Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
IronPort-PHdr: A9a23:IK2BZx87AnhMof9uWD3oyV9kXcBvk6/5Mg4c9twsjLcdOqig/pG3O
 kvZ6L0tiVLSRozU5rpCjPaeqKHvX2EMoPPj+HAPeZBBTVkJ3MMRmQFzGsOJCUTnavXtan9yE
 MFLTlQw+Xa9PABcE9r/YFuHpHq04HYSFxzzOBAzKP7yH9vZjt+80Ka5/JiACzg=
IronPort-HdrOrdr: A9a23:3UhiN6+6LGYrjCn329xuk+ENdb1zdoMgy1knxilNoENuE/Bwxv
 rBoB1E73DJYW4qKQ4dcdDpAtjmfZquz+8K3WB3B8biYOCGghrnEGgG1+vfKlLbalbDH4JmpM
 Jdmu1FeaHN5DtB/IbHCWuDYqwdKbC8mcjC74qzvhQdLz2CKZsQkjuRYTzrdHGeMTM2fabRY6
 Dsn/avyQDQHUg/X4CePD0oTuLDr9rEmNbNehgdHSMq7wGIkHeB9KP6OwLw5GZcbxp/hZMZtU
 TVmQ3w4auu99uhzAXH6mPV55NK3PP819p4AtCWgMR9EESvtu/oXvUlZ1SxhkFznAid0idtrD
 AKmWZ4Ay1H0QKUQohym2q05+Cv6kd015ao8y7ovZKqm72IeNt9MbsauWqcGSGpt3bJe7pHof
 92NiuixulqJAKFkyLn69fSURZ20kKyvHo5iOYWy2dSSI0EddZq3MEiFW5uYdw99RjBmcoa+S
 hVfbfhDf1tAB+nhrDizyFSKfmXLzsO9zu9Mzw/U/2uonFrdSpCvj4lLeQk7wA9HbwGOut529
 g=
From:   "Denys Zagorui -X (dzagorui - GLOBALLOGIC INC at Cisco)" 
        <dzagorui@cisco.com>
To:     Jiri Olsa <jolsa@redhat.com>
CC:     "linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
        "peterz@infradead.org" <peterz@infradead.org>,
        "mingo@redhat.com" <mingo@redhat.com>,
        "acme@kernel.org" <acme@kernel.org>,
        "mark.rutland@arm.com" <mark.rutland@arm.com>,
        "alexander.shishkin@linux.intel.com" 
        <alexander.shishkin@linux.intel.com>,
        "namhyung@kernel.org" <namhyung@kernel.org>
Subject: Re: [PATCH v4 2/3] perf tests: avoid storing an absolute path in perf
 binary
Thread-Topic: [PATCH v4 2/3] perf tests: avoid storing an absolute path in
 perf binary
Thread-Index: AQHXQoXAMcfZOmyFyUqfrBbHL7WC86rX4Udi
Date:   Fri, 7 May 2021 11:25:51 +0000
Message-ID: <BY5PR11MB4024A423BC46263285CC19B4D9579@BY5PR11MB4024.namprd11.prod.outlook.com>
References: <20210430133350.20504-1-dzagorui@cisco.com>
 <20210430133350.20504-2-dzagorui@cisco.com>,<YJP/wDx/cIKLkpLk@krava>
In-Reply-To: <YJP/wDx/cIKLkpLk@krava>
Accept-Language: en-GB, en-US
Content-Language: en-GB
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: quoted-printable
MIME-Version: 1.0
X-MS-Exchange-CrossTenant-Network-Message-Id: e6cdd3dc-f3ea-41c2-2908-08d9114adecd
X-MS-Exchange-CrossTenant-originalarrivaltime: 07 May 2021 11:25:51.1136
 (UTC)
X-MS-Exchange-CrossTenant-fromentityheader: Hosted
X-MS-Exchange-CrossTenant-id: 5ae1af62-9505-4097-a69a-c1553ef7840e
X-MS-Exchange-CrossTenant-mailboxtype: HOSTED
X-MS-Exchange-CrossTenant-userprincipalname: gG1s+U+B9gBITKFCkPbIwQdjZ0NHHTdrG0Mq4Px6NkNY9LkKxFgnRjCiHvOxxH13DDLB1mzNg5JD4tlWAQnZmA==
X-MS-Exchange-Transport-CrossTenantHeadersStamped: BY5PR11MB4069
X-OriginatorOrg: cisco.com
X-Outbound-SMTP-Client: 173.37.102.18, xbe-rcd-003.cisco.com
X-Outbound-Node: rcdn-core-7.cisco.com
Precedence: bulk
List-ID: <linux-kernel.vger.kernel.org>
X-Mailing-List: linux-kernel@vger.kernel.org

>>  int test__python_use(struct test *test __maybe_unused, int subtest __ma=
ybe_unused)
>>  {
>>        char *cmd;
>>        int ret;
>> +     char *exec_path;
>> +     char *pythonpath;
>> +     struct stat sb;
>> +
>> +     exec_path =3D perf_exe_path();
>> +     if (exec_path =3D=3D NULL)
>> +             return -1;
>
> should we return TEST_SKIP in here?

I'm not sure how it differs from a case where asprintf returns -1

>> +     buf =3D perf_exe(buf, PATH_MAX);
>> +
>> +     for (i =3D strlen(buf) - 1; i !=3D 0 && buf[i] !=3D '/'; i--)
>> +             ;
>
> could we call dirname for this?

