[PATCH 0/2] tracing: Fix how the tracer options file ups the trace_array ref count

[PATCH 0/2] tracing: Fix how the tracer options file ups the trace_array ref count

[Steven Rostedt]

Sashiko reported that the tracer option files pass a descriptor to
the file system operations that points to the trace_array it is apart
of as well as information on the tracer option. It uses the trace_array
to up its reference counter so that it can not be deleted when the option
is open. The issue is that this descriptor sent to the open call itself
can be removed by the removal of the instance the trace_array represents.
There's nothing protecting it for going away and dereferencing it after
it has been freed.

Another issue is that the tracer options are an array on an array for
each tracer in the trace_array descriptor. The first patch converts
that to a single array. This is need for the second patch.

The second patch changes the open function to search all the existing
trace_arrays under lock and compare their topts array with the topt
passed to the open function. If the topt is within the range of
a trace_array topts array then that trace_array is considered a match
and its reference is updated.

Steven Rostedt (2):
      tracing: Use a single array to represent tracer options files
      tracing: Take trace_array reference when opening a tracer options file

----
 kernel/trace/trace.c | 81 +++++++++++++++++++++++++++++++++-------------------
 kernel/trace/trace.h | 21 ++++++--------
 2 files changed, 60 insertions(+), 42 deletions(-)

[PATCH 1/2] tracing: Use a single array to represent tracer options files

[Steven Rostedt]

From: Steven Rostedt <rostedt@goodmis.org>

An instance is represented by a trace_array descriptor that has an array
for options of each tracer. That array is a pointer to another array that
represents every option for a specific tracer. Instead of having two
arrays use just a single array for every tracer option within a
trace_array.

This not only simplifies the logic for representing the tracer option in
the callbacks, but also can be used to simplify a way to find the
trace_array descriptor from this one single array when needing to find
which trace_array the option is attached to in order to take its
reference.

Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
---
 kernel/trace/trace.c | 47 ++++++++++++++++++--------------------------
 kernel/trace/trace.h | 21 ++++++++------------
 2 files changed, 27 insertions(+), 41 deletions(-)

diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c
index 722d0ba2d233..f26244253a8b 100644
--- a/kernel/trace/trace.c
+++ b/kernel/trace/trace.c
@@ -7689,7 +7689,7 @@ static ssize_t
 trace_options_read(struct file *filp, char __user *ubuf, size_t cnt,
 			loff_t *ppos)
 {
-	struct trace_option_dentry *topt = filp->private_data;
+	struct trace_options *topt = filp->private_data;
 	char *buf;
 
 	if (topt->flags->val & topt->opt->bit)
@@ -7704,7 +7704,7 @@ static ssize_t
 trace_options_write(struct file *filp, const char __user *ubuf, size_t cnt,
 			 loff_t *ppos)
 {
-	struct trace_option_dentry *topt = filp->private_data;
+	struct trace_options *topt = filp->private_data;
 	unsigned long val;
 	int ret;
 
@@ -7730,7 +7730,7 @@ trace_options_write(struct file *filp, const char __user *ubuf, size_t cnt,
 
 static int tracing_open_options(struct inode *inode, struct file *filp)
 {
-	struct trace_option_dentry *topt = inode->i_private;
+	struct trace_options *topt = inode->i_private;
 	int ret;
 
 	ret = tracing_check_open_get_tr(topt->tr);
@@ -7743,7 +7743,7 @@ static int tracing_open_options(struct inode *inode, struct file *filp)
 
 static int tracing_release_options(struct inode *inode, struct file *file)
 {
-	struct trace_option_dentry *topt = file->private_data;
+	struct trace_options *topt = file->private_data;
 
 	trace_array_put(topt->tr);
 	return 0;
@@ -7946,8 +7946,9 @@ static struct dentry *trace_options_init_dentry(struct trace_array *tr)
 
