURES SET 
TRANSFER MODE semaphore timeout"
>Confidential:   no
>Severity:       critical
>Priority:       high
>Responsible:    freebsd-bugs
>State:          open
>Quarter:        
>Keywords:       
>Date-Required:
>Class:          sw-bug
>Submitter-Id:   current-users
>Arrival-Date:   Fri Jan 06 18:30:04 GMT 2006
>Closed-Date:
>Last-Modified:
>Originator:     John McLaughlin
>Release:        6.0-R
>Organization:
>Environment:
FreeBSD jmcl.shadowbox.ie 6.0-STABLE FreeBSD 6.0-STABLE #0: Fri Jan  6 
13:35:01 GMT 2006 

Interestingly, this bug disappeared for me somewhere around 6.0-STABLE #4 
(i386), but as of 6.0-STABLE #9 or 10, it is back.   

What I've noticed: 
* It only affects drives/filesystems attached via a SATA card
* It occurs with SI 3113 or Promise 20575 SATA controllers, maybe others
* The error starts under heavy file system load
  (What I usually do to trigger it is to start a file system copy of 
  several gigs worth of files from a drive attached to the SATA controller to
  another drive on the motherboard bus (Via 8235).  Sometimes this 
  copy alone will trigger it.  But its much more likely if I start 
  some other read operation on the drive, like 'ls -lR' or a 
  $ 'wc `find . -type file`) 
* it is reproducable

based upon my own experience with it, I'd say that the following three 
bugs may be the same thing: 

kern/91410
kern/91408
i386/86364

Soren, I think this is your domain.  Since it's fairly reproducible, let 
me know what I can do to help you find it.  

E: thede at loder.com









More information about the freebsd-bugs mailing list