drm fixes for 5.10-rc7

amdgpu:
 - SMU11 manual fan fix
 - Renoir display clock fix
 - VCN3 dynamic powergating fix
 
 i915:
 - Program mocs:63 for cache eviction on gen9 (Chris)
 - Protect context lifetime with RCU (Chris)
 - Split the breadcrumb spinlock between global and contexts (Chris)
 - Retain default context state across shrinking (Venkata)
 - Limit frequency drop to RPe on parking (Chris)
 - Return earlier from intel_modeset_init() without display (Jani)
 - Defer initial modeset until after GGTT is initialized (Chris)
 
 nouveau:
 - pre-nv50 regression fix
 
 rockchip:
 - uninitialised LVDS property fix
 
 omap:
 - bridge fix
 
 panel:
 - race fix
 
 mxsfb:
 - fence sync fix
 - modifiers fix
 
 tegra:
 - idr init fix
 - sor fixes
 - output/of cleanup fix
 -----BEGIN PGP SIGNATURE-----
 
 iQIcBAABAgAGBQJfyZorAAoJEAx081l5xIa+inUP/2HXHXeYTUn0dW26GEMM/2XR
 buPNs2f7YmV+dVh2Eb1zM22guTmE8INmVZtaUBlN2L/lmh2kAJqZKn5QU1QTFbmH
 XY7ASKh9HURid8nBt+ZSopRXRKZkjiU/0ii98/1vxmjIGm4hTNVcm4BN82Ed94QK
 KgSYQbqH6gpkwk71Qe9xJhv9d1ySQP/Ubq6+p9qrb8/Cq0xvgLhYjl7bJ4MGzZ3/
 o3ILxnudSFQhDGwU1FxZiHZHjU1iuIZIEC7NauAjppNE1nv8T6JLIXjfRkCnRVWU
 rYMExK3HzZEFiHJU9ebYj3loTjn3PLEhNdy7zs5jRO5pgAkSlf0UM1GIeXe0LwSz
 wxqjrCE22rBlcPJCoXS0JXvFA4O00NbzPSRdQiPdQnMC28rESnTDxUW5vo0KI/tf
 XL63v7q8JHq5wel+8jD9ahBqfMLnXN9AnNzvZDQydYJz5Px2tsryIOaPsELvz/8M
 33iiH8iGfgLYn9fJWMuU7neE+PHsKB+4TsrTMYnOt9Lm1kRRJhpedoWZuyiMmE8Y
 skfj+5XTBceQz5bPr/IjSxiBJtGydpz7GfAkJc1nTgkeZDVT7JSYHQHNLxkylh/v
 SV/6Bu/NwdovCEgjFf3/DIkc13ymom2YgDd+NQFuqJ+nNe0quHThUKUtTE+jFHPr
 Yx6fEUwxn+GaujOfAjFQ
 =uZ33
 -----END PGP SIGNATURE-----

Merge tag 'drm-fixes-2020-12-04' of git://anongit.freedesktop.org/drm/drm

Pull drm fixes from Dave Airlie:
 "This week's regular fixes.

  i915 has fixes for a few races, use-after-free, and gpu hangs. Tegra
  just has some minor fixes that I didn't see much point in hanging on
  to. The nouveau fix is for all pre-nv50 cards and was reported a few
  times. Otherwise it's just some amdgpu, and a few misc fixes.

  Summary:

  amdgpu:
   - SMU11 manual fan fix
   - Renoir display clock fix
   - VCN3 dynamic powergating fix

  i915:
   - Program mocs:63 for cache eviction on gen9 (Chris)
   - Protect context lifetime with RCU (Chris)
   - Split the breadcrumb spinlock between global and contexts (Chris)
   - Retain default context state across shrinking (Venkata)
   - Limit frequency drop to RPe on parking (Chris)
   - Return earlier from intel_modeset_init() without display (Jani)
   - Defer initial modeset until after GGTT is initialized (Chris)

  nouveau:
   - pre-nv50 regression fix

  rockchip:
   - uninitialised LVDS property fix

  omap:
   - bridge fix

  panel:
   - race fix

  mxsfb:
   - fence sync fix
   - modifiers fix

  tegra:
   - idr init fix
   - sor fixes
   - output/of cleanup fix"

* tag 'drm-fixes-2020-12-04' of git://anongit.freedesktop.org/drm/drm: (22 commits)
  drm/amdgpu/vcn3.0: remove old DPG workaround
  drm/amdgpu/vcn3.0: stall DPG when WPTR/RPTR reset
  drm/amd/display: Init clock value by current vbios CLKs
  drm/amdgpu/pm/smu11: Fix fan set speed bug
  drm/i915/display: Defer initial modeset until after GGTT is initialised
  drm/i915/display: return earlier from intel_modeset_init() without display
  drm/i915/gt: Limit frequency drop to RPe on parking
  drm/i915/gt: Retain default context state across shrinking
  drm/i915/gt: Split the breadcrumb spinlock between global and contexts
  drm/i915/gt: Protect context lifetime with RCU
  drm/i915/gt: Program mocs:63 for cache eviction on gen9
  drm/omap: sdi: fix bridge enable/disable
  drm/panel: sony-acx565akm: Fix race condition in probe
  drm/rockchip: Avoid uninitialized use of endpoint id in LVDS
  drm/tegra: sor: Disable clocks on error in tegra_sor_init()
  drm/nouveau: make sure ret is initialized in nouveau_ttm_io_mem_reserve
  drm: mxsfb: Implement .format_mod_supported
  drm: mxsfb: fix fence synchronization
  drm/tegra: output: Do not put OF node twice
  drm/tegra: replace idr_init() by idr_init_base()
  ...

