Merge tag 'drm-fixes-for-v4.13-rc2' of git://people.freedesktop.org/~airlied/linux

Pull drm fixes from Dave Airlie:
 "A bunch of fixes for rc2: two imx regressions, vc4 fix, dma-buf fix,
  some displayport mst fixes, and an amdkfd fix.

  Nothing too crazy, I assume we just haven't see much rc1 testing yet"

* tag 'drm-fixes-for-v4.13-rc2' of git://people.freedesktop.org/~airlied/linux:
  drm/mst: Avoid processing partially received up/down message transactions
  drm/mst: Avoid dereferencing a NULL mstb in drm_dp_mst_handle_up_req()
  drm/mst: Fix error handling during MST sideband message reception
  drm/imx: parallel-display: Accept drm_of_find_panel_or_bridge failure
  drm/imx: fix typo in ipu_plane_formats[]
  drm/vc4: Fix VBLANK handling in crtc->enable() path
  dma-buf/fence: Avoid use of uninitialised timestamp
  drm/amdgpu: Remove unused field kgd2kfd_shared_resources.num_mec
  drm/radeon: Remove initialization of shared_resources.num_mec
  drm/amdkfd: Remove unused references to shared_resources.num_mec
  drm/amdgpu: Fix KFD oversubscription by tracking queues correctly

14 files changed, 95 insertions, 64 deletions

diff --git a/drivers/dma-buf/dma-fence.c b/drivers/dma-buf/dma-fence.c
index 57da14c15987..56e0a0e1b600 100644
--- a/drivers/dma-buf/dma-fence.c
+++ b/drivers/dma-buf/dma-fence.c
@@ -75,11 +75,6 @@ int dma_fence_signal_locked(struct dma_fence *fence)
 	if (WARN_ON(!fence))
 		return -EINVAL;
 
-	if (!ktime_to_ns(fence->timestamp)) {
-		fence->timestamp = ktime_get();
-		smp_mb__before_atomic();
-	}
-
 	if (test_and_set_bit(DMA_FENCE_FLAG_SIGNALED_BIT, &fence->flags)) {
 		ret = -EINVAL;
 
@@ -87,8 +82,11 @@ int dma_fence_signal_locked(struct dma_fence *fence)
 		 * we might have raced with the unlocked dma_fence_signal,
 		 * still run through all callbacks
 		 */
-	} else
+	} else {
+		fence->timestamp = ktime_get();
+		set_bit(DMA_FENCE_FLAG_TIMESTAMP_BIT, &fence->flags);
 		trace_dma_fence_signaled(fence);
+	}
 
 	list_for_each_entry_safe(cur, tmp, &fence->cb_list, node) {
 		list_del_init(&cur->node);
@@ -115,14 +113,11 @@ int dma_fence_signal(struct dma_fence *fence)
 	if (!fence)
 		return -EINVAL;
 
-	if (!ktime_to_ns(fence->timestamp)) {
-		fence->timestamp = ktime_get();
-		smp_mb__before_atomic();
-	}
-
 	if (test_and_set_bit(DMA_FENCE_FLAG_SIGNALED_BIT, &fence->flags))
 		return -EINVAL;
 
+	fence->timestamp = ktime_get();
+	set_bit(DMA_FENCE_FLAG_TIMESTAMP_BIT, &fence->flags);
 	trace_dma_fence_signaled(fence);
 
 	if (test_bit(DMA_FENCE_FLAG_ENABLE_SIGNAL_BIT, &fence->flags)) {
diff --git a/drivers/dma-buf/sync_debug.c b/drivers/dma-buf/sync_debug.c
index 82a6e7f6d37f..59a3b2f8ee91 100644
--- a/drivers/dma-buf/sync_debug.c
+++ b/drivers/dma-buf/sync_debug.c
@@ -84,7 +84,7 @@ static void sync_print_fence(struct seq_file *s,
 		   show ? "_" : "",
 		   sync_status_str(status));
 
