Re: kern/54212: AMD SB700-SB900 SATA Controller has issue with NCQ

[Tobias Ulmer <tobiasu%tmux.org@localhost>]

The following reply was made to PR kern/54212; it has been noted by GNATS.

From: Tobias Ulmer <tobiasu%tmux.org@localhost>
To: gnats-bugs%netbsd.org@localhost
Cc: 
Subject: Re: kern/54212: AMD SB700-SB900 SATA Controller has issue with NCQ
Date: Fri, 17 May 2019 18:57:51 +0200

 On Fri, May 17, 2019 at 04:40:01PM +0000, Michael van Elst wrote:
 > The following reply was made to PR kern/54212; it has been noted by GNATS.
 > 
 > From: mlelstv%serpens.de@localhost (Michael van Elst)
 > To: gnats-bugs%netbsd.org@localhost
 > Cc: 
 > Subject: Re: kern/54212: AMD SB700-SB900 SATA Controller has issue with NCQ
 > Date: Fri, 17 May 2019 16:38:16 -0000 (UTC)
 > 
 >  tobiasu%tmux.org@localhost writes:
 >  
 >  >wd1f: error writing fsbn 193889968 of 193889968-193889975 (wd1 bn 281972400; cn 279734 tn 8 sn 24), slot 1, retry 1
 >  >wd1: (aborted command, interface CRC error)
 >  >wd1f: requeue writing fsbn 193888152 of 193888152-193888159 (wd1 bn 281970584; cn 279732 tn 11 sn 35), slot 2, retry 1
 >  
 >  No idea why it should be related to NCQ, but CRC errors are usually
 >  related to cable/connector problems. Would also explain why changing
 >  the drive itself didn't help.
 
 This occured to me of course.  One of the first things I've tried was changing
 cables and port. That's why the boot disk is currently wd1 :)
 
 Not sure how this can *not* be related to the NCQ merge? I've literally
 tracked it down to that single commit with a whole series of kernels.
 
 It's not like this is a random failure that happens only some of the time.
 I can trigger it within 5 seconds, 100% of the time.
 
 Before NCQ is fine, after that it's always broken.
 
 >  
 >  -- 
 >  -- 
 >                                  Michael van Elst
 >  Internet: mlelstv%serpens.de@localhost
 >                                  "A potential Snark may lurk in every tree."
 >