Skip to contentSkip to conversation

binder: frozen notification

PATCH Superseded Yu-Ting Tseng 4 participants 20 messages Jun 22, 2024 – Jul 4, 2024 1 message not in the archive

Subsystems (2)

v2 of 9 Superseded Latest is v9 →

lore
Share & tools
lore

A message that is not in the archive (sent off-list or not yet received): [email protected]

#1 lore raw

PATCH v2 1/1 binder: frozen notification

Yu-Ting Tseng (1):
binder: frozen notification

 drivers/android/binder.c            | 294 +++++++++++++++++++++++++++-
 drivers/android/binder_internal.h   |  19 +-
 include/uapi/linux/android/binder.h |  35 ++++
 3 files changed, 344 insertions(+), 4 deletions(-)
Quoted 1 line

nit: typo in "state"

Fixed

Quoted 1 line

It might be cleaner to factor out this copy_form_user() part

Done

Quoted 1 line

avoid WARNs

Done

Quoted 1 line

Shouldn't this return -ENOMEM or something

Done

Quoted 1 line

This "BC_REQUEST_FREEZE_NOTIFICATION" can be in the string directly.

Done

Quoted 2 lines
Quoted 1 line

+ struct binder_ref_freeze *freeze = NULL;

This doesn't need to be initialized.

Done

Quoted 1 line

This "BC_CLEAR_FREEZE_NOTIFICATION" can be in the string directly.

Done

Quoted 1 line

"freeze" has not been assigned here, right?

Fixed

Quoted 1 line

The preferred style for comments...

Fixed

Quoted 1 line

reversed christmas tree for variables

Done

Quoted 1 line

Why is entry on a new line

Done

Quoted 1 line

shorten the name of these functions

Done

Quoted 1 line

Do we need two different work types

Merged the two

Quoted 1 line

This should come after the variable declarations

Done

Quoted 1 line

write a short explanation of what this means to userspace instead

Done

Quoted 1 line

Use scripts/get_maintainer.pl

Done

Quoted 1 line

the corrisponding changes to userspace

Please see https://android-review.googlesource.com/c/platform/frameworks/native/+/3070045

Quoted 1 line

Multiple put_user() calls seems slow

Changed to copy_to_user

base-commit: 14d7c92f8df9c0964ae6f8b813c1b3ac38120825

Signature · 1 line
2.45.2.741.gdbec12cfda-goog
#2 lore raw

PATCH v2 1/1 binder: frozen notification

In reply to a message not in the archive

Frozen processes present a significant challenge in binder transactions. When a process is frozen, it cannot, by design, accept and/or respond to binder transactions. As a result, the sender needs to adjust its behavior, such as postponing transactions until the peer process unfreezes. However, there is currently no way to subscribe to these state change events, making it impossible to implement frozen-aware behaviors efficiently.

Introduce a binder API for subscribing to frozen state change events. This allows programs to react to changes in peer process state, mitigating issues related to binder transactions sent to frozen processes.

Implementation details:
For a given binder_ref, the state of frozen notification can be one of the followings:
1. Userspace doesn't want a notification. binder_ref->freeze is null.
2. Userspace wants a notification but none is in flight. list_empty(&binder_ref->freeze->work.entry) = true
3. A notification is in flight and waiting to be read by userspace. binder_ref_freeze.sent is false.
4. A notification was read by userspace and kernel is waiting for an ack. binder_ref_freeze.sent is true.

When a notification is in flight, new state change events are coalesced into the existing binder_ref_freeze struct. If userspace hasn't picked up the notification yet, the driver simply rewrites the state. Otherwise, the notification is flagged as requiring a resend, which will be performed once userspace acks the original notification that's inflight.

See https://android-review.googlesource.com/c/platform/frameworks/native/+/3070045
for how userspace is going to use this feature.

---
V1 -> V2: addressed review comments

 drivers/android/binder.c            | 294 +++++++++++++++++++++++++++-
 drivers/android/binder_internal.h   |  19 +-
 include/uapi/linux/android/binder.h |  35 ++++
 3 files changed, 344 insertions(+), 4 deletions(-)
drivers/android/binder.c +292 -2
index b21a7b246a0d..048f76e5dd35 100644
@@ -3763,6 +3763,168 @@ static void binder_transaction(struct binder_proc *proc,
 	}
 }
 
+static int
+binder_request_freeze_notification(struct binder_proc *proc,
+				   struct binder_thread *thread,
+				   struct binder_handle_cookie *handle_cookie)
+{
+	struct binder_ref_freeze *freeze;
+	struct binder_ref *ref;
+
+	/*
+	 * Allocate memory for freeze notification before taking lock.
+	 */
+	freeze = kzalloc(sizeof(*freeze), GFP_KERNEL);
+	if (!freeze) {
+		thread->return_error.cmd = BR_ERROR;
+		binder_enqueue_thread_work(thread, &thread->return_error.work);
+		binder_debug(BINDER_DEBUG_FAILED_TRANSACTION,
+			     "%d:%d BC_REQUEST_FREEZE_NOTIFICATION failed\n",
+			     proc->pid, thread->pid);
+		return -ENOMEM;
+	}
+	freeze->sent = false;
+	freeze->resend = false;
+
+	binder_proc_lock(proc);
+	ref = binder_get_ref_olocked(proc, handle_cookie->handle, false);
+	if (!ref) {
+		binder_user_error("%d:%d BC_REQUEST_FREEZE_NOTIFICATION invalid ref %d\n",
+				  proc->pid, thread->pid, handle_cookie->handle);
+		binder_proc_unlock(proc);
+		kfree(freeze);
+		return -EINVAL;
+	}
+
+	binder_node_lock(ref->node);
+
+	if (ref->freeze) {
+		binder_user_error("%d:%d BC_REQUEST_FREEZE_NOTIFICATION freeze notification already set\n",
+				  proc->pid, thread->pid);
+		binder_node_unlock(ref->node);
+		binder_proc_unlock(proc);
+		kfree(freeze);
+		return -EINVAL;
+	}
+	binder_stats_created(BINDER_STAT_FREEZE);
+	INIT_LIST_HEAD(&freeze->work.entry);
+	freeze->cookie = handle_cookie->cookie;
+	ref->freeze = freeze;
+	if (ref->node->proc) {
+		ref->freeze->work.type = BINDER_WORK_FROZEN_BINDER;
+		ref->freeze->is_frozen = ref->node->proc->is_frozen;
+		binder_inner_proc_lock(proc);
+		binder_enqueue_work_ilocked(&ref->freeze->work, &proc->todo);
+		binder_wakeup_proc_ilocked(proc);
+		binder_inner_proc_unlock(proc);
+	}
+	binder_node_unlock(ref->node);
+	binder_proc_unlock(proc);
+	return 0;
+}
+
+static int
+binder_clear_freeze_notification(struct binder_proc *proc,
+				 struct binder_thread *thread,
+				 struct binder_handle_cookie *handle_cookie)
+{
+	struct binder_ref_freeze *freeze;
+	struct binder_ref *ref;
+
+	binder_proc_lock(proc);
+	ref = binder_get_ref_olocked(proc, handle_cookie->handle, false);
+	if (!ref) {
+		binder_user_error("%d:%d BC_CLEAR_FREEZE_NOTIFICATION invalid ref %d\n",
+				  proc->pid, thread->pid, handle_cookie->handle);
+		binder_proc_unlock(proc);
+		return -EINVAL;
+	}
+
+	binder_node_lock(ref->node);
+
+	if (!ref->freeze) {
+		binder_user_error("%d:%d BC_CLEAR_FREEZE_NOTIFICATION freeze notification not active\n",
+				  proc->pid, thread->pid);
+		binder_node_unlock(ref->node);
+		binder_proc_unlock(proc);
+		return -EINVAL;
+	}
+	freeze = ref->freeze;
+	if (freeze->cookie != handle_cookie->cookie) {
+		binder_user_error("%d:%d BC_CLEAR_FREEZE_NOTIFICATION freeze notification cookie mismatch %016llx != %016llx\n",
+				  proc->pid, thread->pid, (u64)freeze->cookie,
+				  (u64)handle_cookie->cookie);
+		binder_node_unlock(ref->node);
+		binder_proc_unlock(proc);
+		return -EINVAL;
+	}
+	ref->freeze = NULL;
+	binder_inner_proc_lock(proc);
+	if (list_empty(&freeze->work.entry)) {
+		freeze->work.type = BINDER_WORK_CLEAR_FREEZE_NOTIFICATION;
+		if (thread->looper &
+			(BINDER_LOOPER_STATE_REGISTERED | BINDER_LOOPER_STATE_ENTERED)) {
+			binder_enqueue_thread_work_ilocked(thread, &freeze->work);
+		} else {
+			binder_enqueue_work_ilocked(&freeze->work, &proc->todo);
+			binder_wakeup_proc_ilocked(proc);
+		}
+	} else {
+		/*
+		 * There is already a freeze notification. Take it over and rewrite
+		 * the work type. If it was already sent, flag it for re-sending;
+		 * Otherwise it's pending and will be sent soon.
+		 */
+		freeze->work.type = BINDER_WORK_CLEAR_DEATH_NOTIFICATION;
+		if (freeze->sent)
+			freeze->resend = true;
+	}
+	binder_inner_proc_unlock(proc);
+	binder_node_unlock(ref->node);
+	binder_proc_unlock(proc);
+	return 0;
+}
+
+static int
+binder_freeze_notification_done(struct binder_proc *proc,
+				struct binder_thread *thread,
+				binder_uintptr_t cookie)
+{
+	struct binder_ref_freeze *freeze = NULL;
+	struct binder_work *w;
+
+	binder_inner_proc_lock(proc);
+	list_for_each_entry(w, &proc->delivered_freeze, entry) {
+		struct binder_ref_freeze *tmp_freeze =
+			container_of(w, struct binder_ref_freeze, work);
+
+		if (tmp_freeze->cookie == cookie) {
+			freeze = tmp_freeze;
+			break;
+		}
+	}
+	if (!freeze) {
+		binder_user_error("%d:%d BC_FREEZE_NOTIFICATION_DONE %016llx not found\n",
+				  proc->pid, thread->pid, (u64)cookie);
+		binder_inner_proc_unlock(proc);
+		return -EINVAL;
+	}
+	binder_dequeue_work_ilocked(&freeze->work);
+	freeze->sent = false;
+	if (freeze->resend) {
+		freeze->resend = false;
+		if (thread->looper &
+			(BINDER_LOOPER_STATE_REGISTERED | BINDER_LOOPER_STATE_ENTERED)) {
+			binder_enqueue_thread_work_ilocked(thread, &freeze->work);
+		} else {
+			binder_enqueue_work_ilocked(&freeze->work, &proc->todo);
+			binder_wakeup_proc_ilocked(proc);
+		}
+	}
+	binder_inner_proc_unlock(proc);
+	return 0;
+}
+
 /**
  * binder_free_buf() - free the specified buffer
  * @proc:	binder proc that owns buffer
@@ -4246,6 +4408,44 @@ static int binder_thread_write(struct binder_proc *proc,
 			binder_inner_proc_unlock(proc);
 		} break;
 
