Received-SPF: pass (google.com: best guess record for domain of linux-crypto-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
MIME-Version: 1.0
References: <20191202135918.566dd377@canb.auug.org.au> <5231d2c0-41d9-6721-e15f-a7eedf3ce69e@infradead.org>
In-Reply-To: <5231d2c0-41d9-6721-e15f-a7eedf3ce69e@infradead.org>
From:   Marco Elver <elver@google.com>
Date:   Mon, 2 Dec 2019 17:20:32 +0100
Message-ID: <CANpmjNM2b26Oo6k-4EqfrJf1sBj3WoFf-NQnwsLr3EW9B=G8kw@mail.gmail.com>
Subject: Re: linux-next: Tree for Dec 2 (drivers/crypto/hisilicon/sec2/sec_crypto.o
 +kcsan??)
To:     Randy Dunlap <rdunlap@infradead.org>
Cc:     Stephen Rothwell <sfr@canb.auug.org.au>,
        Linux Next Mailing List <linux-next@vger.kernel.org>,
        Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
        linux-crypto@vger.kernel.org, Zaibo Xu <xuzaibo@huawei.com>,
        Dmitry Vyukov <dvyukov@google.com>
Content-Type: text/plain; charset="UTF-8"
Sender: linux-crypto-owner@vger.kernel.org
Precedence: bulk

On Mon, 2 Dec 2019 at 16:56, Randy Dunlap <rdunlap@infradead.org> wrote:
>
> On 12/1/19 6:59 PM, Stephen Rothwell wrote:
> > Hi all,
> >
> > Please do not add any material for v5.6 to your linux-next included
> > trees until after v5.5-rc1 has been released.
> >
> > Changes since 20191129:
> >
>
> on x86_64:
>
> ld: drivers/crypto/hisilicon/sec2/sec_crypto.o: in function `sec_bd_send':
> sec_crypto.c:(.text+0xae9): undefined reference to `__tsan_atomic64_fetch_add'
> ld: drivers/crypto/hisilicon/sec2/sec_crypto.o: in function `sec_req_cb':
> sec_crypto.c:(.text+0xbf3): undefined reference to `__tsan_atomic64_fetch_add'
> ld: drivers/crypto/hisilicon/sec2/sec_crypto.o: in function `sec_skcipher_callback':
> sec_crypto.c:(.text+0x12d3): undefined reference to `__tsan_atomic32_compare_exchange_strong'
>
>
> Full randconfig file is attached.

Thanks, Randy.

I think uses of __sync compiler builtins are banned in the kernel, so
here compiling with KCSAN is actually acting as a linter showing us
places where they are used in new code. The hisilicon/sec2 code in
question was only recently added.

hisilicon/sec2 maintainers: please use atomic64_t and atomic64_inc()
for send_cnt and recv_cnt.

Thanks,
-- Marco