-	if (status) {
+	if (test_bit(DMA_FENCE_FLAG_TIMESTAMP_BIT, &fence->flags)) {
 		struct timespec64 ts64 =
 			ktime_to_timespec64(fence->timestamp);
 
diff --git a/drivers/dma-buf/sync_file.c b/drivers/dma-buf/sync_file.c
index 545e2c5c4815..d7e219d2669d 100644
--- a/drivers/dma-buf/sync_file.c
+++ b/drivers/dma-buf/sync_file.c
@@ -391,7 +391,13 @@ static void sync_fill_fence_info(struct dma_fence *fence,
 		sizeof(info->driver_name));
 
 	info->status = dma_fence_get_status(fence);
-	info->timestamp_ns = ktime_to_ns(fence->timestamp);
+	while (test_bit(DMA_FENCE_FLAG_SIGNALED_BIT, &fence->flags) &&
+	       !test_bit(DMA_FENCE_FLAG_TIMESTAMP_BIT, &fence->flags))
+		cpu_relax();
+	info->timestamp_ns =
+		test_bit(DMA_FENCE_FLAG_TIMESTAMP_BIT, &fence->flags) ?
+		ktime_to_ns(fence->timestamp) :
+		ktime_set(0, 0);
 }
 
 static long sync_file_ioctl_fence_info(struct sync_file *sync_file,
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c
index 5f8ada1d872b..37971d9402e3 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c
@@ -101,7 +101,6 @@ void amdgpu_amdkfd_device_init(struct amdgpu_device *adev)
 	if (adev->kfd) {
 		struct kgd2kfd_shared_resources gpu_resources = {
 			.compute_vmid_bitmap = 0xFF00,
-			.num_mec = adev->gfx.mec.num_mec,
 			.num_pipe_per_mec = adev->gfx.mec.num_pipe_per_mec,
 			.num_queue_per_pipe = adev->gfx.mec.num_queue_per_pipe
 		};
@@ -122,7 +121,7 @@ void amdgpu_amdkfd_device_init(struct amdgpu_device *adev)
 
 		/* According to linux/bitmap.h we shouldn't use bitmap_clear if
 		 * nbits is not compile time constant */
-		last_valid_bit = adev->gfx.mec.num_mec
+		last_valid_bit = 1 /* only first MEC can have compute queues */
 				* adev->gfx.mec.num_pipe_per_mec
 				* adev->gfx.mec.num_queue_per_pipe;
 		for (i = last_valid_bit; i < KGD_MAX_QUEUES; ++i)
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_device.c b/drivers/gpu/drm/amd/amdkfd/kfd_device.c
index 88187bfc5ea3..3f95f7cb4019 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_device.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_device.c
@@ -226,10 +226,6 @@ bool kgd2kfd_device_init(struct kfd_dev *kfd,
 
 	kfd->shared_resources = *gpu_resources;
 
-	/* We only use the first MEC */
-	if (kfd->shared_resources.num_mec > 1)
-		kfd->shared_resources.num_mec = 1;
-
 	/* calculate max size of mqds needed for queues */
 	size = max_num_of_queues_per_device *
 			kfd->device_info->mqd_size_aligned;
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c
index 955aa304ff48..602769ced3bd 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c
@@ -77,13 +77,6 @@ static bool is_pipe_enabled(struct device_queue_manager *dqm, int mec, int pipe)
 	return false;
 }
 
-unsigned int get_mec_num(struct device_queue_manager *dqm)
-{
-	BUG_ON(!dqm || !dqm->dev);
-
-	return dqm->dev->shared_resources.num_mec;
-}
-
 unsigned int get_queues_num(struct device_queue_manager *dqm)
 {
 	BUG_ON(!dqm || !dqm->dev);
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.h b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.h
index 66b9615bc3c1..faf820a06400 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.h
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.h
@@ -180,7 +180,6 @@ void device_queue_manager_init_cik(struct device_queue_manager_asic_ops *ops);
 void device_queue_manager_init_vi(struct device_queue_manager_asic_ops *ops);
 void program_sh_mem_settings(struct device_queue_manager *dqm,
 					struct qcm_process_device *qpd);
-unsigned int get_mec_num(struct device_queue_manager *dqm);
 unsigned int get_queues_num(struct device_queue_manager *dqm);
 unsigned int get_queues_per_pipe(struct device_queue_manager *dqm);
 unsigned int get_pipes_per_mec(struct device_queue_manager *dqm);
diff --git a/drivers/gpu/drm/amd/include/kgd_kfd_interface.h b/drivers/gpu/drm/amd/include/kgd_kfd_interface.h
index 91ef1484b3bb..36f376677a53 100644
--- a/drivers/gpu/drm/amd/include/kgd_kfd_interface.h
+++ b/drivers/gpu/drm/amd/include/kgd_kfd_interface.h
@@ -63,9 +63,6 @@ struct kgd2kfd_shared_resources {
 	/* Bit n == 1 means VMID n is available for KFD. */
 	unsigned int compute_vmid_bitmap;
 