 static void
 create_trace_option_file(struct trace_array *tr,
-			 struct trace_option_dentry *topt,
+			 struct trace_options *topt,
 			 struct tracer_flags *flags,
+			 struct tracer *tracer,
 			 struct tracer_opt *opt)
 {
 	struct dentry *t_options;
@@ -7959,6 +7960,7 @@ create_trace_option_file(struct trace_array *tr,
 	topt->flags = flags;
 	topt->opt = opt;
 	topt->tr = tr;
+	topt->tracer = tracer;
 
 	topt->entry = trace_create_file(opt->name, TRACE_MODE_WRITE,
 					t_options, topt, &trace_options_fops);
@@ -7968,10 +7970,9 @@ static int
 create_trace_option_files(struct trace_array *tr, struct tracer *tracer,
 			  struct tracer_flags *flags)
 {
-	struct trace_option_dentry *topts;
-	struct trace_options *tr_topts;
+	struct trace_options *topts;
 	struct tracer_opt *opts;
-	int cnt;
+	int i, cnt;
 
 	if (!flags || !flags->opts)
 		return 0;
@@ -7981,29 +7982,22 @@ create_trace_option_files(struct trace_array *tr, struct tracer *tracer,
 	for (cnt = 0; opts[cnt].name; cnt++)
 		;
 
-	topts = kzalloc_objs(*topts, cnt + 1);
-	if (!topts)
-		return 0;
-
-	tr_topts = krealloc_array(tr->topts, tr->nr_topts + 1, sizeof(*tr->topts),
+	topts = krealloc_array(tr->topts, tr->nr_topts + cnt, sizeof(*tr->topts),
 				  GFP_KERNEL);
-	if (!tr_topts) {
-		kfree(topts);
+	if (!topts)
 		return -ENOMEM;
-	}
 
-	tr->topts = tr_topts;
-	tr->topts[tr->nr_topts].tracer = tracer;
-	tr->topts[tr->nr_topts].topts = topts;
-	tr->nr_topts++;
+	tr->topts = topts;
 
-	for (cnt = 0; opts[cnt].name; cnt++) {
-		create_trace_option_file(tr, &topts[cnt], flags,
-					 &opts[cnt]);
-		MEM_FAIL(topts[cnt].entry == NULL,
+	for (topts += tr->nr_topts, i = 0; i < cnt; topts++, i++) {
+		create_trace_option_file(tr, topts, flags, tracer,
+					 &opts[i]);
+		MEM_FAIL(topts->entry == NULL,
 			  "Failed to create trace option: %s",
-			  opts[cnt].name);
+			  opts[i].name);
 	}
+
+	tr->nr_topts += cnt;
 	return 0;
 }
 
@@ -8865,9 +8859,6 @@ static int __remove_instance(struct trace_array *tr)
 	if (tr->flags & TRACE_ARRAY_FL_VMALLOC)
 		vfree((void *)tr->range_addr_start);
 
-	for (i = 0; i < tr->nr_topts; i++) {
-		kfree(tr->topts[i].topts);
-	}
 	kfree(tr->topts);
 
 	free_cpumask_var(tr->pipe_cpumask);
diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h
index 3c111ca88e32..a07802be5673 100644
--- a/kernel/trace/trace.h
+++ b/kernel/trace/trace.h
@@ -224,11 +224,6 @@ struct array_buffer {
 
 #define TRACE_FLAGS_MAX_SIZE		64
 
-struct trace_options {
-	struct tracer			*tracer;
-	struct trace_option_dentry	*topts;
-};
-
 struct trace_pid_list *trace_pid_list_alloc(void);
 void trace_pid_list_free(struct trace_pid_list *pid_list);
 bool trace_pid_list_is_set(struct trace_pid_list *pid_list, unsigned int pid);
@@ -325,6 +320,14 @@ struct trace_module_delta {
 	long		delta[];
 };
 
+struct trace_options {
+	struct tracer_opt		*opt;
+	struct tracer_flags		*flags;
+	struct trace_array		*tr;
+	struct tracer			*tracer;
+	struct dentry			*entry;
+};
+
 /*
  * The trace array - an array of per-CPU trace arrays. This is the
  * highest level data structure that individual tracers deal with.
@@ -605,14 +608,6 @@ struct tracer_flags {
 /* Makes more easy to define a tracer opt */
 #define TRACER_OPT(s, b)	.name = #s, .bit = b
 
