On Sun, Jan 15, 2017 at 11:38:21AM +0000, Laurent Bercot wrote: > Polling is evil for several reasons, the two most important of them being > the following.
Harsh but true. > I have met a real-life example of this at Google, of all places: sending ... > I have also met a real-life example of this, when I was working at > Sagemcom (a French manufacturer of embedded devices), making the base > system for an energy gateway (the thing that's supposed to collect and Thanks for sharing those, I find the Google example surprising given their reputation but once you've built surch a large infrastructure I suppose making changes becomes very hard. > When you're an init system, i.e. the lowest possible level for user-space > software, and you *already* introduce polling, well, it doesn't bode well > for the rest of your software stack. Your energy-saving device is already > screwed, and automation that relies on runsvdir picking up a new service > is already eating an average 7 second delay. As a desktop user, you > obviously don't care; as a software architect, this makes me shake my head. > Even systemd does better on that point. That is the crux of my original motivation, I don't actually know if the power consumption is going to be an issue but when I see CPU time accumulating for a process that should be idle then it's a cause for suspicion. The tradeoff here favors patching it if it's really an intrinsic problem and not some kind of misconfiguration.