Re: CVE-2023-52560: mm/damon/vaddr-test: fix memory leak in damon_do_test_apply_three_regions()

[Michal Hocko]

On Wed 06-03-24 08:42:07, Greg KH wrote:
> On Wed, Mar 06, 2024 at 08:49:42AM +0100, Michal Hocko wrote:
> > On Tue 05-03-24 22:25:11, Greg KH wrote:
> > > On Tue, Mar 05, 2024 at 05:51:11PM +0100, Michal Hocko wrote:
> > > > On Sat 02-03-24 22:59:54, Greg KH wrote:
> > > > > Description
> > > > > ===========
> > > > > 
> > > > > In the Linux kernel, the following vulnerability has been resolved:
> > > > > 
> > > > > mm/damon/vaddr-test: fix memory leak in damon_do_test_apply_three_regions()
> > > > > 
> > > > > When CONFIG_DAMON_VADDR_KUNIT_TEST=y and making CONFIG_DEBUG_KMEMLEAK=y
> > > > > and CONFIG_DEBUG_KMEMLEAK_AUTO_SCAN=y, the below memory leak is detected.
> > > > 
> > > > This is a kunit test case AFAICS. Is this really a CVE material?
> > > 
> > > People run kunit tests on real systems (again, we do not dictate use
> > > cases.)  So yes, fixing a memory leak that can be triggered is resolving
> > > a weakness and so should get a CVE I would think, right?
> > 
> > This is stretching the meaning of CVE beyond my imagination. Up to you
> > to decide but I yet have to see a real production system that casually
> > runs unit test just for <looking for a reason .... but failed>.
> 
> I know of at least one place that uses kunit tests in "production", and
> I know of more that will be enabling them in newer releases, so this is
> a real thing.

I would be really curious to hear more details.

> Again, we just mark "fixes for a weakness" as a CVE and
> let others decide what to do with it.

OK, this is something we have discussed and concluded to disagree. Not
my call though but I would really like to hear _who_ outside of the stable
tree userbase is really appreciating this approach.
-- 
Michal Hocko
SUSE Labs