Skip to contentSkip to conversation

ftrace: fix kmemleak in unregister_ftrace_graph

PATCH In review [email protected] 2 participants 3 messages Sep 12, 2017 – Oct 3, 2017

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

PATCH 1/1 ftrace: fix kmemleak in unregister_ftrace_graph

From: Shu Wang <[email protected]>

The trampoline allocated by function tracer was overwrite by function_graph tracer, and cause memory leak. The save_global_trampoline should save previous trampoline in register_ftrace_graph and restore it in unregister_ftrace_graph. But as it implemented, save_global_trampoline was only used in unregister_ftrace_graph as default value 0, and overwrite the previous trampoline's value.

kmmeleak backtrace:

    kmemleak_vmalloc+0x77/0xc0
    __vmalloc_node_range+0x1b5/0x2c0
    module_alloc+0x7c/0xd0
    arch_ftrace_update_trampoline+0xb5/0x290
    ftrace_startup+0x78/0x210
    register_ftrace_function+0x8b/0xd0
    function_trace_init+0x4f/0x80
    tracing_set_tracer+0xe6/0x170
    tracing_set_trace_write+0x90/0xd0
    __vfs_write+0x37/0x170
    vfs_write+0xb2/0x1b0
    SyS_write+0x55/0xc0
    do_syscall_64+0x67/0x180
    return_from_SYSCALL_64+0x0/0x6a
---
 kernel/trace/ftrace.c | 14 --------------
 1 file changed, 14 deletions(-)
kernel/trace/ftrace.c +0 -14
index 6abfafd7f173..8319e09e15b9 100644
@@ -4954,9 +4954,6 @@ static char ftrace_graph_buf[FTRACE_FILTER_SIZE] __initdata;
 static char ftrace_graph_notrace_buf[FTRACE_FILTER_SIZE] __initdata;
 static int ftrace_graph_set_hash(struct ftrace_hash *hash, char *buffer);
 
-static unsigned long save_global_trampoline;
-static unsigned long save_global_flags;
-
 static int __init set_graph_function(char *str)
 {
 	strlcpy(ftrace_graph_buf, str, FTRACE_FILTER_SIZE);
@@ -6808,17 +6805,6 @@ void unregister_ftrace_graph(void)
 	unregister_pm_notifier(&ftrace_suspend_notifier);
 	unregister_trace_sched_switch(ftrace_graph_probe_sched_switch, NULL);
 
-#ifdef CONFIG_DYNAMIC_FTRACE
-	/*
-	 * Function graph does not allocate the trampoline, but
-	 * other global_ops do. We need to reset the ALLOC_TRAMP flag
-	 * if one was used.
-	 */
-	global_ops.trampoline = save_global_trampoline;
-	if (save_global_flags & FTRACE_OPS_FL_ALLOC_TRAMP)
-		global_ops.flags |= FTRACE_OPS_FL_ALLOC_TRAMP;
-#endif
-
  out:
 	mutex_unlock(&ftrace_lock);
 }
Signature · 1 line
2.13.5
author
#1 lore raw
Quoted 72 lines

From: [email protected]
To: [email protected], [email protected]
Cc: [email protected], [email protected], [email protected], "Shu Wang" <[email protected]>
Sent: Tuesday, September 12, 2017 10:14:54 AM
Subject: [PATCH] ftrace: fix kmemleak in unregister_ftrace_graph

From: Shu Wang <[email protected]>

The trampoline allocated by function tracer was overwrite by function_graph tracer, and cause memory leak. The save_global_trampoline should save previous trampoline in register_ftrace_graph and restore it in unregister_ftrace_graph. But as it implemented, save_global_trampoline was only used in unregister_ftrace_graph as default value 0, and overwrite the previous trampoline's value.