+		case BC_REQUEST_FREEZE_NOTIFICATION: {
+			struct binder_handle_cookie handle_cookie;
+			int error;
+
+			if (copy_from_user(&handle_cookie, ptr, sizeof(handle_cookie)))
+				return -EFAULT;
+			ptr += sizeof(handle_cookie);
+			error = binder_request_freeze_notification(proc, thread,
+								   &handle_cookie);
+			if (error)
+				return error;
+		} break;
+
+		case BC_CLEAR_FREEZE_NOTIFICATION: {
+			struct binder_handle_cookie handle_cookie;
+			int error;
+
+			if (copy_from_user(&handle_cookie, ptr, sizeof(handle_cookie)))
+				return -EFAULT;
+			ptr += sizeof(handle_cookie);
+			error = binder_clear_freeze_notification(proc, thread, &handle_cookie);
+			if (error)
+				return error;
+		} break;
+
+		case BC_FREEZE_NOTIFICATION_DONE: {
+			binder_uintptr_t cookie;
+			int error;
+
+			if (get_user(cookie, (binder_uintptr_t __user *)ptr))
+				return -EFAULT;
+
+			ptr += sizeof(cookie);
+			error = binder_freeze_notification_done(proc, thread, cookie);
+			if (error)
+				return error;
+		} break;
+
 		default:
 			pr_err("%d:%d unknown command %u\n",
 			       proc->pid, thread->pid, cmd);
@@ -4635,6 +4835,45 @@ static int binder_thread_read(struct binder_proc *proc,
 			if (cmd == BR_DEAD_BINDER)
 				goto done; /* DEAD_BINDER notifications can cause transactions */
 		} break;
+
+		case BINDER_WORK_FROZEN_BINDER: {
+			struct binder_ref_freeze *freeze;
+			struct binder_frozen_state_info info;
+
+			freeze = container_of(w, struct binder_ref_freeze, work);
+			info.is_frozen = freeze->is_frozen;
+			info.cookie = freeze->cookie;
+			freeze->sent = true;
+			binder_enqueue_work_ilocked(w, &proc->delivered_freeze);
+			binder_inner_proc_unlock(proc);
+
+			if (put_user(BR_FROZEN_BINDER, (uint32_t __user *)ptr))
+				return -EFAULT;
+			ptr += sizeof(uint32_t);
+			if (copy_to_user(ptr, &info, sizeof(info)))
+				return -EFAULT;
+			ptr += sizeof(info);
+			binder_stat_br(proc, thread, BR_FROZEN_BINDER);
+			goto done; /* BR_FROZEN_BINDER notifications can cause transactions */
+		} break;
+
+		case BINDER_WORK_CLEAR_FREEZE_NOTIFICATION: {
+			struct binder_ref_freeze *freeze =
+			    container_of(w, struct binder_ref_freeze, work);
+			binder_uintptr_t cookie = freeze->cookie;
+
+			binder_inner_proc_unlock(proc);
+			kfree(freeze);
+			binder_stats_deleted(BINDER_STAT_FREEZE);
+			if (put_user(BR_CLEAR_FREEZE_NOTIFICATION_DONE, (uint32_t __user *)ptr))
+				return -EFAULT;
+			ptr += sizeof(uint32_t);
+			if (put_user(cookie, (binder_uintptr_t __user *)ptr))
+				return -EFAULT;
+			ptr += sizeof(binder_uintptr_t);
+			binder_stat_br(proc, thread, BR_CLEAR_FREEZE_NOTIFICATION_DONE);
+		} break;
+
 		default:
 			binder_inner_proc_unlock(proc);
 			pr_err("%d:%d: bad work type %d\n",
@@ -5242,6 +5481,48 @@ static bool binder_txns_pending_ilocked(struct binder_proc *proc)
 	return false;
 }
 
+static void binder_add_freeze_work(struct binder_proc *proc, bool is_frozen)
+{
+	struct rb_node *n;
+	struct binder_ref *ref;
+
+	binder_inner_proc_lock(proc);
+	for (n = rb_first(&proc->nodes); n; n = rb_next(n)) {
+		struct binder_node *node;
+
+		node = rb_entry(n, struct binder_node, rb_node);
+		binder_inner_proc_unlock(proc);
+		binder_node_lock(node);
+		hlist_for_each_entry(ref, &node->refs, node_entry) {
+			/*
+			 * Need the node lock to synchronize
+			 * with new notification requests and the
+			 * inner lock to synchronize with queued
+			 * freeze notifications.
+			 */
+			binder_inner_proc_lock(ref->proc);
+			if (!ref->freeze) {
+				binder_inner_proc_unlock(ref->proc);
+				continue;
+			}
+			ref->freeze->work.type = BINDER_WORK_FROZEN_BINDER;
+			if (list_empty(&ref->freeze->work.entry)) {
+				ref->freeze->is_frozen = is_frozen;
+				binder_enqueue_work_ilocked(&ref->freeze->work, &ref->proc->todo);
+				binder_wakeup_proc_ilocked(ref->proc);
+			} else {
+				if (ref->freeze->sent && ref->freeze->is_frozen != is_frozen)
+					ref->freeze->resend = true;
+				ref->freeze->is_frozen = is_frozen;
+			}
+			binder_inner_proc_unlock(ref->proc);
+		}
+		binder_node_unlock(node);
+		binder_inner_proc_lock(proc);
+	}
+	binder_inner_proc_unlock(proc);
+}
+
 static int binder_ioctl_freeze(struct binder_freeze_info *info,
 			       struct binder_proc *target_proc)
 {
@@ -5253,6 +5534,7 @@ static int binder_ioctl_freeze(struct binder_freeze_info *info,
 		target_proc->async_recv = false;
 		target_proc->is_frozen = false;
 		binder_inner_proc_unlock(target_proc);
+		binder_add_freeze_work(target_proc, false);
 		return 0;
 	}
 
@@ -5266,6 +5548,7 @@ static int binder_ioctl_freeze(struct binder_freeze_info *info,
 	target_proc->async_recv = false;
 	target_proc->is_frozen = true;
 	binder_inner_proc_unlock(target_proc);
+	binder_add_freeze_work(target_proc, true);
 
 	if (info->timeout_ms > 0)
 		ret = wait_event_interruptible_timeout(
@@ -5658,6 +5941,7 @@ static int binder_open(struct inode *nodp, struct file *filp)
 	binder_stats_created(BINDER_STAT_PROC);
 	proc->pid = current->group_leader->pid;
 	INIT_LIST_HEAD(&proc->delivered_death);
+	INIT_LIST_HEAD(&proc->delivered_freeze);
 	INIT_LIST_HEAD(&proc->waiting_threads);
 	filp->private_data = proc;
 
@@ -6209,7 +6493,9 @@ static const char * const binder_return_strings[] = {
 	"BR_FAILED_REPLY",
 	"BR_FROZEN_REPLY",
 	"BR_ONEWAY_SPAM_SUSPECT",
-	"BR_TRANSACTION_PENDING_FROZEN"
+	"BR_TRANSACTION_PENDING_FROZEN",
+	"BR_FROZEN_BINDER",
+	"BR_CLEAR_FREEZE_NOTIFICATION_DONE",
 };
 
 static const char * const binder_command_strings[] = {
@@ -6232,6 +6518,9 @@ static const char * const binder_command_strings[] = {
 	"BC_DEAD_BINDER_DONE",
 	"BC_TRANSACTION_SG",
 	"BC_REPLY_SG",
+	"BC_REQUEST_FREEZE_NOTIFICATION",
+	"BC_CLEAR_FREEZE_NOTIFICATION",
+	"BC_FREEZE_NOTIFICATION_DONE",
 };
 
 static const char * const binder_objstat_strings[] = {
@@ -6241,7 +6530,8 @@ static const char * const binder_objstat_strings[] = {
 	"ref",
 	"death",
 	"transaction",
-	"transaction_complete"
+	"transaction_complete",
+	"freeze",
 };
 
 static void print_binder_stats(struct seq_file *m, const char *prefix,
drivers/android/binder_internal.h +17 -2
index 5b7c80b99ae8..21a4b3c48255 100644
@@ -129,12 +129,13 @@ enum binder_stat_types {
 	BINDER_STAT_DEATH,
 	BINDER_STAT_TRANSACTION,
 	BINDER_STAT_TRANSACTION_COMPLETE,
+	BINDER_STAT_FREEZE,
 	BINDER_STAT_COUNT
 };
 
 struct binder_stats {
-	atomic_t br[_IOC_NR(BR_TRANSACTION_PENDING_FROZEN) + 1];
-	atomic_t bc[_IOC_NR(BC_REPLY_SG) + 1];
+	atomic_t br[_IOC_NR(BR_CLEAR_FREEZE_NOTIFICATION_DONE) + 1];
+	atomic_t bc[_IOC_NR(BC_FREEZE_NOTIFICATION_DONE) + 1];
 	atomic_t obj_created[BINDER_STAT_COUNT];
 	atomic_t obj_deleted[BINDER_STAT_COUNT];
 };
@@ -159,6 +160,8 @@ struct binder_work {
 		BINDER_WORK_DEAD_BINDER,
 		BINDER_WORK_DEAD_BINDER_AND_CLEAR,
 		BINDER_WORK_CLEAR_DEATH_NOTIFICATION,
+		BINDER_WORK_FROZEN_BINDER,
+		BINDER_WORK_CLEAR_FREEZE_NOTIFICATION,
 	} type;
 };
 
@@ -275,6 +278,14 @@ struct binder_ref_death {
 	binder_uintptr_t cookie;
 };
 
+struct binder_ref_freeze {
+	struct binder_work work;
+	binder_uintptr_t cookie;
+	bool is_frozen;
+	bool sent;
+	bool resend;
+};
+
 /**
  * struct binder_ref_data - binder_ref counts and id
  * @debug_id:        unique ID for the ref
@@ -307,6 +318,7 @@ struct binder_ref_data {
  *               @node indicates the node must be freed
  * @death:       pointer to death notification (ref_death) if requested
  *               (protected by @node->lock)
+ * @freeze:      pointer to freeze notification (ref_freeze) if requested (protected by @node->lock)
  *
  * Structure to track references from procA to target node (on procB). This
  * structure is unsafe to access without holding @proc->outer_lock.
@@ -323,6 +335,7 @@ struct binder_ref {
 	struct binder_proc *proc;
 	struct binder_node *node;
 	struct binder_ref_death *death;
+	struct binder_ref_freeze *freeze;
 };
 
