On 21/09/2021 18:05, Pierre-Louis Bossart wrote:
+static void apm_populate_connection_obj(struct apm_module_conn_obj *obj,
+ struct audioreach_module *module)
+{
+ obj->src_mod_inst_id = module->src_mod_inst_id;
+ obj->src_mod_op_port_id = module->src_mod_op_port_id;
+ obj->dst_mod_inst_id = module->instance_id;
+ obj->dst_mod_ip_port_id = module->in_port;
alignment seems off?
Its fixed in next version.
+void *audioreach_alloc_graph_pkt(struct q6apm *apm, struct list_head *sg_list, int graph_id)
+{
+ void *p;
move this as last declaration to have a nice reverse christmas tree style?
Done.
+ int payload_size, sg_sz, cont_sz, ml_sz, mp_sz, mc_sz;
+ struct apm_module_param_data *param_data;
+ struct audioreach_container *container;
+ struct apm_graph_open_params params;
+ struct audioreach_sub_graph *sgs;
+ struct audioreach_module *module;
+ int num_modules_per_list;
+ int num_connections = 0;
+ int num_containers = 0;
+ int num_sub_graphs = 0;
+ int num_modules = 0;
+ int num_modules_list;
+ struct gpr_pkt *pkt;
[...]
+static struct audioreach_graph *q6apm_get_audioreach_graph(struct q6apm *apm, uint32_t graph_id)
+{
+ struct audioreach_graph_info *info;
+ struct audioreach_graph *graph;
+
+ mutex_lock(&apm->lock);
+ graph = idr_find(&apm->graph_idr, graph_id);
+ mutex_unlock(&apm->lock);
+
+ if (graph) {
+ kref_get(&graph->refcount);
+ return graph;
+ }
+
+ info = idr_find(&apm->graph_info_idr, graph_id);
+
+ if (!info)
+ return ERR_PTR(-ENODEV);
+
+ graph = kzalloc(sizeof(*graph), GFP_KERNEL);
+ if (!graph)
+ return ERR_PTR(-ENOMEM);
+
+ graph->apm = apm;
+ graph->info = info;
+ graph->id = graph_id;
+
+ graph->graph = audioreach_alloc_graph_pkt(apm, &info->sg_list, graph_id);
+ if (IS_ERR(graph->graph)) {
+ kfree(graph);
+ return ERR_PTR(-ENOMEM);
why not return graph->graph (store and return the value before freeing
graph)?
Updated something like this now:
if (IS_ERR(graph->graph)) {
void *err = graph->graph;
kfree(graph);
return ERR_CAST(err);
}
+ }
+
+ mutex_lock(&apm->lock);
+ if (idr_alloc(&apm->graph_idr, graph, graph_id, graph_id + 1, GFP_KERNEL) < 0) {
+ dev_err(apm->dev, "Unable to allocate graph id (%d)\n", graph_id);
+ kfree(graph);
+ mutex_unlock(&apm->lock);
+ return ERR_PTR(-ENOMEM);
+ }
+ mutex_unlock(&apm->lock);
+
+ kref_init(&graph->refcount);
+
+ q6apm_send_cmd_sync(apm, graph->graph, 0);
+
+ return graph;
+}
+static int graph_callback(struct gpr_resp_pkt *data, void *priv, int op)
+{
+ struct data_cmd_rsp_rd_sh_mem_ep_data_buffer_done_v2 *rd_done;
+ struct data_cmd_rsp_wr_sh_mem_ep_data_buffer_done_v2 *done;
+ struct apm_cmd_rsp_shared_mem_map_regions *rsp;
+ struct gpr_ibasic_rsp_result_t *result;
+ struct q6apm_graph *graph = priv;
+ struct gpr_hdr *hdr = &data->hdr;
+ struct device *dev = graph->dev;
+ uint32_t client_event;
+ int ret = -EINVAL;
+ phys_addr_t phys;
+ int token;
+
+ result = data->payload;
+
+ switch (hdr->opcode) {
+ case DATA_CMD_RSP_WR_SH_MEM_EP_DATA_BUFFER_DONE_V2:
+ client_event = APM_CLIENT_EVENT_DATA_WRITE_DONE;
+ mutex_lock(&graph->lock);
+ token = hdr->token & APM_WRITE_TOKEN_MASK;
+
+ done = data->payload;
+ phys = graph->rx_data.buf[token].phys;
+
+ if (lower_32_bits(phys) != done->buf_addr_lsw ||
+ upper_32_bits(phys) != done->buf_addr_msw) {
+ dev_err(dev, "WR BUFF Unexpected addr %08x-%08x\n",
+ done->buf_addr_lsw, done->buf_addr_msw);
+ ret = -EINVAL;
since you don't return here you might invoke graph->cb() below, is this
desired?
No, it does not make sense to do the cb() for un-expected/error response.
TBH, rethinking on this..
returning error from these callbacks is totally useless as the apr-bus
will not do anything about this. Only reporting it as dev_err and caller
receiving a command timeout on error is the right thing to do.
