Skip to contentSkip to conversation

x86, ptrace: void dopiness

PATCH In review Markus Metzger 3 participants 7 messages Sep 19, 2008 – Sep 23, 2008

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

PATCH 1/1 x86, ptrace: void dopiness

Remove void cast to explicitly ignore the return value.

Signed-off-by [email protected]

---

arch/x86/kernel/ptrace.c +2 -2
Index: gits.x86/arch/x86/kernel/ptrace.c
===================================================================
@@ -738,7 +738,7 @@
 		unsigned int sig = 0;
 
 		/* we ignore the error in case we were not tracing child */
-		(void)ds_release_bts(child);
+		ds_release_bts(child);
 
 		if (cfg.flags & PTRACE_BTS_O_SIGNAL) {
 			if (!cfg.signal)
@@ -947,7 +947,7 @@
 	clear_tsk_thread_flag(child, TIF_SYSCALL_EMU);
 #endif
 #ifdef CONFIG_X86_PTRACE_BTS
-	(void)ds_release_bts(child);
+	ds_release_bts(child);
 
 	child->thread.debugctlmsr &= ~bts_cfg.debugctl_mask;
 	if (!child->thread.debugctlmsr)

---------------------------------------------------------------------
Intel GmbH
Dornacher Strasse 1
85622 Feldkirchen/Muenchen Germany
Sitz der Gesellschaft: Feldkirchen bei Muenchen
Geschaeftsfuehrer: Douglas Lusk, Peter Gleissner, Hannes Schwaderer
Registergericht: Muenchen HRB 47456 Ust.-IdNr.
VAT Registration No.: DE129385895
Citibank Frankfurt (BLZ 502 109 00) 600119052

This e-mail and any attachments may contain confidential material for the sole use of the intended recipient(s). Any review or distribution by others is strictly prohibited. If you are not the intended recipient, please contact the sender and delete all copies.

#1 lore raw
Quoted 29 lines · Markus Metzger

Remove void cast to explicitly ignore the return value.

Signed-off-by [email protected]

---
Index: gits.x86/arch/x86/kernel/ptrace.c
===================================================================

arch/x86/kernel/ptrace.c +2 -3
@@ -738,7 +738,7 @@
 		unsigned int sig = 0;
 
 		/* we ignore the error in case we were not tracing child */
-		(void)ds_release_bts(child);
+		ds_release_bts(child);
 
 		if (cfg.flags & PTRACE_BTS_O_SIGNAL) {
 			if (!cfg.signal)
@@ -947,7 +947,7 @@
 	clear_tsk_thread_flag(child, TIF_SYSCALL_EMU);
 #endif
 #ifdef CONFIG_X86_PTRACE_BTS
-	(void)ds_release_bts(child);
+	ds_release_bts(child);
 
 	child->thread.debugctlmsr &= ~bts_cfg.debugctl_mask;
 	if (!child->thread.debugctlmsr)
---------------------------------------------------------------------
Quoted 8 lines

Intel GmbH
Dornacher Strasse 1
85622 Feldkirchen/Muenchen Germany
Sitz der Gesellschaft: Feldkirchen bei Muenchen
Geschaeftsfuehrer: Douglas Lusk, Peter Gleissner, Hannes Schwaderer
Registergericht: Muenchen HRB 47456 Ust.-IdNr.
VAT Registration No.: DE129385895
Citibank Frankfurt (BLZ 502 109 00) 600119052

I don't agree. Casting the return value to void, since it takes extra steps on the part of the writer, means that the writer is deliberately ignoring the return value because it has no objective use in the code.

Just letting unused return-values be ignored means that the writer either wanted to ignore it or simply forgot about it. The cast is supposed to remove doubt. Certain user-mode functions that are obviously used for effect like printf(), are exempt.

That said, putting a space between the cast and the function name helps in the use of some text-scanning tools and therefore improves maintainability.

Cheers,
Dick Johnson
Penguin : Linux version 2.6.25.17 on an i686 machine (4786.70 BogoMips). My book : http://www.AbominableFirebug.com/
_

****************************************************************
The information transmitted in this message is confidential and may be privileged. Any review, retransmission, dissemination, or other use of this information by persons or entities other than the intended recipient is prohibited. If you are not the intended recipient, please notify Analogic Corporation immediately - by replying to this message or by sending an email to [email protected] - and destroy all copies of this information, including any attachments, without reading or disclosing them.

Thank you.

#2 lore raw
In reply to Markus Metzger

arch/x86/kernel/ptrace.c line 741

Quoted 7 lines · Markus Metzger

+++ gits.x86/arch/x86/kernel/ptrace.c 2008-09-19 13:53:02.%N +0200

arch/x86/kernel/ptrace.c +1 -1
@@ -738,7 +738,7 @@
 		unsigned int sig = 0;
 
 		/* we ignore the error in case we were not tracing child */
-		(void)ds_release_bts(child);
+		ds_release_bts(child);

hm, here the cast is OK because we actually ignore the return value.

arch/x86/kernel/ptrace.c line 950

Quoted 6 lines
arch/x86/kernel/ptrace.c +1 -1
@@ -947,7 +947,7 @@
 	clear_tsk_thread_flag(child, TIF_SYSCALL_EMU);
 #endif
 #ifdef CONFIG_X86_PTRACE_BTS
-	(void)ds_release_bts(child);
+	ds_release_bts(child);

is it right/intentional here?