 /**
@@ -374,6 +387,7 @@ struct binder_ref {
  *                        (atomics, no lock needed)
  * @delivered_death:      list of delivered death notification
  *                        (protected by @inner_lock)
+ * @delivered_freeze:     list of delivered freeze notification (protected by @inner_lock)
  * @max_threads:          cap on number of binder threads
  *                        (protected by @inner_lock)
  * @requested_threads:    number of binder threads requested but not
@@ -421,6 +435,7 @@ struct binder_proc {
 	struct list_head todo;
 	struct binder_stats stats;
 	struct list_head delivered_death;
+	struct list_head delivered_freeze;
 	u32 max_threads;
 	int requested_threads;
 	int requested_threads_started;
include/uapi/linux/android/binder.h +35 -0
index d44a8118b2ed..a358cde61ee4 100644
@@ -236,6 +236,11 @@ struct binder_frozen_status_info {
 	__u32            async_recv;
 };
 
+struct binder_frozen_state_info {
+	binder_uintptr_t cookie;
+	__u32            is_frozen;
+};
+
 /* struct binder_extened_error - extended error information
  * @id:		identifier for the failed operation
  * @command:	command as defined by binder_driver_return_protocol
@@ -467,6 +472,17 @@ enum binder_driver_return_protocol {
 	/*
 	 * The target of the last async transaction is frozen.  No parameters.
 	 */
+
+	BR_FROZEN_BINDER = _IOR('r', 21, struct binder_frozen_state_info),
+	/*
+	 * The cookie and a boolean (is_frozen) that indicates whether the process
+	 * transitioned into a frozen or an unfrozen state.
+	 */
+
+	BR_CLEAR_FREEZE_NOTIFICATION_DONE = _IOR('r', 22, binder_uintptr_t),
+	/*
+	 * void *: cookie
+	 */
 };
 
 enum binder_driver_command_protocol {
@@ -550,6 +566,25 @@ enum binder_driver_command_protocol {
 	/*
 	 * binder_transaction_data_sg: the sent command.
 	 */
+
+	BC_REQUEST_FREEZE_NOTIFICATION =
+			_IOW('c', 19, struct binder_handle_cookie),
+	/*
+	 * int: handle
+	 * void *: cookie
+	 */
+
+	BC_CLEAR_FREEZE_NOTIFICATION = _IOW('c', 20,
+					    struct binder_handle_cookie),
+	/*
+	 * int: handle
+	 * void *: cookie
+	 */
+
+	BC_FREEZE_NOTIFICATION_DONE = _IOW('c', 21, binder_uintptr_t),
+	/*
+	 * void *: cookie
+	 */
 };
 
 #endif /* _UAPI_LINUX_BINDER_H */
Signature · 1 line
2.45.2.741.gdbec12cfda-goog
maintainer
#3 lore raw
Quoted 35 lines · Yu-Ting Tseng

Frozen processes present a significant challenge in binder transactions. When a process is frozen, it cannot, by design, accept and/or respond to binder transactions. As a result, the sender needs to adjust its behavior, such as postponing transactions until the peer process unfreezes. However, there is currently no way to subscribe to these state change events, making it impossible to implement frozen-aware behaviors efficiently.

Introduce a binder API for subscribing to frozen state change events. This allows programs to react to changes in peer process state, mitigating issues related to binder transactions sent to frozen processes.

Implementation details:
For a given binder_ref, the state of frozen notification can be one of the followings:
1. Userspace doesn't want a notification. binder_ref->freeze is null.
2. Userspace wants a notification but none is in flight. list_empty(&binder_ref->freeze->work.entry) = true
3. A notification is in flight and waiting to be read by userspace. binder_ref_freeze.sent is false.
4. A notification was read by userspace and kernel is waiting for an ack. binder_ref_freeze.sent is true.

When a notification is in flight, new state change events are coalesced into the existing binder_ref_freeze struct. If userspace hasn't picked up the notification yet, the driver simply rewrites the state. Otherwise, the notification is flagged as requiring a resend, which will be performed once userspace acks the original notification that's inflight.

See https://android-review.googlesource.com/c/platform/frameworks/native/+/3070045
for how userspace is going to use this feature.

[...]

drivers/android/binder.c static void binder_transaction(struct binder_proc *proc, · line 3878

Quoted 6 lines
drivers/android/binder.c +6 -0
+               /*
+                * There is already a freeze notification. Take it over and rewrite
+                * the work type. If it was already sent, flag it for re-sending;
+                * Otherwise it's pending and will be sent soon.
+                */
+               freeze->work.type = BINDER_WORK_CLEAR_DEATH_NOTIFICATION;

I have not done a comprehensive review yet, but this looks wrong.

Is there any chance that we could have a test in aosp that would have caught this?

Alice

#4 lore raw
Quoted 40 lines · Alice Ryhl
Quoted 35 lines · Yu-Ting Tseng

Frozen processes present a significant challenge in binder transactions. When a process is frozen, it cannot, by design, accept and/or respond to binder transactions. As a result, the sender needs to adjust its behavior, such as postponing transactions until the peer process unfreezes. However, there is currently no way to subscribe to these state change events, making it impossible to implement frozen-aware behaviors efficiently.

Introduce a binder API for subscribing to frozen state change events. This allows programs to react to changes in peer process state, mitigating issues related to binder transactions sent to frozen processes.

Implementation details:
For a given binder_ref, the state of frozen notification can be one of the followings:
1. Userspace doesn't want a notification. binder_ref->freeze is null.
2. Userspace wants a notification but none is in flight. list_empty(&binder_ref->freeze->work.entry) = true
3. A notification is in flight and waiting to be read by userspace. binder_ref_freeze.sent is false.
4. A notification was read by userspace and kernel is waiting for an ack. binder_ref_freeze.sent is true.

When a notification is in flight, new state change events are coalesced into the existing binder_ref_freeze struct. If userspace hasn't picked up the notification yet, the driver simply rewrites the state. Otherwise, the notification is flagged as requiring a resend, which will be performed once userspace acks the original notification that's inflight.

See https://android-review.googlesource.com/c/platform/frameworks/native/+/3070045
for how userspace is going to use this feature.

[...]

Quoted 8 lines

drivers/android/binder.c static void binder_transaction(struct binder_proc *proc, · line 3878

Quoted 6 lines
drivers/android/binder.c +6 -0
+               /*
+                * There is already a freeze notification. Take it over and rewrite
+                * the work type. If it was already sent, flag it for re-sending;
+                * Otherwise it's pending and will be sent soon.
+                */
+               freeze->work.type = BINDER_WORK_CLEAR_DEATH_NOTIFICATION;

I have not done a comprehensive review yet, but this looks wrong.

[resending as plain text]

Thanks for looking at the change!

The code here seems correct to me. Could you please elaborate why this looks wrong?

This part of code gets executed if freeze->work.entry is in a list. There are two possibilities:
1. The freeze work of type BINDER_WORK_FROZEN_BINDER is in the work queue, waiting to be picked up by binder_thread_read. Since it hasn't been picked up yet, this code rewrites the work type to BINDER_WORK_CLEAR_DEATH_NOTIFICATION, effectively canceling the state change notification and instead making binder_thread_read send a BR_CLEAR_FREEZE_NOTIFICATION_DONE to userspace. The API contract allows coalescing of events. I can explicitly mention this case if it helps.
2. The freeze work of type BINDER_WORK_FROZEN_BINDER is in the proc->delivered_freeze queue. This means a state change notification was just sent to userspace and the kernel is waiting for an ack. freeze->sent is true. In this case we set freeze->resend to true. Once the kernel receives the ack, it would queue up the work again, whose type was already set to BINDER_WORK_CLEAR_DEATH_NOTIFICATION.

Yu-Ting

Quoted 6 lines

Is there any chance that we could have a test in aosp that would have caught this?

Alice

maintainer
#5 lore raw
Quoted 73 lines · Yu-Ting Tseng
Quoted 40 lines · Alice Ryhl
Quoted 35 lines · Yu-Ting Tseng

Frozen processes present a significant challenge in binder transactions. When a process is frozen, it cannot, by design, accept and/or respond to binder transactions. As a result, the sender needs to adjust its behavior, such as postponing transactions until the peer process unfreezes. However, there is currently no way to subscribe to these state change events, making it impossible to implement frozen-aware behaviors efficiently.

Introduce a binder API for subscribing to frozen state change events. This allows programs to react to changes in peer process state, mitigating issues related to binder transactions sent to frozen processes.

Implementation details:
For a given binder_ref, the state of frozen notification can be one of the followings:
1. Userspace doesn't want a notification. binder_ref->freeze is null.
2. Userspace wants a notification but none is in flight. list_empty(&binder_ref->freeze->work.entry) = true
3. A notification is in flight and waiting to be read by userspace. binder_ref_freeze.sent is false.
4. A notification was read by userspace and kernel is waiting for an ack. binder_ref_freeze.sent is true.

When a notification is in flight, new state change events are coalesced into the existing binder_ref_freeze struct. If userspace hasn't picked up the notification yet, the driver simply rewrites the state. Otherwise, the notification is flagged as requiring a resend, which will be performed once userspace acks the original notification that's inflight.

See https://android-review.googlesource.com/c/platform/frameworks/native/+/3070045
for how userspace is going to use this feature.

[...]

Quoted 8 lines

drivers/android/binder.c static void binder_transaction(struct binder_proc *proc, · line 3878

Quoted 6 lines
drivers/android/binder.c +6 -0
+               /*
+                * There is already a freeze notification. Take it over and rewrite
+                * the work type. If it was already sent, flag it for re-sending;
+                * Otherwise it's pending and will be sent soon.
+                */
+               freeze->work.type = BINDER_WORK_CLEAR_DEATH_NOTIFICATION;

I have not done a comprehensive review yet, but this looks wrong.

[resending as plain text]

Thanks for looking at the change!

The code here seems correct to me. Could you please elaborate why this looks wrong?

This part of code gets executed if freeze->work.entry is in a list. There are two possibilities:
1. The freeze work of type BINDER_WORK_FROZEN_BINDER is in the work queue, waiting to be picked up by binder_thread_read. Since it hasn't been picked up yet, this code rewrites the work type to BINDER_WORK_CLEAR_DEATH_NOTIFICATION, effectively canceling the state change notification and instead making binder_thread_read send a BR_CLEAR_FREEZE_NOTIFICATION_DONE to userspace. The API contract allows coalescing of events. I can explicitly mention this case if it helps.
2. The freeze work of type BINDER_WORK_FROZEN_BINDER is in the proc->delivered_freeze queue. This means a state change notification was just sent to userspace and the kernel is waiting for an ack. freeze->sent is true. In this case we set freeze->resend to true. Once the kernel receives the ack, it would queue up the work again, whose type was already set to BINDER_WORK_CLEAR_DEATH_NOTIFICATION.

