svn commit: r232071 - head/sys/vm

[Alexey Dokuchaev]

On Mon, May 27, 2019 at 02:34:09PM -0700, Rodney W. Grimes wrote:
> I do not at all mean to discourage what you are doing, it is good to go
> over static analysis reports, the problem is that there are often many
> false positives,

There are tons of them indeed, but that's normal for any static analysis
tool.  I've attempted to clean up the original log by ignoring the most
noisy and dubious diagnostics and maintaining false positives list,
starting with the ones pointed out by kib at .

> It would be nice if we had a "team" that looked at all the Coverity
> data, and any other data like what you have offered up here.  Part of
> the problem is that few want to do that work, or those that do want to
> think it is low hanging fruit that anyone can do.

That's exactly right, generating the check log is easy, finding people
motivated enough to triage it (or finding the motivation within oneself)
is tough.  Maybe this ~2.4k line log would attract more interest than
the first raw, full >16k line one:

    freefall:/home/danfe/pvs-kernel-filtered-2019-05-28.log.xz

I'd appreciate if experts in their areas (cam/scsi, dev drivers, geom,
kern, netinet, nfs, etc.) glanced over and see if they might help with
anything there.  If you find a false positive, please tell me so I
can add it to the ignore list.  Thanks!
 
./danfe