linux-kernel - Re: selftests: net: l2tp.sh regression starting with 6.1-rc1

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: <ZCRYpDehyDxsrnfi@debian>
Date:   Wed, 29 Mar 2023 17:26:28 +0200
From:   Guillaume Nault <gnault@...hat.com>
To:     "Drewek, Wojciech" <wojciech.drewek@...el.com>
Cc:     Andrea Righi <andrea.righi@...onical.com>,
        "David S. Miller" <davem@...emloft.net>,
        Eric Dumazet <edumazet@...gle.com>,
        Jakub Kicinski <kuba@...nel.org>,
        Paolo Abeni <pabeni@...hat.com>, Shuah Khan <shuah@...nel.org>,
        "netdev@...r.kernel.org" <netdev@...r.kernel.org>,
        "linux-kselftest@...r.kernel.org" <linux-kselftest@...r.kernel.org>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>
Subject: Re: selftests: net: l2tp.sh regression starting with 6.1-rc1

On Wed, Mar 29, 2023 at 02:16:37PM +0000, Drewek, Wojciech wrote:
> Hi,
> 
> Modifying UAPI was not a good idea although the patch should not break userspace (related discussion [1]).
> We could revert this patch with one additional change (include l2tp.h in net/sched/cls_flower.c) but then again,
> modifying UAPI. This patch was mostly cosmetic anyway.
> Second option is to try to fix the automatic load. I'm not an expert but I think
> MODULE_ALIAS_NET_PF_PROTO macro is somehow responsible for that. I noticed some comments saying that
> "__stringify doesn't like enums" (this macro is using _stringify) and my patch defined IPPROTO_L2TP in enum.
> We can just replace IPPROTO_L2TP with 115 (where this macro is used) in order to fix this.
> I'm going to give it a try and will let you know.

Yes, the modules aliases now have symbolic names:

$ modinfo l2tp_ip l2tp_ip6 | grep alias
alias:          net-pf-2-proto-IPPROTO_L2TP
alias:          net-pf-2-proto-2-type-IPPROTO_L2TP
alias:          net-pf-10-proto-IPPROTO_L2TP
alias:          net-pf-10-proto-2-type-IPPROTO_L2TP

Therefore, 'request_module("net-pf-%d-proto-%d-type-%d")' can't find
them.

My personal preference is for the second option: fix module loading by
using plain numbers in MODULE_ALIAS_*. We can always keep the symbolic
names in comments.

---- >8 ----

diff --git a/net/l2tp/l2tp_ip.c b/net/l2tp/l2tp_ip.c
index 4db5a554bdbd..afe94a390ef0 100644
--- a/net/l2tp/l2tp_ip.c
+++ b/net/l2tp/l2tp_ip.c
@@ -680,5 +680,5 @@ MODULE_VERSION("1.0");
 /* Use the value of SOCK_DGRAM (2) directory, because __stringify doesn't like
  * enums
  */
-MODULE_ALIAS_NET_PF_PROTO_TYPE(PF_INET, 2, IPPROTO_L2TP);
-MODULE_ALIAS_NET_PF_PROTO(PF_INET, IPPROTO_L2TP);
+MODULE_ALIAS_NET_PF_PROTO_TYPE(PF_INET, 2, 115 /* IPPROTO_L2TP */);
+MODULE_ALIAS_NET_PF_PROTO(PF_INET, 115 /* IPPROTO_L2TP */);
diff --git a/net/l2tp/l2tp_ip6.c b/net/l2tp/l2tp_ip6.c
index 2478aa60145f..65d106b41951 100644
--- a/net/l2tp/l2tp_ip6.c
+++ b/net/l2tp/l2tp_ip6.c
@@ -809,5 +809,5 @@ MODULE_VERSION("1.0");
 /* Use the value of SOCK_DGRAM (2) directory, because __stringify doesn't like
  * enums
  */
-MODULE_ALIAS_NET_PF_PROTO_TYPE(PF_INET6, 2, IPPROTO_L2TP);
-MODULE_ALIAS_NET_PF_PROTO(PF_INET6, IPPROTO_L2TP);
+MODULE_ALIAS_NET_PF_PROTO_TYPE(PF_INET6, 2, 115 /* IPPROTO_L2TP */);
+MODULE_ALIAS_NET_PF_PROTO(PF_INET6, 115 /* IPPROTO_L2TP */);