[igt-dev] [PATCH i-g-t] tests/i915/gem_exec_basic: Skip subtest gtt if gtt mapping is unavailable

Chris Wilson chris at chris-wilson.co.uk
Fri Jan 31 10:56:58 UTC 2020


Quoting Petri Latvala (2020-01-31 10:50:37)
> On Thu, Jan 30, 2020 at 04:15:16PM -0800, Antonio Argenziano wrote:
> > The subtest tries to do a gtt mapping but if there is no mappable
> > aperture that is not possible.
> > 
> > Signed-off-by: Antonio Argenziano <antonio.argenziano at intel.com>
> > ---
> >  tests/i915/gem_exec_basic.c | 1 +
> >  1 file changed, 1 insertion(+)
> > 
> > diff --git a/tests/i915/gem_exec_basic.c b/tests/i915/gem_exec_basic.c
> > index 70dce34b..08336168 100644
> > --- a/tests/i915/gem_exec_basic.c
> > +++ b/tests/i915/gem_exec_basic.c
> > @@ -93,6 +93,7 @@ static void gtt(int fd, uint64_t flags)
> >       struct drm_i915_gem_exec_object2 *exec;
> >       uint32_t handle;
> >  
> > +     gem_require_mappable_ggtt(fd);
> >       gem_require_ring(fd, flags);
> 
> 
> Hmm, without mappable aperture this ends up yielding a SKIP for each
> dynamic subtest. Note to self: Check that it gets properly handled...
> 
> Anyway, it's useless to loop and yield multiple SKIPs when you know a
> priori you're going to do it for each loop element.

There's no reason to loop over engines at all for this, since it's just
testing passing a pointer to the kernel and copy_from_user(). Sigh. It's
a confused test.
-Chris


More information about the igt-dev mailing list