Re: [PATCH 2/2] usercopy: Convert test_user_copy to KUnit test

[Kees Cook]

On Wed, May 29, 2024 at 01:17:35PM +0100, Ivan Orlov wrote:
> On 5/19/24 20:12, Kees Cook wrote:
> >   #define test(condition, msg, ...)					\
> >   ({									\
> >   	int cond = (condition);						\
> >   	if (cond)							\
> > -		pr_warn("[%d] " msg "\n", __LINE__, ##__VA_ARGS__);	\
> > +		KUNIT_EXPECT_FALSE_MSG(test, cond, msg, ##__VA_ARGS__);	\
> >   	cond;								\
> >   })
> It looks like the 'test' macro is not used anymore, so probably it should be
> removed.

Oops, yes. Thanks!

> > +static int usercopy_test_init(struct kunit *test)
> > +{
> > +	struct usercopy_test_priv *priv;
> > +	unsigned long user_addr;
> > -	if (ret == 0) {
> > -		pr_info("tests passed.\n");
> > -		return 0;
> > -	}
> > +	priv = kunit_kzalloc(test, sizeof(*priv), GFP_KERNEL);
> > +	if (!priv)
> > +		return -ENOMEM;
> 
> Should the check be done with KUNIT_ASSERT_NOT_ERR_OR_NULL here as well, as
> it is done with priv->kmem?

Yes, that's much more idiomatic. I'll adjust this too.

> Other than that,
> 
> Tested-by: Ivan Orlov <ivan.orlov0322@xxxxxxxxx>

Thanks!

-- 
Kees Cook