Re: [PATCH v3 3/4] iio: health: max30102: do not use internal iio_dev lock

[Miquel Raynal <miquel.raynal@bootlin.com>]

Hi Nuno,

nuno.sa@analog.com wrote on Wed, 12 Oct 2022 17:16:19 +0200:

> The pattern used in this device does not quite fit in the
> iio_device_claim_direct_mode() typical usage. In this case, we want to
> know if we are in buffered mode or not to know if the device is powered
> (buffer mode) or not. And depending on that max30102_get_temp() will
> power on the device if needed. Hence, in order to keep the same
> functionality, we try to:
> 
> 1. Claim Buffered mode;
> 2: If 1) succeeds call max30102_get_temp() without powering on the
>    device;
> 3: Release Buffered mode;
> 4: If 1) fails, Claim Direct mode;
> 5: If 4) succeeds call max30102_get_temp() with powering on the device;
> 6: Release Direct mode;
> 7: If 4) fails, goto to 1) and try again.
> 
> This dance between buffered and direct mode is not particularly pretty
> (as well as the loop introduced by the goto statement) but it does allow
> us to get rid of the mlock usage while keeping the same behavior.

What about adding a TODO comment saying something like: "this comes
from static analysis and helped dropping mlock access, but someone with
the device needs to figure out if we can simplify this dance"? Because
the reason behind all this is that we don't want to risk breaking the
driver, but perhaps a simpler approach would work, right?

> 
> Signed-off-by: Nuno Sá <nuno.sa@analog.com>
> ---
>  drivers/iio/health/max30102.c | 19 +++++++++++++++----
>  1 file changed, 15 insertions(+), 4 deletions(-)
> 
> diff --git a/drivers/iio/health/max30102.c
> b/drivers/iio/health/max30102.c index 437298a29f2d..66df4aaa31a7
> 100644 --- a/drivers/iio/health/max30102.c
> +++ b/drivers/iio/health/max30102.c
> @@ -477,12 +477,23 @@ static int max30102_read_raw(struct iio_dev
> *indio_dev,
>  		 * Temperature reading can only be acquired when not
> in
>  		 * shutdown; leave shutdown briefly when buffer not
> running */
> -		mutex_lock(&indio_dev->mlock);
> -		if (!iio_buffer_enabled(indio_dev))
> +any_mode_retry:
> +		if (iio_device_claim_buffer_mode(indio_dev)) {
> +			/*
> +			 * This one is a *bit* hacky. If we cannot
> claim buffer
> +			 * mode, then try direct mode so that we
> make sure
> +			 * things cannot concurrently change. And we
> just keep
> +			 * trying until we get one of the modes...
> +			 */
> +			if (iio_device_claim_direct_mode(indio_dev))
> +				goto any_mode_retry;
> +
>  			ret = max30102_get_temp(data, val, true);
> -		else
> +			iio_device_release_direct_mode(indio_dev);
> +		} else {
>  			ret = max30102_get_temp(data, val, false);
> -		mutex_unlock(&indio_dev->mlock);
> +			iio_device_release_buffer_mode(indio_dev);
> +		}
>  		if (ret)
>  			return ret;
>  


Thanks,
Miquèl

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel