Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
MIME-Version: 1.0
References: <20220625153439.513559-1-tanzixuan.me@gmail.com> <YrhxE4s0hLvbbibp@krava>
In-Reply-To: <YrhxE4s0hLvbbibp@krava>
From:   =?UTF-8?B?6LCt5qKT54WK?= <tanzixuan.me@gmail.com>
Date:   Mon, 27 Jun 2022 11:08:34 +0800
Message-ID: <CABwm_eT_LE6VbLMgT31yqW=tc_obLP=6E0jnMqVn1sMdWrVVNw@mail.gmail.com>
Subject: Re: [PATCH] perf build: Suppress openssl v3 deprecation warnings in
 libcrypto feature test
To:     Jiri Olsa <olsajiri@gmail.com>
Cc:     Zixuan Tan <tanzixuangg@gmail.com>, terrelln@fb.com,
        Alexei Starovoitov <ast@kernel.org>,
        Daniel Borkmann <daniel@iogearbox.net>,
        Andrii Nakryiko <andrii@kernel.org>,
        Martin KaFai Lau <kafai@fb.com>,
        Song Liu <songliubraving@fb.com>, Yonghong Song <yhs@fb.com>,
        John Fastabend <john.fastabend@gmail.com>,
        KP Singh <kpsingh@kernel.org>, linux-kernel@vger.kernel.org,
        netdev@vger.kernel.org, bpf@vger.kernel.org
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

On Sun, Jun 26, 2022 at 10:45 PM Jiri Olsa <olsajiri@gmail.com> wrote:
>
> On Sat, Jun 25, 2022 at 11:34:38PM +0800, Zixuan Tan wrote:
> > With OpenSSL v3 installed, the libcrypto feature check fails as it use the
> > deprecated MD5_* API (and is compiled with -Werror). The error message is
> > as follows.
> >
> > $ make tools/perf
> > ```
> > Makefile.config:778: No libcrypto.h found, disables jitted code injection,
> > please install openssl-devel or libssl-dev
> >
> > Auto-detecting system features:
> > ...                         dwarf: [ on  ]
> > ...            dwarf_getlocations: [ on  ]
> > ...                         glibc: [ on  ]
> > ...                        libbfd: [ on  ]
> > ...                libbfd-buildid: [ on  ]
> > ...                        libcap: [ on  ]
> > ...                        libelf: [ on  ]
> > ...                       libnuma: [ on  ]
> > ...        numa_num_possible_cpus: [ on  ]
> > ...                       libperl: [ on  ]
> > ...                     libpython: [ on  ]
> > ...                     libcrypto: [ OFF ]
> > ...                     libunwind: [ on  ]
> > ...            libdw-dwarf-unwind: [ on  ]
> > ...                          zlib: [ on  ]
> > ...                          lzma: [ on  ]
> > ...                     get_cpuid: [ on  ]
> > ...                           bpf: [ on  ]
> > ...                        libaio: [ on  ]
> > ...                       libzstd: [ on  ]
> > ...        disassembler-four-args: [ on  ]
> > ```
> >
> > This is very confusing because the suggested library (on my Ubuntu 20.04
> > it is libssl-dev) is already installed. As the test only checks for the
> > presence of libcrypto, this commit suppresses the deprecation warning to
> > allow the test to pass.
> >
> > Signed-off-by: Zixuan Tan <tanzixuan.me@gmail.com>
> > ---
> >  tools/build/feature/test-libcrypto.c | 6 ++++++
> >  1 file changed, 6 insertions(+)
> >
> > diff --git a/tools/build/feature/test-libcrypto.c b/tools/build/feature/test-libcrypto.c
> > index a98174e0569c..31afff093d0b 100644
> > --- a/tools/build/feature/test-libcrypto.c
> > +++ b/tools/build/feature/test-libcrypto.c
> > @@ -2,6 +2,12 @@
> >  #include <openssl/sha.h>
> >  #include <openssl/md5.h>
> >
> > +/*
> > + * The MD5_* API have been deprecated since OpenSSL 3.0, which causes the
> > + * feature test to fail silently. This is a workaround.
> > + */
>
> then we use these deprecated MD5 calls in util/genelf.c if libcrypto is detected,
> so I wonder how come the rest of the compilation passed for you.. do you have
> CONFIG_JITDUMP disabled?
>
> thanks,
> jirka
>
No, CONFIG_JITDUMP is not disabled. I am using the default configuration.

Yes, you are right. The rest of the compilation should fail, but it doesn't.
I checked the verbose build commands. This seems to be the result of another
inconsistency.

If libcrypto is detected, the macro "HAVE_LIBCRYPTO_SUPPORT" will be
defined, but in perf/util/genelf.c, "HAVE_LIBCRYPTO" without the "_SUPPORT"
prefix is checked. This causes urandom always be used to create build id
rather than MD5 and SHA1, no matter what the detection result is.

In perf/Makefile.config, from line 776
```
ifndef NO_LIBCRYPTO
  ifneq ($(feature-libcrypto), 1)
    msg := $(warning No libcrypto.h found, disables jitted code injection,
            please install openssl-devel or libssl-dev);
    NO_LIBCRYPTO := 1
  else                                  <-- if libcrypto feature detected
    CFLAGS += -DHAVE_LIBCRYPTO_SUPPORT  <-- define this
    EXTLIBS += -lcrypto
    $(call detected,CONFIG_CRYPTO)
  endif
endif
```

In perf/util/genelf.c, from line 33
```
#ifdef HAVE_LIBCRYPTO                <-- but check this, it's always false

#define BUILD_ID_MD5
#undef BUILD_ID_SHA /* does not seem to work well when linked with Java */
#undef BUILD_ID_URANDOM /* different uuid for each run */

#ifdef BUILD_ID_SHA
#include <openssl/sha.h>
#endif

#ifdef BUILD_ID_MD5
#include <openssl/md5.h>
#endif
#endif                               <-- this block will be skipped
```

Maybe we should fix this, to really make use of libcrypto if it is available?

Links:
 This commit include the genelf.c:
  https://lore.kernel.org/all/1448874143-7269-3-git-send-email-eranian@google.com/T/#mb6d3e18bee4901b71a4d4ef4f406feaaf48346d9
 This commit include the feature test:
  https://lore.kernel.org/all/1448874143-7269-3-git-send-email-eranian@google.com/T/#m12a2ababf8ad3e366d56d9efab870592e6ff60a5

Thanks,
Zixuan

> > +#pragma GCC diagnostic ignored "-Wdeprecated-declarations"
> > +
> >  int main(void)
> >  {
> >       MD5_CTX context;
> > --
> > 2.34.1
> >