Skip to contentSkip to conversation

x86: ftrace.c declare prepare_ftrace_return before they get used

PATCH In review Jaswinder Singh Rajput 3 participants 6 messages Apr 9, 2009 – Apr 10, 2009

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

PATCH 1/1 x86: ftrace.c declare prepare_ftrace_return before they get used

Impact: fix sparse warning:
arch/x86/kernel/ftrace.c:411:6: warning: symbol 'prepare_ftrace_return' was not declared. Should it be static?

---
 arch/x86/include/asm/ftrace.h |    6 ++++++
 1 files changed, 6 insertions(+), 0 deletions(-)
arch/x86/include/asm/ftrace.h +6 -0
index bd2c651..49616d4 100644
@@ -62,4 +62,10 @@ struct dyn_arch_ftrace {
 #endif /* __ASSEMBLY__ */
 #endif /* CONFIG_FUNCTION_TRACER */
 
+#ifdef CONFIG_FUNCTION_GRAPH_TRACER
+#ifndef __ASSEMBLY__
+extern void prepare_ftrace_return(unsigned long *, unsigned long);
+#endif
+#endif
+
 #endif /* _ASM_X86_FTRACE_H */
Signature · 1 line
1.6.0.6
#1 lore raw
Quoted 13 lines · Jaswinder Singh Rajput

Impact: fix sparse warning:
arch/x86/kernel/ftrace.c:411:6: warning: symbol 'prepare_ftrace_return' was not declared. Should it be static?

---
 arch/x86/include/asm/ftrace.h |    6 ++++++
 1 files changed, 6 insertions(+), 0 deletions(-)
arch/x86/include/asm/ftrace.h +0 -0
index bd2c651..49616d4 100644

arch/x86/include/asm/ftrace.h struct dyn_arch_ftrace { · line 67

Quoted 6 lines
arch/x86/include/asm/ftrace.h +3 -0
@@ -62,4 +62,10 @@ struct dyn_arch_ftrace {
 #endif /* __ASSEMBLY__ */
 #endif /* CONFIG_FUNCTION_TRACER */
 
+#ifdef CONFIG_FUNCTION_GRAPH_TRACER
+#ifndef __ASSEMBLY__
+extern void prepare_ftrace_return(unsigned long *, unsigned long);

But it is only used from asm code so there is no need to make its prototype public.

I don't think this is necessary but if you really think it is, then I would prefer you use the already existing #ifndef __ASSEMBLY__ block.

Thanks,
Frederic.

Quoted 14 lines
arch/x86/include/asm/ftrace.h +3 -0
+#endif
+#endif
+
 #endif /* _ASM_X86_FTRACE_H */
Signature · 9 lines
1.6.0.6



--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/
#2 lore raw
Quoted 25 lines · Frederic Weisbecker
Quoted 13 lines · Jaswinder Singh Rajput

Impact: fix sparse warning:
arch/x86/kernel/ftrace.c:411:6: warning: symbol 'prepare_ftrace_return' was not declared. Should it be static?

---
 arch/x86/include/asm/ftrace.h |    6 ++++++
 1 files changed, 6 insertions(+), 0 deletions(-)
arch/x86/include/asm/ftrace.h +0 -0
index bd2c651..49616d4 100644

arch/x86/include/asm/ftrace.h struct dyn_arch_ftrace { · line 67

Quoted 6 lines
arch/x86/include/asm/ftrace.h +3 -0
@@ -62,4 +62,10 @@ struct dyn_arch_ftrace {
 #endif /* __ASSEMBLY__ */
 #endif /* CONFIG_FUNCTION_TRACER */
 
+#ifdef CONFIG_FUNCTION_GRAPH_TRACER
+#ifndef __ASSEMBLY__
+extern void prepare_ftrace_return(unsigned long *, unsigned long);

But it is only used from asm code so there is no need to make its prototype public.

Quoted 3 lines

I don't think this is necessary but if you really think it is, then I would prefer you use the already existing #ifndef __ASSEMBLY__ block.

