[PATCH AUTOSEL 6.14 058/642] drm/amdgpu/discovery: check ip_discovery fw file available

[PATCH AUTOSEL 6.14 058/642] drm/amdgpu/discovery: check ip_discovery fw file available

[Sasha Levin]

From: Flora Cui <flora.cui@amd.com>

[ Upstream commit 017fbb6690c2245b1b4ef39b66c79d2990fe63dd ]

Signed-off-by: Flora Cui <flora.cui@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c | 31 ++++++++++---------
 1 file changed, 16 insertions(+), 15 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c
index 949d74eff2946..6a6dc15273dc7 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c
@@ -113,8 +113,7 @@
 #include "amdgpu_isp.h"
 #endif
 
-#define FIRMWARE_IP_DISCOVERY "amdgpu/ip_discovery.bin"
-MODULE_FIRMWARE(FIRMWARE_IP_DISCOVERY);
+MODULE_FIRMWARE("amdgpu/ip_discovery.bin");
 
 #define mmIP_DISCOVERY_VERSION  0x16A00
 #define mmRCC_CONFIG_MEMSIZE	0xde3
@@ -297,21 +296,13 @@ static int amdgpu_discovery_read_binary_from_mem(struct amdgpu_device *adev,
 	return ret;
 }
 
-static int amdgpu_discovery_read_binary_from_file(struct amdgpu_device *adev, uint8_t *binary)
+static int amdgpu_discovery_read_binary_from_file(struct amdgpu_device *adev,
+							uint8_t *binary,
+							const char *fw_name)
 {
 	const struct firmware *fw;
-	const char *fw_name;
 	int r;
 
-	switch (amdgpu_discovery) {
-	case 2:
-		fw_name = FIRMWARE_IP_DISCOVERY;
-		break;
-	default:
-		dev_warn(adev->dev, "amdgpu_discovery is not set properly\n");
-		return -EINVAL;
-	}
-
 	r = request_firmware(&fw, fw_name, adev->dev);
 	if (r) {
 		dev_err(adev->dev, "can't load firmware \"%s\"\n",
@@ -404,10 +395,19 @@ static int amdgpu_discovery_verify_npsinfo(struct amdgpu_device *adev,
 	return 0;
 }
 
+static const char *amdgpu_discovery_get_fw_name(struct amdgpu_device *adev)
+{
+	if (amdgpu_discovery == 2)
+		return "amdgpu/ip_discovery.bin";
+
+	return NULL;
+}
+
 static int amdgpu_discovery_init(struct amdgpu_device *adev)
 {
 	struct table_info *info;
 	struct binary_header *bhdr;
+	const char *fw_name;
 	uint16_t offset;
 	uint16_t size;
 	uint16_t checksum;
@@ -419,9 +419,10 @@ static int amdgpu_discovery_init(struct amdgpu_device *adev)
 		return -ENOMEM;
 
 	/* Read from file if it is the preferred option */
-	if (amdgpu_discovery == 2) {
+	fw_name = amdgpu_discovery_get_fw_name(adev);
+	if (fw_name != NULL) {
 		dev_info(adev->dev, "use ip discovery information from file");
-		r = amdgpu_discovery_read_binary_from_file(adev, adev->mman.discovery_bin);
+		r = amdgpu_discovery_read_binary_from_file(adev, adev->mman.discovery_bin, fw_name);
 
 		if (r) {
 			dev_err(adev->dev, "failed to read ip discovery binary from file\n");
-- 
2.39.5

[PATCH AUTOSEL 6.14 059/642] drm/amdgpu: rework how the cleaner shader is emitted v3

[Sasha Levin]

From: Christian König <christian.koenig@amd.com>

[ Upstream commit b7fbcd77bb467d09ba14cb4ec3b121dc85bb3100 ]

Instead of emitting the cleaner shader for every job which has the
enforce_isolation flag set only emit it for the first submission from
every client.

v2: add missing NULL check
v3: fix another NULL pointer deref

Signed-off-by: Christian König <christian.koenig@amd.com>
Acked-by: Srinivasan Shanmugam <srinivasan.shanmugam@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c | 27 ++++++++++++++++++++------
 1 file changed, 21 insertions(+), 6 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
index 22aa4a8f11891..f0d675c0fc69c 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
@@ -754,6 +754,7 @@ int amdgpu_vm_flush(struct amdgpu_ring *ring, struct amdgpu_job *job,
 		    bool need_pipe_sync)
 {
 	struct amdgpu_device *adev = ring->adev;
+	struct amdgpu_isolation *isolation = &adev->isolation[ring->xcp_id];
 	unsigned vmhub = ring->vm_hub;
 	struct amdgpu_vmid_mgr *id_mgr = &adev->vm_manager.id_mgr[vmhub];
 	struct amdgpu_vmid *id = &id_mgr->ids[job->vmid];
@@ -761,8 +762,9 @@ int amdgpu_vm_flush(struct amdgpu_ring *ring, struct amdgpu_job *job,
 	bool gds_switch_needed = ring->funcs->emit_gds_switch &&
 		job->gds_switch_needed;
 	bool vm_flush_needed = job->vm_needs_flush;
-	struct dma_fence *fence = NULL;
+	bool cleaner_shader_needed = false;
 	bool pasid_mapping_needed = false;
+	struct dma_fence *fence = NULL;
 	unsigned int patch;
 	int r;
 
@@ -785,8 +787,12 @@ int amdgpu_vm_flush(struct amdgpu_ring *ring, struct amdgpu_job *job,
 	pasid_mapping_needed &= adev->gmc.gmc_funcs->emit_pasid_mapping &&
 		ring->funcs->emit_wreg;
 
+	cleaner_shader_needed = adev->gfx.enable_cleaner_shader &&
+		ring->funcs->emit_cleaner_shader && job->base.s_fence &&
+		&job->base.s_fence->scheduled == isolation->spearhead;
+
 	if (!vm_flush_needed && !gds_switch_needed && !need_pipe_sync &&
-	    !(job->enforce_isolation && !job->vmid))
+	    !cleaner_shader_needed)
 		return 0;
 
 	amdgpu_ring_ib_begin(ring);
@@ -797,9 +803,7 @@ int amdgpu_vm_flush(struct amdgpu_ring *ring, struct amdgpu_job *job,
 	if (need_pipe_sync)
 		amdgpu_ring_emit_pipeline_sync(ring);
 
-	if (adev->gfx.enable_cleaner_shader &&
-	    ring->funcs->emit_cleaner_shader &&
-	    job->enforce_isolation)
+	if (cleaner_shader_needed)
 		ring->funcs->emit_cleaner_shader(ring);
 
 	if (vm_flush_needed) {
@@ -821,7 +825,7 @@ int amdgpu_vm_flush(struct amdgpu_ring *ring, struct amdgpu_job *job,
 					    job->oa_size);
 	}
 
-	if (vm_flush_needed || pasid_mapping_needed) {
+	if (vm_flush_needed || pasid_mapping_needed || cleaner_shader_needed) {
 		r = amdgpu_fence_emit(ring, &fence, NULL, 0);
 		if (r)
 			return r;
@@ -843,6 +847,17 @@ int amdgpu_vm_flush(struct amdgpu_ring *ring, struct amdgpu_job *job,
 		id->pasid_mapping = dma_fence_get(fence);
 		mutex_unlock(&id_mgr->lock);
 	}
+
+	/*
+	 * Make sure that all other submissions wait for the cleaner shader to
+	 * finish before we push them to the HW.
+	 */
+	if (cleaner_shader_needed) {
+		mutex_lock(&adev->enforce_isolation_mutex);
+		dma_fence_put(isolation->spearhead);
+		isolation->spearhead = dma_fence_get(fence);
+		mutex_unlock(&adev->enforce_isolation_mutex);
+	}
 	dma_fence_put(fence);
 
 	amdgpu_ring_patch_cond_exec(ring, patch);
-- 
2.39.5

[PATCH AUTOSEL 6.14 060/642] drm/amdgpu: rework how isolation is enforced v2

[Sasha Levin]

From: Christian König <christian.koenig@amd.com>

[ Upstream commit bd22e44ad415ac22e3a4f9a983d2a085f6cb4427 ]

Limiting the number of available VMIDs to enforce isolation causes some
issues with gang submit and applying certain HW workarounds which
require multiple VMIDs to work correctly.

So instead start to track all submissions to the relevant engines in a
per partition data structure and use the dma_fences of the submissions
to enforce isolation similar to what a VMID limit does.

v2: use ~0l for jobs without isolation to distinct it from kernel
    submissions which uses NULL for the owner. Add some warning when we
    are OOM.

Signed-off-by: Christian König <christian.koenig@amd.com>
Acked-by: Srinivasan Shanmugam <srinivasan.shanmugam@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu.h        | 13 ++-
 drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 98 +++++++++++++++++++++-
 drivers/gpu/drm/amd/amdgpu/amdgpu_ids.c    | 43 ++++------
 drivers/gpu/drm/amd/amdgpu/amdgpu_job.c    | 16 +++-
 drivers/gpu/drm/amd/amdgpu/amdgpu_sync.c   | 19 +++++
 drivers/gpu/drm/amd/amdgpu/amdgpu_sync.h   |  1 +
 6 files changed, 155 insertions(+), 35 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
index 98f0c12df12bc..9a61f5fe3245a 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
@@ -1187,9 +1187,15 @@ struct amdgpu_device {
 	bool                            debug_enable_ras_aca;
 	bool                            debug_exp_resets;
 
-	bool				enforce_isolation[MAX_XCP];
-	/* Added this mutex for cleaner shader isolation between GFX and compute processes */
+	/* Protection for the following isolation structure */
 	struct mutex                    enforce_isolation_mutex;
+	bool				enforce_isolation[MAX_XCP];
+	struct amdgpu_isolation {
+		void			*owner;
+		struct dma_fence	*spearhead;
+		struct amdgpu_sync	active;
+		struct amdgpu_sync	prev;
+	} isolation[MAX_XCP];
 
 	struct amdgpu_init_level *init_lvl;
 };
@@ -1470,6 +1476,9 @@ void amdgpu_device_pcie_port_wreg(struct amdgpu_device *adev,
 struct dma_fence *amdgpu_device_get_gang(struct amdgpu_device *adev);
 struct dma_fence *amdgpu_device_switch_gang(struct amdgpu_device *adev,
 					    struct dma_fence *gang);
+struct dma_fence *amdgpu_device_enforce_isolation(struct amdgpu_device *adev,
+						  struct amdgpu_ring *ring,
+						  struct amdgpu_job *job);
 bool amdgpu_device_has_display_hardware(struct amdgpu_device *adev);
 ssize_t amdgpu_get_soft_full_reset_mask(struct amdgpu_ring *ring);
 ssize_t amdgpu_show_reset_mask(char *buf, uint32_t supported_reset);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
index 71e8a76180ad6..e298b48488c22 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
@@ -4232,6 +4232,11 @@ int amdgpu_device_init(struct amdgpu_device *adev,
 	mutex_init(&adev->gfx.reset_sem_mutex);
 	/* Initialize the mutex for cleaner shader isolation between GFX and compute processes */
 	mutex_init(&adev->enforce_isolation_mutex);
+	for (i = 0; i < MAX_XCP; ++i) {
+		adev->isolation[i].spearhead = dma_fence_get_stub();
+		amdgpu_sync_create(&adev->isolation[i].active);
+		amdgpu_sync_create(&adev->isolation[i].prev);
+	}
 	mutex_init(&adev->gfx.kfd_sch_mutex);
 
 	amdgpu_device_init_apu_flags(adev);
@@ -4731,7 +4736,7 @@ void amdgpu_device_fini_hw(struct amdgpu_device *adev)
 
 void amdgpu_device_fini_sw(struct amdgpu_device *adev)
 {
-	int idx;
+	int i, idx;
 	bool px;
 
 	amdgpu_device_ip_fini(adev);
@@ -4739,6 +4744,11 @@ void amdgpu_device_fini_sw(struct amdgpu_device *adev)
 	amdgpu_ucode_release(&adev->firmware.gpu_info_fw);
 	adev->accel_working = false;
 	dma_fence_put(rcu_dereference_protected(adev->gang_submit, true));
+	for (i = 0; i < MAX_XCP; ++i) {
+		dma_fence_put(adev->isolation[i].spearhead);
+		amdgpu_sync_free(&adev->isolation[i].active);
+		amdgpu_sync_free(&adev->isolation[i].prev);
+	}
 
 	amdgpu_reset_fini(adev);
 
@@ -6875,6 +6885,92 @@ struct dma_fence *amdgpu_device_switch_gang(struct amdgpu_device *adev,
 	return NULL;
 }
 
+/**
+ * amdgpu_device_enforce_isolation - enforce HW isolation
+ * @adev: the amdgpu device pointer
+ * @ring: the HW ring the job is supposed to run on
+ * @job: the job which is about to be pushed to the HW ring
+ *
+ * Makes sure that only one client at a time can use the GFX block.
+ * Returns: The dependency to wait on before the job can be pushed to the HW.
+ * The function is called multiple times until NULL is returned.
+ */
+struct dma_fence *amdgpu_device_enforce_isolation(struct amdgpu_device *adev,
+						  struct amdgpu_ring *ring,
+						  struct amdgpu_job *job)
+{
+	struct amdgpu_isolation *isolation = &adev->isolation[ring->xcp_id];
+	struct drm_sched_fence *f = job->base.s_fence;
+	struct dma_fence *dep;
+	void *owner;
+	int r;
+
+	/*
+	 * For now enforce isolation only for the GFX block since we only need
+	 * the cleaner shader on those rings.
+	 */
+	if (ring->funcs->type != AMDGPU_RING_TYPE_GFX &&
+	    ring->funcs->type != AMDGPU_RING_TYPE_COMPUTE)
+		return NULL;
+
+	/*
+	 * All submissions where enforce isolation is false are handled as if
+	 * they come from a single client. Use ~0l as the owner to distinct it
+	 * from kernel submissions where the owner is NULL.
+	 */
+	owner = job->enforce_isolation ? f->owner : (void *)~0l;
+
+	mutex_lock(&adev->enforce_isolation_mutex);
+
+	/*
+	 * The "spearhead" submission is the first one which changes the
+	 * ownership to its client. We always need to wait for it to be
+	 * pushed to the HW before proceeding with anything.
+	 */
+	if (&f->scheduled != isolation->spearhead &&
+	    !dma_fence_is_signaled(isolation->spearhead)) {
+		dep = isolation->spearhead;
+		goto out_grab_ref;
+	}
+
+	if (isolation->owner != owner) {
+
+		/*
+		 * Wait for any gang to be assembled before switching to a
+		 * different owner or otherwise we could deadlock the
+		 * submissions.
+		 */
+		if (!job->gang_submit) {
+			dep = amdgpu_device_get_gang(adev);
+			if (!dma_fence_is_signaled(dep))
+				goto out_return_dep;
+			dma_fence_put(dep);
+		}
+
+		dma_fence_put(isolation->spearhead);
+		isolation->spearhead = dma_fence_get(&f->scheduled);
+		amdgpu_sync_move(&isolation->active, &isolation->prev);
+		isolation->owner = owner;
+	}
+
+	/*
+	 * Specifying the ring here helps to pipeline submissions even when
+	 * isolation is enabled. If that is not desired for testing NULL can be
+	 * used instead of the ring to enforce a CPU round trip while switching
+	 * between clients.
+	 */
+	dep = amdgpu_sync_peek_fence(&isolation->prev, ring);
+	r = amdgpu_sync_fence(&isolation->active, &f->finished, GFP_NOWAIT);
+	if (r)
+		DRM_WARN("OOM tracking isolation\n");
+
+out_grab_ref:
+	dma_fence_get(dep);
+out_return_dep:
+	mutex_unlock(&adev->enforce_isolation_mutex);
+	return dep;
+}
+
 bool amdgpu_device_has_display_hardware(struct amdgpu_device *adev)
 {
 	switch (adev->asic_type) {
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ids.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ids.c
index 8e712a11aba5d..9008b7388e897 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ids.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ids.c
@@ -287,40 +287,27 @@ static int amdgpu_vmid_grab_reserved(struct amdgpu_vm *vm,
 	    (*id)->flushed_updates < updates ||
 	    !(*id)->last_flush ||
 	    ((*id)->last_flush->context != fence_context &&
-	     !dma_fence_is_signaled((*id)->last_flush))) {
+	     !dma_fence_is_signaled((*id)->last_flush)))
+		needs_flush = true;
+
+	if ((*id)->owner != vm->immediate.fence_context ||
+	    (!adev->vm_manager.concurrent_flush && needs_flush)) {
 		struct dma_fence *tmp;
 
-		/* Wait for the gang to be assembled before using a
-		 * reserved VMID or otherwise the gang could deadlock.
+		/* Don't use per engine and per process VMID at the
+		 * same time
 		 */
-		tmp = amdgpu_device_get_gang(adev);
-		if (!dma_fence_is_signaled(tmp) && tmp != job->gang_submit) {
+		if (adev->vm_manager.concurrent_flush)
+			ring = NULL;
+
+		/* to prevent one context starved by another context */
+		(*id)->pd_gpu_addr = 0;
+		tmp = amdgpu_sync_peek_fence(&(*id)->active, ring);
+		if (tmp) {
 			*id = NULL;
-			*fence = tmp;
+			*fence = dma_fence_get(tmp);
 			return 0;
 		}
-		dma_fence_put(tmp);
-
-		/* Make sure the id is owned by the gang before proceeding */
-		if (!job->gang_submit ||
-		    (*id)->owner != vm->immediate.fence_context) {
-
-			/* Don't use per engine and per process VMID at the
-			 * same time
-			 */
-			if (adev->vm_manager.concurrent_flush)
-				ring = NULL;
-
-			/* to prevent one context starved by another context */
-			(*id)->pd_gpu_addr = 0;
-			tmp = amdgpu_sync_peek_fence(&(*id)->active, ring);
-			if (tmp) {
-				*id = NULL;
-				*fence = dma_fence_get(tmp);
-				return 0;
-			}
-		}
-		needs_flush = true;
 	}
 
 	/* Good we can use this VMID. Remember this submission as
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c
index 100f044759435..685c61a05af85 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c
@@ -342,17 +342,24 @@ amdgpu_job_prepare_job(struct drm_sched_job *sched_job,
 {
 	struct amdgpu_ring *ring = to_amdgpu_ring(s_entity->rq->sched);
 	struct amdgpu_job *job = to_amdgpu_job(sched_job);
-	struct dma_fence *fence = NULL;
+	struct dma_fence *fence;
 	int r;
 
 	r = drm_sched_entity_error(s_entity);
 	if (r)
 		goto error;
 
-	if (job->gang_submit)
+	if (job->gang_submit) {
 		fence = amdgpu_device_switch_gang(ring->adev, job->gang_submit);
+		if (fence)
+			return fence;
+	}
+
+	fence = amdgpu_device_enforce_isolation(ring->adev, ring, job);
+	if (fence)
+		return fence;
 
-	if (!fence && job->vm && !job->vmid) {
+	if (job->vm && !job->vmid) {
 		r = amdgpu_vmid_grab(job->vm, ring, job, &fence);
 		if (r) {
 			dev_err(ring->adev->dev, "Error getting VM ID (%d)\n", r);
@@ -365,9 +372,10 @@ amdgpu_job_prepare_job(struct drm_sched_job *sched_job,
 		 */
 		if (!fence)
 			job->vm = NULL;
+		return fence;
 	}
 
-	return fence;
+	return NULL;
 
 error:
 	dma_fence_set_error(&job->base.s_fence->finished, r);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_sync.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_sync.c
index c586ab4c911bf..d75715b3f1870 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_sync.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_sync.c
@@ -399,6 +399,25 @@ int amdgpu_sync_clone(struct amdgpu_sync *source, struct amdgpu_sync *clone)
 	return 0;
 }
 
+/**
+ * amdgpu_sync_move - move all fences from src to dst
+ *
+ * @src: source of the fences, empty after function
+ * @dst: destination for the fences
+ *
+ * Moves all fences from source to destination. All fences in destination are
+ * freed and source is empty after the function call.
+ */
+void amdgpu_sync_move(struct amdgpu_sync *src, struct amdgpu_sync *dst)
+{
+	unsigned int i;
+
+	amdgpu_sync_free(dst);
+
+	for (i = 0; i < HASH_SIZE(src->fences); ++i)
+		hlist_move_list(&src->fences[i], &dst->fences[i]);
+}
+
 /**
  * amdgpu_sync_push_to_job - push fences into job
  * @sync: sync object to get the fences from
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_sync.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_sync.h
index e3272dce798d7..a91a8eaf808b1 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_sync.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_sync.h
@@ -56,6 +56,7 @@ struct dma_fence *amdgpu_sync_peek_fence(struct amdgpu_sync *sync,
 				     struct amdgpu_ring *ring);
 struct dma_fence *amdgpu_sync_get_fence(struct amdgpu_sync *sync);
 int amdgpu_sync_clone(struct amdgpu_sync *source, struct amdgpu_sync *clone);
+void amdgpu_sync_move(struct amdgpu_sync *src, struct amdgpu_sync *dst);
 int amdgpu_sync_push_to_job(struct amdgpu_sync *sync, struct amdgpu_job *job);
 int amdgpu_sync_wait(struct amdgpu_sync *sync, bool intr);
 void amdgpu_sync_free(struct amdgpu_sync *sync);
-- 
2.39.5

[PATCH AUTOSEL 6.14 061/642] drm/amdgpu: use GFP_NOWAIT for memory allocations

[Sasha Levin]

From: Christian König <christian.koenig@amd.com>

[ Upstream commit 16590745b571c07869ef8958e0bbe44ab6f08d1f ]

In the critical submission path memory allocations can't wait for
reclaim since that can potentially wait for submissions to finish.

Finally clean that up and mark most memory allocations in the critical
path with GFP_NOWAIT. The only exception left is the dma_fence_array()
used when no VMID is available, but that will be cleaned up later on.

Signed-off-by: Christian König <christian.koenig@amd.com>
Acked-by: Srinivasan Shanmugam <srinivasan.shanmugam@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 .../gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c   |  8 ++++----
 drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c         | 18 +++++++++++-------
 drivers/gpu/drm/amd/amdgpu/amdgpu_ids.c        | 11 +++++++----
 drivers/gpu/drm/amd/amdgpu/amdgpu_mes.c        |  4 ++--
 drivers/gpu/drm/amd/amdgpu/amdgpu_sync.c       | 11 ++++++-----
 drivers/gpu/drm/amd/amdgpu/amdgpu_sync.h       |  3 ++-
 6 files changed, 32 insertions(+), 23 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c
index 1e998f972c308..70224b9f54f2f 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c
@@ -499,7 +499,7 @@ static int vm_update_pds(struct amdgpu_vm *vm, struct amdgpu_sync *sync)
 	if (ret)
 		return ret;
 
-	return amdgpu_sync_fence(sync, vm->last_update);
+	return amdgpu_sync_fence(sync, vm->last_update, GFP_KERNEL);
 }
 
 static uint64_t get_pte_flags(struct amdgpu_device *adev, struct kgd_mem *mem)
@@ -1263,7 +1263,7 @@ static int unmap_bo_from_gpuvm(struct kgd_mem *mem,
 
 	(void)amdgpu_vm_clear_freed(adev, vm, &bo_va->last_pt_update);
 
-	(void)amdgpu_sync_fence(sync, bo_va->last_pt_update);
+	(void)amdgpu_sync_fence(sync, bo_va->last_pt_update, GFP_KERNEL);
 
 	return 0;
 }
@@ -1287,7 +1287,7 @@ static int update_gpuvm_pte(struct kgd_mem *mem,
 		return ret;
 	}
 
-	return amdgpu_sync_fence(sync, bo_va->last_pt_update);
+	return amdgpu_sync_fence(sync, bo_va->last_pt_update, GFP_KERNEL);
 }
 
 static int map_bo_to_gpuvm(struct kgd_mem *mem,
@@ -2969,7 +2969,7 @@ int amdgpu_amdkfd_gpuvm_restore_process_bos(void *info, struct dma_fence __rcu *
 		}
 		dma_resv_for_each_fence(&cursor, bo->tbo.base.resv,
 					DMA_RESV_USAGE_KERNEL, fence) {
-			ret = amdgpu_sync_fence(&sync_obj, fence);
+			ret = amdgpu_sync_fence(&sync_obj, fence, GFP_KERNEL);
 			if (ret) {
 				pr_debug("Memory eviction: Sync BO fence failed. Try again\n");
 				goto validate_map_fail;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
index 5cc5f59e30184..4a5b406601fa2 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
@@ -428,7 +428,7 @@ static int amdgpu_cs_p2_dependencies(struct amdgpu_cs_parser *p,
 			dma_fence_put(old);
 		}
 
-		r = amdgpu_sync_fence(&p->sync, fence);
+		r = amdgpu_sync_fence(&p->sync, fence, GFP_KERNEL);
 		dma_fence_put(fence);
 		if (r)
 			return r;
@@ -450,7 +450,7 @@ static int amdgpu_syncobj_lookup_and_add(struct amdgpu_cs_parser *p,
 		return r;
 	}
 
-	r = amdgpu_sync_fence(&p->sync, fence);
+	r = amdgpu_sync_fence(&p->sync, fence, GFP_KERNEL);
 	dma_fence_put(fence);
 	return r;
 }
@@ -1124,7 +1124,8 @@ static int amdgpu_cs_vm_handling(struct amdgpu_cs_parser *p)
 	if (r)
 		return r;
 
-	r = amdgpu_sync_fence(&p->sync, fpriv->prt_va->last_pt_update);
+	r = amdgpu_sync_fence(&p->sync, fpriv->prt_va->last_pt_update,
+			      GFP_KERNEL);
 	if (r)
 		return r;
 
@@ -1135,7 +1136,8 @@ static int amdgpu_cs_vm_handling(struct amdgpu_cs_parser *p)
 		if (r)
 			return r;
 
-		r = amdgpu_sync_fence(&p->sync, bo_va->last_pt_update);
+		r = amdgpu_sync_fence(&p->sync, bo_va->last_pt_update,
+				      GFP_KERNEL);
 		if (r)
 			return r;
 	}
@@ -1154,7 +1156,8 @@ static int amdgpu_cs_vm_handling(struct amdgpu_cs_parser *p)
 		if (r)
 			return r;
 
-		r = amdgpu_sync_fence(&p->sync, bo_va->last_pt_update);
+		r = amdgpu_sync_fence(&p->sync, bo_va->last_pt_update,
+				      GFP_KERNEL);
 		if (r)
 			return r;
 	}
@@ -1167,7 +1170,7 @@ static int amdgpu_cs_vm_handling(struct amdgpu_cs_parser *p)
 	if (r)
 		return r;
 
-	r = amdgpu_sync_fence(&p->sync, vm->last_update);
+	r = amdgpu_sync_fence(&p->sync, vm->last_update, GFP_KERNEL);
 	if (r)
 		return r;
 
@@ -1248,7 +1251,8 @@ static int amdgpu_cs_sync_rings(struct amdgpu_cs_parser *p)
 			continue;
 		}
 
-		r = amdgpu_sync_fence(&p->gang_leader->explicit_sync, fence);
+		r = amdgpu_sync_fence(&p->gang_leader->explicit_sync, fence,
+				      GFP_KERNEL);
 		dma_fence_put(fence);
 		if (r)
 			return r;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ids.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ids.c
index 9008b7388e897..92ab821afc06a 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ids.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ids.c
@@ -209,7 +209,7 @@ static int amdgpu_vmid_grab_idle(struct amdgpu_ring *ring,
 		return 0;
 	}
 
-	fences = kmalloc_array(id_mgr->num_ids, sizeof(void *), GFP_KERNEL);
+	fences = kmalloc_array(id_mgr->num_ids, sizeof(void *), GFP_NOWAIT);
 	if (!fences)
 		return -ENOMEM;
 
@@ -313,7 +313,8 @@ static int amdgpu_vmid_grab_reserved(struct amdgpu_vm *vm,
 	/* Good we can use this VMID. Remember this submission as
 	* user of the VMID.
 	*/
-	r = amdgpu_sync_fence(&(*id)->active, &job->base.s_fence->finished);
+	r = amdgpu_sync_fence(&(*id)->active, &job->base.s_fence->finished,
+			      GFP_NOWAIT);
 	if (r)
 		return r;
 
@@ -372,7 +373,8 @@ static int amdgpu_vmid_grab_used(struct amdgpu_vm *vm,
 		 * user of the VMID.
 		 */
 		r = amdgpu_sync_fence(&(*id)->active,
-				      &job->base.s_fence->finished);
+				      &job->base.s_fence->finished,
+				      GFP_NOWAIT);
 		if (r)
 			return r;
 
@@ -424,7 +426,8 @@ int amdgpu_vmid_grab(struct amdgpu_vm *vm, struct amdgpu_ring *ring,
 
 			/* Remember this submission as user of the VMID */
 			r = amdgpu_sync_fence(&id->active,
-					      &job->base.s_fence->finished);
+					      &job->base.s_fence->finished,
+					      GFP_NOWAIT);
 			if (r)
 				goto error;
 
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.c
index 6fa20980a0b15..e4251d0691c9c 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.c
@@ -1335,14 +1335,14 @@ int amdgpu_mes_ctx_map_meta_data(struct amdgpu_device *adev,
 		DRM_ERROR("failed to do vm_bo_update on meta data\n");
 		goto error_del_bo_va;
 	}
-	amdgpu_sync_fence(&sync, bo_va->last_pt_update);
+	amdgpu_sync_fence(&sync, bo_va->last_pt_update, GFP_KERNEL);
 
 	r = amdgpu_vm_update_pdes(adev, vm, false);
 	if (r) {
 		DRM_ERROR("failed to update pdes on meta data\n");
 		goto error_del_bo_va;
 	}
-	amdgpu_sync_fence(&sync, vm->last_update);
+	amdgpu_sync_fence(&sync, vm->last_update, GFP_KERNEL);
 
 	amdgpu_sync_wait(&sync, false);
 	drm_exec_fini(&exec);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_sync.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_sync.c
index d75715b3f1870..34fc742fda91d 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_sync.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_sync.c
@@ -152,7 +152,8 @@ static bool amdgpu_sync_add_later(struct amdgpu_sync *sync, struct dma_fence *f)
  *
  * Add the fence to the sync object.
  */
-int amdgpu_sync_fence(struct amdgpu_sync *sync, struct dma_fence *f)
+int amdgpu_sync_fence(struct amdgpu_sync *sync, struct dma_fence *f,
+		      gfp_t flags)
 {
 	struct amdgpu_sync_entry *e;
 
@@ -162,7 +163,7 @@ int amdgpu_sync_fence(struct amdgpu_sync *sync, struct dma_fence *f)
 	if (amdgpu_sync_add_later(sync, f))
 		return 0;
 
-	e = kmem_cache_alloc(amdgpu_sync_slab, GFP_KERNEL);
+	e = kmem_cache_alloc(amdgpu_sync_slab, flags);
 	if (!e)
 		return -ENOMEM;
 
@@ -249,7 +250,7 @@ int amdgpu_sync_resv(struct amdgpu_device *adev, struct amdgpu_sync *sync,
 			struct dma_fence *tmp = dma_fence_chain_contained(f);
 
 			if (amdgpu_sync_test_fence(adev, mode, owner, tmp)) {
-				r = amdgpu_sync_fence(sync, f);
+				r = amdgpu_sync_fence(sync, f, GFP_KERNEL);
 				dma_fence_put(f);
 				if (r)
 					return r;
@@ -281,7 +282,7 @@ int amdgpu_sync_kfd(struct amdgpu_sync *sync, struct dma_resv *resv)
 		if (fence_owner != AMDGPU_FENCE_OWNER_KFD)
 			continue;
 
-		r = amdgpu_sync_fence(sync, f);
+		r = amdgpu_sync_fence(sync, f, GFP_KERNEL);
 		if (r)
 			break;
 	}
@@ -388,7 +389,7 @@ int amdgpu_sync_clone(struct amdgpu_sync *source, struct amdgpu_sync *clone)
 	hash_for_each_safe(source->fences, i, tmp, e, node) {
 		f = e->fence;
 		if (!dma_fence_is_signaled(f)) {
-			r = amdgpu_sync_fence(clone, f);
+			r = amdgpu_sync_fence(clone, f, GFP_KERNEL);
 			if (r)
 				return r;
 		} else {
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_sync.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_sync.h
index a91a8eaf808b1..51eb4382c91eb 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_sync.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_sync.h
@@ -47,7 +47,8 @@ struct amdgpu_sync {
 };
 
 void amdgpu_sync_create(struct amdgpu_sync *sync);
-int amdgpu_sync_fence(struct amdgpu_sync *sync, struct dma_fence *f);
+int amdgpu_sync_fence(struct amdgpu_sync *sync, struct dma_fence *f,
+		      gfp_t flags);
 int amdgpu_sync_resv(struct amdgpu_device *adev, struct amdgpu_sync *sync,
 		     struct dma_resv *resv, enum amdgpu_sync_mode mode,
 		     void *owner);
-- 
2.39.5

[PATCH AUTOSEL 6.14 062/642] drm/amdkfd: set precise mem ops caps to disabled for gfx 11 and 12

[Sasha Levin]

From: Jonathan Kim <jonathan.kim@amd.com>

[ Upstream commit f82d27dcff939d3cbecbc60e1b71e2518c37e81d ]

Clause instructions with precise memory enabled currently hang the
shader so set capabilities flag to disabled since it's unsafe to use
for debugging.

Signed-off-by: Jonathan Kim <jonathan.kim@amd.com>
Tested-by: Lancelot Six <lancelot.six@amd.com>
Reviewed-by: Harish Kasiviswanathan <harish.kasiviswanathan@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/gpu/drm/amd/amdkfd/kfd_topology.c | 4 ----
 1 file changed, 4 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_topology.c b/drivers/gpu/drm/amd/amdkfd/kfd_topology.c
index 62a9a9ccf9bb6..334c576a75b14 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_topology.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_topology.c
@@ -2006,10 +2006,6 @@ static void kfd_topology_set_capabilities(struct kfd_topology_device *dev)
 		dev->node_props.debug_prop |= HSA_DBG_WATCH_ADDR_MASK_LO_BIT_GFX10 |
 					HSA_DBG_WATCH_ADDR_MASK_HI_BIT;
 
-		if (KFD_GC_VERSION(dev->gpu) >= IP_VERSION(11, 0, 0))
-			dev->node_props.capability |=
-				HSA_CAP_TRAP_DEBUG_PRECISE_MEMORY_OPERATIONS_SUPPORTED;
-
 		if (KFD_GC_VERSION(dev->gpu) >= IP_VERSION(12, 0, 0))
 			dev->node_props.capability |=
 				HSA_CAP_TRAP_DEBUG_PRECISE_ALU_OPERATIONS_SUPPORTED;
-- 
2.39.5

[PATCH AUTOSEL 6.14 070/642] drm/amdgpu: Skip pcie_replay_count sysfs creation for VF

[Sasha Levin]

From: Victor Skvortsov <victor.skvortsov@amd.com>

[ Upstream commit 9c05636ca72a2dbf41bf0900380f438a0de47319 ]

VFs cannot read the NAK_COUNTER register. This information is only
available through PMFW metrics.

Signed-off-by: Victor Skvortsov <victor.skvortsov@amd.com>
Reviewed-by: Lijo Lazar <lijo.lazar@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 27 ++++++++++++++++------
 1 file changed, 20 insertions(+), 7 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
index e298b48488c22..6992ab4878360 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
@@ -224,6 +224,24 @@ static ssize_t amdgpu_device_get_pcie_replay_count(struct device *dev,
 static DEVICE_ATTR(pcie_replay_count, 0444,
 		amdgpu_device_get_pcie_replay_count, NULL);
 
+static int amdgpu_device_attr_sysfs_init(struct amdgpu_device *adev)
+{
+	int ret = 0;
+
+	if (!amdgpu_sriov_vf(adev))
+		ret = sysfs_create_file(&adev->dev->kobj,
+					&dev_attr_pcie_replay_count.attr);
+
+	return ret;
+}
+
+static void amdgpu_device_attr_sysfs_fini(struct amdgpu_device *adev)
+{
+	if (!amdgpu_sriov_vf(adev))
+		sysfs_remove_file(&adev->dev->kobj,
+				  &dev_attr_pcie_replay_count.attr);
+}
+
 static ssize_t amdgpu_sysfs_reg_state_get(struct file *f, struct kobject *kobj,
 					  struct bin_attribute *attr, char *buf,
 					  loff_t ppos, size_t count)
@@ -4123,11 +4141,6 @@ static bool amdgpu_device_check_iommu_remap(struct amdgpu_device *adev)
 }
 #endif
 
-static const struct attribute *amdgpu_dev_attributes[] = {
-	&dev_attr_pcie_replay_count.attr,
-	NULL
-};
-
 static void amdgpu_device_set_mcbp(struct amdgpu_device *adev)
 {
 	if (amdgpu_mcbp == 1)
@@ -4572,7 +4585,7 @@ int amdgpu_device_init(struct amdgpu_device *adev,
 	} else
 		adev->ucode_sysfs_en = true;
 
-	r = sysfs_create_files(&adev->dev->kobj, amdgpu_dev_attributes);
+	r = amdgpu_device_attr_sysfs_init(adev);
 	if (r)
 		dev_err(adev->dev, "Could not create amdgpu device attr\n");
 
@@ -4709,7 +4722,7 @@ void amdgpu_device_fini_hw(struct amdgpu_device *adev)
 		amdgpu_pm_sysfs_fini(adev);
 	if (adev->ucode_sysfs_en)
 		amdgpu_ucode_sysfs_fini(adev);
-	sysfs_remove_files(&adev->dev->kobj, amdgpu_dev_attributes);
+	amdgpu_device_attr_sysfs_fini(adev);
 	amdgpu_fru_sysfs_fini(adev);
 
 	amdgpu_reg_state_sysfs_fini(adev);
-- 
2.39.5

[PATCH AUTOSEL 6.14 104/642] drm/amdgpu: release xcp_mgr on exit

[Sasha Levin]

From: Flora Cui <flora.cui@amd.com>

[ Upstream commit b5aaa82e2b12feaaa6958f7fa0917ddcc03c24ee ]

Free on driver cleanup.

Reviewed-by: Lijo Lazar <lijo.lazar@amd.com>
Signed-off-by: Flora Cui <flora.cui@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
index 6992ab4878360..b3e4201a405c8 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
@@ -4778,6 +4778,9 @@ void amdgpu_device_fini_sw(struct amdgpu_device *adev)
 	kfree(adev->fru_info);
 	adev->fru_info = NULL;
 
+	kfree(adev->xcp_mgr);
+	adev->xcp_mgr = NULL;
+
 	px = amdgpu_device_supports_px(adev_to_drm(adev));
 
 	if (px || (!dev_is_removable(&adev->pdev->dev) &&
-- 
2.39.5

[PATCH AUTOSEL 6.14 105/642] drm/amd/display: Guard against setting dispclk low for dcn31x

[Sasha Levin]

From: Jing Zhou <Jing.Zhou@amd.com>

[ Upstream commit 9c2f4ae64bb6f6d83a54d88b9ee0f369cdbb9fa8 ]

[WHY]
We should never apply a minimum dispclk value while in
prepare_bandwidth or while displays are active. This is
always an optimizaiton for when all displays are disabled.

[HOW]
Defer dispclk optimization until safe_to_lower = true
and display_count reaches 0.

Since 0 has a special value in this logic (ie. no dispclk
required) we also need adjust the logic that clamps it for
the actual request to PMFW.

Reviewed-by: Charlene Liu <charlene.liu@amd.com>
Reviewed-by: Chris Park <chris.park@amd.com>
Reviewed-by: Eric Yang <eric.yang@amd.com>
Signed-off-by: Jing Zhou <Jing.Zhou@amd.com>
Signed-off-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Signed-off-by: Alex Hung <alex.hung@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 .../dc/clk_mgr/dcn315/dcn315_clk_mgr.c        | 20 +++++++++++--------
 .../dc/clk_mgr/dcn316/dcn316_clk_mgr.c        | 13 +++++++++---
 2 files changed, 22 insertions(+), 11 deletions(-)

diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn315/dcn315_clk_mgr.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn315/dcn315_clk_mgr.c
index a0fb4481d2f1b..827b24b3442ad 100644
--- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn315/dcn315_clk_mgr.c
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn315/dcn315_clk_mgr.c
@@ -130,7 +130,7 @@ static void dcn315_update_clocks(struct clk_mgr *clk_mgr_base,
 	struct clk_mgr_internal *clk_mgr = TO_CLK_MGR_INTERNAL(clk_mgr_base);
 	struct dc_clocks *new_clocks = &context->bw_ctx.bw.dcn.clk;
 	struct dc *dc = clk_mgr_base->ctx->dc;
-	int display_count;
+	int display_count = 0;
 	bool update_dppclk = false;
 	bool update_dispclk = false;
 	bool dpp_clock_lowered = false;
@@ -204,15 +204,19 @@ static void dcn315_update_clocks(struct clk_mgr *clk_mgr_base,
 		update_dppclk = true;
 	}
 
-	if (should_set_clock(safe_to_lower, new_clocks->dispclk_khz, clk_mgr_base->clks.dispclk_khz)) {
-		/* No need to apply the w/a if we haven't taken over from bios yet */
-		if (clk_mgr_base->clks.dispclk_khz)
-			dcn315_disable_otg_wa(clk_mgr_base, context, true);
+	if (should_set_clock(safe_to_lower, new_clocks->dispclk_khz, clk_mgr_base->clks.dispclk_khz) &&
+	    (new_clocks->dispclk_khz > 0 || (safe_to_lower && display_count == 0))) {
+		int requested_dispclk_khz = new_clocks->dispclk_khz;
 
+		dcn315_disable_otg_wa(clk_mgr_base, context, true);
+
+		/* Clamp the requested clock to PMFW based on their limit. */
+		if (dc->debug.min_disp_clk_khz > 0 && requested_dispclk_khz < dc->debug.min_disp_clk_khz)
+			requested_dispclk_khz = dc->debug.min_disp_clk_khz;
+
+		dcn315_smu_set_dispclk(clk_mgr, requested_dispclk_khz);
 		clk_mgr_base->clks.dispclk_khz = new_clocks->dispclk_khz;
-		dcn315_smu_set_dispclk(clk_mgr, clk_mgr_base->clks.dispclk_khz);
-		if (clk_mgr_base->clks.dispclk_khz)
-			dcn315_disable_otg_wa(clk_mgr_base, context, false);
+		dcn315_disable_otg_wa(clk_mgr_base, context, false);
 
 		update_dispclk = true;
 	}
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn316/dcn316_clk_mgr.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn316/dcn316_clk_mgr.c
index c3e50c3aaa609..37c39756fece4 100644
--- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn316/dcn316_clk_mgr.c
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn316/dcn316_clk_mgr.c
@@ -140,7 +140,7 @@ static void dcn316_update_clocks(struct clk_mgr *clk_mgr_base,
 	struct clk_mgr_internal *clk_mgr = TO_CLK_MGR_INTERNAL(clk_mgr_base);
 	struct dc_clocks *new_clocks = &context->bw_ctx.bw.dcn.clk;
 	struct dc *dc = clk_mgr_base->ctx->dc;
-	int display_count;
+	int display_count = 0;
 	bool update_dppclk = false;
 	bool update_dispclk = false;
 	bool dpp_clock_lowered = false;
@@ -211,11 +211,18 @@ static void dcn316_update_clocks(struct clk_mgr *clk_mgr_base,
 		update_dppclk = true;
 	}
 
-	if (should_set_clock(safe_to_lower, new_clocks->dispclk_khz, clk_mgr_base->clks.dispclk_khz)) {
+	if (should_set_clock(safe_to_lower, new_clocks->dispclk_khz, clk_mgr_base->clks.dispclk_khz) &&
+	    (new_clocks->dispclk_khz > 0 || (safe_to_lower && display_count == 0))) {
+		int requested_dispclk_khz = new_clocks->dispclk_khz;
+
 		dcn316_disable_otg_wa(clk_mgr_base, context, safe_to_lower, true);
 
+		/* Clamp the requested clock to PMFW based on their limit. */
+		if (dc->debug.min_disp_clk_khz > 0 && requested_dispclk_khz < dc->debug.min_disp_clk_khz)
+			requested_dispclk_khz = dc->debug.min_disp_clk_khz;
+
+		dcn316_smu_set_dispclk(clk_mgr, requested_dispclk_khz);
 		clk_mgr_base->clks.dispclk_khz = new_clocks->dispclk_khz;
-		dcn316_smu_set_dispclk(clk_mgr, clk_mgr_base->clks.dispclk_khz);
 		dcn316_disable_otg_wa(clk_mgr_base, context, safe_to_lower, false);
 
 		update_dispclk = true;
-- 
2.39.5

[PATCH AUTOSEL 6.14 106/642] drm/amdgpu: don't free conflicting apertures for non-display devices

[Sasha Levin]

From: Alex Deucher <alexander.deucher@amd.com>

[ Upstream commit 9deacd6c55f1b31e5ab20db79df2e14ac480203c ]

PCI_CLASS_ACCELERATOR_PROCESSING devices won't ever be
the sysfb, so there is no need to free conflicting
apertures.

Reviewed-by: Kent Russell <kent.russell@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 15 +++++++++++----
 1 file changed, 11 insertions(+), 4 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
index b3e4201a405c8..9cde62b178c3b 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
@@ -4370,10 +4370,17 @@ int amdgpu_device_init(struct amdgpu_device *adev,
 	if (r)
 		return r;
 
-	/* Get rid of things like offb */
-	r = aperture_remove_conflicting_pci_devices(adev->pdev, amdgpu_kms_driver.name);
-	if (r)
-		return r;
+	/*
+	 * No need to remove conflicting FBs for non-display class devices.
+	 * This prevents the sysfb from being freed accidently.
+	 */
+	if ((pdev->class >> 8) == PCI_CLASS_DISPLAY_VGA ||
+	    (pdev->class >> 8) == PCI_CLASS_DISPLAY_OTHER) {
+		/* Get rid of things like offb */
+		r = aperture_remove_conflicting_pci_devices(adev->pdev, amdgpu_kms_driver.name);
+		if (r)
+			return r;
+	}
 
 	/* Enable TMZ based on IP_VERSION */
 	amdgpu_gmc_tmz_set(adev);
-- 
2.39.5

[PATCH AUTOSEL 6.14 107/642] drm/amdgpu: adjust drm_firmware_drivers_only() handling

[Sasha Levin]

From: Alex Deucher <alexander.deucher@amd.com>

[ Upstream commit e00e5c223878a60e391e5422d173c3382d378f87 ]

Move to probe so we can check the PCI device type and
only apply the drm_firmware_drivers_only() check for
PCI DISPLAY classes.  Also add a module parameter to
override the nomodeset kernel parameter as a workaround
for platforms that have this hardcoded on their kernel
command lines.

Reviewed-by: Kent Russell <kent.russell@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c | 14 ++++++++++++++
 1 file changed, 14 insertions(+)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
index f2d77bc04e4a9..7246c54bd2bbf 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
@@ -173,6 +173,7 @@ uint amdgpu_sdma_phase_quantum = 32;
 char *amdgpu_disable_cu;
 char *amdgpu_virtual_display;
 bool enforce_isolation;
+int amdgpu_modeset = -1;
 
 /* Specifies the default granularity for SVM, used in buffer
  * migration and restoration of backing memory when handling
@@ -1033,6 +1034,13 @@ module_param_named(user_partt_mode, amdgpu_user_partt_mode, uint, 0444);
 module_param(enforce_isolation, bool, 0444);
 MODULE_PARM_DESC(enforce_isolation, "enforce process isolation between graphics and compute . enforce_isolation = on");
 
+/**
+ * DOC: modeset (int)
+ * Override nomodeset (1 = override, -1 = auto). The default is -1 (auto).
+ */
+MODULE_PARM_DESC(modeset, "Override nomodeset (1 = enable, -1 = auto)");
+module_param_named(modeset, amdgpu_modeset, int, 0444);
+
 /**
  * DOC: seamless (int)
  * Seamless boot will keep the image on the screen during the boot process.
@@ -2244,6 +2252,12 @@ static int amdgpu_pci_probe(struct pci_dev *pdev,
 	int ret, retry = 0, i;
 	bool supports_atomic = false;
 
+	if ((pdev->class >> 8) == PCI_CLASS_DISPLAY_VGA ||
+	    (pdev->class >> 8) == PCI_CLASS_DISPLAY_OTHER) {
+		if (drm_firmware_drivers_only() && amdgpu_modeset == -1)
+			return -EINVAL;
+	}
+
 	/* skip devices which are owned by radeon */
 	for (i = 0; i < ARRAY_SIZE(amdgpu_unsupported_pciidlist); i++) {
 		if (amdgpu_unsupported_pciidlist[i] == pdev->device)
-- 
2.39.5

[PATCH AUTOSEL 6.14 150/642] drm/amdgpu/gfx12: don't read registers in mqd init

[Sasha Levin]

From: Alex Deucher <alexander.deucher@amd.com>

[ Upstream commit fc3c139cf0432b79fd08e23100a559ee51cd0be4 ]

Just use the default values.  There's not need to
get the value from hardware and it could cause problems
if we do that at runtime and gfxoff is active.

Reviewed-by: Mukul Joshi <mukul.joshi@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/gpu/drm/amd/amdgpu/gfx_v12_0.c | 48 ++++++++++++++++++--------
 1 file changed, 33 insertions(+), 15 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v12_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v12_0.c
index 0c08785099f32..2ec900d50d7f8 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v12_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v12_0.c
@@ -52,6 +52,24 @@
 
 #define RLCG_UCODE_LOADING_START_ADDRESS	0x00002000L
 
+#define regCP_GFX_MQD_CONTROL_DEFAULT                                             0x00000100
+#define regCP_GFX_HQD_VMID_DEFAULT                                                0x00000000
+#define regCP_GFX_HQD_QUEUE_PRIORITY_DEFAULT                                      0x00000000
+#define regCP_GFX_HQD_QUANTUM_DEFAULT                                             0x00000a01
+#define regCP_GFX_HQD_CNTL_DEFAULT                                                0x00f00000
+#define regCP_RB_DOORBELL_CONTROL_DEFAULT                                         0x00000000
+#define regCP_GFX_HQD_RPTR_DEFAULT                                                0x00000000
+
+#define regCP_HQD_EOP_CONTROL_DEFAULT                                             0x00000006
+#define regCP_HQD_PQ_DOORBELL_CONTROL_DEFAULT                                     0x00000000
+#define regCP_MQD_CONTROL_DEFAULT                                                 0x00000100
+#define regCP_HQD_PQ_CONTROL_DEFAULT                                              0x00308509
+#define regCP_HQD_PQ_DOORBELL_CONTROL_DEFAULT                                     0x00000000
+#define regCP_HQD_PQ_RPTR_DEFAULT                                                 0x00000000
+#define regCP_HQD_PERSISTENT_STATE_DEFAULT                                        0x0be05501
+#define regCP_HQD_IB_CONTROL_DEFAULT                                              0x00300000
+
+
 MODULE_FIRMWARE("amdgpu/gc_12_0_0_pfp.bin");
 MODULE_FIRMWARE("amdgpu/gc_12_0_0_me.bin");
 MODULE_FIRMWARE("amdgpu/gc_12_0_0_mec.bin");
@@ -2891,25 +2909,25 @@ static int gfx_v12_0_gfx_mqd_init(struct amdgpu_device *adev, void *m,
 	mqd->cp_mqd_base_addr_hi = upper_32_bits(prop->mqd_gpu_addr);
 
 	/* set up mqd control */
-	tmp = RREG32_SOC15(GC, 0, regCP_GFX_MQD_CONTROL);
+	tmp = regCP_GFX_MQD_CONTROL_DEFAULT;
 	tmp = REG_SET_FIELD(tmp, CP_GFX_MQD_CONTROL, VMID, 0);
 	tmp = REG_SET_FIELD(tmp, CP_GFX_MQD_CONTROL, PRIV_STATE, 1);
 	tmp = REG_SET_FIELD(tmp, CP_GFX_MQD_CONTROL, CACHE_POLICY, 0);
 	mqd->cp_gfx_mqd_control = tmp;
 
 	/* set up gfx_hqd_vimd with 0x0 to indicate the ring buffer's vmid */
-	tmp = RREG32_SOC15(GC, 0, regCP_GFX_HQD_VMID);
+	tmp = regCP_GFX_HQD_VMID_DEFAULT;
 	tmp = REG_SET_FIELD(tmp, CP_GFX_HQD_VMID, VMID, 0);
 	mqd->cp_gfx_hqd_vmid = 0;
 
 	/* set up default queue priority level
 	 * 0x0 = low priority, 0x1 = high priority */
-	tmp = RREG32_SOC15(GC, 0, regCP_GFX_HQD_QUEUE_PRIORITY);
+	tmp = regCP_GFX_HQD_QUEUE_PRIORITY_DEFAULT;
 	tmp = REG_SET_FIELD(tmp, CP_GFX_HQD_QUEUE_PRIORITY, PRIORITY_LEVEL, 0);
 	mqd->cp_gfx_hqd_queue_priority = tmp;
 
 	/* set up time quantum */
-	tmp = RREG32_SOC15(GC, 0, regCP_GFX_HQD_QUANTUM);
+	tmp = regCP_GFX_HQD_QUANTUM_DEFAULT;
 	tmp = REG_SET_FIELD(tmp, CP_GFX_HQD_QUANTUM, QUANTUM_EN, 1);
 	mqd->cp_gfx_hqd_quantum = tmp;
 
@@ -2931,7 +2949,7 @@ static int gfx_v12_0_gfx_mqd_init(struct amdgpu_device *adev, void *m,
 
 	/* set up the gfx_hqd_control, similar as CP_RB0_CNTL */
 	rb_bufsz = order_base_2(prop->queue_size / 4) - 1;
-	tmp = RREG32_SOC15(GC, 0, regCP_GFX_HQD_CNTL);
+	tmp = regCP_GFX_HQD_CNTL_DEFAULT;
 	tmp = REG_SET_FIELD(tmp, CP_GFX_HQD_CNTL, RB_BUFSZ, rb_bufsz);
 	tmp = REG_SET_FIELD(tmp, CP_GFX_HQD_CNTL, RB_BLKSZ, rb_bufsz - 2);
 #ifdef __BIG_ENDIAN
@@ -2940,7 +2958,7 @@ static int gfx_v12_0_gfx_mqd_init(struct amdgpu_device *adev, void *m,
 	mqd->cp_gfx_hqd_cntl = tmp;
 
 	/* set up cp_doorbell_control */
-	tmp = RREG32_SOC15(GC, 0, regCP_RB_DOORBELL_CONTROL);
+	tmp = regCP_RB_DOORBELL_CONTROL_DEFAULT;
 	if (prop->use_doorbell) {
 		tmp = REG_SET_FIELD(tmp, CP_RB_DOORBELL_CONTROL,
 				    DOORBELL_OFFSET, prop->doorbell_index);
@@ -2952,7 +2970,7 @@ static int gfx_v12_0_gfx_mqd_init(struct amdgpu_device *adev, void *m,
 	mqd->cp_rb_doorbell_control = tmp;
 
 	/* reset read and write pointers, similar to CP_RB0_WPTR/_RPTR */
-	mqd->cp_gfx_hqd_rptr = RREG32_SOC15(GC, 0, regCP_GFX_HQD_RPTR);
+	mqd->cp_gfx_hqd_rptr = regCP_GFX_HQD_RPTR_DEFAULT;
 
 	/* active the queue */
 	mqd->cp_gfx_hqd_active = 1;
@@ -3047,14 +3065,14 @@ static int gfx_v12_0_compute_mqd_init(struct amdgpu_device *adev, void *m,
 	mqd->cp_hqd_eop_base_addr_hi = upper_32_bits(eop_base_addr);
 
 	/* set the EOP size, register value is 2^(EOP_SIZE+1) dwords */
-	tmp = RREG32_SOC15(GC, 0, regCP_HQD_EOP_CONTROL);
+	tmp = regCP_HQD_EOP_CONTROL_DEFAULT;
 	tmp = REG_SET_FIELD(tmp, CP_HQD_EOP_CONTROL, EOP_SIZE,
 			(order_base_2(GFX12_MEC_HPD_SIZE / 4) - 1));
 
 	mqd->cp_hqd_eop_control = tmp;
 
 	/* enable doorbell? */
-	tmp = RREG32_SOC15(GC, 0, regCP_HQD_PQ_DOORBELL_CONTROL);
+	tmp = regCP_HQD_PQ_DOORBELL_CONTROL_DEFAULT;
 
 	if (prop->use_doorbell) {
 		tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_DOORBELL_CONTROL,
@@ -3083,7 +3101,7 @@ static int gfx_v12_0_compute_mqd_init(struct amdgpu_device *adev, void *m,
 	mqd->cp_mqd_base_addr_hi = upper_32_bits(prop->mqd_gpu_addr);
 
 	/* set MQD vmid to 0 */
-	tmp = RREG32_SOC15(GC, 0, regCP_MQD_CONTROL);
+	tmp = regCP_MQD_CONTROL_DEFAULT;
 	tmp = REG_SET_FIELD(tmp, CP_MQD_CONTROL, VMID, 0);
 	mqd->cp_mqd_control = tmp;
 
@@ -3093,7 +3111,7 @@ static int gfx_v12_0_compute_mqd_init(struct amdgpu_device *adev, void *m,
 	mqd->cp_hqd_pq_base_hi = upper_32_bits(hqd_gpu_addr);
 
 	/* set up the HQD, this is similar to CP_RB0_CNTL */
-	tmp = RREG32_SOC15(GC, 0, regCP_HQD_PQ_CONTROL);
+	tmp = regCP_HQD_PQ_CONTROL_DEFAULT;
 	tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, QUEUE_SIZE,
 			    (order_base_2(prop->queue_size / 4) - 1));
 	tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, RPTR_BLOCK_SIZE,
@@ -3118,7 +3136,7 @@ static int gfx_v12_0_compute_mqd_init(struct amdgpu_device *adev, void *m,
 	tmp = 0;
 	/* enable the doorbell if requested */
 	if (prop->use_doorbell) {
-		tmp = RREG32_SOC15(GC, 0, regCP_HQD_PQ_DOORBELL_CONTROL);
+		tmp = regCP_HQD_PQ_DOORBELL_CONTROL_DEFAULT;
 		tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_DOORBELL_CONTROL,
 				DOORBELL_OFFSET, prop->doorbell_index);
 
@@ -3133,17 +3151,17 @@ static int gfx_v12_0_compute_mqd_init(struct amdgpu_device *adev, void *m,
 	mqd->cp_hqd_pq_doorbell_control = tmp;
 
 	/* reset read and write pointers, similar to CP_RB0_WPTR/_RPTR */
-	mqd->cp_hqd_pq_rptr = RREG32_SOC15(GC, 0, regCP_HQD_PQ_RPTR);
+	mqd->cp_hqd_pq_rptr = regCP_HQD_PQ_RPTR_DEFAULT;
 
 	/* set the vmid for the queue */
 	mqd->cp_hqd_vmid = 0;
 
-	tmp = RREG32_SOC15(GC, 0, regCP_HQD_PERSISTENT_STATE);
+	tmp = regCP_HQD_PERSISTENT_STATE_DEFAULT;
 	tmp = REG_SET_FIELD(tmp, CP_HQD_PERSISTENT_STATE, PRELOAD_SIZE, 0x55);
 	mqd->cp_hqd_persistent_state = tmp;
 
 	/* set MIN_IB_AVAIL_SIZE */
-	tmp = RREG32_SOC15(GC, 0, regCP_HQD_IB_CONTROL);
+	tmp = regCP_HQD_IB_CONTROL_DEFAULT;
 	tmp = REG_SET_FIELD(tmp, CP_HQD_IB_CONTROL, MIN_IB_AVAIL_SIZE, 3);
 	mqd->cp_hqd_ib_control = tmp;
 
-- 
2.39.5

[PATCH AUTOSEL 6.14 151/642] drm/amdgpu/gfx11: don't read registers in mqd init

[Sasha Levin]

From: Alex Deucher <alexander.deucher@amd.com>

[ Upstream commit e27b36ea6ba5f29e91fcfb375ea29503708fcf43 ]

Just use the default values.  There's not need to
get the value from hardware and it could cause problems
if we do that at runtime and gfxoff is active.

