drm/i915: Defer final intel_wakeref_put to process context

As we need to acquire a mutex to serialise the final
intel_wakeref_put, we need to ensure that we are in process context at
that time. However, we want to allow operation on the intel_wakeref from
inside timer and other hardirq context, which means that need to defer
that final put to a workqueue.

Inside the final wakeref puts, we are safe to operate in any context, as
we are simply marking up the HW and state tracking for the potential
sleep. It's only the serialisation with the potential sleeping getting
that requires careful wait avoidance. This allows us to retain the
immediate processing as before (we only need to sleep over the same
races as the current mutex_lock).

v2: Add a selftest to ensure we exercise the code while lockdep watches.
v3: That test was extremely loud and complained about many things!
v4: Not a whale!

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=111295
References: https://bugs.freedesktop.org/show_bug.cgi?id=111245
References: https://bugs.freedesktop.org/show_bug.cgi?id=111256
Fixes: 18398904ca9e ("drm/i915: Only recover active engines")
Fixes: 51fbd8de87dc ("drm/i915/pmu: Atomically acquire the gt_pm wakeref")
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Reviewed-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20190808202758.10453-1-chris@chris-wilson.co.uk

1 files changed, 83 insertions, 0 deletions

diff --git a/drivers/gpu/drm/i915/gt/selftest_engine_pm.c b/drivers/gpu/drm/i915/gt/selftest_engine_pm.c
new file mode 100644
index 000000000000..3a1419376912
--- /dev/null
+++ b/drivers/gpu/drm/i915/gt/selftest_engine_pm.c
@@ -0,0 +1,83 @@
+/*
+ * SPDX-License-Identifier: GPL-2.0
+ *
+ * Copyright © 2018 Intel Corporation
+ */
+
+#include "i915_selftest.h"
+#include "selftest_engine.h"
+#include "selftests/igt_atomic.h"
+
+static int live_engine_pm(void *arg)
+{
+	struct intel_gt *gt = arg;
+	struct intel_engine_cs *engine;
+	enum intel_engine_id id;
+
+	/*
+	 * Check we can call intel_engine_pm_put from any context. No
+	 * failures are reported directly, but if we mess up lockdep should
+	 * tell us.
+	 */
+	if (intel_gt_pm_wait_for_idle(gt)) {
+		pr_err("Unable to flush GT pm before test\n");
+		return -EBUSY;
+	}
+
+	GEM_BUG_ON(intel_gt_pm_is_awake(gt));
+	for_each_engine(engine, gt->i915, id) {
+		const typeof(*igt_atomic_phases) *p;
+
+		for (p = igt_atomic_phases; p->name; p++) {
+			/*
+			 * Acquisition is always synchronous, except if we
+			 * know that the engine is already awake, in which
+			 * case we should use intel_engine_pm_get_if_awake()
+			 * to atomically grab the wakeref.
+			 *
+			 * In practice,
+			 *    intel_engine_pm_get();
+			 *    intel_engine_pm_put();
+			 * occurs in one thread, while simultaneously
+			 *    intel_engine_pm_get_if_awake();
+			 *    intel_engine_pm_put();
+			 * occurs from atomic context in another.
+			 */
+			GEM_BUG_ON(intel_engine_pm_is_awake(engine));
+			intel_engine_pm_get(engine);
+
+			p->critical_section_begin();
+			if (!intel_engine_pm_get_if_awake(engine))
+				pr_err("intel_engine_pm_get_if_awake(%s) failed under %s\n",
+				       engine->name, p->name);
+			else
+				intel_engine_pm_put(engine);
+			intel_engine_pm_put(engine);
+			p->critical_section_end();
+
+			/* engine wakeref is sync (instant) */
+			if (intel_engine_pm_is_awake(engine)) {
+				pr_err("%s is still awake after flushing pm\n",
+				       engine->name);
+				return -EINVAL;
+			}
+
+			/* gt wakeref is async (deferred to workqueue) */
+			if (intel_gt_pm_wait_for_idle(gt)) {
+				pr_err("GT failed to idle\n");
+				return -EINVAL;
+			}
+		}
+	}
+
+	return 0;
+}
+
+int live_engine_pm_selftests(struct intel_gt *gt)
+{
+	static const struct i915_subtest tests[] = {
+		SUBTEST(live_engine_pm),
+	};
+
+	return intel_gt_live_subtests(tests, gt);
+}