-	/* number of mec available from the hardware */
-	uint32_t num_mec;
-
 	/* number of pipes per mec */
 	uint32_t num_pipe_per_mec;
 
diff --git a/drivers/gpu/drm/drm_dp_mst_topology.c b/drivers/gpu/drm/drm_dp_mst_topology.c
index bfd237c15e76..ae5f06895562 100644
--- a/drivers/gpu/drm/drm_dp_mst_topology.c
+++ b/drivers/gpu/drm/drm_dp_mst_topology.c
@@ -330,6 +330,13 @@ static bool drm_dp_sideband_msg_build(struct drm_dp_sideband_msg_rx *msg,
 			return false;
 		}
 
+		/*
+		 * ignore out-of-order messages or messages that are part of a
+		 * failed transaction
+		 */
+		if (!recv_hdr.somt && !msg->have_somt)
+			return false;
+
 		/* get length contained in this portion */
 		msg->curchunk_len = recv_hdr.msg_len;
 		msg->curchunk_hdrlen = hdrlen;
@@ -2164,7 +2171,7 @@ out_unlock:
 }
 EXPORT_SYMBOL(drm_dp_mst_topology_mgr_resume);
 
-static void drm_dp_get_one_sb_msg(struct drm_dp_mst_topology_mgr *mgr, bool up)
+static bool drm_dp_get_one_sb_msg(struct drm_dp_mst_topology_mgr *mgr, bool up)
 {
 	int len;
 	u8 replyblock[32];
@@ -2179,12 +2186,12 @@ static void drm_dp_get_one_sb_msg(struct drm_dp_mst_topology_mgr *mgr, bool up)
 			       replyblock, len);
 	if (ret != len) {
 		DRM_DEBUG_KMS("failed to read DPCD down rep %d %d\n", len, ret);
-		return;
+		return false;
 	}
 	ret = drm_dp_sideband_msg_build(msg, replyblock, len, true);
 	if (!ret) {
 		DRM_DEBUG_KMS("sideband msg build failed %d\n", replyblock[0]);
-		return;
+		return false;
 	}
 	replylen = msg->curchunk_len + msg->curchunk_hdrlen;
 
@@ -2196,21 +2203,32 @@ static void drm_dp_get_one_sb_msg(struct drm_dp_mst_topology_mgr *mgr, bool up)
 		ret = drm_dp_dpcd_read(mgr->aux, basereg + curreply,
 				    replyblock, len);
 		if (ret != len) {
-			DRM_DEBUG_KMS("failed to read a chunk\n");
+			DRM_DEBUG_KMS("failed to read a chunk (len %d, ret %d)\n",
+				      len, ret);
+			return false;
 		}
+
 		ret = drm_dp_sideband_msg_build(msg, replyblock, len, false);
-		if (ret == false)
+		if (!ret) {
 			DRM_DEBUG_KMS("failed to build sideband msg\n");
+			return false;
+		}
+
 		curreply += len;
 		replylen -= len;
 	}
+	return true;
 }
 
 static int drm_dp_mst_handle_down_rep(struct drm_dp_mst_topology_mgr *mgr)
 {
 	int ret = 0;
 
-	drm_dp_get_one_sb_msg(mgr, false);
+	if (!drm_dp_get_one_sb_msg(mgr, false)) {
+		memset(&mgr->down_rep_recv, 0,
+		       sizeof(struct drm_dp_sideband_msg_rx));
+		return 0;
+	}
 
 	if (mgr->down_rep_recv.have_eomt) {
 		struct drm_dp_sideband_msg_tx *txmsg;
@@ -2266,7 +2284,12 @@ static int drm_dp_mst_handle_down_rep(struct drm_dp_mst_topology_mgr *mgr)
 static int drm_dp_mst_handle_up_req(struct drm_dp_mst_topology_mgr *mgr)
 {
 	int ret = 0;
-	drm_dp_get_one_sb_msg(mgr, true);
+
+	if (!drm_dp_get_one_sb_msg(mgr, true)) {
+		memset(&mgr->up_req_recv, 0,
+		       sizeof(struct drm_dp_sideband_msg_rx));
+		return 0;
+	}
 
 	if (mgr->up_req_recv.have_eomt) {
 		struct drm_dp_sideband_msg_req_body msg;
@@ -2318,7 +2341,9 @@ static int drm_dp_mst_handle_up_req(struct drm_dp_mst_topology_mgr *mgr)
 			DRM_DEBUG_KMS("Got RSN: pn: %d avail_pbn %d\n", msg.u.resource_stat.port_number, msg.u.resource_stat.available_pbn);
 		}
 
