Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:18 as permitted sender) client-ip=2620:137:e000::1:18;
Date:   Fri, 25 Mar 2022 16:12:03 -0700
From:   Jakub Kicinski <kuba@kernel.org>
To:     Shuah Khan <skhan@linuxfoundation.org>
Cc:     Naresh Kamboju <naresh.kamboju@linaro.org>,
        Shuah Khan <shuah@kernel.org>,
        "David S. Miller" <davem@davemloft.net>,
        Paolo Abeni <pabeni@redhat.com>,
        Linux-Next Mailing List <linux-next@vger.kernel.org>,
        "open list:KERNEL SELFTEST FRAMEWORK" 
        <linux-kselftest@vger.kernel.org>,
        open list <linux-kernel@vger.kernel.org>,
        lkft-triage@lists.linaro.org, Netdev <netdev@vger.kernel.org>
Subject: Re: kselftest: net: tls: hangs
Message-ID: <20220325161203.7000698c@kicinski-fedora-pc1c0hjn.dhcp.thefacebook.com>
In-Reply-To: <8c81e8ad-6741-b5ed-cf0a-5a302d51d40a@linuxfoundation.org>
References: <CA+G9fYsntwPrwk39VfsAjRwoSNnb3nX8kCEUa=Gxit7_pfD6bg@mail.gmail.com>
        <8c81e8ad-6741-b5ed-cf0a-5a302d51d40a@linuxfoundation.org>
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Precedence: bulk

On Fri, 25 Mar 2022 16:13:33 -0600 Shuah Khan wrote:
> > # #  RUN           tls.13_sm4_gcm.splice_cmsg_to_pipe ...
> > # # tls.c:688:splice_cmsg_to_pipe:Expected splice(self->cfd, NULL,
> > p[1], NULL, send_len, 0) (10) == -1 (-1)
> > # # tls.c:689:splice_cmsg_to_pipe:Expected errno (2) == EINVAL (22)
> > # # splice_cmsg_to_pipe: Test terminated by timeout
> > # #          FAIL  tls.13_sm4_gcm.splice_cmsg_to_pipe
> > # not ok 217 tls.13_sm4_gcm.splice_cmsg_to_pipe
> > # #  RUN           tls.13_sm4_gcm.splice_dec_cmsg_to_pipe ...
> > # # tls.c:708:splice_dec_cmsg_to_pipe:Expected recv(self->cfd, buf,
> > send_len, 0) (10) == -1 (-1)
> > # # tls.c:709:splice_dec_cmsg_to_pipe:Expected errno (2) == EIO (5)
> > [  661.901558] kworker/dying (49) used greatest stack depth: 10576 bytes left  
> 
> This seems to be the problem perhaps.
>
> Jakub, any thoughts. The last change to tls.c was a while back.

Yes, sorry, kicked off a build and got distracted.

I can repro the failures, TLS=n in the config I must have not tested
that in the new cases.

But I can't repro the hung, and we have a timer at the hardness level
IIUC so IDK how this could "hang"?

Naresh, is there any stack trace in the logs? Can you repro on Linus's
tree?