Quoted 1 line

Yu-Ting

Alice means you want to use BINDER_WORK_CLEAR_FREEZE_NOTIFICATION and not BINDER_WORK_CLEAR_DEATH_NOTIFICATION.

Quoted 7 lines
Quoted 7 lines

Is there any chance that we could have a test in aosp that would have caught this?

Alice

#6 lore raw
Quoted 75 lines · Carlos Llamas
Quoted 73 lines · Yu-Ting Tseng
Quoted 40 lines · Alice Ryhl
Quoted 35 lines · Yu-Ting Tseng

Frozen processes present a significant challenge in binder transactions. When a process is frozen, it cannot, by design, accept and/or respond to binder transactions. As a result, the sender needs to adjust its behavior, such as postponing transactions until the peer process unfreezes. However, there is currently no way to subscribe to these state change events, making it impossible to implement frozen-aware behaviors efficiently.

Introduce a binder API for subscribing to frozen state change events. This allows programs to react to changes in peer process state, mitigating issues related to binder transactions sent to frozen processes.

Implementation details:
For a given binder_ref, the state of frozen notification can be one of the followings:
1. Userspace doesn't want a notification. binder_ref->freeze is null.
2. Userspace wants a notification but none is in flight. list_empty(&binder_ref->freeze->work.entry) = true
3. A notification is in flight and waiting to be read by userspace. binder_ref_freeze.sent is false.
4. A notification was read by userspace and kernel is waiting for an ack. binder_ref_freeze.sent is true.

When a notification is in flight, new state change events are coalesced into the existing binder_ref_freeze struct. If userspace hasn't picked up the notification yet, the driver simply rewrites the state. Otherwise, the notification is flagged as requiring a resend, which will be performed once userspace acks the original notification that's inflight.

See https://android-review.googlesource.com/c/platform/frameworks/native/+/3070045
for how userspace is going to use this feature.

[...]

Quoted 8 lines

drivers/android/binder.c static void binder_transaction(struct binder_proc *proc, · line 3878

Quoted 6 lines
drivers/android/binder.c +6 -0
+               /*
+                * There is already a freeze notification. Take it over and rewrite
+                * the work type. If it was already sent, flag it for re-sending;
+                * Otherwise it's pending and will be sent soon.
+                */
+               freeze->work.type = BINDER_WORK_CLEAR_DEATH_NOTIFICATION;

I have not done a comprehensive review yet, but this looks wrong.

[resending as plain text]

Thanks for looking at the change!

The code here seems correct to me. Could you please elaborate why this looks wrong?

This part of code gets executed if freeze->work.entry is in a list. There are two possibilities:
1. The freeze work of type BINDER_WORK_FROZEN_BINDER is in the work queue, waiting to be picked up by binder_thread_read. Since it hasn't been picked up yet, this code rewrites the work type to BINDER_WORK_CLEAR_DEATH_NOTIFICATION, effectively canceling the state change notification and instead making binder_thread_read send a BR_CLEAR_FREEZE_NOTIFICATION_DONE to userspace. The API contract allows coalescing of events. I can explicitly mention this case if it helps.
2. The freeze work of type BINDER_WORK_FROZEN_BINDER is in the proc->delivered_freeze queue. This means a state change notification was just sent to userspace and the kernel is waiting for an ack. freeze->sent is true. In this case we set freeze->resend to true. Once the kernel receives the ack, it would queue up the work again, whose type was already set to BINDER_WORK_CLEAR_DEATH_NOTIFICATION.

Quoted 4 lines
Quoted 1 line

Yu-Ting

Alice means you want to use BINDER_WORK_CLEAR_FREEZE_NOTIFICATION and not BINDER_WORK_CLEAR_DEATH_NOTIFICATION.

Ah, I see. Thanks! Will get this corrected and also look into adding a test for it.

Quoted 8 lines
Quoted 8 lines
Quoted 8 lines

Is there any chance that we could have a test in aosp that would have caught this?

Alice

#7 lore raw

PATCH v3 1/1 binder: frozen notification

Yu-Ting Tseng (1):
binder: frozen notification

 drivers/android/binder.c            | 300 +++++++++++++++++++++++++++-
 drivers/android/binder_internal.h   |  23 ++-
 include/uapi/linux/android/binder.h |  35 ++++
 3 files changed, 354 insertions(+), 4 deletions(-)
Quoted 1 line

freeze was allocated with kzalloc(), you could drop the "= false".

Done.

Quoted 1 line

If !node->proc then process is dead. Do we really need to continue?

Update the code to return an error early if the process is already dead.

Quoted 1 line

This access to node->proc->* doesn't seem safe

Added locking.

Quoted 1 line

Why do we queue this notification?

Yes, this is to get the current state back to userspace. The userspace API delivers an initial event for the current state upon a listener registration, which makes it easier to track what the latest state is.

Quoted 1 line

I'm looking at the death notification code and it seems it only queues a

BR_ERROR after failing to allocate a "death" and that other errors are silently ignored?
Sure. Please let me know if you think we need a change here.

Quoted 1 line

these could be just bitfields.

Done

Quoted 1 line

freeze->work.type = BINDER_WORK_CLEAR_DEATH_NOTIFICATION

Fixed. Working on a userspace test. Will post a link when it's ready.

base-commit: 14d7c92f8df9c0964ae6f8b813c1b3ac38120825

Signature · 1 line
2.45.2.741.gdbec12cfda-goog
#8 lore raw
Quoted 26 lines · Yu-Ting Tseng

Yu-Ting Tseng (1):
binder: frozen notification

 drivers/android/binder.c            | 300 +++++++++++++++++++++++++++-
 drivers/android/binder_internal.h   |  23 ++-
 include/uapi/linux/android/binder.h |  35 ++++
 3 files changed, 354 insertions(+), 4 deletions(-)
Quoted 1 line

freeze was allocated with kzalloc(), you could drop the "= false".

Done.

Quoted 1 line

If !node->proc then process is dead. Do we really need to continue?

Update the code to return an error early if the process is already dead.

Quoted 1 line

This access to node->proc->* doesn't seem safe

Added locking.

Quoted 1 line

Why do we queue this notification?

Yes, this is to get the current state back to userspace. The userspace API delivers an initial event for the current state upon a listener registration, which makes it easier to track what the latest state is.

Quoted 1 line

I'm looking at the death notification code and it seems it only queues a

BR_ERROR after failing to allocate a "death" and that other errors are silently ignored?
Sure. Please let me know if you think we need a change here.

Quoted 5 lines
Quoted 1 line

these could be just bitfields.

Done

Quoted 1 line

freeze->work.type = BINDER_WORK_CLEAR_DEATH_NOTIFICATION

Fixed. Working on a userspace test. Will post a link when it's ready.

New test now included in the aosp patch:
https://android-review.googlesource.com/c/platform/frameworks/native/+/3070045/6/libs/binder/tests/binderDriverInterfaceTest.cpp

Quoted 5 lines

base-commit: 14d7c92f8df9c0964ae6f8b813c1b3ac38120825 --
2.45.2.741.gdbec12cfda-goog

#9 lore raw

PATCH v3 1/1 binder: frozen notification

In reply to Yu-Ting Tseng

Frozen processes present a significant challenge in binder transactions. When a process is frozen, it cannot, by design, accept and/or respond to binder transactions. As a result, the sender needs to adjust its behavior, such as postponing transactions until the peer process unfreezes. However, there is currently no way to subscribe to these state change events, making it impossible to implement frozen-aware behaviors efficiently.

Introduce a binder API for subscribing to frozen state change events. This allows programs to react to changes in peer process state, mitigating issues related to binder transactions sent to frozen processes.

Implementation details:
For a given binder_ref, the state of frozen notification can be one of the followings:
1. Userspace doesn't want a notification. binder_ref->freeze is null.
2. Userspace wants a notification but none is in flight. list_empty(&binder_ref->freeze->work.entry) = true
3. A notification is in flight and waiting to be read by userspace. binder_ref_freeze.sent is false.
4. A notification was read by userspace and kernel is waiting for an ack. binder_ref_freeze.sent is true.

When a notification is in flight, new state change events are coalesced into the existing binder_ref_freeze struct. If userspace hasn't picked up the notification yet, the driver simply rewrites the state. Otherwise, the notification is flagged as requiring a resend, which will be performed once userspace acks the original notification that's inflight.

See https://android-review.googlesource.com/c/platform/frameworks/native/+/3070045
for how userspace is going to use this feature.

---
 drivers/android/binder.c            | 300 +++++++++++++++++++++++++++-
 drivers/android/binder_internal.h   |  23 ++-
 include/uapi/linux/android/binder.h |  35 ++++
 3 files changed, 354 insertions(+), 4 deletions(-)
drivers/android/binder.c +298 -2
index b21a7b246a0d..d3dacfe0926d 100644
@@ -3763,6 +3763,174 @@ static void binder_transaction(struct binder_proc *proc,
 	}
 }
 
