accel/amdxdna: Add ioctl DRM_IOCTL_AMDXDNA_GET_ARRAY

Add interface for applications to get information array. The application
provides a buffer pointer along with information type, maximum number of
entries and maximum size of each entry. The buffer may also contain match
conditions based on the information type. After the ioctl completes, the
actual number of entries and entry size are returned. (see [1], used by
driver runtime library)

[1] https://github.com/amd/xdna-driver/blob/main/src/shim/host/platform_host.cpp#L337

Reviewed-by: Mario Limonciello (AMD) <superm1@kernel.org>
Reviewed-by: Maciej Falkowski <maciej.falkowski@linux.intel.com>
Signed-off-by: Lizhi Hou <lizhi.hou@amd.com>
Link: https://lore.kernel.org/r/20250903053402.2103196-1-lizhi.hou@amd.com

3 files changed, 118 insertions, 26 deletions

diff --git a/drivers/accel/amdxdna/aie2_pci.c b/drivers/accel/amdxdna/aie2_pci.c
index 7a3449541107..87c425e3d2b9 100644
--- a/drivers/accel/amdxdna/aie2_pci.c
+++ b/drivers/accel/amdxdna/aie2_pci.c
@@ -785,11 +785,12 @@ static int aie2_get_clock_metadata(struct amdxdna_client *client,
 
 static int aie2_hwctx_status_cb(struct amdxdna_hwctx *hwctx, void *arg)
 {
-	struct amdxdna_drm_query_hwctx *tmp __free(kfree) = NULL;
-	struct amdxdna_drm_get_info *get_info_args = arg;
-	struct amdxdna_drm_query_hwctx __user *buf;
+	struct amdxdna_drm_hwctx_entry *tmp __free(kfree) = NULL;
+	struct amdxdna_drm_get_array *array_args = arg;
+	struct amdxdna_drm_hwctx_entry __user *buf;
+	u32 size;
 
-	if (get_info_args->buffer_size < sizeof(*tmp))
+	if (!array_args->num_element)
 		return -EINVAL;
 
 	tmp = kzalloc(sizeof(*tmp), GFP_KERNEL);
@@ -802,14 +803,23 @@ static int aie2_hwctx_status_cb(struct amdxdna_hwctx *hwctx, void *arg)
 	tmp->num_col = hwctx->num_col;
 	tmp->command_submissions = hwctx->priv->seq;
 	tmp->command_completions = hwctx->priv->completed;
-
-	buf = u64_to_user_ptr(get_info_args->buffer);
-
-	if (copy_to_user(buf, tmp, sizeof(*tmp)))
+	tmp->pasid = hwctx->client->pasid;
+	tmp->priority = hwctx->qos.priority;
+	tmp->gops = hwctx->qos.gops;
+	tmp->fps = hwctx->qos.fps;
+	tmp->dma_bandwidth = hwctx->qos.dma_bandwidth;
+	tmp->latency = hwctx->qos.latency;
+	tmp->frame_exec_time = hwctx->qos.frame_exec_time;
+	tmp->state = AMDXDNA_HWCTX_STATE_ACTIVE;
+
+	buf = u64_to_user_ptr(array_args->buffer);
+	size = min(sizeof(*tmp), array_args->element_size);
+
+	if (copy_to_user(buf, tmp, size))
 		return -EFAULT;
 
-	get_info_args->buffer += sizeof(*tmp);
-	get_info_args->buffer_size -= sizeof(*tmp);
+	array_args->buffer += size;
+	array_args->num_element--;
 
 	return 0;
 }
@@ -817,23 +827,24 @@ static int aie2_hwctx_status_cb(struct amdxdna_hwctx *hwctx, void *arg)
 static int aie2_get_hwctx_status(struct amdxdna_client *client,
 				 struct amdxdna_drm_get_info *args)
 {
+	struct amdxdna_drm_get_array array_args;
 	struct amdxdna_dev *xdna = client->xdna;
-	struct amdxdna_drm_get_info info_args;
 	struct amdxdna_client *tmp_client;
 	int ret;
 
 	drm_WARN_ON(&xdna->ddev, !mutex_is_locked(&xdna->dev_lock));
 