drivers/gpu/drm/amd/amdgpu/vcn_v3_0.c

@@ -1011,6 +1011,11 @@ static int vcn_v3_0_start_dpg_mode(struct amdgpu_device *adev, int inst_idx, boo
 	tmp = REG_SET_FIELD(tmp, UVD_RBC_RB_CNTL, RB_RPTR_WR_EN, 1);
 	WREG32_SOC15(VCN, inst_idx, mmUVD_RBC_RB_CNTL, tmp);
 
+	/* Stall DPG before WPTR/RPTR reset */
+	WREG32_P(SOC15_REG_OFFSET(VCN, inst_idx, mmUVD_POWER_STATUS),
+		UVD_POWER_STATUS__STALL_DPG_POWER_UP_MASK,
+		~UVD_POWER_STATUS__STALL_DPG_POWER_UP_MASK);
+
 	/* set the write pointer delay */
 	WREG32_SOC15(VCN, inst_idx, mmUVD_RBC_RB_WPTR_CNTL, 0);
 
@@ -1033,6 +1038,10 @@ static int vcn_v3_0_start_dpg_mode(struct amdgpu_device *adev, int inst_idx, boo
 	WREG32_SOC15(VCN, inst_idx, mmUVD_RBC_RB_WPTR,
 		lower_32_bits(ring->wptr));
 
+	/* Unstall DPG */
+	WREG32_P(SOC15_REG_OFFSET(VCN, inst_idx, mmUVD_POWER_STATUS),
+		0, ~UVD_POWER_STATUS__STALL_DPG_POWER_UP_MASK);
+
 	return 0;
 }
 
@@ -1556,8 +1565,14 @@ static int vcn_v3_0_pause_dpg_mode(struct amdgpu_device *adev,
 					UVD_DPG_PAUSE__NJ_PAUSE_DPG_ACK_MASK,
 					UVD_DPG_PAUSE__NJ_PAUSE_DPG_ACK_MASK);
 
+				/* Stall DPG before WPTR/RPTR reset */
+				WREG32_P(SOC15_REG_OFFSET(VCN, inst_idx, mmUVD_POWER_STATUS),
+					UVD_POWER_STATUS__STALL_DPG_POWER_UP_MASK,
+					~UVD_POWER_STATUS__STALL_DPG_POWER_UP_MASK);
+
 				/* Restore */
 				ring = &adev->vcn.inst[inst_idx].ring_enc[0];
+				ring->wptr = 0;
 				WREG32_SOC15(VCN, inst_idx, mmUVD_RB_BASE_LO, ring->gpu_addr);
 				WREG32_SOC15(VCN, inst_idx, mmUVD_RB_BASE_HI, upper_32_bits(ring->gpu_addr));
 				WREG32_SOC15(VCN, inst_idx, mmUVD_RB_SIZE, ring->ring_size / 4);
@@ -1565,14 +1580,16 @@ static int vcn_v3_0_pause_dpg_mode(struct amdgpu_device *adev,
 				WREG32_SOC15(VCN, inst_idx, mmUVD_RB_WPTR, lower_32_bits(ring->wptr));
 
 				ring = &adev->vcn.inst[inst_idx].ring_enc[1];
+				ring->wptr = 0;
 				WREG32_SOC15(VCN, inst_idx, mmUVD_RB_BASE_LO2, ring->gpu_addr);
 				WREG32_SOC15(VCN, inst_idx, mmUVD_RB_BASE_HI2, upper_32_bits(ring->gpu_addr));
 				WREG32_SOC15(VCN, inst_idx, mmUVD_RB_SIZE2, ring->ring_size / 4);
 				WREG32_SOC15(VCN, inst_idx, mmUVD_RB_RPTR2, lower_32_bits(ring->wptr));
 				WREG32_SOC15(VCN, inst_idx, mmUVD_RB_WPTR2, lower_32_bits(ring->wptr));
 
-				WREG32_SOC15(VCN, inst_idx, mmUVD_RBC_RB_WPTR,
+				/* Unstall DPG */
-					RREG32_SOC15(VCN, inst_idx, mmUVD_SCRATCH2) & 0x7FFFFFFF);
+				WREG32_P(SOC15_REG_OFFSET(VCN, inst_idx, mmUVD_POWER_STATUS),
+					0, ~UVD_POWER_STATUS__STALL_DPG_POWER_UP_MASK);
 
 				SOC15_WAIT_ON_RREG(VCN, inst_idx, mmUVD_POWER_STATUS,
 					UVD_PGFSM_CONFIG__UVDM_UVDU_PWR_ON, UVD_POWER_STATUS__UVD_POWER_STATUS_MASK);
@@ -1630,10 +1647,6 @@ static void vcn_v3_0_dec_ring_set_wptr(struct amdgpu_ring *ring)
 {
 	struct amdgpu_device *adev = ring->adev;
 
-	if (adev->pg_flags & AMD_PG_SUPPORT_VCN_DPG)
-		WREG32_SOC15(VCN, ring->me, mmUVD_SCRATCH2,
-			lower_32_bits(ring->wptr) | 0x80000000);
-
 	if (ring->use_doorbell) {
 		adev->wb.wb[ring->wptr_offs] = lower_32_bits(ring->wptr);
 		WDOORBELL32(ring->doorbell_index, lower_32_bits(ring->wptr));

drivers/gpu/drm/amd/display/dc/clk_mgr/dcn21/rn_clk_mgr.c

@@ -163,8 +163,17 @@ void rn_update_clocks(struct clk_mgr *clk_mgr_base,
 			new_clocks->dppclk_khz = 100000;
 	}
 
