Re: [PATCH v2] ALSA: hda/tegra: enable clock during probe

[Jon Hunter <jonathanh@nvidia.com>]

On 25/01/2019 12:19, Sameer Pujar wrote:
> 
> On 1/25/2019 5:12 PM, Jon Hunter wrote:
>> On 25/01/2019 11:06, Sameer Pujar wrote:
>>> If CONFIG_PM is disabled or runtime PM calls are forbidden, the clocks
>>> will not be ON. This could cause issue during probe, where hda init
>>> setup is done. This patch enables clocks unconditionally during probe.
>>>
>>> Along with above, follwoing changes are done.
>>>    * enable runtime PM before exiting from probe work. This helps to
>>> avoid
>>>      usage of pm_runtime_get_sync/pm_runtime_put() in probe work.
>>>    * hda_tegra_disable_clocks() is moved out of CONFIG_PM_SLEEP check.
>>>    * runtime PM callbacks moved out of CONFIG_PM check
>>>
>>> Signed-off-by: Sameer Pujar <spujar@nvidia.com>
>>> Reviewed-by: Ravindra Lokhande <rlokhande@nvidia.com>
>>> Reviewed-by: Jon Hunter <jonathanh@nvidia.com>
>>> ---
>>>   sound/pci/hda/hda_tegra.c | 26 +++++++++++++++++---------
>>>   1 file changed, 17 insertions(+), 9 deletions(-)
>>>
>>> diff --git a/sound/pci/hda/hda_tegra.c b/sound/pci/hda/hda_tegra.c
>>> index c8d18dc..ba6175f 100644
>>> --- a/sound/pci/hda/hda_tegra.c
>>> +++ b/sound/pci/hda/hda_tegra.c
>>> @@ -219,7 +219,6 @@ static int hda_tegra_enable_clocks(struct
>>> hda_tegra *data)
>>>       return rc;
>>>   }
>>>   -#ifdef CONFIG_PM_SLEEP
>>>   static void hda_tegra_disable_clocks(struct hda_tegra *data)
>>>   {
>>>       clk_disable_unprepare(data->hda2hdmi_clk);
>>> @@ -227,6 +226,7 @@ static void hda_tegra_disable_clocks(struct
>>> hda_tegra *data)
>>>       clk_disable_unprepare(data->hda_clk);
>>>   }
>>>   +#ifdef CONFIG_PM_SLEEP
>>>   /*
>>>    * power management
>>>    */
>>> @@ -257,7 +257,6 @@ static int hda_tegra_resume(struct device *dev)
>>>   }
>>>   #endif /* CONFIG_PM_SLEEP */
>>>   -#ifdef CONFIG_PM
>>>   static int hda_tegra_runtime_suspend(struct device *dev)
>>>   {
>>>       struct snd_card *card = dev_get_drvdata(dev);
>>> @@ -283,7 +282,7 @@ static int hda_tegra_runtime_resume(struct device
>>> *dev)
>>>       int rc;
>>>         rc = hda_tegra_enable_clocks(hda);
>>> -    if (rc != 0)
>>> +    if (rc)
>>>           return rc;
>>>       if (chip && chip->running) {
>>>           hda_tegra_init(hda);
>>> @@ -292,7 +291,6 @@ static int hda_tegra_runtime_resume(struct device
>>> *dev)
>>>         return 0;
>>>   }
>>> -#endif /* CONFIG_PM */
>>>     static const struct dev_pm_ops hda_tegra_pm = {
>>>       SET_SYSTEM_SLEEP_PM_OPS(hda_tegra_suspend, hda_tegra_resume)
>>> @@ -551,9 +549,9 @@ static int hda_tegra_probe(struct platform_device
>>> *pdev)
>>>         dev_set_drvdata(&pdev->dev, card);
>>>   -    pm_runtime_enable(hda->dev);
>>> -    if (!azx_has_pm_runtime(chip))
>>> -        pm_runtime_forbid(hda->dev);
>>> +    err = hda_tegra_enable_clocks(hda);
>>> +    if (err)
>>> +        goto out_free;
>>>         schedule_work(&hda->probe_work);
>>>   @@ -571,7 +569,6 @@ static void hda_tegra_probe_work(struct
>>> work_struct *work)
>>>       struct platform_device *pdev = to_platform_device(hda->dev);
>>>       int err;
>>>   -    pm_runtime_get_sync(hda->dev);
>>>       err = hda_tegra_first_init(chip, pdev);
>>>       if (err < 0)
>>>           goto out_free;
>>> @@ -592,8 +589,15 @@ static void hda_tegra_probe_work(struct
>>> work_struct *work)
>>>       chip->running = 1;
>>>       snd_hda_set_power_save(&chip->bus, power_save * 1000);
>>>   +    /* set device state as active */
>>> +    if (pm_runtime_set_active(hda->dev) < 0)
>>> +        goto out_free;
>>> +    /* enable runtime PM */
>>> +    pm_runtime_enable(hda->dev);
>>> +    if (!azx_has_pm_runtime(chip))
>>> +        pm_runtime_forbid(hda->dev);
>>> +
>>>    out_free:
>>> -    pm_runtime_put(hda->dev);
>>>       return; /* no error return from async probe */
>>>   }
>>>   @@ -603,6 +607,10 @@ static int hda_tegra_remove(struct
>>> platform_device *pdev)
>>>         ret = snd_card_free(dev_get_drvdata(&pdev->dev));
>>>       pm_runtime_disable(&pdev->dev);
>>> +    if (!pm_runtime_status_suspended(&pdev->dev)) {
>>> +        hda_tegra_runtime_suspend(&pdev->dev);
>>> +        pm_runtime_set_suspended(&pdev->dev);
>>> +    }
>> I think that we need to be consistent in the above with what is done in
>> the probe. If in the probe we call hda_tegra_enable_clocks(), then here
>> we should call hda_tegra_disable_clocks(). However, my preference is
>> still to all hda_tegra_runtime_resume() in probe and then leave the
>> above as-is. Let's see what everyone else thinks.
> Though it is good to have the handling at a single place, it is slightly
> confusing to call
> runtime_resume() in probe. When runtime PM is not yet enabled, why call
> something that is related
> to it(not logically but intuitively)!

I still don't think that it is confusing, the name of the function is
called hda_tegra_runtime_resume and does what is says on the tin (ie.
resumes the device at runtime).

Furthermore, I think it is completely logical that if pm-runtime is not
enabled, we manually call it. The benefit is that we have single
function for resuming the device for all circumstances.

Cheers
Jon

-- 
nvpublic