+static int
+binder_request_freeze_notification(struct binder_proc *proc,
+				   struct binder_thread *thread,
+				   struct binder_handle_cookie *handle_cookie)
+{
+	struct binder_ref_freeze *freeze;
+	struct binder_ref *ref;
+
+	/*
+	 * Allocate memory for freeze notification before taking lock.
+	 */
+	freeze = kzalloc(sizeof(*freeze), GFP_KERNEL);
+	if (!freeze) {
+		thread->return_error.cmd = BR_ERROR;
+		binder_enqueue_thread_work(thread, &thread->return_error.work);
+		binder_debug(BINDER_DEBUG_FAILED_TRANSACTION,
+			     "%d:%d BC_REQUEST_FREEZE_NOTIFICATION failed\n",
+			     proc->pid, thread->pid);
+		return -ENOMEM;
+	}
+
+	binder_proc_lock(proc);
+	ref = binder_get_ref_olocked(proc, handle_cookie->handle, false);
+	if (!ref) {
+		binder_user_error("%d:%d BC_REQUEST_FREEZE_NOTIFICATION invalid ref %d\n",
+				  proc->pid, thread->pid, handle_cookie->handle);
+		binder_proc_unlock(proc);
+		kfree(freeze);
+		return -EINVAL;
+	}
+
+	binder_node_lock(ref->node);
+
+	if (ref->freeze || !ref->node->proc) {
+		if (ref->freeze) {
+			binder_user_error("%d:%d BC_REQUEST_FREEZE_NOTIFICATION freeze notification already set\n",
+					  proc->pid, thread->pid);
+		} else {
+			binder_user_error("%d:%d BC_REQUEST_FREEZE_NOTIFICATION process is already dead\n",
+					  proc->pid, thread->pid);
+		}
+		binder_node_unlock(ref->node);
+		binder_proc_unlock(proc);
+		kfree(freeze);
+		return -EINVAL;
+	}
+	binder_stats_created(BINDER_STAT_FREEZE);
+	INIT_LIST_HEAD(&freeze->work.entry);
+	freeze->cookie = handle_cookie->cookie;
+	ref->freeze = freeze;
+	ref->freeze->work.type = BINDER_WORK_FROZEN_BINDER;
+
+	binder_inner_proc_lock(ref->node->proc);
+	ref->freeze->is_frozen = ref->node->proc->is_frozen;
+	binder_inner_proc_unlock(ref->node->proc);
+
+	binder_inner_proc_lock(proc);
+	binder_enqueue_work_ilocked(&ref->freeze->work, &proc->todo);
+	binder_wakeup_proc_ilocked(proc);
+	binder_inner_proc_unlock(proc);
+
+	binder_node_unlock(ref->node);
+	binder_proc_unlock(proc);
+	return 0;
+}
+
+static int
+binder_clear_freeze_notification(struct binder_proc *proc,
+				 struct binder_thread *thread,
+				 struct binder_handle_cookie *handle_cookie)
+{
+	struct binder_ref_freeze *freeze;
+	struct binder_ref *ref;
+
+	binder_proc_lock(proc);
+	ref = binder_get_ref_olocked(proc, handle_cookie->handle, false);
+	if (!ref) {
+		binder_user_error("%d:%d BC_CLEAR_FREEZE_NOTIFICATION invalid ref %d\n",
+				  proc->pid, thread->pid, handle_cookie->handle);
+		binder_proc_unlock(proc);
+		return -EINVAL;
+	}
+
+	binder_node_lock(ref->node);
+
+	if (!ref->freeze) {
+		binder_user_error("%d:%d BC_CLEAR_FREEZE_NOTIFICATION freeze notification not active\n",
+				  proc->pid, thread->pid);
+		binder_node_unlock(ref->node);
+		binder_proc_unlock(proc);
+		return -EINVAL;
+	}
+	freeze = ref->freeze;
+	if (freeze->cookie != handle_cookie->cookie) {
+		binder_user_error("%d:%d BC_CLEAR_FREEZE_NOTIFICATION freeze notification cookie mismatch %016llx != %016llx\n",
+				  proc->pid, thread->pid, (u64)freeze->cookie,
+				  (u64)handle_cookie->cookie);
+		binder_node_unlock(ref->node);
+		binder_proc_unlock(proc);
+		return -EINVAL;
+	}
+	ref->freeze = NULL;
+	binder_inner_proc_lock(proc);
+	if (list_empty(&freeze->work.entry)) {
+		freeze->work.type = BINDER_WORK_CLEAR_FREEZE_NOTIFICATION;
+		if (thread->looper &
+			(BINDER_LOOPER_STATE_REGISTERED | BINDER_LOOPER_STATE_ENTERED)) {
+			binder_enqueue_thread_work_ilocked(thread, &freeze->work);
+		} else {
+			binder_enqueue_work_ilocked(&freeze->work, &proc->todo);
+			binder_wakeup_proc_ilocked(proc);
+		}
+	} else {
+		/*
+		 * There is already a freeze notification. Take it over and rewrite
+		 * the work type. If it was already sent, flag it for re-sending;
+		 * Otherwise it's pending and will be sent soon.
+		 */
+		freeze->work.type = BINDER_WORK_CLEAR_FREEZE_NOTIFICATION;
+		if (freeze->sent)
+			freeze->resend = true;
+	}
+	binder_inner_proc_unlock(proc);
+	binder_node_unlock(ref->node);
+	binder_proc_unlock(proc);
+	return 0;
+}
+
+static int
+binder_freeze_notification_done(struct binder_proc *proc,
+				struct binder_thread *thread,
+				binder_uintptr_t cookie)
+{
+	struct binder_ref_freeze *freeze = NULL;
+	struct binder_work *w;
+
+	binder_inner_proc_lock(proc);
+	list_for_each_entry(w, &proc->delivered_freeze, entry) {
+		struct binder_ref_freeze *tmp_freeze =
+			container_of(w, struct binder_ref_freeze, work);
+
+		if (tmp_freeze->cookie == cookie) {
+			freeze = tmp_freeze;
+			break;
+		}
+	}
+	if (!freeze) {
+		binder_user_error("%d:%d BC_FREEZE_NOTIFICATION_DONE %016llx not found\n",
+				  proc->pid, thread->pid, (u64)cookie);
+		binder_inner_proc_unlock(proc);
+		return -EINVAL;
+	}
+	binder_dequeue_work_ilocked(&freeze->work);
+	freeze->sent = false;
+	if (freeze->resend) {
+		freeze->resend = false;
+		if (thread->looper &
+			(BINDER_LOOPER_STATE_REGISTERED | BINDER_LOOPER_STATE_ENTERED)) {
+			binder_enqueue_thread_work_ilocked(thread, &freeze->work);
+		} else {
+			binder_enqueue_work_ilocked(&freeze->work, &proc->todo);
+			binder_wakeup_proc_ilocked(proc);
+		}
+	}
+	binder_inner_proc_unlock(proc);
+	return 0;
+}
+
 /**
  * binder_free_buf() - free the specified buffer
  * @proc:	binder proc that owns buffer
@@ -4246,6 +4414,44 @@ static int binder_thread_write(struct binder_proc *proc,
 			binder_inner_proc_unlock(proc);
 		} break;
 
+		case BC_REQUEST_FREEZE_NOTIFICATION: {
+			struct binder_handle_cookie handle_cookie;
+			int error;
+
+			if (copy_from_user(&handle_cookie, ptr, sizeof(handle_cookie)))
+				return -EFAULT;
+			ptr += sizeof(handle_cookie);
+			error = binder_request_freeze_notification(proc, thread,
+								   &handle_cookie);
+			if (error)
+				return error;
+		} break;
+
+		case BC_CLEAR_FREEZE_NOTIFICATION: {
+			struct binder_handle_cookie handle_cookie;
+			int error;
+
+			if (copy_from_user(&handle_cookie, ptr, sizeof(handle_cookie)))
+				return -EFAULT;
+			ptr += sizeof(handle_cookie);
+			error = binder_clear_freeze_notification(proc, thread, &handle_cookie);
+			if (error)
+				return error;
+		} break;
+
+		case BC_FREEZE_NOTIFICATION_DONE: {
+			binder_uintptr_t cookie;
+			int error;
+
+			if (get_user(cookie, (binder_uintptr_t __user *)ptr))
+				return -EFAULT;
+
+			ptr += sizeof(cookie);
+			error = binder_freeze_notification_done(proc, thread, cookie);
+			if (error)
+				return error;
+		} break;
+
 		default:
 			pr_err("%d:%d unknown command %u\n",
 			       proc->pid, thread->pid, cmd);
@@ -4635,6 +4841,45 @@ static int binder_thread_read(struct binder_proc *proc,
 			if (cmd == BR_DEAD_BINDER)
 				goto done; /* DEAD_BINDER notifications can cause transactions */
 		} break;
+
+		case BINDER_WORK_FROZEN_BINDER: {
+			struct binder_ref_freeze *freeze;
+			struct binder_frozen_state_info info;
+
+			freeze = container_of(w, struct binder_ref_freeze, work);
+			info.is_frozen = freeze->is_frozen;
+			info.cookie = freeze->cookie;
+			freeze->sent = true;
+			binder_enqueue_work_ilocked(w, &proc->delivered_freeze);
+			binder_inner_proc_unlock(proc);
+
+			if (put_user(BR_FROZEN_BINDER, (uint32_t __user *)ptr))
+				return -EFAULT;
+			ptr += sizeof(uint32_t);
+			if (copy_to_user(ptr, &info, sizeof(info)))
+				return -EFAULT;
+			ptr += sizeof(info);
+			binder_stat_br(proc, thread, BR_FROZEN_BINDER);
+			goto done; /* BR_FROZEN_BINDER notifications can cause transactions */
+		} break;
+
+		case BINDER_WORK_CLEAR_FREEZE_NOTIFICATION: {
+			struct binder_ref_freeze *freeze =
+			    container_of(w, struct binder_ref_freeze, work);
+			binder_uintptr_t cookie = freeze->cookie;
+
+			binder_inner_proc_unlock(proc);
+			kfree(freeze);
+			binder_stats_deleted(BINDER_STAT_FREEZE);
+			if (put_user(BR_CLEAR_FREEZE_NOTIFICATION_DONE, (uint32_t __user *)ptr))
+				return -EFAULT;
+			ptr += sizeof(uint32_t);
+			if (put_user(cookie, (binder_uintptr_t __user *)ptr))
+				return -EFAULT;
+			ptr += sizeof(binder_uintptr_t);
+			binder_stat_br(proc, thread, BR_CLEAR_FREEZE_NOTIFICATION_DONE);
+		} break;
+
 		default:
 			binder_inner_proc_unlock(proc);
 			pr_err("%d:%d: bad work type %d\n",
@@ -5242,6 +5487,48 @@ static bool binder_txns_pending_ilocked(struct binder_proc *proc)
 	return false;
 }
 
+static void binder_add_freeze_work(struct binder_proc *proc, bool is_frozen)
+{
+	struct rb_node *n;
+	struct binder_ref *ref;
+
+	binder_inner_proc_lock(proc);
+	for (n = rb_first(&proc->nodes); n; n = rb_next(n)) {
+		struct binder_node *node;
+
+		node = rb_entry(n, struct binder_node, rb_node);
+		binder_inner_proc_unlock(proc);
+		binder_node_lock(node);
+		hlist_for_each_entry(ref, &node->refs, node_entry) {
+			/*
+			 * Need the node lock to synchronize
+			 * with new notification requests and the
+			 * inner lock to synchronize with queued
+			 * freeze notifications.
+			 */
+			binder_inner_proc_lock(ref->proc);
+			if (!ref->freeze) {
+				binder_inner_proc_unlock(ref->proc);
+				continue;
+			}
+			ref->freeze->work.type = BINDER_WORK_FROZEN_BINDER;
+			if (list_empty(&ref->freeze->work.entry)) {
+				ref->freeze->is_frozen = is_frozen;
+				binder_enqueue_work_ilocked(&ref->freeze->work, &ref->proc->todo);
+				binder_wakeup_proc_ilocked(ref->proc);
+			} else {
+				if (ref->freeze->sent && ref->freeze->is_frozen != is_frozen)
+					ref->freeze->resend = true;
+				ref->freeze->is_frozen = is_frozen;
+			}
+			binder_inner_proc_unlock(ref->proc);
+		}
+		binder_node_unlock(node);
+		binder_inner_proc_lock(proc);
+	}
+	binder_inner_proc_unlock(proc);
+}
+
 static int binder_ioctl_freeze(struct binder_freeze_info *info,
 			       struct binder_proc *target_proc)
 {
@@ -5253,6 +5540,7 @@ static int binder_ioctl_freeze(struct binder_freeze_info *info,
 		target_proc->async_recv = false;
 		target_proc->is_frozen = false;
 		binder_inner_proc_unlock(target_proc);
+		binder_add_freeze_work(target_proc, false);
 		return 0;
 	}
 
