Skip to contentSkip to conversation

Don't return void types from void functions.

DISCUSSION Jeff Garzik 4 participants 4 messages May 26, 2004 – May 28, 2004 1 message not in the archive

lore
Share & tools
lore

A message that is not in the archive (sent off-list or not yet received): [email protected]

#1 lore raw

drivers/net/tokenring/olympic.c line 1810

Quoted 10 lines · Linux Kernel Mailing List

diff -Nru a/drivers/net/tokenring/olympic.c b/drivers/net/tokenring/olympic.c

drivers/net/tokenring/olympic.c +1 -1
@@ -1806,7 +1806,7 @@
 
 static void __exit olympic_pci_cleanup(void)
 {
-	return pci_unregister_driver(&olympic_driver) ; 
+	pci_unregister_driver(&olympic_driver) ; 
 }	

Can we make gcc error out when it finds this?

Jeff

#2 lore raw
Quoted 7 lines · Jeff Garzik
Quoted 6 lines · Linux Kernel Mailing List

diff -Nru a/drivers/net/tokenring/olympic.c
b/drivers/net/tokenring/olympic.c ---
a/drivers/net/tokenring/olympic.c 2004-05-25 23:06:49 -07:00 +++
b/drivers/net/tokenring/olympic.c 2004-05-25 23:06:49 -07:00 @@ -1806,7
+1806,7 @@

Quoted 7 lines

Review

Quoted 5 lines

static void __exit olympic_pci_cleanup(void)
{

(quoted diff) +1 -1
-	return pci_unregister_driver(&olympic_driver) ;
+	pci_unregister_driver(&olympic_driver) ;
 }

Can we make gcc error out when it finds this?

AFAIK new C++ standard allows this syntax.

typedef int opaque;

opaque f();
opaque g() { return f(); }

Now imagine we need to change
-typedef int opaque;
+typedef void opaque;

Signature · 1 line
vda
#3 lore raw
Quoted 24 lines · Denis Vlasenko
Quoted 7 lines · Jeff Garzik
Quoted 6 lines · Linux Kernel Mailing List

diff -Nru a/drivers/net/tokenring/olympic.c
b/drivers/net/tokenring/olympic.c ---
a/drivers/net/tokenring/olympic.c 2004-05-25 23:06:49 -07:00 +++
b/drivers/net/tokenring/olympic.c 2004-05-25 23:06:49 -07:00 @@ -1806,7
+1806,7 @@

Quoted 7 lines

Review

Quoted 5 lines

static void __exit olympic_pci_cleanup(void)
{

(quoted diff) +1 -1
-	return pci_unregister_driver(&olympic_driver) ;
+	pci_unregister_driver(&olympic_driver) ;
 }

Can we make gcc error out when it finds this?

AFAIK new C++ standard allows this syntax.

typedef int opaque;

opaque f();
opaque g() { return f(); }

Now imagine we need to change

Review

Quoted 2 lines
(quoted diff) +1 -1
-typedef int opaque;
+typedef void opaque;

strict ISO C++ - yes,
strict ISO C - no.

# g++ -Wall -s -c void_ret.cpp -pedantic-errors

# gcc -Wall -s -c void_ret.c -pedantic-errors
void_ret.c: In function `g':
void_ret.c:4: error: `return' with a value, in function returning void
# gcc --version
gcc (GCC) 3.4.0 SSP (PLD Linux)
(...)
Signature · 2 lines
If you think of MS-DOS as mono, and Windows as stereo,
  then Linux is Dolby Digital and all the music is free...
#4 lore raw
In reply to Jeff Garzik
Followup to:  <[email protected]>
By author:    Jeff Garzik <[email protected]>
In newsgroup: linux.dev.kernel
Quoted 7 lines
Quoted 5 lines · Linux Kernel Mailing List

diff -Nru a/drivers/net/tokenring/olympic.c b/drivers/net/tokenring/olympic.c

drivers/net/tokenring/olympic.c +0 -0
@@ -1806,7 +1806,7 @@
 
Quoted 9 lines

drivers/net/tokenring/olympic.c line 1810

Quoted 5 lines
drivers/net/tokenring/olympic.c +1 -1
 static void __exit olympic_pci_cleanup(void)
 {
-	return pci_unregister_driver(&olympic_driver) ; 
+	pci_unregister_driver(&olympic_driver) ; 
 }	

Can we make gcc error out when it finds this?

What's wrong with it? Seriously... "return foo();" is the closest thing to a tailcall syntax as C/C++ has, and I've always considered it a Total Fscking Pain in The Ass[TM] that it somehow thinks functions returning void should have a different syntax than all other functions.

-hpa