On Sun 2019-10-06 19:20:07, Greg Kroah-Hartman wrote:
> From: Jarkko Sakkinen <jarkko.sakkinen@linux.intel.com>
> 
> commit 2677ca98ae377517930c183248221f69f771c921 upstream
> 
> Use tpm_try_get_ops() in tpm-sysfs.c so that we can consider moving
> other decorations (locking, localities, power management for example)
> inside it. This direction can be of course taken only after other call
> sites for tpm_transmit() have been treated in the same way.

This changes locking completely:

> @@ -244,10 +274,12 @@ static ssize_t cancel_store(struct device *dev, struct device_attribute *attr,
>  			    const char *buf, size_t count)
>  {
>  	struct tpm_chip *chip = to_tpm_chip(dev);
> -	if (chip == NULL)
> +
> +	if (tpm_try_get_ops(chip))
>  		return 0;
>  
>  	chip->ops->cancel(chip);
> +	tpm_put_ops(chip);
>  	return count;
>  }
>  static DEVICE_ATTR_WO(cancel);

For example this did not have any locking, and is now protected by

        get_device(&chip->dev);

        down_read(&chip->ops_sem);

. Is that intended? Is this known to fix any bugs?

Best regards,
									Pavel

-- 
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html