Reviewed-by: Mukul Joshi <mukul.joshi@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c | 47 ++++++++++++++++++--------
 1 file changed, 32 insertions(+), 15 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c
index f1f53c7687410..e050c2e4ea734 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c
@@ -64,6 +64,23 @@
 #define regPC_CONFIG_CNTL_1		0x194d
 #define regPC_CONFIG_CNTL_1_BASE_IDX	1
 
+#define regCP_GFX_MQD_CONTROL_DEFAULT                                             0x00000100
+#define regCP_GFX_HQD_VMID_DEFAULT                                                0x00000000
+#define regCP_GFX_HQD_QUEUE_PRIORITY_DEFAULT                                      0x00000000
+#define regCP_GFX_HQD_QUANTUM_DEFAULT                                             0x00000a01
+#define regCP_GFX_HQD_CNTL_DEFAULT                                                0x00a00000
+#define regCP_RB_DOORBELL_CONTROL_DEFAULT                                         0x00000000
+#define regCP_GFX_HQD_RPTR_DEFAULT                                                0x00000000
+
+#define regCP_HQD_EOP_CONTROL_DEFAULT                                             0x00000006
+#define regCP_HQD_PQ_DOORBELL_CONTROL_DEFAULT                                     0x00000000
+#define regCP_MQD_CONTROL_DEFAULT                                                 0x00000100
+#define regCP_HQD_PQ_CONTROL_DEFAULT                                              0x00308509
+#define regCP_HQD_PQ_DOORBELL_CONTROL_DEFAULT                                     0x00000000
+#define regCP_HQD_PQ_RPTR_DEFAULT                                                 0x00000000
+#define regCP_HQD_PERSISTENT_STATE_DEFAULT                                        0x0be05501
+#define regCP_HQD_IB_CONTROL_DEFAULT                                              0x00300000
+
 MODULE_FIRMWARE("amdgpu/gc_11_0_0_pfp.bin");
 MODULE_FIRMWARE("amdgpu/gc_11_0_0_me.bin");
 MODULE_FIRMWARE("amdgpu/gc_11_0_0_mec.bin");
@@ -3958,7 +3975,7 @@ static void gfx_v11_0_gfx_mqd_set_priority(struct amdgpu_device *adev,
 	if (prop->hqd_pipe_priority == AMDGPU_GFX_PIPE_PRIO_HIGH)
 		priority = 1;
 
-	tmp = RREG32_SOC15(GC, 0, regCP_GFX_HQD_QUEUE_PRIORITY);
+	tmp = regCP_GFX_HQD_QUEUE_PRIORITY_DEFAULT;
 	tmp = REG_SET_FIELD(tmp, CP_GFX_HQD_QUEUE_PRIORITY, PRIORITY_LEVEL, priority);
 	mqd->cp_gfx_hqd_queue_priority = tmp;
 }
@@ -3980,14 +3997,14 @@ static int gfx_v11_0_gfx_mqd_init(struct amdgpu_device *adev, void *m,
 	mqd->cp_mqd_base_addr_hi = upper_32_bits(prop->mqd_gpu_addr);
 
 	/* set up mqd control */
-	tmp = RREG32_SOC15(GC, 0, regCP_GFX_MQD_CONTROL);
+	tmp = regCP_GFX_MQD_CONTROL_DEFAULT;
 	tmp = REG_SET_FIELD(tmp, CP_GFX_MQD_CONTROL, VMID, 0);
 	tmp = REG_SET_FIELD(tmp, CP_GFX_MQD_CONTROL, PRIV_STATE, 1);
 	tmp = REG_SET_FIELD(tmp, CP_GFX_MQD_CONTROL, CACHE_POLICY, 0);
 	mqd->cp_gfx_mqd_control = tmp;
 
 	/* set up gfx_hqd_vimd with 0x0 to indicate the ring buffer's vmid */
-	tmp = RREG32_SOC15(GC, 0, regCP_GFX_HQD_VMID);
+	tmp = regCP_GFX_HQD_VMID_DEFAULT;
 	tmp = REG_SET_FIELD(tmp, CP_GFX_HQD_VMID, VMID, 0);
 	mqd->cp_gfx_hqd_vmid = 0;
 
@@ -3995,7 +4012,7 @@ static int gfx_v11_0_gfx_mqd_init(struct amdgpu_device *adev, void *m,
 	gfx_v11_0_gfx_mqd_set_priority(adev, mqd, prop);
 
 	/* set up time quantum */
-	tmp = RREG32_SOC15(GC, 0, regCP_GFX_HQD_QUANTUM);
+	tmp = regCP_GFX_HQD_QUANTUM_DEFAULT;
 	tmp = REG_SET_FIELD(tmp, CP_GFX_HQD_QUANTUM, QUANTUM_EN, 1);
 	mqd->cp_gfx_hqd_quantum = tmp;
 
@@ -4017,7 +4034,7 @@ static int gfx_v11_0_gfx_mqd_init(struct amdgpu_device *adev, void *m,
 
 	/* set up the gfx_hqd_control, similar as CP_RB0_CNTL */
 	rb_bufsz = order_base_2(prop->queue_size / 4) - 1;
-	tmp = RREG32_SOC15(GC, 0, regCP_GFX_HQD_CNTL);
+	tmp = regCP_GFX_HQD_CNTL_DEFAULT;
 	tmp = REG_SET_FIELD(tmp, CP_GFX_HQD_CNTL, RB_BUFSZ, rb_bufsz);
 	tmp = REG_SET_FIELD(tmp, CP_GFX_HQD_CNTL, RB_BLKSZ, rb_bufsz - 2);
 #ifdef __BIG_ENDIAN
@@ -4026,7 +4043,7 @@ static int gfx_v11_0_gfx_mqd_init(struct amdgpu_device *adev, void *m,
 	mqd->cp_gfx_hqd_cntl = tmp;
 
 	/* set up cp_doorbell_control */
-	tmp = RREG32_SOC15(GC, 0, regCP_RB_DOORBELL_CONTROL);
+	tmp = regCP_RB_DOORBELL_CONTROL_DEFAULT;
 	if (prop->use_doorbell) {
 		tmp = REG_SET_FIELD(tmp, CP_RB_DOORBELL_CONTROL,
 				    DOORBELL_OFFSET, prop->doorbell_index);
@@ -4038,7 +4055,7 @@ static int gfx_v11_0_gfx_mqd_init(struct amdgpu_device *adev, void *m,
 	mqd->cp_rb_doorbell_control = tmp;
 
 	/* reset read and write pointers, similar to CP_RB0_WPTR/_RPTR */
-	mqd->cp_gfx_hqd_rptr = RREG32_SOC15(GC, 0, regCP_GFX_HQD_RPTR);
+	mqd->cp_gfx_hqd_rptr = regCP_GFX_HQD_RPTR_DEFAULT;
 
 	/* active the queue */
 	mqd->cp_gfx_hqd_active = 1;
@@ -4124,14 +4141,14 @@ static int gfx_v11_0_compute_mqd_init(struct amdgpu_device *adev, void *m,
 	mqd->cp_hqd_eop_base_addr_hi = upper_32_bits(eop_base_addr);
 
 	/* set the EOP size, register value is 2^(EOP_SIZE+1) dwords */
-	tmp = RREG32_SOC15(GC, 0, regCP_HQD_EOP_CONTROL);
+	tmp = regCP_HQD_EOP_CONTROL_DEFAULT;
 	tmp = REG_SET_FIELD(tmp, CP_HQD_EOP_CONTROL, EOP_SIZE,
 			(order_base_2(GFX11_MEC_HPD_SIZE / 4) - 1));
 
 	mqd->cp_hqd_eop_control = tmp;
 
 	/* enable doorbell? */
-	tmp = RREG32_SOC15(GC, 0, regCP_HQD_PQ_DOORBELL_CONTROL);
+	tmp = regCP_HQD_PQ_DOORBELL_CONTROL_DEFAULT;
 
 	if (prop->use_doorbell) {
 		tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_DOORBELL_CONTROL,
@@ -4160,7 +4177,7 @@ static int gfx_v11_0_compute_mqd_init(struct amdgpu_device *adev, void *m,
 	mqd->cp_mqd_base_addr_hi = upper_32_bits(prop->mqd_gpu_addr);
 
 	/* set MQD vmid to 0 */
-	tmp = RREG32_SOC15(GC, 0, regCP_MQD_CONTROL);
+	tmp = regCP_MQD_CONTROL_DEFAULT;
 	tmp = REG_SET_FIELD(tmp, CP_MQD_CONTROL, VMID, 0);
 	mqd->cp_mqd_control = tmp;
 
@@ -4170,7 +4187,7 @@ static int gfx_v11_0_compute_mqd_init(struct amdgpu_device *adev, void *m,
 	mqd->cp_hqd_pq_base_hi = upper_32_bits(hqd_gpu_addr);
 
 	/* set up the HQD, this is similar to CP_RB0_CNTL */
-	tmp = RREG32_SOC15(GC, 0, regCP_HQD_PQ_CONTROL);
+	tmp = regCP_HQD_PQ_CONTROL_DEFAULT;
 	tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, QUEUE_SIZE,
 			    (order_base_2(prop->queue_size / 4) - 1));
 	tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_CONTROL, RPTR_BLOCK_SIZE,
@@ -4196,7 +4213,7 @@ static int gfx_v11_0_compute_mqd_init(struct amdgpu_device *adev, void *m,
 	tmp = 0;
 	/* enable the doorbell if requested */
 	if (prop->use_doorbell) {
-		tmp = RREG32_SOC15(GC, 0, regCP_HQD_PQ_DOORBELL_CONTROL);
+		tmp = regCP_HQD_PQ_DOORBELL_CONTROL_DEFAULT;
 		tmp = REG_SET_FIELD(tmp, CP_HQD_PQ_DOORBELL_CONTROL,
 				DOORBELL_OFFSET, prop->doorbell_index);
 
@@ -4211,17 +4228,17 @@ static int gfx_v11_0_compute_mqd_init(struct amdgpu_device *adev, void *m,
 	mqd->cp_hqd_pq_doorbell_control = tmp;
 
 	/* reset read and write pointers, similar to CP_RB0_WPTR/_RPTR */
-	mqd->cp_hqd_pq_rptr = RREG32_SOC15(GC, 0, regCP_HQD_PQ_RPTR);
+	mqd->cp_hqd_pq_rptr = regCP_HQD_PQ_RPTR_DEFAULT;
 
 	/* set the vmid for the queue */
 	mqd->cp_hqd_vmid = 0;
 
-	tmp = RREG32_SOC15(GC, 0, regCP_HQD_PERSISTENT_STATE);
+	tmp = regCP_HQD_PERSISTENT_STATE_DEFAULT;
 	tmp = REG_SET_FIELD(tmp, CP_HQD_PERSISTENT_STATE, PRELOAD_SIZE, 0x55);
 	mqd->cp_hqd_persistent_state = tmp;
 
 	/* set MIN_IB_AVAIL_SIZE */
-	tmp = RREG32_SOC15(GC, 0, regCP_HQD_IB_CONTROL);
+	tmp = regCP_HQD_IB_CONTROL_DEFAULT;
 	tmp = REG_SET_FIELD(tmp, CP_HQD_IB_CONTROL, MIN_IB_AVAIL_SIZE, 3);
 	mqd->cp_hqd_ib_control = tmp;
 
-- 
2.39.5

[PATCH AUTOSEL 6.14 152/642] drm/amdgpu: Update SRIOV video codec caps

[Sasha Levin]

From: David Rosca <david.rosca@amd.com>

[ Upstream commit 19478f2011f8b53dee401c91423c4e0b73753e4f ]

There have been multiple fixes to the video caps that are missing for
SRIOV. Update the SRIOV caps with correct values.

Signed-off-by: David Rosca <david.rosca@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Ruijing Dong <ruijing.dong@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/gpu/drm/amd/amdgpu/nv.c    | 16 ++++++++--------
 drivers/gpu/drm/amd/amdgpu/soc21.c | 10 ++--------
 2 files changed, 10 insertions(+), 16 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/nv.c b/drivers/gpu/drm/amd/amdgpu/nv.c
index 95c609317a8d7..efca7dc27d68d 100644
--- a/drivers/gpu/drm/amd/amdgpu/nv.c
+++ b/drivers/gpu/drm/amd/amdgpu/nv.c
@@ -141,23 +141,23 @@ static struct amdgpu_video_codec_info sriov_sc_video_codecs_encode_array[] = {
 };
 
 static struct amdgpu_video_codec_info sriov_sc_video_codecs_decode_array_vcn0[] = {
-	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG2, 4096, 4096, 3)},
-	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4, 4096, 4096, 5)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG2, 1920, 1088, 3)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4, 1920, 1088, 5)},
 	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4_AVC, 4096, 4096, 52)},
-	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_VC1, 4096, 4096, 4)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_VC1, 1920, 1088, 4)},
 	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_HEVC, 8192, 4352, 186)},
-	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_JPEG, 4096, 4096, 0)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_JPEG, 16384, 16384, 0)},
 	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_VP9, 8192, 4352, 0)},
 	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_AV1, 8192, 4352, 0)},
 };
 
 static struct amdgpu_video_codec_info sriov_sc_video_codecs_decode_array_vcn1[] = {
-	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG2, 4096, 4096, 3)},
-	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4, 4096, 4096, 5)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG2, 1920, 1088, 3)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4, 1920, 1088, 5)},
 	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4_AVC, 4096, 4096, 52)},
-	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_VC1, 4096, 4096, 4)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_VC1, 1920, 1088, 4)},
 	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_HEVC, 8192, 4352, 186)},
-	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_JPEG, 4096, 4096, 0)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_JPEG, 16384, 16384, 0)},
 	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_VP9, 8192, 4352, 0)},
 };
 
diff --git a/drivers/gpu/drm/amd/amdgpu/soc21.c b/drivers/gpu/drm/amd/amdgpu/soc21.c
index 62ad67d0b598f..c66cff399f63d 100644
--- a/drivers/gpu/drm/amd/amdgpu/soc21.c
+++ b/drivers/gpu/drm/amd/amdgpu/soc21.c
@@ -117,23 +117,17 @@ static struct amdgpu_video_codecs sriov_vcn_4_0_0_video_codecs_encode_vcn1 = {
 };
 
 static struct amdgpu_video_codec_info sriov_vcn_4_0_0_video_codecs_decode_array_vcn0[] = {
-	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG2, 4096, 4096, 3)},
-	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4, 4096, 4096, 5)},
 	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4_AVC, 4096, 4096, 52)},
-	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_VC1, 4096, 4096, 4)},
 	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_HEVC, 8192, 4352, 186)},
-	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_JPEG, 4096, 4096, 0)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_JPEG, 16384, 16384, 0)},
 	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_VP9, 8192, 4352, 0)},
 	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_AV1, 8192, 4352, 0)},
 };
 
 static struct amdgpu_video_codec_info sriov_vcn_4_0_0_video_codecs_decode_array_vcn1[] = {
-	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG2, 4096, 4096, 3)},
-	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4, 4096, 4096, 5)},
 	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_MPEG4_AVC, 4096, 4096, 52)},
-	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_VC1, 4096, 4096, 4)},
 	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_HEVC, 8192, 4352, 186)},
-	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_JPEG, 4096, 4096, 0)},
+	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_JPEG, 16384, 16384, 0)},
 	{codec_info_build(AMDGPU_INFO_VIDEO_CAPS_CODEC_IDX_VP9, 8192, 4352, 0)},
 };
 
-- 
2.39.5

[PATCH AUTOSEL 6.14 194/642] drm/amd/display: remove minimum Dispclk and apply oem panel timing.

[Sasha Levin]

From: Charlene Liu <Charlene.Liu@amd.com>

[ Upstream commit 756e58e83e89d372b94269c0cde61fe55da76947 ]

[why & how]
1. apply oem panel timing (not only on OLED)
2. remove MIN_DPP_DISP_CLK request in driver.

This fix will apply for dcn31x but not
sync with DML's output.

Reviewed-by: Ovidiu Bunea <ovidiu.bunea@amd.com>
Signed-off-by: Charlene Liu <Charlene.Liu@amd.com>
Signed-off-by: Tom Chung <chiahsuan.chung@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/gpu/drm/amd/display/dc/clk_mgr/dcn315/dcn315_clk_mgr.c | 2 --
 drivers/gpu/drm/amd/display/dc/clk_mgr/dcn316/dcn316_clk_mgr.c | 2 --
 drivers/gpu/drm/amd/display/dc/hwss/dce110/dce110_hwseq.c      | 3 ++-
 3 files changed, 2 insertions(+), 5 deletions(-)

diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn315/dcn315_clk_mgr.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn315/dcn315_clk_mgr.c
index 827b24b3442ad..e4d22f74f9869 100644
--- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn315/dcn315_clk_mgr.c
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn315/dcn315_clk_mgr.c
@@ -194,8 +194,6 @@ static void dcn315_update_clocks(struct clk_mgr *clk_mgr_base,
 	// workaround: Limit dppclk to 100Mhz to avoid lower eDP panel switch to plus 4K monitor underflow.
 	if (new_clocks->dppclk_khz < MIN_DPP_DISP_CLK)
 		new_clocks->dppclk_khz = MIN_DPP_DISP_CLK;
-	if (new_clocks->dispclk_khz < MIN_DPP_DISP_CLK)
-		new_clocks->dispclk_khz = MIN_DPP_DISP_CLK;
 
 	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)
diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn316/dcn316_clk_mgr.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn316/dcn316_clk_mgr.c
index 37c39756fece4..49efea0c8fcff 100644
--- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn316/dcn316_clk_mgr.c
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn316/dcn316_clk_mgr.c
@@ -201,8 +201,6 @@ static void dcn316_update_clocks(struct clk_mgr *clk_mgr_base,
 	// workaround: Limit dppclk to 100Mhz to avoid lower eDP panel switch to plus 4K monitor underflow.
 	if (new_clocks->dppclk_khz < 100000)
 		new_clocks->dppclk_khz = 100000;
-	if (new_clocks->dispclk_khz < 100000)
-		new_clocks->dispclk_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)
diff --git a/drivers/gpu/drm/amd/display/dc/hwss/dce110/dce110_hwseq.c b/drivers/gpu/drm/amd/display/dc/hwss/dce110/dce110_hwseq.c
index 81f4c386c2875..fcbde50213d69 100644
--- a/drivers/gpu/drm/amd/display/dc/hwss/dce110/dce110_hwseq.c
+++ b/drivers/gpu/drm/amd/display/dc/hwss/dce110/dce110_hwseq.c
@@ -1065,7 +1065,8 @@ void dce110_edp_backlight_control(
 			DC_LOG_DC("edp_receiver_ready_T9 skipped\n");
 	}
 
-	if (!enable && link->dpcd_sink_ext_caps.bits.oled) {
+	if (!enable) {
+		/*follow oem panel config's requirement*/
 		pre_T11_delay += link->panel_config.pps.extra_pre_t11_ms;
 		msleep(pre_T11_delay);
 	}
-- 
2.39.5

[PATCH AUTOSEL 6.14 195/642] drm/amd/display: calculate the remain segments for all pipes

[Sasha Levin]

From: Zhikai Zhai <zhikai.zhai@amd.com>

[ Upstream commit d3069feecdb5542604d29b59acfd1fd213bad95b ]

[WHY]
In some cases the remain de-tile buffer segments will be greater
than zero if we don't add the non-top pipe to calculate, at
this time the override de-tile buffer size will be valid and used.
But it makes the de-tile buffer segments used finally for all of pipes
exceed the maximum.

[HOW]
Add the non-top pipe to calculate the remain de-tile buffer segments.
Don't set override size to use the average according to pipe count
if the value exceed the maximum.

Reviewed-by: Charlene Liu <charlene.liu@amd.com>
Signed-off-by: Zhikai Zhai <zhikai.zhai@amd.com>
Signed-off-by: Tom Chung <chiahsuan.chung@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 .../dc/resource/dcn315/dcn315_resource.c      | 42 +++++++++----------
 1 file changed, 20 insertions(+), 22 deletions(-)

diff --git a/drivers/gpu/drm/amd/display/dc/resource/dcn315/dcn315_resource.c b/drivers/gpu/drm/amd/display/dc/resource/dcn315/dcn315_resource.c
index 14acef036b5a0..6c2bb3f63be15 100644
--- a/drivers/gpu/drm/amd/display/dc/resource/dcn315/dcn315_resource.c
+++ b/drivers/gpu/drm/amd/display/dc/resource/dcn315/dcn315_resource.c
@@ -1698,7 +1698,7 @@ static int dcn315_populate_dml_pipes_from_context(
 		pipes[pipe_cnt].dout.dsc_input_bpc = 0;
 		DC_FP_START();
 		dcn31_zero_pipe_dcc_fraction(pipes, pipe_cnt);
-		if (pixel_rate_crb && !pipe->top_pipe && !pipe->prev_odm_pipe) {
+		if (pixel_rate_crb) {
 			int bpp = source_format_to_bpp(pipes[pipe_cnt].pipe.src.source_format);
 			/* Ceil to crb segment size */
 			int approx_det_segs_required_for_pstate = dcn_get_approx_det_segs_required_for_pstate(
@@ -1755,28 +1755,26 @@ static int dcn315_populate_dml_pipes_from_context(
 				continue;
 			}
 
-			if (!pipe->top_pipe && !pipe->prev_odm_pipe) {
-				bool split_required = pipe->stream->timing.pix_clk_100hz >= dcn_get_max_non_odm_pix_rate_100hz(&dc->dml.soc)
-						|| (pipe->plane_state && pipe->plane_state->src_rect.width > 5120);
-
-				if (remaining_det_segs > MIN_RESERVED_DET_SEGS && crb_pipes != 0)
-					pipes[pipe_cnt].pipe.src.det_size_override += (remaining_det_segs - MIN_RESERVED_DET_SEGS) / crb_pipes +
-							(crb_idx < (remaining_det_segs - MIN_RESERVED_DET_SEGS) % crb_pipes ? 1 : 0);
-				if (pipes[pipe_cnt].pipe.src.det_size_override > 2 * DCN3_15_MAX_DET_SEGS) {
-					/* Clamp to 2 pipe split max det segments */
-					remaining_det_segs += pipes[pipe_cnt].pipe.src.det_size_override - 2 * (DCN3_15_MAX_DET_SEGS);
-					pipes[pipe_cnt].pipe.src.det_size_override = 2 * DCN3_15_MAX_DET_SEGS;
-				}
-				if (pipes[pipe_cnt].pipe.src.det_size_override > DCN3_15_MAX_DET_SEGS || split_required) {
-					/* If we are splitting we must have an even number of segments */
-					remaining_det_segs += pipes[pipe_cnt].pipe.src.det_size_override % 2;
-					pipes[pipe_cnt].pipe.src.det_size_override -= pipes[pipe_cnt].pipe.src.det_size_override % 2;
-				}
-				/* Convert segments into size for DML use */
-				pipes[pipe_cnt].pipe.src.det_size_override *= DCN3_15_CRB_SEGMENT_SIZE_KB;
-
-				crb_idx++;
+			bool split_required = pipe->stream->timing.pix_clk_100hz >= dcn_get_max_non_odm_pix_rate_100hz(&dc->dml.soc)
+					|| (pipe->plane_state && pipe->plane_state->src_rect.width > 5120);
+
+			if (remaining_det_segs > MIN_RESERVED_DET_SEGS && crb_pipes != 0)
+				pipes[pipe_cnt].pipe.src.det_size_override += (remaining_det_segs - MIN_RESERVED_DET_SEGS) / crb_pipes +
+						(crb_idx < (remaining_det_segs - MIN_RESERVED_DET_SEGS) % crb_pipes ? 1 : 0);
+			if (pipes[pipe_cnt].pipe.src.det_size_override > 2 * DCN3_15_MAX_DET_SEGS) {
+				/* Clamp to 2 pipe split max det segments */
+				remaining_det_segs += pipes[pipe_cnt].pipe.src.det_size_override - 2 * (DCN3_15_MAX_DET_SEGS);
+				pipes[pipe_cnt].pipe.src.det_size_override = 2 * DCN3_15_MAX_DET_SEGS;
+			}
+			if (pipes[pipe_cnt].pipe.src.det_size_override > DCN3_15_MAX_DET_SEGS || split_required) {
+				/* If we are splitting we must have an even number of segments */
+				remaining_det_segs += pipes[pipe_cnt].pipe.src.det_size_override % 2;
+				pipes[pipe_cnt].pipe.src.det_size_override -= pipes[pipe_cnt].pipe.src.det_size_override % 2;
 			}
+			/* Convert segments into size for DML use */
+			pipes[pipe_cnt].pipe.src.det_size_override *= DCN3_15_CRB_SEGMENT_SIZE_KB;
+
+			crb_idx++;
 			pipe_cnt++;
 		}
 	}
-- 
2.39.5

[PATCH AUTOSEL 6.14 196/642] drm/amd/display: not abort link train when bw is low

[Sasha Levin]

From: Peichen Huang <PeiChen.Huang@amd.com>

[ Upstream commit 8a21da2842bb22b2b80e5902d0438030d729bfd3 ]

[WHY]
DP tunneling should not abort link train even bandwidth become
too low after downgrade. Otherwise, it would fail compliance test.

[HOW}
Do link train with downgrade settings even bandwidth is not enough

Reviewed-by: Cruise Hung <cruise.hung@amd.com>
Reviewed-by: Meenakshikumar Somasundaram <meenakshikumar.somasundaram@amd.com>
Signed-off-by: Peichen Huang <PeiChen.Huang@amd.com>
Signed-off-by: Tom Chung <chiahsuan.chung@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 .../gpu/drm/amd/display/dc/link/protocols/link_dp_training.c | 5 +----
 1 file changed, 1 insertion(+), 4 deletions(-)

diff --git a/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_training.c b/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_training.c
index 751c18e592ea5..7848ddb94456c 100644
--- a/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_training.c
+++ b/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_training.c
@@ -1782,13 +1782,10 @@ bool perform_link_training_with_retries(
 			is_link_bw_min = ((cur_link_settings.link_rate <= LINK_RATE_LOW) &&
 				(cur_link_settings.lane_count <= LANE_COUNT_ONE));
 
-			if (is_link_bw_low) {
+			if (is_link_bw_low)
 				DC_LOG_WARNING(
 					"%s: Link(%d) bandwidth too low after fallback req_bw(%d) > link_bw(%d)\n",
 					__func__, link->link_index, req_bw, link_bw);
-
-				return false;
-			}
 		}
 
 		msleep(delay_between_attempts);
-- 
2.39.5

[PATCH AUTOSEL 6.14 197/642] drm/amd/display: Do not enable replay when vtotal update is pending.

[Sasha Levin]

From: Danny Wang <danny.wang@amd.com>

[ Upstream commit bd00b29b5f236dce677089319176dee5872b5a7a ]

[Why&How]
Vtotal is not applied to HW when handling vsync interrupt.
Make sure vtotal is aligned before enable replay.

Reviewed-by: Anthony Koo <anthony.koo@amd.com>
Reviewed-by: Robin Chen <robin.chen@amd.com>
Signed-off-by: Danny Wang <danny.wang@amd.com>
Signed-off-by: Zhongwei Zhang <Zhongwei.Zhang@amd.com>
Signed-off-by: Tom Chung <chiahsuan.chung@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/gpu/drm/amd/display/dc/core/dc.c          |  9 +++++++--
 .../gpu/drm/amd/display/dc/core/dc_hw_sequencer.c | 15 +++++++++++++++
 drivers/gpu/drm/amd/display/dc/dc_hw_types.h      |  1 +
 .../drm/amd/display/dc/hwss/dce110/dce110_hwseq.c |  7 ++-----
 .../drm/amd/display/dc/hwss/dcn10/dcn10_hwseq.c   |  7 ++-----
 .../drm/amd/display/dc/hwss/dcn20/dcn20_hwseq.c   |  8 ++------
 .../drm/amd/display/dc/hwss/dcn31/dcn31_hwseq.c   |  4 +---
 .../drm/amd/display/dc/hwss/dcn35/dcn35_hwseq.c   |  3 +--
 .../drm/amd/display/dc/hwss/dcn401/dcn401_hwseq.c | 10 +++-------
 .../gpu/drm/amd/display/dc/hwss/hw_sequencer.h    |  6 ++++++
 10 files changed, 40 insertions(+), 30 deletions(-)

diff --git a/drivers/gpu/drm/amd/display/dc/core/dc.c b/drivers/gpu/drm/amd/display/dc/core/dc.c
index 4683c7ef4507f..a444fe1e0838a 100644
--- a/drivers/gpu/drm/amd/display/dc/core/dc.c
+++ b/drivers/gpu/drm/amd/display/dc/core/dc.c
@@ -452,6 +452,7 @@ bool dc_stream_adjust_vmin_vmax(struct dc *dc,
 
 	if (dc->caps.max_v_total != 0 &&
 		(adjust->v_total_max > dc->caps.max_v_total || adjust->v_total_min > dc->caps.max_v_total)) {
+		stream->adjust.timing_adjust_pending = false;
 		if (adjust->allow_otg_v_count_halt)
 			return set_long_vtotal(dc, stream, adjust);
 		else
@@ -465,7 +466,7 @@ bool dc_stream_adjust_vmin_vmax(struct dc *dc,
 			dc->hwss.set_drr(&pipe,
 					1,
 					*adjust);
-
+			stream->adjust.timing_adjust_pending = false;
 			return true;
 		}
 	}
@@ -3127,8 +3128,12 @@ static void copy_stream_update_to_stream(struct dc *dc,
 	if (update->vrr_active_fixed)
 		stream->vrr_active_fixed = *update->vrr_active_fixed;
 
-	if (update->crtc_timing_adjust)
+	if (update->crtc_timing_adjust) {
+		if (stream->adjust.v_total_min != update->crtc_timing_adjust->v_total_min ||
+			stream->adjust.v_total_max != update->crtc_timing_adjust->v_total_max)
+			stream->adjust.timing_adjust_pending = true;
 		stream->adjust = *update->crtc_timing_adjust;
+	}
 
 	if (update->dpms_off)
 		stream->dpms_off = *update->dpms_off;
diff --git a/drivers/gpu/drm/amd/display/dc/core/dc_hw_sequencer.c b/drivers/gpu/drm/amd/display/dc/core/dc_hw_sequencer.c
index 6eb9bae3af912..a49604b7701f7 100644
--- a/drivers/gpu/drm/amd/display/dc/core/dc_hw_sequencer.c
+++ b/drivers/gpu/drm/amd/display/dc/core/dc_hw_sequencer.c
@@ -609,6 +609,21 @@ void set_p_state_switch_method(
 	}
 }
 
+void set_drr_and_clear_adjust_pending(
+		struct pipe_ctx *pipe_ctx,
+		struct dc_stream_state *stream,
+		struct drr_params *params)
+{
+	/* params can be null.*/
+	if (pipe_ctx && pipe_ctx->stream_res.tg &&
+			pipe_ctx->stream_res.tg->funcs->set_drr)
+		pipe_ctx->stream_res.tg->funcs->set_drr(
+				pipe_ctx->stream_res.tg, params);
+
+	if (stream)
+		stream->adjust.timing_adjust_pending = false;
+}
+
 void get_fams2_visual_confirm_color(
 		struct dc *dc,
 		struct dc_state *context,
diff --git a/drivers/gpu/drm/amd/display/dc/dc_hw_types.h b/drivers/gpu/drm/amd/display/dc/dc_hw_types.h
index 5ac55601a6da1..37e381fc7f02a 100644
--- a/drivers/gpu/drm/amd/display/dc/dc_hw_types.h
+++ b/drivers/gpu/drm/amd/display/dc/dc_hw_types.h
@@ -1015,6 +1015,7 @@ struct dc_crtc_timing_adjust {
 	uint32_t v_total_mid;
 	uint32_t v_total_mid_frame_num;
 	uint32_t allow_otg_v_count_halt;
+	uint8_t timing_adjust_pending;
 };
 
 
diff --git a/drivers/gpu/drm/amd/display/dc/hwss/dce110/dce110_hwseq.c b/drivers/gpu/drm/amd/display/dc/hwss/dce110/dce110_hwseq.c
index fcbde50213d69..94ceccfc04982 100644
--- a/drivers/gpu/drm/amd/display/dc/hwss/dce110/dce110_hwseq.c
+++ b/drivers/gpu/drm/amd/display/dc/hwss/dce110/dce110_hwseq.c
@@ -1655,9 +1655,7 @@ enum dc_status dce110_apply_single_controller_ctx_to_hw(
 
 	params.vertical_total_min = stream->adjust.v_total_min;
 	params.vertical_total_max = stream->adjust.v_total_max;
-	if (pipe_ctx->stream_res.tg->funcs->set_drr)
-		pipe_ctx->stream_res.tg->funcs->set_drr(
-			pipe_ctx->stream_res.tg, &params);
+	set_drr_and_clear_adjust_pending(pipe_ctx, stream, &params);
 
 	// DRR should set trigger event to monitor surface update event
 	if (stream->adjust.v_total_min != 0 && stream->adjust.v_total_max != 0)
@@ -2105,8 +2103,7 @@ static void set_drr(struct pipe_ctx **pipe_ctx,
 		struct timing_generator *tg = pipe_ctx[i]->stream_res.tg;
 
 		if ((tg != NULL) && tg->funcs) {
-			if (tg->funcs->set_drr)
-				tg->funcs->set_drr(tg, &params);
+			set_drr_and_clear_adjust_pending(pipe_ctx[i], pipe_ctx[i]->stream, &params);
 			if (adjust.v_total_max != 0 && adjust.v_total_min != 0)
 				if (tg->funcs->set_static_screen_control)
 					tg->funcs->set_static_screen_control(
diff --git a/drivers/gpu/drm/amd/display/dc/hwss/dcn10/dcn10_hwseq.c b/drivers/gpu/drm/amd/display/dc/hwss/dcn10/dcn10_hwseq.c
index 13f9e9b439f6a..4c89bf6725b3b 100644
--- a/drivers/gpu/drm/amd/display/dc/hwss/dcn10/dcn10_hwseq.c
+++ b/drivers/gpu/drm/amd/display/dc/hwss/dcn10/dcn10_hwseq.c
@@ -1112,9 +1112,7 @@ static void dcn10_reset_back_end_for_pipe(
 		pipe_ctx->stream_res.tg->funcs->disable_crtc(pipe_ctx->stream_res.tg);
 
 		pipe_ctx->stream_res.tg->funcs->enable_optc_clock(pipe_ctx->stream_res.tg, false);
-		if (pipe_ctx->stream_res.tg->funcs->set_drr)
-			pipe_ctx->stream_res.tg->funcs->set_drr(
-					pipe_ctx->stream_res.tg, NULL);
+		set_drr_and_clear_adjust_pending(pipe_ctx, pipe_ctx->stream, NULL);
 		if (dc_is_hdmi_tmds_signal(pipe_ctx->stream->signal))
 			pipe_ctx->stream->link->phy_state.symclk_ref_cnts.otg = 0;
 	}
@@ -3217,8 +3215,7 @@ void dcn10_set_drr(struct pipe_ctx **pipe_ctx,
 		struct timing_generator *tg = pipe_ctx[i]->stream_res.tg;
 
 		if ((tg != NULL) && tg->funcs) {
-			if (tg->funcs->set_drr)
-				tg->funcs->set_drr(tg, &params);
+			set_drr_and_clear_adjust_pending(pipe_ctx[i], pipe_ctx[i]->stream, &params);
 			if (adjust.v_total_max != 0 && adjust.v_total_min != 0)
 				if (tg->funcs->set_static_screen_control)
 					tg->funcs->set_static_screen_control(
diff --git a/drivers/gpu/drm/amd/display/dc/hwss/dcn20/dcn20_hwseq.c b/drivers/gpu/drm/amd/display/dc/hwss/dcn20/dcn20_hwseq.c
index b78096a7690ee..f7e31aec32058 100644
--- a/drivers/gpu/drm/amd/display/dc/hwss/dcn20/dcn20_hwseq.c
+++ b/drivers/gpu/drm/amd/display/dc/hwss/dcn20/dcn20_hwseq.c
@@ -952,9 +952,7 @@ enum dc_status dcn20_enable_stream_timing(
 	params.vertical_total_max = stream->adjust.v_total_max;
 	params.vertical_total_mid = stream->adjust.v_total_mid;
 	params.vertical_total_mid_frame_num = stream->adjust.v_total_mid_frame_num;
-	if (pipe_ctx->stream_res.tg->funcs->set_drr)
-		pipe_ctx->stream_res.tg->funcs->set_drr(
-			pipe_ctx->stream_res.tg, &params);
+	set_drr_and_clear_adjust_pending(pipe_ctx, stream, &params);
 
 	// DRR should set trigger event to monitor surface update event
 	if (stream->adjust.v_total_min != 0 && stream->adjust.v_total_max != 0)
@@ -2849,9 +2847,7 @@ void dcn20_reset_back_end_for_pipe(
 			pipe_ctx->stream_res.tg->funcs->set_odm_bypass(
 					pipe_ctx->stream_res.tg, &pipe_ctx->stream->timing);
 
-		if (pipe_ctx->stream_res.tg->funcs->set_drr)
-			pipe_ctx->stream_res.tg->funcs->set_drr(
-					pipe_ctx->stream_res.tg, NULL);
+		set_drr_and_clear_adjust_pending(pipe_ctx, pipe_ctx->stream, NULL);
 		/* TODO - convert symclk_ref_cnts for otg to a bit map to solve
 		 * the case where the same symclk is shared across multiple otg
 		 * instances
diff --git a/drivers/gpu/drm/amd/display/dc/hwss/dcn31/dcn31_hwseq.c b/drivers/gpu/drm/amd/display/dc/hwss/dcn31/dcn31_hwseq.c
index 03ba01f4ace18..38f8898266971 100644
--- a/drivers/gpu/drm/amd/display/dc/hwss/dcn31/dcn31_hwseq.c
+++ b/drivers/gpu/drm/amd/display/dc/hwss/dcn31/dcn31_hwseq.c
@@ -538,9 +538,7 @@ static void dcn31_reset_back_end_for_pipe(
 	if (dc_is_hdmi_tmds_signal(pipe_ctx->stream->signal))
 		pipe_ctx->stream->link->phy_state.symclk_ref_cnts.otg = 0;
 
-	if (pipe_ctx->stream_res.tg->funcs->set_drr)
-		pipe_ctx->stream_res.tg->funcs->set_drr(
-				pipe_ctx->stream_res.tg, NULL);
+	set_drr_and_clear_adjust_pending(pipe_ctx, pipe_ctx->stream, NULL);
 
 	/* DPMS may already disable or */
 	/* dpms_off status is incorrect due to fastboot
diff --git a/drivers/gpu/drm/amd/display/dc/hwss/dcn35/dcn35_hwseq.c b/drivers/gpu/drm/amd/display/dc/hwss/dcn35/dcn35_hwseq.c
index b907ad1acedd9..922b8d71cf1aa 100644
--- a/drivers/gpu/drm/amd/display/dc/hwss/dcn35/dcn35_hwseq.c
+++ b/drivers/gpu/drm/amd/display/dc/hwss/dcn35/dcn35_hwseq.c
@@ -1473,8 +1473,7 @@ void dcn35_set_drr(struct pipe_ctx **pipe_ctx,
 					num_frames = 2 * (frame_rate % 60);
 				}
 			}
-			if (tg->funcs->set_drr)
-				tg->funcs->set_drr(tg, &params);
+			set_drr_and_clear_adjust_pending(pipe_ctx[i], pipe_ctx[i]->stream, &params);
 			if (adjust.v_total_max != 0 && adjust.v_total_min != 0)
 				if (tg->funcs->set_static_screen_control)
 					tg->funcs->set_static_screen_control(
diff --git a/drivers/gpu/drm/amd/display/dc/hwss/dcn401/dcn401_hwseq.c b/drivers/gpu/drm/amd/display/dc/hwss/dcn401/dcn401_hwseq.c
index 89af3e4afbc25..f21e2f7449517 100644
--- a/drivers/gpu/drm/amd/display/dc/hwss/dcn401/dcn401_hwseq.c
+++ b/drivers/gpu/drm/amd/display/dc/hwss/dcn401/dcn401_hwseq.c
@@ -830,10 +830,7 @@ enum dc_status dcn401_enable_stream_timing(
 	}
 
 	hws->funcs.wait_for_blank_complete(pipe_ctx->stream_res.opp);
-
-	if (pipe_ctx->stream_res.tg->funcs->set_drr)
-		pipe_ctx->stream_res.tg->funcs->set_drr(
-			pipe_ctx->stream_res.tg, &params);
+	set_drr_and_clear_adjust_pending(pipe_ctx, stream, &params);
 
 	/* Event triggers and num frames initialized for DRR, but can be
 	 * later updated for PSR use. Note DRR trigger events are generated
@@ -1866,9 +1863,8 @@ void dcn401_reset_back_end_for_pipe(
 			pipe_ctx->stream_res.tg->funcs->set_odm_bypass(
 					pipe_ctx->stream_res.tg, &pipe_ctx->stream->timing);
 
-		if (pipe_ctx->stream_res.tg->funcs->set_drr)
-			pipe_ctx->stream_res.tg->funcs->set_drr(
-					pipe_ctx->stream_res.tg, NULL);
+		set_drr_and_clear_adjust_pending(pipe_ctx, pipe_ctx->stream, NULL);
+
 		/* TODO - convert symclk_ref_cnts for otg to a bit map to solve
 		 * the case where the same symclk is shared across multiple otg
 		 * instances
diff --git a/drivers/gpu/drm/amd/display/dc/hwss/hw_sequencer.h b/drivers/gpu/drm/amd/display/dc/hwss/hw_sequencer.h
index a7d66cfd93c91..16ef5250a02e1 100644
--- a/drivers/gpu/drm/amd/display/dc/hwss/hw_sequencer.h
+++ b/drivers/gpu/drm/amd/display/dc/hwss/hw_sequencer.h
@@ -46,6 +46,7 @@ struct dce_hwseq;
 struct link_resource;
 struct dc_dmub_cmd;
 struct pg_block_update;
+struct drr_params;
 
 struct subvp_pipe_control_lock_fast_params {
 	struct dc *dc;
@@ -521,6 +522,11 @@ void set_p_state_switch_method(
 		struct dc_state *context,
 		struct pipe_ctx *pipe_ctx);
 
+void set_drr_and_clear_adjust_pending(
+		struct pipe_ctx *pipe_ctx,
+		struct dc_stream_state *stream,
+		struct drr_params *params);
+
 void hwss_execute_sequence(struct dc *dc,
 		struct block_sequence block_sequence[],
 		int num_steps);
-- 
2.39.5

[PATCH AUTOSEL 6.14 198/642] drm/amd/display: Fix incorrect DPCD configs while Replay/PSR switch

[Sasha Levin]

From: Leon Huang <Leon.Huang1@amd.com>

[ Upstream commit 0d9cabc8f591ea1cd97c071b853b75b155c13259 ]

[Why]
When switching between PSR/Replay,
the DPCD config of previous mode is not cleared,
resulting in unexpected behavior in TCON.

[How]
Initialize the DPCD in setup function

Reviewed-by: Robin Chen <robin.chen@amd.com>
Signed-off-by: Leon Huang <Leon.Huang1@amd.com>
Signed-off-by: Tom Chung <chiahsuan.chung@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 .../link/protocols/link_edp_panel_control.c   | 25 ++++++++++++++++---
 1 file changed, 22 insertions(+), 3 deletions(-)

diff --git a/drivers/gpu/drm/amd/display/dc/link/protocols/link_edp_panel_control.c b/drivers/gpu/drm/amd/display/dc/link/protocols/link_edp_panel_control.c
index e0e3bb8653595..1e4adbc764ea6 100644
--- a/drivers/gpu/drm/amd/display/dc/link/protocols/link_edp_panel_control.c
+++ b/drivers/gpu/drm/amd/display/dc/link/protocols/link_edp_panel_control.c
@@ -675,6 +675,18 @@ bool edp_setup_psr(struct dc_link *link,
 	if (!link)
 		return false;
 
+	//Clear PSR cfg
+	memset(&psr_configuration, 0, sizeof(psr_configuration));
+	dm_helpers_dp_write_dpcd(
+		link->ctx,
+		link,
+		DP_PSR_EN_CFG,
+		&psr_configuration.raw,
+		sizeof(psr_configuration.raw));
+
+	if (link->psr_settings.psr_version == DC_PSR_VERSION_UNSUPPORTED)
+		return false;
+
 	dc = link->ctx->dc;
 	dmcu = dc->res_pool->dmcu;
 	psr = dc->res_pool->psr;
@@ -685,9 +697,6 @@ bool edp_setup_psr(struct dc_link *link,
 	if (!dc_get_edp_link_panel_inst(dc, link, &panel_inst))
 		return false;
 
-
-	memset(&psr_configuration, 0, sizeof(psr_configuration));
-
 	psr_configuration.bits.ENABLE                    = 1;
 	psr_configuration.bits.CRC_VERIFICATION          = 1;
 	psr_configuration.bits.FRAME_CAPTURE_INDICATION  =
@@ -950,6 +959,16 @@ bool edp_setup_replay(struct dc_link *link, const struct dc_stream_state *stream
 	if (!link)
 		return false;
 
+	//Clear Replay config
+	dm_helpers_dp_write_dpcd(link->ctx, link,
+		DP_SINK_PR_ENABLE_AND_CONFIGURATION,
+		(uint8_t *)&(replay_config.raw), sizeof(uint8_t));
+
+	if (!(link->replay_settings.config.replay_supported))
+		return false;
+
+	link->replay_settings.config.replay_error_status.raw = 0;
+
 	dc = link->ctx->dc;
 
 	replay = dc->res_pool->replay;
-- 
2.39.5

[PATCH AUTOSEL 6.14 218/642] drm/amdkfd: Set per-process flags only once for gfx9/10/11/12

[Sasha Levin]

From: Harish Kasiviswanathan <Harish.Kasiviswanathan@amd.com>

[ Upstream commit 61972cd93af70738a6ad7f93e17cc7f68a01e182 ]

Define set_cache_memory_policy() for these asics and move all static
changes from update_qpd() which is called each time a queue is created
to set_cache_memory_policy() which is called once during process
initialization

Signed-off-by: Harish Kasiviswanathan <Harish.Kasiviswanathan@amd.com>
Reviewed-by: Amber Lin <Amber.Lin@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 .../amd/amdkfd/kfd_device_queue_manager_v10.c | 41 +++++++++++--------
 .../amd/amdkfd/kfd_device_queue_manager_v11.c | 41 +++++++++++--------
 .../amd/amdkfd/kfd_device_queue_manager_v12.c | 41 +++++++++++--------
 .../amd/amdkfd/kfd_device_queue_manager_v9.c  | 36 +++++++++++++++-
 4 files changed, 107 insertions(+), 52 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager_v10.c b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager_v10.c
index 245a90dfc2f6b..b5f5f141353b5 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager_v10.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager_v10.c
@@ -31,10 +31,17 @@ static int update_qpd_v10(struct device_queue_manager *dqm,
 			 struct qcm_process_device *qpd);
 static void init_sdma_vm_v10(struct device_queue_manager *dqm, struct queue *q,
 			    struct qcm_process_device *qpd);
+static bool set_cache_memory_policy_v10(struct device_queue_manager *dqm,
+				   struct qcm_process_device *qpd,
+				   enum cache_policy default_policy,
+				   enum cache_policy alternate_policy,
+				   void __user *alternate_aperture_base,
+				   uint64_t alternate_aperture_size);
 
 void device_queue_manager_init_v10(
 	struct device_queue_manager_asic_ops *asic_ops)
 {
+	asic_ops->set_cache_memory_policy = set_cache_memory_policy_v10;
 	asic_ops->update_qpd = update_qpd_v10;
 	asic_ops->init_sdma_vm = init_sdma_vm_v10;
 	asic_ops->mqd_manager_init = mqd_manager_init_v10;
@@ -49,27 +56,27 @@ static uint32_t compute_sh_mem_bases_64bit(struct kfd_process_device *pdd)
 		private_base;
 }
 
-static int update_qpd_v10(struct device_queue_manager *dqm,
-			 struct qcm_process_device *qpd)
+static bool set_cache_memory_policy_v10(struct device_queue_manager *dqm,
+				   struct qcm_process_device *qpd,
+				   enum cache_policy default_policy,
+				   enum cache_policy alternate_policy,
+				   void __user *alternate_aperture_base,
+				   uint64_t alternate_aperture_size)
 {
-	struct kfd_process_device *pdd;
-
-	pdd = qpd_to_pdd(qpd);
-
-	/* check if sh_mem_config register already configured */
-	if (qpd->sh_mem_config == 0) {
-		qpd->sh_mem_config =
-			(SH_MEM_ALIGNMENT_MODE_UNALIGNED <<
-				SH_MEM_CONFIG__ALIGNMENT_MODE__SHIFT) |
-			(3 << SH_MEM_CONFIG__INITIAL_INST_PREFETCH__SHIFT);
-		qpd->sh_mem_ape1_limit = 0;
-		qpd->sh_mem_ape1_base = 0;
-	}
-
-	qpd->sh_mem_bases = compute_sh_mem_bases_64bit(pdd);
+	qpd->sh_mem_config = (SH_MEM_ALIGNMENT_MODE_UNALIGNED <<
+			      SH_MEM_CONFIG__ALIGNMENT_MODE__SHIFT) |
+			      (3 << SH_MEM_CONFIG__INITIAL_INST_PREFETCH__SHIFT);
+	qpd->sh_mem_ape1_limit = 0;
+	qpd->sh_mem_ape1_base = 0;
+	qpd->sh_mem_bases = compute_sh_mem_bases_64bit(qpd_to_pdd(qpd));
 
 	pr_debug("sh_mem_bases 0x%X\n", qpd->sh_mem_bases);
+	return true;
+}
 
+static int update_qpd_v10(struct device_queue_manager *dqm,
+			 struct qcm_process_device *qpd)
+{
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager_v11.c b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager_v11.c
index 2e129da7acb43..f436878d0d621 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager_v11.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager_v11.c
@@ -30,10 +30,17 @@ static int update_qpd_v11(struct device_queue_manager *dqm,
 			 struct qcm_process_device *qpd);
 static void init_sdma_vm_v11(struct device_queue_manager *dqm, struct queue *q,
 			    struct qcm_process_device *qpd);
+static bool set_cache_memory_policy_v11(struct device_queue_manager *dqm,
+				   struct qcm_process_device *qpd,
+				   enum cache_policy default_policy,
+				   enum cache_policy alternate_policy,
+				   void __user *alternate_aperture_base,
+				   uint64_t alternate_aperture_size);
 
 void device_queue_manager_init_v11(
 	struct device_queue_manager_asic_ops *asic_ops)
 {
+	asic_ops->set_cache_memory_policy = set_cache_memory_policy_v11;
 	asic_ops->update_qpd = update_qpd_v11;
 	asic_ops->init_sdma_vm = init_sdma_vm_v11;
 	asic_ops->mqd_manager_init = mqd_manager_init_v11;
@@ -48,28 +55,28 @@ static uint32_t compute_sh_mem_bases_64bit(struct kfd_process_device *pdd)
 		private_base;
 }
 
-static int update_qpd_v11(struct device_queue_manager *dqm,
-			 struct qcm_process_device *qpd)
+static bool set_cache_memory_policy_v11(struct device_queue_manager *dqm,
+				   struct qcm_process_device *qpd,
+				   enum cache_policy default_policy,
+				   enum cache_policy alternate_policy,
+				   void __user *alternate_aperture_base,
+				   uint64_t alternate_aperture_size)
 {
-	struct kfd_process_device *pdd;
-
-	pdd = qpd_to_pdd(qpd);
-
-	/* check if sh_mem_config register already configured */
-	if (qpd->sh_mem_config == 0) {
-		qpd->sh_mem_config =
-			(SH_MEM_ALIGNMENT_MODE_UNALIGNED <<
-				SH_MEM_CONFIG__ALIGNMENT_MODE__SHIFT) |
-			(3 << SH_MEM_CONFIG__INITIAL_INST_PREFETCH__SHIFT);
-
-		qpd->sh_mem_ape1_limit = 0;
-		qpd->sh_mem_ape1_base = 0;
-	}
+	qpd->sh_mem_config = (SH_MEM_ALIGNMENT_MODE_UNALIGNED <<
+			      SH_MEM_CONFIG__ALIGNMENT_MODE__SHIFT) |
+			      (3 << SH_MEM_CONFIG__INITIAL_INST_PREFETCH__SHIFT);
 
-	qpd->sh_mem_bases = compute_sh_mem_bases_64bit(pdd);
+	qpd->sh_mem_ape1_limit = 0;
+	qpd->sh_mem_ape1_base = 0;
+	qpd->sh_mem_bases = compute_sh_mem_bases_64bit(qpd_to_pdd(qpd));
 
 	pr_debug("sh_mem_bases 0x%X\n", qpd->sh_mem_bases);
+	return true;
+}
 
+static int update_qpd_v11(struct device_queue_manager *dqm,
+			 struct qcm_process_device *qpd)
+{
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager_v12.c b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager_v12.c
index 4f3295b29dfb1..62ca1c8fcbaf9 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager_v12.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager_v12.c
@@ -30,10 +30,17 @@ static int update_qpd_v12(struct device_queue_manager *dqm,
 			 struct qcm_process_device *qpd);
 static void init_sdma_vm_v12(struct device_queue_manager *dqm, struct queue *q,
 			    struct qcm_process_device *qpd);
+static bool set_cache_memory_policy_v12(struct device_queue_manager *dqm,
+				   struct qcm_process_device *qpd,
+				   enum cache_policy default_policy,
+				   enum cache_policy alternate_policy,
+				   void __user *alternate_aperture_base,
+				   uint64_t alternate_aperture_size);
 
 void device_queue_manager_init_v12(
 	struct device_queue_manager_asic_ops *asic_ops)
 {
+	asic_ops->set_cache_memory_policy = set_cache_memory_policy_v12;
 	asic_ops->update_qpd = update_qpd_v12;
 	asic_ops->init_sdma_vm = init_sdma_vm_v12;
 	asic_ops->mqd_manager_init = mqd_manager_init_v12;
@@ -48,28 +55,28 @@ static uint32_t compute_sh_mem_bases_64bit(struct kfd_process_device *pdd)
 		private_base;
 }
 
-static int update_qpd_v12(struct device_queue_manager *dqm,
-			 struct qcm_process_device *qpd)
+static bool set_cache_memory_policy_v12(struct device_queue_manager *dqm,
+				   struct qcm_process_device *qpd,
+				   enum cache_policy default_policy,
+				   enum cache_policy alternate_policy,
+				   void __user *alternate_aperture_base,
+				   uint64_t alternate_aperture_size)
 {
-	struct kfd_process_device *pdd;
-
-	pdd = qpd_to_pdd(qpd);
-
-	/* check if sh_mem_config register already configured */
-	if (qpd->sh_mem_config == 0) {
-		qpd->sh_mem_config =
-			(SH_MEM_ALIGNMENT_MODE_UNALIGNED <<
-				SH_MEM_CONFIG__ALIGNMENT_MODE__SHIFT) |
-			(3 << SH_MEM_CONFIG__INITIAL_INST_PREFETCH__SHIFT);
-
-		qpd->sh_mem_ape1_limit = 0;
-		qpd->sh_mem_ape1_base = 0;
-	}
+	qpd->sh_mem_config = (SH_MEM_ALIGNMENT_MODE_UNALIGNED <<
+			      SH_MEM_CONFIG__ALIGNMENT_MODE__SHIFT) |
+			      (3 << SH_MEM_CONFIG__INITIAL_INST_PREFETCH__SHIFT);
 
-	qpd->sh_mem_bases = compute_sh_mem_bases_64bit(pdd);
+	qpd->sh_mem_ape1_limit = 0;
+	qpd->sh_mem_ape1_base = 0;
+	qpd->sh_mem_bases = compute_sh_mem_bases_64bit(qpd_to_pdd(qpd));
 
 	pr_debug("sh_mem_bases 0x%X\n", qpd->sh_mem_bases);
+	return true;
+}
 
+static int update_qpd_v12(struct device_queue_manager *dqm,
+			 struct qcm_process_device *qpd)
+{
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager_v9.c b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager_v9.c
index 67137e674f1d0..c734eb9b505f8 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager_v9.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager_v9.c
@@ -30,10 +30,17 @@ static int update_qpd_v9(struct device_queue_manager *dqm,
 			 struct qcm_process_device *qpd);
 static void init_sdma_vm_v9(struct device_queue_manager *dqm, struct queue *q,
 			    struct qcm_process_device *qpd);
+static bool set_cache_memory_policy_v9(struct device_queue_manager *dqm,
+				   struct qcm_process_device *qpd,
+				   enum cache_policy default_policy,
+				   enum cache_policy alternate_policy,
+				   void __user *alternate_aperture_base,
+				   uint64_t alternate_aperture_size);
 
 void device_queue_manager_init_v9(
 	struct device_queue_manager_asic_ops *asic_ops)
 {
+	asic_ops->set_cache_memory_policy = set_cache_memory_policy_v9;
 	asic_ops->update_qpd = update_qpd_v9;
 	asic_ops->init_sdma_vm = init_sdma_vm_v9;
 	asic_ops->mqd_manager_init = mqd_manager_init_v9;
@@ -48,10 +55,37 @@ static uint32_t compute_sh_mem_bases_64bit(struct kfd_process_device *pdd)
 		private_base;
 }
 
+static bool set_cache_memory_policy_v9(struct device_queue_manager *dqm,
+				   struct qcm_process_device *qpd,
+				   enum cache_policy default_policy,
+				   enum cache_policy alternate_policy,
+				   void __user *alternate_aperture_base,
+				   uint64_t alternate_aperture_size)
+{
+	qpd->sh_mem_config = SH_MEM_ALIGNMENT_MODE_UNALIGNED <<
+				SH_MEM_CONFIG__ALIGNMENT_MODE__SHIFT;
+
+	if (dqm->dev->kfd->noretry)
+		qpd->sh_mem_config |= 1 << SH_MEM_CONFIG__RETRY_DISABLE__SHIFT;
+
+	if (KFD_GC_VERSION(dqm->dev->kfd) == IP_VERSION(9, 4, 3) ||
+		KFD_GC_VERSION(dqm->dev->kfd) == IP_VERSION(9, 4, 4) ||
+		KFD_GC_VERSION(dqm->dev->kfd) == IP_VERSION(9, 5, 0))
+		qpd->sh_mem_config |= (1 << SH_MEM_CONFIG__F8_MODE__SHIFT);
+
+	qpd->sh_mem_ape1_limit = 0;
+	qpd->sh_mem_ape1_base = 0;
+	qpd->sh_mem_bases = compute_sh_mem_bases_64bit(qpd_to_pdd(qpd));
+
+	pr_debug("sh_mem_bases 0x%X sh_mem_config 0x%X\n", qpd->sh_mem_bases,
+		 qpd->sh_mem_config);
+	return true;
+}
+
 static int update_qpd_v9(struct device_queue_manager *dqm,
 			 struct qcm_process_device *qpd)
 {
-	struct kfd_process_device *pdd;
+	struct kfd_process_device *pdd = qpd_to_pdd(qpd);
 
 	pdd = qpd_to_pdd(qpd);
 
-- 
2.39.5

[PATCH AUTOSEL 6.14 219/642] drm/amdkfd: Set per-process flags only once cik/vi

[Sasha Levin]

From: Harish Kasiviswanathan <Harish.Kasiviswanathan@amd.com>

[ Upstream commit 289e68503a4533b014f8447e2af28ad44c92c221 ]

Set per-process static sh_mem config only once during process
initialization. Move all static changes from update_qpd() which is
called each time a queue is created to set_cache_memory_policy() which
is called once during process initialization.

set_cache_memory_policy() is currently defined only for cik and vi
family. So this commit only focuses on these two. A separate commit will
address other asics.

