[PATCH 0/3] [GIT PULL] tracing: small fixes for v2.6.32

[PATCH 0/3] [GIT PULL] tracing: small fixes for v2.6.32

[Steven Rostedt]

Ingo,

Please pull the latest tip/tracing/core5 tree, which can be found at:

  git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/linux-2.6-trace.git
tip/tracing/core5


Christian Borntraeger (1):
      oprofile: fix oprofile regression: select RING_BUFFER_ALLOW_SWAP

Steven Rostedt (2):
      vsprintf: add %ps that is the same as %pS but is like %pf
      tracing: switch function prints from %pf to %ps

----
 arch/Kconfig                         |    1 +
 kernel/trace/ftrace.c                |    6 +++---
 kernel/trace/trace_functions.c       |    2 +-
 kernel/trace/trace_functions_graph.c |    6 +++---
 lib/vsprintf.c                       |    6 ++++--
 5 files changed, 12 insertions(+), 9 deletions(-)

[PATCH 1/3] [PATCH 1/3] vsprintf: add %ps that is the same as %pS but is like %pf

[Steven Rostedt]

[-- Attachment #1: 0001-vsprintf-add-ps-that-is-the-same-as-pS-but-is-like-p.patch --]
[-- Type: text/plain, Size: 2214 bytes --]

From: Steven Rostedt <srostedt@redhat.com>

On PowerPC64 function pointers do not point directly at the functions,
but instead point to pointers to the functions. The output of %pF expects
to point to a pointer to the function, whereas %pS will show the function
itself.

mcount returns the direct pointer to the function and not the pointer to
the pointer. Thus %pS must be used to show this. The function tracer
requires printing of the functions without offsets and uses the %pf
instead.

 %pF produces run_local_timers+0x4/0x1f
 %pf produces just run_local_timers

For PowerPC64, we need to use the direct pointer, and we only have
%pS which will produce .run_local_timers+0x4/0x1f

This patch creates a %ps that matches the %pf as %pS matches %pF.

Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Zhao Lei <zhaolei@cn.fujitsu.com>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
---
 lib/vsprintf.c |    6 ++++--
 1 files changed, 4 insertions(+), 2 deletions(-)

diff --git a/lib/vsprintf.c b/lib/vsprintf.c
index 756ccaf..c265e75 100644
--- a/lib/vsprintf.c
+++ b/lib/vsprintf.c
@@ -580,7 +580,7 @@ static char *symbol_string(char *buf, char *end, void *ptr,
 	unsigned long value = (unsigned long) ptr;
 #ifdef CONFIG_KALLSYMS
 	char sym[KSYM_SYMBOL_LEN];
-	if (ext != 'f')
+	if (ext != 'f' && ext != 's')
 		sprint_symbol(sym, value);
 	else
 		kallsyms_lookup(value, NULL, NULL, NULL, sym);
@@ -721,6 +721,7 @@ static char *pointer(const char *fmt, char *buf, char *end, void *ptr,
 	case 'F':
 	case 'f':
 		ptr = dereference_function_descriptor(ptr);
+	case 's':
 		/* Fallthrough */
 	case 'S':
 		return symbol_string(buf, end, ptr, spec, *fmt);
@@ -958,7 +959,8 @@ qualifier:
  * @args: Arguments for the format string
  *
  * This function follows C99 vsnprintf, but has some extensions:
- * %pS output the name of a text symbol
+ * %pS output the name of a text symbol with offset
+ * %ps output the name of a text symbol without offset
  * %pF output the name of a function pointer with its offset
  * %pf output the name of a function pointer without its offset
  * %pR output the address range in a struct resource
-- 
1.6.3.3

Re: [PATCH 1/3] [PATCH 1/3] vsprintf: add %ps that is the same as %pS but is like %pf

[Benjamin Herrenschmidt]

On Thu, 2009-09-17 at 00:27 -0400, Steven Rostedt wrote:
> plain text document attachment
> (0001-vsprintf-add-ps-that-is-the-same-as-pS-but-is-like-p.patch)
> From: Steven Rostedt <srostedt@redhat.com>
> 
> On PowerPC64 function pointers do not point directly at the functions,
> but instead point to pointers to the functions. The output of %pF expects
> to point to a pointer to the function, whereas %pS will show the function
> itself.
> 
> mcount returns the direct pointer to the function and not the pointer to
> the pointer. Thus %pS must be used to show this. The function tracer
> requires printing of the functions without offsets and uses the %pf
> instead.
> 
>  %pF produces run_local_timers+0x4/0x1f
>  %pf produces just run_local_timers
> 
> For PowerPC64, we need to use the direct pointer, and we only have
> %pS which will produce .run_local_timers+0x4/0x1f
> 
> This patch creates a %ps that matches the %pf as %pS matches %pF.
> 
> Cc: Linus Torvalds <torvalds@linux-foundation.org>
> Cc: Zhao Lei <zhaolei@cn.fujitsu.com>