@@ -5266,6 +5554,7 @@ static int binder_ioctl_freeze(struct binder_freeze_info *info,
 	target_proc->async_recv = false;
 	target_proc->is_frozen = true;
 	binder_inner_proc_unlock(target_proc);
+	binder_add_freeze_work(target_proc, true);
 
 	if (info->timeout_ms > 0)
 		ret = wait_event_interruptible_timeout(
@@ -5658,6 +5947,7 @@ static int binder_open(struct inode *nodp, struct file *filp)
 	binder_stats_created(BINDER_STAT_PROC);
 	proc->pid = current->group_leader->pid;
 	INIT_LIST_HEAD(&proc->delivered_death);
+	INIT_LIST_HEAD(&proc->delivered_freeze);
 	INIT_LIST_HEAD(&proc->waiting_threads);
 	filp->private_data = proc;
 
@@ -6209,7 +6499,9 @@ static const char * const binder_return_strings[] = {
 	"BR_FAILED_REPLY",
 	"BR_FROZEN_REPLY",
 	"BR_ONEWAY_SPAM_SUSPECT",
-	"BR_TRANSACTION_PENDING_FROZEN"
+	"BR_TRANSACTION_PENDING_FROZEN",
+	"BR_FROZEN_BINDER",
+	"BR_CLEAR_FREEZE_NOTIFICATION_DONE",
 };
 
 static const char * const binder_command_strings[] = {
@@ -6232,6 +6524,9 @@ static const char * const binder_command_strings[] = {
 	"BC_DEAD_BINDER_DONE",
 	"BC_TRANSACTION_SG",
 	"BC_REPLY_SG",
+	"BC_REQUEST_FREEZE_NOTIFICATION",
+	"BC_CLEAR_FREEZE_NOTIFICATION",
+	"BC_FREEZE_NOTIFICATION_DONE",
 };
 
 static const char * const binder_objstat_strings[] = {
@@ -6241,7 +6536,8 @@ static const char * const binder_objstat_strings[] = {
 	"ref",
 	"death",
 	"transaction",
-	"transaction_complete"
+	"transaction_complete",
+	"freeze",
 };
 
 static void print_binder_stats(struct seq_file *m, const char *prefix,
drivers/android/binder_internal.h +21 -2
index 5b7c80b99ae8..2415fe53b3e9 100644
@@ -129,12 +129,13 @@ enum binder_stat_types {
 	BINDER_STAT_DEATH,
 	BINDER_STAT_TRANSACTION,
 	BINDER_STAT_TRANSACTION_COMPLETE,
+	BINDER_STAT_FREEZE,
 	BINDER_STAT_COUNT
 };
 
 struct binder_stats {
-	atomic_t br[_IOC_NR(BR_TRANSACTION_PENDING_FROZEN) + 1];
-	atomic_t bc[_IOC_NR(BC_REPLY_SG) + 1];
+	atomic_t br[_IOC_NR(BR_CLEAR_FREEZE_NOTIFICATION_DONE) + 1];
+	atomic_t bc[_IOC_NR(BC_FREEZE_NOTIFICATION_DONE) + 1];
 	atomic_t obj_created[BINDER_STAT_COUNT];
 	atomic_t obj_deleted[BINDER_STAT_COUNT];
 };
@@ -159,6 +160,8 @@ struct binder_work {
 		BINDER_WORK_DEAD_BINDER,
 		BINDER_WORK_DEAD_BINDER_AND_CLEAR,
 		BINDER_WORK_CLEAR_DEATH_NOTIFICATION,
+		BINDER_WORK_FROZEN_BINDER,
+		BINDER_WORK_CLEAR_FREEZE_NOTIFICATION,
 	} type;
 };
 
@@ -275,6 +278,16 @@ struct binder_ref_death {
 	binder_uintptr_t cookie;
 };
 
+struct binder_ref_freeze {
+	struct binder_work work;
+	binder_uintptr_t cookie;
+	bool is_frozen;
+	struct {
+		u8 sent:1;
+		u8 resend:1;
+	};
+};
+
 /**
  * struct binder_ref_data - binder_ref counts and id
  * @debug_id:        unique ID for the ref
@@ -307,6 +320,8 @@ struct binder_ref_data {
  *               @node indicates the node must be freed
  * @death:       pointer to death notification (ref_death) if requested
  *               (protected by @node->lock)
+ * @freeze:      pointer to freeze notification (ref_freeze) if requested
+ *               (protected by @node->lock)
  *
  * Structure to track references from procA to target node (on procB). This
  * structure is unsafe to access without holding @proc->outer_lock.
@@ -323,6 +338,7 @@ struct binder_ref {
 	struct binder_proc *proc;
 	struct binder_node *node;
 	struct binder_ref_death *death;
+	struct binder_ref_freeze *freeze;
 };
 
 /**
@@ -374,6 +390,8 @@ struct binder_ref {
  *                        (atomics, no lock needed)
  * @delivered_death:      list of delivered death notification
  *                        (protected by @inner_lock)
+ * @delivered_freeze:     list of delivered freeze notification
+ *                        (protected by @inner_lock)
  * @max_threads:          cap on number of binder threads
  *                        (protected by @inner_lock)
  * @requested_threads:    number of binder threads requested but not
@@ -421,6 +439,7 @@ struct binder_proc {
 	struct list_head todo;
 	struct binder_stats stats;
 	struct list_head delivered_death;
+	struct list_head delivered_freeze;
 	u32 max_threads;
 	int requested_threads;
 	int requested_threads_started;
include/uapi/linux/android/binder.h +35 -0
index d44a8118b2ed..a358cde61ee4 100644
@@ -236,6 +236,11 @@ struct binder_frozen_status_info {
 	__u32            async_recv;
 };
 
+struct binder_frozen_state_info {
+	binder_uintptr_t cookie;
+	__u32            is_frozen;
+};
+
 /* struct binder_extened_error - extended error information
  * @id:		identifier for the failed operation
  * @command:	command as defined by binder_driver_return_protocol
@@ -467,6 +472,17 @@ enum binder_driver_return_protocol {
 	/*
 	 * The target of the last async transaction is frozen.  No parameters.
 	 */
+
+	BR_FROZEN_BINDER = _IOR('r', 21, struct binder_frozen_state_info),
+	/*
+	 * The cookie and a boolean (is_frozen) that indicates whether the process
+	 * transitioned into a frozen or an unfrozen state.
+	 */
+
+	BR_CLEAR_FREEZE_NOTIFICATION_DONE = _IOR('r', 22, binder_uintptr_t),
+	/*
+	 * void *: cookie
+	 */
 };
 
 enum binder_driver_command_protocol {
@@ -550,6 +566,25 @@ enum binder_driver_command_protocol {
 	/*
 	 * binder_transaction_data_sg: the sent command.
 	 */
+
+	BC_REQUEST_FREEZE_NOTIFICATION =
+			_IOW('c', 19, struct binder_handle_cookie),
+	/*
+	 * int: handle
+	 * void *: cookie
+	 */
+
+	BC_CLEAR_FREEZE_NOTIFICATION = _IOW('c', 20,
+					    struct binder_handle_cookie),
+	/*
+	 * int: handle
+	 * void *: cookie
+	 */
+
+	BC_FREEZE_NOTIFICATION_DONE = _IOW('c', 21, binder_uintptr_t),
+	/*
+	 * void *: cookie
+	 */
 };
 
 #endif /* _UAPI_LINUX_BINDER_H */
Signature · 1 line
2.45.2.741.gdbec12cfda-goog
maintainer
#10 lore raw
Quoted 24 lines · Yu-Ting Tseng

Frozen processes present a significant challenge in binder transactions. When a process is frozen, it cannot, by design, accept and/or respond to binder transactions. As a result, the sender needs to adjust its behavior, such as postponing transactions until the peer process unfreezes. However, there is currently no way to subscribe to these state change events, making it impossible to implement frozen-aware behaviors efficiently.

Introduce a binder API for subscribing to frozen state change events. This allows programs to react to changes in peer process state, mitigating issues related to binder transactions sent to frozen processes.

Implementation details:
For a given binder_ref, the state of frozen notification can be one of the followings:
1. Userspace doesn't want a notification. binder_ref->freeze is null.
2. Userspace wants a notification but none is in flight. list_empty(&binder_ref->freeze->work.entry) = true
3. A notification is in flight and waiting to be read by userspace. binder_ref_freeze.sent is false.
4. A notification was read by userspace and kernel is waiting for an ack. binder_ref_freeze.sent is true.

Quoted 7 lines

When a notification is in flight, new state change events are coalesced into the existing binder_ref_freeze struct. If userspace hasn't picked up the notification yet, the driver simply rewrites the state. Otherwise, the notification is flagged as requiring a resend, which will be performed once userspace acks the original notification that's inflight.

See https://android-review.googlesource.com/c/platform/frameworks/native/+/3070045

I believe the external link would be https://r.android.com/3070045

Quoted 27 lines

for how userspace is going to use this feature.

---
 drivers/android/binder.c            | 300 +++++++++++++++++++++++++++-
 drivers/android/binder_internal.h   |  23 ++-
 include/uapi/linux/android/binder.h |  35 ++++
 3 files changed, 354 insertions(+), 4 deletions(-)
drivers/android/binder.c +10 -0
index b21a7b246a0d..d3dacfe0926d 100644
@@ -3763,6 +3763,174 @@ static void binder_transaction(struct binder_proc *proc,
 	}
 }
 
+static int
+binder_request_freeze_notification(struct binder_proc *proc,
+				   struct binder_thread *thread,
+				   struct binder_handle_cookie *handle_cookie)
+{
+	struct binder_ref_freeze *freeze;
+	struct binder_ref *ref;
+
+	/*
+	 * Allocate memory for freeze notification before taking lock.

drivers/android/binder.c static void binder_transaction(struct binder_proc *proc, · line 3785

Quoted 10 lines
drivers/android/binder.c +10 -0
+	 */
+	freeze = kzalloc(sizeof(*freeze), GFP_KERNEL);
+	if (!freeze) {
+		thread->return_error.cmd = BR_ERROR;
+		binder_enqueue_thread_work(thread, &thread->return_error.work);
+		binder_debug(BINDER_DEBUG_FAILED_TRANSACTION,
+			     "%d:%d BC_REQUEST_FREEZE_NOTIFICATION failed\n",
+			     proc->pid, thread->pid);
+		return -ENOMEM;
+	}

A couple notes on this allocation:
(1) drop the BR_ERROR, at least until we define a global policy for propagating errors. For now the "-ENOMEM" will be enough.
(2) typically OOM error logs are discouraged, the infra already dumps necessary info. Please remove the debug log.
(3) I don't think the comment is necessary.

