[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20220530070256.GA2517843@gauss3.secunet.de>
Date: Mon, 30 May 2022 09:02:56 +0200
From: Steffen Klassert <steffen.klassert@...unet.com>
To: Maciej Żenczykowski <maze@...gle.com>
CC: Linux NetDev <netdev@...r.kernel.org>,
Benedict Wong <benedictwong@...gle.com>,
Yan Yan <evitayan@...gle.com>
Subject: Re: 5.18 breaks Android net test PFKEY AddSA test case
On Sat, May 28, 2022 at 01:54:25AM -0700, Maciej Żenczykowski wrote:
> I've not gotten to the bottom of the root cause, since I'm hoping
> someone will know off the top of their head,
> why this might now be broken...
>
> ##### ./pf_key_test.py (13/25)
>
> E
> ======================================================================
> ERROR: testAddDelSa (__main__.PfKeyTest)
> ----------------------------------------------------------------------
> Traceback (most recent call last):
> File "./pf_key_test.py", line 42, in testAddDelSa
> pf_key.SADB_X_AALG_SHA2_256HMAC, ENCRYPTION_KEY)
> File "/aosp-tests/net/test/pf_key.py", line 254, in AddSa
> self.SendAndRecv(msg, self.PackPfKeyExtensions(extlist))
> File "/aosp-tests/net/test/pf_key.py", line 218, in SendAndRecv
> return self.Recv()
> File "/aosp-tests/net/test/pf_key.py", line 208, in Recv
> raise OSError(msg.errno, os.strerror(msg.errno))
> OSError: [Errno 3] No such process
>
> The failure is at
> https://cs.android.com/android/platform/superproject/+/master:kernel/tests/net/test/pf_key_test.py;l=42
> ie.
>
> ENCRYPTION_KEY =
> ("308146eb3bd84b044573d60f5a5fd15957c7d4fe567a2120f35bae0f9869ec22".decode("hex"))
>
> src4 = csocket.Sockaddr(("192.0.2.1", 0))
> dst4 = csocket.Sockaddr(("192.0.2.2", 1))
> self.pf_key.AddSa(src4, dst4, 0xdeadbeef, pf_key.SADB_TYPE_ESP,
> pf_key.IPSEC_MODE_TRANSPORT, 54321,
> pf_key.SADB_X_EALG_AESCBC, ENCRYPTION_KEY,
> pf_key.SADB_X_AALG_SHA2_256HMAC, ENCRYPTION_KEY)
I guess that is because of
commit 4dc2a5a8f6754492180741facf2a8787f2c415d7
net: af_key: add check for pfkey_broadcast in function pfkey_process
This is already reverted in the ipsec tree and will go upstream
this week.
Powered by blists - more mailing lists