[v2] kunit/usercopy: Disable testing on !CONFIG_MMU

Since arch_pick_mmap_layout() is an inline for non-MMU systems, disable
this test there.

Reported-by: kernel test robot <lkp@intel.com>
Closes: https://lore.kernel.org/oe-kbuild-all/202406160505.uBge6TMY-lkp@intel.com/
Signed-off-by: Kees Cook <kees@kernel.org>
---
Resending as v2 with Shuah in To:
---
Cc: Shuah Khan <skhan@linuxfoundation.org>
Cc: Brendan Higgins <brendan.higgins@linux.dev>
Cc: David Gow <davidgow@google.com>
Cc: Rae Moar <rmoar@google.com>
Cc: "Gustavo A. R. Silva" <gustavoars@kernel.org>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: linux-kselftest@vger.kernel.org
Cc: kunit-dev@googlegroups.com
Cc: linux-hardening@vger.kernel.org
Cc: linux-mm@kvack.org
---
 lib/kunit/user_alloc.c | 4 ++++
 lib/usercopy_kunit.c   | 5 +++++
 mm/util.c              | 2 ++
 3 files changed, 11 insertions(+)

Hi Shuah,

Can you please add this to your -next tree since it fixes test failures
on non-MMU systems, after commit cf6219ee889f ("usercopy: Convert
test_user_copy to KUnit test").

Thanks!

-Kees

On Thu, Jun 20, 2024 at 05:13:35PM -0400, Rae Moar wrote:
> On Wed, Jun 19, 2024 at 4:25 PM Kees Cook <kees@kernel.org> wrote:
> >
> > Since arch_pick_mmap_layout() is an inline for non-MMU systems, disable
> > this test there.
> >
> > Reported-by: kernel test robot <lkp@intel.com>
> > Closes: https://lore.kernel.org/oe-kbuild-all/202406160505.uBge6TMY-lkp@intel.com/
> > Signed-off-by: Kees Cook <kees@kernel.org>
> 
> Hello!
> 
> This looks good to me. And seems to fix the problem. Thanks for the fix!
> 
> Reviewed-by: Rae Moar <rmoar@google.com>
> 
> -Rae
> 
> > ---
> > Resending as v2 with Shuah in To:
> > ---
> > Cc: Shuah Khan <skhan@linuxfoundation.org>
> > Cc: Brendan Higgins <brendan.higgins@linux.dev>
> > Cc: David Gow <davidgow@google.com>
> > Cc: Rae Moar <rmoar@google.com>
> > Cc: "Gustavo A. R. Silva" <gustavoars@kernel.org>
> > Cc: Andrew Morton <akpm@linux-foundation.org>
> > Cc: linux-kselftest@vger.kernel.org
> > Cc: kunit-dev@googlegroups.com
> > Cc: linux-hardening@vger.kernel.org
> > Cc: linux-mm@kvack.org
> > ---
> >  lib/kunit/user_alloc.c | 4 ++++
> >  lib/usercopy_kunit.c   | 5 +++++
> >  mm/util.c              | 2 ++
> >  3 files changed, 11 insertions(+)
> >
> > diff --git a/lib/kunit/user_alloc.c b/lib/kunit/user_alloc.c
> > index 76d3d1345ed7..ae935df09a5e 100644
> > --- a/lib/kunit/user_alloc.c
> > +++ b/lib/kunit/user_alloc.c
> > @@ -30,6 +30,10 @@ static int kunit_attach_mm(void)
> >         if (current->mm)
> >                 return 0;
> >
> > +       /* arch_pick_mmap_layout() is only sane with MMU systems. */
> > +       if (!IS_ENABLED(CONFIG_MMU))
> > +               return -EINVAL;
> > +
> >         mm = mm_alloc();
> >         if (!mm)
> >                 return -ENOMEM;
> > diff --git a/lib/usercopy_kunit.c b/lib/usercopy_kunit.c
> > index 45f1e558c464..e819561a540d 100644
> > --- a/lib/usercopy_kunit.c
> > +++ b/lib/usercopy_kunit.c
> > @@ -290,6 +290,11 @@ static int usercopy_test_init(struct kunit *test)
> >         struct usercopy_test_priv *priv;
> >         unsigned long user_addr;
> >
> > +       if (!IS_ENABLED(CONFIG_MMU)) {
> > +               kunit_skip(test, "Userspace allocation testing not available on non-MMU systems");
> > +               return 0;
> > +       }
> > +
> >         priv = kunit_kzalloc(test, sizeof(*priv), GFP_KERNEL);
> >         KUNIT_ASSERT_NOT_ERR_OR_NULL(test, priv);
> >         test->priv = priv;
> > diff --git a/mm/util.c b/mm/util.c
> > index df37c47d9374..e70e8e439258 100644
> > --- a/mm/util.c
> > +++ b/mm/util.c
> > @@ -484,7 +484,9 @@ void arch_pick_mmap_layout(struct mm_struct *mm, struct rlimit *rlim_stack)
> >         clear_bit(MMF_TOPDOWN, &mm->flags);
> >  }
> >  #endif
> > +#ifdef CONFIG_MMU
> >  EXPORT_SYMBOL_IF_KUNIT(arch_pick_mmap_layout);
> > +#endif
> >
> >  /**
> >   * __account_locked_vm - account locked pages to an mm's locked_vm
> > --
> > 2.34.1
> >

On Thu, 20 Jun 2024 at 04:25, Kees Cook <kees@kernel.org> wrote:
>
> Since arch_pick_mmap_layout() is an inline for non-MMU systems, disable
> this test there.
>
> Reported-by: kernel test robot <lkp@intel.com>
> Closes: https://lore.kernel.org/oe-kbuild-all/202406160505.uBge6TMY-lkp@intel.com/
> Signed-off-by: Kees Cook <kees@kernel.org>
> ---
> Resending as v2 with Shuah in To:
> ---
> Cc: Shuah Khan <skhan@linuxfoundation.org>
> Cc: Brendan Higgins <brendan.higgins@linux.dev>
> Cc: David Gow <davidgow@google.com>
> Cc: Rae Moar <rmoar@google.com>
> Cc: "Gustavo A. R. Silva" <gustavoars@kernel.org>
> Cc: Andrew Morton <akpm@linux-foundation.org>
> Cc: linux-kselftest@vger.kernel.org
> Cc: kunit-dev@googlegroups.com
> Cc: linux-hardening@vger.kernel.org
> Cc: linux-mm@kvack.org
> ---

Reviewed-by: David Gow <davidgow@google.com>

Cheers,
-- David

On 6/27/24 11:07, Kees Cook wrote:
> Hi Shuah,
> 
> Can you please add this to your -next tree since it fixes test failures
> on non-MMU systems, after commit cf6219ee889f ("usercopy: Convert
> test_user_copy to KUnit test").
> 

Applied to linux-kselftest kunit branch for next.

thanks,
-- Shuah