This should all just be:

	freeze = kzalloc(sizeof(*freeze), GFP_KERNEL);
	if (!freeze)
		return -ENOMEM;
Quoted 21 lines
drivers/android/binder.c +21 -0
+
+	binder_proc_lock(proc);
+	ref = binder_get_ref_olocked(proc, handle_cookie->handle, false);
+	if (!ref) {
+		binder_user_error("%d:%d BC_REQUEST_FREEZE_NOTIFICATION invalid ref %d\n",
+				  proc->pid, thread->pid, handle_cookie->handle);
+		binder_proc_unlock(proc);
+		kfree(freeze);
+		return -EINVAL;
+	}
+
+	binder_node_lock(ref->node);
+
+	if (ref->freeze || !ref->node->proc) {
+		if (ref->freeze) {
+			binder_user_error("%d:%d BC_REQUEST_FREEZE_NOTIFICATION freeze notification already set\n",
+					  proc->pid, thread->pid);
+		} else {
+			binder_user_error("%d:%d BC_REQUEST_FREEZE_NOTIFICATION process is already dead\n",
+					  proc->pid, thread->pid);
+		}

drivers/android/binder.c static void binder_transaction(struct binder_proc *proc, · line 3816

Quoted 10 lines
drivers/android/binder.c +10 -0
+		binder_node_unlock(ref->node);
+		binder_proc_unlock(proc);
+		kfree(freeze);
+		return -EINVAL;
+	}
+	binder_stats_created(BINDER_STAT_FREEZE);
+	INIT_LIST_HEAD(&freeze->work.entry);
+	freeze->cookie = handle_cookie->cookie;
+	ref->freeze = freeze;
+	ref->freeze->work.type = BINDER_WORK_FROZEN_BINDER;

Are we ok modifying the ref->freeze->* space here without the inner_lock? I believe this is OK as we can't race with the other operations?

Quoted 47 lines
drivers/android/binder.c +47 -0
+
+	binder_inner_proc_lock(ref->node->proc);
+	ref->freeze->is_frozen = ref->node->proc->is_frozen;
+	binder_inner_proc_unlock(ref->node->proc);
+
+	binder_inner_proc_lock(proc);
+	binder_enqueue_work_ilocked(&ref->freeze->work, &proc->todo);
+	binder_wakeup_proc_ilocked(proc);
+	binder_inner_proc_unlock(proc);
+
+	binder_node_unlock(ref->node);
+	binder_proc_unlock(proc);
+	return 0;
+}
+
+static int
+binder_clear_freeze_notification(struct binder_proc *proc,
+				 struct binder_thread *thread,
+				 struct binder_handle_cookie *handle_cookie)
+{
+	struct binder_ref_freeze *freeze;
+	struct binder_ref *ref;
+
+	binder_proc_lock(proc);
+	ref = binder_get_ref_olocked(proc, handle_cookie->handle, false);
+	if (!ref) {
+		binder_user_error("%d:%d BC_CLEAR_FREEZE_NOTIFICATION invalid ref %d\n",
+				  proc->pid, thread->pid, handle_cookie->handle);
+		binder_proc_unlock(proc);
+		return -EINVAL;
+	}
+
+	binder_node_lock(ref->node);
+
+	if (!ref->freeze) {
+		binder_user_error("%d:%d BC_CLEAR_FREEZE_NOTIFICATION freeze notification not active\n",
+				  proc->pid, thread->pid);
+		binder_node_unlock(ref->node);
+		binder_proc_unlock(proc);
+		return -EINVAL;
+	}
+	freeze = ref->freeze;
+	if (freeze->cookie != handle_cookie->cookie) {
+		binder_user_error("%d:%d BC_CLEAR_FREEZE_NOTIFICATION freeze notification cookie mismatch %016llx != %016llx\n",
+				  proc->pid, thread->pid, (u64)freeze->cookie,
+				  (u64)handle_cookie->cookie);
+		binder_node_unlock(ref->node);

drivers/android/binder.c static void binder_transaction(struct binder_proc *proc, · line 3873

Quoted 10 lines
drivers/android/binder.c +10 -0
+		binder_proc_unlock(proc);
+		return -EINVAL;
+	}
+	ref->freeze = NULL;
+	binder_inner_proc_lock(proc);
+	if (list_empty(&freeze->work.entry)) {
+		freeze->work.type = BINDER_WORK_CLEAR_FREEZE_NOTIFICATION;
+		if (thread->looper &
+			(BINDER_LOOPER_STATE_REGISTERED | BINDER_LOOPER_STATE_ENTERED)) {
+			binder_enqueue_thread_work_ilocked(thread, &freeze->work);

I'm not entirely sure why we attempt to queue this as thread work. It seems more complicated that it needs to be. Do you know why not handle this as proc work every time instead?

drivers/android/binder.c static void binder_transaction(struct binder_proc *proc, · line 3884

Quoted 11 lines
drivers/android/binder.c +11 -0
+		} else {
+			binder_enqueue_work_ilocked(&freeze->work, &proc->todo);
+			binder_wakeup_proc_ilocked(proc);
+		}
+	} else {
+		/*
+		 * There is already a freeze notification. Take it over and rewrite
+		 * the work type. If it was already sent, flag it for re-sending;
+		 * Otherwise it's pending and will be sent soon.
+		 */
+		freeze->work.type = BINDER_WORK_CLEAR_FREEZE_NOTIFICATION;

We set the work type to CLEAR regardless, might as well factor it out.

Quoted 34 lines
drivers/android/binder.c +34 -0
+		if (freeze->sent)
+			freeze->resend = true;
+	}
+	binder_inner_proc_unlock(proc);
+	binder_node_unlock(ref->node);
+	binder_proc_unlock(proc);
+	return 0;
+}
+
+static int
+binder_freeze_notification_done(struct binder_proc *proc,
+				struct binder_thread *thread,
+				binder_uintptr_t cookie)
+{
+	struct binder_ref_freeze *freeze = NULL;
+	struct binder_work *w;
+
+	binder_inner_proc_lock(proc);
+	list_for_each_entry(w, &proc->delivered_freeze, entry) {
+		struct binder_ref_freeze *tmp_freeze =
+			container_of(w, struct binder_ref_freeze, work);
+
+		if (tmp_freeze->cookie == cookie) {
+			freeze = tmp_freeze;
+			break;
+		}
+	}
+	if (!freeze) {
+		binder_user_error("%d:%d BC_FREEZE_NOTIFICATION_DONE %016llx not found\n",
+				  proc->pid, thread->pid, (u64)cookie);
+		binder_inner_proc_unlock(proc);
+		return -EINVAL;
+	}
+	binder_dequeue_work_ilocked(&freeze->work);

drivers/android/binder.c static void binder_transaction(struct binder_proc *proc, · line 3928

Quoted 10 lines
drivers/android/binder.c +10 -0
+	freeze->sent = false;
+	if (freeze->resend) {
+		freeze->resend = false;
+		if (thread->looper &
+			(BINDER_LOOPER_STATE_REGISTERED | BINDER_LOOPER_STATE_ENTERED)) {
+			binder_enqueue_thread_work_ilocked(thread, &freeze->work);
+		} else {
+			binder_enqueue_work_ilocked(&freeze->work, &proc->todo);
+			binder_wakeup_proc_ilocked(proc);
+		}

Yeah I don't know why we go through the trouble of trying to handle this as thread work.

Quoted 56 lines
drivers/android/binder.c +43 -0
+	}
+	binder_inner_proc_unlock(proc);
+	return 0;
+}
+
 /**
  * binder_free_buf() - free the specified buffer
  * @proc:	binder proc that owns buffer
@@ -4246,6 +4414,44 @@ static int binder_thread_write(struct binder_proc *proc,
 			binder_inner_proc_unlock(proc);
 		} break;
 
+		case BC_REQUEST_FREEZE_NOTIFICATION: {
+			struct binder_handle_cookie handle_cookie;
+			int error;
+
+			if (copy_from_user(&handle_cookie, ptr, sizeof(handle_cookie)))
+				return -EFAULT;
+			ptr += sizeof(handle_cookie);
+			error = binder_request_freeze_notification(proc, thread,
+								   &handle_cookie);
+			if (error)
+				return error;
+		} break;
+
+		case BC_CLEAR_FREEZE_NOTIFICATION: {
+			struct binder_handle_cookie handle_cookie;
+			int error;
+
+			if (copy_from_user(&handle_cookie, ptr, sizeof(handle_cookie)))
+				return -EFAULT;
+			ptr += sizeof(handle_cookie);
+			error = binder_clear_freeze_notification(proc, thread, &handle_cookie);
+			if (error)
+				return error;
+		} break;
+
+		case BC_FREEZE_NOTIFICATION_DONE: {
+			binder_uintptr_t cookie;
+			int error;
+
+			if (get_user(cookie, (binder_uintptr_t __user *)ptr))
+				return -EFAULT;
+
+			ptr += sizeof(cookie);
+			error = binder_freeze_notification_done(proc, thread, cookie);
+			if (error)
+				return error;
+		} break;
+
 		default:
 			pr_err("%d:%d unknown command %u\n",
 			       proc->pid, thread->pid, cmd);
@@ -4635,6 +4841,45 @@ static int binder_thread_read(struct binder_proc *proc,
 			if (cmd == BR_DEAD_BINDER)
 				goto done; /* DEAD_BINDER notifications can cause transactions */

drivers/android/binder.c static int binder_thread_read(struct binder_proc *proc, · line 4852

Quoted 10 lines
drivers/android/binder.c +9 -0
 		} break;
+
+		case BINDER_WORK_FROZEN_BINDER: {
+			struct binder_ref_freeze *freeze;
+			struct binder_frozen_state_info info;
+
+			freeze = container_of(w, struct binder_ref_freeze, work);
+			info.is_frozen = freeze->is_frozen;
+			info.cookie = freeze->cookie;
+			freeze->sent = true;

It's a bit confusing to me the locking here. It seems the freeze->* space is protected by the proc->inner_lock right? We read/write several elements here with only that lock held.

drivers/android/binder.c static int binder_thread_read(struct binder_proc *proc, · line 4860

Quoted 8 lines
drivers/android/binder.c +8 -0
+			binder_enqueue_work_ilocked(w, &proc->delivered_freeze);
+			binder_inner_proc_unlock(proc);
+
+			if (put_user(BR_FROZEN_BINDER, (uint32_t __user *)ptr))
+				return -EFAULT;
+			ptr += sizeof(uint32_t);
+			if (copy_to_user(ptr, &info, sizeof(info)))
+				return -EFAULT;

I believe you should zero the 'info' before copy_to_user(). This might leak kernel's stack memory to userspace if there is any padding.