Like this:

Subject: [PATCH] x86: ftrace.c declare prepare_ftrace_return before they get used

Impact: fix sparse warning:
arch/x86/kernel/ftrace.c:411:6: warning: symbol 'prepare_ftrace_return' was not declared. Should it be static?

---
 arch/x86/include/asm/ftrace.h |    5 +++++
 1 files changed, 5 insertions(+), 0 deletions(-)
arch/x86/include/asm/ftrace.h +5 -0
index bd2c651..ddc9236 100644
@@ -59,6 +59,11 @@ struct dyn_arch_ftrace {
 };
 
 #endif /*  CONFIG_DYNAMIC_FTRACE */
+
+#ifdef CONFIG_FUNCTION_GRAPH_TRACER
+extern void prepare_ftrace_return(unsigned long *, unsigned long);
+#endif
+
 #endif /* __ASSEMBLY__ */
 #endif /* CONFIG_FUNCTION_TRACER */
 
Signature · 1 line
1.6.0.6
maintainer
#3 lore raw
Quoted 26 lines · Jaswinder Singh Rajput
Quoted 25 lines · Frederic Weisbecker
Quoted 13 lines · Jaswinder Singh Rajput

Impact: fix sparse warning:
arch/x86/kernel/ftrace.c:411:6: warning: symbol 'prepare_ftrace_return' was not declared. Should it be static?

---
 arch/x86/include/asm/ftrace.h |    6 ++++++
 1 files changed, 6 insertions(+), 0 deletions(-)
arch/x86/include/asm/ftrace.h +0 -0
index bd2c651..49616d4 100644

arch/x86/include/asm/ftrace.h struct dyn_arch_ftrace { · line 67

Quoted 6 lines
arch/x86/include/asm/ftrace.h +3 -0
@@ -62,4 +62,10 @@ struct dyn_arch_ftrace {
 #endif /* __ASSEMBLY__ */
 #endif /* CONFIG_FUNCTION_TRACER */
 
+#ifdef CONFIG_FUNCTION_GRAPH_TRACER
+#ifndef __ASSEMBLY__
+extern void prepare_ftrace_return(unsigned long *, unsigned long);

But it is only used from asm code so there is no need to make its prototype public.

Quoted 3 lines
Quoted 3 lines

I don't think this is necessary but if you really think it is, then I would prefer you use the already existing #ifndef __ASSEMBLY__ block.

It is not necessary.

Quoted 18 lines

Like this:

Subject: [PATCH] x86: ftrace.c declare prepare_ftrace_return before they get used

Impact: fix sparse warning:
arch/x86/kernel/ftrace.c:411:6: warning: symbol 'prepare_ftrace_return' was not declared. Should it be static?

---
 arch/x86/include/asm/ftrace.h |    5 +++++
 1 files changed, 5 insertions(+), 0 deletions(-)
arch/x86/include/asm/ftrace.h +0 -0
index bd2c651..ddc9236 100644

arch/x86/include/asm/ftrace.h struct dyn_arch_ftrace { · line 66

Quoted 8 lines
arch/x86/include/asm/ftrace.h +5 -0
@@ -59,6 +59,11 @@ struct dyn_arch_ftrace {
 };
 
 #endif /*  CONFIG_DYNAMIC_FTRACE */
+
+#ifdef CONFIG_FUNCTION_GRAPH_TRACER
+extern void prepare_ftrace_return(unsigned long *, unsigned long);
+#endif
+

The only caller is from assembly, so this patch is unnecessary. Unless we have some rule that all functions must have a prototype, even when it is only called from assembly.

-- Steve

Quoted 9 lines
arch/x86/include/asm/ftrace.h +0 -0
 #endif /* __ASSEMBLY__ */
 #endif /* CONFIG_FUNCTION_TRACER */
 
Signature · 1 line
1.6.0.6
#4 lore raw
Quoted 21 lines · Steven Rostedt
Quoted 18 lines · Jaswinder Singh Rajput

Subject: [PATCH] x86: ftrace.c declare prepare_ftrace_return before they get used

Impact: fix sparse warning:
arch/x86/kernel/ftrace.c:411:6: warning: symbol 'prepare_ftrace_return' was not declared. Should it be static?

