[LEAPSECS] How USB bugs are reported versus UTC

[Warner Losh]

On Aug 22, 2013, at 8:40 PM, Rob Seaman wrote:


> So if Linux mis-implements USB, it's reported that way:

> 

> 	http://www.paritynews.com/2013/08/22/2437/misinterpretation-of-standard-probably-causing-usb-disconnects-on-resume-in-linux/


Devices get detached, big deal. No crashes happen here. And resume is an event the user initiates. It doesn't happen semi-randomly like leapseconds. For resume, the user knows they just opened the lid. For leap seconds, the user likely isn't even aware of it.


> ...but if Linux mis-implements UTC, we get:

> 

> 	http://www.wired.com/wiredenterprise/2012/07/leap-second-bug-wreaks-havoc-with-java-linux/

> 

> Hyperbole and hullabaloo redirected to /dev/null.  (Worth a try...)


Systems crashing, no matter the cause, is a big deal and is rightly classified as 'havoc' :) Especially when it happens seemingly at random. Leap seconds are an in-your-face event like resume, so the proximate cause, and the proper work arounds, aren't always readily apparent.

Warner