Re: [PATCH] drm/amdkfd: Check if there are kfd porcesses using adev by kfd_processes_count

["Kuehling, Felix" <felix.kuehling@amd.com>]

On 2026-04-24 15:11, Xiaogang.Chen wrote:
> From: Xiaogang Chen <xiaogang.chen@amd.com>
>
> Fixes: 6cca686dfce7("drm/amdkfd: kfd driver supports hot unplug/replug
> amdgpu devices")
>
> During gpu hot-unplug need check if there are kfd porcesses still using the
> being removed gpu before clean resources of the device. Current driver checks
> if kfd_processes_table is empty. kfd processes are not terminated after
> removed from kfd_processes_table immediately. They are still alive and may
> access the device until kfd_process_wq work queue got ran.
>
> Check kfd->kfd_processes_count value that is updated after kfd process got
> uninitialized when its ref becomes zero.
>
> Signed-off-by: Xiaogang Chen <xiaogang.chen@amd.com>

Reviewed-by: Felix Kuehling <felix.kuehling@amd.com>


> ---
>   drivers/gpu/drm/amd/amdkfd/kfd_device.c | 33 +------------------------
>   1 file changed, 1 insertion(+), 32 deletions(-)
>
> diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_device.c b/drivers/gpu/drm/amd/amdkfd/kfd_device.c
> index 9a66ee661e57..71fd3b1d0b3a 100644
> --- a/drivers/gpu/drm/amd/amdkfd/kfd_device.c
> +++ b/drivers/gpu/drm/amd/amdkfd/kfd_device.c
> @@ -1737,37 +1737,6 @@ bool kgd2kfd_vmfault_fast_path(struct amdgpu_device *adev, struct amdgpu_iv_entr
>   	return false;
>   }
>   
> -/* check if there is kfd process still uses adev */
> -static bool kgd2kfd_check_device_idle(struct amdgpu_device *adev)
> -{
> -	struct kfd_process *p;
> -	struct hlist_node *p_temp;
> -	unsigned int temp;
> -	struct kfd_node *dev;
> -
> -	mutex_lock(&kfd_processes_mutex);
> -
> -	if (hash_empty(kfd_processes_table)) {
> -		mutex_unlock(&kfd_processes_mutex);
> -		return true;
> -	}
> -
> -	/* check if there is device still use adev */
> -	hash_for_each_safe(kfd_processes_table, temp, p_temp, p, kfd_processes) {
> -		for (int i = 0; i < p->n_pdds; i++) {
> -			dev = p->pdds[i]->dev;
> -			if (dev->adev == adev) {
> -				mutex_unlock(&kfd_processes_mutex);
> -				return false;
> -			}
> -		}
> -	}
> -
> -	mutex_unlock(&kfd_processes_mutex);
> -
> -	return true;
> -}
> -
>   /** kgd2kfd_teardown_processes - gracefully tear down existing
>    *  kfd processes that use adev
>    *
> @@ -1800,7 +1769,7 @@ void kgd2kfd_teardown_processes(struct amdgpu_device *adev)
>   	mutex_unlock(&kfd_processes_mutex);
>   
>   	/* wait all kfd processes use adev terminate */
> -	while (!kgd2kfd_check_device_idle(adev))
> +	while (!!atomic_read(&adev->kfd.dev->kfd_processes_count))
>   		cond_resched();
>   }
>