[PATCH igt] tests/kms_flip: Move EBUSY to a new busy-flip subtest

[PATCH igt] tests/kms_flip: Move EBUSY to a new busy-flip subtest

[ville.syrjala]

From: Ville Syrjälä <ville.syrjala@linux.intel.com>

The EBUSY checking is very fragile currently in case there's any kind
extra delay in the test loop. At least the flip-vs-rmfb fails reliably
on my IVB.

So to make the test less fragile remove the EBUSY check from all the
current flip tests, and instead add a specific busy-flip test. To
better guarantee that we get the EBUSY add a blocking vblank wait just
before we issue the first flip. This maximizes the time we have to
submit two flips during one frame.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
---
 tests/kms_flip.c | 15 +++++++++++----
 1 file changed, 11 insertions(+), 4 deletions(-)

diff --git a/tests/kms_flip.c b/tests/kms_flip.c
index 6711f09..46ce2ed 100644
--- a/tests/kms_flip.c
+++ b/tests/kms_flip.c
@@ -894,6 +894,12 @@ static unsigned int run_test_step(struct test_output *o)
 		igt_assert_f(hang, "failed to exercise page flip hang recovery\n");
 	}
 
+	/* try to make sure we can issue two flips during the same frame */
+	if (do_flip && (o->flags & TEST_EBUSY)) {
+		struct vblank_reply reply;
+		igt_assert(__wait_for_vblank(TEST_VBLANK_BLOCK, o->pipe, 1, 0, &reply) == 0);
+	}
+
 	if (do_flip)
 		do_or_die(do_page_flip(o, new_fb_id, !(o->flags & TEST_NOEVENT)));
 
@@ -1536,12 +1542,13 @@ int main(int argc, char **argv)
 		{ 60,  TEST_VBLANK | TEST_MODESET | TEST_WITH_DUMMY_RCS,
 					"rcs-wf_vblank-vs-modeset" },
 
-		{ 30, TEST_FLIP | TEST_EBUSY , "plain-flip" },
+		{ 30, TEST_FLIP , "plain-flip" },
+		{ 30, TEST_FLIP | TEST_EBUSY , "busy-flip" },
 		{ 30, TEST_FLIP | TEST_FENCE_STRESS , "flip-vs-fences" },
-		{ 30, TEST_FLIP | TEST_CHECK_TS | TEST_EBUSY , "plain-flip-ts-check" },
-		{ 30, TEST_FLIP | TEST_CHECK_TS | TEST_EBUSY | TEST_FB_RECREATE,
+		{ 30, TEST_FLIP | TEST_CHECK_TS, "plain-flip-ts-check" },
+		{ 30, TEST_FLIP | TEST_CHECK_TS | TEST_FB_RECREATE,
 			"plain-flip-fb-recreate" },
-		{ 30, TEST_FLIP | TEST_EBUSY | TEST_RMFB | TEST_MODESET , "flip-vs-rmfb" },
+		{ 30, TEST_FLIP | TEST_RMFB | TEST_MODESET , "flip-vs-rmfb" },
 		{ 60, TEST_FLIP | TEST_DPMS | TEST_EINVAL, "flip-vs-dpms" },
 		{ 60, TEST_FLIP | TEST_DPMS | TEST_WITH_DUMMY_BCS, "bcs-flip-vs-dpms" },
 		{ 60, TEST_FLIP | TEST_DPMS | TEST_WITH_DUMMY_RCS, "rcs-flip-vs-dpms" },
-- 
1.8.5.5

_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

Re: [PATCH igt] tests/kms_flip: Move EBUSY to a new busy-flip subtest

[Daniel Vetter]

On Wed, Jun 04, 2014 at 03:56:19PM +0300, ville.syrjala@linux.intel.com wrote:
> From: Ville Syrjälä <ville.syrjala@linux.intel.com>
> 
> The EBUSY checking is very fragile currently in case there's any kind
> extra delay in the test loop. At least the flip-vs-rmfb fails reliably
> on my IVB.
> 
> So to make the test less fragile remove the EBUSY check from all the
> current flip tests, and instead add a specific busy-flip test. To
> better guarantee that we get the EBUSY add a blocking vblank wait just
> before we issue the first flip. This maximizes the time we have to
> submit two flips during one frame.
> 
> Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>

Yeah, no need to test api error cases in funky special situations.
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
> ---
>  tests/kms_flip.c | 15 +++++++++++----
>  1 file changed, 11 insertions(+), 4 deletions(-)
> 
> diff --git a/tests/kms_flip.c b/tests/kms_flip.c
> index 6711f09..46ce2ed 100644
> --- a/tests/kms_flip.c
> +++ b/tests/kms_flip.c
> @@ -894,6 +894,12 @@ static unsigned int run_test_step(struct test_output *o)
>  		igt_assert_f(hang, "failed to exercise page flip hang recovery\n");
>  	}
>  
> +	/* try to make sure we can issue two flips during the same frame */
> +	if (do_flip && (o->flags & TEST_EBUSY)) {
> +		struct vblank_reply reply;
> +		igt_assert(__wait_for_vblank(TEST_VBLANK_BLOCK, o->pipe, 1, 0, &reply) == 0);
> +	}
> +
>  	if (do_flip)
>  		do_or_die(do_page_flip(o, new_fb_id, !(o->flags & TEST_NOEVENT)));
>  
> @@ -1536,12 +1542,13 @@ int main(int argc, char **argv)
>  		{ 60,  TEST_VBLANK | TEST_MODESET | TEST_WITH_DUMMY_RCS,
>  					"rcs-wf_vblank-vs-modeset" },
>  
> -		{ 30, TEST_FLIP | TEST_EBUSY , "plain-flip" },
> +		{ 30, TEST_FLIP , "plain-flip" },
> +		{ 30, TEST_FLIP | TEST_EBUSY , "busy-flip" },
>  		{ 30, TEST_FLIP | TEST_FENCE_STRESS , "flip-vs-fences" },
> -		{ 30, TEST_FLIP | TEST_CHECK_TS | TEST_EBUSY , "plain-flip-ts-check" },
> -		{ 30, TEST_FLIP | TEST_CHECK_TS | TEST_EBUSY | TEST_FB_RECREATE,
> +		{ 30, TEST_FLIP | TEST_CHECK_TS, "plain-flip-ts-check" },
> +		{ 30, TEST_FLIP | TEST_CHECK_TS | TEST_FB_RECREATE,
>  			"plain-flip-fb-recreate" },
> -		{ 30, TEST_FLIP | TEST_EBUSY | TEST_RMFB | TEST_MODESET , "flip-vs-rmfb" },
> +		{ 30, TEST_FLIP | TEST_RMFB | TEST_MODESET , "flip-vs-rmfb" },
>  		{ 60, TEST_FLIP | TEST_DPMS | TEST_EINVAL, "flip-vs-dpms" },
>  		{ 60, TEST_FLIP | TEST_DPMS | TEST_WITH_DUMMY_BCS, "bcs-flip-vs-dpms" },
>  		{ 60, TEST_FLIP | TEST_DPMS | TEST_WITH_DUMMY_RCS, "rcs-flip-vs-dpms" },
> -- 
> 1.8.5.5
> 
> _______________________________________________
> Intel-gfx mailing list
> Intel-gfx@lists.freedesktop.org
> http://lists.freedesktop.org/mailman/listinfo/intel-gfx

-- 
Daniel Vetter
Software Engineer, Intel Corporation
+41 (0) 79 365 57 48 - http://blog.ffwll.ch