This callback will now only return 0 irrespective of error response or
not. CMD will Timeout for the caller in error cases.
--srini
+ } else {
+ ret = 0;
+ graph->result.opcode = hdr->opcode;
+ graph->result.status = done->status;
+ }
+ mutex_unlock(&graph->lock);
+ if (graph->cb)
+ graph->cb(client_event, hdr->token, data->payload,
+ graph->priv);
+
+ break;
+ case APM_CMD_RSP_SHARED_MEM_MAP_REGIONS:
+ graph->result.opcode = hdr->opcode;
+ graph->result.status = 0;
+ rsp = data->payload;
+
+ if (hdr->token == SNDRV_PCM_STREAM_PLAYBACK)
+ graph->rx_data.mem_map_handle = rsp->mem_map_handle;
+ else
+ graph->tx_data.mem_map_handle = rsp->mem_map_handle;
+
+ wake_up(&graph->cmd_wait);
+ ret = 0;
+ break;
+ case DATA_CMD_RSP_RD_SH_MEM_EP_DATA_BUFFER_V2:
+ client_event = APM_CLIENT_EVENT_DATA_READ_DONE;
+ mutex_lock(&graph->lock);
+ rd_done = data->payload;
+ phys = graph->tx_data.buf[hdr->token].phys;
+ if (upper_32_bits(phys) != rd_done->buf_addr_msw ||
+ lower_32_bits(phys) != rd_done->buf_addr_lsw) {
+ dev_err(dev, "RD BUFF Unexpected addr %08x-%08x\n",
+ rd_done->buf_addr_lsw, rd_done->buf_addr_msw);
+ ret = -EINVAL;
same here, you will call wake_up and conditionally the callback.
+ } else {
+ ret = 0;
+ }
+ mutex_unlock(&graph->lock);
+ wake_up(&graph->cmd_wait);
+
+ if (graph->cb)
+ graph->cb(client_event, hdr->token, data->payload,
+ graph->priv);
+ break;
+ case DATA_CMD_WR_SH_MEM_EP_EOS_RENDERED:
+ break;
+ case GPR_BASIC_RSP_RESULT:
+ switch (result->opcode) {
+ case APM_CMD_SHARED_MEM_UNMAP_REGIONS:
+ graph->result.opcode = result->opcode;
+ graph->result.status = 0;
+ if (hdr->token == SNDRV_PCM_STREAM_PLAYBACK)
+ graph->rx_data.mem_map_handle = 0;
+ else
+ graph->tx_data.mem_map_handle = 0;
+
+ wake_up(&graph->cmd_wait);
+ ret = 0;
+ break;
+ case APM_CMD_SHARED_MEM_MAP_REGIONS:
+ case DATA_CMD_WR_SH_MEM_EP_MEDIA_FORMAT:
+ case APM_CMD_SET_CFG:
+ graph->result.opcode = result->opcode;
+ graph->result.status = result->status;
+ if (result->status) {
+ dev_err(dev, "Error (%d) Processing 0x%08x cmd\n",
+ result->status, result->opcode);
+ ret = -EINVAL;
and here as well.
+ } else {
+ ret = 0;
+ }
+ wake_up(&graph->cmd_wait);
+ break;
+ default:
+ break;
+ }
+ break;
+ default:
+ break;
+ }
+
+ return ret;
+}
+static int apm_callback(struct gpr_resp_pkt *data, void *priv, int op)
+{
+ gpr_device_t *gdev = priv;
+ struct q6apm *apm = dev_get_drvdata(&gdev->dev);
+ struct device *dev = &gdev->dev;
+ struct gpr_ibasic_rsp_result_t *result;
+ struct gpr_hdr *hdr = &data->hdr;
+ int ret = 0;
+
+ result = data->payload;
+
+ switch (hdr->opcode) {
+ case APM_CMD_RSP_GET_SPF_STATE:
+ apm->result.opcode = hdr->opcode;
+ apm->result.status = 0;
+ /* First word of result it state */
+ apm->state = result->opcode;
+ wake_up(&apm->wait);
+ break;
+ case GPR_BASIC_RSP_RESULT:
+ switch (result->opcode) {
+ case APM_CMD_GRAPH_START:
+ case APM_CMD_GRAPH_OPEN:
+ case APM_CMD_GRAPH_PREPARE:
+ case APM_CMD_GRAPH_CLOSE:
+ case APM_CMD_GRAPH_FLUSH:
+ case APM_CMD_GRAPH_STOP:
+ case APM_CMD_SET_CFG:
+ apm->result.opcode = result->opcode;
+ apm->result.status = result->status;
+ if (result->status) {
+ dev_err(dev, "Error (%d) Processing 0x%08x cmd\n",
+ result->status, result->opcode);
+ ret = -EINVAL;
same pattern, you will call wake_up even on an error?
+ }
+ wake_up(&apm->wait);
+ break;
+ default:
+ break;
+ }
+ break;
+ default:
+ break;
+ }
+
+ return ret;
+}
