* [PATCH -tip 2/2] Tracing/fastboot: Get the initcall name before it disappears
@ 2008-10-02 11:26 Frederic Weisbecker
2008-10-02 12:02 ` Ingo Molnar
0 siblings, 1 reply; 4+ messages in thread
From: Frederic Weisbecker @ 2008-10-02 11:26 UTC (permalink / raw)
To: Ingo Molnar; +Cc: Steven Rostedt, linux-kernel
After some initcall traces, some initcall names may be inconsistent.
That's because these functions will disappear from the .init section
and also their name from the symbols table.
So we have to copy the name of the function in a buffer large enough
during the trace appending. It is not costly for the ring_buffer because
the number of initcall entries is commonly not really large.
Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com>
---
include/linux/ftrace.h | 7 ++++---
init/main.c | 3 +--
kernel/trace/trace_boot.c | 14 ++++++++++----
3 files changed, 15 insertions(+), 9 deletions(-)
diff --git a/include/linux/ftrace.h b/include/linux/ftrace.h
index 4455490..e672e51 100644
--- a/include/linux/ftrace.h
+++ b/include/linux/ftrace.h
@@ -7,6 +7,7 @@
#include <linux/fs.h>
#include <linux/init.h>
#include <linux/types.h>
+#include <linux/kallsyms.h>
extern int ftrace_enabled;
extern int
@@ -213,7 +214,7 @@ ftrace_init_module(unsigned long *start, unsigned long *end) { }
struct boot_trace {
pid_t caller;
- initcall_t func;
+ char func[KSYM_NAME_LEN];
int result;
unsigned long long duration;
ktime_t calltime;
@@ -221,10 +222,10 @@ struct boot_trace {
};
#ifdef CONFIG_BOOT_TRACER
-extern void trace_boot(struct boot_trace *it);
+extern void trace_boot(struct boot_trace *it, initcall_t fn);
extern void start_boot_trace(void);
#else
-static inline void trace_boot(struct boot_trace *it) { }
+static inline void trace_boot(struct boot_trace *it, initcall_t fn) { }
static inline void start_boot_trace(void) { }
#endif
diff --git a/init/main.c b/init/main.c
index ac3f760..47d621a 100644
--- a/init/main.c
+++ b/init/main.c
@@ -727,7 +727,6 @@ int do_one_initcall(initcall_t fn)
if (initcall_debug) {
it.caller = task_pid_nr(current);
- it.func = fn;
printk("calling %pF @ %i\n", fn, it.caller);
it.calltime = ktime_get();
}
@@ -740,7 +739,7 @@ int do_one_initcall(initcall_t fn)
it.duration = (unsigned long long) delta.tv64 >> 20;
printk("initcall %pF returned %d after %Ld msecs\n", fn,
it.result, it.duration);
- trace_boot(&it);
+ trace_boot(&it, fn);
}
msgbuf[0] = 0;
diff --git a/kernel/trace/trace_boot.c b/kernel/trace/trace_boot.c
index a8cb1c5..114fade 100644
--- a/kernel/trace/trace_boot.c
+++ b/kernel/trace/trace_boot.c
@@ -8,6 +8,7 @@
#include <linux/init.h>
#include <linux/debugfs.h>
#include <linux/ftrace.h>
+#include <linux/kallsyms.h>
#include "trace.h"
@@ -56,17 +57,19 @@ static enum print_line_t initcall_print_line(struct trace_iterator *iter)
struct timespec rettime = ktime_to_timespec(it->rettime);
if (entry->type == TRACE_BOOT) {
- ret = trace_seq_printf(s, "[%5ld.%06ld] calling %pF @ %i\n",
+ ret = trace_seq_printf(s, "[%5ld.%06ld] calling %s @ %i\n",
calltime.tv_sec,
calltime.tv_nsec,
it->func, it->caller);
if (!ret)
return TRACE_TYPE_PARTIAL_LINE;
- ret = trace_seq_printf(s, "[%5ld.%06ld] initcall %pF "
+
+ ret = trace_seq_printf(s, "[%5ld.%06ld] initcall %s "
"returned %d after %lld msecs\n",
rettime.tv_sec,
rettime.tv_nsec,
it->func, it->result, it->duration);
+
if (!ret)
return TRACE_TYPE_PARTIAL_LINE;
return TRACE_TYPE_HANDLED;
@@ -83,8 +86,7 @@ struct tracer boot_tracer __read_mostly =
.print_line = initcall_print_line,
};
-
-void trace_boot(struct boot_trace *it)
+void trace_boot(struct boot_trace *it, initcall_t fn)
{
struct ring_buffer_event *event;
struct trace_boot *entry;
@@ -95,6 +97,10 @@ void trace_boot(struct boot_trace *it)
if (!trace_boot_enabled)
return;
+ /* Get its name now since this function could
+ * disappear because it is in the .init section.
