Bug#738521: xserver-xorg-video-ati: Persistant system lockup with ATI Radeon X300 GPU running firmware-linux-nonfree

[Michel Dänzer]

On 07.03.2015 21:36, Magnus Holmgren wrote:
> Monday 10 February 2014 00.22.05, Paul wrote:
>> Running Debian Wheezy, by installing the package "firmware-linux-nonfree"
>> which contains the firmware to enable 3D acceleration for the ATI card
>> (Radeon X300), when the system starts and user logins, the system locks.
>>
>> 2. Symptoms
>>
>> The input devices are inoperative, the screen image is undistorted but
>> unresponsive, the eth0 connection breaks (ssh is unavailable.) The REISUB
>> keys are ineffective. The only recourse is to do a hard reset by holding
>> the power button. Lockup occurs any where within seconds to within minutes
>> of bootup. Lockup most often occurs when in the graphical desktop
>> environment; rarely when in a virtual terminal (ALT F1-F6).
> 
> I've experienced very similar symptoms on jessie ever since 3D acceleration 
> via Glamor became possible with my Southern Island card (Bonaire XTX [Radeon 
> R7 260X]) about six months ago,

The symptoms may be similar, but it's very unlikely to be one and the
same problem. Please file your own report.


> although not immediately after logging in but 
> rather after several hours, and never when using the computer but when I leave 
> it unattended. I *think* it's locked up once while switched to a text console, 
> but otherwise only while the graphical console has been active.
> 
> Whenever it happens, the system starts spewing out _something_ on the ethernet 
> interface, making it impossible to connect to my server connected to the same 
> switch/AP (the LEDs of both ports flash furiously), but I haven't been able to 
> see anything with tcpdump. No clues can be found in any logs, except sometimes 
> a few "eth0: link up" in the server syslog, and just one time the affected 
> system managed to get off a message about a GPU fault via netconsole:
> 
> Feb  8 01:49:59 johansson [108159.201208] radeon 0000:01:00.0: GPU fault 
> detected: 146 0x08075014
> Feb  8 01:49:59 johansson [108159.201215] radeon 0000:01:00.0:   
> VM_CONTEXT1_PROTECTION_FAULT_ADDR   0x0007B6C0
> Feb  8 01:49:59 johansson [108159.201217] radeon 0000:01:00.0:   
> VM_CONTEXT1_PROTECTION_FAULT_STATUS 0x07050014
> 
> Once, a couple of weeks ago, I came back to the computer and found it 
> unresponsive; the screen had gone black but not into sleep mode and I could 
> still ssh in. chvt would just hang and when I tried restarting X (kdm), the 
> system stopped responding completely as before. So it seems that the system 
> might not lock up completely right away, but that it's a two-step process.

Which version of libgl1-mesa-dri and the kernel are you using?


-- 
Earthling Michel Dänzer               |               http://www.amd.com
Libre software enthusiast             |             Mesa and X developer

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 173 bytes
Desc: OpenPGP digital signature
URL: <http://lists.x.org/archives/xorg-driver-ati/attachments/20150309/203e0ef8/attachment.sig>