drm/i915: Add mechanism to submit a context WA on ring submission

This patch adds framework to submit an arbitrary batchbuffer on each
context switch to clear residual state for render engine on Gen7/7.5
devices.

The idea of always emitting the context and vm setup around each request
is primary to make reset recovery easy, and not require rewriting the
ringbuffer. As each request would set up its own context, leaving it to
the HW to notice and elide no-op context switches, we could restart the
ring at any point, and reorder the requests freely.

However, to avoid emitting clear_residuals() between consecutive requests
in the ringbuffer of the same context, we do want to track the current
context in the ring. In doing so, we need to be careful to only record a
context switch when we are sure the next request will be emitted.

This security mitigation change does not trigger any performance
regression. Performance is on par with current mainline/drm-tip.

v2: Update vm_alias params to point to correct address space "vm" due to
changes made in the patch "f21613797bae98773"

v3-v4: none

Signed-off-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Signed-off-by: Prathap Kumar Valsan <prathap.kumar.valsan@intel.com>
Signed-off-by: Akeem G Abodunrin <akeem.g.abodunrin@intel.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Balestrieri Francesco <francesco.balestrieri@intel.com>
Cc: Bloomfield Jon <jon.bloomfield@intel.com>
Cc: Dutt Sudeep <sudeep.dutt@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Link: https://patchwork.freedesktop.org/patch/msgid/20200306000957.2836150-1-chris@chris-wilson.co.uk

1 files changed, 134 insertions, 4 deletions

diff --git a/drivers/gpu/drm/i915/gt/intel_ring_submission.c b/drivers/gpu/drm/i915/gt/intel_ring_submission.c
index ee241b7eaa3b..f0ce70861e93 100644
--- a/drivers/gpu/drm/i915/gt/intel_ring_submission.c
+++ b/drivers/gpu/drm/i915/gt/intel_ring_submission.c
@@ -1356,7 +1356,9 @@ static int load_pd_dir(struct i915_request *rq,
 	return rq->engine->emit_flush(rq, EMIT_FLUSH);
 }
 
-static inline int mi_set_context(struct i915_request *rq, u32 flags)
+static inline int mi_set_context(struct i915_request *rq,
+				 struct intel_context *ce,
+				 u32 flags)
 {
 	struct drm_i915_private *i915 = rq->i915;
 	struct intel_engine_cs *engine = rq->engine;
@@ -1431,7 +1433,7 @@ static inline int mi_set_context(struct i915_request *rq, u32 flags)
 
 	*cs++ = MI_NOOP;
 	*cs++ = MI_SET_CONTEXT;
-	*cs++ = i915_ggtt_offset(rq->context->state) | flags;
+	*cs++ = i915_ggtt_offset(ce->state) | flags;
 	/*
 	 * w/a: MI_SET_CONTEXT must always be followed by MI_NOOP
 	 * WaMiSetContext_Hang:snb,ivb,vlv
@@ -1546,13 +1548,56 @@ static int switch_mm(struct i915_request *rq, struct i915_address_space *vm)
 	return rq->engine->emit_flush(rq, EMIT_INVALIDATE);
 }
 
+static int clear_residuals(struct i915_request *rq)
+{
+	struct intel_engine_cs *engine = rq->engine;
+	int ret;
+
+	ret = switch_mm(rq, vm_alias(engine->kernel_context->vm));
+	if (ret)
+		return ret;
+
+	if (engine->kernel_context->state) {
+		ret = mi_set_context(rq,
+				     engine->kernel_context,
+				     MI_MM_SPACE_GTT | MI_RESTORE_INHIBIT);
+		if (ret)
+			return ret;
+	}
+
+	ret = engine->emit_bb_start(rq,
+				    engine->wa_ctx.vma->node.start, 0,
+				    0);
+	if (ret)
+		return ret;
+
+	ret = engine->emit_flush(rq, EMIT_FLUSH);
+	if (ret)
+		return ret;
+
+	/* Always invalidate before the next switch_mm() */
+	return engine->emit_flush(rq, EMIT_INVALIDATE);
+}
+
 static int switch_context(struct i915_request *rq)
 {
+	struct intel_engine_cs *engine = rq->engine;
 	struct intel_context *ce = rq->context;
+	void **residuals = NULL;
 	int ret;
 
 	GEM_BUG_ON(HAS_EXECLISTS(rq->i915));
 
+	if (engine->wa_ctx.vma && ce != engine->kernel_context) {
+		if (engine->wa_ctx.vma->private != ce) {
+			ret = clear_residuals(rq);
+			if (ret)
+				return ret;
+
+			residuals = &engine->wa_ctx.vma->private;
+		}
+	}
+
 	ret = switch_mm(rq, vm_alias(ce->vm));
 	if (ret)
 		return ret;
@@ -1560,7 +1605,7 @@ static int switch_context(struct i915_request *rq)
 	if (ce->state) {
 		u32 flags;
 