Signed-off-by: Harish Kasiviswanathan <Harish.Kasiviswanathan@amd.com>
Reviewed-by: Amber Lin <Amber.Lin@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 .../drm/amd/amdkfd/kfd_device_queue_manager.c | 39 +---------
 .../amd/amdkfd/kfd_device_queue_manager_cik.c | 69 ++++++++++++------
 .../amd/amdkfd/kfd_device_queue_manager_vi.c  | 71 ++++++++++++-------
 3 files changed, 94 insertions(+), 85 deletions(-)

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 ad9cb50a9fa38..7f8ec2a152ac6 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c
@@ -2491,14 +2491,6 @@ static int destroy_queue_cpsch(struct device_queue_manager *dqm,
 	return retval;
 }
 
-/*
- * Low bits must be 0000/FFFF as required by HW, high bits must be 0 to
- * stay in user mode.
- */
-#define APE1_FIXED_BITS_MASK 0xFFFF80000000FFFFULL
-/* APE1 limit is inclusive and 64K aligned. */
-#define APE1_LIMIT_ALIGNMENT 0xFFFF
-
 static bool set_cache_memory_policy(struct device_queue_manager *dqm,
 				   struct qcm_process_device *qpd,
 				   enum cache_policy default_policy,
@@ -2513,34 +2505,6 @@ static bool set_cache_memory_policy(struct device_queue_manager *dqm,
 
 	dqm_lock(dqm);
 
-	if (alternate_aperture_size == 0) {
-		/* base > limit disables APE1 */
-		qpd->sh_mem_ape1_base = 1;
-		qpd->sh_mem_ape1_limit = 0;
-	} else {
-		/*
-		 * In FSA64, APE1_Base[63:0] = { 16{SH_MEM_APE1_BASE[31]},
-		 *			SH_MEM_APE1_BASE[31:0], 0x0000 }
-		 * APE1_Limit[63:0] = { 16{SH_MEM_APE1_LIMIT[31]},
-		 *			SH_MEM_APE1_LIMIT[31:0], 0xFFFF }
-		 * Verify that the base and size parameters can be
-		 * represented in this format and convert them.
-		 * Additionally restrict APE1 to user-mode addresses.
-		 */
-
-		uint64_t base = (uintptr_t)alternate_aperture_base;
-		uint64_t limit = base + alternate_aperture_size - 1;
-
-		if (limit <= base || (base & APE1_FIXED_BITS_MASK) != 0 ||
-		   (limit & APE1_FIXED_BITS_MASK) != APE1_LIMIT_ALIGNMENT) {
-			retval = false;
-			goto out;
-		}
-
-		qpd->sh_mem_ape1_base = base >> 16;
-		qpd->sh_mem_ape1_limit = limit >> 16;
-	}
-
 	retval = dqm->asic_ops.set_cache_memory_policy(
 			dqm,
 			qpd,
@@ -2549,6 +2513,9 @@ static bool set_cache_memory_policy(struct device_queue_manager *dqm,
 			alternate_aperture_base,
 			alternate_aperture_size);
 
+	if (retval)
+		goto out;
+
 	if ((dqm->sched_policy == KFD_SCHED_POLICY_NO_HWS) && (qpd->vmid != 0))
 		program_sh_mem_settings(dqm, qpd);
 
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager_cik.c b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager_cik.c
index d4d95c7f2e5d4..32bedef912b3b 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager_cik.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager_cik.c
@@ -27,6 +27,14 @@
 #include "oss/oss_2_4_sh_mask.h"
 #include "gca/gfx_7_2_sh_mask.h"
 
+/*
+ * Low bits must be 0000/FFFF as required by HW, high bits must be 0 to
+ * stay in user mode.
+ */
+#define APE1_FIXED_BITS_MASK 0xFFFF80000000FFFFULL
+/* APE1 limit is inclusive and 64K aligned. */
+#define APE1_LIMIT_ALIGNMENT 0xFFFF
+
 static bool set_cache_memory_policy_cik(struct device_queue_manager *dqm,
 				   struct qcm_process_device *qpd,
 				   enum cache_policy default_policy,
@@ -84,6 +92,36 @@ static bool set_cache_memory_policy_cik(struct device_queue_manager *dqm,
 {
 	uint32_t default_mtype;
 	uint32_t ape1_mtype;
+	unsigned int temp;
+	bool retval = true;
+
+	if (alternate_aperture_size == 0) {
+		/* base > limit disables APE1 */
+		qpd->sh_mem_ape1_base = 1;
+		qpd->sh_mem_ape1_limit = 0;
+	} else {
+		/*
+		 * In FSA64, APE1_Base[63:0] = { 16{SH_MEM_APE1_BASE[31]},
+		 *			SH_MEM_APE1_BASE[31:0], 0x0000 }
+		 * APE1_Limit[63:0] = { 16{SH_MEM_APE1_LIMIT[31]},
+		 *			SH_MEM_APE1_LIMIT[31:0], 0xFFFF }
+		 * Verify that the base and size parameters can be
+		 * represented in this format and convert them.
+		 * Additionally restrict APE1 to user-mode addresses.
+		 */
+
+		uint64_t base = (uintptr_t)alternate_aperture_base;
+		uint64_t limit = base + alternate_aperture_size - 1;
+
+		if (limit <= base || (base & APE1_FIXED_BITS_MASK) != 0 ||
+		   (limit & APE1_FIXED_BITS_MASK) != APE1_LIMIT_ALIGNMENT) {
+			retval = false;
+			goto out;
+		}
+
+		qpd->sh_mem_ape1_base = base >> 16;
+		qpd->sh_mem_ape1_limit = limit >> 16;
+	}
 
 	default_mtype = (default_policy == cache_policy_coherent) ?
 			MTYPE_NONCACHED :
@@ -97,37 +135,22 @@ static bool set_cache_memory_policy_cik(struct device_queue_manager *dqm,
 			| ALIGNMENT_MODE(SH_MEM_ALIGNMENT_MODE_UNALIGNED)
 			| DEFAULT_MTYPE(default_mtype)
 			| APE1_MTYPE(ape1_mtype);
-
-	return true;
-}
-
-static int update_qpd_cik(struct device_queue_manager *dqm,
-			  struct qcm_process_device *qpd)
-{
-	struct kfd_process_device *pdd;
-	unsigned int temp;
-
-	pdd = qpd_to_pdd(qpd);
-
-	/* check if sh_mem_config register already configured */
-	if (qpd->sh_mem_config == 0) {
-		qpd->sh_mem_config =
-			ALIGNMENT_MODE(SH_MEM_ALIGNMENT_MODE_UNALIGNED) |
-			DEFAULT_MTYPE(MTYPE_NONCACHED) |
-			APE1_MTYPE(MTYPE_NONCACHED);
-		qpd->sh_mem_ape1_limit = 0;
-		qpd->sh_mem_ape1_base = 0;
-	}
-
 	/* On dGPU we're always in GPUVM64 addressing mode with 64-bit
 	 * aperture addresses.
 	 */
-	temp = get_sh_mem_bases_nybble_64(pdd);
+	temp = get_sh_mem_bases_nybble_64(qpd_to_pdd(qpd));
 	qpd->sh_mem_bases = compute_sh_mem_bases_64bit(temp);
 
 	pr_debug("is32bit process: %d sh_mem_bases nybble: 0x%X and register 0x%X\n",
 		qpd->pqm->process->is_32bit_user_mode, temp, qpd->sh_mem_bases);
 
+out:
+	return retval;
+}
+
+static int update_qpd_cik(struct device_queue_manager *dqm,
+			  struct qcm_process_device *qpd)
+{
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager_vi.c b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager_vi.c
index b291ee0fab943..320518f418903 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager_vi.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager_vi.c
@@ -27,6 +27,14 @@
 #include "gca/gfx_8_0_sh_mask.h"
 #include "oss/oss_3_0_sh_mask.h"
 
+/*
+ * Low bits must be 0000/FFFF as required by HW, high bits must be 0 to
+ * stay in user mode.
+ */
+#define APE1_FIXED_BITS_MASK 0xFFFF80000000FFFFULL
+/* APE1 limit is inclusive and 64K aligned. */
+#define APE1_LIMIT_ALIGNMENT 0xFFFF
+
 static bool set_cache_memory_policy_vi(struct device_queue_manager *dqm,
 				       struct qcm_process_device *qpd,
 				       enum cache_policy default_policy,
@@ -85,6 +93,36 @@ static bool set_cache_memory_policy_vi(struct device_queue_manager *dqm,
 {
 	uint32_t default_mtype;
 	uint32_t ape1_mtype;
+	unsigned int temp;
+	bool retval = true;
+
+	if (alternate_aperture_size == 0) {
+		/* base > limit disables APE1 */
+		qpd->sh_mem_ape1_base = 1;
+		qpd->sh_mem_ape1_limit = 0;
+	} else {
+		/*
+		 * In FSA64, APE1_Base[63:0] = { 16{SH_MEM_APE1_BASE[31]},
+		 *			SH_MEM_APE1_BASE[31:0], 0x0000 }
+		 * APE1_Limit[63:0] = { 16{SH_MEM_APE1_LIMIT[31]},
+		 *			SH_MEM_APE1_LIMIT[31:0], 0xFFFF }
+		 * Verify that the base and size parameters can be
+		 * represented in this format and convert them.
+		 * Additionally restrict APE1 to user-mode addresses.
+		 */
+
+		uint64_t base = (uintptr_t)alternate_aperture_base;
+		uint64_t limit = base + alternate_aperture_size - 1;
+
+		if (limit <= base || (base & APE1_FIXED_BITS_MASK) != 0 ||
+		   (limit & APE1_FIXED_BITS_MASK) != APE1_LIMIT_ALIGNMENT) {
+			retval = false;
+			goto out;
+		}
+
+		qpd->sh_mem_ape1_base = base >> 16;
+		qpd->sh_mem_ape1_limit = limit >> 16;
+	}
 
 	default_mtype = (default_policy == cache_policy_coherent) ?
 			MTYPE_UC :
@@ -100,40 +138,21 @@ static bool set_cache_memory_policy_vi(struct device_queue_manager *dqm,
 			default_mtype << SH_MEM_CONFIG__DEFAULT_MTYPE__SHIFT |
 			ape1_mtype << SH_MEM_CONFIG__APE1_MTYPE__SHIFT;
 
-	return true;
-}
-
-static int update_qpd_vi(struct device_queue_manager *dqm,
-			 struct qcm_process_device *qpd)
-{
-	struct kfd_process_device *pdd;
-	unsigned int temp;
-
-	pdd = qpd_to_pdd(qpd);
-
-	/* check if sh_mem_config register already configured */
-	if (qpd->sh_mem_config == 0) {
-		qpd->sh_mem_config =
-				SH_MEM_ALIGNMENT_MODE_UNALIGNED <<
-					SH_MEM_CONFIG__ALIGNMENT_MODE__SHIFT |
-				MTYPE_UC <<
-					SH_MEM_CONFIG__DEFAULT_MTYPE__SHIFT |
-				MTYPE_UC <<
-					SH_MEM_CONFIG__APE1_MTYPE__SHIFT;
-
-		qpd->sh_mem_ape1_limit = 0;
-		qpd->sh_mem_ape1_base = 0;
-	}
-
 	/* On dGPU we're always in GPUVM64 addressing mode with 64-bit
 	 * aperture addresses.
 	 */
-	temp = get_sh_mem_bases_nybble_64(pdd);
+	temp = get_sh_mem_bases_nybble_64(qpd_to_pdd(qpd));
 	qpd->sh_mem_bases = compute_sh_mem_bases_64bit(temp);
 
 	pr_debug("sh_mem_bases nybble: 0x%X and register 0x%X\n",
 		temp, qpd->sh_mem_bases);
+out:
+	return retval;
+}
 
+static int update_qpd_vi(struct device_queue_manager *dqm,
+			 struct qcm_process_device *qpd)
+{
 	return 0;
 }
 
-- 
2.39.5

[PATCH AUTOSEL 6.14 220/642] drm/amdkfd: clear F8_MODE for gfx950

[Sasha Levin]

From: Alex Sierra <alex.sierra@amd.com>

[ Upstream commit 59228c6631f902fa826dc61321ab377ba8aadec5 ]

Default F8_MODE should be OCP format on gfx950.

Signed-off-by: Alex Sierra <alex.sierra@amd.com>
Reviewed-by: Harish Kasiviswanathan <Harish.Kasiviswanathan@amd.com>
Signed-off-by: Amber Lin <Amber.Lin@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager_v9.c | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager_v9.c b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager_v9.c
index c734eb9b505f8..3264509408bc8 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager_v9.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager_v9.c
@@ -98,8 +98,7 @@ static int update_qpd_v9(struct device_queue_manager *dqm,
 			qpd->sh_mem_config |= 1 << SH_MEM_CONFIG__RETRY_DISABLE__SHIFT;
 
 		if (KFD_GC_VERSION(dqm->dev->kfd) == IP_VERSION(9, 4, 3) ||
-		    KFD_GC_VERSION(dqm->dev->kfd) == IP_VERSION(9, 4, 4) ||
-		    KFD_GC_VERSION(dqm->dev->kfd) == IP_VERSION(9, 5, 0))
+		    KFD_GC_VERSION(dqm->dev->kfd) == IP_VERSION(9, 4, 4))
 			qpd->sh_mem_config |=
 				(1 << SH_MEM_CONFIG__F8_MODE__SHIFT);
 
-- 
2.39.5

[PATCH AUTOSEL 6.14 221/642] drm/amdgpu: increase RAS bad page threshold

[Sasha Levin]

From: Tao Zhou <tao.zhou1@amd.com>

[ Upstream commit 334dc5fcc3f177823115ec4e075259997c16d4a7 ]

For default policy, driver will issue an RMA event when the number of
bad pages is greater than 8 physical rows, rather than reaches 8
physical rows, don't rely on threshold configurable parameters in
default mode.

Signed-off-by: Tao Zhou <tao.zhou1@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.c
index 52c16bfeccaad..12ffe4a963d31 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.c
@@ -748,7 +748,7 @@ amdgpu_ras_eeprom_update_header(struct amdgpu_ras_eeprom_control *control)
 	/* Modify the header if it exceeds.
 	 */
 	if (amdgpu_bad_page_threshold != 0 &&
-	    control->ras_num_bad_pages >= ras->bad_page_cnt_threshold) {
+	    control->ras_num_bad_pages > ras->bad_page_cnt_threshold) {
 		dev_warn(adev->dev,
 			"Saved bad pages %d reaches threshold value %d\n",
 			control->ras_num_bad_pages, ras->bad_page_cnt_threshold);
@@ -806,7 +806,7 @@ amdgpu_ras_eeprom_update_header(struct amdgpu_ras_eeprom_control *control)
 	 */
 	if (amdgpu_bad_page_threshold != 0 &&
 	    control->tbl_hdr.version == RAS_TABLE_VER_V2_1 &&
-	    control->ras_num_bad_pages < ras->bad_page_cnt_threshold)
+	    control->ras_num_bad_pages <= ras->bad_page_cnt_threshold)
 		control->tbl_rai.health_percent = ((ras->bad_page_cnt_threshold -
 						   control->ras_num_bad_pages) * 100) /
 						   ras->bad_page_cnt_threshold;
@@ -1451,7 +1451,7 @@ int amdgpu_ras_eeprom_check(struct amdgpu_ras_eeprom_control *control)
 				  res);
 			return -EINVAL;
 		}
-		if (ras->bad_page_cnt_threshold > control->ras_num_bad_pages) {
+		if (ras->bad_page_cnt_threshold >= control->ras_num_bad_pages) {
 			/* This means that, the threshold was increased since
 			 * the last time the system was booted, and now,
 			 * ras->bad_page_cnt_threshold - control->num_recs > 0,
-- 
2.39.5

[PATCH AUTOSEL 6.14 222/642] drm/amdgpu: Fix missing drain retry fault the last entry

[Sasha Levin]

From: Emily Deng <Emily.Deng@amd.com>

[ Upstream commit fe2fa3be3d59ba67d6de54a0064441ec233cb50c ]

While the entry get in svm_range_unmap_from_cpu is the last entry, and
the entry is page fault, it also need to be dropped. So for equal case,
it also need to be dropped.

v2:
Only modify the svm_range_restore_pages.

Signed-off-by: Emily Deng <Emily.Deng@amd.com>
Reviewed-by: Xiaogang Chen<xiaogang.chen@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_ih.h | 3 +++
 drivers/gpu/drm/amd/amdkfd/kfd_svm.c   | 2 +-
 2 files changed, 4 insertions(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ih.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ih.h
index 7d4395a5d8ac9..b0a88f92cd821 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ih.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ih.h
@@ -78,6 +78,9 @@ struct amdgpu_ih_ring {
 #define amdgpu_ih_ts_after(t1, t2) \
 		(((int64_t)((t2) << 16) - (int64_t)((t1) << 16)) > 0LL)
 
+#define amdgpu_ih_ts_after_or_equal(t1, t2) \
+		(((int64_t)((t2) << 16) - (int64_t)((t1) << 16)) >= 0LL)
+
 /* provided by the ih block */
 struct amdgpu_ih_funcs {
 	/* ring read/write ptr handling, called from interrupt context */
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_svm.c b/drivers/gpu/drm/amd/amdkfd/kfd_svm.c
index d1cf9dd352904..47189453b20c3 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_svm.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_svm.c
@@ -3024,7 +3024,7 @@ svm_range_restore_pages(struct amdgpu_device *adev, unsigned int pasid,
 
 	/* check if this page fault time stamp is before svms->checkpoint_ts */
 	if (svms->checkpoint_ts[gpuidx] != 0) {
-		if (amdgpu_ih_ts_after(ts,  svms->checkpoint_ts[gpuidx])) {
+		if (amdgpu_ih_ts_after_or_equal(ts,  svms->checkpoint_ts[gpuidx])) {
 			pr_debug("draining retry fault, drop fault 0x%llx\n", addr);
 			r = -EAGAIN;
 			goto out_unlock_svms;
-- 
2.39.5

[PATCH AUTOSEL 6.14 246/642] drm/amdgpu: Do not program AGP BAR regs under SRIOV in gfxhub_v1_0.c

[Sasha Levin]

From: Victor Lu <victorchengchi.lu@amd.com>

[ Upstream commit 057fef20b8401110a7bc1c2fe9d804a8a0bf0d24 ]

SRIOV VF does not have write access to AGP BAR regs.
Skip the writes to avoid a dmesg warning.

Signed-off-by: Victor Lu <victorchengchi.lu@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/gpu/drm/amd/amdgpu/gfxhub_v1_0.c | 10 +++++-----
 1 file changed, 5 insertions(+), 5 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/gfxhub_v1_0.c b/drivers/gpu/drm/amd/amdgpu/gfxhub_v1_0.c
index 0e3ddea7b8e0f..a7bfc9f41d0e3 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfxhub_v1_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfxhub_v1_0.c
@@ -92,12 +92,12 @@ static void gfxhub_v1_0_init_system_aperture_regs(struct amdgpu_device *adev)
 {
 	uint64_t value;
 
-	/* Program the AGP BAR */
-	WREG32_SOC15_RLC(GC, 0, mmMC_VM_AGP_BASE, 0);
-	WREG32_SOC15_RLC(GC, 0, mmMC_VM_AGP_BOT, adev->gmc.agp_start >> 24);
-	WREG32_SOC15_RLC(GC, 0, mmMC_VM_AGP_TOP, adev->gmc.agp_end >> 24);
-
 	if (!amdgpu_sriov_vf(adev) || adev->asic_type <= CHIP_VEGA10) {
+		/* Program the AGP BAR */
+		WREG32_SOC15_RLC(GC, 0, mmMC_VM_AGP_BASE, 0);
+		WREG32_SOC15_RLC(GC, 0, mmMC_VM_AGP_BOT, adev->gmc.agp_start >> 24);
+		WREG32_SOC15_RLC(GC, 0, mmMC_VM_AGP_TOP, adev->gmc.agp_end >> 24);
+
 		/* Program the system aperture low logical page number. */
 		WREG32_SOC15_RLC(GC, 0, mmMC_VM_SYSTEM_APERTURE_LOW_ADDR,
 			min(adev->gmc.fb_start, adev->gmc.agp_start) >> 18);
-- 
2.39.5

[PATCH AUTOSEL 6.14 247/642] drm/amdgpu: Reinit FW shared flags on VCN v5.0.1

[Sasha Levin]

From: Lijo Lazar <lijo.lazar@amd.com>

[ Upstream commit 6ef5ccaad76d907d4257f20de992f89c0f7a7f8e ]

After a full device reset, shared memory region will clear out and it's
not possible to reliably save the region in case of RAS errors.
Reinitialize the flags if required.

Signed-off-by: Lijo Lazar <lijo.lazar@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/gpu/drm/amd/amdgpu/vcn_v5_0_1.c | 28 ++++++++++++++++++-------
 1 file changed, 20 insertions(+), 8 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_1.c b/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_1.c
index 8b0b3739a5377..cdbc10d7c9fb7 100644
--- a/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_1.c
+++ b/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_1.c
@@ -65,6 +65,22 @@ static int vcn_v5_0_1_early_init(struct amdgpu_ip_block *ip_block)
 	return amdgpu_vcn_early_init(adev);
 }
 
+static void vcn_v5_0_1_fw_shared_init(struct amdgpu_device *adev, int inst_idx)
+{
+	struct amdgpu_vcn5_fw_shared *fw_shared;
+
+	fw_shared = adev->vcn.inst[inst_idx].fw_shared.cpu_addr;
+
+	if (fw_shared->sq.is_enabled)
+		return;
+	fw_shared->present_flag_0 =
+		cpu_to_le32(AMDGPU_FW_SHARED_FLAG_0_UNIFIED_QUEUE);
+	fw_shared->sq.is_enabled = 1;
+
+	if (amdgpu_vcnfw_log)
+		amdgpu_vcn_fwlog_init(&adev->vcn.inst[inst_idx]);
+}
+
 /**
  * vcn_v5_0_1_sw_init - sw init for VCN block
  *
@@ -95,8 +111,6 @@ static int vcn_v5_0_1_sw_init(struct amdgpu_ip_block *ip_block)
 		return r;
 
 	for (i = 0; i < adev->vcn.num_vcn_inst; i++) {
-		volatile struct amdgpu_vcn5_fw_shared *fw_shared;
-
 		vcn_inst = GET_INST(VCN, i);
 
 		ring = &adev->vcn.inst[i].ring_enc[0];
@@ -111,12 +125,7 @@ static int vcn_v5_0_1_sw_init(struct amdgpu_ip_block *ip_block)
 		if (r)
 			return r;
 
-		fw_shared = adev->vcn.inst[i].fw_shared.cpu_addr;
-		fw_shared->present_flag_0 = cpu_to_le32(AMDGPU_FW_SHARED_FLAG_0_UNIFIED_QUEUE);
-		fw_shared->sq.is_enabled = true;
-
-		if (amdgpu_vcnfw_log)
-			amdgpu_vcn_fwlog_init(&adev->vcn.inst[i]);
+		vcn_v5_0_1_fw_shared_init(adev, i);
 	}
 
 	/* TODO: Add queue reset mask when FW fully supports it */
@@ -188,6 +197,9 @@ static int vcn_v5_0_1_hw_init(struct amdgpu_ip_block *ip_block)
 				 9 * vcn_inst),
 				adev->vcn.inst[i].aid_id);
 
+		/* Re-init fw_shared, if required */
+		vcn_v5_0_1_fw_shared_init(adev, i);
+
 		r = amdgpu_ring_test_helper(ring);
 		if (r)
 			return r;
-- 
2.39.5

[PATCH AUTOSEL 6.14 248/642] drm/amdgpu: add dce_v6_0_soft_reset() to DCE6

[Sasha Levin]

From: Alexandre Demers <alexandre.f.demers@gmail.com>

[ Upstream commit ab23db6d08efdda5d13d01a66c593d0e57f8917f ]

DCE6 was missing soft reset, but it was easily identifiable under radeon.
This should be it, pretty much as it is done under DCE8 and DCE10.

Signed-off-by: Alexandre Demers <alexandre.f.demers@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/gpu/drm/amd/amdgpu/dce_v6_0.c | 53 ++++++++++++++++++++++++++-
 1 file changed, 51 insertions(+), 2 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/dce_v6_0.c b/drivers/gpu/drm/amd/amdgpu/dce_v6_0.c
index 915804a6a1d7d..ed5e06b677df1 100644
--- a/drivers/gpu/drm/amd/amdgpu/dce_v6_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/dce_v6_0.c
@@ -370,13 +370,41 @@ static u32 dce_v6_0_hpd_get_gpio_reg(struct amdgpu_device *adev)
 	return mmDC_GPIO_HPD_A;
 }
 
+static bool dce_v6_0_is_display_hung(struct amdgpu_device *adev)
+{
+	u32 crtc_hung = 0;
+	u32 crtc_status[6];
+	u32 i, j, tmp;
+
+	for (i = 0; i < adev->mode_info.num_crtc; i++) {
+		if (RREG32(mmCRTC_CONTROL + crtc_offsets[i]) & CRTC_CONTROL__CRTC_MASTER_EN_MASK) {
+			crtc_status[i] = RREG32(mmCRTC_STATUS_HV_COUNT + crtc_offsets[i]);
+			crtc_hung |= (1 << i);
+		}
+	}
+
+	for (j = 0; j < 10; j++) {
+		for (i = 0; i < adev->mode_info.num_crtc; i++) {
+			if (crtc_hung & (1 << i)) {
+				tmp = RREG32(mmCRTC_STATUS_HV_COUNT + crtc_offsets[i]);
+				if (tmp != crtc_status[i])
+					crtc_hung &= ~(1 << i);
+			}
+		}
+		if (crtc_hung == 0)
+			return false;
+		udelay(100);
+	}
+
+	return true;
+}
+
 static void dce_v6_0_set_vga_render_state(struct amdgpu_device *adev,
 					  bool render)
 {
 	if (!render)
 		WREG32(mmVGA_RENDER_CONTROL,
 			RREG32(mmVGA_RENDER_CONTROL) & VGA_VSTATUS_CNTL);
-
 }
 
 static int dce_v6_0_get_num_crtc(struct amdgpu_device *adev)
@@ -2872,7 +2900,28 @@ static bool dce_v6_0_is_idle(void *handle)
 
 static int dce_v6_0_soft_reset(struct amdgpu_ip_block *ip_block)
 {
-	DRM_INFO("xxxx: dce_v6_0_soft_reset --- no impl!!\n");
+	u32 srbm_soft_reset = 0, tmp;
+	struct amdgpu_device *adev = ip_block->adev;
+
+	if (dce_v6_0_is_display_hung(adev))
+		srbm_soft_reset |= SRBM_SOFT_RESET__SOFT_RESET_DC_MASK;
+
+	if (srbm_soft_reset) {
+		tmp = RREG32(mmSRBM_SOFT_RESET);
+		tmp |= srbm_soft_reset;
+		dev_info(adev->dev, "SRBM_SOFT_RESET=0x%08X\n", tmp);
+		WREG32(mmSRBM_SOFT_RESET, tmp);
+		tmp = RREG32(mmSRBM_SOFT_RESET);
+
+		udelay(50);
+
+		tmp &= ~srbm_soft_reset;
+		WREG32(mmSRBM_SOFT_RESET, tmp);
+		tmp = RREG32(mmSRBM_SOFT_RESET);
+
+		/* Wait a little for things to settle down */
+		udelay(50);
+	}
 	return 0;
 }
 
-- 
2.39.5

[PATCH AUTOSEL 6.14 249/642] drm/amd/display: Ensure DMCUB idle before reset on DCN31/DCN35

[Sasha Levin]

From: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>

[ Upstream commit c707ea82c79dbd1d295ec94cc6529a5248c77757 ]

[Why]
If we soft reset before halt finishes and there are outstanding
memory transactions then the memory interface may produce unexpected
results, such as out of order transactions when the firmware next runs.

These can manifest as random or unexpected load/store violations.

[How]
Increase the timeout before soft reset to ensure the DMCUB has quiesced.
This is effectively 1s maximum based on experimentation.

Use the enable bit check on DCN31 like we're doing on DCN35 and reorder
the reset writes to follow the HW programming guide.

Ensure we're reading SCRATCH7 instead of SCRATCH8 for the HALT code.
No current versions of DMCUB firmware use the SCRATCH8 boot bit to
dynamically switch where the HALT code goes to maintain backwards
compatibility with PSP.

Reviewed-by: Dillon Varone <dillon.varone@amd.com>
Signed-off-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Signed-off-by: Wayne Lin <wayne.lin@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 .../gpu/drm/amd/display/dmub/src/dmub_dcn31.c   | 17 +++++++++++------
 .../gpu/drm/amd/display/dmub/src/dmub_dcn35.c   |  4 ++--
 2 files changed, 13 insertions(+), 8 deletions(-)

diff --git a/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn31.c b/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn31.c
index d9f31b191c693..1a68b5782cac6 100644
--- a/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn31.c
+++ b/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn31.c
@@ -83,8 +83,8 @@ static inline void dmub_dcn31_translate_addr(const union dmub_addr *addr_in,
 void dmub_dcn31_reset(struct dmub_srv *dmub)
 {
 	union dmub_gpint_data_register cmd;
-	const uint32_t timeout = 100;
-	uint32_t in_reset, scratch, i, pwait_mode;
+	const uint32_t timeout = 100000;
+	uint32_t in_reset, is_enabled, scratch, i, pwait_mode;
 
 	REG_GET(DMCUB_CNTL2, DMCUB_SOFT_RESET, &in_reset);
 
@@ -108,7 +108,7 @@ void dmub_dcn31_reset(struct dmub_srv *dmub)
 		}
 
 		for (i = 0; i < timeout; ++i) {
-			scratch = dmub->hw_funcs.get_gpint_response(dmub);
+			scratch = REG_READ(DMCUB_SCRATCH7);
 			if (scratch == DMUB_GPINT__STOP_FW_RESPONSE)
 				break;
 
@@ -125,9 +125,14 @@ void dmub_dcn31_reset(struct dmub_srv *dmub)
 		/* Force reset in case we timed out, DMCUB is likely hung. */
 	}
 
-	REG_UPDATE(DMCUB_CNTL2, DMCUB_SOFT_RESET, 1);
-	REG_UPDATE(DMCUB_CNTL, DMCUB_ENABLE, 0);
-	REG_UPDATE(MMHUBBUB_SOFT_RESET, DMUIF_SOFT_RESET, 1);
+	REG_GET(DMCUB_CNTL, DMCUB_ENABLE, &is_enabled);
+
+	if (is_enabled) {
+		REG_UPDATE(DMCUB_CNTL2, DMCUB_SOFT_RESET, 1);
+		REG_UPDATE(MMHUBBUB_SOFT_RESET, DMUIF_SOFT_RESET, 1);
+		REG_UPDATE(DMCUB_CNTL, DMCUB_ENABLE, 0);
+	}
+
 	REG_WRITE(DMCUB_INBOX1_RPTR, 0);
 	REG_WRITE(DMCUB_INBOX1_WPTR, 0);
 	REG_WRITE(DMCUB_OUTBOX1_RPTR, 0);
diff --git a/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn35.c b/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn35.c
index e5e77bd3c31ea..01d013a12b947 100644
--- a/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn35.c
+++ b/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn35.c
@@ -88,7 +88,7 @@ static inline void dmub_dcn35_translate_addr(const union dmub_addr *addr_in,
 void dmub_dcn35_reset(struct dmub_srv *dmub)
 {
 	union dmub_gpint_data_register cmd;
-	const uint32_t timeout = 100;
+	const uint32_t timeout = 100000;
 	uint32_t in_reset, is_enabled, scratch, i, pwait_mode;
 
 	REG_GET(DMCUB_CNTL2, DMCUB_SOFT_RESET, &in_reset);
@@ -113,7 +113,7 @@ void dmub_dcn35_reset(struct dmub_srv *dmub)
 		}
 
 		for (i = 0; i < timeout; ++i) {
-			scratch = dmub->hw_funcs.get_gpint_response(dmub);
+			scratch = REG_READ(DMCUB_SCRATCH7);
 			if (scratch == DMUB_GPINT__STOP_FW_RESPONSE)
 				break;
 
-- 
2.39.5

[PATCH AUTOSEL 6.14 250/642] drm/amd/display: Skip checking FRL_MODE bit for PCON BW determination

[Sasha Levin]

From: George Shen <george.shen@amd.com>

[ Upstream commit 0584bbcf0c53c133081100e4f4c9fe41e598d045 ]

[Why/How]
Certain PCON will clear the FRL_MODE bit despite supporting the link BW
indicated in the other bits.

Thus, skip checking the FRL_MODE bit when interpreting the
hdmi_encoded_link_bw struct.

Reviewed-by: Wenjing Liu <wenjing.liu@amd.com>
Signed-off-by: George Shen <george.shen@amd.com>
Signed-off-by: Wayne Lin <wayne.lin@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 .../dc/link/protocols/link_dp_capability.c    | 30 +++++++++----------
 1 file changed, 15 insertions(+), 15 deletions(-)

diff --git a/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_capability.c b/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_capability.c
index 44f33e3bc1c59..6d7131369f00b 100644
--- a/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_capability.c
+++ b/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_capability.c
@@ -250,21 +250,21 @@ static uint32_t intersect_frl_link_bw_support(
 {
 	uint32_t supported_bw_in_kbps = max_supported_frl_bw_in_kbps;
 
-	// HDMI_ENCODED_LINK_BW bits are only valid if HDMI Link Configuration bit is 1 (FRL mode)
-	if (hdmi_encoded_link_bw.bits.FRL_MODE) {
-		if (hdmi_encoded_link_bw.bits.BW_48Gbps)
-			supported_bw_in_kbps = 48000000;
-		else if (hdmi_encoded_link_bw.bits.BW_40Gbps)
-			supported_bw_in_kbps = 40000000;
-		else if (hdmi_encoded_link_bw.bits.BW_32Gbps)
-			supported_bw_in_kbps = 32000000;
-		else if (hdmi_encoded_link_bw.bits.BW_24Gbps)
-			supported_bw_in_kbps = 24000000;
-		else if (hdmi_encoded_link_bw.bits.BW_18Gbps)
-			supported_bw_in_kbps = 18000000;
-		else if (hdmi_encoded_link_bw.bits.BW_9Gbps)
-			supported_bw_in_kbps = 9000000;
-	}
+	/* Skip checking FRL_MODE bit, as certain PCON will clear
+	 * it despite supporting the link BW indicated in the other bits.
+	 */
+	if (hdmi_encoded_link_bw.bits.BW_48Gbps)
+		supported_bw_in_kbps = 48000000;
+	else if (hdmi_encoded_link_bw.bits.BW_40Gbps)
+		supported_bw_in_kbps = 40000000;
+	else if (hdmi_encoded_link_bw.bits.BW_32Gbps)
+		supported_bw_in_kbps = 32000000;
+	else if (hdmi_encoded_link_bw.bits.BW_24Gbps)
+		supported_bw_in_kbps = 24000000;
+	else if (hdmi_encoded_link_bw.bits.BW_18Gbps)
+		supported_bw_in_kbps = 18000000;
+	else if (hdmi_encoded_link_bw.bits.BW_9Gbps)
+		supported_bw_in_kbps = 9000000;
 
 	return supported_bw_in_kbps;
 }
-- 
2.39.5

[PATCH AUTOSEL 6.14 251/642] drm/amd/display: Fix DMUB reset sequence for DCN401

[Sasha Levin]

From: Dillon Varone <Dillon.Varone@amd.com>

[ Upstream commit 0dfcc2bf269010a6e093793034c048049a40ee93 ]

[WHY]
It should no longer use DMCUB_SOFT_RESET as it can result
in the memory request path becoming desynchronized.

[HOW]
To ensure robustness in the reset sequence:
1) Extend timeout on the "halt" command sent via gpint, and check for
controller to enter "wait" as a stronger guarantee that there are no
requests to memory still in flight.
2) Remove usage of DMCUB_SOFT_RESET
3) Rely on PSP to reset the controller safely

Reviewed-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Signed-off-by: Dillon Varone <Dillon.Varone@amd.com>
Signed-off-by: Wayne Lin <wayne.lin@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 .../drm/amd/display/dmub/src/dmub_dcn401.c    | 47 ++++++++++++-------
 .../drm/amd/display/dmub/src/dmub_dcn401.h    |  3 +-
 2 files changed, 32 insertions(+), 18 deletions(-)

diff --git a/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn401.c b/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn401.c
index 39a8cb6d7523c..e1c4fe1c6e3ee 100644
--- a/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn401.c
+++ b/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn401.c
@@ -63,8 +63,10 @@ static inline void dmub_dcn401_translate_addr(const union dmub_addr *addr_in,
 void dmub_dcn401_reset(struct dmub_srv *dmub)
 {
 	union dmub_gpint_data_register cmd;
-	const uint32_t timeout = 30;
-	uint32_t in_reset, scratch, i;
+	const uint32_t timeout_us = 1 * 1000 * 1000; //1s
+	const uint32_t poll_delay_us = 1; //1us
+	uint32_t i = 0;
+	uint32_t in_reset, scratch, pwait_mode;
 
 	REG_GET(DMCUB_CNTL2, DMCUB_SOFT_RESET, &in_reset);
 
@@ -75,32 +77,35 @@ void dmub_dcn401_reset(struct dmub_srv *dmub)
 
 		dmub->hw_funcs.set_gpint(dmub, cmd);
 
-		/**
-		 * Timeout covers both the ACK and the wait
-		 * for remaining work to finish.
-		 *
-		 * This is mostly bound by the PHY disable sequence.
-		 * Each register check will be greater than 1us, so
-		 * don't bother using udelay.
-		 */
-
-		for (i = 0; i < timeout; ++i) {
+		for (i = 0; i < timeout_us; i++) {
 			if (dmub->hw_funcs.is_gpint_acked(dmub, cmd))
 				break;
+
+			udelay(poll_delay_us);
 		}
 
-		for (i = 0; i < timeout; ++i) {
+		for (; i < timeout_us; i++) {
 			scratch = dmub->hw_funcs.get_gpint_response(dmub);
 			if (scratch == DMUB_GPINT__STOP_FW_RESPONSE)
 				break;
+
+			udelay(poll_delay_us);
 		}
 
-		/* Force reset in case we timed out, DMCUB is likely hung. */
+		for (; i < timeout_us; i++) {
+			REG_GET(DMCUB_CNTL, DMCUB_PWAIT_MODE_STATUS, &pwait_mode);
+			if (pwait_mode & (1 << 0))
+				break;
+
+			udelay(poll_delay_us);
+		}
+	}
+
+	if (i >= timeout_us) {
+		/* timeout should never occur */
+		BREAK_TO_DEBUGGER();
 	}
 
-	REG_UPDATE(DMCUB_CNTL2, DMCUB_SOFT_RESET, 1);
-	REG_UPDATE(DMCUB_CNTL, DMCUB_ENABLE, 0);
-	REG_UPDATE(MMHUBBUB_SOFT_RESET, DMUIF_SOFT_RESET, 1);
 	REG_WRITE(DMCUB_INBOX1_RPTR, 0);
 	REG_WRITE(DMCUB_INBOX1_WPTR, 0);
 	REG_WRITE(DMCUB_OUTBOX1_RPTR, 0);
@@ -131,7 +136,10 @@ void dmub_dcn401_backdoor_load(struct dmub_srv *dmub,
 
 	dmub_dcn401_get_fb_base_offset(dmub, &fb_base, &fb_offset);
 
+	/* reset and disable DMCUB and MMHUBBUB DMUIF */
 	REG_UPDATE(DMCUB_SEC_CNTL, DMCUB_SEC_RESET, 1);
+	REG_UPDATE(MMHUBBUB_SOFT_RESET, DMUIF_SOFT_RESET, 1);
+	REG_UPDATE(DMCUB_CNTL, DMCUB_ENABLE, 0);
 
 	dmub_dcn401_translate_addr(&cw0->offset, fb_base, fb_offset, &offset);
 
@@ -151,6 +159,7 @@ void dmub_dcn401_backdoor_load(struct dmub_srv *dmub,
 			DMCUB_REGION3_CW1_TOP_ADDRESS, cw1->region.top,
 			DMCUB_REGION3_CW1_ENABLE, 1);
 
+	/* release DMCUB reset only to prevent premature execution */
 	REG_UPDATE_2(DMCUB_SEC_CNTL, DMCUB_SEC_RESET, 0, DMCUB_MEM_UNIT_ID,
 			0x20);
 }
@@ -161,7 +170,10 @@ void dmub_dcn401_backdoor_load_zfb_mode(struct dmub_srv *dmub,
 {
 	union dmub_addr offset;
 
+	/* reset and disable DMCUB and MMHUBBUB DMUIF */
 	REG_UPDATE(DMCUB_SEC_CNTL, DMCUB_SEC_RESET, 1);
+	REG_UPDATE(MMHUBBUB_SOFT_RESET, DMUIF_SOFT_RESET, 1);
+	REG_UPDATE(DMCUB_CNTL, DMCUB_ENABLE, 0);
 
 	offset = cw0->offset;
 
@@ -181,6 +193,7 @@ void dmub_dcn401_backdoor_load_zfb_mode(struct dmub_srv *dmub,
 			DMCUB_REGION3_CW1_TOP_ADDRESS, cw1->region.top,
 			DMCUB_REGION3_CW1_ENABLE, 1);
 
+	/* release DMCUB reset only to prevent premature execution */
 	REG_UPDATE_2(DMCUB_SEC_CNTL, DMCUB_SEC_RESET, 0, DMCUB_MEM_UNIT_ID,
 			0x20);
 }
diff --git a/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn401.h b/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn401.h
index 4c8843b796950..31f95b27e227d 100644
--- a/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn401.h
+++ b/drivers/gpu/drm/amd/display/dmub/src/dmub_dcn401.h
@@ -169,7 +169,8 @@ struct dmub_srv;
 	DMUB_SF(HOST_INTERRUPT_CSR, HOST_REG_INBOX0_RSP_INT_EN) \
 	DMUB_SF(HOST_INTERRUPT_CSR, HOST_REG_OUTBOX0_RDY_INT_ACK) \
 	DMUB_SF(HOST_INTERRUPT_CSR, HOST_REG_OUTBOX0_RDY_INT_STAT) \
-	DMUB_SF(HOST_INTERRUPT_CSR, HOST_REG_OUTBOX0_RDY_INT_EN)
+	DMUB_SF(HOST_INTERRUPT_CSR, HOST_REG_OUTBOX0_RDY_INT_EN) \
+	DMUB_SF(DMCUB_CNTL, DMCUB_PWAIT_MODE_STATUS)
 
 struct dmub_srv_dcn401_reg_offset {
 #define DMUB_SR(reg) uint32_t reg;
-- 
2.39.5

[PATCH AUTOSEL 6.14 252/642] drm/amd/display: Fix p-state type when p-state is unsupported

[Sasha Levin]

From: Dillon Varone <Dillon.Varone@amd.com>

[ Upstream commit a025f424af0407b7561bd5e6217295dde3abbc2e ]

[WHY&HOW]
P-state type would remain on previously used when unsupported which
causes confusion in logging and visual confirm, so set back to zero
when unsupported.

Reviewed-by: Aric Cyr <aric.cyr@amd.com>
Signed-off-by: Dillon Varone <Dillon.Varone@amd.com>
Signed-off-by: Wayne Lin <wayne.lin@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/gpu/drm/amd/display/dc/core/dc_hw_sequencer.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/drivers/gpu/drm/amd/display/dc/core/dc_hw_sequencer.c b/drivers/gpu/drm/amd/display/dc/core/dc_hw_sequencer.c
index a49604b7701f7..1406ee4bff801 100644
--- a/drivers/gpu/drm/amd/display/dc/core/dc_hw_sequencer.c
+++ b/drivers/gpu/drm/amd/display/dc/core/dc_hw_sequencer.c
@@ -563,6 +563,7 @@ void set_p_state_switch_method(
 	if (!dc->ctx || !dc->ctx->dmub_srv || !pipe_ctx || !vba)
 		return;
 
+	pipe_ctx->p_state_type = P_STATE_UNKNOWN;
 	if (vba->DRAMClockChangeSupport[vba->VoltageLevel][vba->maxMpcComb] !=
 			dm_dram_clock_change_unsupported) {
 		/* MCLK switching is supported */
-- 
2.39.5

[PATCH AUTOSEL 6.14 253/642] drm/amd/display: Request HW cursor on DCN3.2 with SubVP

[Sasha Levin]

From: Aric Cyr <Aric.Cyr@amd.com>

[ Upstream commit b74f46f3ce1e5f6336645f1e9ff47c56d5dfdef1 ]

[why]
When SubVP is active the HW cursor size is limited to 64x64, and
anything larger will force composition which is bad for gaming on
DCN3.2 if the game uses a larger cursor.

[how]
If HW cursor is requested, typically by a fullscreen game, do not
enable SubVP so that up to 256x256 cursor sizes are available for
DCN3.2.

Reviewed-by: Aric Cyr <aric.cyr@amd.com>
Signed-off-by: Aric Cyr <Aric.Cyr@amd.com>
Signed-off-by: Wayne Lin <wayne.lin@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/gpu/drm/amd/display/dc/core/dc.c             | 3 ++-
 drivers/gpu/drm/amd/display/dc/dml/dcn32/dcn32_fpu.c | 1 +
 2 files changed, 3 insertions(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/amd/display/dc/core/dc.c b/drivers/gpu/drm/amd/display/dc/core/dc.c
index a444fe1e0838a..5cd13aea1f694 100644
--- a/drivers/gpu/drm/amd/display/dc/core/dc.c
+++ b/drivers/gpu/drm/amd/display/dc/core/dc.c
@@ -4907,7 +4907,8 @@ static bool full_update_required(struct dc *dc,
 			stream_update->lut3d_func ||
 			stream_update->pending_test_pattern ||
 			stream_update->crtc_timing_adjust ||
-			stream_update->scaler_sharpener_update))
+			stream_update->scaler_sharpener_update ||
+			stream_update->hw_cursor_req))
 		return true;
 
 	if (stream) {
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn32/dcn32_fpu.c b/drivers/gpu/drm/amd/display/dc/dml/dcn32/dcn32_fpu.c
index 6f490d8d7038c..56dda686e2992 100644
--- a/drivers/gpu/drm/amd/display/dc/dml/dcn32/dcn32_fpu.c
+++ b/drivers/gpu/drm/amd/display/dc/dml/dcn32/dcn32_fpu.c
@@ -626,6 +626,7 @@ static bool dcn32_assign_subvp_pipe(struct dc *dc,
 		 * - Not TMZ surface
 		 */
 		if (pipe->plane_state && !pipe->top_pipe && !pipe->prev_odm_pipe && !dcn32_is_center_timing(pipe) &&
+				!pipe->stream->hw_cursor_req &&
 				!(pipe->stream->timing.pix_clk_100hz / 10000 > DCN3_2_MAX_SUBVP_PIXEL_RATE_MHZ) &&
 				(!dcn32_is_psr_capable(pipe) || (context->stream_count == 1 && dc->caps.dmub_caps.subvp_psr)) &&
 				dc_state_get_pipe_subvp_type(context, pipe) == SUBVP_NONE &&
-- 
2.39.5

[PATCH AUTOSEL 6.14 254/642] drm/amdgpu: Avoid HDP flush on JPEG v5.0.1

[Sasha Levin]

From: Lijo Lazar <lijo.lazar@amd.com>

[ Upstream commit a734a717dcfe1ce618301775034e598cb456665b ]

Similar to JPEG v4.0.3, HDP flush shouldn't be performed by JPEG engine.
Keep it empty.

Signed-off-by: Lijo Lazar <lijo.lazar@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/gpu/drm/amd/amdgpu/jpeg_v4_0_3.c | 2 +-
 drivers/gpu/drm/amd/amdgpu/jpeg_v4_0_3.h | 1 +
 drivers/gpu/drm/amd/amdgpu/jpeg_v5_0_1.c | 1 +
 3 files changed, 3 insertions(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0_3.c b/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0_3.c
index 88f9771c16869..b2904ee494e04 100644
--- a/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0_3.c
+++ b/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0_3.c
@@ -637,7 +637,7 @@ static uint64_t jpeg_v4_0_3_dec_ring_get_wptr(struct amdgpu_ring *ring)
 			ring->pipe ? (0x40 * ring->pipe - 0xc80) : 0);
 }
 
-static void jpeg_v4_0_3_ring_emit_hdp_flush(struct amdgpu_ring *ring)
+void jpeg_v4_0_3_ring_emit_hdp_flush(struct amdgpu_ring *ring)
 {
 	/* JPEG engine access for HDP flush doesn't work when RRMT is enabled.
 	 * This is a workaround to avoid any HDP flush through JPEG ring.
diff --git a/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0_3.h b/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0_3.h
index 747a3e5f68564..a90bf370a0025 100644
--- a/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0_3.h
+++ b/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0_3.h
@@ -56,6 +56,7 @@ void jpeg_v4_0_3_dec_ring_emit_fence(struct amdgpu_ring *ring, u64 addr, u64 seq
 				unsigned int flags);
 void jpeg_v4_0_3_dec_ring_emit_vm_flush(struct amdgpu_ring *ring,
 					unsigned int vmid, uint64_t pd_addr);
+void jpeg_v4_0_3_ring_emit_hdp_flush(struct amdgpu_ring *ring);
 void jpeg_v4_0_3_dec_ring_nop(struct amdgpu_ring *ring, uint32_t count);
 void jpeg_v4_0_3_dec_ring_insert_start(struct amdgpu_ring *ring);
 void jpeg_v4_0_3_dec_ring_insert_end(struct amdgpu_ring *ring);
diff --git a/drivers/gpu/drm/amd/amdgpu/jpeg_v5_0_1.c b/drivers/gpu/drm/amd/amdgpu/jpeg_v5_0_1.c
index 40d4c32a8c2a6..f2cc11b3fd68b 100644
--- a/drivers/gpu/drm/amd/amdgpu/jpeg_v5_0_1.c
+++ b/drivers/gpu/drm/amd/amdgpu/jpeg_v5_0_1.c
@@ -655,6 +655,7 @@ static const struct amdgpu_ring_funcs jpeg_v5_0_1_dec_ring_vm_funcs = {
 	.emit_ib = jpeg_v4_0_3_dec_ring_emit_ib,
 	.emit_fence = jpeg_v4_0_3_dec_ring_emit_fence,
 	.emit_vm_flush = jpeg_v4_0_3_dec_ring_emit_vm_flush,
+	.emit_hdp_flush = jpeg_v4_0_3_ring_emit_hdp_flush,
 	.test_ring = amdgpu_jpeg_dec_ring_test_ring,
 	.test_ib = amdgpu_jpeg_dec_ring_test_ib,
 	.insert_nop = jpeg_v4_0_3_dec_ring_nop,
-- 
2.39.5

[PATCH AUTOSEL 6.14 255/642] drm/amdgpu: Add offset normalization in VCN v5.0.1

[Sasha Levin]

From: Lijo Lazar <lijo.lazar@amd.com>

[ Upstream commit 0b9647d40ef82837d5025de6daad64db775ea1c5 ]

VCN v5.0.1 also will need register offset normalization. Reuse the logic
from VCN v4.0.3. Also, avoid HDP flush similar to VCN v4.0.3

Signed-off-by: Lijo Lazar <lijo.lazar@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/gpu/drm/amd/amdgpu/vcn_v4_0_3.c | 14 ++++++++------
 drivers/gpu/drm/amd/amdgpu/vcn_v4_0_3.h |  9 +++++++++
 drivers/gpu/drm/amd/amdgpu/vcn_v5_0_1.c | 20 +++++++++++---------
 3 files changed, 28 insertions(+), 15 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_3.c b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_3.c
index ecdc027f82203..e6cf21ed8afb8 100644
--- a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_3.c
+++ b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_3.c
@@ -31,6 +31,7 @@
 #include "soc15d.h"
 #include "soc15_hw_ip.h"
 #include "vcn_v2_0.h"
+#include "vcn_v4_0_3.h"
 #include "mmsch_v4_0_3.h"
 
 #include "vcn/vcn_4_0_3_offset.h"
@@ -1461,8 +1462,8 @@ static uint64_t vcn_v4_0_3_unified_ring_get_wptr(struct amdgpu_ring *ring)
 				    regUVD_RB_WPTR);
 }
 
-static void vcn_v4_0_3_enc_ring_emit_reg_wait(struct amdgpu_ring *ring, uint32_t reg,
-				uint32_t val, uint32_t mask)
+void vcn_v4_0_3_enc_ring_emit_reg_wait(struct amdgpu_ring *ring, uint32_t reg,
+				       uint32_t val, uint32_t mask)
 {
 	/* Use normalized offsets when required */
 	if (vcn_v4_0_3_normalizn_reqd(ring->adev))
@@ -1474,7 +1475,8 @@ static void vcn_v4_0_3_enc_ring_emit_reg_wait(struct amdgpu_ring *ring, uint32_t
 	amdgpu_ring_write(ring, val);
 }
 
-static void vcn_v4_0_3_enc_ring_emit_wreg(struct amdgpu_ring *ring, uint32_t reg, uint32_t val)
+void vcn_v4_0_3_enc_ring_emit_wreg(struct amdgpu_ring *ring, uint32_t reg,
+				   uint32_t val)
 {
 	/* Use normalized offsets when required */
 	if (vcn_v4_0_3_normalizn_reqd(ring->adev))
@@ -1485,8 +1487,8 @@ static void vcn_v4_0_3_enc_ring_emit_wreg(struct amdgpu_ring *ring, uint32_t reg
 	amdgpu_ring_write(ring, val);
 }
 
-static void vcn_v4_0_3_enc_ring_emit_vm_flush(struct amdgpu_ring *ring,
-				unsigned int vmid, uint64_t pd_addr)
+void vcn_v4_0_3_enc_ring_emit_vm_flush(struct amdgpu_ring *ring,
+				       unsigned int vmid, uint64_t pd_addr)
 {
 	struct amdgpu_vmhub *hub = &ring->adev->vmhub[ring->vm_hub];
 
@@ -1498,7 +1500,7 @@ static void vcn_v4_0_3_enc_ring_emit_vm_flush(struct amdgpu_ring *ring,
 					lower_32_bits(pd_addr), 0xffffffff);
 }
 
-static void vcn_v4_0_3_ring_emit_hdp_flush(struct amdgpu_ring *ring)
+void vcn_v4_0_3_ring_emit_hdp_flush(struct amdgpu_ring *ring)
 {
 	/* VCN engine access for HDP flush doesn't work when RRMT is enabled.
 	 * This is a workaround to avoid any HDP flush through VCN ring.
diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_3.h b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_3.h
index 0b046114373ae..03572a1d0c9cb 100644
--- a/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_3.h
+++ b/drivers/gpu/drm/amd/amdgpu/vcn_v4_0_3.h
@@ -26,4 +26,13 @@
 
 extern const struct amdgpu_ip_block_version vcn_v4_0_3_ip_block;
 
+void vcn_v4_0_3_enc_ring_emit_reg_wait(struct amdgpu_ring *ring, uint32_t reg,
+				       uint32_t val, uint32_t mask);
+
+void vcn_v4_0_3_enc_ring_emit_wreg(struct amdgpu_ring *ring, uint32_t reg,
+				   uint32_t val);
+void vcn_v4_0_3_enc_ring_emit_vm_flush(struct amdgpu_ring *ring,
+				       unsigned int vmid, uint64_t pd_addr);
+void vcn_v4_0_3_ring_emit_hdp_flush(struct amdgpu_ring *ring);
+
 #endif /* __VCN_V4_0_3_H__ */
diff --git a/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_1.c b/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_1.c
index cdbc10d7c9fb7..f893a84282832 100644
--- a/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_1.c
+++ b/drivers/gpu/drm/amd/amdgpu/vcn_v5_0_1.c
@@ -29,6 +29,7 @@
 #include "soc15d.h"
 #include "soc15_hw_ip.h"
 #include "vcn_v2_0.h"
+#include "vcn_v4_0_3.h"
 
 #include "vcn/vcn_5_0_0_offset.h"
 #include "vcn/vcn_5_0_0_sh_mask.h"
@@ -905,16 +906,17 @@ static const struct amdgpu_ring_funcs vcn_v5_0_1_unified_ring_vm_funcs = {
 	.get_rptr = vcn_v5_0_1_unified_ring_get_rptr,
 	.get_wptr = vcn_v5_0_1_unified_ring_get_wptr,
 	.set_wptr = vcn_v5_0_1_unified_ring_set_wptr,
-	.emit_frame_size =
-		SOC15_FLUSH_GPU_TLB_NUM_WREG * 3 +
-		SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 4 +
-		4 + /* vcn_v2_0_enc_ring_emit_vm_flush */
-		5 + 5 + /* vcn_v2_0_enc_ring_emit_fence x2 vm fence */
-		1, /* vcn_v2_0_enc_ring_insert_end */
+	.emit_frame_size = SOC15_FLUSH_GPU_TLB_NUM_WREG * 3 +
+			   SOC15_FLUSH_GPU_TLB_NUM_REG_WAIT * 4 +
+			   4 + /* vcn_v2_0_enc_ring_emit_vm_flush */
+			   5 +
+			   5 + /* vcn_v2_0_enc_ring_emit_fence x2 vm fence */
+			   1, /* vcn_v2_0_enc_ring_insert_end */
 	.emit_ib_size = 5, /* vcn_v2_0_enc_ring_emit_ib */
 	.emit_ib = vcn_v2_0_enc_ring_emit_ib,
 	.emit_fence = vcn_v2_0_enc_ring_emit_fence,
-	.emit_vm_flush = vcn_v2_0_enc_ring_emit_vm_flush,
+	.emit_vm_flush = vcn_v4_0_3_enc_ring_emit_vm_flush,
+	.emit_hdp_flush = vcn_v4_0_3_ring_emit_hdp_flush,
 	.test_ring = amdgpu_vcn_enc_ring_test_ring,
 	.test_ib = amdgpu_vcn_unified_ring_test_ib,
 	.insert_nop = amdgpu_ring_insert_nop,
@@ -922,8 +924,8 @@ static const struct amdgpu_ring_funcs vcn_v5_0_1_unified_ring_vm_funcs = {
 	.pad_ib = amdgpu_ring_generic_pad_ib,
 	.begin_use = amdgpu_vcn_ring_begin_use,
 	.end_use = amdgpu_vcn_ring_end_use,
-	.emit_wreg = vcn_v2_0_enc_ring_emit_wreg,
-	.emit_reg_wait = vcn_v2_0_enc_ring_emit_reg_wait,
+	.emit_wreg = vcn_v4_0_3_enc_ring_emit_wreg,
+	.emit_reg_wait = vcn_v4_0_3_enc_ring_emit_reg_wait,
 	.emit_reg_write_reg_wait = amdgpu_ring_emit_reg_write_reg_wait_helper,
 };
 
-- 
2.39.5

[PATCH AUTOSEL 6.14 304/642] drm/amdkfd: KFD release_work possible circular locking

[Sasha Levin]

From: Philip Yang <Philip.Yang@amd.com>

[ Upstream commit 1b9366c601039d60546794c63fbb83ce8e53b978 ]

If waiting for gpu reset done in KFD release_work, thers is WARNING:
possible circular locking dependency detected

  #2  kfd_create_process
        kfd_process_mutex
          flush kfd release work

  #1  kfd release work
        wait for amdgpu reset work

  #0  amdgpu_device_gpu_reset
        kgd2kfd_pre_reset
          kfd_process_mutex

  Possible unsafe locking scenario:

        CPU0                    CPU1
        ----                    ----
   lock((work_completion)(&p->release_work));
                  lock((wq_completion)kfd_process_wq);
                  lock((work_completion)(&p->release_work));
   lock((wq_completion)amdgpu-reset-dev);

To fix this, KFD create process move flush release work outside
kfd_process_mutex.