-	if (should_set_clock(safe_to_lower, new_clocks->dppclk_khz, clk_mgr->base.clks.dppclk_khz)) {
+	/*
-		if (clk_mgr->base.clks.dppclk_khz > new_clocks->dppclk_khz)
+	 * Temporally ignore thew 0 cases for disp and dpp clks.
+	 * We may have a new feature that requires 0 clks in the future.
+	 */
+	if (new_clocks->dppclk_khz == 0 || new_clocks->dispclk_khz == 0) {
+		new_clocks->dppclk_khz = clk_mgr_base->clks.dppclk_khz;
+		new_clocks->dispclk_khz = clk_mgr_base->clks.dispclk_khz;
+	}
+
+	if (should_set_clock(safe_to_lower, new_clocks->dppclk_khz, clk_mgr_base->clks.dppclk_khz)) {
+		if (clk_mgr_base->clks.dppclk_khz > new_clocks->dppclk_khz)
 			dpp_clock_lowered = true;
 		clk_mgr_base->clks.dppclk_khz = new_clocks->dppclk_khz;
 		update_dppclk = true;

drivers/gpu/drm/amd/pm/swsmu/smu11/smu_v11_0.c

@@ -1164,7 +1164,12 @@ int smu_v11_0_set_fan_speed_rpm(struct smu_context *smu,
 	if (ret)
 		return ret;
 
-	crystal_clock_freq = amdgpu_asic_get_xclk(adev);
+	/*
+	 * crystal_clock_freq div by 4 is required since the fan control
+	 * module refers to 25MHz
+	 */
+
+	crystal_clock_freq = amdgpu_asic_get_xclk(adev) / 4;
 	tach_period = 60 * crystal_clock_freq * 10000 / (8 * speed);
 	WREG32_SOC15(THM, 0, mmCG_TACH_CTRL,
 		     REG_SET_FIELD(RREG32_SOC15(THM, 0, mmCG_TACH_CTRL),

drivers/gpu/drm/i915/display/intel_display.c

@@ -18021,16 +18021,6 @@ int intel_modeset_init_nogem(struct drm_i915_private *i915)
 	if (!HAS_GMCH(i915))
 		sanitize_watermarks(i915);
 
-	/*
-	 * Force all active planes to recompute their states. So that on
-	 * mode_setcrtc after probe, all the intel_plane_state variables
-	 * are already calculated and there is no assert_plane warnings
-	 * during bootup.
-	 */
-	ret = intel_initial_commit(dev);
-	if (ret)
-		drm_dbg_kms(&i915->drm, "Initial commit in probe failed.\n");
-
 	return 0;
 }
 
@@ -18039,11 +18029,21 @@ int intel_modeset_init(struct drm_i915_private *i915)
 {
 	int ret;
 
-	intel_overlay_setup(i915);
-
 	if (!HAS_DISPLAY(i915))
 		return 0;
 
+	/*
+	 * Force all active planes to recompute their states. So that on
+	 * mode_setcrtc after probe, all the intel_plane_state variables
+	 * are already calculated and there is no assert_plane warnings
+	 * during bootup.
+	 */
+	ret = intel_initial_commit(&i915->drm);
+	if (ret)
+		return ret;
+
+	intel_overlay_setup(i915);
+
 	ret = intel_fbdev_init(&i915->drm);
 	if (ret)
 		return ret;

drivers/gpu/drm/i915/gt/intel_breadcrumbs.c

@@ -101,18 +101,37 @@ static void __intel_breadcrumbs_disarm_irq(struct intel_breadcrumbs *b)
 	intel_gt_pm_put_async(b->irq_engine->gt);
 }
 
+static void intel_breadcrumbs_disarm_irq(struct intel_breadcrumbs *b)
+{
+	spin_lock(&b->irq_lock);
+	if (b->irq_armed)
+		__intel_breadcrumbs_disarm_irq(b);
+	spin_unlock(&b->irq_lock);
+}
+
 static void add_signaling_context(struct intel_breadcrumbs *b,
 				  struct intel_context *ce)
 {
-	intel_context_get(ce);
+	lockdep_assert_held(&ce->signal_lock);
-	list_add_tail(&ce->signal_link, &b->signalers);
+
+	spin_lock(&b->signalers_lock);
+	list_add_rcu(&ce->signal_link, &b->signalers);
+	spin_unlock(&b->signalers_lock);
 }
 
-static void remove_signaling_context(struct intel_breadcrumbs *b,
+static bool remove_signaling_context(struct intel_breadcrumbs *b,
 				     struct intel_context *ce)
 {
-	list_del(&ce->signal_link);
+	lockdep_assert_held(&ce->signal_lock);
-	intel_context_put(ce);
+
+	if (!list_empty(&ce->signals))
+		return false;
+
+	spin_lock(&b->signalers_lock);
+	list_del_rcu(&ce->signal_link);
+	spin_unlock(&b->signalers_lock);
+
+	return true;
 }
 
 static inline bool __request_completed(const struct i915_request *rq)
@@ -175,6 +194,8 @@ static void add_retire(struct intel_breadcrumbs *b, struct intel_timeline *tl)
 
 static bool __signal_request(struct i915_request *rq)
 {
+	GEM_BUG_ON(test_bit(I915_FENCE_FLAG_SIGNAL, &rq->fence.flags));
+
 	if (!__dma_fence_signal(&rq->fence)) {
 		i915_request_put(rq);
 		return false;
@@ -195,15 +216,12 @@ static void signal_irq_work(struct irq_work *work)
 	struct intel_breadcrumbs *b = container_of(work, typeof(*b), irq_work);
 	const ktime_t timestamp = ktime_get();
 	struct llist_node *signal, *sn;
-	struct intel_context *ce, *cn;
+	struct intel_context *ce;
-	struct list_head *pos, *next;
 
 	signal = NULL;
 	if (unlikely(!llist_empty(&b->signaled_requests)))
 		signal = llist_del_all(&b->signaled_requests);
 
