i386/103435: Kernel appears somewhat deadlocked during heavy ATA I/O (post-August 4th)

[Jeremy Chadwick]

The following reply was made to PR i386/103435; it has been noted by GNATS.

From: Jeremy Chadwick <freebsd at jdc.parodius.com>
To: FreeBSD-gnats-submit at FreeBSD.org, freebsd-i386 at FreeBSD.org
Cc:  
Subject: Re: i386/103435: Kernel appears somewhat deadlocked during heavy ATA I/O (post-August 4th)
Date: Fri, 22 Sep 2006 17:22:01 -0700

 Follow-up to this PR (requesting closure):
 
 The system this PR was based on was an Intel D865GLC with
 a P4 CPU; no issues up until my PR was filed.
 
 I am now using an Asus A8N-E board with an AMD X2 CPU,
 and am not experiencing any issues what-so-ever.  I've
 managed to make -j2 buildkernel as well, which is a
 pretty good test.
 
 I'm left believing there's either some recent FreeBSD
 ATA code weirdness going on which affects the ICH5 on
 the Intel board (no idea) that isn't affecting the nVidia
 MCP9 SATA controller, or that the ICH5 on the Intel board
 has simply gone bad.
 
 Since I'm not the only one reporting this problem (but
 no one recently has reported it), I'm left to believe there's
 some general kernel oddness going on -- which I do not have
 the capability/knowledge to debug.
 
 If someone who is willing to work on this problem wishes
 to receive the board + CPU + RAM for it, free of charge
 (shipped within the USA only), I will gladly donate it to
 whomever wishes to debug the problem.
 
 Otherwise, please close this PR.
 
 -- 
 | Jeremy Chadwick                                 jdc at parodius.com |
 | Parodius Networking                        http://www.parodius.com/ |
 | UNIX Systems Administrator                   Mountain View, CA, USA |
 | Making life hard for others since 1977.               PGP: 4BD6C0CB |