-		drm_dp_put_mst_branch_device(mstb);
+		if (mstb)
+			drm_dp_put_mst_branch_device(mstb);
+
 		memset(&mgr->up_req_recv, 0, sizeof(struct drm_dp_sideband_msg_rx));
 	}
 	return ret;
diff --git a/drivers/gpu/drm/imx/ipuv3-plane.c b/drivers/gpu/drm/imx/ipuv3-plane.c
index 49546222c6d3..6276bb834b4f 100644
--- a/drivers/gpu/drm/imx/ipuv3-plane.c
+++ b/drivers/gpu/drm/imx/ipuv3-plane.c
@@ -54,7 +54,7 @@ static const uint32_t ipu_plane_formats[] = {
 	DRM_FORMAT_RGBA8888,
 	DRM_FORMAT_RGBX8888,
 	DRM_FORMAT_BGRA8888,
-	DRM_FORMAT_BGRA8888,
+	DRM_FORMAT_BGRX8888,
 	DRM_FORMAT_UYVY,
 	DRM_FORMAT_VYUY,
 	DRM_FORMAT_YUYV,
diff --git a/drivers/gpu/drm/imx/parallel-display.c b/drivers/gpu/drm/imx/parallel-display.c
index 636031a30e17..8aca20209cb8 100644
--- a/drivers/gpu/drm/imx/parallel-display.c
+++ b/drivers/gpu/drm/imx/parallel-display.c
@@ -237,7 +237,7 @@ static int imx_pd_bind(struct device *dev, struct device *master, void *data)
 
 	/* port@1 is the output port */
 	ret = drm_of_find_panel_or_bridge(np, 1, 0, &imxpd->panel, &imxpd->bridge);
-	if (ret)
+	if (ret && ret != -ENODEV)
 		return ret;
 
 	imxpd->dev = dev;
diff --git a/drivers/gpu/drm/radeon/radeon_kfd.c b/drivers/gpu/drm/radeon/radeon_kfd.c
index 699fe7f9b8bf..a2ab6dcdf4a2 100644
--- a/drivers/gpu/drm/radeon/radeon_kfd.c
+++ b/drivers/gpu/drm/radeon/radeon_kfd.c
@@ -184,7 +184,6 @@ void radeon_kfd_device_init(struct radeon_device *rdev)
 	if (rdev->kfd) {
 		struct kgd2kfd_shared_resources gpu_resources = {
 			.compute_vmid_bitmap = 0xFF00,
-			.num_mec = 1,
 			.num_pipe_per_mec = 4,
 			.num_queue_per_pipe = 8
 		};
diff --git a/drivers/gpu/drm/vc4/vc4_crtc.c b/drivers/gpu/drm/vc4/vc4_crtc.c
index 403bbd5f99a9..a12cc7ea99b6 100644
--- a/drivers/gpu/drm/vc4/vc4_crtc.c
+++ b/drivers/gpu/drm/vc4/vc4_crtc.c
@@ -520,6 +520,34 @@ static void vc4_crtc_disable(struct drm_crtc *crtc)
 		     SCALER_DISPSTATX_EMPTY);
 }
 
+static void vc4_crtc_update_dlist(struct drm_crtc *crtc)
+{
+	struct drm_device *dev = crtc->dev;
+	struct vc4_dev *vc4 = to_vc4_dev(dev);
+	struct vc4_crtc *vc4_crtc = to_vc4_crtc(crtc);
+	struct vc4_crtc_state *vc4_state = to_vc4_crtc_state(crtc->state);
+
+	if (crtc->state->event) {
+		unsigned long flags;
+
+		crtc->state->event->pipe = drm_crtc_index(crtc);
+
+		WARN_ON(drm_crtc_vblank_get(crtc) != 0);
+
+		spin_lock_irqsave(&dev->event_lock, flags);
+		vc4_crtc->event = crtc->state->event;
+		crtc->state->event = NULL;
+
+		HVS_WRITE(SCALER_DISPLISTX(vc4_crtc->channel),
+			  vc4_state->mm.start);
+
+		spin_unlock_irqrestore(&dev->event_lock, flags);
+	} else {
+		HVS_WRITE(SCALER_DISPLISTX(vc4_crtc->channel),
+			  vc4_state->mm.start);
+	}
+}
+
 static void vc4_crtc_enable(struct drm_crtc *crtc)
 {
 	struct drm_device *dev = crtc->dev;
@@ -530,6 +558,12 @@ static void vc4_crtc_enable(struct drm_crtc *crtc)
 
 	require_hvs_enabled(dev);
 
