Skip to contentSkip to conversation

scripts/coccinelle/misc/semicolon.cocci: Add unneeded semicolon test

PATCH In review Peter Senna Tschudin 3 participants 8 messages Dec 8, 2012 – Jan 24, 2013

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 scripts/coccinelle/misc/semicolon.cocci: Add unneeded semicolon test

This semantic patch looks for semicolons that can be removed without changing the semantics of the code. The confidence is moderate because there are some false positives on cases like:

b/drivers/mmc/host/cb710-mmc.c:589

                break;
        case MMC_POWER_UP:
        default:
-               /* ignore */;
        }

There are 37 patches accepted reported by this semantic patch and more than 300 fixes to be applied.

---
Resending unchanged. Last sent on November 13 2012. First sent on October 30 2012.

 scripts/coccinelle/misc/semicolon.cocci | 83 +++++++++++++++++++++++++++++++++
 1 file changed, 83 insertions(+)
 create mode 100644 scripts/coccinelle/misc/semicolon.cocci
scripts/coccinelle/misc/semicolon.cocci new file +83 -0
new file mode 100644
index 0000000..a47eba2
@@ -0,0 +1,83 @@
+///
+/// Removes unneeded semicolon.
+///
+// Confidence: Moderate
+// Copyright: (C) 2012 Peter Senna Tschudin, INRIA/LIP6.  GPLv2.
+// URL: http://coccinelle.lip6.fr/
+// Comments: Some false positives on empty default cases in switch statements.
+// Options: --no-includes --include-headers
+
+virtual patch
+virtual report
+virtual context
+virtual org
+
+ at r_default@
+position p;
+@@
+switch (...)
+{
+default: ...;@p
+}
+
+ at r_case@
+position p;
+@@
+(
+switch (...)
+{
+case ...:;@p
+}
+|
+switch (...)
+{
+case ...:...
+case ...:;@p
+}
+|
+switch (...)
+{
+case ...:...
+case ...:
+case ...:;@p
+}
+)
+
+ at r1@
+statement S;
+position p1;
+position p != {r_default.p, r_case.p};
+identifier label;
+@@
+(
+label:;
+|
+S at p1;@p
+)
+
+ at script:python@
+p << r1.p;
+p1 << r1.p1;
+@@
+if p[0].line != p1[0].line_end:
+	cocci.include_match(False)
+
+ at depends on patch@
+position r1.p;
+@@
+-;@p
+
+ at script:python depends on report@
+p << r1.p;
+@@
+coccilib.report.print_report(p[0],"Unneeded semicolon")
+
+ at depends on context@
+position r1.p;
+@@
+*;@p
+
+@script:python depends on org@
+p << r1.p;
+@@
+cocci.print_main("Unneeded semicolon",p)
Signature · 1 line
1.7.11.7
maintainer
#1 lore raw
Quoted 10 lines · Peter Senna Tschudin

This semantic patch looks for semicolons that can be removed without changing the semantics of the code. The confidence is moderate because there are some false positives on cases like:

b/drivers/mmc/host/cb710-mmc.c:589

                break;
        case MMC_POWER_UP:
        default:
-               /* ignore */;
        }

[]

scripts/coccinelle/misc/semicolon.cocci

Quoted 1 line
scripts/coccinelle/misc/semicolon.cocci +0 -0

[]

scripts/coccinelle/misc/semicolon.cocci line 57

Quoted 12 lines

+ at r1@

scripts/coccinelle/misc/semicolon.cocci +11 -0
+statement S;
+position p1;
+position p != {r_default.p, r_case.p};
+identifier label;
+@@
+(
+label:;
+|
+S at p1;@p
+)
+

I believe this also fails on this case:

	void some_func(...)
	{
		...
		if (foo)
			goto exit;
		...

	exit:
		;
	}

where gcc needs a semicolon after a label before a function exit.

#2 lore raw
Quoted 40 lines · Joe Perches
Quoted 10 lines · Peter Senna Tschudin

This semantic patch looks for semicolons that can be removed without changing the semantics of the code. The confidence is moderate because there are some false positives on cases like:

b/drivers/mmc/host/cb710-mmc.c:589

                break;
        case MMC_POWER_UP:
        default:
-               /* ignore */;
        }

[]

scripts/coccinelle/misc/semicolon.cocci

Quoted 1 line
scripts/coccinelle/misc/semicolon.cocci +0 -0

[]

scripts/coccinelle/misc/semicolon.cocci line 57

Quoted 12 lines

+ at r1@

scripts/coccinelle/misc/semicolon.cocci +11 -0
+statement S;
+position p1;
+position p != {r_default.p, r_case.p};
+identifier label;
+@@
+(
+label:;
+|
+S at p1;@p
+)
+

I believe this also fails on this case:

        void some_func(...)
        {
                ...
                if (foo)
                        goto exit;
                ...

        exit:
                ;
        }
Quoted 1 line

where gcc needs a semicolon after a label before a function exit.

