On 12/13/2017 2:00 AM, Linus Torvalds wrote:

On Mon, Dec 11, 2017 at 9:20 PM, Byungchul Park [off-list ref] wrote:

quoted

The *problem* is false positives, since locks and waiters in
kernel are not classified properly

So the problem is that those false positives apparently end up being a
big deal for the filesystem people.

I personally don't think the code itself has to be removed, but I do
think that it should never have been added on as part of the generic
lock proving, and should always have been a separate config option.

I admit it.

I also feel that you dismiss "false positives" much too easily. A

I don't dismiss the ones easily...

Anyway, I mostly agree with your whole opinion.

Thanks for replying.

false positive is a big problem - because it makes people ignore the
real cases (or just disable the functionality entirely).

It's why I am very quick to disable compiler warnings that have false
positives, for example. Just a couple of "harmless" false positive
warnings will poison the real warnings for people because they'll get
used to seeing warnings while building, and no longer actually look at
them.

                  Linus

-- 
Thanks,
Byungchul