Signed-off-by: Philip Yang <Philip.Yang@amd.com>
Reviewed-by: Felix Kuehling <felix.kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/gpu/drm/amd/amdkfd/kfd_process.c | 16 ++++++++--------
 1 file changed, 8 insertions(+), 8 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_process.c b/drivers/gpu/drm/amd/amdkfd/kfd_process.c
index c3f2c0428e013..c9cbc0ecd9cb2 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_process.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_process.c
@@ -842,6 +842,14 @@ struct kfd_process *kfd_create_process(struct task_struct *thread)
 		return ERR_PTR(-EINVAL);
 	}
 
+	/* If the process just called exec(3), it is possible that the
+	 * cleanup of the kfd_process (following the release of the mm
+	 * of the old process image) is still in the cleanup work queue.
+	 * Make sure to drain any job before trying to recreate any
+	 * resource for this process.
+	 */
+	flush_workqueue(kfd_process_wq);
+
 	/*
 	 * take kfd processes mutex before starting of process creation
 	 * so there won't be a case where two threads of the same process
@@ -862,14 +870,6 @@ struct kfd_process *kfd_create_process(struct task_struct *thread)
 	if (process) {
 		pr_debug("Process already found\n");
 	} else {
-		/* If the process just called exec(3), it is possible that the
-		 * cleanup of the kfd_process (following the release of the mm
-		 * of the old process image) is still in the cleanup work queue.
-		 * Make sure to drain any job before trying to recreate any
-		 * resource for this process.
-		 */
-		flush_workqueue(kfd_process_wq);
-
 		process = create_process(thread);
 		if (IS_ERR(process))
 			goto out;
-- 
2.39.5

[PATCH AUTOSEL 6.14 321/642] drm/amd/display: handle max_downscale_src_width fail check

[Sasha Levin]

From: Yihan Zhu <Yihan.Zhu@amd.com>

[ Upstream commit 02a940da2ccc0cc0299811379580852b405a0ea2 ]

[WHY]
If max_downscale_src_width check fails, we exit early from TAP calculation and left a NULL
value to the scaling data structure to cause the zero divide in the DML validation.

[HOW]
Call set default TAP calculation before early exit in get_optimal_number_of_taps due to
max downscale limit exceed.

Reviewed-by: Samson Tam <samson.tam@amd.com>
Signed-off-by: Yihan Zhu <Yihan.Zhu@amd.com>
Signed-off-by: Zaeem Mohamed <zaeem.mohamed@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/gpu/drm/amd/display/dc/dpp/dcn30/dcn30_dpp.c | 11 ++++++-----
 1 file changed, 6 insertions(+), 5 deletions(-)

diff --git a/drivers/gpu/drm/amd/display/dc/dpp/dcn30/dcn30_dpp.c b/drivers/gpu/drm/amd/display/dc/dpp/dcn30/dcn30_dpp.c
index 40acebd13e46d..abf439e743f23 100644
--- a/drivers/gpu/drm/amd/display/dc/dpp/dcn30/dcn30_dpp.c
+++ b/drivers/gpu/drm/amd/display/dc/dpp/dcn30/dcn30_dpp.c
@@ -425,11 +425,6 @@ bool dpp3_get_optimal_number_of_taps(
 	int min_taps_y, min_taps_c;
 	enum lb_memory_config lb_config;
 
-	if (scl_data->viewport.width > scl_data->h_active &&
-		dpp->ctx->dc->debug.max_downscale_src_width != 0 &&
-		scl_data->viewport.width > dpp->ctx->dc->debug.max_downscale_src_width)
-		return false;
-
 	/*
 	 * Set default taps if none are provided
 	 * From programming guide: taps = min{ ceil(2*H_RATIO,1), 8} for downscaling
@@ -467,6 +462,12 @@ bool dpp3_get_optimal_number_of_taps(
 	else
 		scl_data->taps.h_taps_c = in_taps->h_taps_c;
 
+	// Avoid null data in the scl data with this early return, proceed non-adaptive calcualtion first
+	if (scl_data->viewport.width > scl_data->h_active &&
+		dpp->ctx->dc->debug.max_downscale_src_width != 0 &&
+		scl_data->viewport.width > dpp->ctx->dc->debug.max_downscale_src_width)
+		return false;
+
 	/*Ensure we can support the requested number of vtaps*/
 	min_taps_y = dc_fixpt_ceil(scl_data->ratios.vert);
 	min_taps_c = dc_fixpt_ceil(scl_data->ratios.vert_c);
-- 
2.39.5

[PATCH AUTOSEL 6.14 322/642] drm/amd/display: fix dcn4x init failed

[Sasha Levin]

From: Charlene Liu <Charlene.Liu@amd.com>

[ Upstream commit 23ef388a84c72b0614a6c10f866ffeac7e807719 ]

[why]
failed due to cmdtable not created.
switch atombios cmdtable as default.

Reviewed-by: Alvin Lee <alvin.lee2@amd.com>
Signed-off-by: Charlene Liu <Charlene.Liu@amd.com>
Signed-off-by: Zaeem Mohamed <zaeem.mohamed@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/gpu/drm/amd/display/dc/bios/command_table2.c     | 9 ---------
 .../gpu/drm/amd/display/dc/bios/command_table_helper2.c  | 3 +--
 2 files changed, 1 insertion(+), 11 deletions(-)

diff --git a/drivers/gpu/drm/amd/display/dc/bios/command_table2.c b/drivers/gpu/drm/amd/display/dc/bios/command_table2.c
index 7d18f372ce7ab..6bc59b7ef007b 100644
--- a/drivers/gpu/drm/amd/display/dc/bios/command_table2.c
+++ b/drivers/gpu/drm/amd/display/dc/bios/command_table2.c
@@ -101,7 +101,6 @@ static void init_dig_encoder_control(struct bios_parser *bp)
 		bp->cmd_tbl.dig_encoder_control = encoder_control_digx_v1_5;
 		break;
 	default:
-		dm_output_to_console("Don't have dig_encoder_control for v%d\n", version);
 		bp->cmd_tbl.dig_encoder_control = encoder_control_fallback;
 		break;
 	}
@@ -238,7 +237,6 @@ static void init_transmitter_control(struct bios_parser *bp)
 		bp->cmd_tbl.transmitter_control = transmitter_control_v1_7;
 		break;
 	default:
-		dm_output_to_console("Don't have transmitter_control for v%d\n", crev);
 		bp->cmd_tbl.transmitter_control = transmitter_control_fallback;
 		break;
 	}
@@ -408,8 +406,6 @@ static void init_set_pixel_clock(struct bios_parser *bp)
 		bp->cmd_tbl.set_pixel_clock = set_pixel_clock_v7;
 		break;
 	default:
-		dm_output_to_console("Don't have set_pixel_clock for v%d\n",
-			 BIOS_CMD_TABLE_PARA_REVISION(setpixelclock));
 		bp->cmd_tbl.set_pixel_clock = set_pixel_clock_fallback;
 		break;
 	}
@@ -554,7 +550,6 @@ static void init_set_crtc_timing(struct bios_parser *bp)
 			set_crtc_using_dtd_timing_v3;
 		break;
 	default:
-		dm_output_to_console("Don't have set_crtc_timing for v%d\n", dtd_version);
 		bp->cmd_tbl.set_crtc_timing = NULL;
 		break;
 	}
@@ -671,8 +666,6 @@ static void init_enable_crtc(struct bios_parser *bp)
 		bp->cmd_tbl.enable_crtc = enable_crtc_v1;
 		break;
 	default:
-		dm_output_to_console("Don't have enable_crtc for v%d\n",
-			 BIOS_CMD_TABLE_PARA_REVISION(enablecrtc));
 		bp->cmd_tbl.enable_crtc = NULL;
 		break;
 	}
@@ -864,8 +857,6 @@ static void init_set_dce_clock(struct bios_parser *bp)
 		bp->cmd_tbl.set_dce_clock = set_dce_clock_v2_1;
 		break;
 	default:
-		dm_output_to_console("Don't have set_dce_clock for v%d\n",
-			 BIOS_CMD_TABLE_PARA_REVISION(setdceclock));
 		bp->cmd_tbl.set_dce_clock = NULL;
 		break;
 	}
diff --git a/drivers/gpu/drm/amd/display/dc/bios/command_table_helper2.c b/drivers/gpu/drm/amd/display/dc/bios/command_table_helper2.c
index 73458e2951034..df8139bda142b 100644
--- a/drivers/gpu/drm/amd/display/dc/bios/command_table_helper2.c
+++ b/drivers/gpu/drm/amd/display/dc/bios/command_table_helper2.c
@@ -87,8 +87,7 @@ bool dal_bios_parser_init_cmd_tbl_helper2(
 		return true;
 
 	default:
-		/* Unsupported DCE */
-		BREAK_TO_DEBUGGER();
+		*h = dal_cmd_tbl_helper_dce112_get_table2();
 		return false;
 	}
 }
-- 
2.39.5

[PATCH AUTOSEL 6.14 323/642] drm/amd/display: fix check for identity ratio

[Sasha Levin]

From: Samson Tam <Samson.Tam@amd.com>

[ Upstream commit 0d3004647631aedb713251525a99784661574767 ]

[Why]
IDENTITY_RATIO check uses 2 bits for integer, which only allows
 checking downscale ratios up to 3.  But we support up to 6x
 downscale

[How]
Update IDENTITY_RATIO to check 3 bits for integer
Add ASSERT to catch if we downscale more than 6x

Signed-off-by: Samson Tam <Samson.Tam@amd.com>
Reviewed-by: Jun Lei <jun.lei@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/gpu/drm/amd/display/dc/spl/dc_spl.c | 21 ++++++++++++++++++++-
 1 file changed, 20 insertions(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/amd/display/dc/spl/dc_spl.c b/drivers/gpu/drm/amd/display/dc/spl/dc_spl.c
index 38a9a0d680581..18b423bd302a7 100644
--- a/drivers/gpu/drm/amd/display/dc/spl/dc_spl.c
+++ b/drivers/gpu/drm/amd/display/dc/spl/dc_spl.c
@@ -8,7 +8,7 @@
 #include "dc_spl_isharp_filters.h"
 #include "spl_debug.h"
 
-#define IDENTITY_RATIO(ratio) (spl_fixpt_u2d19(ratio) == (1 << 19))
+#define IDENTITY_RATIO(ratio) (spl_fixpt_u3d19(ratio) == (1 << 19))
 #define MIN_VIEWPORT_SIZE 12
 
 static bool spl_is_yuv420(enum spl_pixel_format format)
@@ -887,6 +887,8 @@ static bool spl_get_isharp_en(struct spl_in *spl_in,
 static void spl_get_taps_non_adaptive_scaler(
 	  struct spl_scratch *spl_scratch, const struct spl_taps *in_taps)
 {
+	bool check_max_downscale = false;
+
 	if (in_taps->h_taps == 0) {
 		if (spl_fixpt_ceil(spl_scratch->scl_data.ratios.horz) > 1)
 			spl_scratch->scl_data.taps.h_taps = spl_min(2 * spl_fixpt_ceil(
@@ -926,6 +928,23 @@ static void spl_get_taps_non_adaptive_scaler(
 	else
 		spl_scratch->scl_data.taps.h_taps_c = in_taps->h_taps_c;
 
+
+	/*
+	 * Max downscale supported is 6.0x.  Add ASSERT to catch if go beyond that
+	 */
+	check_max_downscale = spl_fixpt_le(spl_scratch->scl_data.ratios.horz,
+		spl_fixpt_from_fraction(6, 1));
+	SPL_ASSERT(check_max_downscale);
+	check_max_downscale = spl_fixpt_le(spl_scratch->scl_data.ratios.vert,
+		spl_fixpt_from_fraction(6, 1));
+	SPL_ASSERT(check_max_downscale);
+	check_max_downscale = spl_fixpt_le(spl_scratch->scl_data.ratios.horz_c,
+		spl_fixpt_from_fraction(6, 1));
+	SPL_ASSERT(check_max_downscale);
+	check_max_downscale = spl_fixpt_le(spl_scratch->scl_data.ratios.vert_c,
+		spl_fixpt_from_fraction(6, 1));
+	SPL_ASSERT(check_max_downscale);
+
 	if (IDENTITY_RATIO(spl_scratch->scl_data.ratios.horz))
 		spl_scratch->scl_data.taps.h_taps = 1;
 	if (IDENTITY_RATIO(spl_scratch->scl_data.ratios.vert))
-- 
2.39.5

[PATCH AUTOSEL 6.14 324/642] drm/amd/display: Fix mismatch type comparison

[Sasha Levin]

From: "Assadian, Navid" <navid.assadian@amd.com>

[ Upstream commit 26873260d394b1e33cdd720154aedf0af95327f9 ]

The mismatch type comparison/assignment may cause data loss. Since the
values are always non-negative, it is safe to use unsigned variables to
resolve the mismatch.

Signed-off-by: Navid Assadian <navid.assadian@amd.com>
Reviewed-by: Joshua Aberback <joshua.aberback@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/gpu/drm/amd/display/dc/spl/dc_spl.c       | 4 ++--
 drivers/gpu/drm/amd/display/dc/spl/dc_spl_types.h | 2 +-
 2 files changed, 3 insertions(+), 3 deletions(-)

diff --git a/drivers/gpu/drm/amd/display/dc/spl/dc_spl.c b/drivers/gpu/drm/amd/display/dc/spl/dc_spl.c
index 18b423bd302a7..22602f088553d 100644
--- a/drivers/gpu/drm/amd/display/dc/spl/dc_spl.c
+++ b/drivers/gpu/drm/amd/display/dc/spl/dc_spl.c
@@ -963,8 +963,8 @@ static bool spl_get_optimal_number_of_taps(
 	  bool *enable_isharp)
 {
 	int num_part_y, num_part_c;
-	int max_taps_y, max_taps_c;
-	int min_taps_y, min_taps_c;
+	unsigned int max_taps_y, max_taps_c;
+	unsigned int min_taps_y, min_taps_c;
 	enum lb_memory_config lb_config;
 	bool skip_easf = false;
 	bool is_subsampled = spl_is_subsampled_format(spl_in->basic_in.format);
diff --git a/drivers/gpu/drm/amd/display/dc/spl/dc_spl_types.h b/drivers/gpu/drm/amd/display/dc/spl/dc_spl_types.h
index 467af9dd90ded..5d139cf51e89b 100644
--- a/drivers/gpu/drm/amd/display/dc/spl/dc_spl_types.h
+++ b/drivers/gpu/drm/amd/display/dc/spl/dc_spl_types.h
@@ -484,7 +484,7 @@ struct spl_sharpness_range {
 };
 struct adaptive_sharpness {
 	bool enable;
-	int sharpness_level;
+	unsigned int sharpness_level;
 	struct spl_sharpness_range sharpness_range;
 };
 enum linear_light_scaling	{	// convert it in translation logic
-- 
2.39.5

[PATCH AUTOSEL 6.14 325/642] drm/amd/display: Add opp recout adjustment

[Sasha Levin]

From: Navid Assadian <Navid.Assadian@amd.com>

[ Upstream commit fba4d19f3731483ee8565f9e9bb7ed9fc89479e8 ]

[Why]
For subsampled YUV output formats, more pixels can get fetched and be
used for scaling.

[How]
Add the adjustment to the calculated recout, so the viewport covers the
corresponding pixels on the source plane.

Signed-off-by: Navid Assadian <Navid.Assadian@amd.com>
Reviewed-by: Samson Tam <Samson.Tam@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/gpu/drm/amd/display/dc/spl/dc_spl.c   | 31 +++++++++++++++----
 .../gpu/drm/amd/display/dc/spl/dc_spl_types.h | 10 ++++++
 2 files changed, 35 insertions(+), 6 deletions(-)

diff --git a/drivers/gpu/drm/amd/display/dc/spl/dc_spl.c b/drivers/gpu/drm/amd/display/dc/spl/dc_spl.c
index 22602f088553d..153b7a8904e1e 100644
--- a/drivers/gpu/drm/amd/display/dc/spl/dc_spl.c
+++ b/drivers/gpu/drm/amd/display/dc/spl/dc_spl.c
@@ -76,6 +76,21 @@ static struct spl_rect shift_rec(const struct spl_rect *rec_in, int x, int y)
 	return rec_out;
 }
 
+static void spl_opp_adjust_rect(struct spl_rect *rec, const struct spl_opp_adjust *adjust)
+{
+	if ((rec->x + adjust->x) >= 0)
+		rec->x += adjust->x;
+
+	if ((rec->y + adjust->y) >= 0)
+		rec->y += adjust->y;
+
+	if ((rec->width + adjust->width) >= 1)
+		rec->width += adjust->width;
+
+	if ((rec->height + adjust->height) >= 1)
+		rec->height += adjust->height;
+}
+
 static struct spl_rect calculate_plane_rec_in_timing_active(
 		struct spl_in *spl_in,
 		const struct spl_rect *rec_in)
@@ -723,13 +738,15 @@ static void spl_handle_3d_recout(struct spl_in *spl_in, struct spl_rect *recout)
 	}
 }
 
-static void spl_clamp_viewport(struct spl_rect *viewport)
+static void spl_clamp_viewport(struct spl_rect *viewport, int min_viewport_size)
 {
+	if (min_viewport_size == 0)
+		min_viewport_size = MIN_VIEWPORT_SIZE;
 	/* Clamp minimum viewport size */
-	if (viewport->height < MIN_VIEWPORT_SIZE)
-		viewport->height = MIN_VIEWPORT_SIZE;
-	if (viewport->width < MIN_VIEWPORT_SIZE)
-		viewport->width = MIN_VIEWPORT_SIZE;
+	if (viewport->height < min_viewport_size)
+		viewport->height = min_viewport_size;
+	if (viewport->width < min_viewport_size)
+		viewport->width = min_viewport_size;
 }
 
 static enum scl_mode spl_get_dscl_mode(const struct spl_in *spl_in,
@@ -1800,6 +1817,8 @@ static bool spl_calculate_number_of_taps(struct spl_in *spl_in, struct spl_scrat
 	spl_calculate_recout(spl_in, spl_scratch, spl_out);
 	/* depends on pixel format */
 	spl_calculate_scaling_ratios(spl_in, spl_scratch, spl_out);
+	/* Adjust recout for opp if needed */
+	spl_opp_adjust_rect(&spl_scratch->scl_data.recout, &spl_in->basic_in.opp_recout_adjust);
 	/* depends on scaling ratios and recout, does not calculate offset yet */
 	spl_calculate_viewport_size(spl_in, spl_scratch);
 
@@ -1836,7 +1855,7 @@ bool spl_calculate_scaler_params(struct spl_in *spl_in, struct spl_out *spl_out)
 	// Handle 3d recout
 	spl_handle_3d_recout(spl_in, &spl_scratch.scl_data.recout);
 	// Clamp
-	spl_clamp_viewport(&spl_scratch.scl_data.viewport);
+	spl_clamp_viewport(&spl_scratch.scl_data.viewport, spl_in->min_viewport_size);
 
 	// Save all calculated parameters in dscl_prog_data structure to program hw registers
 	spl_set_dscl_prog_data(spl_in, &spl_scratch, spl_out, enable_easf_v, enable_easf_h, enable_isharp);
diff --git a/drivers/gpu/drm/amd/display/dc/spl/dc_spl_types.h b/drivers/gpu/drm/amd/display/dc/spl/dc_spl_types.h
index 5d139cf51e89b..1c3949b24611f 100644
--- a/drivers/gpu/drm/amd/display/dc/spl/dc_spl_types.h
+++ b/drivers/gpu/drm/amd/display/dc/spl/dc_spl_types.h
@@ -427,6 +427,14 @@ struct spl_out	{
 
 // SPL inputs
 
+// opp extra adjustment for rect
+struct spl_opp_adjust {
+	int x;
+	int y;
+	int width;
+	int height;
+};
+
 // Basic input information
 struct basic_in	{
 	enum spl_pixel_format format; // Pixel Format
@@ -444,6 +452,7 @@ struct basic_in	{
 		} num_slices_recout_width;
 	} num_h_slices_recout_width_align;
 	int mpc_h_slice_index; // previous mpc_combine_v - split_idx
+	struct spl_opp_adjust opp_recout_adjust;
 	// Inputs for adaptive scaler - TODO
 	enum spl_transfer_func_type tf_type; /* Transfer function type */
 	enum spl_transfer_func_predefined tf_predefined_type; /* Transfer function predefined type */
@@ -535,6 +544,7 @@ struct spl_in	{
 	bool is_hdr_on;
 	int h_active;
 	int v_active;
+	int min_viewport_size;
 	int sdr_white_level_nits;
 	enum sharpen_policy sharpen_policy;
 };
-- 
2.39.5

[PATCH AUTOSEL 6.14 326/642] drm/amd/display: Fix mismatch type comparison in custom_float

[Sasha Levin]

From: Samson Tam <Samson.Tam@amd.com>

[ Upstream commit 86f06bcbb54e93f3c7b5e22ae37e72882b74c4b0 ]

[Why & How]
Passing uint into uchar function param.  Pass uint instead

Signed-off-by: Samson Tam <Samson.Tam@amd.com>
Reviewed-by: Alvin Lee <alvin.lee2@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/gpu/drm/amd/display/dc/spl/spl_fixpt31_32.c | 2 +-
 drivers/gpu/drm/amd/display/dc/spl/spl_fixpt31_32.h | 4 ++--
 2 files changed, 3 insertions(+), 3 deletions(-)

diff --git a/drivers/gpu/drm/amd/display/dc/spl/spl_fixpt31_32.c b/drivers/gpu/drm/amd/display/dc/spl/spl_fixpt31_32.c
index 131f1e3949d33..52d97918a3bd2 100644
--- a/drivers/gpu/drm/amd/display/dc/spl/spl_fixpt31_32.c
+++ b/drivers/gpu/drm/amd/display/dc/spl/spl_fixpt31_32.c
@@ -346,7 +346,7 @@ struct spl_fixed31_32 spl_fixpt_exp(struct spl_fixed31_32 arg)
 		if (m > 0)
 			return spl_fixpt_shl(
 				spl_fixed31_32_exp_from_taylor_series(r),
-				(unsigned char)m);
+				(unsigned int)m);
 		else
 			return spl_fixpt_div_int(
 				spl_fixed31_32_exp_from_taylor_series(r),
diff --git a/drivers/gpu/drm/amd/display/dc/spl/spl_fixpt31_32.h b/drivers/gpu/drm/amd/display/dc/spl/spl_fixpt31_32.h
index ed2647f9a0999..9f349ffe91485 100644
--- a/drivers/gpu/drm/amd/display/dc/spl/spl_fixpt31_32.h
+++ b/drivers/gpu/drm/amd/display/dc/spl/spl_fixpt31_32.h
@@ -189,7 +189,7 @@ static inline struct spl_fixed31_32 spl_fixpt_clamp(
  * @brief
  * result = arg << shift
  */
-static inline struct spl_fixed31_32 spl_fixpt_shl(struct spl_fixed31_32 arg, unsigned char shift)
+static inline struct spl_fixed31_32 spl_fixpt_shl(struct spl_fixed31_32 arg, unsigned int shift)
 {
 	SPL_ASSERT(((arg.value >= 0) && (arg.value <= LLONG_MAX >> shift)) ||
 		((arg.value < 0) && (arg.value >= ~(LLONG_MAX >> shift))));
@@ -203,7 +203,7 @@ static inline struct spl_fixed31_32 spl_fixpt_shl(struct spl_fixed31_32 arg, uns
  * @brief
  * result = arg >> shift
  */
-static inline struct spl_fixed31_32 spl_fixpt_shr(struct spl_fixed31_32 arg, unsigned char shift)
+static inline struct spl_fixed31_32 spl_fixpt_shr(struct spl_fixed31_32 arg, unsigned int shift)
 {
 	bool negative = arg.value < 0;
 
-- 
2.39.5

[PATCH AUTOSEL 6.14 355/642] drm/amdgpu: remove all KFD fences from the BO on release

[Sasha Levin]

From: Christian König <christian.koenig@amd.com>

[ Upstream commit cb0de06d1b0afb2d0c600ad748069f5ce27730ec ]

Remove all KFD BOs from the private dma_resv object.

This prevents the KFD from being evict unecessarily when an exported BO
is released.

Signed-off-by: Christian König <christian.koenig@amd.com>
Signed-off-by: James Zhu <James.Zhu@amd.com>
Reviewed-by: Felix Kuehling <felix.kuehling@amd.com>
Reviewed-and-tested-by: James Zhu <James.Zhu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h    |  5 +-
 .../gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c  | 52 ++++++++-----------
 drivers/gpu/drm/amd/amdgpu/amdgpu_object.c    | 38 ++++++++------
 3 files changed, 47 insertions(+), 48 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h
index 8af67f18500a7..2f48dc5747aa2 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h
@@ -192,7 +192,7 @@ int kfd_debugfs_kfd_mem_limits(struct seq_file *m, void *data);
 #if IS_ENABLED(CONFIG_HSA_AMD)
 bool amdkfd_fence_check_mm(struct dma_fence *f, struct mm_struct *mm);
 struct amdgpu_amdkfd_fence *to_amdgpu_amdkfd_fence(struct dma_fence *f);
-int amdgpu_amdkfd_remove_fence_on_pt_pd_bos(struct amdgpu_bo *bo);
+void amdgpu_amdkfd_remove_all_eviction_fences(struct amdgpu_bo *bo);
 int amdgpu_amdkfd_evict_userptr(struct mmu_interval_notifier *mni,
 				unsigned long cur_seq, struct kgd_mem *mem);
 int amdgpu_amdkfd_bo_validate_and_fence(struct amdgpu_bo *bo,
@@ -212,9 +212,8 @@ struct amdgpu_amdkfd_fence *to_amdgpu_amdkfd_fence(struct dma_fence *f)
 }
 
 static inline
-int amdgpu_amdkfd_remove_fence_on_pt_pd_bos(struct amdgpu_bo *bo)
+void amdgpu_amdkfd_remove_all_eviction_fences(struct amdgpu_bo *bo)
 {
-	return 0;
 }
 
 static inline
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c
index 70224b9f54f2f..c0aaa72b6c210 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c
@@ -370,40 +370,32 @@ static int amdgpu_amdkfd_remove_eviction_fence(struct amdgpu_bo *bo,
 	return 0;
 }
 
-int amdgpu_amdkfd_remove_fence_on_pt_pd_bos(struct amdgpu_bo *bo)
+/**
+ * amdgpu_amdkfd_remove_all_eviction_fences - Remove all eviction fences
+ * @bo: the BO where to remove the evictions fences from.
+ *
+ * This functions should only be used on release when all references to the BO
+ * are already dropped. We remove the eviction fence from the private copy of
+ * the dma_resv object here since that is what is used during release to
+ * determine of the BO is idle or not.
+ */
+void amdgpu_amdkfd_remove_all_eviction_fences(struct amdgpu_bo *bo)
 {
-	struct amdgpu_bo *root = bo;
-	struct amdgpu_vm_bo_base *vm_bo;
-	struct amdgpu_vm *vm;
-	struct amdkfd_process_info *info;
-	struct amdgpu_amdkfd_fence *ef;
-	int ret;
-
-	/* we can always get vm_bo from root PD bo.*/
-	while (root->parent)
-		root = root->parent;
+	struct dma_resv *resv = &bo->tbo.base._resv;
+	struct dma_fence *fence, *stub;
+	struct dma_resv_iter cursor;
 
-	vm_bo = root->vm_bo;
-	if (!vm_bo)
-		return 0;
+	dma_resv_assert_held(resv);
 
-	vm = vm_bo->vm;
-	if (!vm)
-		return 0;
-
-	info = vm->process_info;
-	if (!info || !info->eviction_fence)
-		return 0;
-
-	ef = container_of(dma_fence_get(&info->eviction_fence->base),
-			struct amdgpu_amdkfd_fence, base);
-
-	BUG_ON(!dma_resv_trylock(bo->tbo.base.resv));
-	ret = amdgpu_amdkfd_remove_eviction_fence(bo, ef);
-	dma_resv_unlock(bo->tbo.base.resv);
+	stub = dma_fence_get_stub();
+	dma_resv_for_each_fence(&cursor, resv, DMA_RESV_USAGE_BOOKKEEP, fence) {
+		if (!to_amdgpu_amdkfd_fence(fence))
+			continue;
 
-	dma_fence_put(&ef->base);
-	return ret;
+		dma_resv_replace_fences(resv, fence->context, stub,
+					DMA_RESV_USAGE_BOOKKEEP);
+	}
+	dma_fence_put(stub);
 }
 
 static int amdgpu_amdkfd_bo_validate(struct amdgpu_bo *bo, uint32_t domain,
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
index 00752e3f9d8ab..0b9987781f762 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
@@ -1295,28 +1295,36 @@ void amdgpu_bo_release_notify(struct ttm_buffer_object *bo)
 	if (abo->kfd_bo)
 		amdgpu_amdkfd_release_notify(abo);
 
-	/* We only remove the fence if the resv has individualized. */
-	WARN_ON_ONCE(bo->type == ttm_bo_type_kernel
-			&& bo->base.resv != &bo->base._resv);
-	if (bo->base.resv == &bo->base._resv)
-		amdgpu_amdkfd_remove_fence_on_pt_pd_bos(abo);
+	/*
+	 * We lock the private dma_resv object here and since the BO is about to
+	 * be released nobody else should have a pointer to it.
+	 * So when this locking here fails something is wrong with the reference
+	 * counting.
+	 */
+	if (WARN_ON_ONCE(!dma_resv_trylock(&bo->base._resv)))
+		return;
+
+	amdgpu_amdkfd_remove_all_eviction_fences(abo);
 
 	if (!bo->resource || bo->resource->mem_type != TTM_PL_VRAM ||
 	    !(abo->flags & AMDGPU_GEM_CREATE_VRAM_WIPE_ON_RELEASE) ||
 	    adev->in_suspend || drm_dev_is_unplugged(adev_to_drm(adev)))
-		return;
+		goto out;
 
-	if (WARN_ON_ONCE(!dma_resv_trylock(bo->base.resv)))
-		return;
+	r = dma_resv_reserve_fences(&bo->base._resv, 1);
+	if (r)
+		goto out;
 
-	r = amdgpu_fill_buffer(abo, 0, bo->base.resv, &fence, true);
-	if (!WARN_ON(r)) {
-		amdgpu_vram_mgr_set_cleared(bo->resource);
-		amdgpu_bo_fence(abo, fence, false);
-		dma_fence_put(fence);
-	}
+	r = amdgpu_fill_buffer(abo, 0, &bo->base._resv, &fence, true);
+	if (WARN_ON(r))
+		goto out;
+
+	amdgpu_vram_mgr_set_cleared(bo->resource);
+	dma_resv_add_fence(&bo->base._resv, fence, DMA_RESV_USAGE_KERNEL);
+	dma_fence_put(fence);
 
-	dma_resv_unlock(bo->base.resv);
+out:
+	dma_resv_unlock(&bo->base._resv);
 }
 
 /**
-- 
2.39.5

[PATCH AUTOSEL 6.14 383/642] drm/amd/pm: Fetch current power limit from PMFW

[Sasha Levin]

From: Lijo Lazar <lijo.lazar@amd.com>

[ Upstream commit b2a9e562dfa156bd53e62ce571f3f8f65d243f14 ]

On SMU v13.0.12, always query the firmware to get the current power
limit as it could be updated through other means also.

Signed-off-by: Lijo Lazar <lijo.lazar@amd.com>
Reviewed-by: Asad Kamal <asad.kamal@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/gpu/drm/amd/pm/swsmu/amdgpu_smu.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/drivers/gpu/drm/amd/pm/swsmu/amdgpu_smu.c b/drivers/gpu/drm/amd/pm/swsmu/amdgpu_smu.c
index ed9dac00ebfb1..f3f5b7dd15ccc 100644
--- a/drivers/gpu/drm/amd/pm/swsmu/amdgpu_smu.c
+++ b/drivers/gpu/drm/amd/pm/swsmu/amdgpu_smu.c
@@ -2802,6 +2802,7 @@ int smu_get_power_limit(void *handle,
 			switch (amdgpu_ip_version(adev, MP1_HWIP, 0)) {
 			case IP_VERSION(13, 0, 2):
 			case IP_VERSION(13, 0, 6):
+			case IP_VERSION(13, 0, 12):
 			case IP_VERSION(13, 0, 14):
 			case IP_VERSION(11, 0, 7):
 			case IP_VERSION(11, 0, 11):
-- 
2.39.5

[PATCH AUTOSEL 6.14 384/642] drm/amd/display: Add support for disconnected eDP streams

[Sasha Levin]

From: Harry VanZyllDeJong <hvanzyll@amd.com>

[ Upstream commit 6571bef25fe48c642f7a69ccf7c3198b317c136a ]

[Why]
eDP may not be connected to the GPU on driver start causing
fail enumeration.

[How]
Move the virtual signal type check before the eDP connector
signal check.

Reviewed-by: Wenjing Liu <wenjing.liu@amd.com>
Signed-off-by: Harry VanZyllDeJong <hvanzyll@amd.com>
Signed-off-by: Roman Li <roman.li@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 .../drm/amd/display/dc/link/protocols/link_dp_capability.c  | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_capability.c b/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_capability.c
index 6d7131369f00b..28843e9882d39 100644
--- a/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_capability.c
+++ b/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_capability.c
@@ -945,6 +945,9 @@ bool link_decide_link_settings(struct dc_stream_state *stream,
 		 * TODO: add MST specific link training routine
 		 */
 		decide_mst_link_settings(link, link_setting);
+	} else if (stream->signal == SIGNAL_TYPE_VIRTUAL) {
+		link_setting->lane_count = LANE_COUNT_FOUR;
+		link_setting->link_rate = LINK_RATE_HIGH3;
 	} else if (link->connector_signal == SIGNAL_TYPE_EDP) {
 		/* enable edp link optimization for DSC eDP case */
 		if (stream->timing.flags.DSC) {
@@ -967,9 +970,6 @@ bool link_decide_link_settings(struct dc_stream_state *stream,
 		} else {
 			edp_decide_link_settings(link, link_setting, req_bw);
 		}
-	} else if (stream->signal == SIGNAL_TYPE_VIRTUAL) {
-		link_setting->lane_count = LANE_COUNT_FOUR;
-		link_setting->link_rate = LINK_RATE_HIGH3;
 	} else {
 		decide_dp_link_settings(link, link_setting, req_bw);
 	}
-- 
2.39.5

[PATCH AUTOSEL 6.14 385/642] drm/amd/display: Guard against setting dispclk low when active

[Sasha Levin]

From: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>

[ Upstream commit 72d7a7fa1f2404fd31c84a8f808b1b37021a3a9e ]

[Why]
We should never apply a minimum dispclk value while in prepare_bandwidth
or while displays are active. This is always an optimization for when
all displays are disabled.

[How]
Defer dispclk optimization until safe_to_lower = true and display_count
reaches 0.

Since 0 has a special value in this logic (ie. no dispclk required)
we also need adjust the logic that clamps it for the actual request
to PMFW.

Reviewed-by: Gabe Teeger <gabe.teeger@amd.com>
Reviewed-by: Leo Chen <leo.chen@amd.com>
Reviewed-by: Syed Hassan <syed.hassan@amd.com>
Signed-off-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Signed-off-by: Roman Li <roman.li@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 .../amd/display/dc/clk_mgr/dcn35/dcn35_clk_mgr.c    | 13 +++++++++----
 1 file changed, 9 insertions(+), 4 deletions(-)

diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn35/dcn35_clk_mgr.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn35/dcn35_clk_mgr.c
index 1648226586e22..1f47931c2dafc 100644
--- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn35/dcn35_clk_mgr.c
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn35/dcn35_clk_mgr.c
@@ -467,14 +467,19 @@ void dcn35_update_clocks(struct clk_mgr *clk_mgr_base,
 		update_dppclk = true;
 	}
 
-	if (should_set_clock(safe_to_lower, new_clocks->dispclk_khz, clk_mgr_base->clks.dispclk_khz)) {
+	if (should_set_clock(safe_to_lower, new_clocks->dispclk_khz, clk_mgr_base->clks.dispclk_khz) &&
+	    (new_clocks->dispclk_khz > 0 || (safe_to_lower && display_count == 0))) {
+		int requested_dispclk_khz = new_clocks->dispclk_khz;
+
 		dcn35_disable_otg_wa(clk_mgr_base, context, safe_to_lower, true);
 
-		if (dc->debug.min_disp_clk_khz > 0 && new_clocks->dispclk_khz < dc->debug.min_disp_clk_khz)
-			new_clocks->dispclk_khz = dc->debug.min_disp_clk_khz;
+		/* Clamp the requested clock to PMFW based on their limit. */
+		if (dc->debug.min_disp_clk_khz > 0 && requested_dispclk_khz < dc->debug.min_disp_clk_khz)
+			requested_dispclk_khz = dc->debug.min_disp_clk_khz;
 
+		dcn35_smu_set_dispclk(clk_mgr, requested_dispclk_khz);
 		clk_mgr_base->clks.dispclk_khz = new_clocks->dispclk_khz;
-		dcn35_smu_set_dispclk(clk_mgr, clk_mgr_base->clks.dispclk_khz);
+
 		dcn35_disable_otg_wa(clk_mgr_base, context, safe_to_lower, false);
 
 		update_dispclk = true;
-- 
2.39.5

[PATCH AUTOSEL 6.14 386/642] drm/amd/display: Fix BT2020 YCbCr limited/full range input

[Sasha Levin]

From: Ilya Bakoulin <Ilya.Bakoulin@amd.com>

[ Upstream commit 07bc2dcbcf403d47d6f305ef7f0d3d489491c5fb ]

[Why]
BT2020 YCbCr input is not handled properly when full range
quantization is used and limited range is not supported at all.

[How]
- Add enums for BT2020 YCbCr limited/full range
- Add limited range CSC matrix

Reviewed-by: Krunoslav Kovac <krunoslav.kovac@amd.com>
Signed-off-by: Ilya Bakoulin <Ilya.Bakoulin@amd.com>
Signed-off-by: Roman Li <roman.li@amd.com>
Tested-by: Robert Mader <robert.mader@collabora.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c           | 6 +++---
 drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_debugfs.c   | 2 +-
 drivers/gpu/drm/amd/display/dc/basics/dc_common.c           | 3 ++-
 drivers/gpu/drm/amd/display/dc/core/dc_hw_sequencer.c       | 5 +++--
 drivers/gpu/drm/amd/display/dc/core/dc_resource.c           | 4 ++--
 drivers/gpu/drm/amd/display/dc/dc_hw_types.h                | 4 +++-
 drivers/gpu/drm/amd/display/dc/dce/dce_stream_encoder.c     | 3 ++-
 .../gpu/drm/amd/display/dc/dio/dcn10/dcn10_stream_encoder.c | 3 ++-
 .../amd/display/dc/dio/dcn401/dcn401_dio_stream_encoder.c   | 3 ++-
 .../amd/display/dc/hpo/dcn31/dcn31_hpo_dp_stream_encoder.c  | 3 ++-
 drivers/gpu/drm/amd/display/dc/inc/hw/dpp.h                 | 6 +++++-
 .../gpu/drm/amd/display/modules/info_packet/info_packet.c   | 4 ++--
 12 files changed, 29 insertions(+), 17 deletions(-)

diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
index 76c8e6457175f..79c01541a0190 100644
--- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
@@ -5662,9 +5662,9 @@ fill_plane_color_attributes(const struct drm_plane_state *plane_state,
 
 	case DRM_COLOR_YCBCR_BT2020:
 		if (full_range)
-			*color_space = COLOR_SPACE_2020_YCBCR;
+			*color_space = COLOR_SPACE_2020_YCBCR_FULL;
 		else
-			return -EINVAL;
+			*color_space = COLOR_SPACE_2020_YCBCR_LIMITED;
 		break;
 
 	default:
@@ -6160,7 +6160,7 @@ get_output_color_space(const struct dc_crtc_timing *dc_crtc_timing,
 		if (dc_crtc_timing->pixel_encoding == PIXEL_ENCODING_RGB)
 			color_space = COLOR_SPACE_2020_RGB_FULLRANGE;
 		else
-			color_space = COLOR_SPACE_2020_YCBCR;
+			color_space = COLOR_SPACE_2020_YCBCR_LIMITED;
 		break;
 	case DRM_MODE_COLORIMETRY_DEFAULT: // ITU601
 	default:
diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_debugfs.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_debugfs.c
index 049046c604626..c7d13e743e6c8 100644
--- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_debugfs.c
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_debugfs.c
@@ -1169,7 +1169,7 @@ static int amdgpu_current_colorspace_show(struct seq_file *m, void *data)
 	case COLOR_SPACE_2020_RGB_FULLRANGE:
 		seq_puts(m, "BT2020_RGB");
 		break;
-	case COLOR_SPACE_2020_YCBCR:
+	case COLOR_SPACE_2020_YCBCR_LIMITED:
 		seq_puts(m, "BT2020_YCC");
 		break;
 	default:
diff --git a/drivers/gpu/drm/amd/display/dc/basics/dc_common.c b/drivers/gpu/drm/amd/display/dc/basics/dc_common.c
index b2fc4f8e64825..a51c2701da247 100644
--- a/drivers/gpu/drm/amd/display/dc/basics/dc_common.c
+++ b/drivers/gpu/drm/amd/display/dc/basics/dc_common.c
@@ -40,7 +40,8 @@ bool is_rgb_cspace(enum dc_color_space output_color_space)
 	case COLOR_SPACE_YCBCR709:
 	case COLOR_SPACE_YCBCR601_LIMITED:
 	case COLOR_SPACE_YCBCR709_LIMITED:
-	case COLOR_SPACE_2020_YCBCR:
+	case COLOR_SPACE_2020_YCBCR_LIMITED:
+	case COLOR_SPACE_2020_YCBCR_FULL:
 		return false;
 	default:
 		/* Add a case to switch */
diff --git a/drivers/gpu/drm/amd/display/dc/core/dc_hw_sequencer.c b/drivers/gpu/drm/amd/display/dc/core/dc_hw_sequencer.c
index 1406ee4bff801..4f54e75a8f95b 100644
--- a/drivers/gpu/drm/amd/display/dc/core/dc_hw_sequencer.c
+++ b/drivers/gpu/drm/amd/display/dc/core/dc_hw_sequencer.c
@@ -176,7 +176,7 @@ static bool is_ycbcr2020_type(
 {
 	bool ret = false;
 
-	if (color_space == COLOR_SPACE_2020_YCBCR)
+	if (color_space == COLOR_SPACE_2020_YCBCR_LIMITED || color_space == COLOR_SPACE_2020_YCBCR_FULL)
 		ret = true;
 	return ret;
 }
@@ -247,7 +247,8 @@ void color_space_to_black_color(
 	case COLOR_SPACE_YCBCR709_BLACK:
 	case COLOR_SPACE_YCBCR601_LIMITED:
 	case COLOR_SPACE_YCBCR709_LIMITED:
-	case COLOR_SPACE_2020_YCBCR:
+	case COLOR_SPACE_2020_YCBCR_LIMITED:
+	case COLOR_SPACE_2020_YCBCR_FULL:
 		*black_color = black_color_format[BLACK_COLOR_FORMAT_YUV_CV];
 		break;
 
diff --git a/drivers/gpu/drm/amd/display/dc/core/dc_resource.c b/drivers/gpu/drm/amd/display/dc/core/dc_resource.c
index 298668e9729c7..3367030da3414 100644
--- a/drivers/gpu/drm/amd/display/dc/core/dc_resource.c
+++ b/drivers/gpu/drm/amd/display/dc/core/dc_resource.c
@@ -4247,7 +4247,7 @@ static void set_avi_info_frame(
 		break;
 	case COLOR_SPACE_2020_RGB_FULLRANGE:
 	case COLOR_SPACE_2020_RGB_LIMITEDRANGE:
-	case COLOR_SPACE_2020_YCBCR:
+	case COLOR_SPACE_2020_YCBCR_LIMITED:
 		hdmi_info.bits.EC0_EC2 = COLORIMETRYEX_BT2020RGBYCBCR;
 		hdmi_info.bits.C0_C1   = COLORIMETRY_EXTENDED;
 		break;
@@ -4261,7 +4261,7 @@ static void set_avi_info_frame(
 		break;
 	}
 
-	if (pixel_encoding && color_space == COLOR_SPACE_2020_YCBCR &&
+	if (pixel_encoding && color_space == COLOR_SPACE_2020_YCBCR_LIMITED &&
 			stream->out_transfer_func.tf == TRANSFER_FUNCTION_GAMMA22) {
 		hdmi_info.bits.EC0_EC2 = 0;
 		hdmi_info.bits.C0_C1 = COLORIMETRY_ITU709;
diff --git a/drivers/gpu/drm/amd/display/dc/dc_hw_types.h b/drivers/gpu/drm/amd/display/dc/dc_hw_types.h
index 37e381fc7f02a..d562ddeca5126 100644
--- a/drivers/gpu/drm/amd/display/dc/dc_hw_types.h
+++ b/drivers/gpu/drm/amd/display/dc/dc_hw_types.h
@@ -653,7 +653,8 @@ enum dc_color_space {
 	COLOR_SPACE_YCBCR709_LIMITED,
 	COLOR_SPACE_2020_RGB_FULLRANGE,
 	COLOR_SPACE_2020_RGB_LIMITEDRANGE,
-	COLOR_SPACE_2020_YCBCR,
+	COLOR_SPACE_2020_YCBCR_LIMITED,
+	COLOR_SPACE_2020_YCBCR_FULL,
 	COLOR_SPACE_ADOBERGB,
 	COLOR_SPACE_DCIP3,
 	COLOR_SPACE_DISPLAYNATIVE,
@@ -661,6 +662,7 @@ enum dc_color_space {
 	COLOR_SPACE_APPCTRL,
 	COLOR_SPACE_CUSTOMPOINTS,
 	COLOR_SPACE_YCBCR709_BLACK,
+	COLOR_SPACE_2020_YCBCR = COLOR_SPACE_2020_YCBCR_LIMITED,
 };
 
 enum dc_dither_option {
diff --git a/drivers/gpu/drm/amd/display/dc/dce/dce_stream_encoder.c b/drivers/gpu/drm/amd/display/dc/dce/dce_stream_encoder.c
index d199e4ed2e59e..1130d7619b263 100644
--- a/drivers/gpu/drm/amd/display/dc/dce/dce_stream_encoder.c
+++ b/drivers/gpu/drm/amd/display/dc/dce/dce_stream_encoder.c
@@ -418,7 +418,7 @@ static void dce110_stream_encoder_dp_set_stream_attribute(
 			dynamic_range_rgb = 1; /*limited range*/
 			break;
 		case COLOR_SPACE_2020_RGB_FULLRANGE:
-		case COLOR_SPACE_2020_YCBCR:
+		case COLOR_SPACE_2020_YCBCR_LIMITED:
 		case COLOR_SPACE_XR_RGB:
 		case COLOR_SPACE_MSREF_SCRGB:
 		case COLOR_SPACE_ADOBERGB:
@@ -430,6 +430,7 @@ static void dce110_stream_encoder_dp_set_stream_attribute(
 		case COLOR_SPACE_APPCTRL:
 		case COLOR_SPACE_CUSTOMPOINTS:
 		case COLOR_SPACE_UNKNOWN:
+		default:
 			/* do nothing */
 			break;
 		}
diff --git a/drivers/gpu/drm/amd/display/dc/dio/dcn10/dcn10_stream_encoder.c b/drivers/gpu/drm/amd/display/dc/dio/dcn10/dcn10_stream_encoder.c
index d01a8b8f95954..22e66b375a7fe 100644
--- a/drivers/gpu/drm/amd/display/dc/dio/dcn10/dcn10_stream_encoder.c
+++ b/drivers/gpu/drm/amd/display/dc/dio/dcn10/dcn10_stream_encoder.c
@@ -391,7 +391,7 @@ void enc1_stream_encoder_dp_set_stream_attribute(
 		break;
 	case COLOR_SPACE_2020_RGB_LIMITEDRANGE:
 	case COLOR_SPACE_2020_RGB_FULLRANGE:
-	case COLOR_SPACE_2020_YCBCR:
+	case COLOR_SPACE_2020_YCBCR_LIMITED:
 	case COLOR_SPACE_XR_RGB:
 	case COLOR_SPACE_MSREF_SCRGB:
 	case COLOR_SPACE_ADOBERGB:
@@ -404,6 +404,7 @@ void enc1_stream_encoder_dp_set_stream_attribute(
 	case COLOR_SPACE_CUSTOMPOINTS:
 	case COLOR_SPACE_UNKNOWN:
 	case COLOR_SPACE_YCBCR709_BLACK:
+	default:
 		/* do nothing */
 		break;
 	}
diff --git a/drivers/gpu/drm/amd/display/dc/dio/dcn401/dcn401_dio_stream_encoder.c b/drivers/gpu/drm/amd/display/dc/dio/dcn401/dcn401_dio_stream_encoder.c
index 098c2a01a8509..9e5072627ec7b 100644
--- a/drivers/gpu/drm/amd/display/dc/dio/dcn401/dcn401_dio_stream_encoder.c
+++ b/drivers/gpu/drm/amd/display/dc/dio/dcn401/dcn401_dio_stream_encoder.c
@@ -632,7 +632,7 @@ void enc401_stream_encoder_dp_set_stream_attribute(
 		break;
 	case COLOR_SPACE_2020_RGB_LIMITEDRANGE:
 	case COLOR_SPACE_2020_RGB_FULLRANGE:
-	case COLOR_SPACE_2020_YCBCR:
+	case COLOR_SPACE_2020_YCBCR_LIMITED:
 	case COLOR_SPACE_XR_RGB:
 	case COLOR_SPACE_MSREF_SCRGB:
 	case COLOR_SPACE_ADOBERGB:
@@ -645,6 +645,7 @@ void enc401_stream_encoder_dp_set_stream_attribute(
 	case COLOR_SPACE_CUSTOMPOINTS:
 	case COLOR_SPACE_UNKNOWN:
 	case COLOR_SPACE_YCBCR709_BLACK:
+	default:
 		/* do nothing */
 		break;
 	}
diff --git a/drivers/gpu/drm/amd/display/dc/hpo/dcn31/dcn31_hpo_dp_stream_encoder.c b/drivers/gpu/drm/amd/display/dc/hpo/dcn31/dcn31_hpo_dp_stream_encoder.c
index 678db949cfe3c..759b453385c46 100644
--- a/drivers/gpu/drm/amd/display/dc/hpo/dcn31/dcn31_hpo_dp_stream_encoder.c
+++ b/drivers/gpu/drm/amd/display/dc/hpo/dcn31/dcn31_hpo_dp_stream_encoder.c
@@ -323,7 +323,7 @@ static void dcn31_hpo_dp_stream_enc_set_stream_attribute(
 		break;
 	case COLOR_SPACE_2020_RGB_LIMITEDRANGE:
 	case COLOR_SPACE_2020_RGB_FULLRANGE:
-	case COLOR_SPACE_2020_YCBCR:
+	case COLOR_SPACE_2020_YCBCR_LIMITED:
 	case COLOR_SPACE_XR_RGB:
 	case COLOR_SPACE_MSREF_SCRGB:
 	case COLOR_SPACE_ADOBERGB:
@@ -336,6 +336,7 @@ static void dcn31_hpo_dp_stream_enc_set_stream_attribute(
 	case COLOR_SPACE_CUSTOMPOINTS:
 	case COLOR_SPACE_UNKNOWN:
 	case COLOR_SPACE_YCBCR709_BLACK:
+	default:
 		/* do nothing */
 		break;
 	}
diff --git a/drivers/gpu/drm/amd/display/dc/inc/hw/dpp.h b/drivers/gpu/drm/amd/display/dc/inc/hw/dpp.h
index 0150f2581ee4c..0c5675d1c5936 100644
--- a/drivers/gpu/drm/amd/display/dc/inc/hw/dpp.h
+++ b/drivers/gpu/drm/amd/display/dc/inc/hw/dpp.h
@@ -119,10 +119,14 @@ static const struct dpp_input_csc_matrix __maybe_unused dpp_input_csc_matrix[] =
 		{ 0x39a6, 0x2568, 0,      0xe0d6,
 		  0xeedd, 0x2568, 0xf925, 0x9a8,
 		  0,      0x2568, 0x43ee, 0xdbb2 } },
-	{ COLOR_SPACE_2020_YCBCR,
+	{ COLOR_SPACE_2020_YCBCR_FULL,
 		{ 0x2F30, 0x2000, 0,      0xE869,
 		  0xEDB7, 0x2000, 0xFABC, 0xBC6,
 		  0,      0x2000, 0x3C34, 0xE1E6 } },
+	{ COLOR_SPACE_2020_YCBCR_LIMITED,
+		{ 0x35B9, 0x2543, 0,      0xE2B2,
+		  0xEB2F, 0x2543, 0xFA01, 0x0B1F,
+		  0,      0x2543, 0x4489, 0xDB42 } },
 	{ COLOR_SPACE_2020_RGB_LIMITEDRANGE,
 		{ 0x35E0, 0x255F, 0,      0xE2B3,
 		  0xEB20, 0x255F, 0xF9FD, 0xB1E,
diff --git a/drivers/gpu/drm/amd/display/modules/info_packet/info_packet.c b/drivers/gpu/drm/amd/display/modules/info_packet/info_packet.c
index a344e2e49b0ea..b3d55cac35694 100644
--- a/drivers/gpu/drm/amd/display/modules/info_packet/info_packet.c
+++ b/drivers/gpu/drm/amd/display/modules/info_packet/info_packet.c
@@ -383,10 +383,10 @@ void mod_build_vsc_infopacket(const struct dc_stream_state *stream,
 				colorimetryFormat = ColorimetryYCC_DP_ITU709;
 			else if (cs == COLOR_SPACE_ADOBERGB)
 				colorimetryFormat = ColorimetryYCC_DP_AdobeYCC;
-			else if (cs == COLOR_SPACE_2020_YCBCR)
+			else if (cs == COLOR_SPACE_2020_YCBCR_LIMITED)
 				colorimetryFormat = ColorimetryYCC_DP_ITU2020YCbCr;
 
-			if (cs == COLOR_SPACE_2020_YCBCR && tf == TRANSFER_FUNC_GAMMA_22)
+			if (cs == COLOR_SPACE_2020_YCBCR_LIMITED && tf == TRANSFER_FUNC_GAMMA_22)
 				colorimetryFormat = ColorimetryYCC_DP_ITU709;
 			break;
 
-- 
2.39.5

[PATCH AUTOSEL 6.14 387/642] drm/amd/display: Read LTTPR ALPM caps during link cap retrieval

[Sasha Levin]

From: George Shen <george.shen@amd.com>

[ Upstream commit de84d580126eb2214937df755cfec5ef0901479e ]

[Why]
The latest DP spec requires the DP TX to read DPCD F0000h through F0009h
when detecting LTTPR capabilities for the first time.

[How]
Update LTTPR cap retrieval to read up to F0009h (two more bytes than the
previous F0007h), and store the LTTPR ALPM capabilities.

Reviewed-by: Wenjing Liu <wenjing.liu@amd.com>
Signed-off-by: George Shen <george.shen@amd.com>
Signed-off-by: Roman Li <roman.li@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/gpu/drm/amd/display/dc/dc_dp_types.h         | 12 ++++++++++++
 .../display/dc/link/protocols/link_dp_capability.c   |  6 +++++-
 2 files changed, 17 insertions(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/amd/display/dc/dc_dp_types.h b/drivers/gpu/drm/amd/display/dc/dc_dp_types.h
index cc005da75ce4c..8bb628ab78554 100644
--- a/drivers/gpu/drm/amd/display/dc/dc_dp_types.h
+++ b/drivers/gpu/drm/amd/display/dc/dc_dp_types.h
@@ -959,6 +959,14 @@ union dp_128b_132b_supported_lttpr_link_rates {
 	uint8_t raw;
 };
 
+union dp_alpm_lttpr_cap {
+	struct {
+		uint8_t AUX_LESS_ALPM_SUPPORTED	:1;
+		uint8_t RESERVED				:7;
+	} bits;
+	uint8_t raw;
+};
+
 union dp_sink_video_fallback_formats {
 	struct {
 		uint8_t dp_1024x768_60Hz_24bpp_support	:1;
@@ -1118,6 +1126,7 @@ struct dc_lttpr_caps {
 	uint8_t max_ext_timeout;
 	union dp_main_link_channel_coding_lttpr_cap main_link_channel_coding;
 	union dp_128b_132b_supported_lttpr_link_rates supported_128b_132b_rates;
+	union dp_alpm_lttpr_cap alpm;
 	uint8_t aux_rd_interval[MAX_REPEATER_CNT - 1];
 	uint8_t lttpr_ieee_oui[3];
 	uint8_t lttpr_device_id[6];
@@ -1372,6 +1381,9 @@ struct dp_trace {
 #ifndef DPCD_MAX_UNCOMPRESSED_PIXEL_RATE_CAP
 #define DPCD_MAX_UNCOMPRESSED_PIXEL_RATE_CAP    0x221c
 #endif
+#ifndef DP_LTTPR_ALPM_CAPABILITIES
+#define DP_LTTPR_ALPM_CAPABILITIES              0xF0009
+#endif
 #ifndef DP_REPEATER_CONFIGURATION_AND_STATUS_SIZE
 #define DP_REPEATER_CONFIGURATION_AND_STATUS_SIZE	0x50
 #endif
diff --git a/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_capability.c b/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_capability.c
index 28843e9882d39..64e4ae379e346 100644
--- a/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_capability.c
+++ b/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_capability.c
@@ -1502,7 +1502,7 @@ static bool dpcd_read_sink_ext_caps(struct dc_link *link)
 
 enum dc_status dp_retrieve_lttpr_cap(struct dc_link *link)
 {
-	uint8_t lttpr_dpcd_data[8] = {0};
+	uint8_t lttpr_dpcd_data[10] = {0};
 	enum dc_status status;
 	bool is_lttpr_present;
 
@@ -1552,6 +1552,10 @@ enum dc_status dp_retrieve_lttpr_cap(struct dc_link *link)
 			lttpr_dpcd_data[DP_PHY_REPEATER_128B132B_RATES -
 							DP_LT_TUNABLE_PHY_REPEATER_FIELD_DATA_STRUCTURE_REV];
 
+	link->dpcd_caps.lttpr_caps.alpm.raw =
+			lttpr_dpcd_data[DP_LTTPR_ALPM_CAPABILITIES -
+							DP_LT_TUNABLE_PHY_REPEATER_FIELD_DATA_STRUCTURE_REV];
+
 	/* If this chip cap is set, at least one retimer must exist in the chain
 	 * Override count to 1 if we receive a known bad count (0 or an invalid value) */
 	if (((link->chip_caps & AMD_EXT_DISPLAY_PATH_CAPS__EXT_CHIP_MASK) == AMD_EXT_DISPLAY_PATH_CAPS__DP_FIXED_VS_EN) &&
-- 
2.39.5

[PATCH AUTOSEL 6.14 388/642] Revert "drm/amd/display: Request HW cursor on DCN3.2 with SubVP"

[Sasha Levin]

From: Leo Zeng <Leo.Zeng@amd.com>

[ Upstream commit 8ae6dfc0b61b170cf13832d4cfe2a0c744e621a7 ]

This reverts commit 13437c91606c9232c747475e202fe3827cd53264.

Reason to revert: idle power regression found in testing.

