Re: [PATCH v7] checkkconfigsymbols.sh: reimplementation in python

Re: [PATCH v7] checkkconfigsymbols.sh: reimplementation in python

[Paul Bolle]

[Added  linux-kbuild@vger.kernel.org.]

On Sun, 2014-09-28 at 17:55 +0200, Valentin Rothberg wrote:
> The scripts/checkkconfigsymbols.sh script searches Kconfig features
> in the source code that are not defined in Kconfig. Such identifiers
> always evaluate to false and are the source of various kinds of bugs.
> However, the shell script is slow and it does not detect such broken
> references in Kbuild and Kconfig files (e.g., ``depends on UNDEFINED´´).
> Furthermore, it generates false positives. The script is also hard to
> read and understand, and is thereby difficult to maintain.
> 
> This patch replaces the shell script with an implementation in Python,
> which:
>     (a) detects the same bugs, but does not report previous false positives
>     (b) additionally detects broken references in Kconfig and all
>         non-Kconfig files, such as Kbuild, .[cSh], .txt, .sh, defconfig, etc.
>     (c) is up to 75 times faster than the shell script
>     (d) only checks files under version control ('git ls-files')

(The shell script is .git unaware. If you happen to have one or more
branches with "Kconfig" in their name, as I do, it generates a lot of
noise on stderr.)

> The new script reduces the runtime on my machine (i7-2620M, 8GB RAM, SSD)
> from 3m47s to 0m3s, and reports 939 broken identifiers in Linux v3.17-rc1;
> 420 additional reports of which 16 are located in Kconfig files,
> 287 in defconfigs, 63 in ./Documentation, 1 in Kbuild.
> 
> Moreover, we intentionally include references in comments, which have been
> ignored until now. Such comments may be leftovers of features that have
> been removed or renamed in Kconfig (e.g., ``#endif /* CONFIG_MPC52xx */´´).
> These references can be misleading and should be removed or replaced.
> 
> Note that the output format changed from (file list <tab> feature) to
> (feature <tab> file list) as it simplifies the detection of the Kconfig
> feature for long file lists.
> 
> Signed-off-by: Valentin Rothberg <valentinrothberg@gmail.com>
> Signed-off-by: Stefan Hengelein <stefan.hengelein@fau.de>
> ---
> Changelog:
> v2: Fix of regular expressions
> v3: Changelog replacement, and add changes of v2
> v4: Based on comments from Paul Bolle <pebolle@tiscali.nl>
>   - Inclusion of all non-Kconfig files, such as .txt, .sh, etc.
>   - Changes of regular expressions
>   - Increases additional reports from 49 to 229 compared to v3
>   - Change of output format from (file list <tab> feature) to
>         (feature <tab> file list)·
> v5: Only analyze files under version control ('git ls-files')
> v6: Cover features with numbers and small letters (e.g., 4xx)
> v7: Add changes of v6 (lost 'git add') and filter FOO/BAR features
> ---
>  scripts/checkkconfigsymbols.py | 138 +++++++++++++++++++++++++++++++++++++++++
>  scripts/checkkconfigsymbols.sh |  59 ------------------
>  2 files changed, 138 insertions(+), 59 deletions(-)
>  create mode 100644 scripts/checkkconfigsymbols.py
>  delete mode 100755 scripts/checkkconfigsymbols.sh
> 
> diff --git a/scripts/checkkconfigsymbols.py b/scripts/checkkconfigsymbols.py
> new file mode 100644
> index 0000000..01bf9c4
> --- /dev/null
> +++ b/scripts/checkkconfigsymbols.py
> @@ -0,0 +1,138 @@
> +#!/usr/bin/env python
> +
> +"""Find Kconfig identifiers that are referenced but not defined."""
> +
> +# (c) 2014 Valentin Rothberg <valentinrothberg@gmail.com>
> +# (c) 2014 Stefan Hengelein <stefan.hengelein@fau.de>
> +#
> +# Licensed under the terms of the GNU GPL License version 2
> +
> +
> +import os
> +import re
> +from subprocess import Popen, PIPE, STDOUT
> +
> +
> +# regex expressions
> +OPERATORS = r"&|\(|\)|\||\!"
> +FEATURE = r"(?:\w*[A-Z0-9]\w*){2,}"
> +DEF = r"^\s*(?:menu){,1}config\s+(" + FEATURE + r")\s*"
> +EXPR = r"(?:" + OPERATORS + r"|\s|" + FEATURE + r")+"
> +STMT = r"^\s*(?:if|select|depends\s+on)\s+" + EXPR

Could please make that "depends on"? Yes, it seems the yacc grammar
accepts any amount of whitespace, but that doesn't make it right to use
anything other than a single space. (Can the yacc grammar be tweaked to
see "depends on" as one, well, token?)

> +
> +# regex objects
> +REGEX_FILE_KCONFIG = re.compile(r".*Kconfig[\.\w+\-]*$")
> +REGEX_FEATURE = re.compile(r"(" + FEATURE + r")")
> +REGEX_SOURCE_FEATURE = re.compile(r"(?:D|\W|\b)+CONFIG_(" + FEATURE + r")")

That should be "-D", because now you get a hit on
    TPS6507X_ADCONFIG_CONVERT_TS

and friends. What does \W do, by the way?

> +REGEX_KCONFIG_DEF = re.compile(DEF)
> +REGEX_KCONFIG_EXPR = re.compile(EXPR)
> +REGEX_KCONFIG_STMT = re.compile(STMT)
> +REGEX_KCONFIG_HELP = re.compile(r"^\s+(help|---help---)\s*$")
> +REGEX_FILTER_FEATURES = re.compile(r"[A-Za-z0-9]$")
> +
> +
> +def main():
> +    """Main function of this module."""
> +    source_files = []
> +    kconfig_files = []
> +    defined_features = set()
> +    referenced_features = dict()  # {feature: [files]}
> +
> +    # use 'git ls-files' to get the worklist
> +    pop = Popen("git ls-files", stdout=PIPE, stderr=STDOUT, shell=True)
> +    (stdout, _) = pop.communicate()  # wait until finished
> +    if len(stdout) > 0 and stdout[-1] == "\n":
> +        stdout = stdout[:-1]
> +
> +    for gitfile in stdout.rsplit("\n"):
> +        if ".git" in gitfile or "ChangeLog" in gitfile or \
> +                os.path.isdir(gitfile):

(The only directories in the git tree are a few symlinks:
    arch/arm/boot/dts/include/dt-bindings
    arch/metag/boot/dts/include/dt-bindings
    arch/mips/boot/dts/include/dt-bindings
    arch/powerpc/boot/dts/include/dt-bindings

Filtering out symlinks makes sense anyway. But that's probably not worth
the effort.)

You might also consider filtering out "Next/merge.log". It tends to have
references to Kconfig macros. But it's only a few hits anyway.