-	spin_lock(&b->irq_lock);
-
 	/*
 	 * Keep the irq armed until the interrupt after all listeners are gone.
 	 *
@@ -229,47 +247,44 @@ static void signal_irq_work(struct irq_work *work)
 	 * interrupt draw less ire from other users of the system and tools
 	 * like powertop.
 	 */
-	if (!signal && b->irq_armed && list_empty(&b->signalers))
+	if (!signal && READ_ONCE(b->irq_armed) && list_empty(&b->signalers))
-		__intel_breadcrumbs_disarm_irq(b);
+		intel_breadcrumbs_disarm_irq(b);
 
-	list_for_each_entry_safe(ce, cn, &b->signalers, signal_link) {
+	rcu_read_lock();
-		GEM_BUG_ON(list_empty(&ce->signals));
+	list_for_each_entry_rcu(ce, &b->signalers, signal_link) {
+		struct i915_request *rq;
 
-		list_for_each_safe(pos, next, &ce->signals) {
+		list_for_each_entry_rcu(rq, &ce->signals, signal_link) {
-			struct i915_request *rq =
+			bool release;
-				list_entry(pos, typeof(*rq), signal_link);
 
-			GEM_BUG_ON(!check_signal_order(ce, rq));
 			if (!__request_completed(rq))
 				break;
 
+			if (!test_and_clear_bit(I915_FENCE_FLAG_SIGNAL,
+						&rq->fence.flags))
+				break;
+
 			/*
 			 * Queue for execution after dropping the signaling
 			 * spinlock as the callback chain may end up adding
 			 * more signalers to the same context or engine.
 			 */
-			clear_bit(I915_FENCE_FLAG_SIGNAL, &rq->fence.flags);
+			spin_lock(&ce->signal_lock);
+			list_del_rcu(&rq->signal_link);
+			release = remove_signaling_context(b, ce);
+			spin_unlock(&ce->signal_lock);
+
 			if (__signal_request(rq))
 				/* We own signal_node now, xfer to local list */
 				signal = slist_add(&rq->signal_node, signal);
-		}
 
-		/*
+			if (release) {
-		 * We process the list deletion in bulk, only using a list_add
-		 * (not list_move) above but keeping the status of
-		 * rq->signal_link known with the I915_FENCE_FLAG_SIGNAL bit.
-		 */
-		if (!list_is_first(pos, &ce->signals)) {
-			/* Advance the list to the first incomplete request */
-			__list_del_many(&ce->signals, pos);
-			if (&ce->signals == pos) { /* now empty */
 				add_retire(b, ce->timeline);
-				remove_signaling_context(b, ce);
+				intel_context_put(ce);
 			}
 		}
 	}
-
+	rcu_read_unlock();
-	spin_unlock(&b->irq_lock);
 
 	llist_for_each_safe(signal, sn, signal) {
 		struct i915_request *rq =
@@ -298,14 +313,15 @@ intel_breadcrumbs_create(struct intel_engine_cs *irq_engine)
 	if (!b)
 		return NULL;
 
-	spin_lock_init(&b->irq_lock);
+	b->irq_engine = irq_engine;
+
+	spin_lock_init(&b->signalers_lock);
 	INIT_LIST_HEAD(&b->signalers);
 	init_llist_head(&b->signaled_requests);
 
+	spin_lock_init(&b->irq_lock);
 	init_irq_work(&b->irq_work, signal_irq_work);
 
-	b->irq_engine = irq_engine;
-
 	return b;
 }
 
@@ -347,9 +363,9 @@ void intel_breadcrumbs_free(struct intel_breadcrumbs *b)
 	kfree(b);
 }
 
-static void insert_breadcrumb(struct i915_request *rq,
+static void insert_breadcrumb(struct i915_request *rq)
-			      struct intel_breadcrumbs *b)
 {
+	struct intel_breadcrumbs *b = READ_ONCE(rq->engine)->breadcrumbs;
 	struct intel_context *ce = rq->context;
 	struct list_head *pos;
 
@@ -371,6 +387,7 @@ static void insert_breadcrumb(struct i915_request *rq,
 	}
 
 	if (list_empty(&ce->signals)) {
+		intel_context_get(ce);
 		add_signaling_context(b, ce);
 		pos = &ce->signals;
 	} else {
@@ -396,8 +413,9 @@ static void insert_breadcrumb(struct i915_request *rq,
 				break;
 		}
 	}
-	list_add(&rq->signal_link, pos);
+	list_add_rcu(&rq->signal_link, pos);
 	GEM_BUG_ON(!check_signal_order(ce, rq));
+	GEM_BUG_ON(test_bit(DMA_FENCE_FLAG_SIGNALED_BIT, &rq->fence.flags));
 	set_bit(I915_FENCE_FLAG_SIGNAL, &rq->fence.flags);
 
 	/*
@@ -410,7 +428,7 @@ static void insert_breadcrumb(struct i915_request *rq,
 
 bool i915_request_enable_breadcrumb(struct i915_request *rq)
 {
-	struct intel_breadcrumbs *b;
+	struct intel_context *ce = rq->context;
 
 	/* Serialises with i915_request_retire() using rq->lock */
 	if (test_bit(DMA_FENCE_FLAG_SIGNALED_BIT, &rq->fence.flags))
@@ -425,67 +443,30 @@ bool i915_request_enable_breadcrumb(struct i915_request *rq)
 	if (!test_bit(I915_FENCE_FLAG_ACTIVE, &rq->fence.flags))
 		return true;
 
