tracing: Add interface to allow multiple trace buffers

Add the interface ("instances" directory) to add multiple buffers
to ftrace. To create a new instance, simply do a mkdir in the
instances directory:

This will create a directory with the following:

 # cd instances
 # mkdir foo
 # ls foo
buffer_size_kb        free_buffer  trace_clock    trace_pipe
buffer_total_size_kb  set_event    trace_marker   tracing_enabled
events/               trace        trace_options  tracing_on

Currently only events are able to be set, and there isn't a way
to delete a buffer when one is created (yet).

Note, the i_mutex lock is dropped from the parent "instances"
directory during the mkdir operation. As the "instances" directory
can not be renamed or deleted (created on boot), I do not see
any harm in dropping the lock. The creation of the sub directories
is protected by trace_types_lock mutex, which only lets one
instance get into the code path at a time. If two tasks try to
create or delete directories of the same name, only one will occur
and the other will fail with -EEXIST.

Cc: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>

3 files changed, 142 insertions, 1 deletions

diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c
index 406adbc277a0..07a63114d938 100644
--- a/kernel/trace/trace.c
+++ b/kernel/trace/trace.c
@@ -5107,6 +5107,133 @@ static const struct file_operations rb_simple_fops = {
         .llseek         = default_llseek,
 };
 
+struct dentry *trace_instance_dir;
+
+static void
+init_tracer_debugfs(struct trace_array *tr, struct dentry *d_tracer);
+
+static int new_instance_create(const char *name)
+{
+        enum ring_buffer_flags rb_flags;
+        struct trace_array *tr;
+        int ret;
+        int i;
+
+        mutex_lock(&trace_types_lock);
+
+        ret = -EEXIST;
+        list_for_each_entry(tr, &ftrace_trace_arrays, list) {
+                if (tr->name && strcmp(tr->name, name) == 0)
+                        goto out_unlock;
+        }
+
+        ret = -ENOMEM;
+        tr = kzalloc(sizeof(*tr), GFP_KERNEL);
+        if (!tr)
+                goto out_unlock;
+
+        tr->name = kstrdup(name, GFP_KERNEL);
+        if (!tr->name)
+                goto out_free_tr;
+
+        raw_spin_lock_init(&tr->start_lock);
+
+        tr->current_trace = &nop_trace;
+
+        INIT_LIST_HEAD(&tr->systems);
+        INIT_LIST_HEAD(&tr->events);
+
+        rb_flags = trace_flags & TRACE_ITER_OVERWRITE ? RB_FL_OVERWRITE : 0;
+
+        tr->buffer = ring_buffer_alloc(trace_buf_size, rb_flags);
+        if (!tr->buffer)
+                goto out_free_tr;
+
+        tr->data = alloc_percpu(struct trace_array_cpu);
+        if (!tr->data)
+                goto out_free_tr;
+
+        for_each_tracing_cpu(i) {
+                memset(per_cpu_ptr(tr->data, i), 0, sizeof(struct trace_array_cpu));
+                per_cpu_ptr(tr->data, i)->trace_cpu.cpu = i;
+                per_cpu_ptr(tr->data, i)->trace_cpu.tr = tr;
+        }
+
+        /* Holder for file callbacks */
+        tr->trace_cpu.cpu = RING_BUFFER_ALL_CPUS;
+        tr->trace_cpu.tr = tr;
+
+        tr->dir = debugfs_create_dir(name, trace_instance_dir);
+        if (!tr->dir)
+                goto out_free_tr;
+
+        ret = event_trace_add_tracer(tr->dir, tr);
+        if (ret)
+                goto out_free_tr;
+
+        init_tracer_debugfs(tr, tr->dir);
+
+        list_add(&tr->list, &ftrace_trace_arrays);
+
+        mutex_unlock(&trace_types_lock);
+
+        return 0;
+
+ out_free_tr:
+        if (tr->buffer)
+                ring_buffer_free(tr->buffer);
+        kfree(tr->name);
+        kfree(tr);
+
+ out_unlock:
+        mutex_unlock(&trace_types_lock);
+
+        return ret;
+
+}
+
+static int instance_mkdir (struct inode *inode, struct dentry *dentry, umode_t mode)
+{
+        struct dentry *parent;
+        int ret;
+
+        /* Paranoid: Make sure the parent is the "instances" directory */
+        parent = hlist_entry(inode->i_dentry.first, struct dentry, d_alias);
+        if (WARN_ON_ONCE(parent != trace_instance_dir))
+                return -ENOENT;
+
+        /*
+         * The inode mutex is locked, but debugfs_create_dir() will also
+         * take the mutex. As the instances directory can not be destroyed
+         * or changed in any other way, it is safe to unlock it, and
+         * let the dentry try. If two users try to make the same dir at
+         * the same time, then the new_instance_create() will determine the
+         * winner.
+         */
+        mutex_unlock(&inode->i_mutex);
+
+        ret = new_instance_create(dentry->d_iname);
+
+        mutex_lock(&inode->i_mutex);
+
+        return ret;
+}
+
+static const struct inode_operations instance_dir_inode_operations = {
+        .lookup         = simple_lookup,
+        .mkdir          = instance_mkdir,
+};
+
+static __init void create_trace_instances(struct dentry *d_tracer)
+{
+        trace_instance_dir = debugfs_create_dir("instances", d_tracer);
+        if (WARN_ON(!trace_instance_dir))
+                return;
+
+        /* Hijack the dir inode operations, to allow mkdir */
+        trace_instance_dir->d_inode->i_op = &instance_dir_inode_operations;
+}
+
 static void
 init_tracer_debugfs(struct trace_array *tr, struct dentry *d_tracer)
 {
@@ -5183,6 +5310,8 @@ static __init int tracer_init_debugfs(void)
                           (void *)&global_trace.trace_cpu, &snapshot_fops);
 #endif
 
