panic - sleeping thread on FreeBSD 8.0-stable / amd64

[C. P. Ghost]

On Sun, Feb 21, 2010 at 4:12 AM, C. P. Ghost <cpghost at cordula.ws> wrote:
> On Sun, Feb 21, 2010 at 12:35 AM, Jeremy Chadwick <freebsd at jdc.parodius.com> wrote:
>>
>> Some Linux users have reported AHCI-related issues with the SB600
>> southbridge, but the core of the problem turned out to be MSI on certain
>> AMD northbridges (specifically RS480, RS400, and RS200).
>
> Just one more data point.
>
> I have a machine with similar hardware: an MSI K9A2GM-FIH motherboard:
> http://eu.msi.com/index.php?func=proddesc&maincat_no=1&prod_no=1436
> with an AMD 780G chipset and AMD SB700 SATA controller,
> and I experienced freezes after switching to AHCI.
>
> Those freezes happened e.g. after some sustained random disk activity,
> followed by starting 'dvdisaster'. Then the HDD LED started to blink
> slooooowwwwly, every 10 seconds on, then off and again. No more disk
> activity on the aha0 controller was possible. The system remained responsive
> as long as it didn't involve disk activity (i.e. pings, mouse, keyboard etc..,
> but not starting new processes).

I just had a new crash with these symptoms, again exactly when
starting 'dvdisaster'. dmesg then shows a lot of those:

ahcich0: Timeout on slot 20
ahcich0: is 04000000 cs fff00007 ss fff00007 tfd 451 serr 00400000
ahcich0: port is not ready (timeout 10000ms) tfd = 00000480

repeating for multiple ports with different values, but every time
with tfd = 00000480, whatever that may be.

Obviously, the 10000ms timeout corresponds to the 10 seconds
period of the HDD LED going on and off when the machine is in
that state.

I didn't test this yet with msi disabled:

% sysctl -a | grep msi
hw.bce.msi_enable: 1
hw.pci.honor_msi_blacklist: 1
hw.pci.enable_msix: 1
hw.pci.enable_msi: 1
hw.drm.msi: 1

Thanks,
-cpghost.

-- 
Cordula's Web. http://www.cordula.ws/