From: Frederic Weisbecker <fweisbec@gmail.com>
To: Ingo Molnar <mingo@elte.hu>
Cc: Steven Rostedt <rostedt@goodmis.org>,
Andrew Morton <akpm@linux-foundation.org>,
Thomas Gleixner <tglx@linutronix.de>,
Linux Kernel <linux-kernel@vger.kernel.org>
Subject: [PATCH] tracing/function-graph-tracer: prevent from hrtimer interrupt eternal loop
Date: Thu, 18 Dec 2008 00:52:47 +0100 [thread overview]
Message-ID: <494990CF.70504@gmail.com> (raw)
Impact: fix a system hang on slow systems
While testing the function graph tracer on VirtualBox, I had a system hang
immediately after enabling the tracer.
If hrtimer is enabled on kernel, a slow system can spend too much time during
tracing the hrtimer_interrupt which will do eternal loops, assuming it always
have to retry its process because too much time elapsed during its time
update.
Now we provide a feature which lurks at the number of retries on
hrtimer_interrupt. After 10 retries, the function graph tracer will finally stop
its tracing.
Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Steven Rostedt <rostedt@goodmis.org>
---
arch/x86/kernel/ftrace.c | 2 ++
include/linux/ftrace.h | 31 +++++++++++++++++++++++++++++++
kernel/hrtimer.c | 5 +++++
3 files changed, 38 insertions(+), 0 deletions(-)
diff --git a/arch/x86/kernel/ftrace.c b/arch/x86/kernel/ftrace.c
index 1b43086..244f178 100644
--- a/arch/x86/kernel/ftrace.c
+++ b/arch/x86/kernel/ftrace.c
@@ -389,6 +389,8 @@ void ftrace_nmi_exit(void)
#endif /* !CONFIG_DYNAMIC_FTRACE */
+DEFINE_PER_CPU(int, ftrace_graph_hrtimer_loop) = LOCAL_INIT(0);
+
/* Add a function return address to the trace stack on thread info.*/
static int push_return_trace(unsigned long ret, unsigned long long time,
unsigned long func, int *depth)
diff --git a/include/linux/ftrace.h b/include/linux/ftrace.h
index 286af82..39be782 100644
--- a/include/linux/ftrace.h
+++ b/include/linux/ftrace.h
@@ -408,6 +408,34 @@ extern void unregister_ftrace_graph(void);
extern void ftrace_graph_init_task(struct task_struct *t);
extern void ftrace_graph_exit_task(struct task_struct *t);
+/*
+ * Those two functions prevent from eternal loops on hrtimer_interrupt
+ * because the function graph tracer can cause two much time processing
+ * on slow systems, and the hrtimer_interrupt can relaunch its processing
+ * if too much time elapsed during its previous time update.
+ *
+ * Note that we don't need an atomic counter for loops here because
+ * interrupts are disabled in hrtimer_interrupt.
+ */
+
+extern DEFINE_PER_CPU(int, ftrace_graph_hrtimer_loop);
+
+static inline void ftrace_graph_hrtimer_enter(void)
+{
+ __get_cpu_var(ftrace_graph_hrtimer_loop) = 0;
+}
+
+static inline void ftrace_graph_hrtimer_retry(void)
+{
+ __get_cpu_var(ftrace_graph_hrtimer_loop)++;
+
+ if (__get_cpu_var(ftrace_graph_hrtimer_loop) == 10) {
+ ftrace_graph_stop();
+ printk(KERN_WARNING "function-graph-tracer: hrtimer_interrupt "
+ "exceeded 10 loops. Stopping tracing\n");
+ }
+}
+
static inline int task_curr_ret_stack(struct task_struct *t)
{
return t->curr_ret_stack;
@@ -437,6 +465,9 @@ static inline int task_curr_ret_stack(struct task_struct *tsk)
static inline void pause_graph_tracing(void) { }
static inline void unpause_graph_tracing(void) { }
+
+static inline void ftrace_graph_hrtimer_enter(void) { };
+static inline void ftrace_graph_hrtimer_retry(void) { };
#endif
#ifdef CONFIG_TRACING
diff --git a/kernel/hrtimer.c b/kernel/hrtimer.c
index b741f85..662eb53 100644
--- a/kernel/hrtimer.c
+++ b/kernel/hrtimer.c
@@ -44,6 +44,7 @@
#include <linux/seq_file.h>
#include <linux/err.h>
#include <linux/debugobjects.h>
+#include <linux/ftrace.h>
#include <asm/uaccess.h>
@@ -1186,7 +1187,11 @@ void hrtimer_interrupt(struct clock_event_device *dev)
cpu_base->nr_events++;
dev->next_event.tv64 = KTIME_MAX;
+ ftrace_graph_hrtimer_enter();
retry:
+ /* Prevent from eternal loop on slow systems while tracing */
+ ftrace_graph_hrtimer_retry();
+
now = ktime_get();
expires_next.tv64 = KTIME_MAX;
--
1.6.0.4
next reply other threads:[~2008-12-17 23:53 UTC|newest]
Thread overview: 3+ messages / expand[flat|nested] mbox.gz Atom feed top
2008-12-17 23:52 Frederic Weisbecker [this message]
2008-12-17 23:59 ` [PATCH] tracing/function-graph-tracer: prevent from hrtimer interrupt eternal loop Andrew Morton
2008-12-18 0:17 ` Frédéric Weisbecker
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=494990CF.70504@gmail.com \
--to=fweisbec@gmail.com \
--cc=akpm@linux-foundation.org \
--cc=linux-kernel@vger.kernel.org \
--cc=mingo@elte.hu \
--cc=rostedt@goodmis.org \
--cc=tglx@linutronix.de \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
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.