Reviewed-by: Dillon Varone <dillon.varone@amd.com>
Signed-off-by: Leo Zeng <Leo.Zeng@amd.com>
Signed-off-by: Roman Li <roman.li@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/gpu/drm/amd/display/dc/dml/dcn32/dcn32_fpu.c | 1 -
 1 file changed, 1 deletion(-)

diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn32/dcn32_fpu.c b/drivers/gpu/drm/amd/display/dc/dml/dcn32/dcn32_fpu.c
index 56dda686e2992..6f490d8d7038c 100644
--- a/drivers/gpu/drm/amd/display/dc/dml/dcn32/dcn32_fpu.c
+++ b/drivers/gpu/drm/amd/display/dc/dml/dcn32/dcn32_fpu.c
@@ -626,7 +626,6 @@ static bool dcn32_assign_subvp_pipe(struct dc *dc,
 		 * - Not TMZ surface
 		 */
 		if (pipe->plane_state && !pipe->top_pipe && !pipe->prev_odm_pipe && !dcn32_is_center_timing(pipe) &&
-				!pipe->stream->hw_cursor_req &&
 				!(pipe->stream->timing.pix_clk_100hz / 10000 > DCN3_2_MAX_SUBVP_PIXEL_RATE_MHZ) &&
 				(!dcn32_is_psr_capable(pipe) || (context->stream_count == 1 && dc->caps.dmub_caps.subvp_psr)) &&
 				dc_state_get_pipe_subvp_type(context, pipe) == SUBVP_NONE &&
-- 
2.39.5

[PATCH AUTOSEL 6.14 389/642] drm/amd/display: Don't treat wb connector as physical in create_validate_stream_for_sink

[Sasha Levin]

From: Harry Wentland <harry.wentland@amd.com>

[ Upstream commit cbf4890c6f28fb1ad733e14613fbd33c2004bced ]

Don't try to operate on a drm_wb_connector as an amdgpu_dm_connector.
While dereferencing aconnector->base will "work" it's wrong and
might lead to unknown bad things. Just... don't.

Reviewed-by: Alex Hung <alex.hung@amd.com>
Signed-off-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Roman Li <roman.li@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 .../gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c | 26 ++++++++++++-------
 .../gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.h |  2 +-
 .../display/amdgpu_dm/amdgpu_dm_mst_types.c   |  6 ++---
 3 files changed, 20 insertions(+), 14 deletions(-)

diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
index 79c01541a0190..d51dbac3235ca 100644
--- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
@@ -7505,12 +7505,12 @@ static enum dc_status dm_validate_stream_and_context(struct dc *dc,
 }
 
 struct dc_stream_state *
-create_validate_stream_for_sink(struct amdgpu_dm_connector *aconnector,
+create_validate_stream_for_sink(struct drm_connector *connector,
 				const struct drm_display_mode *drm_mode,
 				const struct dm_connector_state *dm_state,
 				const struct dc_stream_state *old_stream)
 {
-	struct drm_connector *connector = &aconnector->base;
+	struct amdgpu_dm_connector *aconnector = NULL;
 	struct amdgpu_device *adev = drm_to_adev(connector->dev);
 	struct dc_stream_state *stream;
 	const struct drm_connector_state *drm_state = dm_state ? &dm_state->base : NULL;
@@ -7521,8 +7521,12 @@ create_validate_stream_for_sink(struct amdgpu_dm_connector *aconnector,
 	if (!dm_state)
 		return NULL;
 
-	if (aconnector->dc_link->connector_signal == SIGNAL_TYPE_HDMI_TYPE_A ||
-	    aconnector->dc_link->dpcd_caps.dongle_type == DISPLAY_DONGLE_DP_HDMI_CONVERTER)
+	if (connector->connector_type != DRM_MODE_CONNECTOR_WRITEBACK)
+		aconnector = to_amdgpu_dm_connector(connector);
+
+	if (aconnector &&
+	    (aconnector->dc_link->connector_signal == SIGNAL_TYPE_HDMI_TYPE_A ||
+	     aconnector->dc_link->dpcd_caps.dongle_type == DISPLAY_DONGLE_DP_HDMI_CONVERTER))
 		bpc_limit = 8;
 
 	do {
@@ -7534,10 +7538,11 @@ create_validate_stream_for_sink(struct amdgpu_dm_connector *aconnector,
 			break;
 		}
 
-		if (aconnector->base.connector_type == DRM_MODE_CONNECTOR_WRITEBACK)
+		dc_result = dc_validate_stream(adev->dm.dc, stream);
+
+		if (!aconnector) /* writeback connector */
 			return stream;
 
-		dc_result = dc_validate_stream(adev->dm.dc, stream);
 		if (dc_result == DC_OK && stream->signal == SIGNAL_TYPE_DISPLAY_PORT_MST)
 			dc_result = dm_dp_mst_is_port_support_mode(aconnector, stream);
 
@@ -7567,7 +7572,7 @@ create_validate_stream_for_sink(struct amdgpu_dm_connector *aconnector,
 				     __func__, __LINE__);
 
 		aconnector->force_yuv420_output = true;
-		stream = create_validate_stream_for_sink(aconnector, drm_mode,
+		stream = create_validate_stream_for_sink(connector, drm_mode,
 						dm_state, old_stream);
 		aconnector->force_yuv420_output = false;
 	}
@@ -7582,6 +7587,9 @@ enum drm_mode_status amdgpu_dm_connector_mode_valid(struct drm_connector *connec
 	struct dc_sink *dc_sink;
 	/* TODO: Unhardcode stream count */
 	struct dc_stream_state *stream;
+	/* we always have an amdgpu_dm_connector here since we got
+	 * here via the amdgpu_dm_connector_helper_funcs
+	 */
 	struct amdgpu_dm_connector *aconnector = to_amdgpu_dm_connector(connector);
 
 	if ((mode->flags & DRM_MODE_FLAG_INTERLACE) ||
@@ -7606,7 +7614,7 @@ enum drm_mode_status amdgpu_dm_connector_mode_valid(struct drm_connector *connec
 
 	drm_mode_set_crtcinfo(mode, 0);
 
-	stream = create_validate_stream_for_sink(aconnector, mode,
+	stream = create_validate_stream_for_sink(connector, mode,
 						 to_dm_connector_state(connector->state),
 						 NULL);
 	if (stream) {
@@ -10658,7 +10666,7 @@ static int dm_update_crtc_state(struct amdgpu_display_manager *dm,
 		if (!drm_atomic_crtc_needs_modeset(new_crtc_state))
 			goto skip_modeset;
 
-		new_stream = create_validate_stream_for_sink(aconnector,
+		new_stream = create_validate_stream_for_sink(connector,
 							     &new_crtc_state->mode,
 							     dm_new_conn_state,
 							     dm_old_crtc_state->stream);
diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.h b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.h
index d2703ca7dff31..195fec9048df7 100644
--- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.h
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.h
@@ -989,7 +989,7 @@ int amdgpu_dm_process_dmub_set_config_sync(struct dc_context *ctx, unsigned int
 					struct set_config_cmd_payload *payload, enum set_config_status *operation_result);
 
 struct dc_stream_state *
-	create_validate_stream_for_sink(struct amdgpu_dm_connector *aconnector,
+	create_validate_stream_for_sink(struct drm_connector *connector,
 					const struct drm_display_mode *drm_mode,
 					const struct dm_connector_state *dm_state,
 					const struct dc_stream_state *old_stream);
diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c
index 07e744da7bf41..056b171989565 100644
--- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c
@@ -1625,7 +1625,6 @@ int pre_validate_dsc(struct drm_atomic_state *state,
 
 		if (ind >= 0) {
 			struct drm_connector *connector;
-			struct amdgpu_dm_connector *aconnector;
 			struct drm_connector_state *drm_new_conn_state;
 			struct dm_connector_state *dm_new_conn_state;
 			struct dm_crtc_state *dm_old_crtc_state;
@@ -1633,15 +1632,14 @@ int pre_validate_dsc(struct drm_atomic_state *state,
 			connector =
 				amdgpu_dm_find_first_crtc_matching_connector(state,
 									     state->crtcs[ind].ptr);
-			aconnector = to_amdgpu_dm_connector(connector);
 			drm_new_conn_state =
 				drm_atomic_get_new_connector_state(state,
-								   &aconnector->base);
+								   connector);
 			dm_new_conn_state = to_dm_connector_state(drm_new_conn_state);
 			dm_old_crtc_state = to_dm_crtc_state(state->crtcs[ind].old_state);
 
 			local_dc_state->streams[i] =
-				create_validate_stream_for_sink(aconnector,
+				create_validate_stream_for_sink(connector,
 								&state->crtcs[ind].new_state->mode,
 								dm_new_conn_state,
 								dm_old_crtc_state->stream);
-- 
2.39.5

[PATCH AUTOSEL 6.14 458/642] drm/amdgpu/mes11: fix set_hw_resources_1 calculation

[Sasha Levin]

From: Alex Deucher <alexander.deucher@amd.com>

[ Upstream commit 1350dd3691b5f757a948e5b9895d62c422baeb90 ]

It's GPU page size not CPU page size.  In most cases they
are the same, but not always.  This can lead to overallocation
on systems with larger pages.

Cc: Srinivasan Shanmugam <srinivasan.shanmugam@amd.com>
Cc: Christian König <christian.koenig@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/gpu/drm/amd/amdgpu/mes_v11_0.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/mes_v11_0.c b/drivers/gpu/drm/amd/amdgpu/mes_v11_0.c
index 0f808ffcab943..68bb334393bb6 100644
--- a/drivers/gpu/drm/amd/amdgpu/mes_v11_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/mes_v11_0.c
@@ -730,7 +730,7 @@ static int mes_v11_0_set_hw_resources(struct amdgpu_mes *mes)
 
 static int mes_v11_0_set_hw_resources_1(struct amdgpu_mes *mes)
 {
-	int size = 128 * PAGE_SIZE;
+	int size = 128 * AMDGPU_GPU_PAGE_SIZE;
 	int ret = 0;
 	struct amdgpu_device *adev = mes->adev;
 	union MESAPI_SET_HW_RESOURCES_1 mes_set_hw_res_pkt;
-- 
2.39.5

[PATCH AUTOSEL 6.14 459/642] drm/amdkfd: fix missing L2 cache info in topology

[Sasha Levin]

From: Eric Huang <jinhuieric.huang@amd.com>

[ Upstream commit 5ffd56822a7159917306d99f18fd15dfd7288f20 ]

In some ASICs L2 cache info may miss in kfd topology,
because the first bitmap may be empty, that means
the first cu may be inactive, so to find the first
active cu will solve the issue.

v2: Only find the first active cu in the first xcc

Signed-off-by: Eric Huang <jinhuieric.huang@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Acked-by: Lijo Lazar <lijo.lazar@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/gpu/drm/amd/amdkfd/kfd_topology.c | 19 +++++++++++++++++--
 1 file changed, 17 insertions(+), 2 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_topology.c b/drivers/gpu/drm/amd/amdkfd/kfd_topology.c
index 334c576a75b14..98317eda2cdb4 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_topology.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_topology.c
@@ -1683,17 +1683,32 @@ static int fill_in_l2_l3_pcache(struct kfd_cache_properties **props_ext,
 				int cache_type, unsigned int cu_processor_id,
 				struct kfd_node *knode)
 {
-	unsigned int cu_sibling_map_mask;
+	unsigned int cu_sibling_map_mask = 0;
 	int first_active_cu;
 	int i, j, k, xcc, start, end;
 	int num_xcc = NUM_XCC(knode->xcc_mask);
 	struct kfd_cache_properties *pcache = NULL;
 	enum amdgpu_memory_partition mode;
 	struct amdgpu_device *adev = knode->adev;
+	bool found = false;
 
 	start = ffs(knode->xcc_mask) - 1;
 	end = start + num_xcc;
-	cu_sibling_map_mask = cu_info->bitmap[start][0][0];
+
+	/* To find the bitmap in the first active cu in the first
+	 * xcc, it is based on the assumption that evrey xcc must
+	 * have at least one active cu.
+	 */
+	for (i = 0; i < gfx_info->max_shader_engines && !found; i++) {
+		for (j = 0; j < gfx_info->max_sh_per_se && !found; j++) {
+			if (cu_info->bitmap[start][i % 4][j % 4]) {
+				cu_sibling_map_mask =
+					cu_info->bitmap[start][i % 4][j % 4];
+				found = true;
+			}
+		}
+	}
+
 	cu_sibling_map_mask &=
 		((1 << pcache_info[cache_type].num_cu_shared) - 1);
 	first_active_cu = ffs(cu_sibling_map_mask);
-- 
2.39.5

[PATCH AUTOSEL 6.14 460/642] drm/amdgpu: Set snoop bit for SDMA for MI series

[Sasha Levin]

From: Harish Kasiviswanathan <Harish.Kasiviswanathan@amd.com>

[ Upstream commit 3394b1f76d3f8adf695ceed350a5dae49003eb37 ]

SDMA writes has to probe invalidate RW lines. Set snoop bit in mmhub for
this to happen.

v2: Missed a few mmhub_v9_4. Added now.
v3: Calculate hub offset once since it doesn't change inside the loop
    Modified function names based on review comments.

Signed-off-by: Harish Kasiviswanathan <Harish.Kasiviswanathan@amd.com>
Reviewed-by: Philip Yang <Philip.Yang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/gpu/drm/amd/amdgpu/mmhub_v1_7.c       | 25 ++++++++++
 drivers/gpu/drm/amd/amdgpu/mmhub_v1_8.c       | 27 +++++++++++
 drivers/gpu/drm/amd/amdgpu/mmhub_v9_4.c       | 31 ++++++++++++
 .../asic_reg/mmhub/mmhub_9_4_1_offset.h       | 32 +++++++++++++
 .../asic_reg/mmhub/mmhub_9_4_1_sh_mask.h      | 48 +++++++++++++++++++
 5 files changed, 163 insertions(+)

diff --git a/drivers/gpu/drm/amd/amdgpu/mmhub_v1_7.c b/drivers/gpu/drm/amd/amdgpu/mmhub_v1_7.c
index 9689e2b5d4e51..2adee2b94c37d 100644
--- a/drivers/gpu/drm/amd/amdgpu/mmhub_v1_7.c
+++ b/drivers/gpu/drm/amd/amdgpu/mmhub_v1_7.c
@@ -172,6 +172,30 @@ static void mmhub_v1_7_init_tlb_regs(struct amdgpu_device *adev)
 	WREG32_SOC15(MMHUB, 0, regMC_VM_MX_L1_TLB_CNTL, tmp);
 }
 
+/* Set snoop bit for SDMA so that SDMA writes probe-invalidates RW lines */
+static void mmhub_v1_7_init_snoop_override_regs(struct amdgpu_device *adev)
+{
+	uint32_t tmp;
+	int i;
+	uint32_t distance = regDAGB1_WRCLI_GPU_SNOOP_OVERRIDE -
+			    regDAGB0_WRCLI_GPU_SNOOP_OVERRIDE;
+
+	for (i = 0; i < 5; i++) { /* DAGB instances */
+		tmp = RREG32_SOC15_OFFSET(MMHUB, 0,
+			regDAGB0_WRCLI_GPU_SNOOP_OVERRIDE, i * distance);
+		tmp |= (1 << 15); /* SDMA client is BIT15 */
+		WREG32_SOC15_OFFSET(MMHUB, 0,
+			regDAGB0_WRCLI_GPU_SNOOP_OVERRIDE, i * distance, tmp);
+
+		tmp = RREG32_SOC15_OFFSET(MMHUB, 0,
+			regDAGB0_WRCLI_GPU_SNOOP_OVERRIDE_VALUE, i * distance);
+		tmp |= (1 << 15);
+		WREG32_SOC15_OFFSET(MMHUB, 0,
+			regDAGB0_WRCLI_GPU_SNOOP_OVERRIDE_VALUE, i * distance, tmp);
+	}
+
+}
+
 static void mmhub_v1_7_init_cache_regs(struct amdgpu_device *adev)
 {
 	uint32_t tmp;
@@ -337,6 +361,7 @@ static int mmhub_v1_7_gart_enable(struct amdgpu_device *adev)
 	mmhub_v1_7_init_system_aperture_regs(adev);
 	mmhub_v1_7_init_tlb_regs(adev);
 	mmhub_v1_7_init_cache_regs(adev);
+	mmhub_v1_7_init_snoop_override_regs(adev);
 
 	mmhub_v1_7_enable_system_domain(adev);
 	mmhub_v1_7_disable_identity_aperture(adev);
diff --git a/drivers/gpu/drm/amd/amdgpu/mmhub_v1_8.c b/drivers/gpu/drm/amd/amdgpu/mmhub_v1_8.c
index e646e5cef0a2e..ce013a715b864 100644
--- a/drivers/gpu/drm/amd/amdgpu/mmhub_v1_8.c
+++ b/drivers/gpu/drm/amd/amdgpu/mmhub_v1_8.c
@@ -213,6 +213,32 @@ static void mmhub_v1_8_init_tlb_regs(struct amdgpu_device *adev)
 	}
 }
 
+/* Set snoop bit for SDMA so that SDMA writes probe-invalidates RW lines */
+static void mmhub_v1_8_init_snoop_override_regs(struct amdgpu_device *adev)
+{
+	uint32_t tmp, inst_mask;
+	int i, j;
+	uint32_t distance = regDAGB1_WRCLI_GPU_SNOOP_OVERRIDE -
+			    regDAGB0_WRCLI_GPU_SNOOP_OVERRIDE;
+
+	inst_mask = adev->aid_mask;
+	for_each_inst(i, inst_mask) {
+		for (j = 0; j < 5; j++) { /* DAGB instances */
+			tmp = RREG32_SOC15_OFFSET(MMHUB, i,
+				regDAGB0_WRCLI_GPU_SNOOP_OVERRIDE, j * distance);
+			tmp |= (1 << 15); /* SDMA client is BIT15 */
+			WREG32_SOC15_OFFSET(MMHUB, i,
+				regDAGB0_WRCLI_GPU_SNOOP_OVERRIDE, j * distance, tmp);
+
+			tmp = RREG32_SOC15_OFFSET(MMHUB, i,
+				regDAGB0_WRCLI_GPU_SNOOP_OVERRIDE_VALUE, j * distance);
+			tmp |= (1 << 15);
+			WREG32_SOC15_OFFSET(MMHUB, i,
+				regDAGB0_WRCLI_GPU_SNOOP_OVERRIDE_VALUE, j * distance, tmp);
+		}
+	}
+}
+
 static void mmhub_v1_8_init_cache_regs(struct amdgpu_device *adev)
 {
 	uint32_t tmp, inst_mask;
@@ -418,6 +444,7 @@ static int mmhub_v1_8_gart_enable(struct amdgpu_device *adev)
 	mmhub_v1_8_init_system_aperture_regs(adev);
 	mmhub_v1_8_init_tlb_regs(adev);
 	mmhub_v1_8_init_cache_regs(adev);
+	mmhub_v1_8_init_snoop_override_regs(adev);
 
 	mmhub_v1_8_enable_system_domain(adev);
 	mmhub_v1_8_disable_identity_aperture(adev);
diff --git a/drivers/gpu/drm/amd/amdgpu/mmhub_v9_4.c b/drivers/gpu/drm/amd/amdgpu/mmhub_v9_4.c
index ff1b58e446892..fe0710b55c3ac 100644
--- a/drivers/gpu/drm/amd/amdgpu/mmhub_v9_4.c
+++ b/drivers/gpu/drm/amd/amdgpu/mmhub_v9_4.c
@@ -198,6 +198,36 @@ static void mmhub_v9_4_init_tlb_regs(struct amdgpu_device *adev, int hubid)
 			    hubid * MMHUB_INSTANCE_REGISTER_OFFSET, tmp);
 }
 
+/* Set snoop bit for SDMA so that SDMA writes probe-invalidates RW lines */
+static void mmhub_v9_4_init_snoop_override_regs(struct amdgpu_device *adev, int hubid)
+{
+	uint32_t tmp;
+	int i;
+	uint32_t distance = mmDAGB1_WRCLI_GPU_SNOOP_OVERRIDE -
+			    mmDAGB0_WRCLI_GPU_SNOOP_OVERRIDE;
+	uint32_t huboffset = hubid * MMHUB_INSTANCE_REGISTER_OFFSET;
+
+	for (i = 0; i < 5 - (2 * hubid); i++) {
+		/* DAGB instances 0 to 4 are in hub0 and 5 to 7 are in hub1 */
+		tmp = RREG32_SOC15_OFFSET(MMHUB, 0,
+			mmDAGB0_WRCLI_GPU_SNOOP_OVERRIDE,
+			huboffset + i * distance);
+		tmp |= (1 << 15); /* SDMA client is BIT15 */
+		WREG32_SOC15_OFFSET(MMHUB, 0,
+			mmDAGB0_WRCLI_GPU_SNOOP_OVERRIDE,
+			huboffset + i * distance, tmp);
+
+		tmp = RREG32_SOC15_OFFSET(MMHUB, 0,
+			mmDAGB0_WRCLI_GPU_SNOOP_OVERRIDE_VALUE,
+			huboffset + i * distance);
+		tmp |= (1 << 15);
+		WREG32_SOC15_OFFSET(MMHUB, 0,
+			mmDAGB0_WRCLI_GPU_SNOOP_OVERRIDE_VALUE,
+			huboffset + i * distance, tmp);
+	}
+
+}
+
 static void mmhub_v9_4_init_cache_regs(struct amdgpu_device *adev, int hubid)
 {
 	uint32_t tmp;
@@ -392,6 +422,7 @@ static int mmhub_v9_4_gart_enable(struct amdgpu_device *adev)
 		if (!amdgpu_sriov_vf(adev))
 			mmhub_v9_4_init_cache_regs(adev, i);
 
+		mmhub_v9_4_init_snoop_override_regs(adev, i);
 		mmhub_v9_4_enable_system_domain(adev, i);
 		if (!amdgpu_sriov_vf(adev))
 			mmhub_v9_4_disable_identity_aperture(adev, i);
diff --git a/drivers/gpu/drm/amd/include/asic_reg/mmhub/mmhub_9_4_1_offset.h b/drivers/gpu/drm/amd/include/asic_reg/mmhub/mmhub_9_4_1_offset.h
index c488d4a50cf46..b2252deabc17a 100644
--- a/drivers/gpu/drm/amd/include/asic_reg/mmhub/mmhub_9_4_1_offset.h
+++ b/drivers/gpu/drm/amd/include/asic_reg/mmhub/mmhub_9_4_1_offset.h
@@ -203,6 +203,10 @@
 #define mmDAGB0_WR_DATA_CREDIT_BASE_IDX                                                                1
 #define mmDAGB0_WR_MISC_CREDIT                                                                         0x0058
 #define mmDAGB0_WR_MISC_CREDIT_BASE_IDX                                                                1
+#define mmDAGB0_WRCLI_GPU_SNOOP_OVERRIDE                                                               0x005b
+#define mmDAGB0_WRCLI_GPU_SNOOP_OVERRIDE_BASE_IDX                                                      1
+#define mmDAGB0_WRCLI_GPU_SNOOP_OVERRIDE_VALUE                                                         0x005c
+#define mmDAGB0_WRCLI_GPU_SNOOP_OVERRIDE_VALUE_BASE_IDX                                                1
 #define mmDAGB0_WRCLI_ASK_PENDING                                                                      0x005d
 #define mmDAGB0_WRCLI_ASK_PENDING_BASE_IDX                                                             1
 #define mmDAGB0_WRCLI_GO_PENDING                                                                       0x005e
@@ -455,6 +459,10 @@
 #define mmDAGB1_WR_DATA_CREDIT_BASE_IDX                                                                1
 #define mmDAGB1_WR_MISC_CREDIT                                                                         0x00d8
 #define mmDAGB1_WR_MISC_CREDIT_BASE_IDX                                                                1
+#define mmDAGB1_WRCLI_GPU_SNOOP_OVERRIDE                                                               0x00db
+#define mmDAGB1_WRCLI_GPU_SNOOP_OVERRIDE_BASE_IDX                                                      1
+#define mmDAGB1_WRCLI_GPU_SNOOP_OVERRIDE_VALUE                                                         0x00dc
+#define mmDAGB1_WRCLI_GPU_SNOOP_OVERRIDE_VALUE_BASE_IDX                                                1
 #define mmDAGB1_WRCLI_ASK_PENDING                                                                      0x00dd
 #define mmDAGB1_WRCLI_ASK_PENDING_BASE_IDX                                                             1
 #define mmDAGB1_WRCLI_GO_PENDING                                                                       0x00de
@@ -707,6 +715,10 @@
 #define mmDAGB2_WR_DATA_CREDIT_BASE_IDX                                                                1
 #define mmDAGB2_WR_MISC_CREDIT                                                                         0x0158
 #define mmDAGB2_WR_MISC_CREDIT_BASE_IDX                                                                1
+#define mmDAGB2_WRCLI_GPU_SNOOP_OVERRIDE                                                               0x015b
+#define mmDAGB2_WRCLI_GPU_SNOOP_OVERRIDE_BASE_IDX                                                      1
+#define mmDAGB2_WRCLI_GPU_SNOOP_OVERRIDE_VALUE                                                         0x015c
+#define mmDAGB2_WRCLI_GPU_SNOOP_OVERRIDE_VALUE_BASE_IDX                                                1
 #define mmDAGB2_WRCLI_ASK_PENDING                                                                      0x015d
 #define mmDAGB2_WRCLI_ASK_PENDING_BASE_IDX                                                             1
 #define mmDAGB2_WRCLI_GO_PENDING                                                                       0x015e
@@ -959,6 +971,10 @@
 #define mmDAGB3_WR_DATA_CREDIT_BASE_IDX                                                                1
 #define mmDAGB3_WR_MISC_CREDIT                                                                         0x01d8
 #define mmDAGB3_WR_MISC_CREDIT_BASE_IDX                                                                1
+#define mmDAGB3_WRCLI_GPU_SNOOP_OVERRIDE                                                               0x01db
+#define mmDAGB3_WRCLI_GPU_SNOOP_OVERRIDE_BASE_IDX                                                      1
+#define mmDAGB3_WRCLI_GPU_SNOOP_OVERRIDE_VALUE                                                         0x01dc
+#define mmDAGB3_WRCLI_GPU_SNOOP_OVERRIDE_VALUE_BASE_IDX                                                1
 #define mmDAGB3_WRCLI_ASK_PENDING                                                                      0x01dd
 #define mmDAGB3_WRCLI_ASK_PENDING_BASE_IDX                                                             1
 #define mmDAGB3_WRCLI_GO_PENDING                                                                       0x01de
@@ -1211,6 +1227,10 @@
 #define mmDAGB4_WR_DATA_CREDIT_BASE_IDX                                                                1
 #define mmDAGB4_WR_MISC_CREDIT                                                                         0x0258
 #define mmDAGB4_WR_MISC_CREDIT_BASE_IDX                                                                1
+#define mmDAGB4_WRCLI_GPU_SNOOP_OVERRIDE                                                               0x025b
+#define mmDAGB4_WRCLI_GPU_SNOOP_OVERRIDE_BASE_IDX                                                      1
+#define mmDAGB4_WRCLI_GPU_SNOOP_OVERRIDE_VALUE                                                         0x025c
+#define mmDAGB4_WRCLI_GPU_SNOOP_OVERRIDE_VALUE_BASE_IDX                                                1
 #define mmDAGB4_WRCLI_ASK_PENDING                                                                      0x025d
 #define mmDAGB4_WRCLI_ASK_PENDING_BASE_IDX                                                             1
 #define mmDAGB4_WRCLI_GO_PENDING                                                                       0x025e
@@ -4793,6 +4813,10 @@
 #define mmDAGB5_WR_DATA_CREDIT_BASE_IDX                                                                1
 #define mmDAGB5_WR_MISC_CREDIT                                                                         0x3058
 #define mmDAGB5_WR_MISC_CREDIT_BASE_IDX                                                                1
+#define mmDAGB5_WRCLI_GPU_SNOOP_OVERRIDE                                                               0x305b
+#define mmDAGB5_WRCLI_GPU_SNOOP_OVERRIDE_BASE_IDX                                                      1
+#define mmDAGB5_WRCLI_GPU_SNOOP_OVERRIDE_VALUE                                                         0x305c
+#define mmDAGB5_WRCLI_GPU_SNOOP_OVERRIDE_VALUE_BASE_IDX                                                1
 #define mmDAGB5_WRCLI_ASK_PENDING                                                                      0x305d
 #define mmDAGB5_WRCLI_ASK_PENDING_BASE_IDX                                                             1
 #define mmDAGB5_WRCLI_GO_PENDING                                                                       0x305e
@@ -5045,6 +5069,10 @@
 #define mmDAGB6_WR_DATA_CREDIT_BASE_IDX                                                                1
 #define mmDAGB6_WR_MISC_CREDIT                                                                         0x30d8
 #define mmDAGB6_WR_MISC_CREDIT_BASE_IDX                                                                1
+#define mmDAGB6_WRCLI_GPU_SNOOP_OVERRIDE                                                               0x30db
+#define mmDAGB6_WRCLI_GPU_SNOOP_OVERRIDE_BASE_IDX                                                      1
+#define mmDAGB6_WRCLI_GPU_SNOOP_OVERRIDE_VALUE                                                         0x30dc
+#define mmDAGB6_WRCLI_GPU_SNOOP_OVERRIDE_VALUE_BASE_IDX                                                1
 #define mmDAGB6_WRCLI_ASK_PENDING                                                                      0x30dd
 #define mmDAGB6_WRCLI_ASK_PENDING_BASE_IDX                                                             1
 #define mmDAGB6_WRCLI_GO_PENDING                                                                       0x30de
@@ -5297,6 +5325,10 @@
 #define mmDAGB7_WR_DATA_CREDIT_BASE_IDX                                                                1
 #define mmDAGB7_WR_MISC_CREDIT                                                                         0x3158
 #define mmDAGB7_WR_MISC_CREDIT_BASE_IDX                                                                1
+#define mmDAGB7_WRCLI_GPU_SNOOP_OVERRIDE                                                               0x315b
+#define mmDAGB7_WRCLI_GPU_SNOOP_OVERRIDE_BASE_IDX                                                      1
+#define mmDAGB7_WRCLI_GPU_SNOOP_OVERRIDE_VALUE                                                         0x315c
+#define mmDAGB7_WRCLI_GPU_SNOOP_OVERRIDE_VALUE_BASE_IDX                                                1
 #define mmDAGB7_WRCLI_ASK_PENDING                                                                      0x315d
 #define mmDAGB7_WRCLI_ASK_PENDING_BASE_IDX                                                             1
 #define mmDAGB7_WRCLI_GO_PENDING                                                                       0x315e
diff --git a/drivers/gpu/drm/amd/include/asic_reg/mmhub/mmhub_9_4_1_sh_mask.h b/drivers/gpu/drm/amd/include/asic_reg/mmhub/mmhub_9_4_1_sh_mask.h
index 2969fbf282b7d..5069d2fd467f2 100644
--- a/drivers/gpu/drm/amd/include/asic_reg/mmhub/mmhub_9_4_1_sh_mask.h
+++ b/drivers/gpu/drm/amd/include/asic_reg/mmhub/mmhub_9_4_1_sh_mask.h
@@ -1532,6 +1532,12 @@
 //DAGB0_WRCLI_DBUS_GO_PENDING
 #define DAGB0_WRCLI_DBUS_GO_PENDING__BUSY__SHIFT                                                              0x0
 #define DAGB0_WRCLI_DBUS_GO_PENDING__BUSY_MASK                                                                0xFFFFFFFFL
+//DAGB0_WRCLI_GPU_SNOOP_OVERRIDE
+#define DAGB0_WRCLI_GPU_SNOOP_OVERRIDE__ENABLE__SHIFT                                                         0x0
+#define DAGB0_WRCLI_GPU_SNOOP_OVERRIDE__ENABLE_MASK                                                           0xFFFFFFFFL
+//DAGB0_WRCLI_GPU_SNOOP_OVERRIDE_VALUE
+#define DAGB0_WRCLI_GPU_SNOOP_OVERRIDE_VALUE__ENABLE__SHIFT                                                   0x0
+#define DAGB0_WRCLI_GPU_SNOOP_OVERRIDE_VALUE__ENABLE_MASK                                                     0xFFFFFFFFL
 //DAGB0_DAGB_DLY
 #define DAGB0_DAGB_DLY__DLY__SHIFT                                                                            0x0
 #define DAGB0_DAGB_DLY__CLI__SHIFT                                                                            0x8
@@ -3207,6 +3213,12 @@
 //DAGB1_WRCLI_DBUS_GO_PENDING
 #define DAGB1_WRCLI_DBUS_GO_PENDING__BUSY__SHIFT                                                              0x0
 #define DAGB1_WRCLI_DBUS_GO_PENDING__BUSY_MASK                                                                0xFFFFFFFFL
+//DAGB1_WRCLI_GPU_SNOOP_OVERRIDE
+#define DAGB1_WRCLI_GPU_SNOOP_OVERRIDE__ENABLE__SHIFT                                                         0x0
+#define DAGB1_WRCLI_GPU_SNOOP_OVERRIDE__ENABLE_MASK                                                           0xFFFFFFFFL
+//DAGB1_WRCLI_GPU_SNOOP_OVERRIDE_VALUE
+#define DAGB1_WRCLI_GPU_SNOOP_OVERRIDE_VALUE__ENABLE__SHIFT                                                   0x0
+#define DAGB1_WRCLI_GPU_SNOOP_OVERRIDE_VALUE__ENABLE_MASK                                                     0xFFFFFFFFL
 //DAGB1_DAGB_DLY
 #define DAGB1_DAGB_DLY__DLY__SHIFT                                                                            0x0
 #define DAGB1_DAGB_DLY__CLI__SHIFT                                                                            0x8
@@ -4882,6 +4894,12 @@
 //DAGB2_WRCLI_DBUS_GO_PENDING
 #define DAGB2_WRCLI_DBUS_GO_PENDING__BUSY__SHIFT                                                              0x0
 #define DAGB2_WRCLI_DBUS_GO_PENDING__BUSY_MASK                                                                0xFFFFFFFFL
+//DAGB2_WRCLI_GPU_SNOOP_OVERRIDE
+#define DAGB2_WRCLI_GPU_SNOOP_OVERRIDE__ENABLE__SHIFT                                                         0x0
+#define DAGB2_WRCLI_GPU_SNOOP_OVERRIDE__ENABLE_MASK                                                           0xFFFFFFFFL
+//DAGB2_WRCLI_GPU_SNOOP_OVERRIDE_VALUE
+#define DAGB2_WRCLI_GPU_SNOOP_OVERRIDE_VALUE__ENABLE__SHIFT                                                   0x0
+#define DAGB2_WRCLI_GPU_SNOOP_OVERRIDE_VALUE__ENABLE_MASK                                                     0xFFFFFFFFL
 //DAGB2_DAGB_DLY
 #define DAGB2_DAGB_DLY__DLY__SHIFT                                                                            0x0
 #define DAGB2_DAGB_DLY__CLI__SHIFT                                                                            0x8
@@ -6557,6 +6575,12 @@
 //DAGB3_WRCLI_DBUS_GO_PENDING
 #define DAGB3_WRCLI_DBUS_GO_PENDING__BUSY__SHIFT                                                              0x0
 #define DAGB3_WRCLI_DBUS_GO_PENDING__BUSY_MASK                                                                0xFFFFFFFFL
+//DAGB3_WRCLI_GPU_SNOOP_OVERRIDE
+#define DAGB3_WRCLI_GPU_SNOOP_OVERRIDE__ENABLE__SHIFT                                                         0x0
+#define DAGB3_WRCLI_GPU_SNOOP_OVERRIDE__ENABLE_MASK                                                           0xFFFFFFFFL
+//DAGB3_WRCLI_GPU_SNOOP_OVERRIDE_VALUE
+#define DAGB3_WRCLI_GPU_SNOOP_OVERRIDE_VALUE__ENABLE__SHIFT                                                   0x0
+#define DAGB3_WRCLI_GPU_SNOOP_OVERRIDE_VALUE__ENABLE_MASK                                                     0xFFFFFFFFL
 //DAGB3_DAGB_DLY
 #define DAGB3_DAGB_DLY__DLY__SHIFT                                                                            0x0
 #define DAGB3_DAGB_DLY__CLI__SHIFT                                                                            0x8
@@ -8232,6 +8256,12 @@
 //DAGB4_WRCLI_DBUS_GO_PENDING
 #define DAGB4_WRCLI_DBUS_GO_PENDING__BUSY__SHIFT                                                              0x0
 #define DAGB4_WRCLI_DBUS_GO_PENDING__BUSY_MASK                                                                0xFFFFFFFFL
+//DAGB4_WRCLI_GPU_SNOOP_OVERRIDE
+#define DAGB4_WRCLI_GPU_SNOOP_OVERRIDE__ENABLE__SHIFT                                                         0x0
+#define DAGB4_WRCLI_GPU_SNOOP_OVERRIDE__ENABLE_MASK                                                           0xFFFFFFFFL
+//DAGB4_WRCLI_GPU_SNOOP_OVERRIDE_VALUE
+#define DAGB4_WRCLI_GPU_SNOOP_OVERRIDE_VALUE__ENABLE__SHIFT                                                   0x0
+#define DAGB4_WRCLI_GPU_SNOOP_OVERRIDE_VALUE__ENABLE_MASK                                                     0xFFFFFFFFL
 //DAGB4_DAGB_DLY
 #define DAGB4_DAGB_DLY__DLY__SHIFT                                                                            0x0
 #define DAGB4_DAGB_DLY__CLI__SHIFT                                                                            0x8
@@ -28737,6 +28767,12 @@
 //DAGB5_WRCLI_DBUS_GO_PENDING
 #define DAGB5_WRCLI_DBUS_GO_PENDING__BUSY__SHIFT                                                              0x0
 #define DAGB5_WRCLI_DBUS_GO_PENDING__BUSY_MASK                                                                0xFFFFFFFFL
+//DAGB5_WRCLI_GPU_SNOOP_OVERRIDE
+#define DAGB5_WRCLI_GPU_SNOOP_OVERRIDE__ENABLE__SHIFT                                                         0x0
+#define DAGB5_WRCLI_GPU_SNOOP_OVERRIDE__ENABLE_MASK                                                           0xFFFFFFFFL
+//DAGB5_WRCLI_GPU_SNOOP_OVERRIDE_VALUE
+#define DAGB5_WRCLI_GPU_SNOOP_OVERRIDE_VALUE__ENABLE__SHIFT                                                   0x0
+#define DAGB5_WRCLI_GPU_SNOOP_OVERRIDE_VALUE__ENABLE_MASK                                                     0xFFFFFFFFL
 //DAGB5_DAGB_DLY
 #define DAGB5_DAGB_DLY__DLY__SHIFT                                                                            0x0
 #define DAGB5_DAGB_DLY__CLI__SHIFT                                                                            0x8
@@ -30412,6 +30448,12 @@
 //DAGB6_WRCLI_DBUS_GO_PENDING
 #define DAGB6_WRCLI_DBUS_GO_PENDING__BUSY__SHIFT                                                              0x0
 #define DAGB6_WRCLI_DBUS_GO_PENDING__BUSY_MASK                                                                0xFFFFFFFFL
+//DAGB6_WRCLI_GPU_SNOOP_OVERRIDE
+#define DAGB6_WRCLI_GPU_SNOOP_OVERRIDE__ENABLE__SHIFT                                                         0x0
+#define DAGB6_WRCLI_GPU_SNOOP_OVERRIDE__ENABLE_MASK                                                           0xFFFFFFFFL
+//DAGB6_WRCLI_GPU_SNOOP_OVERRIDE_VALUE
+#define DAGB6_WRCLI_GPU_SNOOP_OVERRIDE_VALUE__ENABLE__SHIFT                                                   0x0
+#define DAGB6_WRCLI_GPU_SNOOP_OVERRIDE_VALUE__ENABLE_MASK                                                     0xFFFFFFFFL
 //DAGB6_DAGB_DLY
 #define DAGB6_DAGB_DLY__DLY__SHIFT                                                                            0x0
 #define DAGB6_DAGB_DLY__CLI__SHIFT                                                                            0x8
@@ -32087,6 +32129,12 @@
 //DAGB7_WRCLI_DBUS_GO_PENDING
 #define DAGB7_WRCLI_DBUS_GO_PENDING__BUSY__SHIFT                                                              0x0
 #define DAGB7_WRCLI_DBUS_GO_PENDING__BUSY_MASK                                                                0xFFFFFFFFL
+//DAGB7_WRCLI_GPU_SNOOP_OVERRIDE
+#define DAGB7_WRCLI_GPU_SNOOP_OVERRIDE__ENABLE__SHIFT                                                         0x0
+#define DAGB7_WRCLI_GPU_SNOOP_OVERRIDE__ENABLE_MASK                                                           0xFFFFFFFFL
+//DAGB7_WRCLI_GPU_SNOOP_OVERRIDE_VALUE
+#define DAGB7_WRCLI_GPU_SNOOP_OVERRIDE_VALUE__ENABLE__SHIFT                                                   0x0
+#define DAGB7_WRCLI_GPU_SNOOP_OVERRIDE_VALUE__ENABLE_MASK                                                     0xFFFFFFFFL
 //DAGB7_DAGB_DLY
 #define DAGB7_DAGB_DLY__DLY__SHIFT                                                                            0x0
 #define DAGB7_DAGB_DLY__CLI__SHIFT                                                                            0x8
-- 
2.39.5

[PATCH AUTOSEL 6.14 461/642] drm/amd/display: pass calculated dram_speed_mts to dml2

[Sasha Levin]

From: Charlene Liu <Charlene.Liu@amd.com>

[ Upstream commit b40d022ec06ade9f6c809091dc188422a0f0946d ]

[why]
currently dml2 is using a hard coded 16 to convert memclk to dram_speed_mts.
for apu, this depends on wck_ratio.

change to pass the already calculated dram_speed_mts from fpu to dml2.

v2: use existing calculation of dram_speed_mts for now to avoid regression

Signed-off-by: Charlene Liu <Charlene.Liu@amd.com>
Signed-off-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Reviewed-by: Roman Li <Roman.Li@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/gpu/drm/amd/display/dc/dml/dcn35/dcn35_fpu.c   | 2 ++
 drivers/gpu/drm/amd/display/dc/dml/dcn351/dcn351_fpu.c | 1 +
 drivers/gpu/drm/amd/display/dc/dml2/dml2_wrapper.h     | 1 +
 3 files changed, 4 insertions(+)

diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn35/dcn35_fpu.c b/drivers/gpu/drm/amd/display/dc/dml/dcn35/dcn35_fpu.c
index 47d785204f29c..e8efffcc69a16 100644
--- a/drivers/gpu/drm/amd/display/dc/dml/dcn35/dcn35_fpu.c
+++ b/drivers/gpu/drm/amd/display/dc/dml/dcn35/dcn35_fpu.c
@@ -367,6 +367,8 @@ void dcn35_update_bw_bounding_box_fpu(struct dc *dc,
 				clock_limits[i].socclk_mhz;
 			dc->dml2_options.bbox_overrides.clks_table.clk_entries[i].memclk_mhz =
 				clk_table->entries[i].memclk_mhz * clk_table->entries[i].wck_ratio;
+
+			dc->dml2_options.bbox_overrides.clks_table.clk_entries[i].dram_speed_mts = clock_limits[i].dram_speed_mts;
 			dc->dml2_options.bbox_overrides.clks_table.clk_entries[i].dtbclk_mhz =
 				clock_limits[i].dtbclk_mhz;
 			dc->dml2_options.bbox_overrides.clks_table.num_entries_per_clk.num_dcfclk_levels =
diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn351/dcn351_fpu.c b/drivers/gpu/drm/amd/display/dc/dml/dcn351/dcn351_fpu.c
index d9e63c4fdd95c..17d0b4923b0cc 100644
--- a/drivers/gpu/drm/amd/display/dc/dml/dcn351/dcn351_fpu.c
+++ b/drivers/gpu/drm/amd/display/dc/dml/dcn351/dcn351_fpu.c
@@ -401,6 +401,7 @@ void dcn351_update_bw_bounding_box_fpu(struct dc *dc,
 				clock_limits[i].socclk_mhz;
 			dc->dml2_options.bbox_overrides.clks_table.clk_entries[i].memclk_mhz =
 				clk_table->entries[i].memclk_mhz * clk_table->entries[i].wck_ratio;
+			dc->dml2_options.bbox_overrides.clks_table.clk_entries[i].dram_speed_mts = clock_limits[i].dram_speed_mts;
 			dc->dml2_options.bbox_overrides.clks_table.clk_entries[i].dtbclk_mhz =
 				clock_limits[i].dtbclk_mhz;
 			dc->dml2_options.bbox_overrides.clks_table.num_entries_per_clk.num_dcfclk_levels =
diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml2_wrapper.h b/drivers/gpu/drm/amd/display/dc/dml2/dml2_wrapper.h
index 0f944fcfd5a5b..785226945699d 100644
--- a/drivers/gpu/drm/amd/display/dc/dml2/dml2_wrapper.h
+++ b/drivers/gpu/drm/amd/display/dc/dml2/dml2_wrapper.h
@@ -159,6 +159,7 @@ struct dml2_clks_table_entry {
 	unsigned int dtbclk_mhz;
 	unsigned int dispclk_mhz;
 	unsigned int dppclk_mhz;
+	unsigned int dram_speed_mts; /*which is based on wck_ratio*/
 };
 
 struct dml2_clks_num_entries {
-- 
2.39.5

[PATCH AUTOSEL 6.14 462/642] drm/amd/display: remove TF check for LLS policy

[Sasha Levin]

From: Samson Tam <Samson.Tam@amd.com>

[ Upstream commit 2a4519c4e9b2e1f622ab4c5f5841abdb9760cb0b ]

[Why & How]
LLS policy not affected by TF.
Remove check in don't care case and use
 pixel format only.

Reviewed-by: Navid Assadian <navid.assadian@amd.com>
Signed-off-by: Samson Tam <Samson.Tam@amd.com>
Signed-off-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/gpu/drm/amd/display/dc/spl/dc_spl.c | 31 +++++----------------
 1 file changed, 7 insertions(+), 24 deletions(-)

diff --git a/drivers/gpu/drm/amd/display/dc/spl/dc_spl.c b/drivers/gpu/drm/amd/display/dc/spl/dc_spl.c
index 153b7a8904e1e..047f05ab01810 100644
--- a/drivers/gpu/drm/amd/display/dc/spl/dc_spl.c
+++ b/drivers/gpu/drm/amd/display/dc/spl/dc_spl.c
@@ -784,25 +784,13 @@ static enum scl_mode spl_get_dscl_mode(const struct spl_in *spl_in,
 	return SCL_MODE_SCALING_420_YCBCR_ENABLE;
 }
 
-static bool spl_choose_lls_policy(enum spl_pixel_format format,
-	enum spl_transfer_func_type tf_type,
-	enum spl_transfer_func_predefined tf_predefined_type,
+static void spl_choose_lls_policy(enum spl_pixel_format format,
 	enum linear_light_scaling *lls_pref)
 {
-	if (spl_is_video_format(format)) {
+	if (spl_is_subsampled_format(format))
 		*lls_pref = LLS_PREF_NO;
-		if ((tf_type == SPL_TF_TYPE_PREDEFINED) ||
-			(tf_type == SPL_TF_TYPE_DISTRIBUTED_POINTS))
-			return true;
-	} else { /* RGB or YUV444 */
-		if ((tf_type == SPL_TF_TYPE_PREDEFINED) ||
-			(tf_type == SPL_TF_TYPE_BYPASS)) {
-			*lls_pref = LLS_PREF_YES;
-			return true;
-		}
-	}
-	*lls_pref = LLS_PREF_NO;
-	return false;
+	else /* RGB or YUV444 */
+		*lls_pref = LLS_PREF_YES;
 }
 
 /* Enable EASF ?*/
@@ -811,7 +799,6 @@ static bool enable_easf(struct spl_in *spl_in, struct spl_scratch *spl_scratch)
 	int vratio = 0;
 	int hratio = 0;
 	bool skip_easf = false;
-	bool lls_enable_easf = true;
 
 	if (spl_in->disable_easf)
 		skip_easf = true;
@@ -827,17 +814,13 @@ static bool enable_easf(struct spl_in *spl_in, struct spl_scratch *spl_scratch)
 		skip_easf = true;
 
 	/*
-	 * If lls_pref is LLS_PREF_DONT_CARE, then use pixel format and transfer
-	 *  function to determine whether to use LINEAR or NONLINEAR scaling
+	 * If lls_pref is LLS_PREF_DONT_CARE, then use pixel format
+	 *  to determine whether to use LINEAR or NONLINEAR scaling
 	 */
 	if (spl_in->lls_pref == LLS_PREF_DONT_CARE)
-		lls_enable_easf = spl_choose_lls_policy(spl_in->basic_in.format,
-			spl_in->basic_in.tf_type, spl_in->basic_in.tf_predefined_type,
+		spl_choose_lls_policy(spl_in->basic_in.format,
 			&spl_in->lls_pref);
 
-	if (!lls_enable_easf)
-		skip_easf = true;
-
 	/* Check for linear scaling or EASF preferred */
 	if (spl_in->lls_pref != LLS_PREF_YES && !spl_in->prefer_easf)
 		skip_easf = true;
-- 
2.39.5

[PATCH AUTOSEL 6.14 463/642] drm/amd/display: Don't try AUX transactions on disconnected link

[Sasha Levin]

From: Ilya Bakoulin <Ilya.Bakoulin@amd.com>

[ Upstream commit e8bffa52e0253cfd689813a620e64521256bc712 ]

[Why]
Setting link DPMS off in response to HPD disconnect creates AUX
transactions on a link that is supposed to be disconnected. This can
cause issues in some cases when the sink re-asserts HPD and expects
source to re-enable the link.

[How]
Avoid AUX transactions on disconnected link.

Reviewed-by: Wenjing Liu <wenjing.liu@amd.com>
Signed-off-by: Ilya Bakoulin <Ilya.Bakoulin@amd.com>
Signed-off-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 .../gpu/drm/amd/display/dc/link/protocols/link_dp_phy.c   | 8 +++++---
 1 file changed, 5 insertions(+), 3 deletions(-)

diff --git a/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_phy.c b/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_phy.c
index 2c73ac87cd665..c27ffec5d84fb 100644
--- a/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_phy.c
+++ b/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_phy.c
@@ -75,7 +75,8 @@ void dp_disable_link_phy(struct dc_link *link,
 	struct dc  *dc = link->ctx->dc;
 
 	if (!link->wa_flags.dp_keep_receiver_powered &&
-		!link->skip_implict_edp_power_control)
+			!link->skip_implict_edp_power_control &&
+			link->type != dc_connection_none)
 		dpcd_write_rx_power_ctrl(link, false);
 
 	dc->hwss.disable_link_output(link, link_res, signal);
@@ -163,8 +164,9 @@ enum dc_status dp_set_fec_ready(struct dc_link *link, const struct link_resource
 	} else {
 		if (link->fec_state == dc_link_fec_ready) {
 			fec_config = 0;
-			core_link_write_dpcd(link, DP_FEC_CONFIGURATION,
-				&fec_config, sizeof(fec_config));
+			if (link->type != dc_connection_none)
+				core_link_write_dpcd(link, DP_FEC_CONFIGURATION,
+					&fec_config, sizeof(fec_config));
 
 			link_enc->funcs->fec_set_ready(link_enc, false);
 			link->fec_state = dc_link_fec_not_ready;
-- 
2.39.5

[PATCH AUTOSEL 6.14 464/642] drm/amdgpu: reset psp->cmd to NULL after releasing the buffer

[Sasha Levin]

From: Jiang Liu <gerry@linux.alibaba.com>

[ Upstream commit e92f3f94cad24154fd3baae30c6dfb918492278d ]

Reset psp->cmd to NULL after releasing the buffer in function psp_sw_fini().

Reviewed-by: Lijo Lazar <lijo.lazar@amd.com>
Signed-off-by: Jiang Liu <gerry@linux.alibaba.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c | 5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c
index e5fc80ed06eae..665cc277cdc05 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c
@@ -533,7 +533,6 @@ static int psp_sw_fini(struct amdgpu_ip_block *ip_block)
 {
 	struct amdgpu_device *adev = ip_block->adev;
 	struct psp_context *psp = &adev->psp;
-	struct psp_gfx_cmd_resp *cmd = psp->cmd;
 
 	psp_memory_training_fini(psp);
 
@@ -543,8 +542,8 @@ static int psp_sw_fini(struct amdgpu_ip_block *ip_block)
 	amdgpu_ucode_release(&psp->cap_fw);
 	amdgpu_ucode_release(&psp->toc_fw);
 
-	kfree(cmd);
-	cmd = NULL;
+	kfree(psp->cmd);
+	psp->cmd = NULL;
 
 	psp_free_shared_bufs(psp);
 
-- 
2.39.5

[PATCH AUTOSEL 6.14 465/642] drm/amd/pm: Skip P2S load for SMU v13.0.12

[Sasha Levin]

From: Asad Kamal <asad.kamal@amd.com>

[ Upstream commit 1fb85819d629676f1d53f40c3fffa25a33a881e4 ]

Skip P2S table load for SMU v13.0.12

Signed-off-by: Asad Kamal <asad.kamal@amd.com>
Reviewed-by: Lijo Lazar <lijo.lazar@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_6_ppt.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_6_ppt.c b/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_6_ppt.c
index da7bd9227afeb..5f2a824918e3b 100644
--- a/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_6_ppt.c
+++ b/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_6_ppt.c
@@ -450,8 +450,9 @@ static int smu_v13_0_6_init_microcode(struct smu_context *smu)
 	int var = (adev->pdev->device & 0xF);
 	char ucode_prefix[15];
 
-	/* No need to load P2S tables in IOV mode */
-	if (amdgpu_sriov_vf(adev))
+	/* No need to load P2S tables in IOV mode or for smu v13.0.12 */
+	if (amdgpu_sriov_vf(adev) ||
+	    (amdgpu_ip_version(smu->adev, MP1_HWIP, 0) == IP_VERSION(13, 0, 12)))
 		return 0;
 
 	if (!(adev->flags & AMD_IS_APU)) {
-- 
2.39.5

[PATCH AUTOSEL 6.14 466/642] drm/amd/display: Support multiple options during psr entry.

[Sasha Levin]

From: Martin Tsai <Martin.Tsai@amd.com>

[ Upstream commit 3a5fa55455db6a11248a25f24570c365f9246144 ]

[WHY]
Some panels may not handle idle pattern properly during PSR entry.

[HOW]
Add a condition to allow multiple options on power down
sequence during PSR1 entry.

Reviewed-by: Anthony Koo <anthony.koo@amd.com>
Signed-off-by: Martin Tsai <Martin.Tsai@amd.com>
Signed-off-by: Alex Hung <alex.hung@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/gpu/drm/amd/display/dc/dc_types.h       | 7 +++++++
 drivers/gpu/drm/amd/display/dc/dce/dmub_psr.c   | 4 ++++
 drivers/gpu/drm/amd/display/dmub/inc/dmub_cmd.h | 6 ++++++
 3 files changed, 17 insertions(+)

diff --git a/drivers/gpu/drm/amd/display/dc/dc_types.h b/drivers/gpu/drm/amd/display/dc/dc_types.h
index 0c2aa91f0a111..e60898c2df01a 100644
--- a/drivers/gpu/drm/amd/display/dc/dc_types.h
+++ b/drivers/gpu/drm/amd/display/dc/dc_types.h
@@ -1033,6 +1033,13 @@ struct psr_settings {
 	unsigned int psr_sdp_transmit_line_num_deadline;
 	uint8_t force_ffu_mode;
 	unsigned int psr_power_opt;
+
+	/**
+	 * Some panels cannot handle idle pattern during PSR entry.
+	 * To power down phy before disable stream to avoid sending
+	 * idle pattern.
+	 */
+	uint8_t power_down_phy_before_disable_stream;
 };
 
 enum replay_coasting_vtotal_type {
diff --git a/drivers/gpu/drm/amd/display/dc/dce/dmub_psr.c b/drivers/gpu/drm/amd/display/dc/dce/dmub_psr.c
index 88c75c243bf8a..ff3b8244ba3d0 100644
--- a/drivers/gpu/drm/amd/display/dc/dce/dmub_psr.c
+++ b/drivers/gpu/drm/amd/display/dc/dce/dmub_psr.c
@@ -418,6 +418,10 @@ static bool dmub_psr_copy_settings(struct dmub_psr *dmub,
 	copy_settings_data->relock_delay_frame_cnt = 0;
 	if (link->dpcd_caps.sink_dev_id == DP_BRANCH_DEVICE_ID_001CF8)
 		copy_settings_data->relock_delay_frame_cnt = 2;
+
+	copy_settings_data->power_down_phy_before_disable_stream =
+		link->psr_settings.power_down_phy_before_disable_stream;
+
 	copy_settings_data->dsc_slice_height = psr_context->dsc_slice_height;
 
 	dc_wake_and_execute_dmub_cmd(dc, &cmd, DM_DMUB_WAIT_TYPE_WAIT);
diff --git a/drivers/gpu/drm/amd/display/dmub/inc/dmub_cmd.h b/drivers/gpu/drm/amd/display/dmub/inc/dmub_cmd.h
index d0fe324cb5371..8cf89aed024b7 100644
--- a/drivers/gpu/drm/amd/display/dmub/inc/dmub_cmd.h
+++ b/drivers/gpu/drm/amd/display/dmub/inc/dmub_cmd.h
@@ -3118,6 +3118,12 @@ struct dmub_cmd_psr_copy_settings_data {
 	 * Some panels request main link off before xth vertical line
 	 */
 	uint16_t poweroff_before_vertical_line;
+	/**
+	 * Some panels cannot handle idle pattern during PSR entry.
+	 * To power down phy before disable stream to avoid sending
+	 * idle pattern.
+	 */
+	uint8_t power_down_phy_before_disable_stream;
 };
 
 /**
-- 
2.39.5

[PATCH AUTOSEL 6.14 467/642] Revert "drm/amd/display: Exit idle optimizations before attempt to access PHY"

[Sasha Levin]

From: Brandon Syu <Brandon.Syu@amd.com>

[ Upstream commit be704e5ef4bd66dee9bb3f876964327e3a247d31 ]

This reverts commit de612738e9771bd66aeb20044486c457c512f684.

Reason to revert: screen flashes or gray screen appeared half of the
screen after resume from S4/S5.

