Re: [PATCH v3] mtd: rawnand: atmel: fix possible object reference leak

[Miquel Raynal]

Hi Wen,

For the next version can you please post a series with the three
commits which are fixing the same reference leak? No need to add a
cover letter though.

Wen Yang <yellowriver2010@xxxxxxxxxxx> wrote on Tue, 5 Feb 2019
14:32:41 +0000:

> of_find_device_by_node() takes a reference to the struct device
> when it finds a match via get_device, there is no need to call
> get_device() twice.
> We also should make sure to drop the reference to the device
> taken by of_find_device_by_node() on driver unbind.
> 
> Fixes: f88fc122cc34 ("mtd: nand: Cleanup/rework the atmel_nand driver")
> Signed-off-by: Wen Yang <yellowriver2010@xxxxxxxxxxx>
> Suggested-by: Boris Brezillon <bbrezillon@xxxxxxxxxx>
> Reviewed-by: Boris Brezillon <bbrezillon@xxxxxxxxxx>
> Reviewed-by: Miquel Raynal <miquel.raynal@xxxxxxxxxxx>
> Acked-by: Miquel Raynal <miquel.raynal@xxxxxxxxxxx>
> Cc: Tudor Ambarus <tudor.ambarus@xxxxxxxxxxxxx>
> Cc: Boris Brezillon <bbrezillon@xxxxxxxxxx>
> Cc: Miquel Raynal <miquel.raynal@xxxxxxxxxxx>
> Cc: Richard Weinberger <richard@xxxxxx>
> Cc: David Woodhouse <dwmw2@xxxxxxxxxxxxx>
> Cc: Brian Norris <computersforpeace@xxxxxxxxx>
> Cc: Marek Vasut <marek.vasut@xxxxxxxxx>
> Cc: Nicolas Ferre <nicolas.ferre@xxxxxxxxxxxxx>
> Cc: Alexandre Belloni <alexandre.belloni@xxxxxxxxxxx>
> Cc: Ludovic Desroches <ludovic.desroches@xxxxxxxxxxxxx>
> Cc: linux-mtd@xxxxxxxxxxxxxxxxxxx
> Cc: linux-arm-kernel@xxxxxxxxxxxxxxxxxxx
> Cc: linux-kernel@xxxxxxxxxxxxxxx
> ---
>  drivers/mtd/nand/raw/atmel/pmecc.c | 18 ++++++++++++++----
>  1 file changed, 14 insertions(+), 4 deletions(-)
> 
> diff --git a/drivers/mtd/nand/raw/atmel/pmecc.c b/drivers/mtd/nand/raw/atmel/pmecc.c
> index 555a74e..1477368 100644
> --- a/drivers/mtd/nand/raw/atmel/pmecc.c
> +++ b/drivers/mtd/nand/raw/atmel/pmecc.c
> @@ -876,16 +876,22 @@ static struct atmel_pmecc *atmel_pmecc_get_by_node(struct device *userdev,
>  {
>  	struct platform_device *pdev;
>  	struct atmel_pmecc *pmecc, **ptr;
> +	int ret;
>  
>  	pdev = of_find_device_by_node(np);
> -	if (!pdev || !platform_get_drvdata(pdev))
> +	if (!pdev)
>  		return ERR_PTR(-EPROBE_DEFER);
> +	if (!platform_get_drvdata(pdev)) {
> +		ret = -EPROBE_DEFER;
> +		goto err_put_device;
> +	}
>  
>  	ptr = devres_alloc(devm_atmel_pmecc_put, sizeof(*ptr), GFP_KERNEL);
> -	if (!ptr)
> -		return ERR_PTR(-ENOMEM);
> +	if (!ptr) {
> +		ret = -ENOMEM;
> +		goto err_put_device;
> +	}
>  
> -	get_device(&pdev->dev);
>  	pmecc = platform_get_drvdata(pdev);

I just realized in the three cases, a first call
to platform_get_drvdata() is done to check if the returned pointer is
valid, and then a second one is done to actually retrieve the pointer.
Please avoid this repetition.


Thanks,
MiquÃl