-	/*
+	spin_lock(&ce->signal_lock);
-	 * rq->engine is locked by rq->engine->active.lock. That however
-	 * is not known until after rq->engine has been dereferenced and
-	 * the lock acquired. Hence we acquire the lock and then validate
-	 * that rq->engine still matches the lock we hold for it.
-	 *
-	 * Here, we are using the breadcrumb lock as a proxy for the
-	 * rq->engine->active.lock, and we know that since the breadcrumb
-	 * will be serialised within i915_request_submit/i915_request_unsubmit,
-	 * the engine cannot change while active as long as we hold the
-	 * breadcrumb lock on that engine.
-	 *
-	 * From the dma_fence_enable_signaling() path, we are outside of the
-	 * request submit/unsubmit path, and so we must be more careful to
-	 * acquire the right lock.
-	 */
-	b = READ_ONCE(rq->engine)->breadcrumbs;
-	spin_lock(&b->irq_lock);
-	while (unlikely(b != READ_ONCE(rq->engine)->breadcrumbs)) {
-		spin_unlock(&b->irq_lock);
-		b = READ_ONCE(rq->engine)->breadcrumbs;
-		spin_lock(&b->irq_lock);
-	}
-
-	/*
-	 * Now that we are finally serialised with request submit/unsubmit,
-	 * [with b->irq_lock] and with i915_request_retire() [via checking
-	 * SIGNALED with rq->lock] confirm the request is indeed active. If
-	 * it is no longer active, the breadcrumb will be attached upon
-	 * i915_request_submit().
-	 */
 	if (test_bit(I915_FENCE_FLAG_ACTIVE, &rq->fence.flags))
-		insert_breadcrumb(rq, b);
+		insert_breadcrumb(rq);
-
+	spin_unlock(&ce->signal_lock);
-	spin_unlock(&b->irq_lock);
 
 	return true;
 }
 
 void i915_request_cancel_breadcrumb(struct i915_request *rq)
 {
-	struct intel_breadcrumbs *b = rq->engine->breadcrumbs;
+	struct intel_context *ce = rq->context;
+	bool release;
 
-	/*
+	if (!test_and_clear_bit(I915_FENCE_FLAG_SIGNAL, &rq->fence.flags))
-	 * We must wait for b->irq_lock so that we know the interrupt handler
+		return;
-	 * has released its reference to the intel_context and has completed
-	 * the DMA_FENCE_FLAG_SIGNALED_BIT/I915_FENCE_FLAG_SIGNAL dance (if
-	 * required).
-	 */
-	spin_lock(&b->irq_lock);
-	if (test_bit(I915_FENCE_FLAG_SIGNAL, &rq->fence.flags)) {
-		struct intel_context *ce = rq->context;
 
-		list_del(&rq->signal_link);
+	spin_lock(&ce->signal_lock);
-		if (list_empty(&ce->signals))
+	list_del_rcu(&rq->signal_link);
-			remove_signaling_context(b, ce);
+	release = remove_signaling_context(rq->engine->breadcrumbs, ce);
+	spin_unlock(&ce->signal_lock);
+	if (release)
+		intel_context_put(ce);
 
-		clear_bit(I915_FENCE_FLAG_SIGNAL, &rq->fence.flags);
+	i915_request_put(rq);
-		i915_request_put(rq);
-	}
-	spin_unlock(&b->irq_lock);
 }
 
 static void print_signals(struct intel_breadcrumbs *b, struct drm_printer *p)
@@ -495,18 +476,17 @@ static void print_signals(struct intel_breadcrumbs *b, struct drm_printer *p)
 
 	drm_printf(p, "Signals:\n");
 
-	spin_lock_irq(&b->irq_lock);
+	rcu_read_lock();
-	list_for_each_entry(ce, &b->signalers, signal_link) {
+	list_for_each_entry_rcu(ce, &b->signalers, signal_link) {
-		list_for_each_entry(rq, &ce->signals, signal_link) {
+		list_for_each_entry_rcu(rq, &ce->signals, signal_link)
 			drm_printf(p, "\t[%llx:%llx%s] @ %dms\n",
 				   rq->fence.context, rq->fence.seqno,
 				   i915_request_completed(rq) ? "!" :
 				   i915_request_started(rq) ? "*" :
 				   "",
 				   jiffies_to_msecs(jiffies - rq->emitted_jiffies));
-		}
 	}
-	spin_unlock_irq(&b->irq_lock);
+	rcu_read_unlock();
 }
 
 void intel_engine_print_breadcrumbs(struct intel_engine_cs *engine,

drivers/gpu/drm/i915/gt/intel_breadcrumbs_types.h

@@ -29,18 +29,16 @@
  * the overhead of waking that client is much preferred.
  */
 struct intel_breadcrumbs {
-	spinlock_t irq_lock; /* protects the lists used in hardirq context */
-
 	/* Not all breadcrumbs are attached to physical HW */
 	struct intel_engine_cs *irq_engine;
 
+	spinlock_t signalers_lock; /* protects the list of signalers */
 	struct list_head signalers;
 	struct llist_head signaled_requests;
 
+	spinlock_t irq_lock; /* protects the interrupt from hardirq context */
 	struct irq_work irq_work; /* for use from inside irq_lock */
-
 	unsigned int irq_enabled;
-
 	bool irq_armed;
 };
 

drivers/gpu/drm/i915/gt/intel_context.c

@@ -25,9 +25,16 @@ static struct intel_context *intel_context_alloc(void)
 	return kmem_cache_zalloc(global.slab_ce, GFP_KERNEL);
 }
 
+static void rcu_context_free(struct rcu_head *rcu)
+{
+	struct intel_context *ce = container_of(rcu, typeof(*ce), rcu);
+
+	kmem_cache_free(global.slab_ce, ce);
+}
+
 void intel_context_free(struct intel_context *ce)
 {
-	kmem_cache_free(global.slab_ce, ce);
+	call_rcu(&ce->rcu, rcu_context_free);
 }
 
 struct intel_context *
@@ -356,8 +363,7 @@ static int __intel_context_active(struct i915_active *active)
 }
 
 void
-intel_context_init(struct intel_context *ce,
+intel_context_init(struct intel_context *ce, struct intel_engine_cs *engine)
-		   struct intel_engine_cs *engine)
 {
 	GEM_BUG_ON(!engine->cops);
 	GEM_BUG_ON(!engine->gt->vm);
@@ -373,7 +379,8 @@ intel_context_init(struct intel_context *ce,
 
 	ce->vm = i915_vm_get(engine->gt->vm);
 
