Messages in this thread | | | Subject | Re: [PATCH v2 12/16] ASoC: qcom: audioreach: add q6prm support | From | Srinivas Kandagatla <> | Date | Thu, 15 Jul 2021 11:32:33 +0100 |
| |
On 14/07/2021 18:09, Pierre-Louis Bossart wrote: > > > >> +static int q6prm_send_cmd_sync(struct q6prm *prm, struct gpr_pkt *pkt, >> + uint32_t rsp_opcode) >> +{ >> + gpr_device_t *gdev = prm->gdev; >> + struct gpr_hdr *hdr = &pkt->hdr; >> + int rc; >> + >> + mutex_lock(&prm->lock); >> + prm->result.opcode = 0; >> + prm->result.status = 0; >> + >> + rc = gpr_send_pkt(prm->gdev, pkt); >> + if (rc < 0) >> + goto err; >> + >> + if (rsp_opcode) >> + rc = wait_event_timeout(prm->wait, >> + (prm->result.opcode == hdr->opcode) || >> + (prm->result.opcode == rsp_opcode), >> + 5 * HZ); >> + else >> + rc = wait_event_timeout(prm->wait, >> + (prm->result.opcode == hdr->opcode), >> + 5 * HZ); >> + >> + if (!rc) { >> + dev_err(&gdev->dev, "CMD timeout for [%x] opcode\n", >> + hdr->opcode); >> + rc = -ETIMEDOUT; >> + } else if (prm->result.status > 0) { >> + dev_err(&gdev->dev, "DSP returned error[%x] %x\n", hdr->opcode, >> + prm->result.status); >> + rc = -EINVAL; >> + } else { >> + dev_err(&gdev->dev, "DSP returned [%x]\n", >> + prm->result.status); >> + rc = 0; >> + } >> + >> +err: >> + mutex_unlock(&prm->lock); >> + return rc; >> +} > > In patch7 you had more or less the same code. can a helper be used?
Its possible, will abstract this out in audioreach.c.
> > +int audioreach_graph_send_cmd_sync(struct q6apm_graph *graph, > + struct gpr_pkt *pkt, uint32_t rsp_opcode) > +{ > + > + struct device *dev = graph->dev; > + struct gpr_hdr *hdr = &pkt->hdr; > + int rc; > + > + mutex_lock(&graph->cmd_lock); > + graph->result.opcode = 0; > + graph->result.status = 0; > + > + rc = gpr_send_port_pkt(graph->port, pkt); > + if (rc < 0) > + goto err; > + > + if (rsp_opcode) > + rc = wait_event_timeout(graph->cmd_wait, > + (graph->result.opcode == hdr->opcode) || > + (graph->result.opcode == rsp_opcode), > + 5 * HZ); > + else > + rc = wait_event_timeout(graph->cmd_wait, > + (graph->result.opcode == hdr->opcode), > + 5 * HZ); > + > + if (!rc) { > + dev_err(dev, "CMD timeout for [%x] opcode\n", hdr->opcode); > + rc = -ETIMEDOUT; > + } else if (graph->result.status > 0) { > + dev_err(dev, "DSP returned error[%x] %x\n", hdr->opcode, > + graph->result.status); > + rc = -EINVAL; > + } else { > + dev_err(dev, "DSP returned [%x]\n", graph->result.status); > + rc = 0; > + } > + > +err: > + mutex_unlock(&graph->cmd_lock); > + return rc; > +} > +EXPORT_SYMBOL_GPL(audioreach_graph_send_cmd_sync); > > >> +static int q6prm_set_hw_core_req(struct device *dev, uint32_t hw_block_id, bool enable) >> +{ >> + struct prm_cmd_request_hw_core *req; >> + struct apm_module_param_data *param_data; >> + struct gpr_pkt *pkt; >> + struct q6prm *prm = dev_get_drvdata(dev); >> + gpr_device_t *gdev = prm->gdev; >> + void *p; >> + int rc = 0; >> + uint32_t opcode, rsp_opcode; >> + >> + if (enable) { >> + opcode = PRM_CMD_REQUEST_HW_RSC; >> + rsp_opcode = PRM_CMD_RSP_REQUEST_HW_RSC; >> + } else { >> + opcode = PRM_CMD_RELEASE_HW_RSC; >> + rsp_opcode = PRM_CMD_RSP_RELEASE_HW_RSC; >> + } >> + >> + p = audioreach_alloc_cmd_pkt(sizeof(*req), opcode, 0, gdev->svc.id, >> + GPR_PRM_MODULE_IID); >> + if (IS_ERR(p)) >> + return -ENOMEM; >> + >> + pkt = p; >> + req = p + GPR_HDR_SIZE + APM_CMD_HDR_SIZE; >> + >> + param_data = &req->param_data; >> + >> + param_data->module_instance_id = GPR_PRM_MODULE_IID; >> + param_data->error_code = 0; >> + param_data->param_id = PARAM_ID_RSC_HW_CORE; >> + param_data->param_size = sizeof(*req) - APM_MODULE_PARAM_DATA_SIZE; >> + >> + req->hw_clk_id = hw_block_id; >> + >> + q6prm_send_cmd_sync(prm, pkt, rsp_opcode); >> + >> + kfree(pkt); >> + >> + return rc; > > rc is not assigned, should this not trap the result of sending a command? My bad! will fix such instances.
--srini
| |