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] [day] [month] [year] [list]
Message-ID: <20230310143124.36607bb3@kernel.org>
Date:   Fri, 10 Mar 2023 14:31:24 -0800
From:   Jakub Kicinski <kuba@...nel.org>
To:     Chuck Lever III <chuck.lever@...cle.com>
Cc:     Paolo Abeni <pabeni@...hat.com>,
        Eric Dumazet <edumazet@...gle.com>,
        "open list:NETWORKING [GENERAL]" <netdev@...r.kernel.org>,
        "kernel-tls-handshake@...ts.linux.dev" 
        <kernel-tls-handshake@...ts.linux.dev>,
        John Haxby <john.haxby@...cle.com>
Subject: Re: [PATCH v6 2/2] net/tls: Add kernel APIs for requesting a
 TLSv1.3 handshake

On Fri, 10 Mar 2023 15:25:49 +0000 Chuck Lever III wrote:
> When TLS handshake consumers are built-in but TLS is built
> as a module, these API calls become undefined references:
> 
> ld: net/sunrpc/xprtsock.o: in function `xs_tls_handshake_sync':
> /home/cel/src/linux/linux/net/sunrpc/xprtsock.c:2560: undefined reference to `tls_client_hello_x509'
> ld: /home/cel/src/linux/linux/net/sunrpc/xprtsock.c:2552: undefined reference to `tls_client_hello_anon'
> ld: /home/cel/src/linux/linux/net/sunrpc/xprtsock.c:2572: undefined reference to `tls_handshake_cancel'
> ld: net/sunrpc/xprtsock.o: in function `xs_reset_transport':
> /home/cel/src/linux/linux/net/sunrpc/xprtsock.c:1257: undefined reference to `tls_handshake_cancel'
> ld: net/sunrpc/svcsock.o: in function `svc_tcp_handshake':
> /home/cel/src/linux/linux/net/sunrpc/svcsock.c:449: undefined reference to `tls_server_hello_x509'
> ld: /home/cel/src/linux/linux/net/sunrpc/svcsock.c:458: undefined reference to `tls_handshake_cancel'
> 
> This was fine for our prototype: we just don't build it that
> way. But it won't work long-term.
> 
> What is the approach that would be most acceptable to address
> this?

Best to stick to kconfig dependencies enforcing handshake is also built
in if consumers are. If there's a good reason to support loose
dependencies we usually do a built in stub implementation (function
pointer to a set of ops is built in, the module sets it when it loads).

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