[dpdk-dev] [PATCH] usertools: fix python 3 support of dpdk-pmdinfo.py

[dpdk-dev] [PATCH] usertools: fix python 3 support of dpdk-pmdinfo.py

[Robin Jarry]

Running this script on ubuntu-18.04 with python 3 and pyelftools
installed produces no output but no error is reported neither:

  ~$ python3 usertools/dpdk-pmdinfo.py -r build/app/testpmd
  ~$ echo $?
  0

While with python 2, it works:

  ~# python2 usertools/dpdk-pmdinfo.py -r build/app/testpmd
  {"pci_ids": [], "name": "dpio"}
  {"pci_ids": [], "name": "dpbp"}
  {"pci_ids": [], "name": "dpaa2_qdma"}
  .....

Looking in the code of pyelftools, ELFFile.get_section_by_name performs
a lookup in a _section_name_map dictionary which contains a mapping of
section names to section numbers. Section names are strings, not bytes.
We must not encode the string before using it as argument to this
function. Otherwise the lookup fails and the script assumes there is no
ELF section in the file. Then it exits without any error.

Also, section tags are strings, not bytes, fix the following error on
python 3:

  Traceback (most recent call last):
    File "usertools/dpdk-pmdinfo.py", line 612, in <module>
      main()
    File "usertools/dpdk-pmdinfo.py", line 601, in main
      readelf.process_dt_needed_entries()
    File "usertools/dpdk-pmdinfo.py", line 442, in
    process_dt_needed_entries
      rc = tag.needed.find(b"librte_pmd")
  TypeError: must be str, not bytes

Fixes: c67c9a5c646a ("tools: query binaries for HW and other support information")
Cc: Neil Horman <nhorman@tuxdriver.com>
Fixes: 54ca545dce4b ("make python scripts python2/3 compliant")
Cc: John McNamara <john.mcnamara@intel.com>
Cc: stable@dpdk.org

Signed-off-by: Robin Jarry <robin.jarry@6wind.com>
Reviewed-by: Olivier Matz <olivier.matz@6wind.com>
---
 usertools/dpdk-pmdinfo.py | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/usertools/dpdk-pmdinfo.py b/usertools/dpdk-pmdinfo.py
index 03623d5b8b48..d2d4aa3c143f 100755
--- a/usertools/dpdk-pmdinfo.py
+++ b/usertools/dpdk-pmdinfo.py
@@ -14,7 +14,7 @@
 import string
 import sys
 from elftools.common.exceptions import ELFError
-from elftools.common.py3compat import (byte2int, bytes2str, str2bytes)
+from elftools.common.py3compat import (byte2int, bytes2str)
 from elftools.elf.elffile import ELFFile
 from optparse import OptionParser
 
@@ -267,7 +267,7 @@ def _section_from_spec(self, spec):
                 return None
         except ValueError:
             # Not a number. Must be a name then
-            return self.elffile.get_section_by_name(str2bytes(spec))
+            return self.elffile.get_section_by_name(spec)
 
     def pretty_print_pmdinfo(self, pmdinfo):
         global pcidb
@@ -439,7 +439,7 @@ def process_dt_needed_entries(self):
 
         for tag in dynsec.iter_tags():
             if tag.entry.d_tag == 'DT_NEEDED':
-                rc = tag.needed.find(b"librte_pmd")
+                rc = tag.needed.find("librte_pmd")
                 if (rc != -1):
                     library = search_file(tag.needed,
                                           runpath + ":" + ldlibpath +
-- 
2.23.0.rc1

[dpdk-dev] [PATCH v2] usertools: fix py3 support with pyelftools>=0.24

[Robin Jarry]

Running dpdk-pmdinfo.py on Ubuntu 18.04 (bionic) with python 3 and
pyelftools installed produces no output but no error is reported
neither:

  ~$ python3 usertools/dpdk-pmdinfo.py -r build/app/testpmd
  ~$ echo $?
  0

While with python 2, it works:

  ~# python2 usertools/dpdk-pmdinfo.py -r build/app/testpmd
  {"pci_ids": [], "name": "dpio"}
  {"pci_ids": [], "name": "dpbp"}
  {"pci_ids": [], "name": "dpaa2_qdma"}
  .....

