Practical computer tips, with a smattering of digital philosophy
I think I may have finally happened upon a definitive fix for at least one variation of the mysterious Ubuntu system crash that is accompanied by a flashing caps lock key. The problem, I think, can be traced back to the network drivers.
Chances are that if you’ve read this far you have a pretty good idea what I’m talking about, so I’ll refrain from describing the phenomenon in detail here, and simply refer you to this discussion from the Ubuntu forums. In general, however, what happens is that an afflicted computer will somehow kill the power to its own keyboard (and, in some cases, its mouse as well) upon resuming from sleep or screensaver mode. It does this for no apparent reason, and (in my experience) at totally random intervals. The operating system isn’t technically frozen, but it might as well be – particularly if you (like me) have your computer set up to prompt you for a password after being idle for x minutes. The only way to “fix” the problem is to pull the plug on your computer (do a cold reset) – never a particularly pleasant or safe experience, and potentially a truly disastrous one if you happen to have unsaved documents open at the time of the failure.
At any rate, I’ve yet to find any particularly convincing theory as to what might cause (or prevent) this phenomenon, so after a few bouts of research and much tinkering I’ve decided to put forth my own. I sincerely hope someone will (a) benefit from it, (b) test it, and/or (c) provide evidence that it’s totally and completely wrong. My main interest here is just to hone in on what’s actually causing the problem, because although it seems to be fairly widespread, few in the linux development community seem to be paying it much heed.
Anyway, to get to the point:
I realize that this notion may make absolutely no sense from a technical perspective, but I’m going off of pure empirical evidence here, and the evidence stands as a follows:
– This kind of failure always occurs during a period when a given computer is idle – when one would expect a screensaver to be invoked.
– I’ve often noticed various strange screen saver glitches and “artifacts” appearing immediately prior to the failure; sometimes the screen saver might fail to appear entirely; sometimes it might be fragmented, and certain sections of the screen would seem to be covered by black boxes, and so forth.
– The Ubuntu discussion mentioned above, which is by far the longest analysis I’ve been able to find of the issue, ended without really concluding anything, but some participants thought that the problem might be caused by compiz. I’m almost positive this isn’t the case, however, because I’ve encountered the issue under SuSE 10.2, which definately was not using compiz.
– I disabled my screensaver entirely a few days ago on my Ubuntu installation, and my keyboard hasn’t yet failed.
– Even if the GL screen saver engine were known to be buggy, who would make it a priority to rewrite/debug it? My guess is that screen savers probably aren’t on anyone’s list of top 100 (or 1000) development priorities.