On 2014-05-23 16:32, Cathey, Jim wrote:
Another issue that is often not thought about by many embedded developers, is that flash devices slow down with wear. Excessive logging to flash eventually results in a device that might no longer meet its performance objectives. The flashes I'm most familiar with have something like a 20:1 speed ratio between factory fresh and device death. That's usually pretty significant.
Yes. I don't _think_ my changes actually impact that, though--do they? Even though I'm changing the threshold at which the backlog-series is `full' (and at which logs fall off the end), I don't think I'm changing the *rate* of rotation, or of `pruning', or the rate of logging, either--just the initial time that it takes for logs to start being pruned. And, at least in my own use of the feature, I'm decreasing the amount of free space available for use in wear-levelling. Most likely I'm changing the pattern in which blocks are redistributed by wear-levelling.... Do you suppose that's dangerous? -- "'tis an ill wind that blows no minds." _______________________________________________ busybox mailing list busybox@busybox.net http://lists.busybox.net/mailman/listinfo/busybox
