drm/amdgpu: convert ptl_hw_supported to enum

Convert ptl_hw_supported to enum with three states:
- AMDGPU_PTL_HW_UNINIT: not yet initialized
- AMDGPU_PTL_HW_SUPPORTED: initialized and supported
- AMDGPU_PTL_HW_NOT_SUPPORTED: initialized and not supported

This allows skipping PTL initialization attempts when hardware is
known to not support it, avoiding repeated initialization failures
after GPU resets.

v2:move ptl_hw_supported_state to AMDGPU_PTL_HW_NOT_SUPPORTED regardless
of error code during first time initialization. Print init fail log when
error code is not EOPNOTSUPP.

Signed-off-by: Victor Zhao <Victor.Zhao@amd.com>
Reviewed-by: Lijo Lazar <lijo.lazar@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

4 files changed, 33 insertions, 12 deletions

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c
index e03b94dc111d..04f6ebf31cca 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c
@@ -1613,7 +1613,7 @@ static umode_t amdgpu_ptl_is_visible(struct kobject *kobj, struct attribute *att
 	struct amdgpu_device *adev = drm_to_adev(ddev);
 
 	/* Only show PTL sysfs files if PTL hardware is supported */
-	if (!adev->psp.ptl.hw_supported)
+	if (adev->psp.ptl.hw_supported_state != AMDGPU_PTL_HW_SUPPORTED)
 		return 0;
 
 	return attr->mode;
@@ -1624,7 +1624,7 @@ int amdgpu_ptl_sysfs_init(struct amdgpu_device *adev)
 	struct amdgpu_ptl *ptl = &adev->psp.ptl;
 	int ret;
 
-	if (!ptl->hw_supported)
+	if (ptl->hw_supported_state != AMDGPU_PTL_HW_SUPPORTED)
 		return 0;
 
 	if (ptl->ptl_sysfs_created)
@@ -1641,7 +1641,7 @@ void amdgpu_ptl_sysfs_fini(struct amdgpu_device *adev)
 {
 	struct amdgpu_ptl *ptl = &adev->psp.ptl;
 
-	if (!ptl->hw_supported)
+	if (ptl->hw_supported_state != AMDGPU_PTL_HW_SUPPORTED)
 		return;
 
 	if (!ptl->ptl_sysfs_created)
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_3.c b/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_3.c
index 9512fef81d84..b086fabb8e1e 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_3.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_3.c
@@ -2400,7 +2400,10 @@ static int gfx_v9_4_3_perf_monitor_ptl_init(struct amdgpu_device *adev, bool ena
 	if (!adev->psp.funcs)
 		return -EOPNOTSUPP;
 
-	if (!ptl->hw_supported) {
+	if (ptl->hw_supported_state == AMDGPU_PTL_HW_NOT_SUPPORTED)
+		return -EOPNOTSUPP;
+
+	if (ptl->hw_supported_state == AMDGPU_PTL_HW_UNINIT) {
 		fmt1 = GFX_FTYPE_VECTOR;
 		fmt2 = GFX_FTYPE_F8;
 	} else {
@@ -2411,10 +2414,17 @@ static int gfx_v9_4_3_perf_monitor_ptl_init(struct amdgpu_device *adev, bool ena
 	/* initialize PTL with default formats: GFX_FTYPE_VECTOR & GFX_FTYPE_F8 */
 	r = amdgpu_ptl_perf_monitor_ctrl(adev, PSP_PTL_PERF_MON_SET, &ptl_state,
 							&fmt1, &fmt2);
-	if (r)
+	if (r) {
+		if (ptl->hw_supported_state == AMDGPU_PTL_HW_UNINIT)
+			ptl->hw_supported_state = AMDGPU_PTL_HW_NOT_SUPPORTED;
+
+		if (r != -EOPNOTSUPP)
+			dev_err(adev->dev, "PTL initialization failed (%d)\n", r);
+
 		return r;
+	}
 
-	ptl->hw_supported = true;
+	ptl->hw_supported_state = AMDGPU_PTL_HW_SUPPORTED;
 
 	atomic_set(&ptl->disable_ref, 0);
 	if (!enable && !amdgpu_in_reset(adev) && !adev->in_suspend) {
@@ -2459,7 +2469,8 @@ static int gfx_v9_4_3_hw_fini(struct amdgpu_ip_block *ip_block)
 	struct amdgpu_device *adev = ip_block->adev;
 	int i, num_xcc;
 