-	INIT_LIST_HEAD(&ce->signal_link);
+	/* NB ce->signal_link/lock is used under RCU */
+	spin_lock_init(&ce->signal_lock);
 	INIT_LIST_HEAD(&ce->signals);
 
 	mutex_init(&ce->pin_mutex);

drivers/gpu/drm/i915/gt/intel_context_types.h

@@ -25,6 +25,7 @@ DECLARE_EWMA(runtime, 3, 8);
 struct i915_gem_context;
 struct i915_gem_ww_ctx;
 struct i915_vma;
+struct intel_breadcrumbs;
 struct intel_context;
 struct intel_ring;
 
@@ -44,7 +45,16 @@ struct intel_context_ops {
 };
 
 struct intel_context {
-	struct kref ref;
+	/*
+	 * Note: Some fields may be accessed under RCU.
+	 *
+	 * Unless otherwise noted a field can safely be assumed to be protected
+	 * by strong reference counting.
+	 */
+	union {
+		struct kref ref; /* no kref_get_unless_zero()! */
+		struct rcu_head rcu;
+	};
 
 	struct intel_engine_cs *engine;
 	struct intel_engine_cs *inflight;
@@ -54,8 +64,15 @@ struct intel_context {
 	struct i915_address_space *vm;
 	struct i915_gem_context __rcu *gem_context;
 
-	struct list_head signal_link;
+	/*
-	struct list_head signals;
+	 * @signal_lock protects the list of requests that need signaling,
+	 * @signals. While there are any requests that need signaling,
+	 * we add the context to the breadcrumbs worker, and remove it
+	 * upon completion/cancellation of the last request.
+	 */
+	struct list_head signal_link; /* Accessed under RCU */
+	struct list_head signals; /* Guarded by signal_lock */
+	spinlock_t signal_lock; /* protects signals, the list of requests */
 
 	struct i915_vma *state;
 	struct intel_ring *ring;

drivers/gpu/drm/i915/gt/intel_mocs.c

@@ -131,7 +131,19 @@ static const struct drm_i915_mocs_entry skl_mocs_table[] = {
 	GEN9_MOCS_ENTRIES,
 	MOCS_ENTRY(I915_MOCS_CACHED,
 		   LE_3_WB | LE_TC_2_LLC_ELLC | LE_LRUM(3),
-		   L3_3_WB)
+		   L3_3_WB),
+
+	/*
+	 * mocs:63
+	 * - used by the L3 for all of its evictions.
+	 *   Thus it is expected to allow LLC cacheability to enable coherent
+	 *   flows to be maintained.
+	 * - used to force L3 uncachable cycles.
+	 *   Thus it is expected to make the surface L3 uncacheable.
+	 */
+	MOCS_ENTRY(63,
+		   LE_3_WB | LE_TC_1_LLC | LE_LRUM(3),
+		   L3_1_UC)
 };
 
 /* NOTE: the LE_TGT_CACHE is not used on Broxton */

drivers/gpu/drm/i915/gt/intel_rps.c

@@ -883,6 +883,10 @@ void intel_rps_park(struct intel_rps *rps)
 		adj = -2;
 	rps->last_adj = adj;
 	rps->cur_freq = max_t(int, rps->cur_freq + adj, rps->min_freq);
+	if (rps->cur_freq < rps->efficient_freq) {
+		rps->cur_freq = rps->efficient_freq;
+		rps->last_adj = 0;
+	}
 
 	GT_TRACE(rps_to_gt(rps), "park:%x\n", rps->cur_freq);
 }

drivers/gpu/drm/i915/gt/shmem_utils.c

@@ -103,10 +103,13 @@ static int __shmem_rw(struct file *file, loff_t off,
 			return PTR_ERR(page);
 
 		vaddr = kmap(page);
-		if (write)
+		if (write) {
 			memcpy(vaddr + offset_in_page(off), ptr, this);
-		else
+			set_page_dirty(page);
+		} else {
 			memcpy(ptr, vaddr + offset_in_page(off), this);
+		}
+		mark_page_accessed(page);
 		kunmap(page);
 		put_page(page);
 

drivers/gpu/drm/i915/i915_request.h

@@ -177,10 +177,8 @@ struct i915_request {
 	struct intel_ring *ring;
 	struct intel_timeline __rcu *timeline;
 
-	union {
+	struct list_head signal_link;
-		struct list_head signal_link;
+	struct llist_node signal_node;
-		struct llist_node signal_node;
-	};
 
 	/*
 	 * The rcu epoch of when this request was allocated. Used to judiciously

drivers/gpu/drm/mxsfb/mxsfb_kms.c

@@ -22,6 +22,7 @@
 #include <drm/drm_fb_cma_helper.h>
 #include <drm/drm_fourcc.h>
 #include <drm/drm_gem_cma_helper.h>
+#include <drm/drm_gem_framebuffer_helper.h>
 #include <drm/drm_plane.h>
 #include <drm/drm_plane_helper.h>
 #include <drm/drm_vblank.h>
@@ -484,17 +485,27 @@ static void mxsfb_plane_overlay_atomic_update(struct drm_plane *plane,
 	writel(ctrl, mxsfb->base + LCDC_AS_CTRL);
 }
 
