linux-kernel - Re: [PATCH net-next v3 1/2] nfc: llcp_core: Hold a ref to llcp_local->dev when holding a ref to llcp

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: <fd709885-c489-4f84-83ab-53cfb4920094@linaro.org>
Date:   Tue, 5 Dec 2023 17:40:00 +0100
From:   Krzysztof Kozlowski <krzysztof.kozlowski@...aro.org>
To:     Siddh Raman Pant <code@...dh.me>,
        "David S. Miller" <davem@...emloft.net>,
        Eric Dumazet <edumazet@...gle.com>,
        Jakub Kicinski <kuba@...nel.org>,
        Paolo Abeni <pabeni@...hat.com>,
        Suman Ghosh <sumang@...vell.com>
Cc:     netdev@...r.kernel.org, linux-kernel@...r.kernel.org,
        syzbot+bbe84a4010eeea00982d@...kaller.appspotmail.com
Subject: Re: [PATCH net-next v3 1/2] nfc: llcp_core: Hold a ref to
 llcp_local->dev when holding a ref to llcp_local

On 04/12/2023 14:08, Siddh Raman Pant wrote:
> llcp_sock_sendmsg() calls nfc_llcp_send_ui_frame() which in turn calls
> nfc_alloc_send_skb(), which accesses the nfc_dev from the llcp_sock for
> getting the headroom and tailroom needed for skb allocation.
> 
> Parallelly the nfc_dev can be freed, as the refcount is decreased via
> nfc_free_device(), leading to a UAF reported by Syzkaller, which can
> be summarized as follows:
> 
> (1) llcp_sock_sendmsg() -> nfc_llcp_send_ui_frame()
> 	-> nfc_alloc_send_skb() -> Dereference *nfc_dev
> (2) virtual_ncidev_close() -> nci_free_device() -> nfc_free_device()
> 	-> put_device() -> nfc_release() -> Free *nfc_dev
> 
> When a reference to llcp_local is acquired, we do not acquire the same
> for the nfc_dev. This leads to freeing even when the llcp_local is in
> use, and this is the case with the UAF described above too.
> 
> Thus, when we acquire a reference to llcp_local, we should acquire a
> reference to nfc_dev, and release the references appropriately later.
> 
> References for llcp_local is initialized in nfc_llcp_register_device()
> (which is called by nfc_register_device()). Thus, we should acquire a
> reference to nfc_dev there.
> 
> nfc_unregister_device() calls nfc_llcp_unregister_device() which in
> turn calls nfc_llcp_local_put(). Thus, the reference to nfc_dev is
> appropriately released later.
> 
> Reported-and-tested-by: syzbot+bbe84a4010eeea00982d@...kaller.appspotmail.com
> Closes: https://syzkaller.appspot.com/bug?extid=bbe84a4010eeea00982d
> Fixes: c7aa12252f51 ("NFC: Take a reference on the LLCP local pointer when creating a socket")
> Signed-off-by: Siddh Raman Pant <code@...dh.me>
> ---
>  net/nfc/llcp_core.c | 24 ++++++++++++++++++++++--
>  1 file changed, 22 insertions(+), 2 deletions(-)
> 
> diff --git a/net/nfc/llcp_core.c b/net/nfc/llcp_core.c
> index 1dac28136e6a..9d45ce6dcdca 100644
> --- a/net/nfc/llcp_core.c
> +++ b/net/nfc/llcp_core.c
> @@ -145,6 +145,9 @@ static void nfc_llcp_socket_release(struct nfc_llcp_local *local, bool device,
>  
>  static struct nfc_llcp_local *nfc_llcp_local_get(struct nfc_llcp_local *local)
>  {
> +	if (!nfc_get_device(local->dev->idx))
> +		return NULL;
> +
>  	kref_get(&local->ref);
>  
>  	return local;
> @@ -180,6 +183,7 @@ int nfc_llcp_local_put(struct nfc_llcp_local *local)
>  	if (local == NULL)
>  		return 0;
>  
> +	nfc_put_device(local->dev);

Mismatched order with get. Unwinding is always in reversed order. Or
maybe other order is here on purpose? Then it needs to be explained.

>  	return kref_put(&local->ref, local_release);
>  }
>  
> @@ -959,8 +963,18 @@ static void nfc_llcp_recv_connect(struct nfc_llcp_local *local,
>  	}
>  
>  	new_sock = nfc_llcp_sock(new_sk);
> -	new_sock->dev = local->dev;
> +
>  	new_sock->local = nfc_llcp_local_get(local);
> +	if (!new_sock->local) {

There is already an cleanup path/label, so extend it. Existing code
needs some improvements in that matter as well.

> +		reason = LLCP_DM_REJ;
> +		release_sock(&sock->sk);
> +		sock_put(&sock->sk);
> +		sock_put(&new_sock->sk);
> +		nfc_llcp_sock_free(new_sock);
> +		goto fail;
> +	}
> +
> +	new_sock->dev = local->dev;
>  	new_sock->rw = sock->rw;
>  	new_sock->miux = sock->miux;
>  	new_sock->nfc_protocol = sock->nfc_protocol;
> @@ -1597,7 +1611,13 @@ int nfc_llcp_register_device(struct nfc_dev *ndev)
>  	if (local == NULL)
>  		return -ENOMEM;
>  
> -	local->dev = ndev;
> +	/* Hold a reference to the device. */

That's obvious. Instead write something not obvious - why you call
nfc_get_device() while not incrementing reference to llcp_local.

> +	local->dev = nfc_get_device(ndev->idx);

This looks confusing. If you can access ndev->idx, then ndev reference
was already increased. In such case iterating through all devices to
find it, is unnecessary and confusing.



Best regards,
Krzysztof