Re: [PATCH] usb: gadget: f_phonet: fix use-after-free in pn_bind

On Tue, Aug 04, 2026 at 11:33:41AM +0800, Nguyen Quang Le Kien wrote:
> phonet_free_inst() checks opts->bound to decide whether to call
> gphonet_cleanup() or free_netdev(). If opts->bound is false,
> free_netdev(opts->net) is called immediately.
> 
> However, pn_bind() can race with phonet_free_inst() when a configfs
> entry is removed while binding is in progress. pn_bind() reads
> opts->bound and calls gphonet_set_gadget(opts->net, ...) before
> setting opts->bound = true. If phonet_free_inst() runs concurrently
> after the !opts->bound check in pn_bind() but before opts->bound is
> set, it will free opts->net, causing a use-after-free when pn_bind()
> subsequently writes to net->dev.parent via gphonet_set_gadget().
> 
> Fix this by adding a mutex to f_phonet_opts and holding it in both
> pn_bind() and phonet_free_inst() when accessing opts->bound and
> opts->net. This is consistent with how other gadget functions such
> as f_eem protect their opts->bound flag.
> 
> Fixes: 00a2430ff07d ("usb: gadget: Gadget directory cleanup - group usb functions")
> Reported-by: [email protected]
> Closes: https://syzkaller.appspot.com/bug?extid=098999e05b6b877c01b3
> Signed-off-by: Nguyen Quang Le Kien <[email protected]>
> ---
>  drivers/usb/gadget/function/f_phonet.c | 16 ++++++++--------
>  drivers/usb/gadget/function/u_phonet.h |  1 +
>  2 files changed, 9 insertions(+), 8 deletions(-)
> 
> diff --git a/drivers/usb/gadget/function/f_phonet.c b/drivers/usb/gadget/function/f_phonet.c
> index b1ee9a7c2..b7eae14e8 100644
> --- a/drivers/usb/gadget/function/f_phonet.c
> +++ b/drivers/usb/gadget/function/f_phonet.c
> @@ -499,20 +499,17 @@ static int pn_bind(struct usb_configuration *c, struct usb_function *f)
>  
>  	phonet_opts = container_of(f->fi, struct f_phonet_opts, func_inst);
>  
> -	/*
> -	 * in drivers/usb/gadget/configfs.c:configfs_composite_bind()
> -	 * configurations are bound in sequence with list_for_each_entry,
> -	 * in each configuration its functions are bound in sequence
> -	 * with list_for_each_entry, so we assume no race condition
> -	 * with regard to phonet_opts->bound access
> -	 */
> +	mutex_lock(&phonet_opts->lock);

So the coomment lied?

And why not use guard()?  Was an LLM used to generate this changelog and
code change?


>  	if (!phonet_opts->bound) {
>  		gphonet_set_gadget(phonet_opts->net, gadget);
>  		status = gphonet_register_netdev(phonet_opts->net);
> -		if (status)
> +		if (status) {
> +			mutex_unlock(&phonet_opts->lock);
>  			return status;
> +		}
>  		phonet_opts->bound = true;
>  	}
> +	mutex_unlock(&phonet_opts->lock);
>  
>  	/* Reserve interface IDs */
>  	status = usb_interface_id(c, f);
> @@ -621,10 +618,12 @@ static void phonet_free_inst(struct usb_function_instance *f)
>  	struct f_phonet_opts *opts;
>  
>  	opts = container_of(f, struct f_phonet_opts, func_inst);
> +	mutex_lock(&opts->lock);
>  	if (opts->bound)
>  		gphonet_cleanup(opts->net);
>  	else
>  		free_netdev(opts->net);
> +	mutex_unlock(&opts->lock);
>  	kfree(opts);
>  }
>  
> @@ -636,6 +635,7 @@ static struct usb_function_instance *phonet_alloc_inst(void)
>  	if (!opts)
>  		return ERR_PTR(-ENOMEM);
>  
> +	mutex_init(&opts->lock);
>  	opts->func_inst.free_func_inst = phonet_free_inst;
>  	opts->net = gphonet_setup_default();
>  	if (IS_ERR(opts->net)) {
> diff --git a/drivers/usb/gadget/function/u_phonet.h b/drivers/usb/gadget/function/u_phonet.h
> index ff62ca22c..4666413fc 100644
> --- a/drivers/usb/gadget/function/u_phonet.h
> +++ b/drivers/usb/gadget/function/u_phonet.h
> @@ -13,6 +13,7 @@
>  
>  struct f_phonet_opts {
>  	struct usb_function_instance func_inst;
> +	struct mutex lock;

No comment as to what this lock protects?

thanks,

greg k-h