On Ubuntu 18.04, pyelftools is version 0.24. The change log of
pyelftools v0.24 says:

 - Symbol/section names are strings internally now, not bytestrings
   (this may affect API usage in Python 3) (#76).

We cannot guess which version of pyelftools is actually being used. The
elftools.__version__ symbol is not consistent with each distro's package
version. For example, on Ubuntu 16.04 (xenial), the .deb package version
is '0.23-2' but elftools.__version__ contains '0.25'. This is certainly
due to partial backports.

To have a more consistent behaviour of this script across all versions
of python, add the unicode_literals future import so that literal
strings are now always "unicode".

Add 2 utility functions to force a string into bytes or bytes into an
unicode string.

Force pyelftools return values to unicode strings (will do nothing with
recent version of pyelftools).

If elffile.get_section_by_name returns None with a unicode section name,
try with the same one encoded as bytes.

Also, replace all open() calls by io.open() which behaves like the
builtin open in python 3. The only non-binary opened file is
/usr/share/hwdata/pci.ids which is UTF-8 encoded text. Explicitly
specify that encoding.

Link: https://github.com/eliben/pyelftools/blob/v0.24/CHANGES#L7
Link: https://github.com/eliben/pyelftools/commit/108eaea9e75a8b5a
Signed-off-by: Robin Jarry <robin.jarry@6wind.com>
Reviewed-by: Olivier Matz <olivier.matz@6wind.com>
---
Change log:

v2

  * Behavior is different depending on pyelftools versions. Handle that
    properly.
  * More bytes/unicode fixes.

 usertools/dpdk-pmdinfo.py | 65 +++++++++++++++++++++++++++------------
 1 file changed, 46 insertions(+), 19 deletions(-)

diff --git a/usertools/dpdk-pmdinfo.py b/usertools/dpdk-pmdinfo.py
index 03623d5b8b48..069a3bf124b2 100755
--- a/usertools/dpdk-pmdinfo.py
+++ b/usertools/dpdk-pmdinfo.py
@@ -8,13 +8,15 @@
 #
 # -------------------------------------------------------------------------
 from __future__ import print_function
+from __future__ import unicode_literals
 import json
+import io
 import os
 import platform
 import string
 import sys
 from elftools.common.exceptions import ELFError
-from elftools.common.py3compat import (byte2int, bytes2str, str2bytes)
+from elftools.common.py3compat import byte2int
 from elftools.elf.elffile import ELFFile
 from optparse import OptionParser
 
@@ -213,7 +215,8 @@ def readLocal(self, filename):
         """
         Reads the local file
         """
-        self.contents = open(filename).readlines()
+        with io.open(filename, 'r', encoding='utf-8') as f:
+            self.contents = f.readlines()
         self.date = self.findDate(self.contents)
 
     def loadLocal(self):
@@ -267,7 +270,13 @@ def _section_from_spec(self, spec):
                 return None
         except ValueError:
             # Not a number. Must be a name then
-            return self.elffile.get_section_by_name(str2bytes(spec))
+            section = self.elffile.get_section_by_name(force_unicode(spec))
+            if section is None:
+                # No match with a unicode name.
+                # Some versions of pyelftools (<= 0.23) store internal strings
+                # as bytes. Try again with the name encoded as bytes.
+                section = self.elffile.get_section_by_name(force_bytes(spec))
+            return section
 
     def pretty_print_pmdinfo(self, pmdinfo):
         global pcidb
@@ -339,7 +348,8 @@ def display_pmd_info_strings(self, section_spec):
             while endptr < len(data) and byte2int(data[endptr]) != 0:
                 endptr += 1
 
