Skip to contentSkip to conversation

tracing/syscalls: fix missing release of tracing

tip:tracing/syscalls Frederic Weisbecker 1 participant 1 message Mar 16, 2009 1 message not in the archive Tracing

lore
Share & tools
lore
#1 lore raw
In reply to a message not in the archive
Commit-ID:  ac99c58c9e56967037382e31f865b72b10127965
Gitweb:     http://git.kernel.org/tip/ac99c58c9e56967037382e31f865b72b10127965
Author:     Frederic Weisbecker <[email protected]>
AuthorDate: Sun, 15 Mar 2009 22:10:35 +0100
Commit:     Ingo Molnar <[email protected]>
CommitDate: Mon, 16 Mar 2009 09:13:15 +0100

tracing/syscalls: fix missing release of tracing

Impact: fix 'stuck' syscall tracer

The syscall tracer uses a refcounter to enable several users simultaneously.

But the refcounter did not behave correctly and always restored its value to 0 after calling start_syscall_tracing(). Therefore, stop_syscall_tracing() couldn't release correctly the tasks from tracing.

Also the tracer forgot to reset the buffer when it is released.

Drop the pointless refcount decrement on start_syscall_tracing() and reset the buffer when we release the tracer.

This fixes two reported issue:

- when we switch from syscall tracer to another tracer, syscall tracing continued.

- incorrect use of the refcount.

LKML-Reference: <[email protected]>

---
 kernel/trace/trace_syscalls.c |   11 +++++------
 1 files changed, 5 insertions(+), 6 deletions(-)
kernel/trace/trace_syscalls.c +5 -6
index c72e599..c5fc1d8 100644
@@ -96,8 +96,9 @@ void start_ftrace_syscalls(void)
 	unsigned long flags;
 	struct task_struct *g, *t;
 
+	/* Don't enable the flag on the tasks twice */
 	if (atomic_inc_return(&refcount) != 1)
-		goto out;
+		return;
 
 	arch_init_ftrace_syscalls();
 	read_lock_irqsave(&tasklist_lock, flags);
@@ -107,8 +108,6 @@ void start_ftrace_syscalls(void)
 	} while_each_thread(g, t);
 
 	read_unlock_irqrestore(&tasklist_lock, flags);
-out:
-	atomic_dec(&refcount);
 }
 
 void stop_ftrace_syscalls(void)
@@ -116,8 +115,9 @@ void stop_ftrace_syscalls(void)
 	unsigned long flags;
 	struct task_struct *g, *t;
 
+	/* There are perhaps still some users */
 	if (atomic_dec_return(&refcount))
-		goto out;
+		return;
 
 	read_lock_irqsave(&tasklist_lock, flags);
 
@@ -126,8 +126,6 @@ void stop_ftrace_syscalls(void)
 	} while_each_thread(g, t);
 
 	read_unlock_irqrestore(&tasklist_lock, flags);
-out:
-	atomic_inc(&refcount);
 }
 
 void ftrace_syscall_enter(struct pt_regs *regs)
@@ -201,6 +199,7 @@ static int init_syscall_tracer(struct trace_array *tr)
 static void reset_syscall_tracer(struct trace_array *tr)
 {
 	stop_ftrace_syscalls();
+	tracing_reset_online_cpus(tr);
 }
 
 static struct trace_event syscall_enter_event = {