[Public]
Maybe we need a PTE schedule check in the amdgpu_vm_generation() for the userq
only case.
This could be a sperate patch for handling that, for this patch is
Reviewed-by: Prike Liang <prike.li...@amd.com>
Regards,
Prike
> -----Original Message-----
> From: Liang, Prike <prike.li...@amd.com>
> Sent: Wednesday, September 10, 2025 7:37 PM
> To: amd-gfx@lists.freedesktop.org
> Cc: Deucher, Alexander <alexander.deuc...@amd.com>; Koenig, Christian
> <christian.koe...@amd.com>; Liang, Prike <prike.li...@amd.com>; Khatri, Sunil
> <sunil.kha...@amd.com>; Deucher, Alexander <alexander.deuc...@amd.com>
> Subject: [PATCH 2/9] drm/amdgpu/userq: implement support for query status
>
> Query the status of the user queue, currently whether the queue is hung and
> whether
> or not VRAM is lost.
>
> v2: Misc cleanups
>
> Reviewed-by: Sunil Khatri <sunil.kha...@amd.com>
> Signed-off-by: Alex Deucher <alexander.deuc...@amd.com>
> ---
> drivers/gpu/drm/amd/amdgpu/amdgpu_userq.c | 35 ++++++++++++++++++++++-
> drivers/gpu/drm/amd/amdgpu/amdgpu_userq.h | 1 +
> 2 files changed, 35 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_userq.c
> b/drivers/gpu/drm/amd/amdgpu/amdgpu_userq.c
> index 9608fe3b5a9e..83f0ecdaa0b7 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_userq.c
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_userq.c
> @@ -472,6 +472,7 @@ amdgpu_userq_create(struct drm_file *filp, union
> drm_amdgpu_userq *args)
> queue->queue_type = args->in.ip_type;
> queue->vm = &fpriv->vm;
> queue->priority = priority;
> + queue->generation = amdgpu_vm_generation(adev, &fpriv->vm);
>
> db_info.queue_type = queue->queue_type;
> db_info.doorbell_handle = queue->doorbell_handle; @@ -553,6 +554,34
> @@ amdgpu_userq_create(struct drm_file *filp, union drm_amdgpu_userq *args)
> return r;
> }
>
> +static int
> +amdgpu_userq_query_status(struct drm_file *filp, union drm_amdgpu_userq
> +*args) {
> + struct amdgpu_fpriv *fpriv = filp->driver_priv;
> + struct amdgpu_userq_mgr *uq_mgr = &fpriv->userq_mgr;
> + struct amdgpu_device *adev = uq_mgr->adev;
> + struct amdgpu_usermode_queue *queue;
> + int queue_id = args->in.queue_id;
> +
> + mutex_lock(&uq_mgr->userq_mutex);
> +
> + queue = amdgpu_userq_find(uq_mgr, queue_id);
> + if (!queue) {
> + dev_dbg(adev->dev, "Invalid queue id to query\n");
> + mutex_unlock(&uq_mgr->userq_mutex);
> + return -EINVAL;
> + }
> + args->out_qs.flags = 0;
> + if (queue->state == AMDGPU_USERQ_STATE_HUNG)
> + args->out_qs.flags |=
> AMDGPU_USERQ_QUERY_STATUS_FLAGS_HUNG;
> + if (queue->generation != amdgpu_vm_generation(adev, &fpriv->vm))
> + args->out_qs.flags |=
> AMDGPU_USERQ_QUERY_STATUS_FLAGS_VRAMLOST;
> +
> + mutex_unlock(&uq_mgr->userq_mutex);
> +
> + return 0;
> +}
> +
> static int amdgpu_userq_input_args_validate(struct drm_device *dev,
> union drm_amdgpu_userq *args,
> struct drm_file *filp)
> @@ -633,7 +662,11 @@ int amdgpu_userq_ioctl(struct drm_device *dev, void
> *data,
> if (r)
> drm_file_err(filp, "Failed to destroy usermode
> queue\n");
> break;
> -
> + case AMDGPU_USERQ_OP_QUERY_STATUS:
> + r = amdgpu_userq_query_status(filp, args);
> + if (r)
> + drm_file_err(filp, "Failed to query usermode queue
> status\n");
> + break;
> default:
> drm_dbg_driver(dev, "Invalid user queue op specified: %d\n",
> args-
> >in.op);
> return -EINVAL;
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_userq.h
> b/drivers/gpu/drm/amd/amdgpu/amdgpu_userq.h
> index c027dd916672..2260b1fb8a22 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_userq.h
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_userq.h
> @@ -66,6 +66,7 @@ struct amdgpu_usermode_queue {
> u32 xcp_id;
> int priority;
> struct dentry *debugfs_queue;
> + uint64_t generation;
> };
>
> struct amdgpu_userq_funcs {
> --
> 2.34.1
