[Ksummit-2012-discuss] Coverity "reports"

Jason Wessel jason.wessel at windriver.com
Tue Jun 19 19:30:44 UTC 2012

Dave pointed out something that was buried in the writeback and kernel
testing thread, and for those who already deleted it here is a pointer:


It was very fast to get setup by e-maling scan-admin at coverity.com,
just tell them what you maintain for kernel.org in your request.  In
fact the person who set up my account responded:

---- From the Coverity folks ----

We got so many request from Linux maintainer in last two days to setup
an account for Linux Kernel.  I was wondering what made that drive, is
it something that was found on Internet or did you hear it from other
maintainer?  I want to make sure if there are going to be huge
request, we can provide some easy interface to setup account.


The coverity for dummies basically amounts to:

1) login with the URL and user name they give you
2) Click on Linux-2.6
3) Scroll down on the left hand side to "File:"
4) Input a file filter for what you are interested in:
5) Scan through errors

I picked an arbitrary false positive.


And of course it gives the link as to how to resolve it by adding a
/* fallthrough */ at:


The tool seems very "picky" about things.  Given that we already have
problems with not enough review cycles, fixing some of these trivial
errors would further take away from those resources.  But there is
probably an army of newbies waiting to fix more spelling errors and
the like...

Sure enough, even in code I maintain there are a few of these, and
I'll probably fix them in a future merge window, because at least then
I can very easily know about the new problems and keep the "trivial
changers" at bay.  The question at hand is what kind of value do we as
a community want to place on getting rid of nit picks?


BTW Thanks for the pointers Dave, I'm interested in hearing more about
trinity and if you believe it is prime time ready where you think
other distro maintainers should be putting this into the range of
system tests?

More information about the Ksummit-2012-discuss mailing list