Skip site navigation (1)Skip section navigation (2)
Date:      Fri, 26 Feb 2010 03:55:36 -0800
From:      Jeremy Chadwick <freebsd@jdc.parodius.com>
To:        Torfinn Ingolfsen <torfing@broadpark.no>
Cc:        freebsd-stable@freebsd.org
Subject:   Re: panic - sleeping thread on FreeBSD 8.0-stable / amd64
Message-ID:  <20100226115536.GA17798@icarus.home.lan>
In-Reply-To: <20100226110337.70d1a758.torfinn.ingolfsen@broadpark.no>
References:  <20100131144217.ca08e965.torfinn.ingolfsen@broadpark.no> <20100131175639.86ba9aee.torfinn.ingolfsen@broadpark.no> <20100207163631.da7205fc.torfinn.ingolfsen@broadpark.no> <20100213192404.5e15b5eb.torfinn.ingolfsen@broadpark.no> <20100217091625.d0e74570.torfinn.ingolfsen@broadpark.no> <20100220202108.e1dd1b74.torfinn.ingolfsen@broadpark.no> <20100220193718.GA33214@icarus.home.lan> <20100220224959.c424dd9e.torfinn.ingolfsen@broadpark.no> <20100220233546.GA36973@icarus.home.lan> <20100226110337.70d1a758.torfinn.ingolfsen@broadpark.no>

next in thread | previous in thread | raw e-mail | index | archive | help
On Fri, Feb 26, 2010 at 11:03:37AM +0100, Torfinn Ingolfsen wrote:
> > What exact disks (e.g. adX) are attached to ata5 and ata6?
> 
> root@kg-f2# dmesg | grep ata5
> ata5: <ATA channel 3> on atapci0
> ata5: [ITHREAD]
> ad10: 953869MB <SAMSUNG HD103SJ 1AJ100E4> at ata5-master UDMA100 SATA 3Gb/s
> root@kg-f2# dmesg | grep ata6
> ata6: <ATA channel 4> on atapci0
> ata6: [ITHREAD]
> ad12: 953869MB <SAMSUNG HD103SJ 1AJ100E4> at ata6-master UDMA100 SATA 3Gb/s
> ...snip...
> No, I didn't. I did state that full dmesg's and more info was available on the freebsd web page[1] for the machine 
> in one of my first posts.

Okay, so the breakdown for those following is:

http://sites.google.com/site/tingox/f2-dmesg-8.0-stable-20100131.txt?attredirects=0

atapci0: <ATI IXP700/800 SATA300 controller> port 0xff00-0xff07,0xfe00-0xfe03,0xfd00-0xfd07,0xfc00-0xfc03,0xfb00-0xfb0f mem 0xfe02f000-0xfe02f3ff irq 22 at device 17.0 on pci0
atapci0: [ITHREAD]
atapci0: AHCI v1.10 controller with 6 3Gbps ports, PM supported
ata2: <ATA channel 0> on atapci0
ata3: <ATA channel 1> on atapci0
ata4: <ATA channel 2> on atapci0
ata5: <ATA channel 3> on atapci0
ata6: <ATA channel 4> on atapci0
ata7: <ATA channel 5> on atapci0

ad6: 238475MB <SAMSUNG HD252HJ 1AC01118> at ata3-master UDMA100 SATA 3Gb/s
ad8: 953869MB <SAMSUNG HD103SJ 1AJ100E4> at ata4-master UDMA100 SATA 3Gb/s
ad10: 953869MB <SAMSUNG HD103SJ 1AJ100E4> at ata5-master UDMA100 SATA 3Gb/s
ad12: 953869MB <SAMSUNG HD103SJ 1AJ100E4> at ata6-master UDMA100 SATA 3Gb/s
ad14: 953869MB <SAMSUNG HD103SJ 1AJ100E4> at ata7-master UDMA100 SATA 3Gb/s

But the only ports which are having issues are ata5 and ata6, which
hosts disks ad10 and ad12 respectively.

SMART stats for ad10 and ad12 look fantastic, aside from slightly long
spin-up times (claiming over 8 seconds), but that wouldn't cause what's
seen here.  Both disks have used for nearly 1700 hours.  No SMART error
log entries exist on either disk, which means the timeouts seen when
speaking to the controller are very likely when talking to the
controller itself (and not when waiting for the controller to submit a
request to the disk and that piece stalling).

I'm out of ideas aside from the following:

1) Disabling MSI/MSIX, which at this point I'm doubting will fix
anything (but you never know), since I'd expect it to affect the
entire controller and not just specific ports on the controller.

2) Replacing the SATA cables used between ata5<-->ad10 and ata6<-->ad12.

3) Getting mav@ to talk to AMD to find out if there's any AHCI quirks in
the IXP700 or IXP800 SATA controllers, as there could be some weird
driver bug/quirk on FreeBSD which is needed.

Mainly for mav@: verbose boot messages for this system are here, in case
any SATA register details are of help:

http://sites.google.com/site/tingox/f2-dmesg-8.0-stable-20100131_verb1.txt?attredirects=0
http://sites.google.com/site/tingox/f2-dmesg-8.0-stable-20100131_verb2.txt?attredirects=0

-- 
| Jeremy Chadwick                                   jdc@parodius.com |
| Parodius Networking                       http://www.parodius.com/ |
| UNIX Systems Administrator                  Mountain View, CA, USA |
| Making life hard for others since 1977.              PGP: 4BD6C0CB |




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20100226115536.GA17798>