Reviewed-by: Charlene Liu <charlene.liu@amd.com>
Signed-off-by: Brandon Syu <Brandon.Syu@amd.com>
Signed-off-by: Alex Hung <alex.hung@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/gpu/drm/amd/display/dc/hwss/dce110/dce110_hwseq.c | 5 -----
 1 file changed, 5 deletions(-)

diff --git a/drivers/gpu/drm/amd/display/dc/hwss/dce110/dce110_hwseq.c b/drivers/gpu/drm/amd/display/dc/hwss/dce110/dce110_hwseq.c
index 94ceccfc04982..2f5f3e749a1ab 100644
--- a/drivers/gpu/drm/amd/display/dc/hwss/dce110/dce110_hwseq.c
+++ b/drivers/gpu/drm/amd/display/dc/hwss/dce110/dce110_hwseq.c
@@ -1889,7 +1889,6 @@ void dce110_enable_accelerated_mode(struct dc *dc, struct dc_state *context)
 	bool can_apply_edp_fast_boot = false;
 	bool can_apply_seamless_boot = false;
 	bool keep_edp_vdd_on = false;
-	struct dc_bios *dcb = dc->ctx->dc_bios;
 	DC_LOGGER_INIT();
 
 
@@ -1966,8 +1965,6 @@ void dce110_enable_accelerated_mode(struct dc *dc, struct dc_state *context)
 			hws->funcs.edp_backlight_control(edp_link_with_sink, false);
 		}
 		/*resume from S3, no vbios posting, no need to power down again*/
-		if (dcb && dcb->funcs && !dcb->funcs->is_accelerated_mode(dcb))
-			clk_mgr_exit_optimized_pwr_state(dc, dc->clk_mgr);
 
 		power_down_all_hw_blocks(dc);
 
@@ -1980,8 +1977,6 @@ void dce110_enable_accelerated_mode(struct dc *dc, struct dc_state *context)
 		disable_vga_and_power_gate_all_controllers(dc);
 		if (edp_link_with_sink && !keep_edp_vdd_on)
 			dc->hwss.edp_power_control(edp_link_with_sink, false);
-		if (dcb && dcb->funcs && !dcb->funcs->is_accelerated_mode(dcb))
-			clk_mgr_optimize_pwr_state(dc, dc->clk_mgr);
 	}
 	bios_set_scratch_acc_mode_change(dc->ctx->dc_bios, 1);
 }
-- 
2.39.5

[PATCH AUTOSEL 6.14 468/642] drm/amd/display: Fixes for mcache programming in DML21

[Sasha Levin]

From: Dillon Varone <dillon.varone@amd.com>

[ Upstream commit c909a49128a31bced8cfbd2dfb0a4fe56e01a6d0 ]

[WHY & HOW]
- Fix indexing phantom planes for mcache programming in the wrapper
- Fix phantom mcache allocations to align with HW guidance
- Fix mcache assignment for chroma plane for multi-planar formats

Reviewed-by: Austin Zheng <Austin.Zheng@amd.com>
Signed-off-by: Dillon Varone <dillon.varone@amd.com>
Signed-off-by: Alex Hung <alex.hung@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 .../amd/display/dc/dml2/dml21/dml21_utils.c   |  1 -
 .../amd/display/dc/dml2/dml21/dml21_wrapper.c | 14 +++++++++
 .../display/dc/dml2/dml21/inc/dml_top_types.h |  1 +
 .../dml2/dml21/src/dml2_core/dml2_core_dcn4.c | 30 ++++++++++++++++++-
 .../src/dml2_core/dml2_core_dcn4_calcs.c      |  3 ++
 .../dml21/src/dml2_pmo/dml2_pmo_dcn4_fams2.c  | 21 +++++++++++++
 .../dml2/dml21/src/dml2_top/dml2_top_soc15.c  |  8 -----
 7 files changed, 68 insertions(+), 10 deletions(-)

diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/dml21_utils.c b/drivers/gpu/drm/amd/display/dc/dml2/dml21/dml21_utils.c
index 1e56d995cd0e7..930e86cdb88a2 100644
--- a/drivers/gpu/drm/amd/display/dc/dml2/dml21/dml21_utils.c
+++ b/drivers/gpu/drm/amd/display/dc/dml2/dml21/dml21_utils.c
@@ -232,7 +232,6 @@ void dml21_program_dc_pipe(struct dml2_context *dml_ctx, struct dc_state *contex
 		context->bw_ctx.bw.dcn.clk.dppclk_khz = pipe_ctx->plane_res.bw.dppclk_khz;
 
 	dml21_populate_mall_allocation_size(context, dml_ctx, pln_prog, pipe_ctx);
-	memcpy(&context->bw_ctx.bw.dcn.mcache_allocations[pipe_ctx->pipe_idx], &pln_prog->mcache_allocation, sizeof(struct dml2_mcache_surface_allocation));
 
 	bool sub_vp_enabled = is_sub_vp_enabled(pipe_ctx->stream->ctx->dc, context);
 
diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/dml21_wrapper.c b/drivers/gpu/drm/amd/display/dc/dml2/dml21/dml21_wrapper.c
index d6fd13f43c08f..5d16f36ec95c8 100644
--- a/drivers/gpu/drm/amd/display/dc/dml2/dml21/dml21_wrapper.c
+++ b/drivers/gpu/drm/amd/display/dc/dml2/dml21/dml21_wrapper.c
@@ -129,6 +129,7 @@ static void dml21_calculate_rq_and_dlg_params(const struct dc *dc, struct dc_sta
 	struct pipe_ctx *dc_main_pipes[__DML2_WRAPPER_MAX_STREAMS_PLANES__];
 	struct pipe_ctx *dc_phantom_pipes[__DML2_WRAPPER_MAX_STREAMS_PLANES__] = {0};
 	int num_pipes;
+	unsigned int dml_phantom_prog_idx;
 
 	context->bw_ctx.bw.dcn.clk.dppclk_khz = 0;
 
@@ -142,6 +143,9 @@ static void dml21_calculate_rq_and_dlg_params(const struct dc *dc, struct dc_sta
 	context->bw_ctx.bw.dcn.mall_ss_psr_active_size_bytes = 0;
 	context->bw_ctx.bw.dcn.mall_subvp_size_bytes = 0;
 
+	/* phantom's start after main planes */
+	dml_phantom_prog_idx = in_ctx->v21.mode_programming.programming->display_config.num_planes;
+
 	for (dml_prog_idx = 0; dml_prog_idx < DML2_MAX_PLANES; dml_prog_idx++) {
 		pln_prog = &in_ctx->v21.mode_programming.programming->plane_programming[dml_prog_idx];
 
@@ -167,6 +171,16 @@ static void dml21_calculate_rq_and_dlg_params(const struct dc *dc, struct dc_sta
 				dml21_program_dc_pipe(in_ctx, context, dc_phantom_pipes[dc_pipe_index], pln_prog, stream_prog);
 			}
 		}
+
+		/* copy per plane mcache allocation */
+		memcpy(&context->bw_ctx.bw.dcn.mcache_allocations[dml_prog_idx], &pln_prog->mcache_allocation, sizeof(struct dml2_mcache_surface_allocation));
+		if (pln_prog->phantom_plane.valid) {
+			memcpy(&context->bw_ctx.bw.dcn.mcache_allocations[dml_phantom_prog_idx],
+					&pln_prog->phantom_plane.mcache_allocation,
+					sizeof(struct dml2_mcache_surface_allocation));
+
+			dml_phantom_prog_idx++;
+		}
 	}
 
 	/* assign global clocks */
diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/inc/dml_top_types.h b/drivers/gpu/drm/amd/display/dc/dml2/dml21/inc/dml_top_types.h
index d2d053f2354d0..0ab19cf4d2421 100644
--- a/drivers/gpu/drm/amd/display/dc/dml2/dml21/inc/dml_top_types.h
+++ b/drivers/gpu/drm/amd/display/dc/dml2/dml21/inc/dml_top_types.h
@@ -245,6 +245,7 @@ struct dml2_per_plane_programming {
 	struct {
 		bool valid;
 		struct dml2_plane_parameters descriptor;
+		struct dml2_mcache_surface_allocation mcache_allocation;
 		struct dml2_dchub_per_pipe_register_set *pipe_regs[DML2_MAX_PLANES];
 	} phantom_plane;
 };
diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_dcn4.c b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_dcn4.c
index 7216d25c783e6..44d2969a904ea 100644
--- a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_dcn4.c
+++ b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_dcn4.c
@@ -253,7 +253,8 @@ static void expand_implict_subvp(const struct display_configuation_with_meta *di
 static void pack_mode_programming_params_with_implicit_subvp(struct dml2_core_instance *core, const struct display_configuation_with_meta *display_cfg,
 	const struct dml2_display_cfg *svp_expanded_display_cfg, struct dml2_display_cfg_programming *programming, struct dml2_core_scratch *scratch)
 {
-	unsigned int stream_index, plane_index, pipe_offset, stream_already_populated_mask, main_plane_index;
+	unsigned int stream_index, plane_index, pipe_offset, stream_already_populated_mask, main_plane_index, mcache_index;
+	unsigned int total_main_mcaches_required = 0;
 	int total_pipe_regs_copied = 0;
 	int dml_internal_pipe_index = 0;
 	const struct dml2_plane_parameters *main_plane;
@@ -324,6 +325,13 @@ static void pack_mode_programming_params_with_implicit_subvp(struct dml2_core_in
 
 		dml2_core_calcs_get_mall_allocation(&core->clean_me_up.mode_lib, &programming->plane_programming[plane_index].surface_size_mall_bytes, dml_internal_pipe_index);
 
+		memcpy(&programming->plane_programming[plane_index].mcache_allocation,
+				&display_cfg->stage2.mcache_allocations[plane_index],
+				sizeof(struct dml2_mcache_surface_allocation));
+		total_main_mcaches_required += programming->plane_programming[plane_index].mcache_allocation.num_mcaches_plane0 +
+				programming->plane_programming[plane_index].mcache_allocation.num_mcaches_plane1 -
+				(programming->plane_programming[plane_index].mcache_allocation.last_slice_sharing.plane0_plane1 ? 1 : 0);
+
 		for (pipe_offset = 0; pipe_offset < programming->plane_programming[plane_index].num_dpps_required; pipe_offset++) {
 			// Assign storage for this pipe's register values
 			programming->plane_programming[plane_index].pipe_regs[pipe_offset] = &programming->pipe_regs[total_pipe_regs_copied];
@@ -362,6 +370,22 @@ static void pack_mode_programming_params_with_implicit_subvp(struct dml2_core_in
 		memcpy(&programming->plane_programming[main_plane_index].phantom_plane.descriptor, phantom_plane, sizeof(struct dml2_plane_parameters));
 
 		dml2_core_calcs_get_mall_allocation(&core->clean_me_up.mode_lib, &programming->plane_programming[main_plane_index].svp_size_mall_bytes, dml_internal_pipe_index);
+
+		/* generate mcache allocation, phantoms use identical mcache configuration, but in the MALL set and unique mcache ID's beginning after all main ID's */
+		memcpy(&programming->plane_programming[main_plane_index].phantom_plane.mcache_allocation,
+				&programming->plane_programming[main_plane_index].mcache_allocation,
+				sizeof(struct dml2_mcache_surface_allocation));
+		for (mcache_index = 0; mcache_index < programming->plane_programming[main_plane_index].phantom_plane.mcache_allocation.num_mcaches_plane0; mcache_index++) {
+			programming->plane_programming[main_plane_index].phantom_plane.mcache_allocation.global_mcache_ids_plane0[mcache_index] += total_main_mcaches_required;
+			programming->plane_programming[main_plane_index].phantom_plane.mcache_allocation.global_mcache_ids_mall_plane0[mcache_index] =
+					programming->plane_programming[main_plane_index].phantom_plane.mcache_allocation.global_mcache_ids_plane0[mcache_index];
+		}
+		for (mcache_index = 0; mcache_index < programming->plane_programming[main_plane_index].phantom_plane.mcache_allocation.num_mcaches_plane1; mcache_index++) {
+			programming->plane_programming[main_plane_index].phantom_plane.mcache_allocation.global_mcache_ids_plane1[mcache_index] += total_main_mcaches_required;
+			programming->plane_programming[main_plane_index].phantom_plane.mcache_allocation.global_mcache_ids_mall_plane1[mcache_index] =
+					programming->plane_programming[main_plane_index].phantom_plane.mcache_allocation.global_mcache_ids_plane1[mcache_index];
+		}
+
 		for (pipe_offset = 0; pipe_offset < programming->plane_programming[main_plane_index].num_dpps_required; pipe_offset++) {
 			// Assign storage for this pipe's register values
 			programming->plane_programming[main_plane_index].phantom_plane.pipe_regs[pipe_offset] = &programming->pipe_regs[total_pipe_regs_copied];
@@ -571,6 +595,10 @@ bool core_dcn4_mode_programming(struct dml2_core_mode_programming_in_out *in_out
 
 				dml2_core_calcs_get_mall_allocation(&core->clean_me_up.mode_lib, &in_out->programming->plane_programming[plane_index].surface_size_mall_bytes, dml_internal_pipe_index);
 
+				memcpy(&in_out->programming->plane_programming[plane_index].mcache_allocation,
+						&in_out->display_cfg->stage2.mcache_allocations[plane_index],
+						sizeof(struct dml2_mcache_surface_allocation));
+
 				for (pipe_offset = 0; pipe_offset < in_out->programming->plane_programming[plane_index].num_dpps_required; pipe_offset++) {
 					in_out->programming->plane_programming[plane_index].plane_descriptor = &in_out->programming->display_config.plane_descriptors[plane_index];
 
diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_dcn4_calcs.c b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_dcn4_calcs.c
index c1ff869512f27..8ad7704b76691 100644
--- a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_dcn4_calcs.c
+++ b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_dcn4_calcs.c
@@ -2638,6 +2638,9 @@ static void calculate_mcache_setting(
 	// Luma/Chroma combine in the last mcache
 	// In the case of Luma/Chroma combine-mCache (with lc_comb_mcache==1), all mCaches except the last segment are filled as much as possible, when stay aligned to mvmpg boundary
 	if (*p->lc_comb_mcache && l->is_dual_plane) {
+		/* if luma and chroma planes share an mcache, increase total chroma mcache count */
+		*p->num_mcaches_c = *p->num_mcaches_c + 1;
+
 		for (n = 0; n < *p->num_mcaches_l - 1; n++)
 			p->mcache_offsets_l[n] = (n + 1) * l->mvmpg_per_mcache_lb_l * l->mvmpg_access_width_l;
 		p->mcache_offsets_l[*p->num_mcaches_l - 1] = l->full_vp_access_width_l;
diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_pmo/dml2_pmo_dcn4_fams2.c b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_pmo/dml2_pmo_dcn4_fams2.c
index a3324f7b9ba68..15c906c42ec45 100644
--- a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_pmo/dml2_pmo_dcn4_fams2.c
+++ b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_pmo/dml2_pmo_dcn4_fams2.c
@@ -1082,12 +1082,21 @@ static bool all_timings_support_svp(const struct dml2_pmo_instance *pmo,
 	const struct dml2_fams2_meta *stream_fams2_meta;
 	unsigned int microschedule_vlines;
 	unsigned int i;
+	unsigned int mcaches_per_plane;
+	unsigned int total_mcaches_required = 0;
 
 	unsigned int num_planes_per_stream[DML2_MAX_PLANES] = { 0 };
 
 	/* confirm timing it is not a centered timing */
 	for (i = 0; i < display_config->display_config.num_planes; i++) {
 		plane_descriptor = &display_config->display_config.plane_descriptors[i];
+		mcaches_per_plane = 0;
+
+		if (plane_descriptor->surface.dcc.enable) {
+			mcaches_per_plane += display_config->stage2.mcache_allocations[i].num_mcaches_plane0 +
+					display_config->stage2.mcache_allocations[i].num_mcaches_plane1 -
+					(display_config->stage2.mcache_allocations[i].last_slice_sharing.plane0_plane1 ? 1 : 0);
+		}
 
 		if (is_bit_set_in_bitfield(mask, (unsigned char)plane_descriptor->stream_index)) {
 			num_planes_per_stream[plane_descriptor->stream_index]++;
@@ -1098,7 +1107,19 @@ static bool all_timings_support_svp(const struct dml2_pmo_instance *pmo,
 					plane_descriptor->composition.rotation_angle != dml2_rotation_0) {
 				return false;
 			}
+
+			/* phantom requires same number of mcaches as main */
+			if (plane_descriptor->surface.dcc.enable) {
+				mcaches_per_plane *= 2;
+			}
 		}
+
+		total_mcaches_required += mcaches_per_plane;
+	}
+
+	if (total_mcaches_required > pmo->soc_bb->num_dcc_mcaches) {
+		/* too many mcaches required */
+		return false;
 	}
 
 	for (i = 0; i < DML2_MAX_PLANES; i++) {
diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_top/dml2_top_soc15.c b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_top/dml2_top_soc15.c
index a8f58f8448e42..dc2ce5e77f579 100644
--- a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_top/dml2_top_soc15.c
+++ b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_top/dml2_top_soc15.c
@@ -831,7 +831,6 @@ static bool dml2_top_soc15_build_mode_programming(struct dml2_build_mode_program
 	bool uclk_pstate_success = false;
 	bool vmin_success = false;
 	bool stutter_success = false;
-	unsigned int i;
 
 	memset(l, 0, sizeof(struct dml2_build_mode_programming_locals));
 	memset(in_out->programming, 0, sizeof(struct dml2_display_cfg_programming));
@@ -976,13 +975,6 @@ static bool dml2_top_soc15_build_mode_programming(struct dml2_build_mode_program
 		l->base_display_config_with_meta.stage5.success = true;
 	}
 
-	/*
-	* Populate mcache programming
-	*/
-	for (i = 0; i < in_out->display_config->num_planes; i++) {
-		in_out->programming->plane_programming[i].mcache_allocation = l->base_display_config_with_meta.stage2.mcache_allocations[i];
-	}
-
 	/*
 	* Call DPMM to map all requirements to minimum clock state
 	*/
-- 
2.39.5

[PATCH AUTOSEL 6.14 469/642] drm/amd/display: Ammend DCPG IP control sequences to align with HW guidance

[Sasha Levin]

From: Dillon Varone <dillon.varone@amd.com>

[ Upstream commit cbd97d621ece1d92c3542e52f8af7c04cd2c6afb ]

[WHY&HOW]
IP_REQUEST_CNTL should only be toggled off when it was originally, never
unconditionally.

Reviewed-by: Alvin Lee <alvin.lee2@amd.com>
Signed-off-by: Dillon Varone <dillon.varone@amd.com>
Signed-off-by: Zaeem Mohamed <zaeem.mohamed@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 .../amd/display/dc/hwss/dcn20/dcn20_hwseq.c   | 14 +++++---
 .../amd/display/dc/hwss/dcn401/dcn401_hwseq.c | 34 +++++++++++++++++++
 .../amd/display/dc/hwss/dcn401/dcn401_hwseq.h |  3 ++
 .../amd/display/dc/hwss/dcn401/dcn401_init.c  |  2 +-
 4 files changed, 48 insertions(+), 5 deletions(-)

diff --git a/drivers/gpu/drm/amd/display/dc/hwss/dcn20/dcn20_hwseq.c b/drivers/gpu/drm/amd/display/dc/hwss/dcn20/dcn20_hwseq.c
index f7e31aec32058..1a07973ead4f5 100644
--- a/drivers/gpu/drm/amd/display/dc/hwss/dcn20/dcn20_hwseq.c
+++ b/drivers/gpu/drm/amd/display/dc/hwss/dcn20/dcn20_hwseq.c
@@ -1264,14 +1264,18 @@ static void dcn20_power_on_plane_resources(
 	struct dce_hwseq *hws,
 	struct pipe_ctx *pipe_ctx)
 {
+	uint32_t org_ip_request_cntl = 0;
+
 	DC_LOGGER_INIT(hws->ctx->logger);
 
 	if (hws->funcs.dpp_root_clock_control)
 		hws->funcs.dpp_root_clock_control(hws, pipe_ctx->plane_res.dpp->inst, true);
 
 	if (REG(DC_IP_REQUEST_CNTL)) {
-		REG_SET(DC_IP_REQUEST_CNTL, 0,
-				IP_REQUEST_EN, 1);
+		REG_GET(DC_IP_REQUEST_CNTL, IP_REQUEST_EN, &org_ip_request_cntl);
+		if (org_ip_request_cntl == 0)
+			REG_SET(DC_IP_REQUEST_CNTL, 0,
+					IP_REQUEST_EN, 1);
 
 		if (hws->funcs.dpp_pg_control)
 			hws->funcs.dpp_pg_control(hws, pipe_ctx->plane_res.dpp->inst, true);
@@ -1279,8 +1283,10 @@ static void dcn20_power_on_plane_resources(
 		if (hws->funcs.hubp_pg_control)
 			hws->funcs.hubp_pg_control(hws, pipe_ctx->plane_res.hubp->inst, true);
 
-		REG_SET(DC_IP_REQUEST_CNTL, 0,
-				IP_REQUEST_EN, 0);
+		if (org_ip_request_cntl == 0)
+			REG_SET(DC_IP_REQUEST_CNTL, 0,
+					IP_REQUEST_EN, 0);
+
 		DC_LOG_DEBUG(
 				"Un-gated front end for pipe %d\n", pipe_ctx->plane_res.hubp->inst);
 	}
diff --git a/drivers/gpu/drm/amd/display/dc/hwss/dcn401/dcn401_hwseq.c b/drivers/gpu/drm/amd/display/dc/hwss/dcn401/dcn401_hwseq.c
index f21e2f7449517..50e710b4ce651 100644
--- a/drivers/gpu/drm/amd/display/dc/hwss/dcn401/dcn401_hwseq.c
+++ b/drivers/gpu/drm/amd/display/dc/hwss/dcn401/dcn401_hwseq.c
@@ -2655,3 +2655,37 @@ void dcn401_detect_pipe_changes(struct dc_state *old_state,
 		new_pipe->update_flags.bits.test_pattern_changed = 1;
 	}
 }
+
+void dcn401_plane_atomic_power_down(struct dc *dc,
+		struct dpp *dpp,
+		struct hubp *hubp)
+{
+	struct dce_hwseq *hws = dc->hwseq;
+	uint32_t org_ip_request_cntl = 0;
+
+	DC_LOGGER_INIT(dc->ctx->logger);
+
+	REG_GET(DC_IP_REQUEST_CNTL, IP_REQUEST_EN, &org_ip_request_cntl);
+	if (org_ip_request_cntl == 0)
+		REG_SET(DC_IP_REQUEST_CNTL, 0,
+			IP_REQUEST_EN, 1);
+
+	if (hws->funcs.dpp_pg_control)
+		hws->funcs.dpp_pg_control(hws, dpp->inst, false);
+
+	if (hws->funcs.hubp_pg_control)
+		hws->funcs.hubp_pg_control(hws, hubp->inst, false);
+
+	hubp->funcs->hubp_reset(hubp);
+	dpp->funcs->dpp_reset(dpp);
+
+	if (org_ip_request_cntl == 0)
+		REG_SET(DC_IP_REQUEST_CNTL, 0,
+			IP_REQUEST_EN, 0);
+
+	DC_LOG_DEBUG(
+			"Power gated front end %d\n", hubp->inst);
+
+	if (hws->funcs.dpp_root_clock_control)
+		hws->funcs.dpp_root_clock_control(hws, dpp->inst, false);
+}
diff --git a/drivers/gpu/drm/amd/display/dc/hwss/dcn401/dcn401_hwseq.h b/drivers/gpu/drm/amd/display/dc/hwss/dcn401/dcn401_hwseq.h
index 17cea748789e1..dbd69d215b8bc 100644
--- a/drivers/gpu/drm/amd/display/dc/hwss/dcn401/dcn401_hwseq.h
+++ b/drivers/gpu/drm/amd/display/dc/hwss/dcn401/dcn401_hwseq.h
@@ -102,4 +102,7 @@ void dcn401_detect_pipe_changes(
 	struct dc_state *new_state,
 	struct pipe_ctx *old_pipe,
 	struct pipe_ctx *new_pipe);
+void dcn401_plane_atomic_power_down(struct dc *dc,
+		struct dpp *dpp,
+		struct hubp *hubp);
 #endif /* __DC_HWSS_DCN401_H__ */
diff --git a/drivers/gpu/drm/amd/display/dc/hwss/dcn401/dcn401_init.c b/drivers/gpu/drm/amd/display/dc/hwss/dcn401/dcn401_init.c
index 44cb376f97c17..a4e3501fadbbe 100644
--- a/drivers/gpu/drm/amd/display/dc/hwss/dcn401/dcn401_init.c
+++ b/drivers/gpu/drm/amd/display/dc/hwss/dcn401/dcn401_init.c
@@ -123,7 +123,7 @@ static const struct hwseq_private_funcs dcn401_private_funcs = {
 	.disable_vga = dcn20_disable_vga,
 	.bios_golden_init = dcn10_bios_golden_init,
 	.plane_atomic_disable = dcn20_plane_atomic_disable,
-	.plane_atomic_power_down = dcn10_plane_atomic_power_down,
+	.plane_atomic_power_down = dcn401_plane_atomic_power_down,
 	.enable_power_gating_plane = dcn32_enable_power_gating_plane,
 	.hubp_pg_control = dcn32_hubp_pg_control,
 	.program_all_writeback_pipes_in_tree = dcn30_program_all_writeback_pipes_in_tree,
-- 
2.39.5

[PATCH AUTOSEL 6.14 470/642] drm/amd/display: Account For OTO Prefetch Bandwidth When Calculating Urgent Bandwidth

[Sasha Levin]

From: Austin Zheng <Austin.Zheng@amd.com>

[ Upstream commit 36681f15bb12b5c01df924379cdab9234259825c ]

[Why]
1) The current calculations for OTO prefetch bandwidth do not consider the number of DPP pipes in use.
As a result, OTO prefetch bandwidth may be larger than the vactive bandwidth if multiple DPP pipes are used.
OTO prefetch bandwidth should never exceed the vactive bandwidth.

2) Mode programming may be mismatched with mode support
In cases where mode support has chosen to use the equalized (equ) prefetch schedule,
mode programming may end up using oto prefetch schedule instead.
The bandwidth required to do the oto schedule may end up being higher than the equ schedule.
This can cause the required urgent bandwidth to exceed the available urgent bandwidth.

[How]
Output the oto prefetch bandwidth and incorperate it into the urgent bandwidth calculations
even if the prefetch schedule being used is not the oto schedule.

Reviewed-by: Dillon Varone <dillon.varone@amd.com>
Signed-off-by: Austin Zheng <Austin.Zheng@amd.com>
Signed-off-by: Zaeem Mohamed <zaeem.mohamed@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 .../src/dml2_core/dml2_core_dcn4_calcs.c      | 25 ++++++++++++++++++-
 .../src/dml2_core/dml2_core_shared_types.h    |  5 ++++
 2 files changed, 29 insertions(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_dcn4_calcs.c b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_dcn4_calcs.c
index 8ad7704b76691..913f33c31cf38 100644
--- a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_dcn4_calcs.c
+++ b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_dcn4_calcs.c
@@ -4915,6 +4915,7 @@ static double get_urgent_bandwidth_required(
 	double ReadBandwidthChroma[],
 	double PrefetchBandwidthLuma[],
 	double PrefetchBandwidthChroma[],
+	double PrefetchBandwidthOto[],
 	double excess_vactive_fill_bw_l[],
 	double excess_vactive_fill_bw_c[],
 	double cursor_bw[],
@@ -4978,8 +4979,9 @@ static double get_urgent_bandwidth_required(
 			l->vm_row_bw = NumberOfDPP[k] * prefetch_vmrow_bw[k];
 			l->flip_and_active_bw = l->per_plane_flip_bw[k] + ReadBandwidthLuma[k] * l->adj_factor_p0 + ReadBandwidthChroma[k] * l->adj_factor_p1 + cursor_bw[k] * l->adj_factor_cur;
 			l->flip_and_prefetch_bw = l->per_plane_flip_bw[k] + NumberOfDPP[k] * (PrefetchBandwidthLuma[k] * l->adj_factor_p0_pre + PrefetchBandwidthChroma[k] * l->adj_factor_p1_pre) + prefetch_cursor_bw[k] * l->adj_factor_cur_pre;
+			l->flip_and_prefetch_bw_oto = l->per_plane_flip_bw[k] + NumberOfDPP[k] * (PrefetchBandwidthOto[k] * l->adj_factor_p0_pre + PrefetchBandwidthChroma[k] * l->adj_factor_p1_pre) + prefetch_cursor_bw[k] * l->adj_factor_cur_pre;
 			l->active_and_excess_bw = (ReadBandwidthLuma[k] + excess_vactive_fill_bw_l[k]) * l->tmp_nom_adj_factor_p0 + (ReadBandwidthChroma[k] + excess_vactive_fill_bw_c[k]) * l->tmp_nom_adj_factor_p1 + dpte_row_bw[k] + meta_row_bw[k];
-			surface_required_bw[k] = math_max4(l->vm_row_bw, l->flip_and_active_bw, l->flip_and_prefetch_bw, l->active_and_excess_bw);
+			surface_required_bw[k] = math_max5(l->vm_row_bw, l->flip_and_active_bw, l->flip_and_prefetch_bw, l->active_and_excess_bw, l->flip_and_prefetch_bw_oto);
 
 			/* export peak required bandwidth for the surface */
 			surface_peak_required_bw[k] = math_max2(surface_required_bw[k], surface_peak_required_bw[k]);
@@ -5177,6 +5179,7 @@ static bool CalculatePrefetchSchedule(struct dml2_core_internal_scratch *scratch
 	s->Tsw_est3 = 0.0;
 	s->cursor_prefetch_bytes = 0;
 	*p->prefetch_cursor_bw = 0;
+	*p->RequiredPrefetchBWOTO = 0.0;
 
 	dcc_mrq_enable = (p->dcc_enable && p->mrq_present);
 
@@ -5390,6 +5393,9 @@ static bool CalculatePrefetchSchedule(struct dml2_core_internal_scratch *scratch
 		s->prefetch_bw_oto += (p->swath_width_chroma_ub * p->myPipe->BytePerPixelC) / s->LineTime;
 	}
 
+	/* oto prefetch bw should be always be less than total vactive bw */
+	DML2_ASSERT(s->prefetch_bw_oto < s->per_pipe_vactive_sw_bw * p->myPipe->DPPPerSurface);
+
 	s->prefetch_bw_oto = math_max2(s->per_pipe_vactive_sw_bw, s->prefetch_bw_oto) * p->mall_prefetch_sdp_overhead_factor;
 
 	s->prefetch_bw_oto = math_min2(s->prefetch_bw_oto, *p->prefetch_sw_bytes/(s->min_Lsw_oto*s->LineTime));
@@ -5400,6 +5406,12 @@ static bool CalculatePrefetchSchedule(struct dml2_core_internal_scratch *scratch
 					p->vm_bytes * p->HostVMInefficiencyFactor / (31 * s->LineTime) - *p->Tno_bw,
 					(p->PixelPTEBytesPerRow * p->HostVMInefficiencyFactor + p->meta_row_bytes + tdlut_row_bytes) / (15 * s->LineTime));
 
+	/* oto bw needs to be outputted even if the oto schedule isn't being used to avoid ms/mp mismatch.
+	 * mp will fail if ms decides to use equ schedule and mp decides to use oto schedule
+	 * and the required bandwidth increases when going from ms to mp
+	 */
+	*p->RequiredPrefetchBWOTO = s->prefetch_bw_oto;
+
 #ifdef __DML_VBA_DEBUG__
 	dml2_printf("DML::%s: vactive_sw_bw_l = %f\n", __func__, p->vactive_sw_bw_l);
 	dml2_printf("DML::%s: vactive_sw_bw_c = %f\n", __func__, p->vactive_sw_bw_c);
@@ -6160,6 +6172,7 @@ static void calculate_peak_bandwidth_required(
 				p->surface_read_bandwidth_c,
 				l->zero_array, //PrefetchBandwidthLuma,
 				l->zero_array, //PrefetchBandwidthChroma,
+				l->zero_array, //PrefetchBWOTO
 				l->zero_array,
 				l->zero_array,
 				l->zero_array,
@@ -6196,6 +6209,7 @@ static void calculate_peak_bandwidth_required(
 				p->surface_read_bandwidth_c,
 				l->zero_array, //PrefetchBandwidthLuma,
 				l->zero_array, //PrefetchBandwidthChroma,
+				l->zero_array, //PrefetchBWOTO
 				p->excess_vactive_fill_bw_l,
 				p->excess_vactive_fill_bw_c,
 				p->cursor_bw,
@@ -6232,6 +6246,7 @@ static void calculate_peak_bandwidth_required(
 				p->surface_read_bandwidth_c,
 				p->prefetch_bandwidth_l,
 				p->prefetch_bandwidth_c,
+				p->prefetch_bandwidth_oto, // to prevent ms/mp mismatch when oto bw > total vactive bw
 				p->excess_vactive_fill_bw_l,
 				p->excess_vactive_fill_bw_c,
 				p->cursor_bw,
@@ -6268,6 +6283,7 @@ static void calculate_peak_bandwidth_required(
 				p->surface_read_bandwidth_c,
 				p->prefetch_bandwidth_l,
 				p->prefetch_bandwidth_c,
+				p->prefetch_bandwidth_oto, // to prevent ms/mp mismatch when oto bw > total vactive bw
 				p->excess_vactive_fill_bw_l,
 				p->excess_vactive_fill_bw_c,
 				p->cursor_bw,
@@ -6304,6 +6320,7 @@ static void calculate_peak_bandwidth_required(
 				p->surface_read_bandwidth_c,
 				p->prefetch_bandwidth_l,
 				p->prefetch_bandwidth_c,
+				p->prefetch_bandwidth_oto, // to prevent ms/mp mismatch when oto bw > total vactive bw
 				p->excess_vactive_fill_bw_l,
 				p->excess_vactive_fill_bw_c,
 				p->cursor_bw,
@@ -9066,6 +9083,7 @@ static bool dml_core_mode_support(struct dml2_core_calcs_mode_support_ex *in_out
 				CalculatePrefetchSchedule_params->VRatioPrefetchC = &mode_lib->ms.VRatioPreC[k];
 				CalculatePrefetchSchedule_params->RequiredPrefetchPixelDataBWLuma = &mode_lib->ms.RequiredPrefetchPixelDataBWLuma[k]; // prefetch_sw_bw_l
 				CalculatePrefetchSchedule_params->RequiredPrefetchPixelDataBWChroma = &mode_lib->ms.RequiredPrefetchPixelDataBWChroma[k]; // prefetch_sw_bw_c
+				CalculatePrefetchSchedule_params->RequiredPrefetchBWOTO = &mode_lib->ms.RequiredPrefetchBWOTO[k];
 				CalculatePrefetchSchedule_params->NotEnoughTimeForDynamicMetadata = &mode_lib->ms.NoTimeForDynamicMetadata[k];
 				CalculatePrefetchSchedule_params->Tno_bw = &mode_lib->ms.Tno_bw[k];
 				CalculatePrefetchSchedule_params->Tno_bw_flip = &mode_lib->ms.Tno_bw_flip[k];
@@ -9210,6 +9228,7 @@ static bool dml_core_mode_support(struct dml2_core_calcs_mode_support_ex *in_out
 				calculate_peak_bandwidth_params->surface_read_bandwidth_c = mode_lib->ms.vactive_sw_bw_c;
 				calculate_peak_bandwidth_params->prefetch_bandwidth_l = mode_lib->ms.RequiredPrefetchPixelDataBWLuma;
 				calculate_peak_bandwidth_params->prefetch_bandwidth_c = mode_lib->ms.RequiredPrefetchPixelDataBWChroma;
+				calculate_peak_bandwidth_params->prefetch_bandwidth_oto = mode_lib->ms.RequiredPrefetchBWOTO;
 				calculate_peak_bandwidth_params->excess_vactive_fill_bw_l = mode_lib->ms.excess_vactive_fill_bw_l;
 				calculate_peak_bandwidth_params->excess_vactive_fill_bw_c = mode_lib->ms.excess_vactive_fill_bw_c;
 				calculate_peak_bandwidth_params->cursor_bw = mode_lib->ms.cursor_bw;
@@ -9376,6 +9395,7 @@ static bool dml_core_mode_support(struct dml2_core_calcs_mode_support_ex *in_out
 			calculate_peak_bandwidth_params->surface_read_bandwidth_c = mode_lib->ms.vactive_sw_bw_c;
 			calculate_peak_bandwidth_params->prefetch_bandwidth_l = mode_lib->ms.RequiredPrefetchPixelDataBWLuma;
 			calculate_peak_bandwidth_params->prefetch_bandwidth_c = mode_lib->ms.RequiredPrefetchPixelDataBWChroma;
+			calculate_peak_bandwidth_params->prefetch_bandwidth_oto = mode_lib->ms.RequiredPrefetchBWOTO;
 			calculate_peak_bandwidth_params->excess_vactive_fill_bw_l = mode_lib->ms.excess_vactive_fill_bw_l;
 			calculate_peak_bandwidth_params->excess_vactive_fill_bw_c = mode_lib->ms.excess_vactive_fill_bw_c;
 			calculate_peak_bandwidth_params->cursor_bw = mode_lib->ms.cursor_bw;
@@ -11292,6 +11312,7 @@ static bool dml_core_mode_programming(struct dml2_core_calcs_mode_programming_ex
 			CalculatePrefetchSchedule_params->VRatioPrefetchC = &mode_lib->mp.VRatioPrefetchC[k];
 			CalculatePrefetchSchedule_params->RequiredPrefetchPixelDataBWLuma = &mode_lib->mp.RequiredPrefetchPixelDataBWLuma[k];
 			CalculatePrefetchSchedule_params->RequiredPrefetchPixelDataBWChroma = &mode_lib->mp.RequiredPrefetchPixelDataBWChroma[k];
+			CalculatePrefetchSchedule_params->RequiredPrefetchBWOTO = &s->dummy_single_array[0][k];
 			CalculatePrefetchSchedule_params->NotEnoughTimeForDynamicMetadata = &mode_lib->mp.NotEnoughTimeForDynamicMetadata[k];
 			CalculatePrefetchSchedule_params->Tno_bw = &mode_lib->mp.Tno_bw[k];
 			CalculatePrefetchSchedule_params->Tno_bw_flip = &mode_lib->mp.Tno_bw_flip[k];
@@ -11434,6 +11455,7 @@ static bool dml_core_mode_programming(struct dml2_core_calcs_mode_programming_ex
 			calculate_peak_bandwidth_params->surface_read_bandwidth_c = mode_lib->mp.vactive_sw_bw_c;
 			calculate_peak_bandwidth_params->prefetch_bandwidth_l = mode_lib->mp.RequiredPrefetchPixelDataBWLuma;
 			calculate_peak_bandwidth_params->prefetch_bandwidth_c = mode_lib->mp.RequiredPrefetchPixelDataBWChroma;
+			calculate_peak_bandwidth_params->prefetch_bandwidth_oto = s->dummy_single_array[0];
 			calculate_peak_bandwidth_params->excess_vactive_fill_bw_l = mode_lib->mp.excess_vactive_fill_bw_l;
 			calculate_peak_bandwidth_params->excess_vactive_fill_bw_c = mode_lib->mp.excess_vactive_fill_bw_c;
 			calculate_peak_bandwidth_params->cursor_bw = mode_lib->mp.cursor_bw;
@@ -11566,6 +11588,7 @@ static bool dml_core_mode_programming(struct dml2_core_calcs_mode_programming_ex
 			calculate_peak_bandwidth_params->surface_read_bandwidth_c = mode_lib->mp.vactive_sw_bw_c;
 			calculate_peak_bandwidth_params->prefetch_bandwidth_l = mode_lib->mp.RequiredPrefetchPixelDataBWLuma;
 			calculate_peak_bandwidth_params->prefetch_bandwidth_c = mode_lib->mp.RequiredPrefetchPixelDataBWChroma;
+			calculate_peak_bandwidth_params->prefetch_bandwidth_oto = s->dummy_single_array[k];
 			calculate_peak_bandwidth_params->excess_vactive_fill_bw_l = mode_lib->mp.excess_vactive_fill_bw_l;
 			calculate_peak_bandwidth_params->excess_vactive_fill_bw_c = mode_lib->mp.excess_vactive_fill_bw_c;
 			calculate_peak_bandwidth_params->cursor_bw = mode_lib->mp.cursor_bw;
diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_shared_types.h b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_shared_types.h
index 23c0fca5515fe..b7cb017b59baa 100644
--- a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_shared_types.h
+++ b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_shared_types.h
@@ -484,6 +484,8 @@ struct dml2_core_internal_mode_support {
 	double WriteBandwidth[DML2_MAX_PLANES][DML2_MAX_WRITEBACK];
 	double RequiredPrefetchPixelDataBWLuma[DML2_MAX_PLANES];
 	double RequiredPrefetchPixelDataBWChroma[DML2_MAX_PLANES];
+	/* oto bw should also be considered when calculating urgent bw to avoid situations oto/equ mismatches between ms and mp */
+	double RequiredPrefetchBWOTO[DML2_MAX_PLANES];
 	double cursor_bw[DML2_MAX_PLANES];
 	double prefetch_cursor_bw[DML2_MAX_PLANES];
 	double prefetch_vmrow_bw[DML2_MAX_PLANES];
@@ -1381,6 +1383,7 @@ struct dml2_core_shared_get_urgent_bandwidth_required_locals {
 	double vm_row_bw;
 	double flip_and_active_bw;
 	double flip_and_prefetch_bw;
+	double flip_and_prefetch_bw_oto;
 	double active_and_excess_bw;
 };
 
@@ -1792,6 +1795,7 @@ struct dml2_core_calcs_CalculatePrefetchSchedule_params {
 	double *VRatioPrefetchC;
 	double *RequiredPrefetchPixelDataBWLuma;
 	double *RequiredPrefetchPixelDataBWChroma;
+	double *RequiredPrefetchBWOTO;
 	bool *NotEnoughTimeForDynamicMetadata;
 	double *Tno_bw;
 	double *Tno_bw_flip;
@@ -2025,6 +2029,7 @@ struct dml2_core_calcs_calculate_peak_bandwidth_required_params {
 	double *surface_read_bandwidth_c;
 	double *prefetch_bandwidth_l;
 	double *prefetch_bandwidth_c;
+	double *prefetch_bandwidth_oto;
 	double *excess_vactive_fill_bw_l;
 	double *excess_vactive_fill_bw_c;
 	double *cursor_bw;
-- 
2.39.5

[PATCH AUTOSEL 6.14 471/642] drm/amd/display: Update CR AUX RD interval interpretation

[Sasha Levin]

From: George Shen <george.shen@amd.com>

[ Upstream commit 6a7fde433231c18164c117592d3e18ced648ad58 ]

[Why]
DP spec updated to have the CR AUX RD interval match the EQ AUX RD
interval interpretation of DPCD 0000Eh/0220Eh for 8b/10b non-LTTPR mode
and LTTPR transparent mode cases.

[How]
Update interpretation of DPCD 0000Eh/0220Eh for CR AUX RD interval
during 8b/10b link training.

Reviewed-by: Michael Strauss <michael.strauss@amd.com>
Reviewed-by: Wenjing Liu <wenjing.liu@amd.com>
Signed-off-by: George Shen <george.shen@amd.com>
Signed-off-by: Zaeem Mohamed <zaeem.mohamed@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 .../display/dc/link/protocols/link_dp_training_8b_10b.c    | 7 +++++--
 1 file changed, 5 insertions(+), 2 deletions(-)

diff --git a/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_training_8b_10b.c b/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_training_8b_10b.c
index 3bdce32a85e3c..ae95ec48e5721 100644
--- a/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_training_8b_10b.c
+++ b/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_training_8b_10b.c
@@ -36,7 +36,8 @@
 	link->ctx->logger
 
 static int32_t get_cr_training_aux_rd_interval(struct dc_link *link,
-		const struct dc_link_settings *link_settings)
+		const struct dc_link_settings *link_settings,
+		enum lttpr_mode lttpr_mode)
 {
 	union training_aux_rd_interval training_rd_interval;
 	uint32_t wait_in_micro_secs = 100;
@@ -49,6 +50,8 @@ static int32_t get_cr_training_aux_rd_interval(struct dc_link *link,
 				DP_TRAINING_AUX_RD_INTERVAL,
 				(uint8_t *)&training_rd_interval,
 				sizeof(training_rd_interval));
+		if (lttpr_mode != LTTPR_MODE_NON_TRANSPARENT)
+			wait_in_micro_secs = 400;
 		if (training_rd_interval.bits.TRAINIG_AUX_RD_INTERVAL)
 			wait_in_micro_secs = training_rd_interval.bits.TRAINIG_AUX_RD_INTERVAL * 4000;
 	}
@@ -110,7 +113,6 @@ void decide_8b_10b_training_settings(
 	 */
 	lt_settings->link_settings.link_spread = link->dp_ss_off ?
 			LINK_SPREAD_DISABLED : LINK_SPREAD_05_DOWNSPREAD_30KHZ;
-	lt_settings->cr_pattern_time = get_cr_training_aux_rd_interval(link, link_setting);
 	lt_settings->eq_pattern_time = get_eq_training_aux_rd_interval(link, link_setting);
 	lt_settings->pattern_for_cr = decide_cr_training_pattern(link_setting);
 	lt_settings->pattern_for_eq = decide_eq_training_pattern(link, link_setting);
@@ -119,6 +121,7 @@ void decide_8b_10b_training_settings(
 	lt_settings->disallow_per_lane_settings = true;
 	lt_settings->always_match_dpcd_with_hw_lane_settings = true;
 	lt_settings->lttpr_mode = dp_decide_8b_10b_lttpr_mode(link);
+	lt_settings->cr_pattern_time = get_cr_training_aux_rd_interval(link, link_setting, lt_settings->lttpr_mode);
 	dp_hw_to_dpcd_lane_settings(lt_settings, lt_settings->hw_lane_settings, lt_settings->dpcd_lane_settings);
 }
 
-- 
2.39.5

[PATCH AUTOSEL 6.14 472/642] drm/amd/display: Initial psr_version with correct setting

[Sasha Levin]

From: Tom Chung <chiahsuan.chung@amd.com>

[ Upstream commit d8c782cac5007e68e7484d420168f12d3490def6 ]

[Why & How]
The initial setting for psr_version is not correct while
create a virtual link.

The default psr_version should be DC_PSR_VERSION_UNSUPPORTED.

Reviewed-by: Roman Li <roman.li@amd.com>
Signed-off-by: Tom Chung <chiahsuan.chung@amd.com>
Signed-off-by: Zaeem Mohamed <zaeem.mohamed@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/gpu/drm/amd/display/dc/core/dc.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/drivers/gpu/drm/amd/display/dc/core/dc.c b/drivers/gpu/drm/amd/display/dc/core/dc.c
index 5cd13aea1f694..19fc0cd6a9d69 100644
--- a/drivers/gpu/drm/amd/display/dc/core/dc.c
+++ b/drivers/gpu/drm/amd/display/dc/core/dc.c
@@ -276,6 +276,7 @@ static bool create_links(
 		link->link_id.type = OBJECT_TYPE_CONNECTOR;
 		link->link_id.id = CONNECTOR_ID_VIRTUAL;
 		link->link_id.enum_id = ENUM_ID_1;
+		link->psr_settings.psr_version = DC_PSR_VERSION_UNSUPPORTED;
 		link->link_enc = kzalloc(sizeof(*link->link_enc), GFP_KERNEL);
 
 		if (!link->link_enc) {
-- 
2.39.5

[PATCH AUTOSEL 6.14 473/642] drm/amdgpu/gfx10: Add cleaner shader for GFX10.1.10

[Sasha Levin]

From: Srinivasan Shanmugam <srinivasan.shanmugam@amd.com>

[ Upstream commit 25961bad9212476983c570438366e1f5e9a9cf21 ]

This commit adds the cleaner shader microcode for GFX10.1.0 GPUs. The
cleaner shader is a piece of GPU code that is used to clear or
initialize certain GPU resources, such as Local Data Share (LDS), Vector
General Purpose Registers (VGPRs), and Scalar General Purpose Registers
(SGPRs).

Clearing these resources is important for ensuring data isolation
between different workloads running on the GPU. Without the cleaner
shader, residual data from a previous workload could potentially be
accessed by a subsequent workload, leading to data leaks and incorrect
computation results.

The cleaner shader microcode is represented as an array of 32-bit words
(`gfx_10_1_0_cleaner_shader_hex`). This array is the binary
representation of the cleaner shader code, which is written in a
low-level GPU instruction set.

When the cleaner shader feature is enabled, the AMDGPU driver loads this
array into a specific location in the GPU memory. The GPU then reads
this memory location to fetch and execute the cleaner shader
instructions.

The cleaner shader is executed automatically by the GPU at the end of
each workload, before the next workload starts. This ensures that all
GPU resources are in a clean state before the start of each workload.

This addition is part of the cleaner shader feature implementation. The
cleaner shader feature helps resource utilization by cleaning up GPU
resources after they are used. It also enhances security and reliability
by preventing data leaks between workloads.

Cc: Christian König <christian.koenig@amd.com>
Cc: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Srinivasan Shanmugam <srinivasan.shanmugam@amd.com>
Suggested-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c        |  14 ++
 .../drm/amd/amdgpu/gfx_v10_0_cleaner_shader.h |  35 +++++
 .../amdgpu/gfx_v10_1_10_cleaner_shader.asm    | 126 ++++++++++++++++++
 3 files changed, 175 insertions(+)
 create mode 100644 drivers/gpu/drm/amd/amdgpu/gfx_v10_1_10_cleaner_shader.asm

diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c
index 1f32c531f610e..09178d56afbf6 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c
@@ -4739,6 +4739,20 @@ static int gfx_v10_0_sw_init(struct amdgpu_ip_block *ip_block)
 		break;
 	}
 	switch (amdgpu_ip_version(adev, GC_HWIP, 0)) {
+	case IP_VERSION(10, 1, 10):
+		adev->gfx.cleaner_shader_ptr = gfx_10_1_10_cleaner_shader_hex;
+		adev->gfx.cleaner_shader_size = sizeof(gfx_10_1_10_cleaner_shader_hex);
+		if (adev->gfx.me_fw_version >= 101 &&
+		    adev->gfx.pfp_fw_version  >= 158 &&
+		    adev->gfx.mec_fw_version >= 152) {
+			adev->gfx.enable_cleaner_shader = true;
+			r = amdgpu_gfx_cleaner_shader_sw_init(adev, adev->gfx.cleaner_shader_size);
+			if (r) {
+				adev->gfx.enable_cleaner_shader = false;
+				dev_err(adev->dev, "Failed to initialize cleaner shader\n");
+			}
+		}
+		break;
 	case IP_VERSION(10, 3, 0):
 	case IP_VERSION(10, 3, 2):
 	case IP_VERSION(10, 3, 4):
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v10_0_cleaner_shader.h b/drivers/gpu/drm/amd/amdgpu/gfx_v10_0_cleaner_shader.h
index 663c2572d440a..5255378af53c0 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v10_0_cleaner_shader.h
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v10_0_cleaner_shader.h
@@ -21,6 +21,41 @@
  * OTHER DEALINGS IN THE SOFTWARE.
  */
 
+/* Define the cleaner shader gfx_10_1_10 */
+static const u32 gfx_10_1_10_cleaner_shader_hex[] = {
+	0xb0804004, 0xbf8a0000,
+	0xbf068100, 0xbf840023,
+	0xbe8203b8, 0xbefc0380,
+	0x7e008480, 0x7e028480,
+	0x7e048480, 0x7e068480,
+	0x7e088480, 0x7e0a8480,
+	0x7e0c8480, 0x7e0e8480,
+	0xbefc0302, 0x80828802,
+	0xbf84fff5, 0xbe8203ff,
+	0x80000000, 0x87020102,
+	0xbf840012, 0xbefe03c1,
+	0xbeff03c1, 0xd7650001,
+	0x0001007f, 0xd7660001,
+	0x0002027e, 0x16020288,
+	0xbe8203bf, 0xbefc03c1,
+	0xd9382000, 0x00020201,
+	0xd9386040, 0x00040401,
+	0xd70f6a01, 0x000202ff,
+	0x00000400, 0x80828102,
+	0xbf84fff7, 0xbefc03ff,
+	0x00000068, 0xbe803080,
+	0xbe813080, 0xbe823080,
+	0xbe833080, 0x80fc847c,
+	0xbf84fffa, 0xbeea0480,
+	0xbeec0480, 0xbeee0480,
+	0xbef00480, 0xbef20480,
+	0xbef40480, 0xbef60480,
+	0xbef80480, 0xbefa0480,
+	0xbf810000, 0xbf9f0000,
+	0xbf9f0000, 0xbf9f0000,
+	0xbf9f0000, 0xbf9f0000,
+};
+
 /* Define the cleaner shader gfx_10_3_0 */
 static const u32 gfx_10_3_0_cleaner_shader_hex[] = {
 	0xb0804004, 0xbf8a0000,
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v10_1_10_cleaner_shader.asm b/drivers/gpu/drm/amd/amdgpu/gfx_v10_1_10_cleaner_shader.asm
new file mode 100644
index 0000000000000..9ba3359253c95
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v10_1_10_cleaner_shader.asm
@@ -0,0 +1,126 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright 2025 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+// This shader is to clean LDS, SGPRs and VGPRs. It is  first 64 Dwords or 256 bytes of 256 Dwords cleaner shader.
+
+// GFX10.1 : Clear SGPRs, VGPRs and LDS
+//   Launch 32 waves per CU (16 per SIMD) as a workgroup (threadgroup) to fill every wave slot
+//   Waves are "wave32" and have 64 VGPRs each, which uses all 1024 VGPRs per SIMD
+//   Waves are launched in "CU" mode, and the workgroup shares 64KB of LDS (half of the WGP's LDS)
+//      It takes 2 workgroups to use all of LDS: one on each CU of the WGP
+//   Each wave clears SGPRs 0 - 107
+//   Each wave clears VGPRs 0 - 63
+//   The first wave of the workgroup clears its 64KB of LDS
+//   The shader starts with "S_BARRIER" to ensure SPI has launched all waves of the workgroup
+//       before any wave in the workgroup could end.  Without this, it is possible not all SGPRs get cleared.
+
+
+shader main
+  asic(GFX10.1)
+  type(CS)
+  wave_size(32)
+// Note: original source code from SQ team
+
+//
+// Create 32 waves in a threadgroup (CS waves)
+// Each allocates 64 VGPRs
+// The workgroup allocates all of LDS (64kbytes)
+//
+// Takes about 2500 clocks to run.
+//   (theorhetical fastest = 1024clks vgpr + 640lds = 1660 clks)
+//
+  S_BARRIER
+  s_cmp_eq_u32 s0, 1                                // Bit0 is set, sgpr0 is set then clear VGPRS and LDS as FW set COMPUTE_USER_DATA_0
+  s_cbranch_scc0  label_0023                        // Clean VGPRs and LDS if sgpr0 of wave is set, scc = (s0 == 1)
+
+  s_mov_b32     s2, 0x00000038  // Loop 64/8=8 times  (loop unrolled for performance)
+  s_mov_b32     m0, 0
+  //
+  // CLEAR VGPRs
+  //
+label_0005:
+  v_movreld_b32     v0, 0
+  v_movreld_b32     v1, 0
+  v_movreld_b32     v2, 0
+  v_movreld_b32     v3, 0
+  v_movreld_b32     v4, 0
+  v_movreld_b32     v5, 0
+  v_movreld_b32     v6, 0
+  v_movreld_b32     v7, 0
+  s_mov_b32         m0, s2
+  s_sub_u32     s2, s2, 8
+  s_cbranch_scc0  label_0005
+  //
+  s_mov_b32     s2, 0x80000000                     // Bit31 is first_wave
+  s_and_b32     s2, s2, s0                                  // sgpr0 has tg_size (first_wave) term as in ucode only COMPUTE_PGM_RSRC2.tg_size_en is set
+  s_cbranch_scc0  label_0023                         // Clean LDS if its first wave of ThreadGroup/WorkGroup
+  // CLEAR LDS
+  //
+  s_mov_b32 exec_lo, 0xffffffff
+  s_mov_b32 exec_hi, 0xffffffff
+  v_mbcnt_lo_u32_b32  v1, exec_hi, 0          // Set V1 to thread-ID (0..63)
+  v_mbcnt_hi_u32_b32  v1, exec_lo, v1        // Set V1 to thread-ID (0..63)
+  v_mul_u32_u24  v1, 0x00000008, v1          // * 8, so each thread is a double-dword address (8byte)
+  s_mov_b32     s2, 0x00000003f                    // 64 loop iterations
+  s_mov_b32     m0, 0xffffffff
+  // Clear all of LDS space
+  // Each FirstWave of WorkGroup clears 64kbyte block
+
+label_001F:
+  ds_write2_b64  v1, v[2:3], v[2:3] offset1:32
+  ds_write2_b64  v1, v[4:5], v[4:5] offset0:64 offset1:96
+  v_add_co_u32     v1, vcc, 0x00000400, v1
+  s_sub_u32     s2, s2, 1
+  s_cbranch_scc0  label_001F
+
+  //
+  // CLEAR SGPRs
+  //
+label_0023:
+  s_mov_b32     m0, 0x00000068  // Loop 108/4=27 times  (loop unrolled for performance)
+label_sgpr_loop:
+  s_movreld_b32     s0, 0
+  s_movreld_b32     s1, 0
+  s_movreld_b32     s2, 0
+  s_movreld_b32     s3, 0
+  s_sub_u32         m0, m0, 4
+  s_cbranch_scc0  label_sgpr_loop
+
+  //clear vcc
+  s_mov_b64 vcc, 0          //clear vcc
+  //s_setreg_imm32_b32 hw_reg_shader_flat_scratch_lo, 0   //clear  flat scratch lo SGPR
+  //s_setreg_imm32_b32 hw_reg_shader_flat_scratch_hi, 0    //clear  flat scratch hi SGPR
+  s_mov_b64 ttmp0, 0        //Clear ttmp0 and ttmp1
+  s_mov_b64 ttmp2, 0        //Clear ttmp2 and ttmp3
+  s_mov_b64 ttmp4, 0        //Clear ttmp4 and ttmp5
+  s_mov_b64 ttmp6, 0        //Clear ttmp6 and ttmp7
+  s_mov_b64 ttmp8, 0        //Clear ttmp8 and ttmp9
+  s_mov_b64 ttmp10, 0       //Clear ttmp10 and ttmp11
+  s_mov_b64 ttmp12, 0       //Clear ttmp12 and ttmp13
+  s_mov_b64 ttmp14, 0       //Clear ttmp14 and ttmp15
+
+ s_endpgm
+
+end
+
+
-- 
2.39.5

[PATCH AUTOSEL 6.14 474/642] drm/amdgpu: Skip err_count sysfs creation on VF unsupported RAS blocks

[Sasha Levin]

From: Victor Skvortsov <victor.skvortsov@amd.com>

[ Upstream commit 04893397766a2b2f1bc7fe5c6414e4c0846ed171 ]

VFs are not able to query error counts for all RAS blocks. Rather than
returning error for queries on these blocks, skip sysfs the creation
all together.

Signed-off-by: Victor Skvortsov <victor.skvortsov@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c  |  3 +++
 drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c | 17 ++++++++++++++++-
 drivers/gpu/drm/amd/amdgpu/amdgpu_virt.h |  2 ++
 3 files changed, 21 insertions(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c
index f0924aa3f4e48..0c338dcdde48a 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c
@@ -1864,6 +1864,9 @@ int amdgpu_ras_sysfs_create(struct amdgpu_device *adev,
 	if (!obj || obj->attr_inuse)
 		return -EINVAL;
 
+	if (amdgpu_sriov_vf(adev) && !amdgpu_virt_ras_telemetry_block_en(adev, head->block))
+		return 0;
+
 	get_obj(obj);
 
 	snprintf(obj->fs_data.sysfs_name, sizeof(obj->fs_data.sysfs_name),
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c
index 13e5709ea1caa..e6f0152e5b087 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c
@@ -1247,7 +1247,8 @@ amdgpu_ras_block_to_sriov(struct amdgpu_device *adev, enum amdgpu_ras_block bloc
 	case AMDGPU_RAS_BLOCK__MPIO:
 		return RAS_TELEMETRY_GPU_BLOCK_MPIO;
 	default:
-		dev_err(adev->dev, "Unsupported SRIOV RAS telemetry block 0x%x\n", block);
+		DRM_WARN_ONCE("Unsupported SRIOV RAS telemetry block 0x%x\n",
+			      block);
 		return RAS_TELEMETRY_GPU_BLOCK_COUNT;
 	}
 }
@@ -1332,3 +1333,17 @@ int amdgpu_virt_ras_telemetry_post_reset(struct amdgpu_device *adev)
 
 	return 0;
 }
+
+bool amdgpu_virt_ras_telemetry_block_en(struct amdgpu_device *adev,
+					enum amdgpu_ras_block block)
+{
+	enum amd_sriov_ras_telemetry_gpu_block sriov_block;
+
+	sriov_block = amdgpu_ras_block_to_sriov(adev, block);
+
+	if (sriov_block >= RAS_TELEMETRY_GPU_BLOCK_COUNT ||
+	    !amdgpu_sriov_ras_telemetry_block_en(adev, sriov_block))
+		return false;
+
+	return true;
+}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.h
index 0ca73343a7689..0f3ccae5c1ab3 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.h
@@ -407,4 +407,6 @@ bool amdgpu_virt_get_ras_capability(struct amdgpu_device *adev);
 int amdgpu_virt_req_ras_err_count(struct amdgpu_device *adev, enum amdgpu_ras_block block,
 				  struct ras_err_data *err_data);
 int amdgpu_virt_ras_telemetry_post_reset(struct amdgpu_device *adev);
+bool amdgpu_virt_ras_telemetry_block_en(struct amdgpu_device *adev,
+					enum amdgpu_ras_block block);
 #endif
-- 
2.39.5

[PATCH AUTOSEL 6.14 475/642] amdgpu/soc15: enable asic reset for dGPU in case of suspend abort

[Sasha Levin]

From: Jiang Liu <gerry@linux.alibaba.com>

[ Upstream commit 38e8ca3e4b6de1c6e49d0140264cfc8d314a5f70 ]

When GPU suspend is aborted, do the same for dGPU as APU to reset
soc15 asic. Otherwise it may cause following errors:
[  547.229463] amdgpu 0001:81:00.0: [drm:amdgpu_ring_test_helper [amdgpu]] *ERROR* ring kiq_0.2.1.0 test failed (-110)

[  555.126827] amdgpu 0000:0a:00.0: [drm:amdgpu_ring_test_helper [amdgpu]] *ERROR* ring kiq_0.2.1.0 test failed (-110)
[  555.126901] [drm:amdgpu_gfx_enable_kcq [amdgpu]] *ERROR* KCQ enable failed
[  555.126957] [drm:amdgpu_device_ip_resume_phase2 [amdgpu]] *ERROR* resume of IP block <gfx_v9_4_3> failed -110
[  555.126959] amdgpu 0000:0a:00.0: amdgpu: amdgpu_device_ip_resume failed (-110).
[  555.126965] PM: dpm_run_callback(): pci_pm_resume+0x0/0xe0 returns -110
[  555.126966] PM: Device 0000:0a:00.0 failed to resume async: error -110

This fix has been tested on Mi308X.

