i915/i915_pm_rps: Tighten fence generation for waitboost

We need the object we wait on to be clean, and free of stray fences as
the waitboosting only applies to i915_request that have not been
started. If we use an exclusive fence (from a write dependency) that
will strip off any foreign fences, allowing us to trick the driver into
behaving as we expect.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Acked-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>

tests/i915/i915_pm_rps.c

@@ -538,26 +538,15 @@ static void stabilize_check(int *out)
 	igt_debug("Waited %d msec to stabilize cur\n", wait);
 }
 
-static void resubmit_batch(int fd, uint32_t handle, int count)
-{
-	struct drm_i915_gem_exec_object2 obj = {
-		.handle = handle
-	};
-	struct drm_i915_gem_execbuffer2 eb = {
-		.buffer_count = 1,
-		.buffers_ptr = to_user_pointer(&obj),
-	};
-	while (count--)
-		gem_execbuf(fd, &eb);
-}
-
 static void boost_freq(int fd, int *boost_freqs)
 {
 	int64_t timeout = 1;
 	igt_spin_t *load;
 
 	load = igt_spin_new(fd);
-	resubmit_batch(fd, load->handle, 16);
+
+	/* Strip off extra fences from the object, and keep it from starting */
+	igt_spin_free(fd, igt_spin_new(fd, .dependency = load->handle));
 
 	/* Waiting will grant us a boost to maximum */
 	gem_wait(fd, load->handle, &timeout);