Yes

Thanks,
Denys

________________________________________
From: Jiri Olsa <jolsa@redhat.com>
Sent: 06 May 2021 17:40
To: Denys Zagorui -X (dzagorui - GLOBALLOGIC INC at Cisco)
Cc: linux-kernel@vger.kernel.org; peterz@infradead.org; mingo@redhat.com; a=
cme@kernel.org; mark.rutland@arm.com; alexander.shishkin@linux.intel.com; n=
amhyung@kernel.org
Subject: Re: [PATCH v4 2/3] perf tests: avoid storing an absolute path in p=
erf binary

On Fri, Apr 30, 2021 at 06:33:49AM -0700, Denys Zagorui wrote:
> python binding test uses PYTHONPATH definition to find python/perf.so
> library. This definition is an absolute path that makes perf binary
> unreproducible. This path can be found during runtime execution.
>
> Signed-off-by: Denys Zagorui <dzagorui@cisco.com>
> ---
>  tools/perf/tests/Build        |  2 +-
>  tools/perf/tests/python-use.c | 19 ++++++++++++++++++-
>  tools/perf/util/util.c        | 21 +++++++++++++++++++++
>  tools/perf/util/util.h        |  2 ++
>  4 files changed, 42 insertions(+), 2 deletions(-)
>
> diff --git a/tools/perf/tests/Build b/tools/perf/tests/Build
> index 650aec19d490..a20098dcdbc4 100644
> --- a/tools/perf/tests/Build
> +++ b/tools/perf/tests/Build
> @@ -98,5 +98,5 @@ perf-$(CONFIG_DWARF_UNWIND) +=3D dwarf-unwind.o
>  endif
>
>  CFLAGS_attr.o         +=3D -DBINDIR=3D"BUILD_STR($(bindir_SQ))" -DPYTHON=
=3D"BUILD_STR($(PYTHON_WORD))"
> -CFLAGS_python-use.o   +=3D -DPYTHONPATH=3D"BUILD_STR($(OUTPUT)python)" -=
DPYTHON=3D"BUILD_STR($(PYTHON_WORD))"
> +CFLAGS_python-use.o   +=3D -DPYTHON=3D"BUILD_STR($(PYTHON_WORD))"
>  CFLAGS_dwarf-unwind.o +=3D -fno-optimize-sibling-calls
> diff --git a/tools/perf/tests/python-use.c b/tools/perf/tests/python-use.=
c
> index 98c6d474aa6f..c7a0c9b5366f 100644
> --- a/tools/perf/tests/python-use.c
> +++ b/tools/perf/tests/python-use.c
> @@ -8,18 +8,35 @@
>  #include <linux/compiler.h>
>  #include "tests.h"
>  #include "util/debug.h"
> +#include "util/util.h"
> +#include <sys/stat.h>
>
>  int test__python_use(struct test *test __maybe_unused, int subtest __may=
be_unused)
>  {
>       char *cmd;
>       int ret;
> +     char *exec_path;
> +     char *pythonpath;
> +     struct stat sb;
> +
> +     exec_path =3D perf_exe_path();
> +     if (exec_path =3D=3D NULL)
> +             return -1;

should we return TEST_SKIP in here?

> +
> +     if (asprintf(&pythonpath, "%spython", exec_path) < 0)
> +             return -1;

leaking exec_path

> +
> +     if (stat(pythonpath, &sb) || !S_ISDIR(sb.st_mode))
> +             pythonpath[0] =3D 0;
>
>       if (asprintf(&cmd, "echo \"import sys ; sys.path.append('%s'); impo=
rt perf\" | %s %s",
> -                  PYTHONPATH, PYTHON, verbose > 0 ? "" : "2> /dev/null")=
 < 0)
> +                  pythonpath, PYTHON, verbose > 0 ? "" : "2> /dev/null")=
 < 0)
>               return -1;

leaking exec_path and pythonpath

>
>       pr_debug("python usage test: \"%s\"\n", cmd);
>       ret =3D system(cmd) ? -1 : 0;
>       free(cmd);
> +     free(exec_path);
> +     free(pythonpath);
>       return ret;
>  }
> diff --git a/tools/perf/util/util.c b/tools/perf/util/util.c
> index 3bba74e431ed..54e80452887c 100644
> --- a/tools/perf/util/util.c
> +++ b/tools/perf/util/util.c
> @@ -388,3 +388,24 @@ char *perf_exe(char *buf, int len)
>       }
>       return strcpy(buf, "perf");
>  }
> +
> +char *perf_exe_path(void)
> +{
> +     int i;
> +     char *buf;
> +
> +     buf =3D malloc(PATH_MAX);

need to check buf !=3D NULL

> +     buf =3D perf_exe(buf, PATH_MAX);
> +
> +     for (i =3D strlen(buf) - 1; i !=3D 0 && buf[i] !=3D '/'; i--)
> +             ;

could we call dirname for this?

thanks,
jirka

> +
> +     if (!i) {
> +             free(buf);
> +             return NULL;
> +     }
> +
> +     buf[i + 1] =3D 0;
> +
> +     return buf;
> +}
> diff --git a/tools/perf/util/util.h b/tools/perf/util/util.h
> index 80b194ee6c7d..4e871e890ef8 100644
> --- a/tools/perf/util/util.h
> +++ b/tools/perf/util/util.h
> @@ -49,6 +49,8 @@ void perf_set_singlethreaded(void);
>  void perf_set_multithreaded(void);
>
>  char *perf_exe(char *buf, int len);
> +/* perf_exe_path return malloc'd string on success, caller must free it =
*/
> +char *perf_exe_path(void);
>
>  #ifndef O_CLOEXEC
>  #ifdef __sparc__
> --
> 2.26.2.Cisco
>