I like it when I can report something improving, as opposed to just making notes about things that don’t work.
And even better this time, it’s a rather cranky piece of hardware that I wanted very badly to be able to use about six months ago, but was having next to no luck — the sassy Corega wireless card that was spewing forth “noise floor calibration errors” at every turn.
I never really did find out what the error was, or if it was something I was doing, but I can say that since I moved up with kernels — I’m running 126.96.36.199 at the moment — the errors and dropped connections and spooling messages have all disappeared without sending a postcard.
And good riddance. “Annoying” would be an understatement, given that I couldn’t figure out what to do to shut it up. I do remember tracking down a developer’s note about those messages, that they were somehow enabled or triggered at the code level and ought to be turned off. But I did nothing (that I am aware of) other than install newer kernels, and I haven’t seen a complaint from it in the week I’ve been testing it, so probably somebody else, somewhere on the planet, commented out the code that was making it chatter at me.
Yes, I plugged it in sometime a week ago, when I wanted to use the WPC11 card as a crutch for installing the Broadcom-based wireless in the 600m. So basically, sometime while I was playing musical chairs with my spare networking connections, the Atheros 5K-based card suddenly started behaving itself.
And since it sat quiet and polite for a week without giving me any issue, I’m content to call it “improved.” Of course, now that I’ve typed those words, it will probably start hiccupping again. That’s how these things work, you know.