Re: [PATCH V3 6/8] soundwire: amd: add runtime pm ops for AMD SoundWire manager driver

["Mukunda,Vijendar" <vijendar.mukunda@amd.com>]

On 21/02/23 21:40, Pierre-Louis Bossart wrote:
>> +static int amd_deinit_sdw_manager(struct amd_sdw_manager *amd_manager)
>> +{
>> +	int ret;
>> +
>> +	amd_disable_sdw_interrupts(amd_manager);
>> +	ret = amd_disable_sdw_manager(amd_manager);
>> +	return ret;
> return amd_disable_sdw_manager(amd_manager); ?
will fix it.
>> +}
>> +static int amd_sdw_clock_stop(struct amd_sdw_manager *amd_manager)
>> +{
>> +	u32 val;
>> +	u32 retry_count = 0;
>> +	int ret;
>> +
>> +	ret = sdw_bus_prep_clk_stop(&amd_manager->bus);
>> +	if (ret < 0 && ret != -ENODATA) {
>> +		dev_err(amd_manager->dev, "prepare clock stop failed %d", ret);
>> +		return 0;
>> +	}
>> +	ret = sdw_bus_clk_stop(&amd_manager->bus);
>> +	if (ret < 0 && ret != -ENODATA) {
>> +		dev_err(amd_manager->dev, "bus clock stop failed %d", ret);
>> +		return 0;
>> +	}
>> +
>> +	do {
>> +		val = acp_reg_readl(amd_manager->mmio + ACP_SW_CLK_RESUME_CTRL);
> don't you need a minimal usleep_range to avoid re-reading the same
> register over and over?
>
> This is tied to the frame duration, isn't it? this is 20us typically.
I agree. will update the code with usleep_range(10,20)
>
>> +		if (val & AMD_SDW_CLK_STOP_DONE) {
>> +			amd_manager->clk_stopped = true;
>> +			break;
>> +		}
>> +	} while (retry_count++ < AMD_SDW_CLK_STOP_MAX_RETRY_COUNT);
>> +
>> +	if (!amd_manager->clk_stopped) {
>> +		dev_err(amd_manager->dev, "SDW%x clock stop failed\n", amd_manager->instance);
>> +		return 0;
>> +	}
>> +
>> +	if (amd_manager->wake_en_mask)
>> +		acp_reg_writel(0x01, amd_manager->acp_mmio + ACP_SW_WAKE_EN(amd_manager->instance));
>> +
>> +	dev_dbg(amd_manager->dev, "SDW%x clock stop successful\n", amd_manager->instance);
>> +	return 0;
>> +}
>> +
>> +static int amd_sdw_clock_stop_exit(struct amd_sdw_manager *amd_manager)
>> +{
>> +	int ret;
>> +	u32 val;
>> +	u32 retry_count = 0;
>> +
>> +	if (amd_manager->clk_stopped) {
>> +		val = acp_reg_readl(amd_manager->mmio + ACP_SW_CLK_RESUME_CTRL);
>> +		val |= AMD_SDW_CLK_RESUME_REQ;
>> +		acp_reg_writel(val, amd_manager->mmio + ACP_SW_CLK_RESUME_CTRL);
>> +		do {
>> +			val = acp_reg_readl(amd_manager->mmio + ACP_SW_CLK_RESUME_CTRL);
>> +			if (val & AMD_SDW_CLK_RESUME_DONE)
>> +				break;
>> +			usleep_range(10, 100);
> that's 10x range for sleep, that sounds a vague and suspicious, no?
Agreed . usleep_range(10, 20) is enough in this case.
will fix it.
>
>> +		} while (retry_count++ < AMD_SDW_CLK_STOP_MAX_RETRY_COUNT);
>> +		if (val & AMD_SDW_CLK_RESUME_DONE) {
>> +			acp_reg_writel(0, amd_manager->mmio + ACP_SW_CLK_RESUME_CTRL);
>> +			ret = sdw_bus_exit_clk_stop(&amd_manager->bus);
>> +			if (ret < 0)
>> +				dev_err(amd_manager->dev, "bus failed to exit clock stop %d\n",
>> +					ret);
>> +			amd_manager->clk_stopped = false;
>> +		}
>> +	}
>> +	if (amd_manager->clk_stopped) {
>> +		dev_err(amd_manager->dev, "SDW%x clock stop exit failed\n", amd_manager->instance);
>> +		return 0;
>> +	}
>> +	dev_dbg(amd_manager->dev, "SDW%x clock stop exit successful\n", amd_manager->instance);
>> +	return 0;