-            mystring = bytes2str(data[dataptr:endptr])
+            # pyelftools may return byte-strings, force decode them
+            mystring = force_unicode(data[dataptr:endptr])
             rc = mystring.find("PMD_INFO_STRING")
             if (rc != -1):
                 self.parse_pmd_info_string(mystring)
@@ -348,9 +358,10 @@ def display_pmd_info_strings(self, section_spec):
 
     def find_librte_eal(self, section):
         for tag in section.iter_tags():
-            if tag.entry.d_tag == 'DT_NEEDED':
-                if "librte_eal" in tag.needed:
-                    return tag.needed
+            # pyelftools may return byte-strings, force decode them
+            if force_unicode(tag.entry.d_tag) == 'DT_NEEDED':
+                if "librte_eal" in force_unicode(tag.needed):
+                    return force_unicode(tag.needed)
         return None
 
     def search_for_autoload_path(self):
@@ -373,7 +384,7 @@ def search_for_autoload_path(self):
                     return (None, None)
                 if raw_output is False:
                     print("Scanning for autoload path in %s" % library)
-                scanfile = open(library, 'rb')
+                scanfile = io.open(library, 'rb')
                 scanelf = ReadElf(scanfile, sys.stdout)
         except AttributeError:
             # Not a dynamic binary
@@ -403,7 +414,8 @@ def search_for_autoload_path(self):
             while endptr < len(data) and byte2int(data[endptr]) != 0:
                 endptr += 1
 
-            mystring = bytes2str(data[dataptr:endptr])
+            # pyelftools may return byte-strings, force decode them
+            mystring = force_unicode(data[dataptr:endptr])
             rc = mystring.find("DPDK_PLUGIN_PATH")
             if (rc != -1):
                 rc = mystring.find("=")
@@ -416,8 +428,9 @@ def search_for_autoload_path(self):
 
     def get_dt_runpath(self, dynsec):
         for tag in dynsec.iter_tags():
-            if tag.entry.d_tag == 'DT_RUNPATH':
-                return tag.runpath
+            # pyelftools may return byte-strings, force decode them
+            if force_unicode(tag.entry.d_tag) == 'DT_RUNPATH':
+                return force_unicode(tag.runpath)
         return ""
 
     def process_dt_needed_entries(self):
@@ -438,16 +451,16 @@ def process_dt_needed_entries(self):
             return
 
         for tag in dynsec.iter_tags():
-            if tag.entry.d_tag == 'DT_NEEDED':
-                rc = tag.needed.find(b"librte_pmd")
-                if (rc != -1):
-                    library = search_file(tag.needed,
+            # pyelftools may return byte-strings, force decode them
+            if force_unicode(tag.entry.d_tag) == 'DT_NEEDED':
+                if 'librte_pmd' in force_unicode(tag.needed):
+                    library = search_file(force_unicode(tag.needed),
                                           runpath + ":" + ldlibpath +
                                           ":/usr/lib64:/lib64:/usr/lib:/lib")
                     if library is not None:
                         if raw_output is False:
                             print("Scanning %s for pmd information" % library)
-                        with open(library, 'rb') as file:
+                        with io.open(library, 'rb') as file:
                             try:
                                 libelf = ReadElf(file, sys.stdout)
                             except ELFError:
@@ -458,6 +471,20 @@ def process_dt_needed_entries(self):
                             file.close()
 
 
+# compat: remove force_unicode & force_bytes when pyelftools<=0.23 support is
+# dropped.
+def force_unicode(s):
+    if hasattr(s, 'decode') and callable(s.decode):
+        s = s.decode('latin-1')  # same encoding used in pyelftools py3compat
+    return s
+
+
+def force_bytes(s):
+    if hasattr(s, 'encode') and callable(s.encode):
+        s = s.encode('latin-1')  # same encoding used in pyelftools py3compat
+    return s
+
+
 def scan_autoload_path(autoload_path):
     global raw_output
 
@@ -476,7 +503,7 @@ def scan_autoload_path(autoload_path):
             scan_autoload_path(dpath)
         if os.path.isfile(dpath):
             try:
-                file = open(dpath, 'rb')
+                file = io.open(dpath, 'rb')
                 readelf = ReadElf(file, sys.stdout)
             except ELFError:
                 # this is likely not an elf file, skip it
@@ -503,7 +530,7 @@ def scan_for_autoload_pmds(dpdk_path):
             print("Must specify a file name")
         return
 
-    file = open(dpdk_path, 'rb')
+    file = io.open(dpdk_path, 'rb')
     try:
         readelf = ReadElf(file, sys.stdout)
     except ElfError:
@@ -595,7 +622,7 @@ def main(stream=None):
         print("File not found")
         sys.exit(1)
 