+	/* Enable vblank irq handling before crtc is started otherwise
+	 * drm_crtc_get_vblank() fails in vc4_crtc_update_dlist().
+	 */
+	drm_crtc_vblank_on(crtc);
+	vc4_crtc_update_dlist(crtc);
+
 	/* Turn on the scaler, which will wait for vstart to start
 	 * compositing.
 	 */
@@ -541,9 +575,6 @@ static void vc4_crtc_enable(struct drm_crtc *crtc)
 	/* Turn on the pixel valve, which will emit the vstart signal. */
 	CRTC_WRITE(PV_V_CONTROL,
 		   CRTC_READ(PV_V_CONTROL) | PV_VCONTROL_VIDEN);
-
-	/* Enable vblank irq handling after crtc is started. */
-	drm_crtc_vblank_on(crtc);
 }
 
 static bool vc4_crtc_mode_fixup(struct drm_crtc *crtc,
@@ -598,7 +629,6 @@ static void vc4_crtc_atomic_flush(struct drm_crtc *crtc,
 {
 	struct drm_device *dev = crtc->dev;
 	struct vc4_dev *vc4 = to_vc4_dev(dev);
-	struct vc4_crtc *vc4_crtc = to_vc4_crtc(crtc);
 	struct vc4_crtc_state *vc4_state = to_vc4_crtc_state(crtc->state);
 	struct drm_plane *plane;
 	bool debug_dump_regs = false;
@@ -620,25 +650,15 @@ static void vc4_crtc_atomic_flush(struct drm_crtc *crtc,
 
 	WARN_ON_ONCE(dlist_next - dlist_start != vc4_state->mm.size);
 
-	if (crtc->state->event) {
-		unsigned long flags;
-
-		crtc->state->event->pipe = drm_crtc_index(crtc);
-
-		WARN_ON(drm_crtc_vblank_get(crtc) != 0);
-
-		spin_lock_irqsave(&dev->event_lock, flags);
-		vc4_crtc->event = crtc->state->event;
-		crtc->state->event = NULL;
-
-		HVS_WRITE(SCALER_DISPLISTX(vc4_crtc->channel),
-			  vc4_state->mm.start);
-
-		spin_unlock_irqrestore(&dev->event_lock, flags);
-	} else {
-		HVS_WRITE(SCALER_DISPLISTX(vc4_crtc->channel),
-			  vc4_state->mm.start);
-	}
+	/* Only update DISPLIST if the CRTC was already running and is not
+	 * being disabled.
+	 * vc4_crtc_enable() takes care of updating the dlist just after
+	 * re-enabling VBLANK interrupts and before enabling the engine.
+	 * If the CRTC is being disabled, there's no point in updating this
+	 * information.
+	 */
+	if (crtc->state->active && old_state->active)
+		vc4_crtc_update_dlist(crtc);
 
 	if (debug_dump_regs) {
 		DRM_INFO("CRTC %d HVS after:\n", drm_crtc_index(crtc));
diff --git a/include/linux/dma-fence.h b/include/linux/dma-fence.h
index a5195a7d6f77..0a186c4f3981 100644
--- a/include/linux/dma-fence.h
+++ b/include/linux/dma-fence.h
@@ -55,6 +55,7 @@ struct dma_fence_cb;
  * of the time.
  *
  * DMA_FENCE_FLAG_SIGNALED_BIT - fence is already signaled
+ * DMA_FENCE_FLAG_TIMESTAMP_BIT - timestamp recorded for fence signaling
  * DMA_FENCE_FLAG_ENABLE_SIGNAL_BIT - enable_signaling might have been called
  * DMA_FENCE_FLAG_USER_BITS - start of the unused bits, can be used by the
  * implementer of the fence for its own purposes. Can be used in different
@@ -84,6 +85,7 @@ struct dma_fence {
 
 enum dma_fence_flag_bits {
 	DMA_FENCE_FLAG_SIGNALED_BIT,
+	DMA_FENCE_FLAG_TIMESTAMP_BIT,
 	DMA_FENCE_FLAG_ENABLE_SIGNAL_BIT,
 	DMA_FENCE_FLAG_USER_BITS, /* must always be last member */
 };