+ */
+ sprint_symbol(it->func, (unsigned long)fn);
preempt_disable();
data = tr->data[smp_processor_id()];
^ permalink raw reply related [flat|nested] 4+ messages in thread
* Re: [PATCH -tip 2/2] Tracing/fastboot: Get the initcall name before it disappears
2008-10-02 11:26 [PATCH -tip 2/2] Tracing/fastboot: Get the initcall name before it disappears Frederic Weisbecker
@ 2008-10-02 12:02 ` Ingo Molnar
2008-10-02 12:18 ` Frédéric Weisbecker
2008-10-02 15:47 ` [PATCH] tracing/fastboot: build fix Ingo Molnar
0 siblings, 2 replies; 4+ messages in thread
From: Ingo Molnar @ 2008-10-02 12:02 UTC (permalink / raw)
To: Frederic Weisbecker; +Cc: Steven Rostedt, linux-kernel
* Frederic Weisbecker <fweisbec@gmail.com> wrote:
> After some initcall traces, some initcall names may be inconsistent.
> That's because these functions will disappear from the .init section
> and also their name from the symbols table.
> So we have to copy the name of the function in a buffer large enough
> during the trace appending. It is not costly for the ring_buffer because
> the number of initcall entries is commonly not really large.
>
> Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com>
applied to tip/tracing/fastboot, thanks Frederic,
Ingo
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: [PATCH -tip 2/2] Tracing/fastboot: Get the initcall name before it disappears
2008-10-02 12:02 ` Ingo Molnar
@ 2008-10-02 12:18 ` Frédéric Weisbecker
2008-10-02 15:47 ` [PATCH] tracing/fastboot: build fix Ingo Molnar
1 sibling, 0 replies; 4+ messages in thread
From: Frédéric Weisbecker @ 2008-10-02 12:18 UTC (permalink / raw)
To: Ingo Molnar; +Cc: Steven Rostedt, linux-kernel, Arjan van de Ven
2008/10/2 Ingo Molnar <mingo@elte.hu>:
>
> * Frederic Weisbecker <fweisbec@gmail.com> wrote:
>
>> After some initcall traces, some initcall names may be inconsistent.
>> That's because these functions will disappear from the .init section
>> and also their name from the symbols table.
>> So we have to copy the name of the function in a buffer large enough
>> during the trace appending. It is not costly for the ring_buffer because
>> the number of initcall entries is commonly not really large.
>>
>> Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com>
>
> applied to tip/tracing/fastboot, thanks Frederic,
>
> Ingo
>
BTW I forgot to cc Arjan.
Arjan if you have some time to test it, I would be interested by your
feeling about it since
the boot tracer has now an output which is parsable by bootgraph.pl.
Thanks!
^ permalink raw reply [flat|nested] 4+ messages in thread
* [PATCH] tracing/fastboot: build fix
2008-10-02 12:02 ` Ingo Molnar
2008-10-02 12:18 ` Frédéric Weisbecker
@ 2008-10-02 15:47 ` Ingo Molnar
1 sibling, 0 replies; 4+ messages in thread
From: Ingo Molnar @ 2008-10-02 15:47 UTC (permalink / raw)
To: Frederic Weisbecker; +Cc: Steven Rostedt, linux-kernel
>From 7d14307eeaa710427436c09c6f77dde1974621cf Mon Sep 17 00:00:00 2001
From: Ingo Molnar <mingo@elte.hu>
Date: Thu, 2 Oct 2008 17:45:47 +0200
Subject: [PATCH] tracing/fastboot: build fix
fix:
In file included from kernel/sysctl.c:52:
include/linux/ftrace.h:217: error: 'KSYM_NAME_LEN' undeclared here (not in a function)
Signed-off-by: Ingo Molnar <mingo@elte.hu>
---
include/linux/ftrace.h | 4 ++--
1 files changed, 2 insertions(+), 2 deletions(-)
diff --git a/include/linux/ftrace.h b/include/linux/ftrace.h
index e672e51..deded11 100644
--- a/include/linux/ftrace.h
+++ b/include/linux/ftrace.h
@@ -1,14 +1,14 @@
#ifndef _LINUX_FTRACE_H
#define _LINUX_FTRACE_H
-#ifdef CONFIG_FTRACE
-
#include <linux/linkage.h>
#include <linux/fs.h>
#include <linux/init.h>
#include <linux/types.h>
#include <linux/kallsyms.h>
+#ifdef CONFIG_FTRACE
+
extern int ftrace_enabled;
extern int
ftrace_enable_sysctl(struct ctl_table *table, int write,
^ permalink raw reply related [flat|nested] 4+ messages in thread
end of thread, other threads:[~2008-10-02 15:47 UTC | newest]
Thread overview: 4+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2008-10-02 11:26 [PATCH -tip 2/2] Tracing/fastboot: Get the initcall name before it disappears Frederic Weisbecker
2008-10-02 12:02 ` Ingo Molnar
2008-10-02 12:18 ` Frédéric Weisbecker
2008-10-02 15:47 ` [PATCH] tracing/fastboot: build fix Ingo Molnar
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.