-    with open(myelffile, 'rb') as file:
+    with io.open(myelffile, 'rb') as file:
         try:
             readelf = ReadElf(file, sys.stdout)
             readelf.process_dt_needed_entries()
-- 
2.23.0

[dpdk-dev] [PATCH v3] usertools: fix py3 support with pyelftools>=0.24

[Robin Jarry]

Running dpdk-pmdinfo.py on Ubuntu 18.04 (bionic) with python 3 and
pyelftools installed produces no output but no error is reported
neither:

  ~$ python3 usertools/dpdk-pmdinfo.py -r build/app/testpmd
  ~$ echo $?
  0

While with python 2, it works:

  ~# python2 usertools/dpdk-pmdinfo.py -r build/app/testpmd
  {"pci_ids": [], "name": "dpio"}
  {"pci_ids": [], "name": "dpbp"}
  {"pci_ids": [], "name": "dpaa2_qdma"}
  .....

On Ubuntu 18.04, pyelftools is version 0.24. The change log of
pyelftools v0.24 says:

 - Symbol/section names are strings internally now, not bytestrings
   (this may affect API usage in Python 3) (#76).

We cannot guess which version of pyelftools is actually being used. The
elftools.__version__ symbol is not consistent with each distro's package
version. For example, on Ubuntu 16.04 (xenial), the .deb package version
is '0.23-2' but elftools.__version__ contains '0.25'. This is certainly
due to partial backports.

To have a more consistent behaviour of this script across all versions
of python, add the unicode_literals future import so that literal
strings are now always "unicode".

Add 2 utility functions to force a string into bytes or bytes into an
unicode string.

Force pyelftools return values to unicode strings (will do nothing with
recent version of pyelftools).

If elffile.get_section_by_name returns None with a unicode section name,
try with the same one encoded as bytes.

Also, replace all open() calls by io.open() which behaves like the
builtin open in python 3. The only non-binary opened file is
/usr/share/hwdata/pci.ids which is UTF-8 encoded text. Explicitly
specify that encoding.

Link: https://github.com/eliben/pyelftools/blob/v0.24/CHANGES#L7
Link: https://github.com/eliben/pyelftools/commit/108eaea9e75a8b5a
Fixes: 54ca545dce4b ("make python scripts python2/3 compliant")
Cc: John McNamara <john.mcnamara@intel.com>
Cc: stable@dpdk.org
Signed-off-by: Robin Jarry <robin.jarry@6wind.com>
Reviewed-by: Olivier Matz <olivier.matz@6wind.com>
---
Change log:

v2

  * Behavior is different depending on pyelftools versions. Handle that
    properly.
  * More bytes/unicode fixes.

v3

  * Add missing Fixes line
  * Add Cc: stable@dpdk.org

 usertools/dpdk-pmdinfo.py | 65 +++++++++++++++++++++++++++------------
 1 file changed, 46 insertions(+), 19 deletions(-)

diff --git a/usertools/dpdk-pmdinfo.py b/usertools/dpdk-pmdinfo.py
index 03623d5b8b48..069a3bf124b2 100755
--- a/usertools/dpdk-pmdinfo.py
+++ b/usertools/dpdk-pmdinfo.py
@@ -8,13 +8,15 @@
 #
 # -------------------------------------------------------------------------
 from __future__ import print_function
+from __future__ import unicode_literals
 import json
+import io
 import os
 import platform
 import string
 import sys
 from elftools.common.exceptions import ELFError
-from elftools.common.py3compat import (byte2int, bytes2str, str2bytes)
+from elftools.common.py3compat import byte2int
 from elftools.elf.elffile import ELFFile
 from optparse import OptionParser
 