Signed-off-by: Jiang Liu <gerry@linux.alibaba.com>
Tested-by: Shuo Liu <shuox.liu@linux.alibaba.com>
Reviewed-by: Mario Limonciello <mario.limonciello@amd.com>
Link: https://lore.kernel.org/r/2462b4b12eb9d025e82525178d568cbaa4c223ff.1736739303.git.gerry@linux.alibaba.com
Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/gpu/drm/amd/amdgpu/soc15.c | 8 +++-----
 1 file changed, 3 insertions(+), 5 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/soc15.c b/drivers/gpu/drm/amd/amdgpu/soc15.c
index e98fb3fa36a88..6e09613de8cd2 100644
--- a/drivers/gpu/drm/amd/amdgpu/soc15.c
+++ b/drivers/gpu/drm/amd/amdgpu/soc15.c
@@ -604,12 +604,10 @@ soc15_asic_reset_method(struct amdgpu_device *adev)
 static bool soc15_need_reset_on_resume(struct amdgpu_device *adev)
 {
 	/* Will reset for the following suspend abort cases.
-	 * 1) Only reset on APU side, dGPU hasn't checked yet.
-	 * 2) S3 suspend aborted in the normal S3 suspend or
-	 *    performing pm core test.
+	 * 1) S3 suspend aborted in the normal S3 suspend
+	 * 2) S3 suspend aborted in performing pm core test.
 	 */
-	if (adev->flags & AMD_IS_APU && adev->in_s3 &&
-			!pm_resume_via_firmware())
+	if (adev->in_s3 && !pm_resume_via_firmware())
 		return true;
 	else
 		return false;
-- 
2.39.5

[PATCH AUTOSEL 6.14 476/642] drm/amd/display: Reverse the visual confirm recouts

[Sasha Levin]

From: Peterson Guo <peterson.guo@amd.com>

[ Upstream commit 3c50bf2196aaddcaffe2c7a1a7080470380cbfdd ]

[WHY]
When checking if a pipe can disable cursor to prevent duplicate cursors,
having visual confirm on will prevent disabling cursors on planes which
cover the bottom of the screen.

[HOW]
When checking if a plane can disable visual confirm, the pipe first
reverses these calculations before doing the checks.

Reviewed-by: Alvin Lee <alvin.lee2@amd.com>
Signed-off-by: Peterson Guo <peterson.guo@amd.com>
Signed-off-by: Wayne Lin <wayne.lin@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 .../gpu/drm/amd/display/dc/core/dc_resource.c | 67 +++++++++++++++++++
 drivers/gpu/drm/amd/display/dc/dc_dmub_srv.c  | 50 +-------------
 .../amd/display/dc/hwss/dcn10/dcn10_hwseq.c   | 48 +------------
 .../amd/display/dc/hwss/dcn401/dcn401_hwseq.c | 48 +------------
 drivers/gpu/drm/amd/display/dc/inc/resource.h |  2 +
 5 files changed, 73 insertions(+), 142 deletions(-)

diff --git a/drivers/gpu/drm/amd/display/dc/core/dc_resource.c b/drivers/gpu/drm/amd/display/dc/core/dc_resource.c
index 3367030da3414..375b3b1d1d182 100644
--- a/drivers/gpu/drm/amd/display/dc/core/dc_resource.c
+++ b/drivers/gpu/drm/amd/display/dc/core/dc_resource.c
@@ -941,6 +941,17 @@ static void calculate_adjust_recout_for_visual_confirm(struct pipe_ctx *pipe_ctx
 		*base_offset = VISUAL_CONFIRM_BASE_DEFAULT;
 }
 
+static void reverse_adjust_recout_for_visual_confirm(struct rect *recout,
+		struct pipe_ctx *pipe_ctx)
+{
+	int dpp_offset, base_offset;
+
+	calculate_adjust_recout_for_visual_confirm(pipe_ctx, &base_offset,
+		&dpp_offset);
+	recout->height += base_offset;
+	recout->height += dpp_offset;
+}
+
 static void adjust_recout_for_visual_confirm(struct rect *recout,
 		struct pipe_ctx *pipe_ctx)
 {
@@ -1642,6 +1653,62 @@ bool resource_build_scaling_params(struct pipe_ctx *pipe_ctx)
 	return res;
 }
 
+bool resource_can_pipe_disable_cursor(struct pipe_ctx *pipe_ctx)
+{
+	struct pipe_ctx *test_pipe, *split_pipe;
+	struct rect r1 = pipe_ctx->plane_res.scl_data.recout;
+	int r1_right, r1_bottom;
+	int cur_layer = pipe_ctx->plane_state->layer_index;
+
+	reverse_adjust_recout_for_visual_confirm(&r1, pipe_ctx);
+	r1_right = r1.x + r1.width;
+	r1_bottom = r1.y + r1.height;
+
+	/**
+	 * Disable the cursor if there's another pipe above this with a
+	 * plane that contains this pipe's viewport to prevent double cursor
+	 * and incorrect scaling artifacts.
+	 */
+	for (test_pipe = pipe_ctx->top_pipe; test_pipe;
+	     test_pipe = test_pipe->top_pipe) {
+		struct rect r2;
+		int r2_right, r2_bottom;
+		// Skip invisible layer and pipe-split plane on same layer
+		if (!test_pipe->plane_state ||
+		    !test_pipe->plane_state->visible ||
+		    test_pipe->plane_state->layer_index == cur_layer)
+			continue;
+
+		r2 = test_pipe->plane_res.scl_data.recout;
+		reverse_adjust_recout_for_visual_confirm(&r2, test_pipe);
+		r2_right = r2.x + r2.width;
+		r2_bottom = r2.y + r2.height;
+
+		/**
+		 * There is another half plane on same layer because of
+		 * pipe-split, merge together per same height.
+		 */
+		for (split_pipe = pipe_ctx->top_pipe; split_pipe;
+		     split_pipe = split_pipe->top_pipe)
+			if (split_pipe->plane_state->layer_index == test_pipe->plane_state->layer_index) {
+				struct rect r2_half;
+
+				r2_half = split_pipe->plane_res.scl_data.recout;
+				reverse_adjust_recout_for_visual_confirm(&r2_half, split_pipe);
+				r2.x = min(r2_half.x, r2.x);
+				r2.width = r2.width + r2_half.width;
+				r2_right = r2.x + r2.width;
+				r2_bottom = min(r2_bottom, r2_half.y + r2_half.height);
+				break;
+			}
+
+		if (r1.x >= r2.x && r1.y >= r2.y && r1_right <= r2_right && r1_bottom <= r2_bottom)
+			return true;
+	}
+
+	return false;
+}
+
 
 enum dc_status resource_build_scaling_params_for_context(
 	const struct dc  *dc,
diff --git a/drivers/gpu/drm/amd/display/dc/dc_dmub_srv.c b/drivers/gpu/drm/amd/display/dc/dc_dmub_srv.c
index 44ff9abe2880f..87b4c2793df3c 100644
--- a/drivers/gpu/drm/amd/display/dc/dc_dmub_srv.c
+++ b/drivers/gpu/drm/amd/display/dc/dc_dmub_srv.c
@@ -991,57 +991,11 @@ void dc_dmub_srv_log_diagnostic_data(struct dc_dmub_srv *dc_dmub_srv)
 	DC_LOG_DEBUG("    is_cw6_en          : %d", diag_data.is_cw6_enabled);
 }
 
-static bool dc_can_pipe_disable_cursor(struct pipe_ctx *pipe_ctx)
-{
-	struct pipe_ctx *test_pipe, *split_pipe;
-	const struct scaler_data *scl_data = &pipe_ctx->plane_res.scl_data;
-	struct rect r1 = scl_data->recout, r2, r2_half;
-	int r1_r = r1.x + r1.width, r1_b = r1.y + r1.height, r2_r, r2_b;
-	int cur_layer = pipe_ctx->plane_state->layer_index;
-
-	/**
-	 * Disable the cursor if there's another pipe above this with a
-	 * plane that contains this pipe's viewport to prevent double cursor
-	 * and incorrect scaling artifacts.
-	 */
-	for (test_pipe = pipe_ctx->top_pipe; test_pipe;
-	     test_pipe = test_pipe->top_pipe) {
-		// Skip invisible layer and pipe-split plane on same layer
-		if (!test_pipe->plane_state->visible || test_pipe->plane_state->layer_index == cur_layer)
-			continue;
-
-		r2 = test_pipe->plane_res.scl_data.recout;
-		r2_r = r2.x + r2.width;
-		r2_b = r2.y + r2.height;
-
-		/**
-		 * There is another half plane on same layer because of
-		 * pipe-split, merge together per same height.
-		 */
-		for (split_pipe = pipe_ctx->top_pipe; split_pipe;
-		     split_pipe = split_pipe->top_pipe)
-			if (split_pipe->plane_state->layer_index == test_pipe->plane_state->layer_index) {
-				r2_half = split_pipe->plane_res.scl_data.recout;
-				r2.x = (r2_half.x < r2.x) ? r2_half.x : r2.x;
-				r2.width = r2.width + r2_half.width;
-				r2_r = r2.x + r2.width;
-				break;
-			}
-
-		if (r1.x >= r2.x && r1.y >= r2.y && r1_r <= r2_r && r1_b <= r2_b)
-			return true;
-	}
-
-	return false;
-}
-
 static bool dc_dmub_should_update_cursor_data(struct pipe_ctx *pipe_ctx)
 {
 	if (pipe_ctx->plane_state != NULL) {
-		if (pipe_ctx->plane_state->address.type == PLN_ADDR_TYPE_VIDEO_PROGRESSIVE)
-			return false;
-
-		if (dc_can_pipe_disable_cursor(pipe_ctx))
+		if (pipe_ctx->plane_state->address.type == PLN_ADDR_TYPE_VIDEO_PROGRESSIVE ||
+			resource_can_pipe_disable_cursor(pipe_ctx))
 			return false;
 	}
 
diff --git a/drivers/gpu/drm/amd/display/dc/hwss/dcn10/dcn10_hwseq.c b/drivers/gpu/drm/amd/display/dc/hwss/dcn10/dcn10_hwseq.c
index 4c89bf6725b3b..bbeaefe1ef0db 100644
--- a/drivers/gpu/drm/amd/display/dc/hwss/dcn10/dcn10_hwseq.c
+++ b/drivers/gpu/drm/amd/display/dc/hwss/dcn10/dcn10_hwseq.c
@@ -3424,52 +3424,6 @@ void dcn10_update_dchub(struct dce_hwseq *hws, struct dchub_init_data *dh_data)
 	hubbub->funcs->update_dchub(hubbub, dh_data);
 }
 
-static bool dcn10_can_pipe_disable_cursor(struct pipe_ctx *pipe_ctx)
-{
-	struct pipe_ctx *test_pipe, *split_pipe;
-	const struct scaler_data *scl_data = &pipe_ctx->plane_res.scl_data;
-	struct rect r1 = scl_data->recout, r2, r2_half;
-	int r1_r = r1.x + r1.width, r1_b = r1.y + r1.height, r2_r, r2_b;
-	int cur_layer = pipe_ctx->plane_state->layer_index;
-
-	/**
-	 * Disable the cursor if there's another pipe above this with a
-	 * plane that contains this pipe's viewport to prevent double cursor
-	 * and incorrect scaling artifacts.
-	 */
-	for (test_pipe = pipe_ctx->top_pipe; test_pipe;
-	     test_pipe = test_pipe->top_pipe) {
-		// Skip invisible layer and pipe-split plane on same layer
-		if (!test_pipe->plane_state ||
-		    !test_pipe->plane_state->visible ||
-		    test_pipe->plane_state->layer_index == cur_layer)
-			continue;
-
-		r2 = test_pipe->plane_res.scl_data.recout;
-		r2_r = r2.x + r2.width;
-		r2_b = r2.y + r2.height;
-
-		/**
-		 * There is another half plane on same layer because of
-		 * pipe-split, merge together per same height.
-		 */
-		for (split_pipe = pipe_ctx->top_pipe; split_pipe;
-		     split_pipe = split_pipe->top_pipe)
-			if (split_pipe->plane_state->layer_index == test_pipe->plane_state->layer_index) {
-				r2_half = split_pipe->plane_res.scl_data.recout;
-				r2.x = (r2_half.x < r2.x) ? r2_half.x : r2.x;
-				r2.width = r2.width + r2_half.width;
-				r2_r = r2.x + r2.width;
-				break;
-			}
-
-		if (r1.x >= r2.x && r1.y >= r2.y && r1_r <= r2_r && r1_b <= r2_b)
-			return true;
-	}
-
-	return false;
-}
-
 void dcn10_set_cursor_position(struct pipe_ctx *pipe_ctx)
 {
 	struct dc_cursor_position pos_cpy = pipe_ctx->stream->cursor_position;
@@ -3569,7 +3523,7 @@ void dcn10_set_cursor_position(struct pipe_ctx *pipe_ctx)
 			== PLN_ADDR_TYPE_VIDEO_PROGRESSIVE)
 		pos_cpy.enable = false;
 
-	if (pos_cpy.enable && dcn10_can_pipe_disable_cursor(pipe_ctx))
+	if (pos_cpy.enable && resource_can_pipe_disable_cursor(pipe_ctx))
 		pos_cpy.enable = false;
 
 
diff --git a/drivers/gpu/drm/amd/display/dc/hwss/dcn401/dcn401_hwseq.c b/drivers/gpu/drm/amd/display/dc/hwss/dcn401/dcn401_hwseq.c
index 50e710b4ce651..47132e40aa2d2 100644
--- a/drivers/gpu/drm/amd/display/dc/hwss/dcn401/dcn401_hwseq.c
+++ b/drivers/gpu/drm/amd/display/dc/hwss/dcn401/dcn401_hwseq.c
@@ -972,52 +972,6 @@ void dcn401_setup_hpo_hw_control(const struct dce_hwseq *hws, bool enable)
 	REG_UPDATE(HPO_TOP_HW_CONTROL, HPO_IO_EN, enable);
 }
 
-static bool dcn401_can_pipe_disable_cursor(struct pipe_ctx *pipe_ctx)
-{
-	struct pipe_ctx *test_pipe, *split_pipe;
-	const struct scaler_data *scl_data = &pipe_ctx->plane_res.scl_data;
-	struct rect r1 = scl_data->recout, r2, r2_half;
-	int r1_r = r1.x + r1.width, r1_b = r1.y + r1.height, r2_r, r2_b;
-	int cur_layer = pipe_ctx->plane_state->layer_index;
-
-	/**
-	 * Disable the cursor if there's another pipe above this with a
-	 * plane that contains this pipe's viewport to prevent double cursor
-	 * and incorrect scaling artifacts.
-	 */
-	for (test_pipe = pipe_ctx->top_pipe; test_pipe;
-		test_pipe = test_pipe->top_pipe) {
-		// Skip invisible layer and pipe-split plane on same layer
-		if (!test_pipe->plane_state ||
-			!test_pipe->plane_state->visible ||
-			test_pipe->plane_state->layer_index == cur_layer)
-			continue;
-
-		r2 = test_pipe->plane_res.scl_data.recout;
-		r2_r = r2.x + r2.width;
-		r2_b = r2.y + r2.height;
-
-		/**
-		 * There is another half plane on same layer because of
-		 * pipe-split, merge together per same height.
-		 */
-		for (split_pipe = pipe_ctx->top_pipe; split_pipe;
-			split_pipe = split_pipe->top_pipe)
-			if (split_pipe->plane_state->layer_index == test_pipe->plane_state->layer_index) {
-				r2_half = split_pipe->plane_res.scl_data.recout;
-				r2.x = (r2_half.x < r2.x) ? r2_half.x : r2.x;
-				r2.width = r2.width + r2_half.width;
-				r2_r = r2.x + r2.width;
-				break;
-			}
-
-		if (r1.x >= r2.x && r1.y >= r2.y && r1_r <= r2_r && r1_b <= r2_b)
-			return true;
-	}
-
-	return false;
-}
-
 void adjust_hotspot_between_slices_for_2x_magnify(uint32_t cursor_width, struct dc_cursor_position *pos_cpy)
 {
 	if (cursor_width <= 128) {
@@ -1208,7 +1162,7 @@ void dcn401_set_cursor_position(struct pipe_ctx *pipe_ctx)
 	pos_cpy.x = (uint32_t)x_pos;
 	pos_cpy.y = (uint32_t)y_pos;
 
-	if (pos_cpy.enable && dcn401_can_pipe_disable_cursor(pipe_ctx))
+	if (pos_cpy.enable && resource_can_pipe_disable_cursor(pipe_ctx))
 		pos_cpy.enable = false;
 
 	x_pos = pos_cpy.x - param.recout.x;
diff --git a/drivers/gpu/drm/amd/display/dc/inc/resource.h b/drivers/gpu/drm/amd/display/dc/inc/resource.h
index cd1157d225abe..b32d07ce0f087 100644
--- a/drivers/gpu/drm/amd/display/dc/inc/resource.h
+++ b/drivers/gpu/drm/amd/display/dc/inc/resource.h
@@ -152,6 +152,8 @@ bool resource_attach_surfaces_to_context(
 		struct dc_state *context,
 		const struct resource_pool *pool);
 
+bool resource_can_pipe_disable_cursor(struct pipe_ctx *pipe_ctx);
+
 #define FREE_PIPE_INDEX_NOT_FOUND -1
 
 /*
-- 
2.39.5

[PATCH AUTOSEL 6.14 477/642] drm/amd/display: Increase block_sequence array size

[Sasha Levin]

From: Joshua Aberback <joshua.aberback@amd.com>

[ Upstream commit 3a7810c212bcf2f722671dadf4b23ff70a7d23ee ]

[Why]
It's possible to generate more than 50 steps in hwss_build_fast_sequence,
for example with a 6-pipe asic where all pipes are in one MPC chain. This
overflows the block_sequence buffer and corrupts block_sequence_steps,
causing a crash.

[How]
Expand block_sequence to 100 items. A naive upper bound on the possible
number of steps for a 6-pipe asic, ignoring the potential for steps to be
mutually exclusive, is 91 with current code, therefore 100 is sufficient.

Reviewed-by: Alvin Lee <alvin.lee2@amd.com>
Signed-off-by: Joshua Aberback <joshua.aberback@amd.com>
Signed-off-by: Wayne Lin <wayne.lin@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/gpu/drm/amd/display/dc/inc/core_types.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/amd/display/dc/inc/core_types.h b/drivers/gpu/drm/amd/display/dc/inc/core_types.h
index d558efc6e12f9..652d52040f4e6 100644
--- a/drivers/gpu/drm/amd/display/dc/inc/core_types.h
+++ b/drivers/gpu/drm/amd/display/dc/inc/core_types.h
@@ -627,7 +627,7 @@ struct dc_state {
 	 */
 	struct bw_context bw_ctx;
 
-	struct block_sequence block_sequence[50];
+	struct block_sequence block_sequence[100];
 	unsigned int block_sequence_steps;
 	struct dc_dmub_cmd dc_dmub_cmd[10];
 	unsigned int dmub_cmd_count;
-- 
2.39.5

[PATCH AUTOSEL 6.14 478/642] drm/amd/display: Use Nominal vBlank If Provided Instead Of Capping It

[Sasha Levin]

From: Austin Zheng <Austin.Zheng@amd.com>

[ Upstream commit 41df56b1fc24cc36fffb10e437385b3a49fbb5e2 ]

[Why/How]
vBlank used to determine the max vStartup is based on the smallest between
the vblank provided by the timing and vblank in ip_caps.
Extra vblank time is not considered if the vblank provided by the timing ends
up being higher than what's defined by the ip_caps

Use 1 less than the vblank size in case the timing is interlaced
so vstartup will always be less than vblank_nom.

Reviewed-by: Dillon Varone <dillon.varone@amd.com>
Signed-off-by: Austin Zheng <Austin.Zheng@amd.com>
Signed-off-by: Wayne Lin <wayne.lin@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 .../dc/dml2/dml21/src/dml2_core/dml2_core_dcn4_calcs.c       | 5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)

diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_dcn4_calcs.c b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_dcn4_calcs.c
index 913f33c31cf38..a72b4c05e1fbf 100644
--- a/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_dcn4_calcs.c
+++ b/drivers/gpu/drm/amd/display/dc/dml2/dml21/src/dml2_core/dml2_core_dcn4_calcs.c
@@ -3713,13 +3713,12 @@ static unsigned int CalculateMaxVStartup(
 	double line_time_us = (double)timing->h_total / ((double)timing->pixel_clock_khz / 1000);
 	unsigned int vblank_actual = timing->v_total - timing->v_active;
 	unsigned int vblank_nom_default_in_line = (unsigned int)math_floor2((double)vblank_nom_default_us / line_time_us, 1.0);
-	unsigned int vblank_nom_input = (unsigned int)math_min2(timing->vblank_nom, vblank_nom_default_in_line);
-	unsigned int vblank_avail = (vblank_nom_input == 0) ? vblank_nom_default_in_line : vblank_nom_input;
+	unsigned int vblank_avail = (timing->vblank_nom == 0) ? vblank_nom_default_in_line : (unsigned int)timing->vblank_nom;
 
 	vblank_size = (unsigned int)math_min2(vblank_actual, vblank_avail);
 
 	if (timing->interlaced && !ptoi_supported)
-		max_vstartup_lines = (unsigned int)(math_floor2(vblank_size / 2.0, 1.0));
+		max_vstartup_lines = (unsigned int)(math_floor2((vblank_size - 1) / 2.0, 1.0));
 	else
 		max_vstartup_lines = vblank_size - (unsigned int)math_max2(1.0, math_ceil2(write_back_delay_us / line_time_us, 1.0));
 #ifdef __DML_VBA_DEBUG__
-- 
2.39.5

[PATCH AUTOSEL 6.14 479/642] drm/amd/display: Populate register address for dentist for dcn401

[Sasha Levin]

From: Dillon Varone <dillon.varone@amd.com>

[ Upstream commit 5f0d1ef6f16e150ee46cc00b8d233d9d271fe39e ]

[WHY&HOW]
Address was not previously populated which can result in incorrect
clock frequencies being read on boot.

Reviewed-by: Alvin Lee <alvin.lee2@amd.com>
Signed-off-by: Dillon Varone <dillon.varone@amd.com>
Signed-off-by: Wayne Lin <wayne.lin@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/gpu/drm/amd/display/dc/clk_mgr/dcn401/dcn401_clk_mgr.c | 2 ++
 drivers/gpu/drm/amd/display/dc/inc/hw/clk_mgr_internal.h       | 1 +
 2 files changed, 3 insertions(+)

diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn401/dcn401_clk_mgr.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn401/dcn401_clk_mgr.c
index 8082bb8776114..a3b8e3d4a429e 100644
--- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn401/dcn401_clk_mgr.c
+++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn401/dcn401_clk_mgr.c
@@ -24,6 +24,8 @@
 
 #include "dml/dcn401/dcn401_fpu.h"
 
+#define DCN_BASE__INST0_SEG1                       0x000000C0
+
 #define mmCLK01_CLK0_CLK_PLL_REQ                        0x16E37
 #define mmCLK01_CLK0_CLK0_DFS_CNTL                      0x16E69
 #define mmCLK01_CLK0_CLK1_DFS_CNTL                      0x16E6C
diff --git a/drivers/gpu/drm/amd/display/dc/inc/hw/clk_mgr_internal.h b/drivers/gpu/drm/amd/display/dc/inc/hw/clk_mgr_internal.h
index 7a1ca1e98059b..221645c023b50 100644
--- a/drivers/gpu/drm/amd/display/dc/inc/hw/clk_mgr_internal.h
+++ b/drivers/gpu/drm/amd/display/dc/inc/hw/clk_mgr_internal.h
@@ -221,6 +221,7 @@ enum dentist_divider_range {
 	CLK_SF(CLK0_CLK_PLL_REQ, FbMult_frac, mask_sh)
 
 #define CLK_REG_LIST_DCN401()	  \
+	SR(DENTIST_DISPCLK_CNTL), \
 	CLK_SR_DCN401(CLK0_CLK_PLL_REQ,   CLK01, 0), \
 	CLK_SR_DCN401(CLK0_CLK0_DFS_CNTL, CLK01, 0), \
 	CLK_SR_DCN401(CLK0_CLK1_DFS_CNTL,  CLK01, 0), \
-- 
2.39.5

[PATCH AUTOSEL 6.14 480/642] drm/amdgpu: Use active umc info from discovery

[Sasha Levin]

From: Lijo Lazar <lijo.lazar@amd.com>

[ Upstream commit f7a594e40517fa2ab25d5ca10e7b6a158f529fb5 ]

There could be configs where some UMC instances are harvested. This
information is obtained through discovery data and populated in
umc.active_mask. Avoid reassigning this as AID mask, instead use the
mask directly while iterating through umc instances. This is to avoid
accesses to harvested UMC instances.

v2: fix warning (Alex)

Signed-off-by: Lijo Lazar <lijo.lazar@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_umc.c | 42 +++++++++++++++++++++++++
 drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c   |  1 -
 2 files changed, 42 insertions(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.c
index eafe20d8fe0b6..0a1ef95b28668 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.c
@@ -387,6 +387,45 @@ int amdgpu_umc_fill_error_record(struct ras_err_data *err_data,
 	return 0;
 }
 
+static int amdgpu_umc_loop_all_aid(struct amdgpu_device *adev, umc_func func,
+				   void *data)
+{
+	uint32_t umc_node_inst;
+	uint32_t node_inst;
+	uint32_t umc_inst;
+	uint32_t ch_inst;
+	int ret;
+
+	/*
+	 * This loop is done based on the following -
+	 * umc.active mask = mask of active umc instances across all nodes
+	 * umc.umc_inst_num = maximum number of umc instancess per node
+	 * umc.node_inst_num = maximum number of node instances
+	 * Channel instances are not assumed to be harvested.
+	 */
+	dev_dbg(adev->dev, "active umcs :%lx umc_inst per node: %d",
+		adev->umc.active_mask, adev->umc.umc_inst_num);
+	for_each_set_bit(umc_node_inst, &(adev->umc.active_mask),
+			 adev->umc.node_inst_num * adev->umc.umc_inst_num) {
+		node_inst = umc_node_inst / adev->umc.umc_inst_num;
+		umc_inst = umc_node_inst % adev->umc.umc_inst_num;
+		LOOP_UMC_CH_INST(ch_inst) {
+			dev_dbg(adev->dev,
+				"node_inst :%d umc_inst: %d ch_inst: %d",
+				node_inst, umc_inst, ch_inst);
+			ret = func(adev, node_inst, umc_inst, ch_inst, data);
+			if (ret) {
+				dev_err(adev->dev,
+					"Node %d umc %d ch %d func returns %d\n",
+					node_inst, umc_inst, ch_inst, ret);
+				return ret;
+			}
+		}
+	}
+
+	return 0;
+}
+
 int amdgpu_umc_loop_channels(struct amdgpu_device *adev,
 			umc_func func, void *data)
 {
@@ -395,6 +434,9 @@ int amdgpu_umc_loop_channels(struct amdgpu_device *adev,
 	uint32_t ch_inst         = 0;
 	int ret = 0;
 
+	if (adev->aid_mask)
+		return amdgpu_umc_loop_all_aid(adev, func, data);
+
 	if (adev->umc.node_inst_num) {
 		LOOP_UMC_EACH_NODE_INST_AND_CH(node_inst, umc_inst, ch_inst) {
 			ret = func(adev, node_inst, umc_inst, ch_inst, data);
diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c
index 5250b470e5ef3..f1dc9e50d67e7 100644
--- a/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c
@@ -1504,7 +1504,6 @@ static void gmc_v9_0_set_umc_funcs(struct amdgpu_device *adev)
 		adev->umc.umc_inst_num = UMC_V12_0_UMC_INSTANCE_NUM;
 		adev->umc.node_inst_num /= UMC_V12_0_UMC_INSTANCE_NUM;
 		adev->umc.channel_offs = UMC_V12_0_PER_CHANNEL_OFFSET;
-		adev->umc.active_mask = adev->aid_mask;
 		adev->umc.retire_unit = UMC_V12_0_BAD_PAGE_NUM_PER_CHANNEL;
 		if (!adev->gmc.xgmi.connected_to_cpu && !adev->gmc.is_app_apu)
 			adev->umc.ras = &umc_v12_0_ras;
-- 
2.39.5

[PATCH AUTOSEL 6.14 481/642] drm/amdgpu: enlarge the VBIOS binary size limit

[Sasha Levin]

From: Shiwu Zhang <shiwu.zhang@amd.com>

[ Upstream commit 667b96134c9e206aebe40985650bf478935cbe04 ]

Some chips have a larger VBIOS file so raise the size limit to support
the flashing tool.

Signed-off-by: Shiwu Zhang <shiwu.zhang@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c
index 665cc277cdc05..6dded11a23acf 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c
@@ -44,7 +44,7 @@
 #include "amdgpu_securedisplay.h"
 #include "amdgpu_atomfirmware.h"
 
-#define AMD_VBIOS_FILE_MAX_SIZE_B      (1024*1024*3)
+#define AMD_VBIOS_FILE_MAX_SIZE_B      (1024*1024*16)
 
 static int psp_load_smu_fw(struct psp_context *psp);
 static int psp_rap_terminate(struct psp_context *psp);
-- 
2.39.5

[PATCH AUTOSEL 6.14 482/642] drm/amdkfd: Have kfd driver use same PASID values from graphic driver

[Sasha Levin]

From: Xiaogang Chen <xiaogang.chen@amd.com>

[ Upstream commit 8544374c0f82edb285779f21b149826fe2c2977c ]

Current kfd driver has its own PASID value for a kfd process and uses it to
locate vm at interrupt handler or mapping between kfd process and vm. That
design is not working when a physical gpu device has multiple spatial
partitions, ex: adev in CPX mode. This patch has kfd driver use same pasid
values that graphic driver generated which is per vm per pasid.

These pasid values are passed to fw/hardware. We do not need change interrupt
handler though more pasid values are used. Also, pasid values at log are
replaced by user process pid; pasid values are not exposed to user. Users see
their process pids that have meaning in user space.

Signed-off-by: Xiaogang Chen <xiaogang.chen@amd.com>
Reviewed-by: Felix Kuehling <felix.kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h    |   3 +-
 .../gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c  |  21 ----
 .../gpu/drm/amd/amdkfd/cik_event_interrupt.c  |  18 ++-
 drivers/gpu/drm/amd/amdkfd/kfd_chardev.c      |  25 ++--
 drivers/gpu/drm/amd/amdkfd/kfd_debug.c        |  14 +--
 drivers/gpu/drm/amd/amdkfd/kfd_device.c       |   2 +-
 .../drm/amd/amdkfd/kfd_device_queue_manager.c |  85 +++++++------
 drivers/gpu/drm/amd/amdkfd/kfd_events.c       |  43 ++++---
 .../gpu/drm/amd/amdkfd/kfd_int_process_v11.c  |   2 +-
 .../gpu/drm/amd/amdkfd/kfd_int_process_v9.c   |   2 +-
 .../drm/amd/amdkfd/kfd_packet_manager_v9.c    |   4 +-
 .../drm/amd/amdkfd/kfd_packet_manager_vi.c    |   3 +-
 drivers/gpu/drm/amd/amdkfd/kfd_priv.h         |  11 +-
 drivers/gpu/drm/amd/amdkfd/kfd_process.c      | 115 ++++++++++--------
 .../amd/amdkfd/kfd_process_queue_manager.c    |  10 +-
 drivers/gpu/drm/amd/amdkfd/kfd_svm.c          |  19 +--
 16 files changed, 196 insertions(+), 181 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h
index 2f48dc5747aa2..a160ffe80b3de 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h
@@ -47,6 +47,7 @@ enum TLB_FLUSH_TYPE {
 };
 
 struct amdgpu_device;
+struct kfd_process_device;
 struct amdgpu_reset_context;
 
 enum kfd_mem_attachment_type {
@@ -298,8 +299,6 @@ bool amdgpu_amdkfd_compute_active(struct amdgpu_device *adev, uint32_t node_id);
 	(&((struct amdgpu_fpriv *)					\
 		((struct drm_file *)(drm_priv))->driver_priv)->vm)
 
-int amdgpu_amdkfd_gpuvm_set_vm_pasid(struct amdgpu_device *adev,
-				     struct amdgpu_vm *avm, u32 pasid);
 int amdgpu_amdkfd_gpuvm_acquire_process_vm(struct amdgpu_device *adev,
 					struct amdgpu_vm *avm,
 					void **process_info,
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c
index c0aaa72b6c210..0c8129f8d3635 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c
@@ -1521,27 +1521,6 @@ static void amdgpu_amdkfd_gpuvm_unpin_bo(struct amdgpu_bo *bo)
 	amdgpu_bo_unreserve(bo);
 }
 
-int amdgpu_amdkfd_gpuvm_set_vm_pasid(struct amdgpu_device *adev,
-				     struct amdgpu_vm *avm, u32 pasid)
-
-{
-	int ret;
-
-	/* Free the original amdgpu allocated pasid,
-	 * will be replaced with kfd allocated pasid.
-	 */
-	if (avm->pasid) {
-		amdgpu_pasid_free(avm->pasid);
-		amdgpu_vm_set_pasid(adev, avm, 0);
-	}
-
-	ret = amdgpu_vm_set_pasid(adev, avm, pasid);
-	if (ret)
-		return ret;
-
-	return 0;
-}
-
 int amdgpu_amdkfd_gpuvm_acquire_process_vm(struct amdgpu_device *adev,
 					   struct amdgpu_vm *avm,
 					   void **process_info,
diff --git a/drivers/gpu/drm/amd/amdkfd/cik_event_interrupt.c b/drivers/gpu/drm/amd/amdkfd/cik_event_interrupt.c
index 795382b55e0a9..981d9adcc5e1d 100644
--- a/drivers/gpu/drm/amd/amdkfd/cik_event_interrupt.c
+++ b/drivers/gpu/drm/amd/amdkfd/cik_event_interrupt.c
@@ -107,20 +107,30 @@ static void cik_event_interrupt_wq(struct kfd_node *dev,
 		kfd_signal_hw_exception_event(pasid);
 	else if (ihre->source_id == CIK_INTSRC_GFX_PAGE_INV_FAULT ||
 		ihre->source_id == CIK_INTSRC_GFX_MEM_PROT_FAULT) {
+		struct kfd_process_device *pdd = NULL;
 		struct kfd_vm_fault_info info;
+		struct kfd_process *p;
 
 		kfd_smi_event_update_vmfault(dev, pasid);
-		kfd_dqm_evict_pasid(dev->dqm, pasid);
+		p = kfd_lookup_process_by_pasid(pasid, &pdd);
+		if (!pdd)
+			return;
+
+		kfd_evict_process_device(pdd);
 
 		memset(&info, 0, sizeof(info));
 		amdgpu_amdkfd_gpuvm_get_vm_fault_info(dev->adev, &info);
-		if (!info.page_addr && !info.status)
+		if (!info.page_addr && !info.status) {
+			kfd_unref_process(p);
 			return;
+		}
 
 		if (info.vmid == vmid)
-			kfd_signal_vm_fault_event(dev, pasid, &info, NULL);
+			kfd_signal_vm_fault_event(pdd, &info, NULL);
 		else
-			kfd_signal_vm_fault_event(dev, pasid, NULL, NULL);
+			kfd_signal_vm_fault_event(pdd, &info, NULL);
+
+		kfd_unref_process(p);
 	}
 }
 
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c b/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c
index 33df35cab4679..8c2e92378b491 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c
@@ -155,8 +155,8 @@ static int kfd_open(struct inode *inode, struct file *filep)
 	/* filep now owns the reference returned by kfd_create_process */
 	filep->private_data = process;
 
-	dev_dbg(kfd_device, "process %d opened, compat mode (32 bit) - %d\n",
-		process->pasid, process->is_32bit_user_mode);
+	dev_dbg(kfd_device, "process pid %d opened kfd node, compat mode (32 bit) - %d\n",
+		process->lead_thread->pid, process->is_32bit_user_mode);
 
 	return 0;
 }
@@ -366,8 +366,8 @@ static int kfd_ioctl_create_queue(struct file *filep, struct kfd_process *p,
 		goto err_acquire_queue_buf;
 	}
 
-	pr_debug("Creating queue for PASID 0x%x on gpu 0x%x\n",
-			p->pasid,
+	pr_debug("Creating queue for process pid %d on gpu 0x%x\n",
+			p->lead_thread->pid,
 			dev->id);
 
 	err = pqm_create_queue(&p->pqm, dev, &q_properties, &queue_id,
@@ -420,9 +420,9 @@ static int kfd_ioctl_destroy_queue(struct file *filp, struct kfd_process *p,
 	int retval;
 	struct kfd_ioctl_destroy_queue_args *args = data;
 
-	pr_debug("Destroying queue id %d for pasid 0x%x\n",
+	pr_debug("Destroying queue id %d for process pid %d\n",
 				args->queue_id,
-				p->pasid);
+				p->lead_thread->pid);
 
 	mutex_lock(&p->mutex);
 
@@ -478,8 +478,8 @@ static int kfd_ioctl_update_queue(struct file *filp, struct kfd_process *p,
 	properties.pm4_target_xcc = (args->queue_percentage >> 8) & 0xFF;
 	properties.priority = args->queue_priority;
 
-	pr_debug("Updating queue id %d for pasid 0x%x\n",
-			args->queue_id, p->pasid);
+	pr_debug("Updating queue id %d for process pid %d\n",
+			args->queue_id, p->lead_thread->pid);
 
 	mutex_lock(&p->mutex);
 
@@ -705,7 +705,7 @@ static int kfd_ioctl_get_process_apertures(struct file *filp,
 	struct kfd_process_device_apertures *pAperture;
 	int i;
 
-	dev_dbg(kfd_device, "get apertures for PASID 0x%x", p->pasid);
+	dev_dbg(kfd_device, "get apertures for process pid %d", p->lead_thread->pid);
 
 	args->num_of_nodes = 0;
 
@@ -757,7 +757,8 @@ static int kfd_ioctl_get_process_apertures_new(struct file *filp,
 	int ret;
 	int i;
 
-	dev_dbg(kfd_device, "get apertures for PASID 0x%x", p->pasid);
+	dev_dbg(kfd_device, "get apertures for process pid %d",
+			p->lead_thread->pid);
 
 	if (args->num_of_nodes == 0) {
 		/* Return number of nodes, so that user space can alloacate
@@ -3375,12 +3376,12 @@ static int kfd_mmio_mmap(struct kfd_node *dev, struct kfd_process *process,
 
 	vma->vm_page_prot = pgprot_noncached(vma->vm_page_prot);
 
-	pr_debug("pasid 0x%x mapping mmio page\n"
+	pr_debug("process pid %d mapping mmio page\n"
 		 "     target user address == 0x%08llX\n"
 		 "     physical address    == 0x%08llX\n"
 		 "     vm_flags            == 0x%04lX\n"
 		 "     size                == 0x%04lX\n",
-		 process->pasid, (unsigned long long) vma->vm_start,
+		 process->lead_thread->pid, (unsigned long long) vma->vm_start,
 		 address, vma->vm_flags, PAGE_SIZE);
 
 	return io_remap_pfn_range(vma,
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_debug.c b/drivers/gpu/drm/amd/amdkfd/kfd_debug.c
index a8abc30918013..12456c61ffa54 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_debug.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_debug.c
@@ -204,11 +204,12 @@ bool kfd_set_dbg_ev_from_interrupt(struct kfd_node *dev,
 				   size_t exception_data_size)
 {
 	struct kfd_process *p;
+	struct kfd_process_device *pdd = NULL;
 	bool signaled_to_debugger_or_runtime = false;
 
-	p = kfd_lookup_process_by_pasid(pasid);
+	p = kfd_lookup_process_by_pasid(pasid, &pdd);
 
-	if (!p)
+	if (!pdd)
 		return false;
 
 	if (!kfd_dbg_ev_raise(trap_mask, p, dev, doorbell_id, true,
@@ -238,9 +239,8 @@ bool kfd_set_dbg_ev_from_interrupt(struct kfd_node *dev,
 
 			mutex_unlock(&p->mutex);
 		} else if (trap_mask & KFD_EC_MASK(EC_DEVICE_MEMORY_VIOLATION)) {
-			kfd_dqm_evict_pasid(dev->dqm, p->pasid);
-			kfd_signal_vm_fault_event(dev, p->pasid, NULL,
-							exception_data);
+			kfd_evict_process_device(pdd);
+			kfd_signal_vm_fault_event(pdd, NULL, exception_data);
 
 			signaled_to_debugger_or_runtime = true;
 		}
@@ -276,8 +276,8 @@ int kfd_dbg_send_exception_to_runtime(struct kfd_process *p,
 		data = (struct kfd_hsa_memory_exception_data *)
 						pdd->vm_fault_exc_data;
 
-		kfd_dqm_evict_pasid(pdd->dev->dqm, p->pasid);
-		kfd_signal_vm_fault_event(pdd->dev, p->pasid, NULL, data);
+		kfd_evict_process_device(pdd);
+		kfd_signal_vm_fault_event(pdd, NULL, data);
 		error_reason &= ~KFD_EC_MASK(EC_DEVICE_MEMORY_VIOLATION);
 	}
 
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_device.c b/drivers/gpu/drm/amd/amdkfd/kfd_device.c
index 6cefd338f23de..bf978b368f6a5 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_device.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_device.c
@@ -1558,7 +1558,7 @@ bool kgd2kfd_vmfault_fast_path(struct amdgpu_device *adev, struct amdgpu_iv_entr
 	u32 cam_index;
 
 	if (entry->ih == &adev->irq.ih_soft || entry->ih == &adev->irq.ih1) {
-		p = kfd_lookup_process_by_pasid(entry->pasid);
+		p = kfd_lookup_process_by_pasid(entry->pasid, NULL);
 		if (!p)
 			return true;
 
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 7f8ec2a152ac6..35ae3c55a97fa 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c
@@ -208,7 +208,7 @@ static int add_queue_mes(struct device_queue_manager *dqm, struct queue *q,
 		return -EIO;
 
 	memset(&queue_input, 0x0, sizeof(struct mes_add_queue_input));
-	queue_input.process_id = qpd->pqm->process->pasid;
+	queue_input.process_id = pdd->pasid;
 	queue_input.page_table_base_addr =  qpd->page_table_base;
 	queue_input.process_va_start = 0;
 	queue_input.process_va_end = adev->vm_manager.max_pfn - 1;
@@ -527,6 +527,7 @@ static int allocate_vmid(struct device_queue_manager *dqm,
 			struct qcm_process_device *qpd,
 			struct queue *q)
 {
+	struct kfd_process_device *pdd = qpd_to_pdd(qpd);
 	struct device *dev = dqm->dev->adev->dev;
 	int allocated_vmid = -1, i;
 
@@ -545,9 +546,9 @@ static int allocate_vmid(struct device_queue_manager *dqm,
 
 	pr_debug("vmid allocated: %d\n", allocated_vmid);
 
-	dqm->vmid_pasid[allocated_vmid] = q->process->pasid;
+	dqm->vmid_pasid[allocated_vmid] = pdd->pasid;
 
-	set_pasid_vmid_mapping(dqm, q->process->pasid, allocated_vmid);
+	set_pasid_vmid_mapping(dqm, pdd->pasid, allocated_vmid);
 
 	qpd->vmid = allocated_vmid;
 	q->properties.vmid = allocated_vmid;
@@ -799,6 +800,11 @@ static int dbgdev_wave_reset_wavefronts(struct kfd_node *dev, struct kfd_process
 		return -EOPNOTSUPP;
 	}
 
+	/* taking the VMID for that process on the safe way using PDD */
+	pdd = kfd_get_process_device_data(dev, p);
+	if (!pdd)
+		return -EFAULT;
+
 	/* Scan all registers in the range ATC_VMID8_PASID_MAPPING ..
 	 * ATC_VMID15_PASID_MAPPING
 	 * to check which VMID the current process is mapped to.
@@ -808,23 +814,19 @@ static int dbgdev_wave_reset_wavefronts(struct kfd_node *dev, struct kfd_process
 		status = dev->kfd2kgd->get_atc_vmid_pasid_mapping_info
 				(dev->adev, vmid, &queried_pasid);
 
-		if (status && queried_pasid == p->pasid) {
-			pr_debug("Killing wave fronts of vmid %d and pasid 0x%x\n",
-					vmid, p->pasid);
+		if (status && queried_pasid == pdd->pasid) {
+			pr_debug("Killing wave fronts of vmid %d and process pid %d\n",
+					vmid, p->lead_thread->pid);
 			break;
 		}
 	}
 
 	if (vmid > last_vmid_to_scan) {
-		dev_err(dev->adev->dev, "Didn't find vmid for pasid 0x%x\n", p->pasid);
+		dev_err(dev->adev->dev, "Didn't find vmid for process pid %d\n",
+				p->lead_thread->pid);
 		return -EFAULT;
 	}
 
-	/* taking the VMID for that process on the safe way using PDD */
-	pdd = kfd_get_process_device_data(dev, p);
-	if (!pdd)
-		return -EFAULT;
-
 	reg_gfx_index.bits.sh_broadcast_writes = 1;
 	reg_gfx_index.bits.se_broadcast_writes = 1;
 	reg_gfx_index.bits.instance_broadcast_writes = 1;
@@ -1060,8 +1062,8 @@ static int suspend_single_queue(struct device_queue_manager *dqm,
 	if (q->properties.is_suspended)
 		return 0;
 
-	pr_debug("Suspending PASID %u queue [%i]\n",
-			pdd->process->pasid,
+	pr_debug("Suspending process pid %d queue [%i]\n",
+			pdd->process->lead_thread->pid,
 			q->properties.queue_id);
 
 	is_new = q->properties.exception_status & KFD_EC_MASK(EC_QUEUE_NEW);
@@ -1108,8 +1110,8 @@ static int resume_single_queue(struct device_queue_manager *dqm,
 
 	pdd = qpd_to_pdd(qpd);
 
-	pr_debug("Restoring from suspend PASID %u queue [%i]\n",
-			    pdd->process->pasid,
+	pr_debug("Restoring from suspend process pid %d queue [%i]\n",
+			    pdd->process->lead_thread->pid,
 			    q->properties.queue_id);
 
 	q->properties.is_suspended = false;
@@ -1142,8 +1144,8 @@ static int evict_process_queues_nocpsch(struct device_queue_manager *dqm,
 		goto out;
 
 	pdd = qpd_to_pdd(qpd);
-	pr_debug_ratelimited("Evicting PASID 0x%x queues\n",
-			    pdd->process->pasid);
+	pr_debug_ratelimited("Evicting process pid %d queues\n",
+			    pdd->process->lead_thread->pid);
 
 	pdd->last_evict_timestamp = get_jiffies_64();
 	/* Mark all queues as evicted. Deactivate all active queues on
@@ -1200,8 +1202,8 @@ static int evict_process_queues_cpsch(struct device_queue_manager *dqm,
 	if (!pdd->drm_priv)
 		goto out;
 
-	pr_debug_ratelimited("Evicting PASID 0x%x queues\n",
-			    pdd->process->pasid);
+	pr_debug_ratelimited("Evicting process pid %d queues\n",
+			    pdd->process->lead_thread->pid);
 
 	/* Mark all queues as evicted. Deactivate all active queues on
 	 * the qpd.
@@ -1261,8 +1263,8 @@ static int restore_process_queues_nocpsch(struct device_queue_manager *dqm,
 		goto out;
 	}
 
-	pr_debug_ratelimited("Restoring PASID 0x%x queues\n",
-			    pdd->process->pasid);
+	pr_debug_ratelimited("Restoring process pid %d queues\n",
+			    pdd->process->lead_thread->pid);
 
 	/* Update PD Base in QPD */
 	qpd->page_table_base = pd_base;
@@ -1345,8 +1347,8 @@ static int restore_process_queues_cpsch(struct device_queue_manager *dqm,
 	if (!pdd->drm_priv)
 		goto vm_not_acquired;
 
-	pr_debug_ratelimited("Restoring PASID 0x%x queues\n",
-			    pdd->process->pasid);
+	pr_debug_ratelimited("Restoring process pid %d queues\n",
+			    pdd->process->lead_thread->pid);
 
 	/* Update PD Base in QPD */
 	qpd->page_table_base = amdgpu_amdkfd_gpuvm_get_process_page_dir(pdd->drm_priv);
@@ -2137,8 +2139,8 @@ static void set_queue_as_reset(struct device_queue_manager *dqm, struct queue *q
 {
 	struct kfd_process_device *pdd = qpd_to_pdd(qpd);
 
-	dev_err(dqm->dev->adev->dev, "queue id 0x%0x at pasid 0x%0x is reset\n",
-		q->properties.queue_id, q->process->pasid);
+	dev_err(dqm->dev->adev->dev, "queue id 0x%0x at pasid %d is reset\n",
+		q->properties.queue_id, pdd->process->lead_thread->pid);
 
 	pdd->has_reset_queue = true;
 	if (q->properties.is_active) {
@@ -2941,20 +2943,19 @@ void device_queue_manager_uninit(struct device_queue_manager *dqm)
 
 int kfd_dqm_suspend_bad_queue_mes(struct kfd_node *knode, u32 pasid, u32 doorbell_id)
 {
-	struct kfd_process_device *pdd;
-	struct kfd_process *p = kfd_lookup_process_by_pasid(pasid);
+	struct kfd_process_device *pdd = NULL;
+	struct kfd_process *p = kfd_lookup_process_by_pasid(pasid, &pdd);
 	struct device_queue_manager *dqm = knode->dqm;
 	struct device *dev = dqm->dev->adev->dev;
 	struct qcm_process_device *qpd;
 	struct queue *q = NULL;
 	int ret = 0;
 
-	if (!p)
+	if (!pdd)
 		return -EINVAL;
 
 	dqm_lock(dqm);
 
-	pdd = kfd_get_process_device_data(dqm->dev, p);
 	if (pdd) {
 		qpd = &pdd->qpd;
 
@@ -2987,6 +2988,7 @@ int kfd_dqm_suspend_bad_queue_mes(struct kfd_node *knode, u32 pasid, u32 doorbel
 
 out:
 	dqm_unlock(dqm);
+	kfd_unref_process(p);
 	return ret;
 }
 
@@ -3028,24 +3030,21 @@ static int kfd_dqm_evict_pasid_mes(struct device_queue_manager *dqm,
 	return ret;
 }
 
-int kfd_dqm_evict_pasid(struct device_queue_manager *dqm, u32 pasid)
+int kfd_evict_process_device(struct kfd_process_device *pdd)
 {
-	struct kfd_process_device *pdd;
-	struct kfd_process *p = kfd_lookup_process_by_pasid(pasid);
+	struct device_queue_manager *dqm;
+	struct kfd_process *p;
 	int ret = 0;
 
-	if (!p)
-		return -EINVAL;
+	p = pdd->process;
+	dqm = pdd->dev->dqm;
+
 	WARN(debug_evictions, "Evicting pid %d", p->lead_thread->pid);
-	pdd = kfd_get_process_device_data(dqm->dev, p);
-	if (pdd) {
-		if (dqm->dev->kfd->shared_resources.enable_mes)
-			ret = kfd_dqm_evict_pasid_mes(dqm, &pdd->qpd);
-		else
-			ret = dqm->ops.evict_process_queues(dqm, &pdd->qpd);
-	}
 
-	kfd_unref_process(p);
+	if (dqm->dev->kfd->shared_resources.enable_mes)
+		ret = kfd_dqm_evict_pasid_mes(dqm, &pdd->qpd);
+	else
+		ret = dqm->ops.evict_process_queues(dqm, &pdd->qpd);
 
 	return ret;
 }
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_events.c b/drivers/gpu/drm/amd/amdkfd/kfd_events.c
index d075f24e5f9f3..fecdb67940750 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_events.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_events.c
@@ -727,7 +727,7 @@ void kfd_signal_event_interrupt(u32 pasid, uint32_t partial_id,
 	 * to process context, kfd_process could attempt to exit while we are
 	 * running so the lookup function increments the process ref count.
 	 */
-	struct kfd_process *p = kfd_lookup_process_by_pasid(pasid);
+	struct kfd_process *p = kfd_lookup_process_by_pasid(pasid, NULL);
 
 	if (!p)
 		return; /* Presumably process exited. */
@@ -1139,8 +1139,8 @@ static void lookup_events_by_type_and_signal(struct kfd_process *p,
 
 	if (type == KFD_EVENT_TYPE_MEMORY) {
 		dev_warn(kfd_device,
-			"Sending SIGSEGV to process %d (pasid 0x%x)",
-				p->lead_thread->pid, p->pasid);
+			"Sending SIGSEGV to process pid %d",
+				p->lead_thread->pid);
 		send_sig(SIGSEGV, p->lead_thread, 0);
 	}
 
@@ -1148,13 +1148,13 @@ static void lookup_events_by_type_and_signal(struct kfd_process *p,
 	if (send_signal) {
 		if (send_sigterm) {
 			dev_warn(kfd_device,
-				"Sending SIGTERM to process %d (pasid 0x%x)",
-					p->lead_thread->pid, p->pasid);
+				"Sending SIGTERM to process pid %d",
+					p->lead_thread->pid);
 			send_sig(SIGTERM, p->lead_thread, 0);
 		} else {
 			dev_err(kfd_device,
-				"Process %d (pasid 0x%x) got unhandled exception",
-				p->lead_thread->pid, p->pasid);
+				"Process pid %d got unhandled exception",
+				p->lead_thread->pid);
 		}
 	}
 
@@ -1168,7 +1168,7 @@ void kfd_signal_hw_exception_event(u32 pasid)
 	 * to process context, kfd_process could attempt to exit while we are
 	 * running so the lookup function increments the process ref count.
 	 */
-	struct kfd_process *p = kfd_lookup_process_by_pasid(pasid);
+	struct kfd_process *p = kfd_lookup_process_by_pasid(pasid, NULL);
 
 	if (!p)
 		return; /* Presumably process exited. */
@@ -1177,22 +1177,20 @@ void kfd_signal_hw_exception_event(u32 pasid)
 	kfd_unref_process(p);
 }
 
-void kfd_signal_vm_fault_event(struct kfd_node *dev, u32 pasid,
+void kfd_signal_vm_fault_event(struct kfd_process_device *pdd,
 				struct kfd_vm_fault_info *info,
 				struct kfd_hsa_memory_exception_data *data)
 {
 	struct kfd_event *ev;
 	uint32_t id;
-	struct kfd_process *p = kfd_lookup_process_by_pasid(pasid);
+	struct kfd_process *p = pdd->process;
 	struct kfd_hsa_memory_exception_data memory_exception_data;
 	int user_gpu_id;
 
-	if (!p)
-		return; /* Presumably process exited. */
-
-	user_gpu_id = kfd_process_get_user_gpu_id(p, dev->id);
+	user_gpu_id = kfd_process_get_user_gpu_id(p, pdd->dev->id);
 	if (unlikely(user_gpu_id == -EINVAL)) {
-		WARN_ONCE(1, "Could not get user_gpu_id from dev->id:%x\n", dev->id);
+		WARN_ONCE(1, "Could not get user_gpu_id from dev->id:%x\n",
+			  pdd->dev->id);
 		return;
 	}
 
@@ -1229,7 +1227,6 @@ void kfd_signal_vm_fault_event(struct kfd_node *dev, u32 pasid,
 		}
 
 	rcu_read_unlock();
-	kfd_unref_process(p);
 }
 
 void kfd_signal_reset_event(struct kfd_node *dev)
@@ -1264,7 +1261,8 @@ void kfd_signal_reset_event(struct kfd_node *dev)
 		}
 
 		if (unlikely(!pdd)) {
-			WARN_ONCE(1, "Could not get device data from pasid:0x%x\n", p->pasid);
+			WARN_ONCE(1, "Could not get device data from process pid:%d\n",
+				  p->lead_thread->pid);
 			continue;
 		}
 
@@ -1273,8 +1271,15 @@ void kfd_signal_reset_event(struct kfd_node *dev)
 
 		if (dev->dqm->detect_hang_count) {
 			struct amdgpu_task_info *ti;
+			struct amdgpu_fpriv *drv_priv;
+
+			if (unlikely(amdgpu_file_to_fpriv(pdd->drm_file, &drv_priv))) {
+				WARN_ONCE(1, "Could not get vm for device %x from pid:%d\n",
+					  dev->id, p->lead_thread->pid);
+				continue;
+			}
 
