[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <YcL2KpAFPOuWYlei@hovoldconsulting.com>
Date: Wed, 22 Dec 2021 10:55:54 +0100
From: Johan Hovold <johan@...nel.org>
To: Greg Kroah-Hartman <gregkh@...uxfoundation.org>
Cc: Rafał Miłecki <zajec5@...il.com>,
linux-kernel@...r.kernel.org
Subject: Re: [PATCH] driver core: platform: document registration-failure
requirement
On Wed, Dec 22, 2021 at 09:52:29AM +0100, Johan Hovold wrote:
> Add an explicit comment to document that the reference initialised by
> platform_device_register() needs to be released by a call to
> platform_device_put() also when registration fails (cf.
> device_register()).
>
> Signed-off-by: Johan Hovold <johan@...nel.org>
> ---
> drivers/base/platform.c | 4 ++++
> 1 file changed, 4 insertions(+)
>
> diff --git a/drivers/base/platform.c b/drivers/base/platform.c
> index 598acf93a360..38fdcbdba4a4 100644
> --- a/drivers/base/platform.c
> +++ b/drivers/base/platform.c
> @@ -762,6 +762,10 @@ EXPORT_SYMBOL_GPL(platform_device_del);
> /**
> * platform_device_register - add a platform-level device
> * @pdev: platform device we're adding
> + *
> + * NOTE: _Never_ directly free @pdev after calling this function, even if it
> + * returned an error! Always use platform_put_device() to give up the
This should have been platform_device_put().
Greg, I'll send a v2 for you to consider since this documents the
current behaviour even if you wanted to look into ways to change this.
> + * reference initialised in this function instead.
> */
> int platform_device_register(struct platform_device *pdev)
> {
Johan
Powered by blists - more mailing lists