-	info_args.buffer = args->buffer;
-	info_args.buffer_size = args->buffer_size;
-
+	array_args.element_size = sizeof(struct amdxdna_drm_query_hwctx);
+	array_args.buffer = args->buffer;
+	array_args.num_element = args->buffer_size / array_args.element_size;
 	list_for_each_entry(tmp_client, &xdna->client_list, node) {
-		ret = amdxdna_hwctx_walk(tmp_client, &info_args, aie2_hwctx_status_cb);
+		ret = amdxdna_hwctx_walk(tmp_client, &array_args,
+					 aie2_hwctx_status_cb);
 		if (ret)
 			break;
 	}
 
-	args->buffer_size = (u32)(info_args.buffer - args->buffer);
+	args->buffer_size -= (u32)(array_args.buffer - args->buffer);
 	return ret;
 }
 
@@ -877,6 +888,58 @@ static int aie2_get_info(struct amdxdna_client *client, struct amdxdna_drm_get_i
 	return ret;
 }
 
+static int aie2_query_ctx_status_array(struct amdxdna_client *client,
+				       struct amdxdna_drm_get_array *args)
+{
+	struct amdxdna_drm_get_array array_args;
+	struct amdxdna_dev *xdna = client->xdna;
+	struct amdxdna_client *tmp_client;
+	int ret;
+
+	drm_WARN_ON(&xdna->ddev, !mutex_is_locked(&xdna->dev_lock));
+
+	array_args.element_size = min(args->element_size,
+				      sizeof(struct amdxdna_drm_hwctx_entry));
+	array_args.buffer = args->buffer;
+	array_args.num_element = args->num_element * args->element_size /
+				array_args.element_size;
+	list_for_each_entry(tmp_client, &xdna->client_list, node) {
+		ret = amdxdna_hwctx_walk(tmp_client, &array_args,
+					 aie2_hwctx_status_cb);
+		if (ret)
+			break;
+	}
+
+	args->element_size = array_args.element_size;
+	args->num_element = (u32)((array_args.buffer - args->buffer) /
+				  args->element_size);
+
+	return ret;
+}
+
+static int aie2_get_array(struct amdxdna_client *client,
+			  struct amdxdna_drm_get_array *args)
+{
+	struct amdxdna_dev *xdna = client->xdna;
+	int ret, idx;
+
+	if (!drm_dev_enter(&xdna->ddev, &idx))
+		return -ENODEV;
+
+	switch (args->param) {
+	case DRM_AMDXDNA_HW_CONTEXT_ALL:
+		ret = aie2_query_ctx_status_array(client, args);
+		break;
+	default:
+		XDNA_ERR(xdna, "Not supported request parameter %u", args->param);
+		ret = -EOPNOTSUPP;
+	}
+	XDNA_DBG(xdna, "Got param %d", args->param);
+
+	drm_dev_exit(idx);
+	return ret;
+}
+
 static int aie2_set_power_mode(struct amdxdna_client *client,
 			       struct amdxdna_drm_set_state *args)
 {
@@ -926,15 +989,16 @@ static int aie2_set_state(struct amdxdna_client *client,
 }
 
 const struct amdxdna_dev_ops aie2_ops = {
-	.init           = aie2_init,
-	.fini           = aie2_fini,
-	.resume         = aie2_hw_resume,
-	.suspend        = aie2_hw_suspend,
-	.get_aie_info   = aie2_get_info,
-	.set_aie_state	= aie2_set_state,
-	.hwctx_init     = aie2_hwctx_init,
-	.hwctx_fini     = aie2_hwctx_fini,
-	.hwctx_config   = aie2_hwctx_config,
-	.cmd_submit     = aie2_cmd_submit,
+	.init = aie2_init,
+	.fini = aie2_fini,
+	.resume = aie2_hw_resume,
+	.suspend = aie2_hw_suspend,
+	.get_aie_info = aie2_get_info,
+	.set_aie_state = aie2_set_state,
+	.hwctx_init = aie2_hwctx_init,
+	.hwctx_fini = aie2_hwctx_fini,
+	.hwctx_config = aie2_hwctx_config,
+	.cmd_submit = aie2_cmd_submit,
 	.hmm_invalidate = aie2_hmm_invalidate,
+	.get_array = aie2_get_array,
 };
diff --git a/drivers/accel/amdxdna/amdxdna_pci_drv.c b/drivers/accel/amdxdna/amdxdna_pci_drv.c
index 8ef5e4f27f5e..569cd703729d 100644
--- a/drivers/accel/amdxdna/amdxdna_pci_drv.c
+++ b/drivers/accel/amdxdna/amdxdna_pci_drv.c
@@ -27,6 +27,13 @@ MODULE_FIRMWARE("amdnpu/17f0_11/npu.sbin");
 MODULE_FIRMWARE("amdnpu/17f0_20/npu.sbin");
 