No it does not fail. This issue is switch/case specific. See how I've tested:

$ cat /tmp/test.c
void some_func(void){
        int *foo;
        if (foo)
        goto exit;

exit:
;
}

void few_func (void){
        int a;
        switch (a){
                case 1:
                        a=1;
                        break;
                case 2:
                default:
                        ;
        }
}

void main (void)

{
        int a;
        some_func();
        few_func();
}
$ spatch -D patch scripts/coccinelle/misc/semicolon.cocci /tmp/test.c
init_defs_builtins: /usr/local/share/coccinelle/standard.h
HANDLING: /tmp/test.c
diff =
--- /tmp/test.c
+++ /tmp/cocci-output-2570-defed8-test.c
@@ -15,7 +15,6 @@ void few_func (void){
  break;
  case 2:
  default:
- ;
  }
 }
Quoted 1 line

--
Peter

maintainer
#3 lore raw
Quoted 15 lines · Peter Senna Tschudin
Quoted 14 lines · Joe Perches
Quoted 10 lines · Peter Senna Tschudin

This semantic patch looks for semicolons that can be removed without changing the semantics of the code. The confidence is moderate because there are some false positives on cases like:

b/drivers/mmc/host/cb710-mmc.c:589

                break;
        case MMC_POWER_UP:
        default:
-               /* ignore */;
        }

[]

scripts/coccinelle/misc/semicolon.cocci

Quoted 1 line
scripts/coccinelle/misc/semicolon.cocci +0 -0

[]

Quoted 14 lines
Quoted 6 lines
Quoted 6 lines

+ at r1@

scripts/coccinelle/misc/semicolon.cocci +5 -0
+statement S;
+position p1;
+position p != {r_default.p, r_case.p};
+identifier label;
+@@
Quoted 8 lines

scripts/coccinelle/misc/semicolon.cocci line 57

Quoted 6 lines
scripts/coccinelle/misc/semicolon.cocci +6 -0
+(
+label:;
+|
+S at p1;@p
+)
+

I believe this also fails on this case:

[]

Quoted 3 lines
Quoted 1 line

where gcc needs a semicolon after a label before a function exit.

No it does not fail. This issue is switch/case specific. See how I've tested:

Thanks Peter, I didn't notice the switch requirement.

In this case, I'd suggest replacement of the
nominally false positive ; with break;

cheers, Joe

#4 lore raw
Quoted 31 lines · Joe Perches
Quoted 15 lines · Peter Senna Tschudin
Quoted 14 lines · Joe Perches
Quoted 10 lines · Peter Senna Tschudin

This semantic patch looks for semicolons that can be removed without changing the semantics of the code. The confidence is moderate because there are some false positives on cases like:

b/drivers/mmc/host/cb710-mmc.c:589

                break;
        case MMC_POWER_UP:
        default:
-               /* ignore */;
        }

[]

scripts/coccinelle/misc/semicolon.cocci

Quoted 1 line
scripts/coccinelle/misc/semicolon.cocci +0 -0

[]

Quoted 14 lines
Quoted 6 lines
Quoted 6 lines

+ at r1@

scripts/coccinelle/misc/semicolon.cocci +5 -0
+statement S;
+position p1;
+position p != {r_default.p, r_case.p};
+identifier label;
+@@
Quoted 8 lines

scripts/coccinelle/misc/semicolon.cocci line 57

Quoted 6 lines
scripts/coccinelle/misc/semicolon.cocci +6 -0
+(
+label:;
+|
+S at p1;@p
+)
+

I believe this also fails on this case:

[]

Quoted 8 lines
Quoted 3 lines
Quoted 1 line

where gcc needs a semicolon after a label before a function exit.

No it does not fail. This issue is switch/case specific. See how I've tested:

Thanks Peter, I didn't notice the switch requirement.

In this case, I'd suggest replacement of the
nominally false positive ; with break;

I agree with you that the best is to replace ; with break; but I prefer to not change the unneeded semicolon test semantic patch to add breaks before ;. I can make other semantic patch for that. The point is that adding break is more than detecting unneeded semicolons, so I prefer to not mix the two.

But even with the moderate confidence this semantic patch seems to be useful as new unneeded semicolons are being added to the Kernel.