> +            continue
> +        if REGEX_FILE_KCONFIG.match(gitfile):
> +            kconfig_files.append(gitfile)
> +        else:
> +            # all non-Kconfig files are checked for consistency
> +            source_files.append(gitfile)
> +
> +    for sfile in source_files:
> +        parse_source_file(sfile, referenced_features)
> +
> +    for kfile in kconfig_files:
> +        parse_kconfig_file(kfile, defined_features, referenced_features)
> +
> +    print "Undefined symbol used\tFile list"
> +    for feature in sorted(referenced_features):
> +        # filter some false positives
> +        if feature == "FOO" or feature == "BAR" or \
> +                feature == "FOO_BAR":
> +            continue
> +        if feature not in defined_features:
> +            if feature.endswith("_MODULE"):
> +                # avoid false positives for kernel modules
> +                if feature[:-len("_MODULE")] in defined_features:
> +                    continue
> +            files = referenced_features.get(feature)
> +            print "%s\t%s" % (feature, ", ".join(files))
> +
> +
> +def parse_source_file(sfile, referenced_features):
> +    """Parse @sfile for referenced Kconfig features."""
> +    lines = []
> +    with open(sfile, "r") as stream:
> +        lines = stream.readlines()
> +
> +    for line in lines:
> +        if not "CONFIG_" in line:
> +            continue
> +        features = REGEX_SOURCE_FEATURE.findall(line)
> +        for feature in features:
> +            if not REGEX_FILTER_FEATURES.search(feature):
> +                continue
> +            sfiles = referenced_features.get(feature, set())
> +            sfiles.add(sfile)
> +            referenced_features[feature] = sfiles
> +
> +
> +def get_features_in_line(line):
> +    """Return mentioned Kconfig features in @line."""
> +    return REGEX_FEATURE.findall(line)
> +
> +
> +def parse_kconfig_file(kfile, defined_features, referenced_features):
> +    """Parse @kfile and update feature definitions and references."""
> +    lines = []
> +    skip = False
> +
> +    with open(kfile, "r") as stream:
> +        lines = stream.readlines()
> +
> +    for i in range(len(lines)):
> +        line = lines[i]
> +        line = line.strip('\n')
> +        line = line.split("#")[0]  # ignore comments
> +
> +        if REGEX_KCONFIG_DEF.match(line):
> +            feature_def = REGEX_KCONFIG_DEF.findall(line)
> +            defined_features.add(feature_def[0])
> +            skip = False
> +        elif REGEX_KCONFIG_HELP.match(line):
> +            skip = True
> +        elif skip:
> +            # ignore content of help messages
> +            pass
> +        elif REGEX_KCONFIG_STMT.match(line):
> +            features = get_features_in_line(line)
> +            # multi-line statements
> +            while line.endswith("\\"):
> +                i += 1
> +                line = lines[i]
> +                line = line.strip('\n')
> +                features.extend(get_features_in_line(line))
> +            for feature in set(features):
> +                paths = referenced_features.get(feature, set())
> +                paths.add(kfile)
> +                referenced_features[feature] = paths
> +
> +
> +if __name__ == "__main__":
> +    main()
>[...]

This seems to find, roughly, what my local perl script finds. It does
skip references to Kconfig macros in the Kconfig help texts and
comments: grep for CONFIG_ITANIUM_ASTEP_SPECIFIC as an example. But
there are so few of those that it's probable not worth the trouble to
check for them too.

A few test show the speedup is especially nice with the entire tree in
cache: run time drops from over four minutes to just under five seconds
here. Provided you look into my comments, this is:

Acked-by: Paul Bolle <pebolle@tiscali.nl>

Thanks,


Paul Bolle

Re: [PATCH v7] checkkconfigsymbols.sh: reimplementation in python

[Valentin Rothberg]

On Mon, Sep 29, 2014 at 12:28 PM, Paul Bolle <pebolle@tiscali.nl> wrote:
> [Added  linux-kbuild@vger.kernel.org.]
>
> On Sun, 2014-09-28 at 17:55 +0200, Valentin Rothberg wrote:
>> The scripts/checkkconfigsymbols.sh script searches Kconfig features
>> in the source code that are not defined in Kconfig. Such identifiers
>> always evaluate to false and are the source of various kinds of bugs.
>> However, the shell script is slow and it does not detect such broken
>> references in Kbuild and Kconfig files (e.g., ``depends on UNDEFINED´´).
>> Furthermore, it generates false positives. The script is also hard to
>> read and understand, and is thereby difficult to maintain.
>>
>> This patch replaces the shell script with an implementation in Python,
>> which:
>>     (a) detects the same bugs, but does not report previous false positives
>>     (b) additionally detects broken references in Kconfig and all
>>         non-Kconfig files, such as Kbuild, .[cSh], .txt, .sh, defconfig, etc.
>>     (c) is up to 75 times faster than the shell script
>>     (d) only checks files under version control ('git ls-files')
>
> (The shell script is .git unaware. If you happen to have one or more
> branches with "Kconfig" in their name, as I do, it generates a lot of
> noise on stderr.)

Do you prefer to use os.walk() then? I just don't want to assume that
the script is called in a clean tree. 'git ls-files' avoids to filter
files. How do you solve this issue in your script?

>> The new script reduces the runtime on my machine (i7-2620M, 8GB RAM, SSD)
>> from 3m47s to 0m3s, and reports 939 broken identifiers in Linux v3.17-rc1;
>> 420 additional reports of which 16 are located in Kconfig files,
>> 287 in defconfigs, 63 in ./Documentation, 1 in Kbuild.
>>
>> Moreover, we intentionally include references in comments, which have been
>> ignored until now. Such comments may be leftovers of features that have
>> been removed or renamed in Kconfig (e.g., ``#endif /* CONFIG_MPC52xx */´´).
>> These references can be misleading and should be removed or replaced.
>>
>> Note that the output format changed from (file list <tab> feature) to
>> (feature <tab> file list) as it simplifies the detection of the Kconfig
>> feature for long file lists.
>>
>> Signed-off-by: Valentin Rothberg <valentinrothberg@gmail.com>
>> Signed-off-by: Stefan Hengelein <stefan.hengelein@fau.de>
>> ---
>> Changelog:
>> v2: Fix of regular expressions
>> v3: Changelog replacement, and add changes of v2
>> v4: Based on comments from Paul Bolle <pebolle@tiscali.nl>
>>   - Inclusion of all non-Kconfig files, such as .txt, .sh, etc.
>>   - Changes of regular expressions
>>   - Increases additional reports from 49 to 229 compared to v3
>>   - Change of output format from (file list <tab> feature) to
>>         (feature <tab> file list)·
>> v5: Only analyze files under version control ('git ls-files')
>> v6: Cover features with numbers and small letters (e.g., 4xx)
>> v7: Add changes of v6 (lost 'git add') and filter FOO/BAR features
>> ---
>>  scripts/checkkconfigsymbols.py | 138 +++++++++++++++++++++++++++++++++++++++++
>>  scripts/checkkconfigsymbols.sh |  59 ------------------
>>  2 files changed, 138 insertions(+), 59 deletions(-)
>>  create mode 100644 scripts/checkkconfigsymbols.py
>>  delete mode 100755 scripts/checkkconfigsymbols.sh
>>
>> diff --git a/scripts/checkkconfigsymbols.py b/scripts/checkkconfigsymbols.py
>> new file mode 100644
>> index 0000000..01bf9c4
>> --- /dev/null
>> +++ b/scripts/checkkconfigsymbols.py
>> @@ -0,0 +1,138 @@
>> +#!/usr/bin/env python
>> +
>> +"""Find Kconfig identifiers that are referenced but not defined."""
>> +
>> +# (c) 2014 Valentin Rothberg <valentinrothberg@gmail.com>
>> +# (c) 2014 Stefan Hengelein <stefan.hengelein@fau.de>
>> +#
>> +# Licensed under the terms of the GNU GPL License version 2
>> +
>> +
>> +import os
>> +import re
>> +from subprocess import Popen, PIPE, STDOUT
>> +
>> +
>> +# regex expressions
>> +OPERATORS = r"&|\(|\)|\||\!"
>> +FEATURE = r"(?:\w*[A-Z0-9]\w*){2,}"
>> +DEF = r"^\s*(?:menu){,1}config\s+(" + FEATURE + r")\s*"
>> +EXPR = r"(?:" + OPERATORS + r"|\s|" + FEATURE + r")+"
>> +STMT = r"^\s*(?:if|select|depends\s+on)\s+" + EXPR
>
> Could please make that "depends on"? Yes, it seems the yacc grammar
> accepts any amount of whitespace, but that doesn't make it right to use
> anything other than a single space. (Can the yacc grammar be tweaked to
> see "depends on" as one, well, token?)

I don't know if yacc can do that. Usually the lexer trims whitespaces.
If I change the regex to "depends on", we will miss potential
statements as Kconfig accepts multiple spaces between the "depends"
and the "on".

>> +
>> +# regex objects
>> +REGEX_FILE_KCONFIG = re.compile(r".*Kconfig[\.\w+\-]*$")
>> +REGEX_FEATURE = re.compile(r"(" + FEATURE + r")")
>> +REGEX_SOURCE_FEATURE = re.compile(r"(?:D|\W|\b)+CONFIG_(" + FEATURE + r")")
>
> That should be "-D", because now you get a hit on
>     TPS6507X_ADCONFIG_CONVERT_TS
>
> and friends. What does \W do, by the way?

(?:\W|\b)+[D]{,1}CONFIG_... will do the trick. "-D" would fail in
getting DCONFIG_ in the build files. "\W" matches not "\w", which is
[A-Za-z0-9_].

>> +REGEX_KCONFIG_DEF = re.compile(DEF)
>> +REGEX_KCONFIG_EXPR = re.compile(EXPR)
>> +REGEX_KCONFIG_STMT = re.compile(STMT)
>> +REGEX_KCONFIG_HELP = re.compile(r"^\s+(help|---help---)\s*$")
>> +REGEX_FILTER_FEATURES = re.compile(r"[A-Za-z0-9]$")
>> +
>> +
>> +def main():
>> +    """Main function of this module."""
>> +    source_files = []
>> +    kconfig_files = []
>> +    defined_features = set()
>> +    referenced_features = dict()  # {feature: [files]}
>> +
>> +    # use 'git ls-files' to get the worklist
>> +    pop = Popen("git ls-files", stdout=PIPE, stderr=STDOUT, shell=True)
>> +    (stdout, _) = pop.communicate()  # wait until finished
>> +    if len(stdout) > 0 and stdout[-1] == "\n":
>> +        stdout = stdout[:-1]
>> +
>> +    for gitfile in stdout.rsplit("\n"):
>> +        if ".git" in gitfile or "ChangeLog" in gitfile or \
>> +                os.path.isdir(gitfile):
>
> (The only directories in the git tree are a few symlinks:
>     arch/arm/boot/dts/include/dt-bindings
>     arch/metag/boot/dts/include/dt-bindings
>     arch/mips/boot/dts/include/dt-bindings
>     arch/powerpc/boot/dts/include/dt-bindings
>
> Filtering out symlinks makes sense anyway. But that's probably not worth
> the effort.)
>
> You might also consider filtering out "Next/merge.log". It tends to have
> references to Kconfig macros. But it's only a few hits anyway.

Thanks. I will put ".log" in

>> +            continue
>> +        if REGEX_FILE_KCONFIG.match(gitfile):
>> +            kconfig_files.append(gitfile)
>> +        else:
>> +            # all non-Kconfig files are checked for consistency
>> +            source_files.append(gitfile)
>> +
>> +    for sfile in source_files:
>> +        parse_source_file(sfile, referenced_features)
>> +
>> +    for kfile in kconfig_files:
>> +        parse_kconfig_file(kfile, defined_features, referenced_features)
>> +
>> +    print "Undefined symbol used\tFile list"
>> +    for feature in sorted(referenced_features):
>> +        # filter some false positives
>> +        if feature == "FOO" or feature == "BAR" or \
>> +                feature == "FOO_BAR":
>> +            continue
>> +        if feature not in defined_features:
>> +            if feature.endswith("_MODULE"):
>> +                # avoid false positives for kernel modules
>> +                if feature[:-len("_MODULE")] in defined_features:
>> +                    continue
>> +            files = referenced_features.get(feature)
>> +            print "%s\t%s" % (feature, ", ".join(files))
>> +
>> +
>> +def parse_source_file(sfile, referenced_features):
>> +    """Parse @sfile for referenced Kconfig features."""
>> +    lines = []
>> +    with open(sfile, "r") as stream:
>> +        lines = stream.readlines()
>> +
>> +    for line in lines:
>> +        if not "CONFIG_" in line:
>> +            continue
>> +        features = REGEX_SOURCE_FEATURE.findall(line)
>> +        for feature in features:
>> +            if not REGEX_FILTER_FEATURES.search(feature):
>> +                continue
>> +            sfiles = referenced_features.get(feature, set())
>> +            sfiles.add(sfile)
>> +            referenced_features[feature] = sfiles
>> +
>> +
>> +def get_features_in_line(line):
>> +    """Return mentioned Kconfig features in @line."""
>> +    return REGEX_FEATURE.findall(line)
>> +
>> +
>> +def parse_kconfig_file(kfile, defined_features, referenced_features):
>> +    """Parse @kfile and update feature definitions and references."""
>> +    lines = []
>> +    skip = False
>> +
>> +    with open(kfile, "r") as stream:
>> +        lines = stream.readlines()
>> +
>> +    for i in range(len(lines)):
>> +        line = lines[i]
>> +        line = line.strip('\n')
>> +        line = line.split("#")[0]  # ignore comments
>> +
>> +        if REGEX_KCONFIG_DEF.match(line):
>> +            feature_def = REGEX_KCONFIG_DEF.findall(line)
>> +            defined_features.add(feature_def[0])
>> +            skip = False
>> +        elif REGEX_KCONFIG_HELP.match(line):
>> +            skip = True
>> +        elif skip:
>> +            # ignore content of help messages
>> +            pass
>> +        elif REGEX_KCONFIG_STMT.match(line):
>> +            features = get_features_in_line(line)
>> +            # multi-line statements
>> +            while line.endswith("\\"):
>> +                i += 1
>> +                line = lines[i]
>> +                line = line.strip('\n')
>> +                features.extend(get_features_in_line(line))
>> +            for feature in set(features):
>> +                paths = referenced_features.get(feature, set())
>> +                paths.add(kfile)
>> +                referenced_features[feature] = paths
>> +
>> +
>> +if __name__ == "__main__":
>> +    main()
>>[...]
>
> This seems to find, roughly, what my local perl script finds. It does
> skip references to Kconfig macros in the Kconfig help texts and
> comments: grep for CONFIG_ITANIUM_ASTEP_SPECIFIC as an example. But
> there are so few of those that it's probable not worth the trouble to
> check for them too.
>
> A few test show the speedup is especially nice with the entire tree in
> cache: run time drops from over four minutes to just under five seconds
> here. Provided you look into my comments, this is:
>
> Acked-by: Paul Bolle <pebolle@tiscali.nl>

Thank you,
 Valentin Rothberg

>
> Thanks,
>
>
> Paul Bolle
>

Re: [PATCH v7] checkkconfigsymbols.sh: reimplementation in python

[Paul Bolle]

On Mon, 2014-09-29 at 14:08 +0200, Valentin Rothberg wrote:
> On Mon, Sep 29, 2014 at 12:28 PM, Paul Bolle <pebolle@tiscali.nl> wrote:
> > On Sun, 2014-09-28 at 17:55 +0200, Valentin Rothberg wrote:
> >>     (d) only checks files under version control ('git ls-files')
> >
> > (The shell script is .git unaware. If you happen to have one or more
> > branches with "Kconfig" in their name, as I do, it generates a lot of
> > noise on stderr.)
> 
> Do you prefer to use os.walk() then?

It seems I was unclear: this described a problem with the shell script
and not a problem with your python rewrite.

> I just don't want to assume that
> the script is called in a clean tree. 'git ls-files' avoids to filter
> files. How do you solve this issue in your script?

This is getting off topic, but I choose to not care about the state of
the tree on disk. My script basically does "git ls-tree -r $SOME_TAG",
filters out symlinks, and generates a long list of (blob, path) pairs to
work with.

(You may remember from https://lkml.org/lkml/2014/9/26/485 that my
script uses git notes for each blob it parses. Different approach,
different pros and cons.)

> >> +STMT = r"^\s*(?:if|select|depends\s+on)\s+" + EXPR
> >
> > Could please make that "depends on"? Yes, it seems the yacc grammar
> > accepts any amount of whitespace, but that doesn't make it right to use
> > anything other than a single space. (Can the yacc grammar be tweaked to
> > see "depends on" as one, well, token?)
> 
> I don't know if yacc can do that. Usually the lexer trims whitespaces.
> If I change the regex to "depends on", we will miss potential
> statements as Kconfig accepts multiple spaces between the "depends"
> and the "on".

Your choice. (The expectation behind my request is that bogus results
from this script might help catch uses of this misfeature. But every use
of "depends on" is currently sane, anyway.)

> >> +REGEX_SOURCE_FEATURE = re.compile(r"(?:D|\W|\b)+CONFIG_(" + FEATURE + r")")
> >
> > That should be "-D", because now you get a hit on
> >     TPS6507X_ADCONFIG_CONVERT_TS
> >
> > and friends. What does \W do, by the way?
> 
> (?:\W|\b)+[D]{,1}CONFIG_... will do the trick. "-D" would fail in
> getting DCONFIG_ in the build files. "\W" matches not "\w", which is
> [A-Za-z0-9_].

The joy of regular expressions! You now know what wrong, and you'll
surely come up with something better than my idea to fix it.

Thanks,


Paul Bolle

Re: [PATCH v7] checkkconfigsymbols.sh: reimplementation in python

[Michal Marek]

On 2014-09-29 12:28, Paul Bolle wrote:
>> +STMT = r"^\s*(?:if|select|depends\s+on)\s+" + EXPR
> 
> Could please make that "depends on"? Yes, it seems the yacc grammar
> accepts any amount of whitespace, but that doesn't make it right to use
> anything other than a single space.

But then lines that violate coding style would not be checked for real
errors.


> (Can the yacc grammar be tweaked to
> see "depends on" as one, well, token?)

I don't think this is a good idea. This is a style issue, why make it a
grammar issue.

Michal

Re: [PATCH v7] checkkconfigsymbols.sh: reimplementation in python

[Paul Bolle]

Hi Michal,

[This point I already conceded to Valentin, so my remarks are moot for
Valentin's script.]

On Mon, 2014-09-29 at 16:47 +0200, Michal Marek wrote:
> On 2014-09-29 12:28, Paul Bolle wrote:
> >> +STMT = r"^\s*(?:if|select|depends\s+on)\s+" + EXPR
> > 
> > Could please make that "depends on"? Yes, it seems the yacc grammar
> > accepts any amount of whitespace, but that doesn't make it right to use
> > anything other than a single space.
> 
> But then lines that violate coding style would not be checked for real
> errors.

Perhaps my dislike of this element of the grammar won here. There are
probably better ways to enforce proper use of "depends on".

> > (Can the yacc grammar be tweaked to
> > see "depends on" as one, well, token?)
> 
> I don't think this is a good idea. This is a style issue, why make it a
> grammar issue.

Well, a grammar that allows one of its keywords to be written in
different ways makes style and grammar issues overlap. Whatever. I guess
it's just me being annoyed with writing
    git grep -w "depends\s\+on\s\+FOO"

instead of
    git grep -w "depends on FOO"

Ditto for writing a parser for Kconfig files. (Real world examples are
more complicated, but you catch my drift.)


Paul Bolle

[PATCH v8] checkkconfigsymbols.sh: reimplementation in python

[Valentin Rothberg]

The scripts/checkkconfigsymbols.sh script searches Kconfig features
in the source code that are not defined in Kconfig. Such identifiers
always evaluate to false and are the source of various kinds of bugs.
However, the shell script is slow and it does not detect such broken
references in Kbuild and Kconfig files (e.g., ``depends on UNDEFINED´´).
Furthermore, it generates false positives. The script is also hard to
read and understand, and is thereby difficult to maintain.

This patch replaces the shell script with an implementation in Python,
which:
    (a) detects the same bugs, but does not report previous false positives
    (b) additionally detects broken references in Kconfig and all
        non-Kconfig files, such as Kbuild, .[cSh], .txt, .sh, defconfig, etc.
    (c) is up to 75 times faster than the shell script
    (d) only checks files under version control ('git ls-files')

The new script reduces the runtime on my machine (i7-2620M, 8GB RAM, SSD)
from 3m47s to 0m3s, and reports 912 broken references in Linux v3.17-rc1;
424 additional reports of which 16 are located in Kconfig files,
287 in defconfigs, 63 in ./Documentation, 1 in Kbuild.

Moreover, we intentionally include references in comments, which have been
ignored until now. Such comments may be leftovers of features that have
been removed or renamed in Kconfig (e.g., ``#endif /* CONFIG_MPC52xx */´´).
These references can be misleading and should be removed or replaced.

Note that the output format changed from (file list <tab> feature) to
(feature <tab> file list) as it simplifies the detection of the Kconfig
feature for long file lists.

Signed-off-by: Valentin Rothberg <valentinrothberg@gmail.com>
Signed-off-by: Stefan Hengelein <stefan.hengelein@fau.de>
Acked-by: Paul Bolle <pebolle@tiscali.nl>
---
Changelog:
v2: Fix of regular expressions
v3: Changelog replacement, and add changes of v2
v4: Based on comments from Paul Bolle <pebolle@tiscali.nl>
  - Inclusion of all non-Kconfig files, such as .txt, .sh, etc.
  - Changes of regular expressions
  - Increases additional reports from 49 to 229 compared to v3
  - Change of output format from (file list <tab> feature) to
        (feature <tab> file list)
v5: Only analyze files under version control ('git ls-files')
v6: Cover features with numbers and small letters (e.g., 4xx)
v7: Add changes of v6 (lost 'git add') and filter FOO/BAR features
v8: Based on comments from Paul Bolle <pebolle@tiscali.nl>
  - Fix of [D]{,1}CONFIG_ regex to exclude false positives
  - Exclude ".log" files of analysis
  - Filter "XXX" feature
---
 scripts/checkkconfigsymbols.py | 139 +++++++++++++++++++++++++++++++++++++++++
 scripts/checkkconfigsymbols.sh |  59 -----------------
 2 files changed, 139 insertions(+), 59 deletions(-)
 create mode 100644 scripts/checkkconfigsymbols.py
 delete mode 100755 scripts/checkkconfigsymbols.sh

diff --git a/scripts/checkkconfigsymbols.py b/scripts/checkkconfigsymbols.py
new file mode 100644
index 0000000..e9cc689
--- /dev/null
+++ b/scripts/checkkconfigsymbols.py
@@ -0,0 +1,139 @@
+#!/usr/bin/env python
+
+"""Find Kconfig identifiers that are referenced but not defined."""
+
+# (c) 2014 Valentin Rothberg <valentinrothberg@gmail.com>
+# (c) 2014 Stefan Hengelein <stefan.hengelein@fau.de>
+#
+# Licensed under the terms of the GNU GPL License version 2
+
+
+import os
+import re
+from subprocess import Popen, PIPE, STDOUT
+
+
+# regex expressions
+OPERATORS = r"&|\(|\)|\||\!"
+FEATURE = r"(?:\w*[A-Z0-9]\w*){2,}"
+DEF = r"^\s*(?:menu){,1}config\s+(" + FEATURE + r")\s*"
+EXPR = r"(?:" + OPERATORS + r"|\s|" + FEATURE + r")+"
+STMT = r"^\s*(?:if|select|depends\s+on)\s+" + EXPR
+SOURCE_FEATURE = r"(?:\W|\b)+[D]{,1}CONFIG_(" + FEATURE + r")"
+
+# regex objects
+REGEX_FILE_KCONFIG = re.compile(r".*Kconfig[\.\w+\-]*$")
+REGEX_FEATURE = re.compile(r"(" + FEATURE + r")")
+REGEX_SOURCE_FEATURE = re.compile(SOURCE_FEATURE)
+REGEX_KCONFIG_DEF = re.compile(DEF)
+REGEX_KCONFIG_EXPR = re.compile(EXPR)
+REGEX_KCONFIG_STMT = re.compile(STMT)
+REGEX_KCONFIG_HELP = re.compile(r"^\s+(help|---help---)\s*$")
+REGEX_FILTER_FEATURES = re.compile(r"[A-Za-z0-9]$")
+
+
+def main():
+    """Main function of this module."""
+    source_files = []
+    kconfig_files = []
+    defined_features = set()
+    referenced_features = dict()  # {feature: [files]}
+
+    # use 'git ls-files' to get the worklist
+    pop = Popen("git ls-files", stdout=PIPE, stderr=STDOUT, shell=True)
+    (stdout, _) = pop.communicate()  # wait until finished
+    if len(stdout) > 0 and stdout[-1] == "\n":
+        stdout = stdout[:-1]
+
+    for gitfile in stdout.rsplit("\n"):
+        if ".git" in gitfile or "ChangeLog" in gitfile or \
+                ".log" in gitfile or os.path.isdir(gitfile):
+            continue
+        if REGEX_FILE_KCONFIG.match(gitfile):
+            kconfig_files.append(gitfile)
+        else:
+            # all non-Kconfig files are checked for consistency
+            source_files.append(gitfile)
+
+    for sfile in source_files:
+        parse_source_file(sfile, referenced_features)
+
+    for kfile in kconfig_files:
+        parse_kconfig_file(kfile, defined_features, referenced_features)
+
+    print "Undefined symbol used\tFile list"
+    for feature in sorted(referenced_features):
+        # filter some false positives
+        if feature == "FOO" or feature == "BAR" or \
+                feature == "FOO_BAR" or feature == "XXX":
+            continue
+        if feature not in defined_features:
+            if feature.endswith("_MODULE"):
+                # avoid false positives for kernel modules
+                if feature[:-len("_MODULE")] in defined_features:
+                    continue
+            files = referenced_features.get(feature)
+            print "%s\t%s" % (feature, ", ".join(files))
+
+
+def parse_source_file(sfile, referenced_features):
+    """Parse @sfile for referenced Kconfig features."""
+    lines = []
+    with open(sfile, "r") as stream:
+        lines = stream.readlines()
+
+    for line in lines:
+        if not "CONFIG_" in line:
+            continue
+        features = REGEX_SOURCE_FEATURE.findall(line)
+        for feature in features:
+            if not REGEX_FILTER_FEATURES.search(feature):
+                continue
+            sfiles = referenced_features.get(feature, set())
+            sfiles.add(sfile)
+            referenced_features[feature] = sfiles
+
+
+def get_features_in_line(line):
+    """Return mentioned Kconfig features in @line."""
+    return REGEX_FEATURE.findall(line)
+
+
+def parse_kconfig_file(kfile, defined_features, referenced_features):
+    """Parse @kfile and update feature definitions and references."""
+    lines = []
+    skip = False
+
+    with open(kfile, "r") as stream:
+        lines = stream.readlines()
+
+    for i in range(len(lines)):
+        line = lines[i]
+        line = line.strip('\n')
+        line = line.split("#")[0]  # ignore comments
+
+        if REGEX_KCONFIG_DEF.match(line):
+            feature_def = REGEX_KCONFIG_DEF.findall(line)
+            defined_features.add(feature_def[0])
+            skip = False
+        elif REGEX_KCONFIG_HELP.match(line):
+            skip = True
+        elif skip:
+            # ignore content of help messages
+            pass
+        elif REGEX_KCONFIG_STMT.match(line):
+            features = get_features_in_line(line)
+            # multi-line statements
+            while line.endswith("\\"):
+                i += 1
+                line = lines[i]
+                line = line.strip('\n')
+                features.extend(get_features_in_line(line))
+            for feature in set(features):
+                paths = referenced_features.get(feature, set())
+                paths.add(kfile)
+                referenced_features[feature] = paths
+
+
+if __name__ == "__main__":
+    main()
diff --git a/scripts/checkkconfigsymbols.sh b/scripts/checkkconfigsymbols.sh
deleted file mode 100755
index ccb3391..0000000
--- a/scripts/checkkconfigsymbols.sh
+++ /dev/null
@@ -1,59 +0,0 @@
-#!/bin/sh
-# Find Kconfig variables used in source code but never defined in Kconfig
-# Copyright (C) 2007, Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it>
-
-# Tested with dash.
-paths="$@"
-[ -z "$paths" ] && paths=.
-
-# Doing this once at the beginning saves a lot of time, on a cache-hot tree.
-Kconfigs="`find . -name 'Kconfig' -o -name 'Kconfig*[^~]'`"
-
-printf "File list \tundefined symbol used\n"
-find $paths -name '*.[chS]' -o -name 'Makefile' -o -name 'Makefile*[^~]'| while read i
-do
-	# Output the bare Kconfig variable and the filename; the _MODULE part at
-	# the end is not removed here (would need perl an not-hungry regexp for that).
-	sed -ne 's!^.*\<\(UML_\)\?CONFIG_\([0-9A-Za-z_]\+\).*!\2 '$i'!p' < $i
-done | \
-# Smart "sort|uniq" implemented in awk and tuned to collect the names of all
-# files which use a given symbol
-awk '{map[$1, count[$1]++] = $2; }
-END {
-	for (combIdx in map) {
-		split(combIdx, separate, SUBSEP);
-		# The value may have been removed.
-		if (! ( (separate[1], separate[2]) in map ) )
-			continue;
-		symb=separate[1];
-		printf "%s ", symb;
-		#Use gawk extension to delete the names vector
-		delete names;
-		#Portably delete the names vector
-		#split("", names);
-		for (i=0; i < count[symb]; i++) {
-			names[map[symb, i]] = 1;
-			# Unfortunately, we may still encounter symb, i in the
-			# outside iteration.
-			delete map[symb, i];
-		}
-		i=0;
-		for (name in names) {
-			if (i > 0)
-				printf ", %s", name;
-			else
-				printf "%s", name;
-			i++;
-		}
-		printf "\n";
-	}
-}' |
-while read symb files; do
-	# Remove the _MODULE suffix when checking the variable name. This should
-	# be done only on tristate symbols, actually, but Kconfig parsing is
-	# beyond the purpose of this script.
-	symb_bare=`echo $symb | sed -e 's/_MODULE//'`
-	if ! grep -q "\<$symb_bare\>" $Kconfigs; then
-		printf "$files: \t$symb\n"
-	fi
-done|sort
--
1.9.3

Re: [PATCH v8] checkkconfigsymbols.sh: reimplementation in python

[Michal Marek]

On 2014-09-29 19:05, Valentin Rothberg wrote:
> The scripts/checkkconfigsymbols.sh script searches Kconfig features
> in the source code that are not defined in Kconfig. Such identifiers
> always evaluate to false and are the source of various kinds of bugs.
> However, the shell script is slow and it does not detect such broken
> references in Kbuild and Kconfig files (e.g., ``depends on UNDEFINED´´).
> Furthermore, it generates false positives. The script is also hard to
> read and understand, and is thereby difficult to maintain.
> 
> This patch replaces the shell script with an implementation in Python,
> which:
>     (a) detects the same bugs, but does not report previous false positives
>     (b) additionally detects broken references in Kconfig and all
>         non-Kconfig files, such as Kbuild, .[cSh], .txt, .sh, defconfig, etc.
>     (c) is up to 75 times faster than the shell script
>     (d) only checks files under version control ('git ls-files')
> 
> The new script reduces the runtime on my machine (i7-2620M, 8GB RAM, SSD)
> from 3m47s to 0m3s, and reports 912 broken references in Linux v3.17-rc1;
> 424 additional reports of which 16 are located in Kconfig files,
> 287 in defconfigs, 63 in ./Documentation, 1 in Kbuild.
> 
> Moreover, we intentionally include references in comments, which have been
> ignored until now. Such comments may be leftovers of features that have
> been removed or renamed in Kconfig (e.g., ``#endif /* CONFIG_MPC52xx */´´).
> These references can be misleading and should be removed or replaced.
> 
> Note that the output format changed from (file list <tab> feature) to
> (feature <tab> file list) as it simplifies the detection of the Kconfig
> feature for long file lists.
> 
> Signed-off-by: Valentin Rothberg <valentinrothberg@gmail.com>
> Signed-off-by: Stefan Hengelein <stefan.hengelein@fau.de>
> Acked-by: Paul Bolle <pebolle@tiscali.nl>
> ---
> Changelog:
> v2: Fix of regular expressions
> v3: Changelog replacement, and add changes of v2
> v4: Based on comments from Paul Bolle <pebolle@tiscali.nl>
>   - Inclusion of all non-Kconfig files, such as .txt, .sh, etc.
>   - Changes of regular expressions
>   - Increases additional reports from 49 to 229 compared to v3
>   - Change of output format from (file list <tab> feature) to
>         (feature <tab> file list)
> v5: Only analyze files under version control ('git ls-files')
> v6: Cover features with numbers and small letters (e.g., 4xx)
> v7: Add changes of v6 (lost 'git add') and filter FOO/BAR features
> v8: Based on comments from Paul Bolle <pebolle@tiscali.nl>
>   - Fix of [D]{,1}CONFIG_ regex to exclude false positives
>   - Exclude ".log" files of analysis
>   - Filter "XXX" feature
> ---
>  scripts/checkkconfigsymbols.py | 139 +++++++++++++++++++++++++++++++++++++++++
>  scripts/checkkconfigsymbols.sh |  59 -----------------
>  2 files changed, 139 insertions(+), 59 deletions(-)
>  create mode 100644 scripts/checkkconfigsymbols.py
>  delete mode 100755 scripts/checkkconfigsymbols.sh

Please make the new file executable as well.


> +    for gitfile in stdout.rsplit("\n"):
> +        if ".git" in gitfile or "ChangeLog" in gitfile or \
> +                ".log" in gitfile or os.path.isdir(gitfile):
> +            continue

Can you also skip arch/*/configs? A significant part of the output are
defconfig files, but there is little value in reporting them. The stale
options will go away automatically as soon as the defconfig is refreshed.

Michal

Re: [PATCH v8] checkkconfigsymbols.sh: reimplementation in python

[Valentin Rothberg]

On Wed, Oct 1, 2014 at 4:58 PM, Michal Marek <mmarek@suse.cz> wrote:
> On 2014-09-29 19:05, Valentin Rothberg wrote:
>> The scripts/checkkconfigsymbols.sh script searches Kconfig features
>> in the source code that are not defined in Kconfig. Such identifiers
>> always evaluate to false and are the source of various kinds of bugs.
>> However, the shell script is slow and it does not detect such broken
>> references in Kbuild and Kconfig files (e.g., ``depends on UNDEFINED´´).
>> Furthermore, it generates false positives. The script is also hard to
>> read and understand, and is thereby difficult to maintain.
>>
>> This patch replaces the shell script with an implementation in Python,
>> which:
>>     (a) detects the same bugs, but does not report previous false positives
>>     (b) additionally detects broken references in Kconfig and all
>>         non-Kconfig files, such as Kbuild, .[cSh], .txt, .sh, defconfig, etc.
>>     (c) is up to 75 times faster than the shell script
>>     (d) only checks files under version control ('git ls-files')
>>
>> The new script reduces the runtime on my machine (i7-2620M, 8GB RAM, SSD)
>> from 3m47s to 0m3s, and reports 912 broken references in Linux v3.17-rc1;
>> 424 additional reports of which 16 are located in Kconfig files,
>> 287 in defconfigs, 63 in ./Documentation, 1 in Kbuild.
>>
>> Moreover, we intentionally include references in comments, which have been
>> ignored until now. Such comments may be leftovers of features that have
>> been removed or renamed in Kconfig (e.g., ``#endif /* CONFIG_MPC52xx */´´).
>> These references can be misleading and should be removed or replaced.
>>
>> Note that the output format changed from (file list <tab> feature) to
>> (feature <tab> file list) as it simplifies the detection of the Kconfig
>> feature for long file lists.
>>
>> Signed-off-by: Valentin Rothberg <valentinrothberg@gmail.com>
>> Signed-off-by: Stefan Hengelein <stefan.hengelein@fau.de>
>> Acked-by: Paul Bolle <pebolle@tiscali.nl>
>> ---
>> Changelog:
>> v2: Fix of regular expressions
>> v3: Changelog replacement, and add changes of v2
>> v4: Based on comments from Paul Bolle <pebolle@tiscali.nl>
>>   - Inclusion of all non-Kconfig files, such as .txt, .sh, etc.
>>   - Changes of regular expressions
>>   - Increases additional reports from 49 to 229 compared to v3
>>   - Change of output format from (file list <tab> feature) to
>>         (feature <tab> file list)
>> v5: Only analyze files under version control ('git ls-files')
>> v6: Cover features with numbers and small letters (e.g., 4xx)
>> v7: Add changes of v6 (lost 'git add') and filter FOO/BAR features
>> v8: Based on comments from Paul Bolle <pebolle@tiscali.nl>
>>   - Fix of [D]{,1}CONFIG_ regex to exclude false positives
>>   - Exclude ".log" files of analysis
>>   - Filter "XXX" feature
>> ---
>>  scripts/checkkconfigsymbols.py | 139 +++++++++++++++++++++++++++++++++++++++++
>>  scripts/checkkconfigsymbols.sh |  59 -----------------
>>  2 files changed, 139 insertions(+), 59 deletions(-)
>>  create mode 100644 scripts/checkkconfigsymbols.py
>>  delete mode 100755 scripts/checkkconfigsymbols.sh
>
> Please make the new file executable as well.
>
>
>> +    for gitfile in stdout.rsplit("\n"):
>> +        if ".git" in gitfile or "ChangeLog" in gitfile or \
>> +                ".log" in gitfile or os.path.isdir(gitfile):
>> +            continue
>
> Can you also skip arch/*/configs? A significant part of the output are
> defconfig files, but there is little value in reporting them. The stale
> options will go away automatically as soon as the defconfig is refreshed.

What do you mean by "refreshed"? And how often are they refreshed? I
think that reporting defconfigs helps to point to a problem, namely
that features are just not present anymore.

Thanks,
 Valentin

Re: [PATCH v8] checkkconfigsymbols.sh: reimplementation in python

[Michal Marek]

On 2014-10-04 11:29, Valentin Rothberg wrote:
> On Wed, Oct 1, 2014 at 4:58 PM, Michal Marek <mmarek@suse.cz> wrote:
>> On 2014-09-29 19:05, Valentin Rothberg wrote:
>>> The scripts/checkkconfigsymbols.sh script searches Kconfig features
>>> in the source code that are not defined in Kconfig. Such identifiers
>>> always evaluate to false and are the source of various kinds of bugs.
>>> However, the shell script is slow and it does not detect such broken
>>> references in Kbuild and Kconfig files (e.g., ``depends on UNDEFINED´´).
>>> Furthermore, it generates false positives. The script is also hard to
>>> read and understand, and is thereby difficult to maintain.
>>>
>>> This patch replaces the shell script with an implementation in Python,
>>> which:
>>>     (a) detects the same bugs, but does not report previous false positives
>>>     (b) additionally detects broken references in Kconfig and all
>>>         non-Kconfig files, such as Kbuild, .[cSh], .txt, .sh, defconfig, etc.
>>>     (c) is up to 75 times faster than the shell script
>>>     (d) only checks files under version control ('git ls-files')
>>>
>>> The new script reduces the runtime on my machine (i7-2620M, 8GB RAM, SSD)
>>> from 3m47s to 0m3s, and reports 912 broken references in Linux v3.17-rc1;
>>> 424 additional reports of which 16 are located in Kconfig files,
>>> 287 in defconfigs, 63 in ./Documentation, 1 in Kbuild.
>>>
>>> Moreover, we intentionally include references in comments, which have been
>>> ignored until now. Such comments may be leftovers of features that have
>>> been removed or renamed in Kconfig (e.g., ``#endif /* CONFIG_MPC52xx */´´).
>>> These references can be misleading and should be removed or replaced.
>>>
>>> Note that the output format changed from (file list <tab> feature) to
>>> (feature <tab> file list) as it simplifies the detection of the Kconfig
>>> feature for long file lists.
>>>
>>> Signed-off-by: Valentin Rothberg <valentinrothberg@gmail.com>
>>> Signed-off-by: Stefan Hengelein <stefan.hengelein@fau.de>
>>> Acked-by: Paul Bolle <pebolle@tiscali.nl>
>>> ---
>>> Changelog:
>>> v2: Fix of regular expressions
>>> v3: Changelog replacement, and add changes of v2
>>> v4: Based on comments from Paul Bolle <pebolle@tiscali.nl>
>>>   - Inclusion of all non-Kconfig files, such as .txt, .sh, etc.
>>>   - Changes of regular expressions
>>>   - Increases additional reports from 49 to 229 compared to v3
>>>   - Change of output format from (file list <tab> feature) to
>>>         (feature <tab> file list)
>>> v5: Only analyze files under version control ('git ls-files')
>>> v6: Cover features with numbers and small letters (e.g., 4xx)
>>> v7: Add changes of v6 (lost 'git add') and filter FOO/BAR features
>>> v8: Based on comments from Paul Bolle <pebolle@tiscali.nl>
>>>   - Fix of [D]{,1}CONFIG_ regex to exclude false positives
>>>   - Exclude ".log" files of analysis
>>>   - Filter "XXX" feature
>>> ---
>>>  scripts/checkkconfigsymbols.py | 139 +++++++++++++++++++++++++++++++++++++++++
>>>  scripts/checkkconfigsymbols.sh |  59 -----------------
>>>  2 files changed, 139 insertions(+), 59 deletions(-)
>>>  create mode 100644 scripts/checkkconfigsymbols.py
>>>  delete mode 100755 scripts/checkkconfigsymbols.sh
>>
>> Please make the new file executable as well.
>>
>>
>>> +    for gitfile in stdout.rsplit("\n"):
>>> +        if ".git" in gitfile or "ChangeLog" in gitfile or \
>>> +                ".log" in gitfile or os.path.isdir(gitfile):
>>> +            continue
>>
>> Can you also skip arch/*/configs? A significant part of the output are
>> defconfig files, but there is little value in reporting them. The stale
>> options will go away automatically as soon as the defconfig is refreshed.
> 
> What do you mean by "refreshed"?

Generated again with new Kconfig data.


> And how often are they refreshed? I
> think that reporting defconfigs helps to point to a problem, namely
> that features are just not present anymore.

Any given defconfig is outdated most of the time, so you could as well
list them all :). In most of the cases, it does not matter though,
thanks to the defaults in Kconfig files.

Michal

Re: [PATCH v8] checkkconfigsymbols.sh: reimplementation in python

[Valentin Rothberg]

On Wed, Oct 8, 2014 at 3:39 PM, Michal Marek <mmarek@suse.cz> wrote:
> On 2014-10-04 11:29, Valentin Rothberg wrote:
>> On Wed, Oct 1, 2014 at 4:58 PM, Michal Marek <mmarek@suse.cz> wrote:
>>> On 2014-09-29 19:05, Valentin Rothberg wrote:
>>>> The scripts/checkkconfigsymbols.sh script searches Kconfig features
>>>> in the source code that are not defined in Kconfig. Such identifiers
>>>> always evaluate to false and are the source of various kinds of bugs.
>>>> However, the shell script is slow and it does not detect such broken
>>>> references in Kbuild and Kconfig files (e.g., ``depends on UNDEFINED´´).
>>>> Furthermore, it generates false positives. The script is also hard to
>>>> read and understand, and is thereby difficult to maintain.
>>>>
>>>> This patch replaces the shell script with an implementation in Python,
>>>> which:
>>>>     (a) detects the same bugs, but does not report previous false positives
>>>>     (b) additionally detects broken references in Kconfig and all
>>>>         non-Kconfig files, such as Kbuild, .[cSh], .txt, .sh, defconfig, etc.
>>>>     (c) is up to 75 times faster than the shell script
>>>>     (d) only checks files under version control ('git ls-files')
>>>>
>>>> The new script reduces the runtime on my machine (i7-2620M, 8GB RAM, SSD)
>>>> from 3m47s to 0m3s, and reports 912 broken references in Linux v3.17-rc1;
>>>> 424 additional reports of which 16 are located in Kconfig files,
>>>> 287 in defconfigs, 63 in ./Documentation, 1 in Kbuild.
>>>>
>>>> Moreover, we intentionally include references in comments, which have been
>>>> ignored until now. Such comments may be leftovers of features that have
>>>> been removed or renamed in Kconfig (e.g., ``#endif /* CONFIG_MPC52xx */´´).
>>>> These references can be misleading and should be removed or replaced.
>>>>
>>>> Note that the output format changed from (file list <tab> feature) to
>>>> (feature <tab> file list) as it simplifies the detection of the Kconfig
>>>> feature for long file lists.
>>>>
>>>> Signed-off-by: Valentin Rothberg <valentinrothberg@gmail.com>
>>>> Signed-off-by: Stefan Hengelein <stefan.hengelein@fau.de>
>>>> Acked-by: Paul Bolle <pebolle@tiscali.nl>
>>>> ---
>>>> Changelog:
>>>> v2: Fix of regular expressions
>>>> v3: Changelog replacement, and add changes of v2
>>>> v4: Based on comments from Paul Bolle <pebolle@tiscali.nl>
>>>>   - Inclusion of all non-Kconfig files, such as .txt, .sh, etc.
>>>>   - Changes of regular expressions
>>>>   - Increases additional reports from 49 to 229 compared to v3
>>>>   - Change of output format from (file list <tab> feature) to
>>>>         (feature <tab> file list)
>>>> v5: Only analyze files under version control ('git ls-files')
>>>> v6: Cover features with numbers and small letters (e.g., 4xx)
>>>> v7: Add changes of v6 (lost 'git add') and filter FOO/BAR features
>>>> v8: Based on comments from Paul Bolle <pebolle@tiscali.nl>
>>>>   - Fix of [D]{,1}CONFIG_ regex to exclude false positives
>>>>   - Exclude ".log" files of analysis
>>>>   - Filter "XXX" feature
>>>> ---
>>>>  scripts/checkkconfigsymbols.py | 139 +++++++++++++++++++++++++++++++++++++++++
>>>>  scripts/checkkconfigsymbols.sh |  59 -----------------
>>>>  2 files changed, 139 insertions(+), 59 deletions(-)
>>>>  create mode 100644 scripts/checkkconfigsymbols.py
>>>>  delete mode 100755 scripts/checkkconfigsymbols.sh
>>>
>>> Please make the new file executable as well.
>>>
>>>
>>>> +    for gitfile in stdout.rsplit("\n"):
>>>> +        if ".git" in gitfile or "ChangeLog" in gitfile or \
>>>> +                ".log" in gitfile or os.path.isdir(gitfile):
>>>> +            continue
>>>
>>> Can you also skip arch/*/configs? A significant part of the output are
>>> defconfig files, but there is little value in reporting them. The stale
>>> options will go away automatically as soon as the defconfig is refreshed.
>>
>> What do you mean by "refreshed"?
>
> Generated again with new Kconfig data.
>
>
>> And how often are they refreshed? I
>> think that reporting defconfigs helps to point to a problem, namely
>> that features are just not present anymore.
>
> Any given defconfig is outdated most of the time, so you could as well
> list them all :). In most of the cases, it does not matter though,
> thanks to the defaults in Kconfig files.

I talked with Greg about this issue on Linux Plumbers last week in
Dusseldorf. It seems that there are people sending and accepting
patches which address exactly the problem of outdated Kconfig options
in defconfigs.

Michal, I understand your arguments, but I feel that reporting
defconfigs would help people doing this kind of work. As a
consequence, I suggest to keep the patch as it is and wait for
feedback. In case defconfigs are too annoying, we can still change it
later.

What do you think?

Valentin