-			ti = amdgpu_vm_get_task_info_pasid(dev->adev, p->pasid);
+			ti = amdgpu_vm_get_task_info_vm(&drv_priv->vm);
 			if (ti) {
 				dev_err(dev->adev->dev,
 					"Queues reset on process %s tid %d thread %s pid %d\n",
@@ -1311,7 +1316,7 @@ void kfd_signal_reset_event(struct kfd_node *dev)
 
 void kfd_signal_poison_consumed_event(struct kfd_node *dev, u32 pasid)
 {
-	struct kfd_process *p = kfd_lookup_process_by_pasid(pasid);
+	struct kfd_process *p = kfd_lookup_process_by_pasid(pasid, NULL);
 	struct kfd_hsa_memory_exception_data memory_exception_data;
 	struct kfd_hsa_hw_exception_data hw_exception_data;
 	struct kfd_event *ev;
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_int_process_v11.c b/drivers/gpu/drm/amd/amdkfd/kfd_int_process_v11.c
index b3f988b275a88..c5f97e6e36ff5 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_int_process_v11.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_int_process_v11.c
@@ -194,7 +194,7 @@ static void event_interrupt_poison_consumption_v11(struct kfd_node *dev,
 	enum amdgpu_ras_block block = 0;
 	int ret = -EINVAL;
 	uint32_t reset = 0;
-	struct kfd_process *p = kfd_lookup_process_by_pasid(pasid);
+	struct kfd_process *p = kfd_lookup_process_by_pasid(pasid, NULL);
 
 	if (!p)
 		return;
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_int_process_v9.c b/drivers/gpu/drm/amd/amdkfd/kfd_int_process_v9.c
index 0cb5c582ce7dc..b8a91bf4ef307 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_int_process_v9.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_int_process_v9.c
@@ -146,7 +146,7 @@ static void event_interrupt_poison_consumption_v9(struct kfd_node *dev,
 {
 	enum amdgpu_ras_block block = 0;
 	uint32_t reset = 0;
-	struct kfd_process *p = kfd_lookup_process_by_pasid(pasid);
+	struct kfd_process *p = kfd_lookup_process_by_pasid(pasid, NULL);
 	enum ras_event_type type = RAS_EVENT_TYPE_POISON_CONSUMPTION;
 	u64 event_id;
 	int old_poison, ret;
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_packet_manager_v9.c b/drivers/gpu/drm/amd/amdkfd/kfd_packet_manager_v9.c
index 1f9f5bfeaf868..d56525201155a 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_packet_manager_v9.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_packet_manager_v9.c
@@ -47,7 +47,7 @@ static int pm_map_process_v9(struct packet_manager *pm,
 		packet->bitfields2.exec_cleaner_shader = 1;
 	packet->bitfields2.diq_enable = (qpd->is_debug) ? 1 : 0;
 	packet->bitfields2.process_quantum = 10;
-	packet->bitfields2.pasid = qpd->pqm->process->pasid;
+	packet->bitfields2.pasid = pdd->pasid;
 	packet->bitfields14.gds_size = qpd->gds_size & 0x3F;
 	packet->bitfields14.gds_size_hi = (qpd->gds_size >> 6) & 0xF;
 	packet->bitfields14.num_gws = (qpd->mapped_gws_queue) ? qpd->num_gws : 0;
@@ -106,7 +106,7 @@ static int pm_map_process_aldebaran(struct packet_manager *pm,
 		packet->bitfields2.exec_cleaner_shader = 1;
 	packet->bitfields2.diq_enable = (qpd->is_debug) ? 1 : 0;
 	packet->bitfields2.process_quantum = 10;
-	packet->bitfields2.pasid = qpd->pqm->process->pasid;
+	packet->bitfields2.pasid = pdd->pasid;
 	packet->bitfields14.gds_size = qpd->gds_size & 0x3F;
 	packet->bitfields14.gds_size_hi = (qpd->gds_size >> 6) & 0xF;
 	packet->bitfields14.num_gws = (qpd->mapped_gws_queue) ? qpd->num_gws : 0;
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_packet_manager_vi.c b/drivers/gpu/drm/amd/amdkfd/kfd_packet_manager_vi.c
index c1199d06d131b..347c86e1c378f 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_packet_manager_vi.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_packet_manager_vi.c
@@ -42,6 +42,7 @@ unsigned int pm_build_pm4_header(unsigned int opcode, size_t packet_size)
 static int pm_map_process_vi(struct packet_manager *pm, uint32_t *buffer,
 				struct qcm_process_device *qpd)
 {
+	struct kfd_process_device *pdd = qpd_to_pdd(qpd);
 	struct pm4_mes_map_process *packet;
 
 	packet = (struct pm4_mes_map_process *)buffer;
@@ -52,7 +53,7 @@ static int pm_map_process_vi(struct packet_manager *pm, uint32_t *buffer,
 					sizeof(struct pm4_mes_map_process));
 	packet->bitfields2.diq_enable = (qpd->is_debug) ? 1 : 0;
 	packet->bitfields2.process_quantum = 10;
-	packet->bitfields2.pasid = qpd->pqm->process->pasid;
+	packet->bitfields2.pasid = pdd->pasid;
 	packet->bitfields3.page_table_base = qpd->page_table_base;
 	packet->bitfields10.gds_size = qpd->gds_size;
 	packet->bitfields10.num_gws = qpd->num_gws;
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_priv.h b/drivers/gpu/drm/amd/amdkfd/kfd_priv.h
index d8cd913aa772b..0a99c5c9cadc0 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_priv.h
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_priv.h
@@ -851,6 +851,8 @@ struct kfd_process_device {
 
 	/* Tracks queue reset status */
 	bool has_reset_queue;
+
+	u32 pasid;
 };
 
 #define qpd_to_pdd(x) container_of(x, struct kfd_process_device, qpd)
@@ -910,8 +912,6 @@ struct kfd_process {
 	/* We want to receive a notification when the mm_struct is destroyed */
 	struct mmu_notifier mmu_notifier;
 
-	u32 pasid;
-
 	/*
 	 * Array of kfd_process_device pointers,
 	 * one for each device the process is using.
@@ -1039,7 +1039,8 @@ void kfd_process_destroy_wq(void);
 void kfd_cleanup_processes(void);
 struct kfd_process *kfd_create_process(struct task_struct *thread);
 struct kfd_process *kfd_get_process(const struct task_struct *task);
-struct kfd_process *kfd_lookup_process_by_pasid(u32 pasid);
+struct kfd_process *kfd_lookup_process_by_pasid(u32 pasid,
+						 struct kfd_process_device **pdd);
 struct kfd_process *kfd_lookup_process_by_mm(const struct mm_struct *mm);
 
 int kfd_process_gpuidx_from_gpuid(struct kfd_process *p, uint32_t gpu_id);
@@ -1337,7 +1338,7 @@ void device_queue_manager_uninit(struct device_queue_manager *dqm);
 struct kernel_queue *kernel_queue_init(struct kfd_node *dev,
 					enum kfd_queue_type type);
 void kernel_queue_uninit(struct kernel_queue *kq);
-int kfd_dqm_evict_pasid(struct device_queue_manager *dqm, u32 pasid);
+int kfd_evict_process_device(struct kfd_process_device *pdd);
 int kfd_dqm_suspend_bad_queue_mes(struct kfd_node *knode, u32 pasid, u32 doorbell_id);
 
 /* Process Queue Manager */
@@ -1492,7 +1493,7 @@ int kfd_event_create(struct file *devkfd, struct kfd_process *p,
 int kfd_get_num_events(struct kfd_process *p);
 int kfd_event_destroy(struct kfd_process *p, uint32_t event_id);
 
-void kfd_signal_vm_fault_event(struct kfd_node *dev, u32 pasid,
+void kfd_signal_vm_fault_event(struct kfd_process_device *pdd,
 				struct kfd_vm_fault_info *info,
 				struct kfd_hsa_memory_exception_data *data);
 
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_process.c b/drivers/gpu/drm/amd/amdkfd/kfd_process.c
index c9cbc0ecd9cb2..a7e0a16dac47b 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_process.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_process.c
@@ -283,8 +283,8 @@ static int kfd_get_cu_occupancy(struct attribute *attr, char *buffer)
 	cu_cnt = 0;
 	proc = pdd->process;
 	if (pdd->qpd.queue_count == 0) {
-		pr_debug("Gpu-Id: %d has no active queues for process %d\n",
-			 dev->id, proc->pasid);
+		pr_debug("Gpu-Id: %d has no active queues for process pid %d\n",
+			 dev->id, (int)proc->lead_thread->pid);
 		return snprintf(buffer, PAGE_SIZE, "%d\n", cu_cnt);
 	}
 
@@ -328,12 +328,9 @@ static int kfd_get_cu_occupancy(struct attribute *attr, char *buffer)
 static ssize_t kfd_procfs_show(struct kobject *kobj, struct attribute *attr,
 			       char *buffer)
 {
-	if (strcmp(attr->name, "pasid") == 0) {
-		struct kfd_process *p = container_of(attr, struct kfd_process,
-						     attr_pasid);
-
-		return snprintf(buffer, PAGE_SIZE, "%d\n", p->pasid);
-	} else if (strncmp(attr->name, "vram_", 5) == 0) {
+	if (strcmp(attr->name, "pasid") == 0)
+		return snprintf(buffer, PAGE_SIZE, "%d\n", 0);
+	else if (strncmp(attr->name, "vram_", 5) == 0) {
 		struct kfd_process_device *pdd = container_of(attr, struct kfd_process_device,
 							      attr_vram);
 		return snprintf(buffer, PAGE_SIZE, "%llu\n", atomic64_read(&pdd->vram_usage));
@@ -1057,9 +1054,8 @@ static void kfd_process_destroy_pdds(struct kfd_process *p)
 	for (i = 0; i < p->n_pdds; i++) {
 		struct kfd_process_device *pdd = p->pdds[i];
 
-		pr_debug("Releasing pdd (topology id %d) for process (pasid 0x%x)\n",
-				pdd->dev->id, p->pasid);
-
+		pr_debug("Releasing pdd (topology id %d, for pid %d)\n",
+			pdd->dev->id, p->lead_thread->pid);
 		kfd_process_device_destroy_cwsr_dgpu(pdd);
 		kfd_process_device_destroy_ib_mem(pdd);
 
@@ -1191,7 +1187,6 @@ static void kfd_process_wq_release(struct work_struct *work)
 
 	kfd_event_free_process(p);
 
-	kfd_pasid_free(p->pasid);
 	mutex_destroy(&p->mutex);
 
 	put_task_struct(p->lead_thread);
@@ -1542,12 +1537,6 @@ static struct kfd_process *create_process(const struct task_struct *thread)
 	atomic_set(&process->debugged_process_count, 0);
 	sema_init(&process->runtime_enable_sema, 0);
 
-	process->pasid = kfd_pasid_alloc();
-	if (process->pasid == 0) {
-		err = -ENOSPC;
-		goto err_alloc_pasid;
-	}
-
 	err = pqm_init(&process->pqm, process);
 	if (err != 0)
 		goto err_process_pqm_init;
@@ -1601,8 +1590,6 @@ static struct kfd_process *create_process(const struct task_struct *thread)
 err_init_apertures:
 	pqm_uninit(&process->pqm);
 err_process_pqm_init:
-	kfd_pasid_free(process->pasid);
-err_alloc_pasid:
 	kfd_event_free_process(process);
 err_event_init:
 	mutex_destroy(&process->mutex);
@@ -1721,15 +1708,18 @@ int kfd_process_device_init_vm(struct kfd_process_device *pdd,
 	if (ret)
 		goto err_init_cwsr;
 
-	ret = amdgpu_amdkfd_gpuvm_set_vm_pasid(dev->adev, avm, p->pasid);
-	if (ret)
-		goto err_set_pasid;
+	if (unlikely(!avm->pasid)) {
+		dev_warn(pdd->dev->adev->dev, "WARN: vm %p has no pasid associated",
+				 avm);
+		goto err_get_pasid;
+	}
 
+	pdd->pasid = avm->pasid;
 	pdd->drm_file = drm_file;
 
 	return 0;
 
-err_set_pasid:
+err_get_pasid:
 	kfd_process_device_destroy_cwsr_dgpu(pdd);
 err_init_cwsr:
 	kfd_process_device_destroy_ib_mem(pdd);
@@ -1815,25 +1805,50 @@ void kfd_process_device_remove_obj_handle(struct kfd_process_device *pdd,
 		idr_remove(&pdd->alloc_idr, handle);
 }
 
-/* This increments the process->ref counter. */
-struct kfd_process *kfd_lookup_process_by_pasid(u32 pasid)
+static struct kfd_process_device *kfd_lookup_process_device_by_pasid(u32 pasid)
 {
-	struct kfd_process *p, *ret_p = NULL;
+	struct kfd_process_device *ret_p = NULL;
+	struct kfd_process *p;
 	unsigned int temp;
-
-	int idx = srcu_read_lock(&kfd_processes_srcu);
+	int i;
 
 	hash_for_each_rcu(kfd_processes_table, temp, p, kfd_processes) {
-		if (p->pasid == pasid) {
-			kref_get(&p->ref);
-			ret_p = p;
-			break;
+		for (i = 0; i < p->n_pdds; i++) {
+			if (p->pdds[i]->pasid == pasid) {
+				ret_p = p->pdds[i];
+				break;
+			}
 		}
+		if (ret_p)
+			break;
+	}
+	return ret_p;
+}
+
+/* This increments the process->ref counter. */
+struct kfd_process *kfd_lookup_process_by_pasid(u32 pasid,
+						struct kfd_process_device **pdd)
+{
+	struct kfd_process_device *ret_p;
+
+	int idx = srcu_read_lock(&kfd_processes_srcu);
+
+	ret_p = kfd_lookup_process_device_by_pasid(pasid);
+	if (ret_p) {
+		if (pdd)
+			*pdd = ret_p;
+		kref_get(&ret_p->process->ref);
+
+		srcu_read_unlock(&kfd_processes_srcu, idx);
+		return ret_p->process;
 	}
 
 	srcu_read_unlock(&kfd_processes_srcu, idx);
 
-	return ret_p;
+	if (pdd)
+		*pdd = NULL;
+
+	return NULL;
 }
 
 /* This increments the process->ref counter. */
@@ -1989,7 +2004,7 @@ static void evict_process_worker(struct work_struct *work)
 	 */
 	p = container_of(dwork, struct kfd_process, eviction_work);
 
-	pr_debug("Started evicting pasid 0x%x\n", p->pasid);
+	pr_debug("Started evicting process pid %d\n", p->lead_thread->pid);
 	ret = kfd_process_evict_queues(p, KFD_QUEUE_EVICTION_TRIGGER_TTM);
 	if (!ret) {
 		/* If another thread already signaled the eviction fence,
@@ -2001,9 +2016,9 @@ static void evict_process_worker(struct work_struct *work)
 				     msecs_to_jiffies(PROCESS_RESTORE_TIME_MS)))
 			kfd_process_restore_queues(p);
 
-		pr_debug("Finished evicting pasid 0x%x\n", p->pasid);
+		pr_debug("Finished evicting process pid %d\n", p->lead_thread->pid);
 	} else
-		pr_err("Failed to evict queues of pasid 0x%x\n", p->pasid);
+		pr_err("Failed to evict queues of process pid %d\n", p->lead_thread->pid);
 }
 
 static int restore_process_helper(struct kfd_process *p)
@@ -2020,9 +2035,11 @@ static int restore_process_helper(struct kfd_process *p)
 
 	ret = kfd_process_restore_queues(p);
 	if (!ret)
-		pr_debug("Finished restoring pasid 0x%x\n", p->pasid);
+		pr_debug("Finished restoring process pid %d\n",
+			p->lead_thread->pid);
 	else
-		pr_err("Failed to restore queues of pasid 0x%x\n", p->pasid);
+		pr_err("Failed to restore queues of process pid %d\n",
+		      p->lead_thread->pid);
 
 	return ret;
 }
@@ -2039,7 +2056,7 @@ static void restore_process_worker(struct work_struct *work)
 	 * lifetime of this thread, kfd_process p will be valid
 	 */
 	p = container_of(dwork, struct kfd_process, restore_work);
-	pr_debug("Started restoring pasid 0x%x\n", p->pasid);
+	pr_debug("Started restoring process pasid %d\n", (int)p->lead_thread->pid);
 
 	/* Setting last_restore_timestamp before successful restoration.
 	 * Otherwise this would have to be set by KGD (restore_process_bos)
@@ -2055,8 +2072,8 @@ static void restore_process_worker(struct work_struct *work)
 
 	ret = restore_process_helper(p);
 	if (ret) {
-		pr_debug("Failed to restore BOs of pasid 0x%x, retry after %d ms\n",
-			 p->pasid, PROCESS_BACK_OFF_TIME_MS);
+		pr_debug("Failed to restore BOs of process pid %d, retry after %d ms\n",
+			 p->lead_thread->pid, PROCESS_BACK_OFF_TIME_MS);
 		if (mod_delayed_work(kfd_restore_wq, &p->restore_work,
 				     msecs_to_jiffies(PROCESS_RESTORE_TIME_MS)))
 			kfd_process_restore_queues(p);
@@ -2072,7 +2089,7 @@ void kfd_suspend_all_processes(void)
 	WARN(debug_evictions, "Evicting all processes");
 	hash_for_each_rcu(kfd_processes_table, temp, p, kfd_processes) {
 		if (kfd_process_evict_queues(p, KFD_QUEUE_EVICTION_TRIGGER_SUSPEND))
-			pr_err("Failed to suspend process 0x%x\n", p->pasid);
+			pr_err("Failed to suspend process pid %d\n", p->lead_thread->pid);
 		signal_eviction_fence(p);
 	}
 	srcu_read_unlock(&kfd_processes_srcu, idx);
@@ -2086,8 +2103,8 @@ int kfd_resume_all_processes(void)
 
 	hash_for_each_rcu(kfd_processes_table, temp, p, kfd_processes) {
 		if (restore_process_helper(p)) {
-			pr_err("Restore process %d failed during resume\n",
-			       p->pasid);
+			pr_err("Restore process pid %d failed during resume\n",
+			      p->lead_thread->pid);
 			ret = -EFAULT;
 		}
 	}
@@ -2142,7 +2159,7 @@ int kfd_process_drain_interrupts(struct kfd_process_device *pdd)
 	memset(irq_drain_fence, 0, sizeof(irq_drain_fence));
 	irq_drain_fence[0] = (KFD_IRQ_FENCE_SOURCEID << 8) |
 							KFD_IRQ_FENCE_CLIENTID;
-	irq_drain_fence[3] = pdd->process->pasid;
+	irq_drain_fence[3] = pdd->pasid;
 
 	/*
 	 * For GFX 9.4.3/9.5.0, send the NodeId also in IH cookie DW[3]
@@ -2173,7 +2190,7 @@ void kfd_process_close_interrupt_drain(unsigned int pasid)
 {
 	struct kfd_process *p;
 
-	p = kfd_lookup_process_by_pasid(pasid);
+	p = kfd_lookup_process_by_pasid(pasid, NULL);
 
 	if (!p)
 		return;
@@ -2294,8 +2311,8 @@ int kfd_debugfs_mqds_by_process(struct seq_file *m, void *data)
 	int idx = srcu_read_lock(&kfd_processes_srcu);
 
 	hash_for_each_rcu(kfd_processes_table, temp, p, kfd_processes) {
-		seq_printf(m, "Process %d PASID 0x%x:\n",
-			   p->lead_thread->tgid, p->pasid);
+		seq_printf(m, "Process %d PASID %d:\n",
+			   p->lead_thread->tgid, p->lead_thread->pid);
 
 		mutex_lock(&p->mutex);
 		r = pqm_debugfs_mqds(m, &p->pqm);
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_process_queue_manager.c b/drivers/gpu/drm/amd/amdkfd/kfd_process_queue_manager.c
index d79caa1a68676..662c595ce7838 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_process_queue_manager.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_process_queue_manager.c
@@ -69,8 +69,8 @@ static int find_available_queue_slot(struct process_queue_manager *pqm,
 	pr_debug("The new slot id %lu\n", found);
 
 	if (found >= KFD_MAX_NUM_OF_QUEUES_PER_PROCESS) {
-		pr_info("Cannot open more queues for process with pasid 0x%x\n",
-				pqm->process->pasid);
+		pr_info("Cannot open more queues for process with pid %d\n",
+			pqm->process->lead_thread->pid);
 		return -ENOMEM;
 	}
 
@@ -451,8 +451,8 @@ int pqm_create_queue(struct process_queue_manager *pqm,
 	}
 
 	if (retval != 0) {
-		pr_err("Pasid 0x%x DQM create queue type %d failed. ret %d\n",
-			pqm->process->pasid, type, retval);
+		pr_err("process pid %d DQM create queue type %d failed. ret %d\n",
+			pqm->process->lead_thread->pid, type, retval);
 		goto err_create_queue;
 	}
 
@@ -546,7 +546,7 @@ int pqm_destroy_queue(struct process_queue_manager *pqm, unsigned int qid)
 		retval = dqm->ops.destroy_queue(dqm, &pdd->qpd, pqn->q);
 		if (retval) {
 			pr_err("Pasid 0x%x destroy queue %d failed, ret %d\n",
-				pqm->process->pasid,
+				pdd->pasid,
 				pqn->q->properties.queue_id, retval);
 			if (retval != -ETIME && retval != -EIO)
 				goto err_destroy_queue;
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_svm.c b/drivers/gpu/drm/amd/amdkfd/kfd_svm.c
index 47189453b20c3..116116a9f5781 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_svm.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_svm.c
@@ -563,7 +563,8 @@ svm_range_vram_node_new(struct kfd_node *node, struct svm_range *prange,
 	int r;
 
 	p = container_of(prange->svms, struct kfd_process, svms);
-	pr_debug("pasid: %x svms 0x%p [0x%lx 0x%lx]\n", p->pasid, prange->svms,
+	pr_debug("process pid: %d svms 0x%p [0x%lx 0x%lx]\n",
+		 p->lead_thread->pid, prange->svms,
 		 prange->start, prange->last);
 
 	if (svm_range_validate_svm_bo(node, prange))
@@ -2973,7 +2974,7 @@ svm_range_restore_pages(struct amdgpu_device *adev, unsigned int pasid,
 		return -EFAULT;
 	}
 
-	p = kfd_lookup_process_by_pasid(pasid);
+	p = kfd_lookup_process_by_pasid(pasid, NULL);
 	if (!p) {
 		pr_debug("kfd process not founded pasid 0x%x\n", pasid);
 		return 0;
@@ -3239,7 +3240,8 @@ void svm_range_list_fini(struct kfd_process *p)
 	struct svm_range *prange;
 	struct svm_range *next;
 
-	pr_debug("pasid 0x%x svms 0x%p\n", p->pasid, &p->svms);
+	pr_debug("process pid %d svms 0x%p\n", p->lead_thread->pid,
+		 &p->svms);
 
 	cancel_delayed_work_sync(&p->svms.restore_work);
 
@@ -3262,7 +3264,8 @@ void svm_range_list_fini(struct kfd_process *p)
 
 	mutex_destroy(&p->svms.lock);
 
-	pr_debug("pasid 0x%x svms 0x%p done\n", p->pasid, &p->svms);
+	pr_debug("process pid %d svms 0x%p done\n",
+		p->lead_thread->pid, &p->svms);
 }
 
 int svm_range_list_init(struct kfd_process *p)
@@ -3625,8 +3628,8 @@ svm_range_set_attr(struct kfd_process *p, struct mm_struct *mm,
 	bool flush_tlb;
 	int r, ret = 0;
 
-	pr_debug("pasid 0x%x svms 0x%p [0x%llx 0x%llx] pages 0x%llx\n",
-		 p->pasid, &p->svms, start, start + size - 1, size);
+	pr_debug("process pid %d svms 0x%p [0x%llx 0x%llx] pages 0x%llx\n",
+		 p->lead_thread->pid, &p->svms, start, start + size - 1, size);
 
 	r = svm_range_check_attr(p, nattr, attrs);
 	if (r)
@@ -3734,8 +3737,8 @@ svm_range_set_attr(struct kfd_process *p, struct mm_struct *mm,
 out:
 	mutex_unlock(&process_info->lock);
 
-	pr_debug("pasid 0x%x svms 0x%p [0x%llx 0x%llx] done, r=%d\n", p->pasid,
-		 &p->svms, start, start + size - 1, r);
+	pr_debug("process pid %d svms 0x%p [0x%llx 0x%llx] done, r=%d\n",
+		 p->lead_thread->pid, &p->svms, start, start + size - 1, r);
 
 	return ret ? ret : r;
 }
-- 
2.39.5

[PATCH AUTOSEL 6.14 483/642] drm/amd/display/dc: enable oem i2c support for DCE 12.x

[Sasha Levin]

From: Alex Deucher <alexander.deucher@amd.com>

[ Upstream commit 2ed83f2cc41e8f7ced1c0610ec2b0821c5522ed5 ]

Use the value pulled from the vbios just like newer chips.

Reviewed-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 .../dc/resource/dce120/dce120_resource.c        | 17 +++++++++++++++++
 1 file changed, 17 insertions(+)

diff --git a/drivers/gpu/drm/amd/display/dc/resource/dce120/dce120_resource.c b/drivers/gpu/drm/amd/display/dc/resource/dce120/dce120_resource.c
index c63c596234333..eb1e158d34361 100644
--- a/drivers/gpu/drm/amd/display/dc/resource/dce120/dce120_resource.c
+++ b/drivers/gpu/drm/amd/display/dc/resource/dce120/dce120_resource.c
@@ -67,6 +67,7 @@
 #include "reg_helper.h"
 
 #include "dce100/dce100_resource.h"
+#include "link.h"
 
 #ifndef mmDP0_DP_DPHY_INTERNAL_CTRL
 	#define mmDP0_DP_DPHY_INTERNAL_CTRL		0x210f
@@ -659,6 +660,12 @@ static void dce120_resource_destruct(struct dce110_resource_pool *pool)
 
 	if (pool->base.dmcu != NULL)
 		dce_dmcu_destroy(&pool->base.dmcu);
+
+	if (pool->base.oem_device != NULL) {
+		struct dc *dc = pool->base.oem_device->ctx->dc;
+
+		dc->link_srv->destroy_ddc_service(&pool->base.oem_device);
+	}
 }
 
 static void read_dce_straps(
@@ -1054,6 +1061,7 @@ static bool dce120_resource_construct(
 	struct dc *dc,
 	struct dce110_resource_pool *pool)
 {
+	struct ddc_service_init_data ddc_init_data = {0};
 	unsigned int i;
 	int j;
 	struct dc_context *ctx = dc->ctx;
@@ -1257,6 +1265,15 @@ static bool dce120_resource_construct(
 
 	bw_calcs_data_update_from_pplib(dc);
 
+	if (dc->ctx->dc_bios->fw_info.oem_i2c_present) {
+		ddc_init_data.ctx = dc->ctx;
+		ddc_init_data.link = NULL;
+		ddc_init_data.id.id = dc->ctx->dc_bios->fw_info.oem_i2c_obj_id;
+		ddc_init_data.id.enum_id = 0;
+		ddc_init_data.id.type = OBJECT_TYPE_GENERIC;
+		pool->base.oem_device = dc->link_srv->create_ddc_service(&ddc_init_data);
+	}
+
 	return true;
 
 irqs_create_fail:
-- 
2.39.5

[PATCH AUTOSEL 6.14 484/642] drm/amd/display/dm: drop hw_support check in amdgpu_dm_i2c_xfer()

[Sasha Levin]

From: Alex Deucher <alexander.deucher@amd.com>

[ Upstream commit 33da70bd1e115d7d73f45fb1c09f5ecc448f3f13 ]

DC supports SW i2c as well.  Drop the check.

Reviewed-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
index d51dbac3235ca..5f25fe7721c17 100644
--- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
@@ -8394,7 +8394,7 @@ static int amdgpu_dm_i2c_xfer(struct i2c_adapter *i2c_adap,
 	int i;
 	int result = -EIO;
 
-	if (!ddc_service->ddc_pin || !ddc_service->ddc_pin->hw_info.hw_supported)
+	if (!ddc_service->ddc_pin)
 		return result;
 
 	cmd.payloads = kcalloc(num, sizeof(struct i2c_payload), GFP_KERNEL);
-- 
2.39.5

[PATCH AUTOSEL 6.14 627/642] drm/amdkfd: Correct F8_MODE for gfx950

[Sasha Levin]

From: Amber Lin <Amber.Lin@amd.com>

[ Upstream commit 0c7e053448945e5a4379dc4396c762d7422b11ca ]

Correct F8_MODE setting for gfx950 that was removed

Fixes: 61972cd93af7 ("drm/amdkfd: Set per-process flags only once for gfx9/10/11/12")
Signed-off-by: Amber Lin <Amber.Lin@amd.com>
Reviewed-by: Harish Kasiviswanathan <Harish.Kasiviwanathan@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager_v9.c | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager_v9.c b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager_v9.c
index 3264509408bc8..d85eadaa1e11b 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager_v9.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager_v9.c
@@ -69,8 +69,7 @@ static bool set_cache_memory_policy_v9(struct device_queue_manager *dqm,
 		qpd->sh_mem_config |= 1 << SH_MEM_CONFIG__RETRY_DISABLE__SHIFT;
 
 	if (KFD_GC_VERSION(dqm->dev->kfd) == IP_VERSION(9, 4, 3) ||
-		KFD_GC_VERSION(dqm->dev->kfd) == IP_VERSION(9, 4, 4) ||
-		KFD_GC_VERSION(dqm->dev->kfd) == IP_VERSION(9, 5, 0))
+		KFD_GC_VERSION(dqm->dev->kfd) == IP_VERSION(9, 4, 4))
 		qpd->sh_mem_config |= (1 << SH_MEM_CONFIG__F8_MODE__SHIFT);
 
 	qpd->sh_mem_ape1_limit = 0;
-- 
2.39.5

[PATCH AUTOSEL 6.14 637/642] drm/amd/display: Exit idle optimizations before accessing PHY

[Sasha Levin]

From: Ovidiu Bunea <Ovidiu.Bunea@amd.com>

[ Upstream commit c488967488d7eff7b9c527d5469c424c15377502 ]

[why & how]
By default, DCN HW is in idle optimized state which does not allow access
to PHY registers. If BIOS powers up the DCN, it is fine because they will
power up everything. Only exit idle optimized state when not taking control
from VBIOS.

Fixes: be704e5ef4bd ("Revert "drm/amd/display: Exit idle optimizations before attempt to access PHY"")
Reviewed-by: Charlene Liu <charlene.liu@amd.com>
Signed-off-by: Ovidiu Bunea <Ovidiu.Bunea@amd.com>
Signed-off-by: Roman Li <roman.li@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/gpu/drm/amd/display/dc/hwss/dce110/dce110_hwseq.c | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/drivers/gpu/drm/amd/display/dc/hwss/dce110/dce110_hwseq.c b/drivers/gpu/drm/amd/display/dc/hwss/dce110/dce110_hwseq.c
index 2f5f3e749a1ab..94ceccfc04982 100644
--- a/drivers/gpu/drm/amd/display/dc/hwss/dce110/dce110_hwseq.c
+++ b/drivers/gpu/drm/amd/display/dc/hwss/dce110/dce110_hwseq.c
@@ -1889,6 +1889,7 @@ void dce110_enable_accelerated_mode(struct dc *dc, struct dc_state *context)
 	bool can_apply_edp_fast_boot = false;
 	bool can_apply_seamless_boot = false;
 	bool keep_edp_vdd_on = false;
+	struct dc_bios *dcb = dc->ctx->dc_bios;
 	DC_LOGGER_INIT();
 
 
@@ -1965,6 +1966,8 @@ void dce110_enable_accelerated_mode(struct dc *dc, struct dc_state *context)
 			hws->funcs.edp_backlight_control(edp_link_with_sink, false);
 		}
 		/*resume from S3, no vbios posting, no need to power down again*/
+		if (dcb && dcb->funcs && !dcb->funcs->is_accelerated_mode(dcb))
+			clk_mgr_exit_optimized_pwr_state(dc, dc->clk_mgr);
 
 		power_down_all_hw_blocks(dc);
 
@@ -1977,6 +1980,8 @@ void dce110_enable_accelerated_mode(struct dc *dc, struct dc_state *context)
 		disable_vga_and_power_gate_all_controllers(dc);
 		if (edp_link_with_sink && !keep_edp_vdd_on)
 			dc->hwss.edp_power_control(edp_link_with_sink, false);
+		if (dcb && dcb->funcs && !dcb->funcs->is_accelerated_mode(dcb))
+			clk_mgr_optimize_pwr_state(dc, dc->clk_mgr);
 	}
 	bios_set_scratch_acc_mode_change(dc->ctx->dc_bios, 1);
 }
-- 
2.39.5

[PATCH AUTOSEL 6.14 638/642] drm/amdkfd: Fix error handling for missing PASID in 'kfd_process_device_init_vm'

[Sasha Levin]

From: Srinivasan Shanmugam <srinivasan.shanmugam@amd.com>

[ Upstream commit 2b04d04de956b44cc140d45cf8ebccfb378ce3bf ]

In the kfd_process_device_init_vm function, a valid error code is now
returned when the associated Process Address Space ID (PASID) is not
present.

If the address space virtual memory (avm) does not have an associated
PASID, the function sets the ret variable to -EINVAL before proceeding
to the error handling section. This ensures that the calling function,
such as kfd_ioctl_acquire_vm, can appropriately handle the error,
thereby preventing any issues during virtual memory initialization.

Fixes the below:
drivers/gpu/drm/amd/amdgpu/../amdkfd/kfd_process.c:1694 kfd_process_device_init_vm()
warn: missing error code 'ret'

drivers/gpu/drm/amd/amdgpu/../amdkfd/kfd_process.c
    1647 int kfd_process_device_init_vm(struct kfd_process_device *pdd,
    1648                                struct file *drm_file)
    1649 {
    ...
    1690
    1691         if (unlikely(!avm->pasid)) {
    1692                 dev_warn(pdd->dev->adev->dev, "WARN: vm %p has no pasid associated",
    1693                                  avm);
--> 1694                 goto err_get_pasid;

ret = -EINVAL?

    1695         }

Fixes: 8544374c0f82 ("drm/amdkfd: Have kfd driver use same PASID values from graphic driver")
Reported by: Dan Carpenter <dan.carpenter@linaro.org>
Cc: Xiaogang Chen <xiaogang.chen@amd.com>
Cc: Felix Kuehling <felix.kuehling@amd.com>
Cc: Christian König <christian.koenig@amd.com>
Cc: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Srinivasan Shanmugam <srinivasan.shanmugam@amd.com>
Reviewed-by: Felix Kuehling <felix.kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/gpu/drm/amd/amdkfd/kfd_process.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_process.c b/drivers/gpu/drm/amd/amdkfd/kfd_process.c
index a7e0a16dac47b..3f411922534b3 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_process.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_process.c
@@ -1711,6 +1711,7 @@ int kfd_process_device_init_vm(struct kfd_process_device *pdd,
 	if (unlikely(!avm->pasid)) {
 		dev_warn(pdd->dev->adev->dev, "WARN: vm %p has no pasid associated",
 				 avm);
+		ret = -EINVAL;
 		goto err_get_pasid;
 	}
 
-- 
2.39.5

[PATCH AUTOSEL 6.14 639/642] drm/amdkfd: Fix pasid value leak

[Sasha Levin]

From: Xiaogang Chen <xiaogang.chen@amd.com>

[ Upstream commit 10e08943caedfb4b0b95933d248503a6f6b9fef6 ]

Curret kfd does not allocate pasid values, instead uses pasid value for each
vm from graphic driver. So should not prevent graphic driver from releasing
pasid values since the values are allocated by graphic driver, not kfd driver
anymore. This patch does not stop graphic driver release pasid values.

Fixes: 8544374c0f82 ("drm/amdkfd: Have kfd driver use same PASID values from graphic driver")
Signed-off-by: Xiaogang Chen <xiaogang.chen@amd.com>
Reviewed-by: Felix Kuehling <felix.kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h    |  2 --
 .../gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c  | 21 -------------------
 drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c        | 14 -------------
 drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h        |  1 -
 drivers/gpu/drm/amd/amdkfd/kfd_process.c      |  5 +----
 5 files changed, 1 insertion(+), 42 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h
index a160ffe80b3de..55d5399676951 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h
@@ -303,8 +303,6 @@ int amdgpu_amdkfd_gpuvm_acquire_process_vm(struct amdgpu_device *adev,
 					struct amdgpu_vm *avm,
 					void **process_info,
 					struct dma_fence **ef);
-void amdgpu_amdkfd_gpuvm_release_process_vm(struct amdgpu_device *adev,
-					void *drm_priv);
 uint64_t amdgpu_amdkfd_gpuvm_get_process_page_dir(void *drm_priv);
 size_t amdgpu_amdkfd_get_available_memory(struct amdgpu_device *adev,
 					uint8_t xcp_id);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c
index 0c8129f8d3635..b3c8eae460425 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c
@@ -1578,27 +1578,6 @@ void amdgpu_amdkfd_gpuvm_destroy_cb(struct amdgpu_device *adev,
 	}
 }
 
-void amdgpu_amdkfd_gpuvm_release_process_vm(struct amdgpu_device *adev,
-					    void *drm_priv)
-{
-	struct amdgpu_vm *avm;
-
-	if (WARN_ON(!adev || !drm_priv))
-		return;
-
-	avm = drm_priv_to_vm(drm_priv);
-
-	pr_debug("Releasing process vm %p\n", avm);
-
-	/* The original pasid of amdgpu vm has already been
-	 * released during making a amdgpu vm to a compute vm
-	 * The current pasid is managed by kfd and will be
-	 * released on kfd process destroy. Set amdgpu pasid
-	 * to 0 to avoid duplicate release.
-	 */
-	amdgpu_vm_release_compute(adev, avm);
-}
-
 uint64_t amdgpu_amdkfd_gpuvm_get_process_page_dir(void *drm_priv)
 {
 	struct amdgpu_vm *avm = drm_priv_to_vm(drm_priv);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
index f0d675c0fc69c..21be10d46cf9c 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
@@ -2687,20 +2687,6 @@ int amdgpu_vm_make_compute(struct amdgpu_device *adev, struct amdgpu_vm *vm)
 	return r;
 }
 
-/**
- * amdgpu_vm_release_compute - release a compute vm
- * @adev: amdgpu_device pointer
- * @vm: a vm turned into compute vm by calling amdgpu_vm_make_compute
- *
- * This is a correspondant of amdgpu_vm_make_compute. It decouples compute
- * pasid from vm. Compute should stop use of vm after this call.
- */
-void amdgpu_vm_release_compute(struct amdgpu_device *adev, struct amdgpu_vm *vm)
-{
-	amdgpu_vm_set_pasid(adev, vm, 0);
-	vm->is_compute_context = false;
-}
-
 static int amdgpu_vm_stats_is_zero(struct amdgpu_vm *vm)
 {
 	for (int i = 0; i < __AMDGPU_PL_NUM; ++i) {
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h
index 5010a3107bf89..f3ad687125ad6 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h
@@ -489,7 +489,6 @@ int amdgpu_vm_set_pasid(struct amdgpu_device *adev, struct amdgpu_vm *vm,
 long amdgpu_vm_wait_idle(struct amdgpu_vm *vm, long timeout);
 int amdgpu_vm_init(struct amdgpu_device *adev, struct amdgpu_vm *vm, int32_t xcp_id);
 int amdgpu_vm_make_compute(struct amdgpu_device *adev, struct amdgpu_vm *vm);
-void amdgpu_vm_release_compute(struct amdgpu_device *adev, struct amdgpu_vm *vm);
 void amdgpu_vm_fini(struct amdgpu_device *adev, struct amdgpu_vm *vm);
 int amdgpu_vm_lock_pd(struct amdgpu_vm *vm, struct drm_exec *exec,
 		      unsigned int num_fences);
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_process.c b/drivers/gpu/drm/amd/amdkfd/kfd_process.c
index 3f411922534b3..7c0c24732481e 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_process.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_process.c
@@ -1059,11 +1059,8 @@ static void kfd_process_destroy_pdds(struct kfd_process *p)
 		kfd_process_device_destroy_cwsr_dgpu(pdd);
 		kfd_process_device_destroy_ib_mem(pdd);
 
-		if (pdd->drm_file) {
-			amdgpu_amdkfd_gpuvm_release_process_vm(
-					pdd->dev->adev, pdd->drm_priv);
+		if (pdd->drm_file)
 			fput(pdd->drm_file);
-		}
 
 		if (pdd->qpd.cwsr_kaddr && !pdd->qpd.cwsr_base)
 			free_pages((unsigned long)pdd->qpd.cwsr_kaddr,
-- 
2.39.5

Re: [PATCH AUTOSEL 6.14 483/642] drm/amd/display/dc: enable oem i2c support for DCE 12.x

[Alex Deucher]

On Mon, May 5, 2025 at 6:34 PM Sasha Levin <sashal@kernel.org> wrote:
>
> From: Alex Deucher <alexander.deucher@amd.com>
>
> [ Upstream commit 2ed83f2cc41e8f7ced1c0610ec2b0821c5522ed5 ]
>
> Use the value pulled from the vbios just like newer chips.
>
> Reviewed-by: Harry Wentland <harry.wentland@amd.com>
> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
> Signed-off-by: Sasha Levin <sashal@kernel.org>

This isn't a bug fix.  This is a new feature and depends on other
changes to actually utilize.

Alex

> ---
>  .../dc/resource/dce120/dce120_resource.c        | 17 +++++++++++++++++
>  1 file changed, 17 insertions(+)
>
> diff --git a/drivers/gpu/drm/amd/display/dc/resource/dce120/dce120_resource.c b/drivers/gpu/drm/amd/display/dc/resource/dce120/dce120_resource.c
> index c63c596234333..eb1e158d34361 100644
> --- a/drivers/gpu/drm/amd/display/dc/resource/dce120/dce120_resource.c
> +++ b/drivers/gpu/drm/amd/display/dc/resource/dce120/dce120_resource.c
> @@ -67,6 +67,7 @@
>  #include "reg_helper.h"
>
>  #include "dce100/dce100_resource.h"
> +#include "link.h"
>
>  #ifndef mmDP0_DP_DPHY_INTERNAL_CTRL
>         #define mmDP0_DP_DPHY_INTERNAL_CTRL             0x210f
> @@ -659,6 +660,12 @@ static void dce120_resource_destruct(struct dce110_resource_pool *pool)
>
>         if (pool->base.dmcu != NULL)
>                 dce_dmcu_destroy(&pool->base.dmcu);
> +
> +       if (pool->base.oem_device != NULL) {
> +               struct dc *dc = pool->base.oem_device->ctx->dc;
> +
> +               dc->link_srv->destroy_ddc_service(&pool->base.oem_device);
> +       }
>  }
>
>  static void read_dce_straps(
> @@ -1054,6 +1061,7 @@ static bool dce120_resource_construct(
>         struct dc *dc,
>         struct dce110_resource_pool *pool)
>  {
> +       struct ddc_service_init_data ddc_init_data = {0};
>         unsigned int i;
>         int j;
>         struct dc_context *ctx = dc->ctx;
> @@ -1257,6 +1265,15 @@ static bool dce120_resource_construct(
>
>         bw_calcs_data_update_from_pplib(dc);
>
> +       if (dc->ctx->dc_bios->fw_info.oem_i2c_present) {
> +               ddc_init_data.ctx = dc->ctx;
> +               ddc_init_data.link = NULL;
> +               ddc_init_data.id.id = dc->ctx->dc_bios->fw_info.oem_i2c_obj_id;
> +               ddc_init_data.id.enum_id = 0;
> +               ddc_init_data.id.type = OBJECT_TYPE_GENERIC;
> +               pool->base.oem_device = dc->link_srv->create_ddc_service(&ddc_init_data);
> +       }
> +
>         return true;
>
>  irqs_create_fail:
> --
> 2.39.5
>

Re: [PATCH AUTOSEL 6.14 248/642] drm/amdgpu: add dce_v6_0_soft_reset() to DCE6

[Alex Deucher]

On Mon, May 5, 2025 at 6:24 PM Sasha Levin <sashal@kernel.org> wrote:
>
> From: Alexandre Demers <alexandre.f.demers@gmail.com>
>
> [ Upstream commit ab23db6d08efdda5d13d01a66c593d0e57f8917f ]
>
> DCE6 was missing soft reset, but it was easily identifiable under radeon.
> This should be it, pretty much as it is done under DCE8 and DCE10.
>
> Signed-off-by: Alexandre Demers <alexandre.f.demers@gmail.com>
> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
> Signed-off-by: Sasha Levin <sashal@kernel.org>

This is not stable material.

Alex

> ---
>  drivers/gpu/drm/amd/amdgpu/dce_v6_0.c | 53 ++++++++++++++++++++++++++-
>  1 file changed, 51 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/gpu/drm/amd/amdgpu/dce_v6_0.c b/drivers/gpu/drm/amd/amdgpu/dce_v6_0.c
> index 915804a6a1d7d..ed5e06b677df1 100644
> --- a/drivers/gpu/drm/amd/amdgpu/dce_v6_0.c
> +++ b/drivers/gpu/drm/amd/amdgpu/dce_v6_0.c
> @@ -370,13 +370,41 @@ static u32 dce_v6_0_hpd_get_gpio_reg(struct amdgpu_device *adev)
>         return mmDC_GPIO_HPD_A;
>  }
>
> +static bool dce_v6_0_is_display_hung(struct amdgpu_device *adev)
> +{
> +       u32 crtc_hung = 0;
> +       u32 crtc_status[6];
> +       u32 i, j, tmp;
> +
> +       for (i = 0; i < adev->mode_info.num_crtc; i++) {
> +               if (RREG32(mmCRTC_CONTROL + crtc_offsets[i]) & CRTC_CONTROL__CRTC_MASTER_EN_MASK) {
> +                       crtc_status[i] = RREG32(mmCRTC_STATUS_HV_COUNT + crtc_offsets[i]);
> +                       crtc_hung |= (1 << i);
> +               }
> +       }
> +
> +       for (j = 0; j < 10; j++) {
> +               for (i = 0; i < adev->mode_info.num_crtc; i++) {
> +                       if (crtc_hung & (1 << i)) {
> +                               tmp = RREG32(mmCRTC_STATUS_HV_COUNT + crtc_offsets[i]);
> +                               if (tmp != crtc_status[i])
> +                                       crtc_hung &= ~(1 << i);
> +                       }
> +               }
> +               if (crtc_hung == 0)
> +                       return false;
> +               udelay(100);
> +       }
> +
> +       return true;
> +}
> +
>  static void dce_v6_0_set_vga_render_state(struct amdgpu_device *adev,
>                                           bool render)
>  {
>         if (!render)
>                 WREG32(mmVGA_RENDER_CONTROL,
>                         RREG32(mmVGA_RENDER_CONTROL) & VGA_VSTATUS_CNTL);
> -
>  }
>
>  static int dce_v6_0_get_num_crtc(struct amdgpu_device *adev)
> @@ -2872,7 +2900,28 @@ static bool dce_v6_0_is_idle(void *handle)
>
>  static int dce_v6_0_soft_reset(struct amdgpu_ip_block *ip_block)
>  {
> -       DRM_INFO("xxxx: dce_v6_0_soft_reset --- no impl!!\n");
> +       u32 srbm_soft_reset = 0, tmp;
> +       struct amdgpu_device *adev = ip_block->adev;
> +
> +       if (dce_v6_0_is_display_hung(adev))
> +               srbm_soft_reset |= SRBM_SOFT_RESET__SOFT_RESET_DC_MASK;
> +
> +       if (srbm_soft_reset) {
> +               tmp = RREG32(mmSRBM_SOFT_RESET);
> +               tmp |= srbm_soft_reset;
> +               dev_info(adev->dev, "SRBM_SOFT_RESET=0x%08X\n", tmp);
> +               WREG32(mmSRBM_SOFT_RESET, tmp);
> +               tmp = RREG32(mmSRBM_SOFT_RESET);
> +
> +               udelay(50);
> +
> +               tmp &= ~srbm_soft_reset;
> +               WREG32(mmSRBM_SOFT_RESET, tmp);
> +               tmp = RREG32(mmSRBM_SOFT_RESET);
> +
> +               /* Wait a little for things to settle down */
> +               udelay(50);
> +       }
>         return 0;
>  }
>
> --
> 2.39.5
>

Re: [PATCH AUTOSEL 6.14 248/642] drm/amdgpu: add dce_v6_0_soft_reset() to DCE6

[Sasha Levin]

On Tue, May 06, 2025 at 11:05:15AM -0400, Alex Deucher wrote:
>On Mon, May 5, 2025 at 6:24 PM Sasha Levin <sashal@kernel.org> wrote:
>>
>> From: Alexandre Demers <alexandre.f.demers@gmail.com>
>>
>> [ Upstream commit ab23db6d08efdda5d13d01a66c593d0e57f8917f ]
>>
>> DCE6 was missing soft reset, but it was easily identifiable under radeon.
>> This should be it, pretty much as it is done under DCE8 and DCE10.
>>
>> Signed-off-by: Alexandre Demers <alexandre.f.demers@gmail.com>
>> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
>> Signed-off-by: Sasha Levin <sashal@kernel.org>
>
>This is not stable material.

Dropped, thanks!

-- 
Thanks,
Sasha

Re: [PATCH AUTOSEL 6.14 107/642] drm/amdgpu: adjust drm_firmware_drivers_only() handling

[Thomas Zimmermann]

Hi

Am 06.05.25 um 00:05 schrieb Sasha Levin:
> From: Alex Deucher <alexander.deucher@amd.com>
>
> [ Upstream commit e00e5c223878a60e391e5422d173c3382d378f87 ]
>
> Move to probe so we can check the PCI device type and
> only apply the drm_firmware_drivers_only() check for
> PCI DISPLAY classes.  Also add a module parameter to
> override the nomodeset kernel parameter as a workaround
> for platforms that have this hardcoded on their kernel
> command lines.

I just came across this patch because it got backported into various 
older releases. It was part of the series at [1]. From the cover letter:

 >>>

There are a number of systems and cloud providers out there
that have nomodeset hardcoded in their kernel parameters
to block nouveau for the nvidia driver.  This prevents the
amdgpu driver from loading. Unfortunately the end user cannot
easily change this.  The preferred way to block modules from
loading is to use modprobe.blacklist=<driver>.  That is what
providers should be using to block specific drivers.

Drop the check to allow the driver to load even when nomodeset
is specified on the kernel command line.

<<<

Why was that series never on dri-devel?

Why is this necessary in the upstream kernel? It works around a problem 
with the user's configuration. The series' cover letter already states 
the correct solution.

Firmware-only parameters affect all drivers; why not try for a common 
solution? At least the test against the PCI class appears useful in the 
common case.

Best regards
Thomas


>
> Reviewed-by: Kent Russell <kent.russell@amd.com>
> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
> Signed-off-by: Sasha Levin <sashal@kernel.org>
> ---
>   drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c | 14 ++++++++++++++
>   1 file changed, 14 insertions(+)
>
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
> index f2d77bc04e4a9..7246c54bd2bbf 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
> @@ -173,6 +173,7 @@ uint amdgpu_sdma_phase_quantum = 32;
>   char *amdgpu_disable_cu;
>   char *amdgpu_virtual_display;
>   bool enforce_isolation;
> +int amdgpu_modeset = -1;
>   
>   /* Specifies the default granularity for SVM, used in buffer
>    * migration and restoration of backing memory when handling
> @@ -1033,6 +1034,13 @@ module_param_named(user_partt_mode, amdgpu_user_partt_mode, uint, 0444);
>   module_param(enforce_isolation, bool, 0444);
>   MODULE_PARM_DESC(enforce_isolation, "enforce process isolation between graphics and compute . enforce_isolation = on");
>   
> +/**
> + * DOC: modeset (int)
> + * Override nomodeset (1 = override, -1 = auto). The default is -1 (auto).
> + */
> +MODULE_PARM_DESC(modeset, "Override nomodeset (1 = enable, -1 = auto)");
> +module_param_named(modeset, amdgpu_modeset, int, 0444);
> +
>   /**
>    * DOC: seamless (int)
>    * Seamless boot will keep the image on the screen during the boot process.
> @@ -2244,6 +2252,12 @@ static int amdgpu_pci_probe(struct pci_dev *pdev,
>   	int ret, retry = 0, i;
>   	bool supports_atomic = false;
>   
> +	if ((pdev->class >> 8) == PCI_CLASS_DISPLAY_VGA ||
> +	    (pdev->class >> 8) == PCI_CLASS_DISPLAY_OTHER) {
> +		if (drm_firmware_drivers_only() && amdgpu_modeset == -1)
> +			return -EINVAL;
> +	}
> +
>   	/* skip devices which are owned by radeon */
>   	for (i = 0; i < ARRAY_SIZE(amdgpu_unsupported_pciidlist); i++) {
>   		if (amdgpu_unsupported_pciidlist[i] == pdev->device)

-- 
--
Thomas Zimmermann
Graphics Driver Developer
SUSE Software Solutions Germany GmbH
Frankenstrasse 146, 90461 Nuernberg, Germany
GF: Ivo Totev, Andrew Myers, Andrew McDonald, Boudien Moerman
HRB 36809 (AG Nuernberg)

Re: [PATCH AUTOSEL 6.14 107/642] drm/amdgpu: adjust drm_firmware_drivers_only() handling

[Thomas Zimmermann]

Am 21.08.25 um 11:32 schrieb Thomas Zimmermann:
> Hi
>
> Am 06.05.25 um 00:05 schrieb Sasha Levin:
>> From: Alex Deucher <alexander.deucher@amd.com>
>>
>> [ Upstream commit e00e5c223878a60e391e5422d173c3382d378f87 ]
>>
>> Move to probe so we can check the PCI device type and
>> only apply the drm_firmware_drivers_only() check for
>> PCI DISPLAY classes.  Also add a module parameter to
>> override the nomodeset kernel parameter as a workaround
>> for platforms that have this hardcoded on their kernel
>> command lines.
>
> I just came across this patch because it got backported into various 
> older releases. It was part of the series at [1]. From the cover letter:

[1] 
https://lore.kernel.org/all/20250314010152.1503510-1-alexander.deucher@amd.com/

>
> >>>
>
> There are a number of systems and cloud providers out there
> that have nomodeset hardcoded in their kernel parameters
> to block nouveau for the nvidia driver.  This prevents the
> amdgpu driver from loading. Unfortunately the end user cannot
> easily change this.  The preferred way to block modules from
> loading is to use modprobe.blacklist=<driver>.  That is what
> providers should be using to block specific drivers.
>
> Drop the check to allow the driver to load even when nomodeset
> is specified on the kernel command line.
>
> <<<
>
> Why was that series never on dri-devel?
>
> Why is this necessary in the upstream kernel? It works around a 
> problem with the user's configuration. The series' cover letter 
> already states the correct solution.
>
> Firmware-only parameters affect all drivers; why not try for a common 
> solution? At least the test against the PCI class appears useful in 
> the common case.
>
> Best regards
> Thomas
>
>
>>
>> Reviewed-by: Kent Russell <kent.russell@amd.com>
>> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
>> Signed-off-by: Sasha Levin <sashal@kernel.org>
>> ---
>>   drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c | 14 ++++++++++++++
>>   1 file changed, 14 insertions(+)
>>
>> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c 
>> b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
>> index f2d77bc04e4a9..7246c54bd2bbf 100644
>> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
>> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
>> @@ -173,6 +173,7 @@ uint amdgpu_sdma_phase_quantum = 32;
>>   char *amdgpu_disable_cu;
>>   char *amdgpu_virtual_display;
>>   bool enforce_isolation;
>> +int amdgpu_modeset = -1;
>>     /* Specifies the default granularity for SVM, used in buffer
>>    * migration and restoration of backing memory when handling
>> @@ -1033,6 +1034,13 @@ module_param_named(user_partt_mode, 
>> amdgpu_user_partt_mode, uint, 0444);
>>   module_param(enforce_isolation, bool, 0444);
>>   MODULE_PARM_DESC(enforce_isolation, "enforce process isolation 
>> between graphics and compute . enforce_isolation = on");
>>   +/**
>> + * DOC: modeset (int)
>> + * Override nomodeset (1 = override, -1 = auto). The default is -1 
>> (auto).
>> + */
>> +MODULE_PARM_DESC(modeset, "Override nomodeset (1 = enable, -1 = 
>> auto)");
>> +module_param_named(modeset, amdgpu_modeset, int, 0444);
>> +
>>   /**
>>    * DOC: seamless (int)
>>    * Seamless boot will keep the image on the screen during the boot 
>> process.
>> @@ -2244,6 +2252,12 @@ static int amdgpu_pci_probe(struct pci_dev *pdev,
>>       int ret, retry = 0, i;
>>       bool supports_atomic = false;
>>   +    if ((pdev->class >> 8) == PCI_CLASS_DISPLAY_VGA ||
>> +        (pdev->class >> 8) == PCI_CLASS_DISPLAY_OTHER) {
>> +        if (drm_firmware_drivers_only() && amdgpu_modeset == -1)
>> +            return -EINVAL;
>> +    }
>> +
>>       /* skip devices which are owned by radeon */
>>       for (i = 0; i < ARRAY_SIZE(amdgpu_unsupported_pciidlist); i++) {
>>           if (amdgpu_unsupported_pciidlist[i] == pdev->device)
>

-- 
--
Thomas Zimmermann
Graphics Driver Developer
SUSE Software Solutions Germany GmbH
Frankenstrasse 146, 90461 Nuernberg, Germany
GF: Ivo Totev, Andrew Myers, Andrew McDonald, Boudien Moerman
HRB 36809 (AG Nuernberg)

Re: [PATCH AUTOSEL 6.14 107/642] drm/amdgpu: adjust drm_firmware_drivers_only() handling

[Alex Deucher]

On Thu, Aug 21, 2025 at 5:33 AM Thomas Zimmermann <tzimmermann@suse.de> wrote:
>
> Hi
>
> Am 06.05.25 um 00:05 schrieb Sasha Levin:
> > From: Alex Deucher <alexander.deucher@amd.com>
> >
> > [ Upstream commit e00e5c223878a60e391e5422d173c3382d378f87 ]
> >
> > Move to probe so we can check the PCI device type and
> > only apply the drm_firmware_drivers_only() check for
> > PCI DISPLAY classes.  Also add a module parameter to
> > override the nomodeset kernel parameter as a workaround
> > for platforms that have this hardcoded on their kernel
> > command lines.
>
> I just came across this patch because it got backported into various
> older releases. It was part of the series at [1]. From the cover letter:
>
>  >>>
>
> There are a number of systems and cloud providers out there
> that have nomodeset hardcoded in their kernel parameters
> to block nouveau for the nvidia driver.  This prevents the
> amdgpu driver from loading. Unfortunately the end user cannot
> easily change this.  The preferred way to block modules from
> loading is to use modprobe.blacklist=<driver>.  That is what
> providers should be using to block specific drivers.
>
> Drop the check to allow the driver to load even when nomodeset
> is specified on the kernel command line.
>
> <<<
>
> Why was that series never on dri-devel?

I guess I should have sent these to dri-devel as well.

>
> Why is this necessary in the upstream kernel? It works around a problem
> with the user's configuration. The series' cover letter already states
> the correct solution.

IIRC, the customers were not willing to change their kernel
configurations across their fleet, but required a way to load the
amdgpu driver, but keep nouveau blocked.  That said, doing this in the
core would also not solve the problem since the goal of nomodeset was
to block nouveau.

>
> Firmware-only parameters affect all drivers; why not try for a common
> solution? At least the test against the PCI class appears useful in the
> common case.

I can port the changes to the core if there is interest.

Alex

>
> Best regards
> Thomas
>
>
> >
> > Reviewed-by: Kent Russell <kent.russell@amd.com>
> > Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
> > Signed-off-by: Sasha Levin <sashal@kernel.org>
> > ---
> >   drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c | 14 ++++++++++++++
> >   1 file changed, 14 insertions(+)
> >
> > diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
> > index f2d77bc04e4a9..7246c54bd2bbf 100644
> > --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
> > +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
> > @@ -173,6 +173,7 @@ uint amdgpu_sdma_phase_quantum = 32;
> >   char *amdgpu_disable_cu;
> >   char *amdgpu_virtual_display;
> >   bool enforce_isolation;
> > +int amdgpu_modeset = -1;
> >
> >   /* Specifies the default granularity for SVM, used in buffer
> >    * migration and restoration of backing memory when handling
> > @@ -1033,6 +1034,13 @@ module_param_named(user_partt_mode, amdgpu_user_partt_mode, uint, 0444);
> >   module_param(enforce_isolation, bool, 0444);
> >   MODULE_PARM_DESC(enforce_isolation, "enforce process isolation between graphics and compute . enforce_isolation = on");
> >
> > +/**
> > + * DOC: modeset (int)
> > + * Override nomodeset (1 = override, -1 = auto). The default is -1 (auto).
> > + */
> > +MODULE_PARM_DESC(modeset, "Override nomodeset (1 = enable, -1 = auto)");
> > +module_param_named(modeset, amdgpu_modeset, int, 0444);
> > +
> >   /**
> >    * DOC: seamless (int)
> >    * Seamless boot will keep the image on the screen during the boot process.
> > @@ -2244,6 +2252,12 @@ static int amdgpu_pci_probe(struct pci_dev *pdev,
> >       int ret, retry = 0, i;
> >       bool supports_atomic = false;
> >
> > +     if ((pdev->class >> 8) == PCI_CLASS_DISPLAY_VGA ||
> > +         (pdev->class >> 8) == PCI_CLASS_DISPLAY_OTHER) {
> > +             if (drm_firmware_drivers_only() && amdgpu_modeset == -1)
> > +                     return -EINVAL;
> > +     }
> > +
> >       /* skip devices which are owned by radeon */
> >       for (i = 0; i < ARRAY_SIZE(amdgpu_unsupported_pciidlist); i++) {
> >               if (amdgpu_unsupported_pciidlist[i] == pdev->device)
>
> --
> --
> Thomas Zimmermann
> Graphics Driver Developer
> SUSE Software Solutions Germany GmbH
> Frankenstrasse 146, 90461 Nuernberg, Germany
> GF: Ivo Totev, Andrew Myers, Andrew McDonald, Boudien Moerman
> HRB 36809 (AG Nuernberg)
>
>