@@ -213,7 +215,8 @@ def readLocal(self, filename):
         """
         Reads the local file
         """
-        self.contents = open(filename).readlines()
+        with io.open(filename, 'r', encoding='utf-8') as f:
+            self.contents = f.readlines()
         self.date = self.findDate(self.contents)
 
     def loadLocal(self):
@@ -267,7 +270,13 @@ def _section_from_spec(self, spec):
                 return None
         except ValueError:
             # Not a number. Must be a name then
-            return self.elffile.get_section_by_name(str2bytes(spec))
+            section = self.elffile.get_section_by_name(force_unicode(spec))
+            if section is None:
+                # No match with a unicode name.
+                # Some versions of pyelftools (<= 0.23) store internal strings
+                # as bytes. Try again with the name encoded as bytes.
+                section = self.elffile.get_section_by_name(force_bytes(spec))
+            return section
 
     def pretty_print_pmdinfo(self, pmdinfo):
         global pcidb
@@ -339,7 +348,8 @@ def display_pmd_info_strings(self, section_spec):
             while endptr < len(data) and byte2int(data[endptr]) != 0:
                 endptr += 1
 
-            mystring = bytes2str(data[dataptr:endptr])
+            # pyelftools may return byte-strings, force decode them
+            mystring = force_unicode(data[dataptr:endptr])
             rc = mystring.find("PMD_INFO_STRING")
             if (rc != -1):
                 self.parse_pmd_info_string(mystring)
@@ -348,9 +358,10 @@ def display_pmd_info_strings(self, section_spec):
 
     def find_librte_eal(self, section):
         for tag in section.iter_tags():
-            if tag.entry.d_tag == 'DT_NEEDED':
-                if "librte_eal" in tag.needed:
-                    return tag.needed
+            # pyelftools may return byte-strings, force decode them
+            if force_unicode(tag.entry.d_tag) == 'DT_NEEDED':
+                if "librte_eal" in force_unicode(tag.needed):
+                    return force_unicode(tag.needed)
         return None
 
     def search_for_autoload_path(self):
@@ -373,7 +384,7 @@ def search_for_autoload_path(self):
                     return (None, None)
                 if raw_output is False:
                     print("Scanning for autoload path in %s" % library)
-                scanfile = open(library, 'rb')
+                scanfile = io.open(library, 'rb')
                 scanelf = ReadElf(scanfile, sys.stdout)
         except AttributeError:
             # Not a dynamic binary
@@ -403,7 +414,8 @@ def search_for_autoload_path(self):
             while endptr < len(data) and byte2int(data[endptr]) != 0:
                 endptr += 1
 
-            mystring = bytes2str(data[dataptr:endptr])
+            # pyelftools may return byte-strings, force decode them
+            mystring = force_unicode(data[dataptr:endptr])
             rc = mystring.find("DPDK_PLUGIN_PATH")
             if (rc != -1):
                 rc = mystring.find("=")
@@ -416,8 +428,9 @@ def search_for_autoload_path(self):
 
     def get_dt_runpath(self, dynsec):
         for tag in dynsec.iter_tags():
-            if tag.entry.d_tag == 'DT_RUNPATH':
-                return tag.runpath
+            # pyelftools may return byte-strings, force decode them
+            if force_unicode(tag.entry.d_tag) == 'DT_RUNPATH':
+                return force_unicode(tag.runpath)
         return ""
 
     def process_dt_needed_entries(self):
@@ -438,16 +451,16 @@ def process_dt_needed_entries(self):
             return
 
         for tag in dynsec.iter_tags():
-            if tag.entry.d_tag == 'DT_NEEDED':
-                rc = tag.needed.find(b"librte_pmd")
-                if (rc != -1):
-                    library = search_file(tag.needed,
+            # pyelftools may return byte-strings, force decode them
+            if force_unicode(tag.entry.d_tag) == 'DT_NEEDED':
+                if 'librte_pmd' in force_unicode(tag.needed):
+                    library = search_file(force_unicode(tag.needed),
                                           runpath + ":" + ldlibpath +
                                           ":/usr/lib64:/lib64:/usr/lib:/lib")
                     if library is not None:
                         if raw_output is False:
                             print("Scanning %s for pmd information" % library)
-                        with open(library, 'rb') as file:
+                        with io.open(library, 'rb') as file:
                             try:
                                 libelf = ReadElf(file, sys.stdout)
                             except ELFError:
@@ -458,6 +471,20 @@ def process_dt_needed_entries(self):
                             file.close()
 