Acked-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>

> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
> ---
>  lib/vsprintf.c |    6 ++++--
>  1 files changed, 4 insertions(+), 2 deletions(-)
> 
> diff --git a/lib/vsprintf.c b/lib/vsprintf.c
> index 756ccaf..c265e75 100644
> --- a/lib/vsprintf.c
> +++ b/lib/vsprintf.c
> @@ -580,7 +580,7 @@ static char *symbol_string(char *buf, char *end, void *ptr,
>  	unsigned long value = (unsigned long) ptr;
>  #ifdef CONFIG_KALLSYMS
>  	char sym[KSYM_SYMBOL_LEN];
> -	if (ext != 'f')
> +	if (ext != 'f' && ext != 's')
>  		sprint_symbol(sym, value);
>  	else
>  		kallsyms_lookup(value, NULL, NULL, NULL, sym);
> @@ -721,6 +721,7 @@ static char *pointer(const char *fmt, char *buf, char *end, void *ptr,
>  	case 'F':
>  	case 'f':
>  		ptr = dereference_function_descriptor(ptr);
> +	case 's':
>  		/* Fallthrough */
>  	case 'S':
>  		return symbol_string(buf, end, ptr, spec, *fmt);
> @@ -958,7 +959,8 @@ qualifier:
>   * @args: Arguments for the format string
>   *
>   * This function follows C99 vsnprintf, but has some extensions:
> - * %pS output the name of a text symbol
> + * %pS output the name of a text symbol with offset
> + * %ps output the name of a text symbol without offset
>   * %pF output the name of a function pointer with its offset
>   * %pf output the name of a function pointer without its offset
>   * %pR output the address range in a struct resource

Re: [PATCH 1/3] [PATCH 1/3] vsprintf: add %ps that is the same as %pS but is like %pf

[Zhaolei]

Steven Rostedt wrote:
> From: Steven Rostedt <srostedt@redhat.com>
> 
> On PowerPC64 function pointers do not point directly at the functions,
> but instead point to pointers to the functions. The output of %pF expects
> to point to a pointer to the function, whereas %pS will show the function
> itself.
> 
> mcount returns the direct pointer to the function and not the pointer to
> the pointer. Thus %pS must be used to show this. The function tracer
> requires printing of the functions without offsets and uses the %pf
> instead.
> 
>  %pF produces run_local_timers+0x4/0x1f
>  %pf produces just run_local_timers
> 
> For PowerPC64, we need to use the direct pointer, and we only have
> %pS which will produce .run_local_timers+0x4/0x1f
> 
> This patch creates a %ps that matches the %pf as %pS matches %pF.
> 
> Cc: Linus Torvalds <torvalds@linux-foundation.org>
> Cc: Zhao Lei <zhaolei@cn.fujitsu.com>
> Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
> ---
>  lib/vsprintf.c |    6 ++++--
>  1 files changed, 4 insertions(+), 2 deletions(-)
> 
> diff --git a/lib/vsprintf.c b/lib/vsprintf.c
> index 756ccaf..c265e75 100644
> --- a/lib/vsprintf.c
> +++ b/lib/vsprintf.c
> @@ -580,7 +580,7 @@ static char *symbol_string(char *buf, char *end, void *ptr,
>  	unsigned long value = (unsigned long) ptr;
>  #ifdef CONFIG_KALLSYMS
>  	char sym[KSYM_SYMBOL_LEN];
> -	if (ext != 'f')
> +	if (ext != 'f' && ext != 's')
>  		sprint_symbol(sym, value);
>  	else
>  		kallsyms_lookup(value, NULL, NULL, NULL, sym);
> @@ -721,6 +721,7 @@ static char *pointer(const char *fmt, char *buf, char *end, void *ptr,
>  	case 'F':
>  	case 'f':
>  		ptr = dereference_function_descriptor(ptr);
> +	case 's':
>  		/* Fallthrough */
>  	case 'S':
>  		return symbol_string(buf, end, ptr, spec, *fmt);
> @@ -958,7 +959,8 @@ qualifier:
>   * @args: Arguments for the format string
>   *
>   * This function follows C99 vsnprintf, but has some extensions:
> - * %pS output the name of a text symbol
> + * %pS output the name of a text symbol with offset
> + * %ps output the name of a text symbol without offset

Comments of bstr_printf() also need to be updated.
or remove duplation and just say "refer to vsnprintf()" in bstr_printf()'s
comment.

Thanks
Zhaolei

[PATCH 2/3] [PATCH 2/3] tracing: switch function prints from %pf to %ps

[Steven Rostedt]

[-- Attachment #1: 0002-tracing-switch-function-prints-from-pf-to-ps.patch --]
[-- Type: text/plain, Size: 3081 bytes --]

From: Steven Rostedt <srostedt@redhat.com>

For direct function pointers (like what mcount provides) PowerPC64
requires the use of %ps, otherwise nothing is printed.

This patch converts all prints of functions retrieved through mcount
to use the %ps format from the %pf.

Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
---
 kernel/trace/ftrace.c                |    6 +++---
 kernel/trace/trace_functions.c       |    2 +-
 kernel/trace/trace_functions_graph.c |    6 +++---
 3 files changed, 7 insertions(+), 7 deletions(-)

diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c
index f7ab7fc..cc615f8 100644
--- a/kernel/trace/ftrace.c
+++ b/kernel/trace/ftrace.c
@@ -1405,7 +1405,7 @@ static int t_hash_show(struct seq_file *m, void *v)
 	if (rec->ops->print)
 		return rec->ops->print(m, rec->ip, rec->ops, rec->data);
 
-	seq_printf(m, "%pf:%pf", (void *)rec->ip, (void *)rec->ops->func);
+	seq_printf(m, "%ps:%ps", (void *)rec->ip, (void *)rec->ops->func);
 
 	if (rec->data)
 		seq_printf(m, ":%p", rec->data);
@@ -1515,7 +1515,7 @@ static int t_show(struct seq_file *m, void *v)
 	if (!rec)
 		return 0;
 
-	seq_printf(m, "%pf\n", (void *)rec->ip);
+	seq_printf(m, "%ps\n", (void *)rec->ip);
 
 	return 0;
 }
@@ -2456,7 +2456,7 @@ static int g_show(struct seq_file *m, void *v)
 		return 0;
 	}
 
-	seq_printf(m, "%pf\n", (void *)*ptr);
+	seq_printf(m, "%ps\n", (void *)*ptr);
 
 	return 0;
 }
diff --git a/kernel/trace/trace_functions.c b/kernel/trace/trace_functions.c
index 5b01b94..b3f3776 100644
--- a/kernel/trace/trace_functions.c
+++ b/kernel/trace/trace_functions.c
@@ -290,7 +290,7 @@ ftrace_trace_onoff_print(struct seq_file *m, unsigned long ip,
 {
 	long count = (long)data;
 
-	seq_printf(m, "%pf:", (void *)ip);
+	seq_printf(m, "%ps:", (void *)ip);
 
 	if (ops == &traceon_probe_ops)
 		seq_printf(m, "traceon");
diff --git a/kernel/trace/trace_functions_graph.c b/kernel/trace/trace_functions_graph.c
index 61f1667..45e6c01 100644
--- a/kernel/trace/trace_functions_graph.c
+++ b/kernel/trace/trace_functions_graph.c
@@ -124,7 +124,7 @@ ftrace_pop_return_trace(struct ftrace_graph_ret *trace, unsigned long *ret,
 	if (unlikely(current->ret_stack[index].fp != frame_pointer)) {
 		ftrace_graph_stop();
 		WARN(1, "Bad frame pointer: expected %lx, received %lx\n"
-		     "  from func %pF return to %lx\n",
+		     "  from func %ps return to %lx\n",
 		     current->ret_stack[index].fp,
 		     frame_pointer,
 		     (void *)current->ret_stack[index].func,
@@ -669,7 +669,7 @@ print_graph_entry_leaf(struct trace_iterator *iter,
 			return TRACE_TYPE_PARTIAL_LINE;
 	}
 
-	ret = trace_seq_printf(s, "%pf();\n", (void *)call->func);
+	ret = trace_seq_printf(s, "%ps();\n", (void *)call->func);
 	if (!ret)
 		return TRACE_TYPE_PARTIAL_LINE;
 
@@ -712,7 +712,7 @@ print_graph_entry_nested(struct trace_iterator *iter,
 			return TRACE_TYPE_PARTIAL_LINE;
 	}
 
-	ret = trace_seq_printf(s, "%pf() {\n", (void *)call->func);
+	ret = trace_seq_printf(s, "%ps() {\n", (void *)call->func);
 	if (!ret)
 		return TRACE_TYPE_PARTIAL_LINE;
 
-- 
1.6.3.3

[PATCH 3/3] [PATCH 3/3] oprofile: fix oprofile regression: select RING_BUFFER_ALLOW_SWAP

[Steven Rostedt]

[-- Attachment #1: 0003-oprofile-fix-oprofile-regression-select-RING_BUFFER_.patch --]
[-- Type: text/plain, Size: 1048 bytes --]

From: Christian Borntraeger <borntraeger@de.ibm.com>

commit 85bac32c4a52c592b857f2c360cc5ec93a097d70
    ring-buffer: only enable ring_buffer_swap_cpu when needed
broke oprofile (at least on s390, but likely on all platforms).

this patch lets oprofile select RING_BUFER_ALLOW_SWAP to make
ring_buffer_swap_cpu usable for oprofile.

Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
LKML-Reference: <200909162156.49239.borntraeger@de.ibm.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Robert Richter <robert.richter@amd.com>
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
---
 arch/Kconfig |    1 +
 1 files changed, 1 insertions(+), 0 deletions(-)

diff --git a/arch/Kconfig b/arch/Kconfig
index 99193b1..4dd737a 100644
--- a/arch/Kconfig
+++ b/arch/Kconfig
@@ -9,6 +9,7 @@ config OPROFILE
 	depends on TRACING_SUPPORT
 	select TRACING
 	select RING_BUFFER
+	select RING_BUFFER_ALLOW_SWAP
 	help
 	  OProfile is a profiling system capable of profiling the
 	  whole system, include the kernel, kernel modules, libraries,
-- 
1.6.3.3