linux-kernel - Re: [PATCH v2 1/5] extcon: Make the allocation and freeing to be private calls

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <806934f5-2992-6085-ab75-909e11e7a95c@gmail.com>
Date:   Fri, 7 Apr 2023 04:19:08 +0900
From:   Chanwoo Choi <cwchoi00@...il.com>
To:     Andy Shevchenko <andriy.shevchenko@...ux.intel.com>,
        Chanwoo Choi <cw00.choi@...sung.com>,
        linux-kernel@...r.kernel.org
Cc:     MyungJoo Ham <myungjoo.ham@...sung.com>
Subject: Re: [PATCH v2 1/5] extcon: Make the allocation and freeing to be
 private calls

Hi,

On 23. 4. 6. 00:27, Andy Shevchenko wrote:
> The extcon_dev_allocate() and extcon_dev_free() are not used
> outside of the extcon framework. Moreover, the struct extcon_dev
> can't be filled outside of the framework either after allocation.
> The registration part, for instance, requires a parent device to
> be set and that's done in the devm_extcon_dev_allocate() wrapper.
> 
> Taking the above into account, sumply move the mentioned APIs to
> the private headers.
> 
> Alternatively, the pointer to the parent device can be added to
> the extcon_dev_allocate(), but since there are no users and magnitude
> of the change it makes a little sense to go this way.
> 
> Signed-off-by: Andy Shevchenko <andriy.shevchenko@...ux.intel.com>

The role of extcon_dev_allocate is not that supporting to touch
the internal data of struct extcon_dev. Some driver needs
the removing sequence between extcon device and their used resource
like irq/opp/regulator or others.

When they need some sequence of resource freeing, it should be supported.
I think that there is no benefit to move into driver/extcon. It might
limit the various use-case of extcon_dev_allocate.


> ---
>  drivers/extcon/extcon.h         | 4 ++++
>  include/linux/extcon-provider.h | 9 ---------
>  2 files changed, 4 insertions(+), 9 deletions(-)
> 
> diff --git a/drivers/extcon/extcon.h b/drivers/extcon/extcon.h
> index 15616446140d..49e4ed9f6450 100644
> --- a/drivers/extcon/extcon.h
> +++ b/drivers/extcon/extcon.h
> @@ -63,4 +63,8 @@ struct extcon_dev {
>  	struct device_attribute *d_attrs_muex;
>  };
>  
> +/* Following APIs allocate/free the memory of the extcon device. */
> +struct extcon_dev *extcon_dev_allocate(const unsigned int *cable);
> +void extcon_dev_free(struct extcon_dev *edev);
> +
>  #endif /* __LINUX_EXTCON_INTERNAL_H__ */
> diff --git a/include/linux/extcon-provider.h b/include/linux/extcon-provider.h
> index fa70945f4e6b..db474ae3c711 100644
> --- a/include/linux/extcon-provider.h
> +++ b/include/linux/extcon-provider.h
> @@ -25,8 +25,6 @@ void devm_extcon_dev_unregister(struct device *dev,
>  				struct extcon_dev *edev);
>  
>  /* Following APIs allocate/free the memory of the extcon device. */
> -struct extcon_dev *extcon_dev_allocate(const unsigned int *cable);
> -void extcon_dev_free(struct extcon_dev *edev);
>  struct extcon_dev *devm_extcon_dev_allocate(struct device *dev,
>  				const unsigned int *cable);
>  void devm_extcon_dev_free(struct device *dev, struct extcon_dev *edev);
> @@ -78,13 +76,6 @@ static inline int devm_extcon_dev_register(struct device *dev,
>  static inline void devm_extcon_dev_unregister(struct device *dev,
>  				struct extcon_dev *edev) { }
>  
> -static inline struct extcon_dev *extcon_dev_allocate(const unsigned int *cable)
> -{
> -	return ERR_PTR(-ENOSYS);
> -}
> -
> -static inline void extcon_dev_free(struct extcon_dev *edev) { }
> -
>  static inline struct extcon_dev *devm_extcon_dev_allocate(struct device *dev,
>  				const unsigned int *cable)
>  {

-- 
Best Regards,
Samsung Electronics
Chanwoo Choi