Re: [igt-dev] [PATCH i-g-t] igt/kms_vblank: Exercise some EINVAL

[Daniel Vetter <daniel@ffwll.ch>]

On Tue, Sep 04, 2018 at 07:51:59PM +0100, Chris Wilson wrote:
> Feed some impossible garbage into drmWaitVblank and expect the kernel to
> report EINVAL.
> 
> Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
> Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
> ---
>  tests/kms_vblank.c | 45 +++++++++++++++++++++++++++++++++++++++++++++
>  1 file changed, 45 insertions(+)
> 
> diff --git a/tests/kms_vblank.c b/tests/kms_vblank.c
> index b3cd2d93c..3e7aff3fe 100644
> --- a/tests/kms_vblank.c
> +++ b/tests/kms_vblank.c
> @@ -455,6 +455,48 @@ static void run_subtests_for_pipe(data_t *data)
>  	}
>  }
>  
> +static void invalid_subtest(data_t *data, int fd)
> +{
> +	union drm_wait_vblank vbl;
> +	unsigned long invalid_flags;
> +
> +	igt_display_require_output_on_pipe(&data->display, 0);
> +
> +	/* First check all is well with a simple query */
> +	memset(&vbl, 0, sizeof(vbl));
> +	vbl.request.type = DRM_VBLANK_RELATIVE;
> +	igt_assert_eq(wait_vblank(fd, &vbl), 0);
> +
> +	invalid_flags = ~(_DRM_VBLANK_TYPES_MASK |
> +			  _DRM_VBLANK_FLAGS_MASK |
> +			  _DRM_VBLANK_HIGH_CRTC_MASK);
> +
> +	/* pick some interesting invalid permutations */
> +	memset(&vbl, 0, sizeof(vbl));
> +	vbl.request.type = _DRM_VBLANK_RELATIVE | invalid_flags;
> +	igt_assert_eq(wait_vblank(fd, &vbl), -EINVAL);
> +	for (int bit = 0; bit < 32; bit++) {
> +		int err;
> +
> +		if (~invalid_flags & (1 << bit))

Double-take on the ~~, but looks correct. I expect CI to complain about
the dmesg-error on this one, but it won't report because it's not yet
vetted. But should be in the full results somewhere.

> +			continue;
> +
> +		memset(&vbl, 0, sizeof(vbl));
> +		vbl.request.type = _DRM_VBLANK_RELATIVE | (1 << bit);
> +		err = wait_vblank(fd, &vbl);
> +		igt_assert_f(err == -EINVAL,
> +			     "vblank wait with invalid request.type bit %d [0x%08x] did not report -EINVAL, got %d\n",
> +			     bit, 1 << bit, err);
> +	}
> +
> +	/* check the maximum pipe, nobody should have that many pipes! */
> +	memset(&vbl, 0, sizeof(vbl));
> +	vbl.request.type = _DRM_VBLANK_RELATIVE;
> +	vbl.request.type |= _DRM_VBLANK_SECONDARY;
> +	vbl.request.type |= _DRM_VBLANK_FLAGS_MASK;
> +	igt_assert_eq(wait_vblank(fd, &vbl), -EINVAL);

Hm just realized that the kernel lets _DRM_VBLANK_SECONDARY and
_DRM_VBLANK_HIGH_CRTC_MASK through at the same time. Not really an issue,
just a bit silly. But for another day.

Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>

> +}
> +
>  igt_main
>  {
>  	int fd;
> @@ -469,6 +511,9 @@ igt_main
>  		igt_display_require_output(&data.display);
>  	}
>  
> +	igt_subtest("invalid")
> +		invalid_subtest(&data, fd);
> +
>  	igt_subtest("crtc-id")
>  		crtc_id_subtest(&data, fd);
>  
> -- 
> 2.19.0.rc1
> 

-- 
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch
_______________________________________________
igt-dev mailing list
igt-dev@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/igt-dev