[PATCH 1/2] drm/amdkfd: add an interface to query whether is KFD is active

[SRINIVASAN SHANMUGAM]

The series is:
Acked-by: Srinivasan Shanmugam <srinivasan.shanmugam at amd.com>

On 10/28/2024 8:22 PM, Alex Deucher wrote:
> Add an interface to query whether KFD has any active queues.
>
> Signed-off-by: Alex Deucher <alexander.deucher at amd.com>
> ---
>   drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c |  9 ++++++++
>   drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h |  7 ++++++
>   drivers/gpu/drm/amd/amdkfd/kfd_device.c    | 25 ++++++++++++++++++++++
>   drivers/gpu/drm/amd/amdkfd/kfd_priv.h      |  1 +
>   4 files changed, 42 insertions(+)
>
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c
> index b545940e512b..82f1c6a19d79 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c
> @@ -890,6 +890,15 @@ int amdgpu_amdkfd_start_sched(struct amdgpu_device *adev, uint32_t node_id)
>   	return kgd2kfd_start_sched(adev->kfd.dev, node_id);
>   }
>   
> +/* check if there are KFD queues active */
> +bool amdgpu_amdkfd_compute_active(struct amdgpu_device *adev, uint32_t node_id)
> +{
> +	if (!adev->kfd.init_complete)
> +		return 0;
> +
> +	return kgd2kfd_compute_active(adev->kfd.dev, node_id);
> +}
> +
>   /* Config CGTT_SQ_CLK_CTRL */
>   int amdgpu_amdkfd_config_sq_perfmon(struct amdgpu_device *adev, uint32_t xcp_id,
>   	bool core_override_enable, bool reg_override_enable, bool perfmon_override_enable)
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h
> index 7e0a22072536..f47f4555437b 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h
> @@ -268,6 +268,7 @@ int amdgpu_amdkfd_start_sched(struct amdgpu_device *adev, uint32_t node_id);
>   int amdgpu_amdkfd_stop_sched(struct amdgpu_device *adev, uint32_t node_id);
>   int amdgpu_amdkfd_config_sq_perfmon(struct amdgpu_device *adev, uint32_t xcp_id,
>   	bool core_override_enable, bool reg_override_enable, bool perfmon_override_enable);
> +bool amdgpu_amdkfd_compute_active(struct amdgpu_device *adev, uint32_t node_id);
>   
>   
>   /* Read user wptr from a specified user address space with page fault
> @@ -431,6 +432,7 @@ int kgd2kfd_check_and_lock_kfd(void);
>   void kgd2kfd_unlock_kfd(void);
>   int kgd2kfd_start_sched(struct kfd_dev *kfd, uint32_t node_id);
>   int kgd2kfd_stop_sched(struct kfd_dev *kfd, uint32_t node_id);
> +bool kgd2kfd_compute_active(struct kfd_dev *kfd, uint32_t node_id);
>   #else
>   static inline int kgd2kfd_init(void)
>   {
> @@ -511,5 +513,10 @@ static inline int kgd2kfd_stop_sched(struct kfd_dev *kfd, uint32_t node_id)
>   {
>   	return 0;
>   }
> +
> +bool kgd2kfd_compute_active(struct kfd_dev *kfd, uint32_t node_id)
> +{
> +	return false;
> +}
>   #endif
>   #endif /* AMDGPU_AMDKFD_H_INCLUDED */
> diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_device.c b/drivers/gpu/drm/amd/amdkfd/kfd_device.c
> index b1ce73c7a9ab..9d8bdead6e65 100644
> --- a/drivers/gpu/drm/amd/amdkfd/kfd_device.c
> +++ b/drivers/gpu/drm/amd/amdkfd/kfd_device.c
> @@ -1404,6 +1404,13 @@ void kfd_dec_compute_active(struct kfd_node *node)
>   	WARN_ONCE(count < 0, "Compute profile ref. count error");
>   }
>   
> +bool kfd_compute_active(struct kfd_node *node)
> +{
> +	if (atomic_read(&node->kfd->compute_profile))
> +		return true;
> +	return false;
> +}
> +
>   void kgd2kfd_smi_event_throttle(struct kfd_dev *kfd, uint64_t throttle_bitmask)
>   {
>   	/*
> @@ -1497,6 +1504,24 @@ int kgd2kfd_stop_sched(struct kfd_dev *kfd, uint32_t node_id)
>   	return node->dqm->ops.halt(node->dqm);
>   }
>   
> +bool kgd2kfd_compute_active(struct kfd_dev *kfd, uint32_t node_id)
> +{
> +	struct kfd_node *node;
> +
> +	if (!kfd->init_complete)
> +		return false;
> +
> +	if (node_id >= kfd->num_nodes) {
> +		dev_warn(kfd->adev->dev, "Invalid node ID: %u exceeds %u\n",
> +			 node_id, kfd->num_nodes - 1);
> +		return false;
> +	}
> +
> +	node = kfd->nodes[node_id];
> +
> +	return kfd_compute_active(node);
> +}
> +
>   #if defined(CONFIG_DEBUG_FS)
>   
>   /* This function will send a package to HIQ to hang the HWS
> diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_priv.h b/drivers/gpu/drm/amd/amdkfd/kfd_priv.h
> index 9e5ca0b93b2a..f9140fdb8132 100644
> --- a/drivers/gpu/drm/amd/amdkfd/kfd_priv.h
> +++ b/drivers/gpu/drm/amd/amdkfd/kfd_priv.h
> @@ -1518,6 +1518,7 @@ bool kfd_is_locked(void);
>   /* Compute profile */
>   void kfd_inc_compute_active(struct kfd_node *dev);
>   void kfd_dec_compute_active(struct kfd_node *dev);
> +bool kfd_compute_active(struct kfd_node *dev);
>   
>   /* Cgroup Support */
>   /* Check with device cgroup if @kfd device is accessible */