[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: System freeze (again)



Richard L. Mace wrote:
X-Loop: debian-amd64@lists.debian.org
List-Id: <debian-amd64.lists.debian.org>
List-Post: <mailto:debian-amd64@lists.debian.org>
List-Help: <mailto:debian-amd64-request@lists.debian.org?subject=help>
List-Subscribe: <mailto:debian-amd64-request@lists.debian.org?subject=subscribe>
List-Unsubscribe: <mailto:debian-amd64-request@lists.debian.org?subject=unsubscribe>
Precedence: list
Resent-Sender: debian-amd64-request@lists.debian.org
Resent-Date: Thu, 16 Nov 2006 12:40:19 -0600 (CST)

I was following with interest the thread about system freezes, but it seems to have dried up. So, I was wondering if people have found a fix. Sorry for this
long post...

My system (HP nx6125 laptop, AMD Turion 64, ATI chipset, ATI Radeon Mobility XPRESS 200M) started to have "random" freezes shortly after the xorg 7.0 -> 7.1 transition in etch. The symptoms seem to be very similar to those described in the "System freeze" thread of a few weeks ago. I first noticed this a few weeks ago when I was viewing my system logs and scrolling by holding down the arrow keys. Sudden freeze during scroll. Reboot and execute "less /var/log/syslog". I scrolled down a bit more and freeze again (during scroll). I did several experiments in the interim. I just booted the machine and left it on. Eventually the screensaver would come on, work for a few minutes and then freeze. Sometimes I could prolong the time before a freeze by interrupting the screensaver with a move of the mouse shortly after it came on, but eventually the system would freeze, typically within a timescale of hours. This happens with the radeon xorg drivers, the ati xorg drivers and proprietary fglrx 8.30.3.

Today, I decided to use the laptop for useful purposes (i.e., not diagnose bugs ;-)). I was editing a moderate sized file with kate and doing the typical, scrolling up and down with mouse and keyboard. I experienced two freezes while doing a long scroll by holding the numpad arrow key down (so that the keyboard repeats) and one while scrolling with the synaptics touchpad (mouse pointer). I briefly switched from fglrx to the ati driver, rebooted, turned away from my laptop to do something else and it froze up.

In every case there is nothing written to the system logs. The machine will not even respond to ping requests once frozen. It needs a complete reboot to get it going again.

I have run memtest86+ for hours without errors. The probability that this is a software/Xorg bug is high, given the sudden increase in reports of this kind of late (so I'm not ready to remove memory chips from my laptop yet...). Google has been unhelpful. This seems to be a common problem on various distributions including ubuntu. The causes seem to be many and varied: presence of powernow-k8 module (I use this), network and usb mouse related (http://ubuntuforums.org/showthread.php?p=1738666), Nvidia related (see previous thread), clock-skew related (powernow-k8 ?), etc. However, there seems to be no definitive answer as to what the cause is, or what the fix is.

In my experience, it seems that it might be triggered by prolonged scrolling or mouse activity (or lack thereof?). But this is such a common user activity that my experiences could just be coincidence. I have not managed to get kernel 2.6.19-rc5 to work properly on my laptop, so that fix (mentioned above) is not an option for me now (does it really work?). I am currently running debian etch (amd 64) with xorg 7.1 and a custom built 2.6.18.2 kernel. I autoload the powernow-k8 module and use the kernel "ondemand" governor by default. I have tried with a range of older (custom built) kernels going back to 2.6.17.x and nothing changes. None of my 32bit debian etch desktops (which do not use frequency scaling) exhibit this. None of my colleagues' 32 bit debian etch desktops (which do not use frequency scaling) exhibit this.
Any help would be appreciated.... My apologies again for the long post.

Richard


As mentioned earlier I would try to downgrade the python-central to 0.5.8. Seems like you started to have the problem at the same time as when 0.5.9 hit etch. When I upgraded to 0.5.9 I could not compile a kernel without system freeze.

I also found that the system sometimes froze after running a screensaver for several hours - when reactivating the system. So now I have disabled the screensaver.

I haven't had a system freeze since.

Anders



Reply to: