Skip to contentSkip to conversation

checkpatch: add shebang check to EXECUTE_PERMISSIONS

RFC PATCH In review Ujjwal Kumar 2 participants 7 messages Oct 11, 2020 Checkpatch

lore
Share & tools
lore
#0 lore raw

RFC PATCH 1/1 checkpatch: add shebang check to EXECUTE_PERMISSIONS

checkpatch script checks for invalid EXECUTE_PERMISSIONS on source files. The script leverages filename extensions and its path in the repository to decide whether to allow execute permissions on the file or not.

Based on current check conditions, a perl script file without '.pl' extension in its filename and not belonging to 'scripts/' directory is reported as ERROR which is a false-positive.

The script can correctly handle patches with mode changes and shebang line if shebang is taken into account. So, along with the current check conditions, adding the shebang check in the check conditions can improve the reports of the script.

---
 scripts/checkpatch.pl | 19 +++++++++++++++++++
 1 file changed, 19 insertions(+)
scripts/checkpatch.pl +19 -0
index fab38b493cef..e596d30794bf 100755
@@ -1795,6 +1795,23 @@ sub get_stat_here {
 	return $herectx;
 }
 
+sub get_shebang {
+	my ($linenr, $realfile) = @_;
+	my $rawline = "";
+	my $shebang = "";
+
+	$rawline = raw_line($linenr, 3);
+	if (defined $rawline &&
+		$rawline =~ /^\@\@ -\d+(?:,\d+)? \+(\d+)(,(\d+))? \@\@/) {
+		if (defined $1 && $1 == 1) {
+			$shebang = raw_line($linenr, 4);
+			$shebang = substr $shebang, 1;
+		}
+	}
+
+	return $shebang;
+}
+
 sub cat_vet {
 	my ($vet) = @_;
 	my ($res, $coded);
@@ -2680,7 +2697,9 @@ sub process {
 # Check for incorrect file permissions
 		if ($line =~ /^new (file )?mode.*[7531]\d{0,2}$/) {
 			my $permhere = $here . "FILE: $realfile\n";
+			my $shebang = get_shebang($linenr, $realfile);
 			if ($realfile !~ m@scripts/@ &&
+			    $shebang !~ /^#!\s*(\/\w)+.*/ &&
 			    $realfile !~ /\.(py|pl|awk|sh)$/) {
 				ERROR("EXECUTE_PERMISSIONS",
 				      "do not set execute permissions for source files\n" . $permhere);

base-commit: d67bc7812221606e1886620a357b13f906814af7

Signature · 6 lines
2.26.2

_______________________________________________
Linux-kernel-mentees mailing list
[email protected]
https://lists.linuxfoundation.org/mailman/listinfo/linux-kernel-mentees
#1 lore raw
Quoted 14 lines · Ujjwal Kumar

checkpatch script checks for invalid EXECUTE_PERMISSIONS on source files. The script leverages filename extensions and its path in the repository to decide whether to allow execute permissions on the file or not.

Based on current check conditions, a perl script file without '.pl' extension in its filename and not belonging to 'scripts/' directory is reported as ERROR which is a false-positive.

The script can correctly handle patches with mode changes and shebang line if shebang is taken into account. So, along with the current check conditions, adding the shebang check in the check conditions can improve the reports of the script.

Quoted 4 lines
---
 scripts/checkpatch.pl | 19 +++++++++++++++++++
 1 file changed, 19 insertions(+)

To test the changes, I used the following scripts:

for i in $(find -executable -type f | grep "\.[a-z]" -v | grep "scripts/" -v); \
do \
hs=$(git log --reverse --format=%H $i | head -1); \
echo $hs; \
done | sort | uniq | \
xargs -n1 -P7 scripts/checkpatch.pl --show-types --types EXECUTE_PERMISSIONS --terse -g

Before patching, the script shows 21 false-positives. After applying the patch, the ERRORS are gone.

Thanks
Ujjwal Kumar
_______________________________________________
Linux-kernel-mentees mailing list
[email protected]
https://lists.linuxfoundation.org/mailman/listinfo/linux-kernel-mentees

#2 lore raw
In reply to Ujjwal Kumar
Quoted 5 lines · Ujjwal Kumar

checkpatch script checks for invalid EXECUTE_PERMISSIONS on source files. The script leverages filename extensions and its path in the repository to decide whether to allow execute permissions on the file or not.

Quoted 9 lines

Based on current check conditions, a perl script file without '.pl' extension in its filename and not belonging to 'scripts/' directory is reported as ERROR which is a false-positive.

The script can correctly handle patches with mode changes and shebang line if shebang is taken into account. So, along with the current check conditions, adding the shebang check in the check conditions can improve the reports of the script.

I think one of the core design decisions of checkpatch.pl is:

checkpatch.pl can run on a patch, even if the patch does not apply to the current repository version that is checked out.

It solely uses the information in the patch, and does not try to guess how it could be applied etc.

This patch violates that core design decisions.

You can propose to Joe Perches and lkml, but do not be surprised if that is rejected because of this reason above.

I would be interested in the discussion.

Lukas

Quoted 44 lines
---
 scripts/checkpatch.pl | 19 +++++++++++++++++++
 1 file changed, 19 insertions(+)
scripts/checkpatch.pl +19 -0
index fab38b493cef..e596d30794bf 100755
@@ -1795,6 +1795,23 @@ sub get_stat_here {
 	return $herectx;
 }
 
+sub get_shebang {
+	my ($linenr, $realfile) = @_;
+	my $rawline = "";
+	my $shebang = "";
+
+	$rawline = raw_line($linenr, 3);
+	if (defined $rawline &&
+		$rawline =~ /^\@\@ -\d+(?:,\d+)? \+(\d+)(,(\d+))? \@\@/) {
+		if (defined $1 && $1 == 1) {
+			$shebang = raw_line($linenr, 4);
+			$shebang = substr $shebang, 1;
+		}
+	}
+
+	return $shebang;
+}
+
 sub cat_vet {
 	my ($vet) = @_;
 	my ($res, $coded);
@@ -2680,7 +2697,9 @@ sub process {
 # Check for incorrect file permissions
 		if ($line =~ /^new (file )?mode.*[7531]\d{0,2}$/) {
 			my $permhere = $here . "FILE: $realfile\n";
+			my $shebang = get_shebang($linenr, $realfile);
 			if ($realfile !~ m@scripts/@ &&
+			    $shebang !~ /^#!\s*(\/\w)+.*/ &&
 			    $realfile !~ /\.(py|pl|awk|sh)$/) {
 				ERROR("EXECUTE_PERMISSIONS",
 				      "do not set execute permissions for source files\n" . $permhere);
Quoted 5 lines

base-commit: d67bc7812221606e1886620a357b13f906814af7

Signature · 1 line
2.26.2

_______________________________________________
Linux-kernel-mentees mailing list
[email protected]
https://lists.linuxfoundation.org/mailman/listinfo/linux-kernel-mentees

#3 lore raw
Quoted 19 lines · Lukas Bulwahn
Quoted 5 lines · Ujjwal Kumar

checkpatch script checks for invalid EXECUTE_PERMISSIONS on source files. The script leverages filename extensions and its path in the repository to decide whether to allow execute permissions on the file or not.

Quoted 9 lines

Based on current check conditions, a perl script file without '.pl' extension in its filename and not belonging to 'scripts/' directory is reported as ERROR which is a false-positive.

The script can correctly handle patches with mode changes and shebang line if shebang is taken into account. So, along with the current check conditions, adding the shebang check in the check conditions can improve the reports of the script.

Quoted 4 lines

I think one of the core design decisions of checkpatch.pl is:

checkpatch.pl can run on a patch, even if the patch does not apply to the current repository version that is checked out.

From our past conversation I remember about this particular point.

Quoted 3 lines

It solely uses the information in the patch, and does not try to guess how it could be applied etc.

I am fetching the 'shebang' from the patch itself (therefore I do not understand how does the proposed change violate that design decision?).

Quoted 2 lines

This patch violates that core design decisions.

Can you please point out the exact change that violates the design decisions?

Quoted 9 lines

You can propose to Joe Perches and lkml, but do not be surprised if that is rejected because of this reason above.

I would be interested in the discussion.

Lukas

Thanks
Ujjwal Kumar
_______________________________________________
Linux-kernel-mentees mailing list
[email protected]
https://lists.linuxfoundation.org/mailman/listinfo/linux-kernel-mentees

#4 lore raw
Quoted 25 lines · Ujjwal Kumar
Quoted 19 lines · Lukas Bulwahn
Quoted 5 lines · Ujjwal Kumar

checkpatch script checks for invalid EXECUTE_PERMISSIONS on source files. The script leverages filename extensions and its path in the repository to decide whether to allow execute permissions on the file or not.

Quoted 9 lines

Based on current check conditions, a perl script file without '.pl' extension in its filename and not belonging to 'scripts/' directory is reported as ERROR which is a false-positive.

The script can correctly handle patches with mode changes and shebang line if shebang is taken into account. So, along with the current check conditions, adding the shebang check in the check conditions can improve the reports of the script.

Quoted 4 lines

I think one of the core design decisions of checkpatch.pl is:

checkpatch.pl can run on a patch, even if the patch does not apply to the current repository version that is checked out.

Quoted 9 lines

From our past conversation I remember about this particular point.

Quoted 3 lines

It solely uses the information in the patch, and does not try to guess how it could be applied etc.