 
+# compat: remove force_unicode & force_bytes when pyelftools<=0.23 support is
+# dropped.
+def force_unicode(s):
+    if hasattr(s, 'decode') and callable(s.decode):
+        s = s.decode('latin-1')  # same encoding used in pyelftools py3compat
+    return s
+
+
+def force_bytes(s):
+    if hasattr(s, 'encode') and callable(s.encode):
+        s = s.encode('latin-1')  # same encoding used in pyelftools py3compat
+    return s
+
+
 def scan_autoload_path(autoload_path):
     global raw_output
 
@@ -476,7 +503,7 @@ def scan_autoload_path(autoload_path):
             scan_autoload_path(dpath)
         if os.path.isfile(dpath):
             try:
-                file = open(dpath, 'rb')
+                file = io.open(dpath, 'rb')
                 readelf = ReadElf(file, sys.stdout)
             except ELFError:
                 # this is likely not an elf file, skip it
@@ -503,7 +530,7 @@ def scan_for_autoload_pmds(dpdk_path):
             print("Must specify a file name")
         return
 
-    file = open(dpdk_path, 'rb')
+    file = io.open(dpdk_path, 'rb')
     try:
         readelf = ReadElf(file, sys.stdout)
     except ElfError:
@@ -595,7 +622,7 @@ def main(stream=None):
         print("File not found")
         sys.exit(1)
 
-    with open(myelffile, 'rb') as file:
+    with io.open(myelffile, 'rb') as file:
         try:
             readelf = ReadElf(file, sys.stdout)
             readelf.process_dt_needed_entries()
-- 
2.23.0

Re: [dpdk-dev] [PATCH v3] usertools: fix py3 support with pyelftools>=0.24

[Thomas Monjalon]

15/10/2019 14:39, Robin Jarry:
> Running dpdk-pmdinfo.py on Ubuntu 18.04 (bionic) with python 3 and
> pyelftools installed produces no output but no error is reported
> neither:
> 
>   ~$ python3 usertools/dpdk-pmdinfo.py -r build/app/testpmd
>   ~$ echo $?
>   0
> 
> While with python 2, it works:
> 
>   ~# python2 usertools/dpdk-pmdinfo.py -r build/app/testpmd
>   {"pci_ids": [], "name": "dpio"}
>   {"pci_ids": [], "name": "dpbp"}
>   {"pci_ids": [], "name": "dpaa2_qdma"}
>   .....
> 
> On Ubuntu 18.04, pyelftools is version 0.24. The change log of
> pyelftools v0.24 says:
> 
>  - Symbol/section names are strings internally now, not bytestrings
>    (this may affect API usage in Python 3) (#76).
> 
> We cannot guess which version of pyelftools is actually being used. The
> elftools.__version__ symbol is not consistent with each distro's package
> version. For example, on Ubuntu 16.04 (xenial), the .deb package version
> is '0.23-2' but elftools.__version__ contains '0.25'. This is certainly
> due to partial backports.
> 
> To have a more consistent behaviour of this script across all versions
> of python, add the unicode_literals future import so that literal
> strings are now always "unicode".
> 
> Add 2 utility functions to force a string into bytes or bytes into an
> unicode string.
> 
> Force pyelftools return values to unicode strings (will do nothing with
> recent version of pyelftools).
> 
> If elffile.get_section_by_name returns None with a unicode section name,
> try with the same one encoded as bytes.
> 
> Also, replace all open() calls by io.open() which behaves like the
> builtin open in python 3. The only non-binary opened file is
> /usr/share/hwdata/pci.ids which is UTF-8 encoded text. Explicitly
> specify that encoding.
> 
> Link: https://github.com/eliben/pyelftools/blob/v0.24/CHANGES#L7
> Link: https://github.com/eliben/pyelftools/commit/108eaea9e75a8b5a
> Fixes: 54ca545dce4b ("make python scripts python2/3 compliant")
> Cc: John McNamara <john.mcnamara@intel.com>
> Cc: stable@dpdk.org
> Signed-off-by: Robin Jarry <robin.jarry@6wind.com>
> Reviewed-by: Olivier Matz <olivier.matz@6wind.com>

I am really scary by all of this,
but I trust you to fix it properly.

Applied, thanks