-		GEM_BUG_ON(rq->engine->id != RCS0);
+		GEM_BUG_ON(engine->id != RCS0);
 
 		/* For resource streamer on HSW+ and power context elsewhere */
 		BUILD_BUG_ON(HSW_MI_RS_SAVE_STATE_EN != MI_SAVE_EXT_STATE_EN);
@@ -1572,7 +1617,7 @@ static int switch_context(struct i915_request *rq)
 		else
 			flags |= MI_RESTORE_INHIBIT;
 
-		ret = mi_set_context(rq, flags);
+		ret = mi_set_context(rq, ce, flags);
 		if (ret)
 			return ret;
 	}
@@ -1581,6 +1626,20 @@ static int switch_context(struct i915_request *rq)
 	if (ret)
 		return ret;
 
+	/*
+	 * Now past the point of no return, this request _will_ be emitted.
+	 *
+	 * Or at least this preamble will be emitted, the request may be
+	 * interrupted prior to submitting the user payload. If so, we
+	 * still submit the "empty" request in order to preserve global
+	 * state tracking such as this, our tracking of the current
+	 * dirty context.
+	 */
+	if (residuals) {
+		intel_context_put(*residuals);
+		*residuals = intel_context_get(ce);
+	}
+
 	return 0;
 }
 
@@ -1765,6 +1824,11 @@ static void ring_release(struct intel_engine_cs *engine)
 
 	intel_engine_cleanup_common(engine);
 
+	if (engine->wa_ctx.vma) {
+		intel_context_put(engine->wa_ctx.vma->private);
+		i915_vma_unpin_and_release(&engine->wa_ctx.vma, 0);
+	}
+
 	intel_ring_unpin(engine->legacy.ring);
 	intel_ring_put(engine->legacy.ring);
 
@@ -1912,6 +1976,60 @@ static void setup_vecs(struct intel_engine_cs *engine)
 	engine->emit_fini_breadcrumb = gen7_xcs_emit_breadcrumb;
 }
 
+static int gen7_ctx_switch_bb_setup(struct intel_engine_cs * const engine,
+				    struct i915_vma * const vma)
+{
+	return 0;
+}
+
+static int gen7_ctx_switch_bb_init(struct intel_engine_cs *engine)
+{
+	struct drm_i915_gem_object *obj;
+	struct i915_vma *vma;
+	int size;
+	int err;
+
+	size = gen7_ctx_switch_bb_setup(engine, NULL /* probe size */);
+	if (size <= 0)
+		return size;
+
+	size = ALIGN(size, PAGE_SIZE);
+	obj = i915_gem_object_create_internal(engine->i915, size);
+	if (IS_ERR(obj))
+		return PTR_ERR(obj);
+
+	vma = i915_vma_instance(obj, engine->gt->vm, NULL);
+	if (IS_ERR(vma)) {
+		err = PTR_ERR(vma);
+		goto err_obj;
+	}
+
+	vma->private = intel_context_create(engine); /* dummy residuals */
+	if (IS_ERR(vma->private)) {
+		err = PTR_ERR(vma->private);
+		goto err_obj;
+	}
+
+	err = i915_vma_pin(vma, 0, 0, PIN_USER | PIN_HIGH);
+	if (err)
+		goto err_private;
+
+	err = gen7_ctx_switch_bb_setup(engine, vma);
+	if (err)
+		goto err_unpin;
+
+	engine->wa_ctx.vma = vma;
+	return 0;
+
+err_unpin:
+	i915_vma_unpin(vma);
+err_private:
+	intel_context_put(vma->private);
+err_obj:
+	i915_gem_object_put(obj);
+	return err;
+}
+
 int intel_ring_submission_setup(struct intel_engine_cs *engine)
 {
 	struct intel_timeline *timeline;
@@ -1965,11 +2083,19 @@ int intel_ring_submission_setup(struct intel_engine_cs *engine)
 
 	GEM_BUG_ON(timeline->hwsp_ggtt != engine->status_page.vma);
 
+	if (IS_GEN(engine->i915, 7) && engine->class == RENDER_CLASS) {
+		err = gen7_ctx_switch_bb_init(engine);
+		if (err)
+			goto err_ring_unpin;
+	}
+
 	/* Finally, take ownership and responsibility for cleanup! */
 	engine->release = ring_release;
 
 	return 0;
 
+err_ring_unpin:
+	intel_ring_unpin(ring);
 err_ring:
 	intel_ring_put(ring);
 err_timeline_unpin:
@@ -1980,3 +2106,7 @@ err:
 	intel_engine_cleanup_common(engine);
 	return err;
 }
+
+#if IS_ENABLED(CONFIG_DRM_I915_SELFTEST)
+#include "selftest_ring_submission.c"
+#endif