Re: [PATCH RESEND] sparc: fix error handling in scan_one_device()

[Rob Gardner <rob.gardner@oracle.com>]

On 4/14/25 05:18, Ma Ke wrote:
> Once of_device_register() failed, we should call put_device() to
> decrement reference count for cleanup. Or it could cause memory leak.
> So fix this by calling put_device(), then the name can be freed in
> kobject_cleanup().

This seems ok but it's not clear why you delete "kfree(op)" here.

scan_one_device() allocates the memory for op via kzalloc, and in the 
case where of_device_register() fails, it sets op = NULL. This 
superficially looks like you *create* a memory leak here, since I don't 
see any other obvious reference to that kzalloc'd memory after 
scan_one_device() returns. Is the "op" pointer saved somewhere quietly 
by build_device_resources()?

I think you should still free the memory (that you allocated) after the 
call to put_device(), or explain why this is not necessary.


>
> As comment of device_add() says, 'if device_add() succeeds, you should
> call device_del() when you want to get rid of it. If device_add() has
> not succeeded, use only put_device() to drop the reference count'.
>
> Found by code review.
>
> Cc: stable@vger.kernel.org
> Fixes: cf44bbc26cf1 ("[SPARC]: Beginnings of generic of_device framework.")
> Signed-off-by: Ma Ke <make24@iscas.ac.cn>
> ---
>   arch/sparc/kernel/of_device_64.c | 2 +-
>   1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/arch/sparc/kernel/of_device_64.c b/arch/sparc/kernel/of_device_64.c
> index f98c2901f335..4272746d7166 100644
> --- a/arch/sparc/kernel/of_device_64.c
> +++ b/arch/sparc/kernel/of_device_64.c
> @@ -677,7 +677,7 @@ static struct platform_device * __init scan_one_device(struct device_node *dp,
>   
>   	if (of_device_register(op)) {
>   		printk("%pOF: Could not register of device.\n", dp);
> -		kfree(op);
> +		put_device(&op->dev);
>   		op = NULL;
>   	}
>