[Ksummit-2012-discuss] [ATTEND] stress testing, regressions, bugs
davej at redhat.com
Fri Jun 15 14:39:23 UTC 2012
I've spent much of the past year finding bugs in dozens of parts of the kernel
using my stress testing tool 'trinity'. I'd like to take some time discussing..
- what this tool is/does, as there still seem to be some people unaware.
- possibly talk about other tools I'm working on if I get them finished in time.
- get feedback for ways to enhance these tools.
and perhaps as a separate follow-up session, I'd like to talk about the rate at which
we're introducing regressions. Many of the bugs trinity finds aren't found during
testing on next, but show up within a day of code appearing in Linus' tree during -rc.
Where is the process failing ?
Finally, as another possible session, I'd like to present some anecdotal data
on the bugs we've been getting against the Fedora kernel. Over the last year I've
been collecting a lot of data on this, and while the results aren't particularly
surprising, they do highlight we still have problems in a number of areas.
More information about the Ksummit-2012-discuss