+static bool mxsfb_format_mod_supported(struct drm_plane *plane,
+				       uint32_t format,
+				       uint64_t modifier)
+{
+	return modifier == DRM_FORMAT_MOD_LINEAR;
+}
+
 static const struct drm_plane_helper_funcs mxsfb_plane_primary_helper_funcs = {
+	.prepare_fb = drm_gem_fb_prepare_fb,
 	.atomic_check = mxsfb_plane_atomic_check,
 	.atomic_update = mxsfb_plane_primary_atomic_update,
 };
 
 static const struct drm_plane_helper_funcs mxsfb_plane_overlay_helper_funcs = {
+	.prepare_fb = drm_gem_fb_prepare_fb,
 	.atomic_check = mxsfb_plane_atomic_check,
 	.atomic_update = mxsfb_plane_overlay_atomic_update,
 };
 
 static const struct drm_plane_funcs mxsfb_plane_funcs = {
+	.format_mod_supported	= mxsfb_format_mod_supported,
 	.update_plane		= drm_atomic_helper_update_plane,
 	.disable_plane		= drm_atomic_helper_disable_plane,
 	.destroy		= drm_plane_cleanup,

drivers/gpu/drm/nouveau/nouveau_bo.c

@@ -1214,8 +1214,8 @@ nouveau_ttm_io_mem_reserve(struct ttm_bo_device *bdev, struct ttm_resource *reg)
 			}
 
 			reg->bus.offset = handle;
-			ret = 0;
 		}
+		ret = 0;
 		break;
 	default:
 		ret = -EINVAL;

drivers/gpu/drm/omapdrm/dss/sdi.c

@@ -195,8 +195,7 @@ static void sdi_bridge_mode_set(struct drm_bridge *bridge,
 	sdi->pixelclock = adjusted_mode->clock * 1000;
 }
 
-static void sdi_bridge_enable(struct drm_bridge *bridge,
+static void sdi_bridge_enable(struct drm_bridge *bridge)
-			      struct drm_bridge_state *bridge_state)
 {
 	struct sdi_device *sdi = drm_bridge_to_sdi(bridge);
 	struct dispc_clock_info dispc_cinfo;
@@ -259,8 +258,7 @@ static void sdi_bridge_enable(struct drm_bridge *bridge,
 	regulator_disable(sdi->vdds_sdi_reg);
 }
 
-static void sdi_bridge_disable(struct drm_bridge *bridge,
+static void sdi_bridge_disable(struct drm_bridge *bridge)
-			       struct drm_bridge_state *bridge_state)
 {
 	struct sdi_device *sdi = drm_bridge_to_sdi(bridge);
 
@@ -278,8 +276,8 @@ static const struct drm_bridge_funcs sdi_bridge_funcs = {
 	.mode_valid = sdi_bridge_mode_valid,
 	.mode_fixup = sdi_bridge_mode_fixup,
 	.mode_set = sdi_bridge_mode_set,
-	.atomic_enable = sdi_bridge_enable,
+	.enable = sdi_bridge_enable,
-	.atomic_disable = sdi_bridge_disable,
+	.disable = sdi_bridge_disable,
 };
 
 static void sdi_bridge_init(struct sdi_device *sdi)

drivers/gpu/drm/panel/panel-sony-acx565akm.c

@@ -629,7 +629,7 @@ static int acx565akm_probe(struct spi_device *spi)
 	lcd->spi = spi;
 	mutex_init(&lcd->mutex);
 
-	lcd->reset_gpio = devm_gpiod_get(&spi->dev, "reset", GPIOD_OUT_LOW);
+	lcd->reset_gpio = devm_gpiod_get(&spi->dev, "reset", GPIOD_OUT_HIGH);
 	if (IS_ERR(lcd->reset_gpio)) {
 		dev_err(&spi->dev, "failed to get reset GPIO\n");
 		return PTR_ERR(lcd->reset_gpio);

drivers/gpu/drm/rockchip/rockchip_lvds.c

@@ -544,7 +544,7 @@ static int rockchip_lvds_bind(struct device *dev, struct device *master,
 	struct device_node  *port, *endpoint;
 	int ret = 0, child_count = 0;
 	const char *name;
-	u32 endpoint_id;
+	u32 endpoint_id = 0;
 
 	lvds->drm_dev = drm_dev;
 	port = of_graph_get_port_by_id(dev->of_node, 1);

drivers/gpu/drm/tegra/drm.c

@@ -90,7 +90,7 @@ static int tegra_drm_open(struct drm_device *drm, struct drm_file *filp)
 	if (!fpriv)
 		return -ENOMEM;
 
-	idr_init(&fpriv->contexts);
+	idr_init_base(&fpriv->contexts, 1);
 	mutex_init(&fpriv->lock);
 	filp->driver_priv = fpriv;
 

drivers/gpu/drm/tegra/output.c

@@ -129,7 +129,6 @@ int tegra_output_probe(struct tegra_output *output)
 
 		if (!output->ddc) {
 			err = -EPROBE_DEFER;
-			of_node_put(ddc);
 			return err;
 		}
 	}

drivers/gpu/drm/tegra/sor.c

@@ -397,7 +397,6 @@ struct tegra_sor;
 struct tegra_sor_ops {
 	const char *name;
 	int (*probe)(struct tegra_sor *sor);
-	int (*remove)(struct tegra_sor *sor);
 	void (*audio_enable)(struct tegra_sor *sor);
 	void (*audio_disable)(struct tegra_sor *sor);
 };
@@ -2942,6 +2941,24 @@ static const struct drm_encoder_helper_funcs tegra_sor_dp_helpers = {
 	.atomic_check = tegra_sor_encoder_atomic_check,
 };
 
+static void tegra_sor_disable_regulator(void *data)
+{
+	struct regulator *reg = data;
+
+	regulator_disable(reg);
+}
+
+static int tegra_sor_enable_regulator(struct tegra_sor *sor, struct regulator *reg)
+{
+	int err;
+
+	err = regulator_enable(reg);
+	if (err)
+		return err;
+
+	return devm_add_action_or_reset(sor->dev, tegra_sor_disable_regulator, reg);
+}
+
 static int tegra_sor_hdmi_probe(struct tegra_sor *sor)
 {
 	int err;
@@ -2953,7 +2970,7 @@ static int tegra_sor_hdmi_probe(struct tegra_sor *sor)
 		return PTR_ERR(sor->avdd_io_supply);
 	}
 
