mbox series

[v2,0/4] Switch on IMX577 on RB5 with a new CCI fix

Message ID 20220524140207.2758605-1-bryan.odonoghue@linaro.org
Headers show
Series Switch on IMX577 on RB5 with a new CCI fix | expand

Message

Bryan O'Donoghue May 24, 2022, 2:02 p.m. UTC 
V2:

- Adds fix for bug identified by Vladimir
  The CCI i2c_adapter_add() and pm_runtime_enable() are racy.
  This is a generic problem not related to the rb5/imx577 but, for the sake
  of our conversation/review's context I'll add it into this series.
- Include Vladimir's camcc patch
  I've also opted to include Vladimir's disable of camcc to make the enable
  of it in my patchset logical.
- Move address/size cells Konrad
- Remove newline in pin definitions - Konrad
- Remove sensor 'status = "okay"' - Konrad
- Add comment to qrb5165-rb5.dts re: imx412 and imx577 difference - Konrad
- Move pin definitions to 8250 dtsi - Vladimir
- Drop power domain from sensor definition - Vladimir
- Correct to "add to cam2" not "cam1" in commit log - bod

To make verification of the CCI race eaiser I've provided a defconfig both
with and without modules enabled.

Link: https://git.linaro.org/people/bryan.odonoghue/kernel.git/log/?h=linux-next-24-05-22%2bimx577-rb5
Link: https://git.linaro.org/people/bryan.odonoghue/kernel.git/log/?h=linux-next-24-05-22%2bimx577-rb5-compiled-in

git diff linaro/linux-next-22-05-22+imx577-rb5 linaro/linux-next-24-05-22+imx577-rb5

V1:
Linux-next now has everything we need to switch on this sensor both in the
qcom DTS and in the imx412 driver.

After this, no further dts or driver work is required to capture images on
the RB5.

Here's a bootable linux-next with a kernel config. I added Vladimir's
power-domain changes on-top to verify nothing breaks for me.

https://git.linaro.org/people/bryan.odonoghue/kernel.git/log/?h=linux-next-18-05-22%2bimx577-rb5

Bryan O'Donoghue (3):
  i2c: qcom-cci: Fix ordering of pm_runtime_xx and i2c_add_adapter
  arm64: dts: qcom: sm8250: camss: Define ports address/size cells
  arm64: dts: qcom: qrb5165-rb5: Enable the IMX577 on cam2

Vladimir Zapolskiy (1):
  arm64: dts: qcom: sm8250: Disable camcc by default

 arch/arm64/boot/dts/qcom/qrb5165-rb5.dts | 60 ++++++++++++++++++++++++
 arch/arm64/boot/dts/qcom/sm8250.dtsi     | 39 +++++++++++++++
 drivers/i2c/busses/i2c-qcom-cci.c        | 14 ++++--
 3 files changed, 108 insertions(+), 5 deletions(-)

Comments

Vladimir Zapolskiy May 25, 2022, 7:27 a.m. UTC | #1 
Hi Bryan,

On 5/24/22 17:02, Bryan O'Donoghue wrote:
> When we compile-in the CCI along with the imx412 driver and run on the RB5
> we see that i2c_add_adapter() causes the probe of the imx412 driver to
> happen.
> 
> This probe tries to perform an i2c xfer() and the xfer() in i2c-qcom-cci.c
> fails on pm_runtime_get() because the i2c-qcom-cci.c::probe() function has
> not completed to pm_runtime_enable(dev).
> 
> Fix this sequence by ensuring pm_runtime_xxx() calls happen prior to adding
> the i2c adapter.
> 

it's a proper root cause disclosure and the fix, thank you.

> Fixes: e517526195de ("i2c: Add Qualcomm CCI I2C driver")
> Reported-by: Vladimir Zapolskiy <vladimir.zapolskiy@linaro.org>
> Signed-off-by: Bryan O'Donoghue <bryan.odonoghue@linaro.org>
> ---
>   drivers/i2c/busses/i2c-qcom-cci.c | 14 +++++++++-----
>   1 file changed, 9 insertions(+), 5 deletions(-)
> 
> diff --git a/drivers/i2c/busses/i2c-qcom-cci.c b/drivers/i2c/busses/i2c-qcom-cci.c
> index 5c7cc862f08f..90d02effeae9 100644
> --- a/drivers/i2c/busses/i2c-qcom-cci.c
> +++ b/drivers/i2c/busses/i2c-qcom-cci.c
> @@ -638,6 +638,11 @@ static int cci_probe(struct platform_device *pdev)
>   	if (ret < 0)
>   		goto error;
>   
> +	pm_runtime_set_autosuspend_delay(dev, MSEC_PER_SEC);
> +	pm_runtime_use_autosuspend(dev);
> +	pm_runtime_set_active(dev);
> +	pm_runtime_enable(dev);
> +
>   	for (i = 0; i < cci->data->num_masters; i++) {
>   		if (!cci->master[i].cci)
>   			continue;
> @@ -649,14 +654,13 @@ static int cci_probe(struct platform_device *pdev)
>   		}
>   	}
>   
> -	pm_runtime_set_autosuspend_delay(dev, MSEC_PER_SEC);
> -	pm_runtime_use_autosuspend(dev);
> -	pm_runtime_set_active(dev);
> -	pm_runtime_enable(dev);
> -
>   	return 0;
>   
>   error_i2c:
> +	pm_runtime_put(dev);

Here pm_runtime_put(dev) should be removed, there is no pm_runtime_get*(dev) above,
otherwise PM usage counter underflow issues are expected.

> +	pm_runtime_disable(dev);
> +	pm_runtime_dont_use_autosuspend(dev);
> +
>   	for (--i ; i >= 0; i--) {
>   		if (cci->master[i].cci) {
>   			i2c_del_adapter(&cci->master[i].adap);

With the correction stated above please feel free to add my tags:

Reviewed-by: Vladimir Zapolskiy <vladimir.zapolskiy@linaro.org>
Tested-by: Vladimir Zapolskiy <vladimir.zapolskiy@linaro.org>

--
Best wishes,
Vladimir