Le 11/06/2024 à 09:45, Mohammad Rafi Shaik a écrit :
From: Prasad Kumpatla quic_pkumpatl@quicinc.com
This patch adds basic SoundWire codec driver to support for WCD9370/WCD9375 TX and RX devices.
The WCD9370/WCD9375 has Multi Button Headset Control hardware to support Headset insertion, type detection, 8 headset buttons detection, Over Current detection and Impedence measurements. This patch adds support for this using wcd-mbhc apis.
Co-developed-by: Konrad Dybcio konrad.dybcio@linaro.org Signed-off-by: Konrad Dybcio konrad.dybcio@linaro.org Signed-off-by: Prasad Kumpatla quic_pkumpatl@quicinc.com Co-developed-by: Mohammad Rafi Shaik quic_mohs@quicinc.com Signed-off-by: Mohammad Rafi Shaik quic_mohs@quicinc.com
Hi,
this patch has reached -next, but I have a question.
If I'm correct, I can send a patch, but if the fix can be folded somewhere, this is also fine for me.
...
+static int wcd937x_probe(struct platform_device *pdev) +{
- struct component_match *match = NULL;
- struct device *dev = &pdev->dev;
- struct wcd937x_priv *wcd937x;
- struct wcd_mbhc_config *cfg;
- int ret;
- wcd937x = devm_kzalloc(dev, sizeof(*wcd937x), GFP_KERNEL);
- if (!wcd937x)
return -ENOMEM;- dev_set_drvdata(dev, wcd937x);
- mutex_init(&wcd937x->micb_lock);
- wcd937x->reset_gpio = devm_gpiod_get(dev, "reset", GPIOD_OUT_LOW);
- if (IS_ERR(wcd937x->reset_gpio))
return dev_err_probe(dev, PTR_ERR(wcd937x->reset_gpio),"failed to reset wcd gpio\n");- wcd937x->us_euro_gpio = devm_gpiod_get_optional(dev, "us-euro", GPIOD_OUT_LOW);
- if (IS_ERR(wcd937x->us_euro_gpio))
return dev_err_probe(dev, PTR_ERR(wcd937x->us_euro_gpio),"us-euro swap Control GPIO not found\n");- cfg = &wcd937x->mbhc_cfg;
- cfg->swap_gnd_mic = wcd937x_swap_gnd_mic;
- wcd937x->supplies[0].supply = "vdd-rxtx";
- wcd937x->supplies[1].supply = "vdd-px";
- wcd937x->supplies[2].supply = "vdd-mic-bias";
- ret = devm_regulator_bulk_get(dev, WCD937X_MAX_BULK_SUPPLY, wcd937x->supplies);
- if (ret)
return dev_err_probe(dev, ret, "Failed to get supplies\n");- ret = regulator_bulk_enable(WCD937X_MAX_BULK_SUPPLY, wcd937x->supplies);
- if (ret)
return dev_err_probe(dev, ret, "Failed to enable supplies\n");- /* Get the buck separately, as it needs special handling */
- wcd937x->buck_supply = devm_regulator_get(dev, "vdd-buck");
- if (IS_ERR(wcd937x->buck_supply))
regulator_bulk_disable() is missing here...
return dev_err_probe(dev, PTR_ERR(wcd937x->buck_supply),"Failed to get buck supply\n");- ret = regulator_enable(wcd937x->buck_supply);
- if (ret)
return dev_err_probe(dev, ret, "Failed to enable buck supply\n");
... and here...
Also, should 'buck_supply' be disabled in the error handling path of the probe and in the remove function? (and maybe even somewhere else related to wcd937x_codec_enable_vdd_buck())
- wcd937x_dt_parse_micbias_info(dev, wcd937x);
- cfg->mbhc_micbias = MIC_BIAS_2;
- cfg->anc_micbias = MIC_BIAS_2;
- cfg->v_hs_max = WCD_MBHC_HS_V_MAX;
- cfg->num_btn = WCD937X_MBHC_MAX_BUTTONS;
- cfg->micb_mv = wcd937x->micb2_mv;
- cfg->linein_th = 5000;
- cfg->hs_thr = 1700;
- cfg->hph_thr = 50;
- wcd_dt_parse_mbhc_data(dev, &wcd937x->mbhc_cfg);
- ret = wcd937x_add_slave_components(wcd937x, dev, &match);
- if (ret)
return ret;
... and here...
- wcd937x_reset(wcd937x);
- ret = component_master_add_with_match(dev, &wcd937x_comp_ops, match);
- if (ret)
return ret;
... and here.
Maybe a devm_add_action_ior_reset() could help?
- pm_runtime_set_autosuspend_delay(dev, 1000);
- pm_runtime_use_autosuspend(dev);
- pm_runtime_mark_last_busy(dev);
- pm_runtime_set_active(dev);
- pm_runtime_enable(dev);
- pm_runtime_idle(dev);
- return ret;
+}
+static void wcd937x_remove(struct platform_device *pdev) +{
- struct device *dev = &pdev->dev;
- struct wcd937x_priv *wcd937x = dev_get_drvdata(dev);
- component_master_del(&pdev->dev, &wcd937x_comp_ops);
- pm_runtime_disable(dev);
- pm_runtime_set_suspended(dev);
- pm_runtime_dont_use_autosuspend(dev);
- regulator_bulk_disable(WCD937X_MAX_BULK_SUPPLY, wcd937x->supplies);
- regulator_bulk_free(WCD937X_MAX_BULK_SUPPLY, wcd937x->supplies);
This has been allocated with devm_regulator_bulk_get(), so this call looks redundant (but harmless).
+}
...