-
-struct trace_option_dentry {
-	struct tracer_opt		*opt;
-	struct tracer_flags		*flags;
-	struct trace_array		*tr;
-	struct dentry			*entry;
-};
-
 /**
  * struct tracer - a specific tracer and its callbacks to interact with tracefs
  * @name: the name chosen to select it on the available_tracers file
-- 
2.53.0

[PATCH 2/2] tracing: Take trace_array reference when opening a tracer options file

[Steven Rostedt]

From: Steven Rostedt <rostedt@goodmis.org>

When a tracer option file is opened, it is passed a descriptor that points
to an element on the trace_array's topts array. This element has
information to find the trace array and other information. It uses this
element to take a reference of the trace_array so that the trace_array
does not get removed while this file is opened.

Unfortunately, there's a race condition where the element itself could be
freed by the removal of the instance the trace_array represents causing a
use-after-free as this element that is used to find the trace_array to
increment its reference counter is also freed when the instance is
removed.

To solve this, add a trace_array_tracer_options_get() helper function that
will take the address of the element that is passed to the open function
by the inode->i_private pointer and search all the trace_arrays under a
lock to find the one that the element's address is in the range of the
trace_arrays topts array. When a match happens, that trace_array's
reference would be increased.

Note, there's a race where if an admin was deleting and creating trace
instances at the same time and the memory of the old trace_array's array
matched the memory of the new trace_array that it could in theory open the
option from the wrong trace array. But we do not care because it would be
stupid to perform that kind of action. As long as the only thing that can
happen is that the option from the wrong trace array is used and doesn't
crash the kernel it will only make the user confused. But if they are
doing something stupid like this, they are already confused, so no harm
done.

Cc: stable@vger.kernel.org
Fixes: 7e2cfbd2d3c86 ("tracing: Have option files inc the trace array ref count")
Reported-by: sashiko-bot@kernel.org
Closes: https://lore.kernel.org/linux-trace-kernel/20260902121918.5a9e9d1b@gandalf.local.home/
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
---
 kernel/trace/trace.c | 34 +++++++++++++++++++++++++++++++++-
 1 file changed, 33 insertions(+), 1 deletion(-)

diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c
index f26244253a8b..03080fcf7f04 100644
--- a/kernel/trace/trace.c
+++ b/kernel/trace/trace.c
@@ -7728,12 +7728,44 @@ trace_options_write(struct file *filp, const char __user *ubuf, size_t cnt,
 	return cnt;
 }
 
+/*
+ * The topt is the address of a trace_array->topts[] element that holds the
+ * the tracer options descriptor. But since the trace_array reference has not
+ * been taken yet, it cannot be dereferenced as it could have been freed by
+ * a rmdir of the instance the trace_array represents.
+ *
+ * Search the list of trace_arrays and compare the topt to the address of
+ * the entire trace_array topts array for each trace_array in the list.
+ * If one is matched, then take the reference and return it. If not, the
+ * trace_array no longer exits.
+ */
+static int trace_array_tracer_options_get(void *topt)
+{
+	struct trace_array *tr;
+	int ret;
+
+	ret = security_locked_down(LOCKDOWN_TRACEFS);
+	if (ret)
+		return ret;
+
+	if (tracing_disabled)
+		return -ENODEV;
+
+	guard(mutex)(&trace_types_lock);
+	list_for_each_entry(tr, &ftrace_trace_arrays, list) {
+		if (topt >= (void *)&tr->topts[0] &&
+		    topt < (void *)&tr->topts[tr->nr_topts])
+			return __trace_array_get(tr);
+	}
+	return -ENODEV;
+}
+
 static int tracing_open_options(struct inode *inode, struct file *filp)
 {
 	struct trace_options *topt = inode->i_private;
 	int ret;
 
-	ret = tracing_check_open_get_tr(topt->tr);
+	ret = trace_array_tracer_options_get(topt);
 	if (ret)
 		return ret;
 
-- 
2.53.0