Hmm. OK, so could these USB "hiccups" potentially be associated with the WiFi dongle itself? That might explain why (in at least some cases) the accept()s seem likely to be also associated with WiFi link loss. Maybe there are two kinds of USB hiccups? One associated with (say) the IK 0/0/4, and another with the WiFi dongle?
Just blabbing, you da man on this. As I said I know essentially nothing about USB. Maybe what I just said above about the hiccups doesn't even makes sense, I dunno.
Anyway, as I type here, I think I see that the parallel-ping experiment just stopped so let me see if it WSJSed (not sure yet). If so, I will post the logs later this evening or tmw morning...
Btw: Thanks much for all your attention on this, it's great to be working with you on it. I know you're busy and this obviously isn't the only thing on your plate, to put it mildly, so very much appreciate your time. Hopefully there will be a worthwhile payoff for you and Phidgets Inc if we can get to the bottom of this and get a robust source fix in place, maybe even rewrite some sections of the code to fix some of the worst offenses. That would take some pressure off you from folks (like moi and Andre
) who've been busting your chops about it for a while.
Anyway, let me check this expt, and will be back at ya with the logs and a posting before too long.