Ingo

#3 lore raw
Quoted 23 lines

-----Original Message-----
From: Ingo Molnar [mailto:[email protected]]
Sent: Montag, 22. September 2008 13:51
To: Metzger, Markus T
Cc: [email protected]; [email protected]; [email protected]; Roland McGrath
Subject: Re: [patch] x86, ptrace: void dopiness

Quoted 1 line · Markus Metzger

+++ gits.x86/arch/x86/kernel/ptrace.c 2008-09-19

13:53:02.%N +0200

arch/x86/kernel/ptrace.c line 740

Quoted 4 lines
arch/x86/kernel/ptrace.c +0 -0
@@ -738,7 +738,7 @@
 		unsigned int sig = 0;
 
 		/* we ignore the error in case we were not 

tracing child */

Review line 741

Quoted 2 lines
(quoted diff) +1 -1
-		(void)ds_release_bts(child);
+		ds_release_bts(child);

hm, here the cast is OK because we actually ignore the return value.

Quoted 8 lines

arch/x86/kernel/ptrace.c line 950

Quoted 6 lines
arch/x86/kernel/ptrace.c +1 -1
@@ -947,7 +947,7 @@
 	clear_tsk_thread_flag(child, TIF_SYSCALL_EMU);
 #endif
 #ifdef CONFIG_X86_PTRACE_BTS
-	(void)ds_release_bts(child);
+	ds_release_bts(child);

is it right/intentional here?

The void-cast is intentional in both cases.

I thought it a question of style, i.e. that we don't want void casts just like we want NULL instead of 0.

regards,
markus.

---------------------------------------------------------------------
Intel GmbH
Dornacher Strasse 1
85622 Feldkirchen/Muenchen Germany
Sitz der Gesellschaft: Feldkirchen bei Muenchen
Geschaeftsfuehrer: Douglas Lusk, Peter Gleissner, Hannes Schwaderer
Registergericht: Muenchen HRB 47456 Ust.-IdNr.
VAT Registration No.: DE129385895
Citibank Frankfurt (BLZ 502 109 00) 600119052

This e-mail and any attachments may contain confidential material for the sole use of the intended recipient(s). Any review or distribution by others is strictly prohibited. If you are not the intended recipient, please contact the sender and delete all copies.

#4 lore raw
Quoted 32 lines · Metzger, Markus T
Quoted 23 lines

-----Original Message-----
From: Ingo Molnar [mailto:[email protected]]
Sent: Montag, 22. September 2008 13:51
To: Metzger, Markus T
Cc: [email protected]; [email protected]; [email protected]; Roland McGrath
Subject: Re: [patch] x86, ptrace: void dopiness

Quoted 1 line · Markus Metzger

+++ gits.x86/arch/x86/kernel/ptrace.c 2008-09-19

13:53:02.%N +0200

arch/x86/kernel/ptrace.c line 740

Quoted 4 lines
arch/x86/kernel/ptrace.c +0 -0
@@ -738,7 +738,7 @@
 		unsigned int sig = 0;
 
 		/* we ignore the error in case we were not 

tracing child */

Review line 741

Quoted 2 lines
(quoted diff) +1 -1
-		(void)ds_release_bts(child);
+		ds_release_bts(child);

hm, here the cast is OK because we actually ignore the return value.

Quoted 8 lines

arch/x86/kernel/ptrace.c line 950

Quoted 6 lines
arch/x86/kernel/ptrace.c +1 -1
@@ -947,7 +947,7 @@
 	clear_tsk_thread_flag(child, TIF_SYSCALL_EMU);
 #endif
 #ifdef CONFIG_X86_PTRACE_BTS
-	(void)ds_release_bts(child);
+	ds_release_bts(child);

is it right/intentional here?

Quoted 4 lines

The void-cast is intentional in both cases.

I thought it a question of style, i.e. that we don't want void casts just like we want NULL instead of 0.

ok.

But you could mark ds_release_bts() as a __must_check function, in that case the (void) has functional aspects as well: the kernel build will complain if a return value is ignored unintentionally.

So i think the code might be fine as-is after all :-/

Ingo

#5 lore raw
Quoted 10 lines

-----Original Message-----
From: Ingo Molnar [mailto:[email protected]]
Sent: Montag, 22. September 2008 14:08
To: Metzger, Markus T

Quoted 4 lines

The void-cast is intentional in both cases.

I thought it a question of style, i.e. that we don't want void casts just like we want NULL instead of 0.

Quoted 6 lines

ok.

But you could mark ds_release_bts() as a __must_check function, in that
case the (void) has functional aspects as well: the kernel build will complain if a return value is ignored unintentionally.

Wouldn't every non-void function be __must_check?

regards,
markus.
---------------------------------------------------------------------
Intel GmbH
Dornacher Strasse 1
85622 Feldkirchen/Muenchen Germany
Sitz der Gesellschaft: Feldkirchen bei Muenchen
Geschaeftsfuehrer: Douglas Lusk, Peter Gleissner, Hannes Schwaderer
Registergericht: Muenchen HRB 47456 Ust.-IdNr.
VAT Registration No.: DE129385895
Citibank Frankfurt (BLZ 502 109 00) 600119052

This e-mail and any attachments may contain confidential material for the sole use of the intended recipient(s). Any review or distribution by others is strictly prohibited. If you are not the intended recipient, please contact the sender and delete all copies.

#6 lore raw
Quoted 10 lines · Metzger, Markus T
Quoted 10 lines

-----Original Message-----
From: Ingo Molnar [mailto:[email protected]]
Sent: Montag, 22. September 2008 14:08
To: Metzger, Markus T

Quoted 4 lines

The void-cast is intentional in both cases.

I thought it a question of style, i.e. that we don't want void casts just like we want NULL instead of 0.

Quoted 8 lines
Quoted 6 lines

ok.

But you could mark ds_release_bts() as a __must_check function, in that
case the (void) has functional aspects as well: the kernel build will complain if a return value is ignored unintentionally.

Wouldn't every non-void function be __must_check?

no, only those which we mark so explicitly.

Ingo