Description of Problem: When XFree has decides that it is time to transition from xscreensaver to DPMS standby mode, the OS system can hang if the mouse is moved during the transition. Version-Release number of selected component (if applicable): Red Hat 7.2 Kernel 2.4.9-31 XFree86-4.1.0-15 xscreensaver-3.33-4 How Reproducible: Tricky. It's happened to me twice of the last 3 or 4 months, although I haven't tried too hard to reproduce it. Steps to Reproduce: 1. xscreensaver options: Blanking & Locking: Blank After = 10:00, Cycle After = 3:00 Display Power Management: Enabled Standby After = 20:00 Suspend After = 25:00 Off After = 30:00 2. Wait for DPMS to enter standby mode. Sometimes the monitor will display an "entering standby mode" message. 3. As this happens, move the mouse as though to wake up the system and leave the screen saver. Actual Results: Screen enters standby mode. Stays that way. Disk activity light stays off. Pressing Num Lock or Caps Lock results in no change in the keyboard status lights. System seems hung. Unfortunately, I do not have another system available with which to attempt network access. But it sure seems dead to me. Expected Results: I expected my Gnome desktop to be redisplayed, as it typically does when the screen saver is running, or the monitor is in power saving mode, and I move the mouse. Additional Information: Whatever is happening when the system hangs, it seems to make the CPU very busy. I have lm_sensors running to monitor my system temperature, and after letting the system sit in its hung state for 20 minutes and then restarting, the CPU temp was significantly higher than before the hang. Once rebooted, the temperature stabilized at the original temperature as expected. Nothing special in the system log at the time of failure. There are a bunch of these suspicious messages, however: kernel: [drm:radeon_freelist_get] *ERROR* returning NULL! (sometimes repeated hundreds of times!) In XF86Config-4, I have: Section "Device" Identifier "Radeon" VendorName "ATI" Driver "ati" Option "AGPMode" "2" EndSection My video card: ATI Radeon 7000 (PCI).
The problem you're describing sounds like it would be incredibly difficult to try and reproduce in order to gather enough data to take a stab at isolating the problem. In order to look into this, I will need you to supply detailed steps to be able to reproduce it easily with a high chance of reproduceability. Once every 2 months is something impossible to debug. If you can start up X, and do a series of individual steps that cause the problem to occur 100% of the time, or even 80-90% of the time, in short order, then it would be much more debuggable. Please try to do this and provided the detailed steps you've used to reproduce. Also, once you've been able to reproduce it, can you disable DRI, and try to reproduce it again many times to see if disabling DRI makes the problem go away?
Hello?
I'm still here. Sorry, I have not yet had the time to investigate this further. Similar, though not identical, weirdness has me thinking that I may be on the wrong track (example: OS hang while screen saver is running, well before DPMS transition). Also, it could be that my XFree device configuration is incomplete. Anyway, I apologize for the delay.
I spent some time trying to reproduce this bug. Couldn't do it, not even once. I changed my screen saver start and DPMS standby times to 1 and 2 minutes respectively to expedite the testing. Turns out my XFree configuration was OK. All of this is with DRI enabled.
Sounds like a random weirdness that disappeared. I'm unable to reproduce, and you appear to be unable to reproduce now also, so I'm closing the bug as WORKSFORME (WORKSFORUS).