binder: print warnings when detecting oneway spamming.

The most common cause of the binder transaction buffer filling up is a
client rapidly firing oneway transactions into a process, before it has
a chance to handle them. Yet the root cause of this is often hard to
debug, because either the system or the app will stop, and by that time
binder debug information we dump in bugreports is no longer relevant.

This change warns as soon as a process dips below 80% of its oneway
space (less than 100kB available in the configuration), when any one
process is responsible for either more than 50 transactions, or more
than 50% of the oneway space.

Signed-off-by: Martijn Coenen <maco@android.com>
Acked-by: Todd Kjos <tkjos@google.com>
Link: https://lore.kernel.org/r/20200821122544.1277051-1-maco@android.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

1 files changed, 52 insertions, 3 deletions

diff --git a/drivers/android/binder_alloc.c b/drivers/android/binder_alloc.c
index 8c98d12b7beb..910c53ba2c91 100644
--- a/drivers/android/binder_alloc.c
+++ b/drivers/android/binder_alloc.c
@@ -338,12 +338,50 @@ static inline struct vm_area_struct *binder_alloc_get_vma(
 	return vma;
 }
 
+static void debug_low_async_space_locked(struct binder_alloc *alloc, int pid)
+{
+	/*
+	 * Find the amount and size of buffers allocated by the current caller;
+	 * The idea is that once we cross the threshold, whoever is responsible
+	 * for the low async space is likely to try to send another async txn,
+	 * and at some point we'll catch them in the act. This is more efficient
+	 * than keeping a map per pid.
+	 */
+	struct rb_node *n = alloc->free_buffers.rb_node;
+	struct binder_buffer *buffer;
+	size_t total_alloc_size = 0;
+	size_t num_buffers = 0;
+
+	for (n = rb_first(&alloc->allocated_buffers); n != NULL;
+		 n = rb_next(n)) {
+		buffer = rb_entry(n, struct binder_buffer, rb_node);
+		if (buffer->pid != pid)
+			continue;
+		if (!buffer->async_transaction)
+			continue;
+		total_alloc_size += binder_alloc_buffer_size(alloc, buffer)
+			+ sizeof(struct binder_buffer);
+		num_buffers++;
+	}
+
+	/*
+	 * Warn if this pid has more than 50 transactions, or more than 50% of
+	 * async space (which is 25% of total buffer size).
+	 */
+	if (num_buffers > 50 || total_alloc_size > alloc->buffer_size / 4) {
+		binder_alloc_debug(BINDER_DEBUG_USER_ERROR,
+			     "%d: pid %d spamming oneway? %zd buffers allocated for a total size of %zd\n",
+			      alloc->pid, pid, num_buffers, total_alloc_size);
+	}
+}
+
 static struct binder_buffer *binder_alloc_new_buf_locked(
 				struct binder_alloc *alloc,
 				size_t data_size,
 				size_t offsets_size,
 				size_t extra_buffers_size,
-				int is_async)
+				int is_async,
+				int pid)
 {
 	struct rb_node *n = alloc->free_buffers.rb_node;
 	struct binder_buffer *buffer;
@@ -486,11 +524,20 @@ static struct binder_buffer *binder_alloc_new_buf_locked(
 	buffer->offsets_size = offsets_size;
 	buffer->async_transaction = is_async;
 	buffer->extra_buffers_size = extra_buffers_size;
+	buffer->pid = pid;
 	if (is_async) {
 		alloc->free_async_space -= size + sizeof(struct binder_buffer);
 		binder_alloc_debug(BINDER_DEBUG_BUFFER_ALLOC_ASYNC,
 			     "%d: binder_alloc_buf size %zd async free %zd\n",
 			      alloc->pid, size, alloc->free_async_space);
+		if (alloc->free_async_space < alloc->buffer_size / 10) {
+			/*
+			 * Start detecting spammers once we have less than 20%
+			 * of async space left (which is less than 10% of total
+			 * buffer size).
+			 */
+			debug_low_async_space_locked(alloc, pid);
+		}
 	}
 	return buffer;
 
@@ -508,6 +555,7 @@ err_alloc_buf_struct_failed:
  * @offsets_size:       user specified buffer offset
  * @extra_buffers_size: size of extra space for meta-data (eg, security context)
  * @is_async:           buffer for async transaction
+ * @pid:				pid to attribute allocation to (used for debugging)
  *
  * Allocate a new buffer given the requested sizes. Returns
  * the kernel version of the buffer pointer. The size allocated
@@ -520,13 +568,14 @@ struct binder_buffer *binder_alloc_new_buf(struct binder_alloc *alloc,
 					   size_t data_size,
 					   size_t offsets_size,
 					   size_t extra_buffers_size,
-					   int is_async)
+					   int is_async,
+					   int pid)
 {
 	struct binder_buffer *buffer;
 
 	mutex_lock(&alloc->mutex);
 	buffer = binder_alloc_new_buf_locked(alloc, data_size, offsets_size,
-					     extra_buffers_size, is_async);
+					     extra_buffers_size, is_async, pid);
 	mutex_unlock(&alloc->mutex);
 	return buffer;
 }