-	err = regulator_enable(sor->avdd_io_supply);
+	err = tegra_sor_enable_regulator(sor, sor->avdd_io_supply);
 	if (err < 0) {
 		dev_err(sor->dev, "failed to enable AVDD I/O supply: %d\n",
 			err);
@@ -2967,7 +2984,7 @@ static int tegra_sor_hdmi_probe(struct tegra_sor *sor)
 		return PTR_ERR(sor->vdd_pll_supply);
 	}
 
-	err = regulator_enable(sor->vdd_pll_supply);
+	err = tegra_sor_enable_regulator(sor, sor->vdd_pll_supply);
 	if (err < 0) {
 		dev_err(sor->dev, "failed to enable VDD PLL supply: %d\n",
 			err);
@@ -2981,7 +2998,7 @@ static int tegra_sor_hdmi_probe(struct tegra_sor *sor)
 		return PTR_ERR(sor->hdmi_supply);
 	}
 
-	err = regulator_enable(sor->hdmi_supply);
+	err = tegra_sor_enable_regulator(sor, sor->hdmi_supply);
 	if (err < 0) {
 		dev_err(sor->dev, "failed to enable HDMI supply: %d\n", err);
 		return err;
@@ -2992,19 +3009,9 @@ static int tegra_sor_hdmi_probe(struct tegra_sor *sor)
 	return 0;
 }
 
-static int tegra_sor_hdmi_remove(struct tegra_sor *sor)
-{
-	regulator_disable(sor->hdmi_supply);
-	regulator_disable(sor->vdd_pll_supply);
-	regulator_disable(sor->avdd_io_supply);
-
-	return 0;
-}
-
 static const struct tegra_sor_ops tegra_sor_hdmi_ops = {
 	.name = "HDMI",
 	.probe = tegra_sor_hdmi_probe,
-	.remove = tegra_sor_hdmi_remove,
 	.audio_enable = tegra_sor_hdmi_audio_enable,
 	.audio_disable = tegra_sor_hdmi_audio_disable,
 };
@@ -3017,7 +3024,7 @@ static int tegra_sor_dp_probe(struct tegra_sor *sor)
 	if (IS_ERR(sor->avdd_io_supply))
 		return PTR_ERR(sor->avdd_io_supply);
 
-	err = regulator_enable(sor->avdd_io_supply);
+	err = tegra_sor_enable_regulator(sor, sor->avdd_io_supply);
 	if (err < 0)
 		return err;
 
@@ -3025,25 +3032,16 @@ static int tegra_sor_dp_probe(struct tegra_sor *sor)
 	if (IS_ERR(sor->vdd_pll_supply))
 		return PTR_ERR(sor->vdd_pll_supply);
 
-	err = regulator_enable(sor->vdd_pll_supply);
+	err = tegra_sor_enable_regulator(sor, sor->vdd_pll_supply);
 	if (err < 0)
 		return err;
 
 	return 0;
 }
 
-static int tegra_sor_dp_remove(struct tegra_sor *sor)
-{
-	regulator_disable(sor->vdd_pll_supply);
-	regulator_disable(sor->avdd_io_supply);
-
-	return 0;
-}
-
 static const struct tegra_sor_ops tegra_sor_dp_ops = {
 	.name = "DP",
 	.probe = tegra_sor_dp_probe,
-	.remove = tegra_sor_dp_remove,
 };
 
 static int tegra_sor_init(struct host1x_client *client)
@@ -3145,6 +3143,7 @@ static int tegra_sor_init(struct host1x_client *client)
 		if (err < 0) {
 			dev_err(sor->dev, "failed to deassert SOR reset: %d\n",
 				err);
+			clk_disable_unprepare(sor->clk);
 			return err;
 		}
 
@@ -3152,12 +3151,17 @@ static int tegra_sor_init(struct host1x_client *client)
 	}
 
 	err = clk_prepare_enable(sor->clk_safe);
-	if (err < 0)
+	if (err < 0) {
+		clk_disable_unprepare(sor->clk);
 		return err;
+	}
 
 	err = clk_prepare_enable(sor->clk_dp);
-	if (err < 0)
+	if (err < 0) {
+		clk_disable_unprepare(sor->clk_safe);
+		clk_disable_unprepare(sor->clk);
 		return err;
+	}
 
 	return 0;
 }
@@ -3764,17 +3768,16 @@ static int tegra_sor_probe(struct platform_device *pdev)
 		return err;
 
 	err = tegra_output_probe(&sor->output);
-	if (err < 0) {
+	if (err < 0)
-		dev_err(&pdev->dev, "failed to probe output: %d\n", err);
+		return dev_err_probe(&pdev->dev, err,
-		return err;
+				     "failed to probe output\n");
-	}
 
 	if (sor->ops && sor->ops->probe) {
 		err = sor->ops->probe(sor);
 		if (err < 0) {
 			dev_err(&pdev->dev, "failed to probe %s: %d\n",
 				sor->ops->name, err);
-			goto output;
+			goto remove;
 		}
 	}
 
@@ -3955,9 +3958,6 @@ static int tegra_sor_probe(struct platform_device *pdev)
 rpm_disable:
 	pm_runtime_disable(&pdev->dev);
 remove:
-	if (sor->ops && sor->ops->remove)
-		sor->ops->remove(sor);
-output:
 	tegra_output_remove(&sor->output);
 	return err;
 }
@@ -3976,12 +3976,6 @@ static int tegra_sor_remove(struct platform_device *pdev)
 
 	pm_runtime_disable(&pdev->dev);
 
-	if (sor->ops && sor->ops->remove) {
-		err = sor->ops->remove(sor);
-		if (err < 0)
-			dev_err(&pdev->dev, "failed to remove SOR: %d\n", err);
-	}
-
 	tegra_output_remove(&sor->output);
 
 	return 0;