+        create_trace_instances(d_tracer);
+
         create_trace_options_dir(&global_trace);
 
         for_each_tracing_cpu(cpu)
diff --git a/kernel/trace/trace.h b/kernel/trace/trace.h
index 68cad7a9e089..883fe0b62f0a 100644
--- a/kernel/trace/trace.h
+++ b/kernel/trace/trace.h
@@ -175,6 +175,7 @@ struct tracer;
 struct trace_array {
         struct ring_buffer      *buffer;
         struct list_head        list;
+        char                    *name;
         int                     cpu;
         int                     buffer_disabled;
         struct trace_cpu        trace_cpu;      /* place holder */
@@ -999,6 +1000,7 @@ filter_check_discard(struct ftrace_event_call *call, void *rec,
 }
 
 extern void trace_event_enable_cmd_record(bool enable);
+extern int event_trace_add_tracer(struct dentry *parent, struct trace_array *tr);
 
 extern struct mutex event_mutex;
 extern struct list_head ftrace_events;
diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c
index 439955239bae..58a61302a733 100644
--- a/kernel/trace/trace_events.c
+++ b/kernel/trace/trace_events.c
@@ -1754,16 +1754,22 @@ int event_trace_add_tracer(struct dentry *parent, struct trace_array *tr)
         struct dentry *d_events;
         struct dentry *entry;
 
+        mutex_lock(&event_mutex);
+
         entry = debugfs_create_file("set_event", 0644, parent,
                                     tr, &ftrace_set_event_fops);
         if (!entry) {
                 pr_warning("Could not create debugfs 'set_event' entry\n");
+                mutex_unlock(&event_mutex);
                 return -ENOMEM;
         }
 
         d_events = debugfs_create_dir("events", parent);
-        if (!d_events)
+        if (!d_events) {
                 pr_warning("Could not create debugfs 'events' directory\n");
+                mutex_unlock(&event_mutex);
+                return -ENOMEM;
+        }
 
         /* ring buffer internal formats */
         trace_create_file("header_page", 0444, d_events,
@@ -1778,7 +1784,11 @@ int event_trace_add_tracer(struct dentry *parent, struct trace_array *tr)
                           tr, &ftrace_tr_enable_fops);
 
         tr->event_dir = d_events;
+        down_write(&trace_event_mutex);
         __trace_add_event_dirs(tr);
+        up_write(&trace_event_mutex);
+
+        mutex_unlock(&event_mutex);
 
         return 0;
 }