-	if (adev->psp.ptl.hw_supported && !amdgpu_in_reset(adev))
+	if (adev->psp.ptl.hw_supported_state == AMDGPU_PTL_HW_SUPPORTED &&
+			!amdgpu_in_reset(adev))
 		gfx_v9_4_3_perf_monitor_ptl_init(adev, false);
 
 	amdgpu_irq_put(adev, &adev->gfx.bad_op_irq, 0);
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c b/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c
index 6b681d948c04..dd7a7a18466d 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_chardev.c
@@ -1783,6 +1783,9 @@ static int kfd_ptl_control(struct kfd_process_device *pdd, bool enable)
 	uint32_t ptl_state = enable ? 1 : 0;
 	int ret;
 
+	if (ptl->hw_supported_state != AMDGPU_PTL_HW_SUPPORTED)
+		return -EOPNOTSUPP;
+
 	if (!pdd->dev->kfd2kgd || !pdd->dev->kfd2kgd->ptl_ctrl)
 		return -EOPNOTSUPP;
 
@@ -1801,7 +1804,7 @@ int kfd_ptl_disable_request(struct kfd_process_device *pdd,
 	struct amdgpu_ptl *ptl = &adev->psp.ptl;
 	int ret = 0;
 
-	if (!ptl->hw_supported)
+	if (ptl->hw_supported_state != AMDGPU_PTL_HW_SUPPORTED)
 		return -EOPNOTSUPP;
 
 	mutex_lock(&ptl->mutex);
@@ -1833,7 +1836,7 @@ int kfd_ptl_disable_release(struct kfd_process_device *pdd,
 	struct amdgpu_ptl *ptl = &adev->psp.ptl;
 	int ret = 0;
 
-	if (!ptl->hw_supported)
+	if (ptl->hw_supported_state != AMDGPU_PTL_HW_SUPPORTED)
 		return -EOPNOTSUPP;
 
 	mutex_lock(&ptl->mutex);
@@ -3353,7 +3356,7 @@ static inline uint32_t profile_lock_device(struct kfd_process *p,
 			kfd->profiler_process = p;
 			status = 0;
 			mutex_unlock(&kfd->profiler_lock);
-			if (ptl->hw_supported) {
+			if (ptl->hw_supported_state == AMDGPU_PTL_HW_SUPPORTED) {
 				status = kfd_ptl_disable_request(pdd, p);
 				if (status != 0)
 					dev_err(kfd_device,
@@ -3371,7 +3374,7 @@ static inline uint32_t profile_lock_device(struct kfd_process *p,
 		status = 0;
 		mutex_unlock(&kfd->profiler_lock);
 
-		if (ptl->hw_supported) {
+		if (ptl->hw_supported_state == AMDGPU_PTL_HW_SUPPORTED) {
 			status = kfd_ptl_disable_release(pdd, p);
 			if (status)
 				dev_err(kfd_device,
diff --git a/drivers/gpu/drm/amd/include/amdgpu_ptl.h b/drivers/gpu/drm/amd/include/amdgpu_ptl.h
index 154b8da3bfa9..eead62b4b327 100644
--- a/drivers/gpu/drm/amd/include/amdgpu_ptl.h
+++ b/drivers/gpu/drm/amd/include/amdgpu_ptl.h
@@ -39,11 +39,18 @@ enum amdgpu_ptl_disable_source {
 	AMDGPU_PTL_DISABLE_PROFILER,
 	AMDGPU_PTL_DISABLE_MAX,
 };
+
+enum amdgpu_ptl_hw_supported_state {
+	AMDGPU_PTL_HW_UNINIT = 0,       /* Not yet initialized */
+	AMDGPU_PTL_HW_SUPPORTED,        /* Initialized and supported */
+	AMDGPU_PTL_HW_NOT_SUPPORTED,    /* Initialized and not supported */
+};
+
 struct amdgpu_ptl {
 	enum amdgpu_ptl_fmt		fmt1;
 	enum amdgpu_ptl_fmt		fmt2;
 	bool				enabled;
-	bool				hw_supported;
+	enum amdgpu_ptl_hw_supported_state hw_supported_state;
 	bool				permanently_disabled;
 	/* PTL disable reference counting */
 	atomic_t			disable_ref;