kmmeleak backtrace:

    kmemleak_vmalloc+0x77/0xc0
    __vmalloc_node_range+0x1b5/0x2c0
    module_alloc+0x7c/0xd0
    arch_ftrace_update_trampoline+0xb5/0x290
    ftrace_startup+0x78/0x210
    register_ftrace_function+0x8b/0xd0
    function_trace_init+0x4f/0x80
    tracing_set_tracer+0xe6/0x170
    tracing_set_trace_write+0x90/0xd0
    __vfs_write+0x37/0x170
    vfs_write+0xb2/0x1b0
    SyS_write+0x55/0xc0
    do_syscall_64+0x67/0x180
    return_from_SYSCALL_64+0x0/0x6a
---
 kernel/trace/ftrace.c | 14 --------------
 1 file changed, 14 deletions(-)
kernel/trace/ftrace.c +0 -3
index 6abfafd7f173..8319e09e15b9 100644
@@ -4954,9 +4954,6 @@ static char ftrace_graph_buf[FTRACE_FILTER_SIZE]
 __initdata;
 static char ftrace_graph_notrace_buf[FTRACE_FILTER_SIZE] __initdata;
 static int ftrace_graph_set_hash(struct ftrace_hash *hash, char *buffer);
 
-static unsigned long save_global_trampoline;
-static unsigned long save_global_flags;
-
 static int __init set_graph_function(char *str)
 {
kernel/trace/ftrace.c +0 -12
 	strlcpy(ftrace_graph_buf, str, FTRACE_FILTER_SIZE);
@@ -6808,17 +6805,6 @@ void unregister_ftrace_graph(void)
 	unregister_pm_notifier(&ftrace_suspend_notifier);
 	unregister_trace_sched_switch(ftrace_graph_probe_sched_switch, NULL);
 
-#ifdef CONFIG_DYNAMIC_FTRACE
-	/*
-	 * Function graph does not allocate the trampoline, but
-	 * other global_ops do. We need to reset the ALLOC_TRAMP flag
-	 * if one was used.
-	 */
-	global_ops.trampoline = save_global_trampoline;
-	if (save_global_flags & FTRACE_OPS_FL_ALLOC_TRAMP)
-		global_ops.flags |= FTRACE_OPS_FL_ALLOC_TRAMP;
-#endif
-
  out:
 	mutex_unlock(&ftrace_lock);
 }
--
Quoted 3 lines

2.13.5

ping

maintainer
#2 lore raw
Quoted 33 lines · On Tue, 26 Sep 2017 08:01:26 -0400 (EDT) Shu Wang <[email protected]> wrote:
Quoted 33 lines

From: [email protected]
To: [email protected], [email protected]
Cc: [email protected], [email protected], [email protected], "Shu Wang" <[email protected]>
Sent: Tuesday, September 12, 2017 10:14:54 AM
Subject: [PATCH] ftrace: fix kmemleak in unregister_ftrace_graph

From: Shu Wang <[email protected]>

The trampoline allocated by function tracer was overwrite by function_graph tracer, and cause memory leak. The save_global_trampoline should save previous trampoline in register_ftrace_graph and restore it in unregister_ftrace_graph. But as it implemented, save_global_trampoline was only used in unregister_ftrace_graph as default value 0, and overwrite the previous trampoline's value.

kmmeleak backtrace:

    kmemleak_vmalloc+0x77/0xc0
    __vmalloc_node_range+0x1b5/0x2c0
    module_alloc+0x7c/0xd0
    arch_ftrace_update_trampoline+0xb5/0x290
    ftrace_startup+0x78/0x210
    register_ftrace_function+0x8b/0xd0
    function_trace_init+0x4f/0x80
    tracing_set_tracer+0xe6/0x170
    tracing_set_trace_write+0x90/0xd0
    __vfs_write+0x37/0x170
    vfs_write+0xb2/0x1b0
    SyS_write+0x55/0xc0
    do_syscall_64+0x67/0x180
    return_from_SYSCALL_64+0x0/0x6a
Quoted 7 lines
Quoted 5 lines

ping

Thanks for the reminder. I pulled it in and will send this out to Linus after the testing finishes. I also marked it for stable.

-- Steve