Quoted 3 lines

cheers, Joe

Thanks for your help!

Peter

--
Peter

maintainer
#5 lore raw

PATCH 1/1 checkpatch: Add --strict test for switch/default missing break

switch default case is sometimes written
as "default:;". This can cause new cases
added below the default to be defective.

Suggest adding a break; after empty default
cases to avoid fallthrough defects.

Fixed indentation in the other semicolon
test above it.

---
 scripts/checkpatch.pl |   18 ++++++++++++++++--
 1 files changed, 16 insertions(+), 2 deletions(-)
scripts/checkpatch.pl +16 -2
index f18750e..73302d8 100755
@@ -3450,8 +3450,22 @@ sub process {
 
 # check for multiple semicolons
 		if ($line =~ /;\s*;\s*$/) {
-		    WARN("ONE_SEMICOLON",
-			 "Statements terminations use 1 semicolon\n" . $herecurr);
+			WARN("ONE_SEMICOLON",
+			     "Statements terminations use 1 semicolon\n" . $herecurr);
+		}
+
+# check for switch/default statements without a break;
+		if ($^V && $^V ge 5.10.0 &&
+		    defined $stat &&
+		    $stat =~ /^\+[$;\s]*(?:case[$;\s]+\w+[$;\s]*:[$;\s]*|)*[$;\s]*\bdefault[$;\s]*:[$;\s]*;/g) {
+			my $ctx = '';
+			my $herectx = $here . "\n";
+			my $cnt = statement_rawlines($stat);
+			for (my $n = 0; $n < $cnt; $n++) {
+				$herectx .= raw_line($linenr, $n) . "\n";
+			}
+			WARN("DEFAULT_NO_BREAK",
+			     "switch default: should use break\n" . $herectx);
 		}
 
 # check for gcc specific __FUNCTION__
Signature · 1 line
1.7.8.112.g3fd21
#6 lore raw
In reply to Peter Senna Tschudin

ping

Quoted 114 lines · Peter Senna Tschudin

This semantic patch looks for semicolons that can be removed without changing the semantics of the code. The confidence is moderate because there are some false positives on cases like:

b/drivers/mmc/host/cb710-mmc.c:589

                break;
        case MMC_POWER_UP:
        default:
-               /* ignore */;
        }

There are 37 patches accepted reported by this semantic patch and more than 300 fixes to be applied.

---
Resending unchanged. Last sent on November 13 2012. First sent on October 30 2012.

 scripts/coccinelle/misc/semicolon.cocci | 83 +++++++++++++++++++++++++++++++++
 1 file changed, 83 insertions(+)
 create mode 100644 scripts/coccinelle/misc/semicolon.cocci
scripts/coccinelle/misc/semicolon.cocci new file +83 -1
new file mode 100644
index 0000000..a47eba2
@@ -0,0 +1,83 @@
+///
+/// Removes unneeded semicolon.
+///
+// Confidence: Moderate
+// Copyright: (C) 2012 Peter Senna Tschudin, INRIA/LIP6.  GPLv2.
+// URL: http://coccinelle.lip6.fr/
+// Comments: Some false positives on empty default cases in switch statements.
+// Options: --no-includes --include-headers
+
+virtual patch
+virtual report
+virtual context
+virtual org
+
+ at r_default@
+position p;
+@@
+switch (...)
+{
+default: ...;@p
+}
+
+ at r_case@
+position p;
+@@
+(
+switch (...)
+{
+case ...:;@p
+}
+|
+switch (...)
+{
+case ...:...
+case ...:;@p
+}
+|
+switch (...)
+{
+case ...:...
+case ...:
+case ...:;@p
+}
+)
+
+ at r1@
+statement S;
+position p1;
+position p != {r_default.p, r_case.p};
+identifier label;
+@@
+(
+label:;
+|
+S at p1;@p
+)
+
+ at script:python@
+p << r1.p;
+p1 << r1.p1;
+@@
+if p[0].line != p1[0].line_end:
+       cocci.include_match(False)
+
+ at depends on patch@
+position r1.p;
+@@
+-;@p
+
+ at script:python depends on report@
+p << r1.p;
+@@
+coccilib.report.print_report(p[0],"Unneeded semicolon")
+
+ at depends on context@
+position r1.p;
+@@
+*;@p
+
+ at script:python depends on org@
+p << r1.p;
+@@
+cocci.print_main("Unneeded semicolon",p)
--

1.7.11.7

Signature · 1 line
Peter
#7 lore raw
Quoted 1 line · Peter Senna Tschudin

ping

pong

I added the patch to kbuild.git#misc now, sorry for the delay.

Michal