---
 arch/x86/include/asm/ftrace.h |    5 +++++
 1 files changed, 5 insertions(+), 0 deletions(-)
arch/x86/include/asm/ftrace.h +0 -0
index bd2c651..ddc9236 100644
@@ -59,6 +59,11 @@ struct dyn_arch_ftrace {
 };
 
Quoted 11 lines

arch/x86/include/asm/ftrace.h struct dyn_arch_ftrace { · line 66

Quoted 6 lines
arch/x86/include/asm/ftrace.h +5 -0
 #endif /*  CONFIG_DYNAMIC_FTRACE */
+
+#ifdef CONFIG_FUNCTION_GRAPH_TRACER
+extern void prepare_ftrace_return(unsigned long *, unsigned long);
+#endif
+

The only caller is from assembly, so this patch is unnecessary. Unless we have some rule that all functions must have a prototype, even when it is only called from assembly.

Then:

arch/x86/kernel/ftrace.c +1 -0
index 18dfa30..75c0682 100644
@@ -408,6 +408,7 @@ int ftrace_disable_ftrace_graph_caller(void)
  * Hook the return address and push it in the stack of return addrs
  * in current thread info.
  */
+asmlinkage
 void prepare_ftrace_return(unsigned long *parent, unsigned long self_addr)
 {
        unsigned long old;

Josh: Do you think sparse should not warn like for this case. What will be the best solution for this case.

--
JSR

#5 lore raw
Quoted 41 lines · Jaswinder Singh Rajput
Quoted 21 lines · Steven Rostedt
Quoted 18 lines · Jaswinder Singh Rajput

Subject: [PATCH] x86: ftrace.c declare prepare_ftrace_return before they get used

Impact: fix sparse warning:
arch/x86/kernel/ftrace.c:411:6: warning: symbol 'prepare_ftrace_return' was not declared. Should it be static?

---
 arch/x86/include/asm/ftrace.h |    5 +++++
 1 files changed, 5 insertions(+), 0 deletions(-)
arch/x86/include/asm/ftrace.h +0 -0
index bd2c651..ddc9236 100644
@@ -59,6 +59,11 @@ struct dyn_arch_ftrace {
 };
 
Quoted 11 lines

arch/x86/include/asm/ftrace.h struct dyn_arch_ftrace { · line 66

Quoted 6 lines
arch/x86/include/asm/ftrace.h +5 -0
 #endif /*  CONFIG_DYNAMIC_FTRACE */
+
+#ifdef CONFIG_FUNCTION_GRAPH_TRACER
+extern void prepare_ftrace_return(unsigned long *, unsigned long);
+#endif
+

The only caller is from assembly, so this patch is unnecessary. Unless we have some rule that all functions must have a prototype, even when it is only called from assembly.

Then:

arch/x86/kernel/ftrace.c +0 -0
index 18dfa30..75c0682 100644

arch/x86/kernel/ftrace.c int ftrace_disable_ftrace_graph_caller(void) · line 411

Quoted 4 lines
arch/x86/kernel/ftrace.c +1 -0
@@ -408,6 +408,7 @@ int ftrace_disable_ftrace_graph_caller(void)
  * Hook the return address and push it in the stack of return addrs
  * in current thread info.
  */
+asmlinkage

No. This is how prepare_ftrace_return was called at first. But note that prepare_ftrace_return may be called for every kernel function while tracing, hence every pico optimization is important. asmlinkage will push the arguments to the stack while we want to keep the fastcall here by passing them to the registers.

Sparse is a useful tool, but sometime there are exceptions when its warnings should be ignored IMO.

Frederic.

Quoted 12 lines
arch/x86/kernel/ftrace.c +0 -0
 void prepare_ftrace_return(unsigned long *parent, unsigned long self_addr)
 {
        unsigned long old;

Josh: Do you think sparse should not warn like for this case. What will be the best solution for this case.

--
JSR