netdev - Re: [PATCH v1 net-next 4/6] socket: Remove kernel socket conversion except for net/rds/.

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: <7a965a97-a6d0-462f-b7dd-8833605ea7c9@redhat.com>
Date: Thu, 22 May 2025 10:55:47 +0200
From: Paolo Abeni <pabeni@...hat.com>
To: Kuniyuki Iwashima <kuniyu@...zon.com>,
 "David S. Miller" <davem@...emloft.net>, Eric Dumazet <edumazet@...gle.com>,
 Jakub Kicinski <kuba@...nel.org>, Willem de Bruijn <willemb@...gle.com>
Cc: Simon Horman <horms@...nel.org>, Kuniyuki Iwashima <kuni1840@...il.com>,
 netdev@...r.kernel.org, Chuck Lever <chuck.lever@...cle.com>,
 Jeff Layton <jlayton@...nel.org>, linux-nfs@...r.kernel.org,
 Keith Busch <kbusch@...nel.org>, Jens Axboe <axboe@...nel.dk>,
 Christoph Hellwig <hch@....de>, Wenjia Zhang <wenjia@...ux.ibm.com>,
 Jan Karcher <jaka@...ux.ibm.com>, Steve French <sfrench@...ba.org>,
 "linux-rdma@...r.kernel.org" <linux-rdma@...r.kernel.org>,
 linux-nvme@...ts.infradead.org, Matthieu Baerts <matttbe@...nel.org>,
 MPTCP Linux <mptcp@...ts.linux.dev>
Subject: Re: [PATCH v1 net-next 4/6] socket: Remove kernel socket conversion
 except for net/rds/.

On 5/17/25 5:50 AM, Kuniyuki Iwashima wrote:
> Since commit 26abe14379f8 ("net: Modify sk_alloc to not reference
> count the netns of kernel sockets."), TCP kernel socket has caused
> many UAF.
> 
> We have converted such sockets to hold netns refcnt, and we have
> the same pattern in cifs, mptcp, nvme, rds, smc, and sunrpc.
> 
>   __sock_create_kern(..., &sock);
>   sk_net_refcnt_upgrade(sock->sk);
> 
> Let's drop the conversion and use sock_create_kern() instead.
> 
> The changes for cifs, mptcp, nvme, and smc are straightforward.
> 
> For sunrpc, we call sock_create_net() for IPPROTO_TCP only and still
> call __sock_create_kern() for others.
> 
> For rds, we cannot drop sk_net_refcnt_upgrade() for accept()ed
> sockets.
> 
> Signed-off-by: Kuniyuki Iwashima <kuniyu@...zon.com>

This LGTM, but is touching a few other subsystems, it would be great to
collect acks from the relevant maintainers: I'm adding a few CCs.

Direct link to the series:

https://lore.kernel.org/all/20250517035120.55560-1-kuniyu@amazon.com/#t

> diff --git a/fs/smb/client/connect.c b/fs/smb/client/connect.c
> index 37a2ba38f10e..c7b4f5a7cca1 100644
> --- a/fs/smb/client/connect.c
> +++ b/fs/smb/client/connect.c
> @@ -3348,21 +3348,14 @@ generic_ip_connect(struct TCP_Server_Info *server)
>  		socket = server->ssocket;
>  	} else {
>  		struct net *net = cifs_net_ns(server);
> -		struct sock *sk;
>  
> -		rc = __sock_create_kern(net, sfamily, SOCK_STREAM,
> -					IPPROTO_TCP, &server->ssocket);
> +		rc = sock_create_kern(net, sfamily, SOCK_STREAM,
> +				      IPPROTO_TCP, &server->ssocket);
>  		if (rc < 0) {
>  			cifs_server_dbg(VFS, "Error %d creating socket\n", rc);
>  			return rc;
>  		}
>  
> -		sk = server->ssocket->sk;
> -		__netns_tracker_free(net, &sk->ns_tracker, false);
> -		sk->sk_net_refcnt = 1;
> -		get_net_track(net, &sk->ns_tracker, GFP_KERNEL);
> -		sock_inuse_add(net, 1);

AFAICS the above implicitly adds a missing net_passive_dec(net), which
in turns looks like a separate bugfix. What about adding a separate
patch introducing that line? Could be in the same series to simplify the
processing.

Thanks,

Paolo