 /*
+ * 0.0: Initial version
+ * 0.1: Support getting all hardware contexts by DRM_IOCTL_AMDXDNA_GET_ARRAY
+ */
+#define AMDXDNA_DRIVER_MAJOR		0
+#define AMDXDNA_DRIVER_MINOR		1
+
+/*
  * Bind the driver base on (vendor_id, device_id) pair and later use the
  * (device_id, rev_id) pair as a key to select the devices. The devices with
  * same device_id have very similar interface to host driver.
@@ -164,6 +171,23 @@ static int amdxdna_drm_get_info_ioctl(struct drm_device *dev, void *data, struct
 	return ret;
 }
 
+static int amdxdna_drm_get_array_ioctl(struct drm_device *dev, void *data,
+				       struct drm_file *filp)
+{
+	struct amdxdna_client *client = filp->driver_priv;
+	struct amdxdna_dev *xdna = to_xdna_dev(dev);
+	struct amdxdna_drm_get_array *args = data;
+
+	if (!xdna->dev_info->ops->get_array)
+		return -EOPNOTSUPP;
+
+	if (args->pad || !args->num_element || !args->element_size)
+		return -EINVAL;
+
+	guard(mutex)(&xdna->dev_lock);
+	return xdna->dev_info->ops->get_array(client, args);
+}
+
 static int amdxdna_drm_set_state_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
 {
 	struct amdxdna_client *client = filp->driver_priv;
@@ -195,6 +219,7 @@ static const struct drm_ioctl_desc amdxdna_drm_ioctls[] = {
 	DRM_IOCTL_DEF_DRV(AMDXDNA_EXEC_CMD, amdxdna_drm_submit_cmd_ioctl, 0),
 	/* AIE hardware */
 	DRM_IOCTL_DEF_DRV(AMDXDNA_GET_INFO, amdxdna_drm_get_info_ioctl, 0),
+	DRM_IOCTL_DEF_DRV(AMDXDNA_GET_ARRAY, amdxdna_drm_get_array_ioctl, 0),
 	DRM_IOCTL_DEF_DRV(AMDXDNA_SET_STATE, amdxdna_drm_set_state_ioctl, DRM_ROOT_ONLY),
 };
 
@@ -218,6 +243,8 @@ const struct drm_driver amdxdna_drm_drv = {
 	.fops = &amdxdna_fops,
 	.name = "amdxdna_accel_driver",
 	.desc = "AMD XDNA DRM implementation",
+	.major = AMDXDNA_DRIVER_MAJOR,
+	.minor = AMDXDNA_DRIVER_MINOR,
 	.open = amdxdna_drm_open,
 	.postclose = amdxdna_drm_close,
 	.ioctls = amdxdna_drm_ioctls,
diff --git a/drivers/accel/amdxdna/amdxdna_pci_drv.h b/drivers/accel/amdxdna/amdxdna_pci_drv.h
index b6b3b424d1d5..72d6696d49da 100644
--- a/drivers/accel/amdxdna/amdxdna_pci_drv.h
+++ b/drivers/accel/amdxdna/amdxdna_pci_drv.h
@@ -58,6 +58,7 @@ struct amdxdna_dev_ops {
 	int (*cmd_submit)(struct amdxdna_hwctx *hwctx, struct amdxdna_sched_job *job, u64 *seq);
 	int (*get_aie_info)(struct amdxdna_client *client, struct amdxdna_drm_get_info *args);
 	int (*set_aie_state)(struct amdxdna_client *client, struct amdxdna_drm_set_state *args);
+	int (*get_array)(struct amdxdna_client *client, struct amdxdna_drm_get_array *args);
 };
 
 /*