I am fetching the 'shebang' from the patch itself (therefore I do not understand how does the proposed change violate that design decision?).

Okay, maybe I misread the patch; so, where those the first line come from? What if that first line is not part of the patch?

Quoted 9 lines
Quoted 2 lines

This patch violates that core design decisions.

Can you please point out the exact change that violates the design decisions?

Quoted 4 lines

You can propose to Joe Perches and lkml, but do not be surprised if that is rejected because of this reason above.

Quoted 9 lines
Quoted 5 lines

I would be interested in the discussion.

Lukas

Thanks
Ujjwal Kumar

_______________________________________________
Linux-kernel-mentees mailing list
[email protected]
https://lists.linuxfoundation.org/mailman/listinfo/linux-kernel-mentees

#5 lore raw
Quoted 35 lines · Lukas Bulwahn
Quoted 31 lines · Ujjwal Kumar
Quoted 19 lines · Lukas Bulwahn
Quoted 5 lines · Ujjwal Kumar

checkpatch script checks for invalid EXECUTE_PERMISSIONS on source files. The script leverages filename extensions and its path in the repository to decide whether to allow execute permissions on the file or not.

Quoted 9 lines

Based on current check conditions, a perl script file without '.pl' extension in its filename and not belonging to 'scripts/' directory is reported as ERROR which is a false-positive.

The script can correctly handle patches with mode changes and shebang line if shebang is taken into account. So, along with the current check conditions, adding the shebang check in the check conditions can improve the reports of the script.

Quoted 4 lines

I think one of the core design decisions of checkpatch.pl is:

checkpatch.pl can run on a patch, even if the patch does not apply to the current repository version that is checked out.

From our past conversation I remember about this particular point.

Quoted 3 lines

It solely uses the information in the patch, and does not try to guess how it could be applied etc.

Quoted 6 lines
Quoted 3 lines

I am fetching the 'shebang' from the patch itself (therefore I do not understand how does the proposed change violate that design decision?).

Okay, maybe I misread the patch; so, where those the first line come from? What if that first line is not part of the patch?

In that case there might be a false-positive. But I tried to handle such cases where the first line is in the patch itself.

Okay, so there are total of 4 cases with file mode change:
1. mode change + create + modify
2. mode change + create
3. mode change + modify
4. mode change

the patch will handle all cases of type 1 and some cases of type 3. Other cases won't be handled because they cannot be handled without breaking the core design decision.

I do have references to say that the cases 2 and 4 are

Which I thought is an improvement over current logic used for the check.

Correct me if I am wrong.

Thanks
Ujjwal Kumar
_______________________________________________
Linux-kernel-mentees mailing list
[email protected]
https://lists.linuxfoundation.org/mailman/listinfo/linux-kernel-mentees

#6 lore raw
Quoted 53 lines · Ujjwal Kumar
Quoted 35 lines · Lukas Bulwahn
Quoted 31 lines · Ujjwal Kumar
Quoted 19 lines · Lukas Bulwahn
Quoted 5 lines · Ujjwal Kumar

checkpatch script checks for invalid EXECUTE_PERMISSIONS on source files. The script leverages filename extensions and its path in the repository to decide whether to allow execute permissions on the file or not.

Quoted 9 lines

Based on current check conditions, a perl script file without '.pl' extension in its filename and not belonging to 'scripts/' directory is reported as ERROR which is a false-positive.

The script can correctly handle patches with mode changes and shebang line if shebang is taken into account. So, along with the current check conditions, adding the shebang check in the check conditions can improve the reports of the script.

Quoted 4 lines

I think one of the core design decisions of checkpatch.pl is:

checkpatch.pl can run on a patch, even if the patch does not apply to the current repository version that is checked out.

From our past conversation I remember about this particular point.

Quoted 3 lines

It solely uses the information in the patch, and does not try to guess how it could be applied etc.

Quoted 6 lines
Quoted 3 lines

I am fetching the 'shebang' from the patch itself (therefore I do not understand how does the proposed change violate that design decision?).

Okay, maybe I misread the patch; so, where those the first line come from? What if that first line is not part of the patch?

In that case there might be a false-positive. But I tried to handle such cases where the first line is in the patch itself.

Okay, so there are total of 4 cases with file mode change:
1. mode change + create + modify
2. mode change + create
3. mode change + modify
4. mode change

Quoted 8 lines

the patch will handle all cases of type 1 and some cases of type 3. Other cases won't be handled because they cannot be handled without breaking the core design decision.

I do have references to say that the cases 2 and 4 are

Which I thought is an improvement over current logic used for the check.

Sounds good, maybe you can explain that in the commit message and then send that to the mailing list and Joe Perches.

Lukas
_______________________________________________
Linux-kernel-mentees mailing list
[email protected]
https://lists.linuxfoundation.org/mailman/listinfo/linux-kernel-mentees