Quoted 32 lines
drivers/android/binder.c +25 -0
+			ptr += sizeof(info);
+			binder_stat_br(proc, thread, BR_FROZEN_BINDER);
+			goto done; /* BR_FROZEN_BINDER notifications can cause transactions */
+		} break;
+
+		case BINDER_WORK_CLEAR_FREEZE_NOTIFICATION: {
+			struct binder_ref_freeze *freeze =
+			    container_of(w, struct binder_ref_freeze, work);
+			binder_uintptr_t cookie = freeze->cookie;
+
+			binder_inner_proc_unlock(proc);
+			kfree(freeze);
+			binder_stats_deleted(BINDER_STAT_FREEZE);
+			if (put_user(BR_CLEAR_FREEZE_NOTIFICATION_DONE, (uint32_t __user *)ptr))
+				return -EFAULT;
+			ptr += sizeof(uint32_t);
+			if (put_user(cookie, (binder_uintptr_t __user *)ptr))
+				return -EFAULT;
+			ptr += sizeof(binder_uintptr_t);
+			binder_stat_br(proc, thread, BR_CLEAR_FREEZE_NOTIFICATION_DONE);
+		} break;
+
 		default:
 			binder_inner_proc_unlock(proc);
 			pr_err("%d:%d: bad work type %d\n",
@@ -5242,6 +5487,48 @@ static bool binder_txns_pending_ilocked(struct binder_proc *proc)
 	return false;
 }
 
+static void binder_add_freeze_work(struct binder_proc *proc, bool is_frozen)
+{
+	struct rb_node *n;

drivers/android/binder.c static bool binder_txns_pending_ilocked(struct binder_proc *proc) · line 5502

Quoted 10 lines
drivers/android/binder.c +10 -0
+	struct binder_ref *ref;
+
+	binder_inner_proc_lock(proc);
+	for (n = rb_first(&proc->nodes); n; n = rb_next(n)) {
+		struct binder_node *node;
+
+		node = rb_entry(n, struct binder_node, rb_node);
+		binder_inner_proc_unlock(proc);
+		binder_node_lock(node);
+		hlist_for_each_entry(ref, &node->refs, node_entry) {

So we traverse every single reference of every single node in this proc looking for references subscribed to freeze, correct? I guess we don't have the structures to provide something faster. Speed isn't critical either here.

Quoted 41 lines
drivers/android/binder.c +30 -0
+			/*
+			 * Need the node lock to synchronize
+			 * with new notification requests and the
+			 * inner lock to synchronize with queued
+			 * freeze notifications.
+			 */
+			binder_inner_proc_lock(ref->proc);
+			if (!ref->freeze) {
+				binder_inner_proc_unlock(ref->proc);
+				continue;
+			}
+			ref->freeze->work.type = BINDER_WORK_FROZEN_BINDER;
+			if (list_empty(&ref->freeze->work.entry)) {
+				ref->freeze->is_frozen = is_frozen;
+				binder_enqueue_work_ilocked(&ref->freeze->work, &ref->proc->todo);
+				binder_wakeup_proc_ilocked(ref->proc);
+			} else {
+				if (ref->freeze->sent && ref->freeze->is_frozen != is_frozen)
+					ref->freeze->resend = true;
+				ref->freeze->is_frozen = is_frozen;
+			}
+			binder_inner_proc_unlock(ref->proc);
+		}
+		binder_node_unlock(node);
+		binder_inner_proc_lock(proc);
+	}
+	binder_inner_proc_unlock(proc);
+}
+
 static int binder_ioctl_freeze(struct binder_freeze_info *info,
 			       struct binder_proc *target_proc)
 {
@@ -5253,6 +5540,7 @@ static int binder_ioctl_freeze(struct binder_freeze_info *info,
 		target_proc->async_recv = false;
 		target_proc->is_frozen = false;
 		binder_inner_proc_unlock(target_proc);
+		binder_add_freeze_work(target_proc, false);
 		return 0;
 	}
 

drivers/android/binder.c static int binder_ioctl_freeze(struct binder_freeze_info *info, · line 5557

Quoted 4 lines
drivers/android/binder.c +1 -0
@@ -5266,6 +5554,7 @@ static int binder_ioctl_freeze(struct binder_freeze_info *info,
 	target_proc->async_recv = false;
 	target_proc->is_frozen = true;
 	binder_inner_proc_unlock(target_proc);
+	binder_add_freeze_work(target_proc, true);

What if there a EAGAIN error following this below? I think you want to update the freeze work about the 'proc->is_frozen = false' change.

Quoted 72 lines
drivers/android/binder.c +9 -2
 	if (info->timeout_ms > 0)
 		ret = wait_event_interruptible_timeout(
@@ -5658,6 +5947,7 @@ static int binder_open(struct inode *nodp, struct file *filp)
 	binder_stats_created(BINDER_STAT_PROC);
 	proc->pid = current->group_leader->pid;
 	INIT_LIST_HEAD(&proc->delivered_death);
+	INIT_LIST_HEAD(&proc->delivered_freeze);
 	INIT_LIST_HEAD(&proc->waiting_threads);
 	filp->private_data = proc;
 
@@ -6209,7 +6499,9 @@ static const char * const binder_return_strings[] = {
 	"BR_FAILED_REPLY",
 	"BR_FROZEN_REPLY",
 	"BR_ONEWAY_SPAM_SUSPECT",
-	"BR_TRANSACTION_PENDING_FROZEN"
+	"BR_TRANSACTION_PENDING_FROZEN",
+	"BR_FROZEN_BINDER",
+	"BR_CLEAR_FREEZE_NOTIFICATION_DONE",
 };
 
 static const char * const binder_command_strings[] = {
@@ -6232,6 +6524,9 @@ static const char * const binder_command_strings[] = {
 	"BC_DEAD_BINDER_DONE",
 	"BC_TRANSACTION_SG",
 	"BC_REPLY_SG",
+	"BC_REQUEST_FREEZE_NOTIFICATION",
+	"BC_CLEAR_FREEZE_NOTIFICATION",
+	"BC_FREEZE_NOTIFICATION_DONE",
 };
 
 static const char * const binder_objstat_strings[] = {
@@ -6241,7 +6536,8 @@ static const char * const binder_objstat_strings[] = {
 	"ref",
 	"death",
 	"transaction",
-	"transaction_complete"
+	"transaction_complete",
+	"freeze",
 };
 
 static void print_binder_stats(struct seq_file *m, const char *prefix,
drivers/android/binder_internal.h +5 -2
index 5b7c80b99ae8..2415fe53b3e9 100644
@@ -129,12 +129,13 @@ enum binder_stat_types {
 	BINDER_STAT_DEATH,
 	BINDER_STAT_TRANSACTION,
 	BINDER_STAT_TRANSACTION_COMPLETE,
+	BINDER_STAT_FREEZE,
 	BINDER_STAT_COUNT
 };
 
 struct binder_stats {
-	atomic_t br[_IOC_NR(BR_TRANSACTION_PENDING_FROZEN) + 1];
-	atomic_t bc[_IOC_NR(BC_REPLY_SG) + 1];
+	atomic_t br[_IOC_NR(BR_CLEAR_FREEZE_NOTIFICATION_DONE) + 1];
+	atomic_t bc[_IOC_NR(BC_FREEZE_NOTIFICATION_DONE) + 1];
 	atomic_t obj_created[BINDER_STAT_COUNT];
 	atomic_t obj_deleted[BINDER_STAT_COUNT];
 };
@@ -159,6 +160,8 @@ struct binder_work {
 		BINDER_WORK_DEAD_BINDER,
 		BINDER_WORK_DEAD_BINDER_AND_CLEAR,
 		BINDER_WORK_CLEAR_DEATH_NOTIFICATION,
+		BINDER_WORK_FROZEN_BINDER,
+		BINDER_WORK_CLEAR_FREEZE_NOTIFICATION,
 	} type;
 };
 

drivers/android/binder_internal.h struct binder_ref_death { · line 284

Quoted 7 lines
drivers/android/binder_internal.h +4 -0
@@ -275,6 +278,16 @@ struct binder_ref_death {
 	binder_uintptr_t cookie;
 };
 
+struct binder_ref_freeze {
+	struct binder_work work;
+	binder_uintptr_t cookie;
+	bool is_frozen;

any reason why is_frozen didn't make it to the bitfields?

drivers/android/binder_internal.h struct binder_ref_death { · line 285

Quoted 1 line
drivers/android/binder_internal.h +1 -0
+	struct {

No need for a struct

drivers/android/binder_internal.h struct binder_ref_death { · line 287

Quoted 2 lines
drivers/android/binder_internal.h +2 -0
+		u8 sent:1;
+		u8 resend:1;

You could still keep the original bool for bitfields too if you want.

Quoted 103 lines
drivers/android/binder_internal.h +9 -0
+	};
+};
+
 /**
  * struct binder_ref_data - binder_ref counts and id
  * @debug_id:        unique ID for the ref
@@ -307,6 +320,8 @@ struct binder_ref_data {
  *               @node indicates the node must be freed
  * @death:       pointer to death notification (ref_death) if requested
  *               (protected by @node->lock)
+ * @freeze:      pointer to freeze notification (ref_freeze) if requested
+ *               (protected by @node->lock)
  *
  * Structure to track references from procA to target node (on procB). This
  * structure is unsafe to access without holding @proc->outer_lock.
@@ -323,6 +338,7 @@ struct binder_ref {
 	struct binder_proc *proc;
 	struct binder_node *node;
 	struct binder_ref_death *death;
+	struct binder_ref_freeze *freeze;
 };
 
 /**
@@ -374,6 +390,8 @@ struct binder_ref {
  *                        (atomics, no lock needed)
  * @delivered_death:      list of delivered death notification
  *                        (protected by @inner_lock)
+ * @delivered_freeze:     list of delivered freeze notification
+ *                        (protected by @inner_lock)
  * @max_threads:          cap on number of binder threads
  *                        (protected by @inner_lock)
  * @requested_threads:    number of binder threads requested but not
@@ -421,6 +439,7 @@ struct binder_proc {
 	struct list_head todo;
 	struct binder_stats stats;
 	struct list_head delivered_death;
+	struct list_head delivered_freeze;
 	u32 max_threads;
 	int requested_threads;
 	int requested_threads_started;
include/uapi/linux/android/binder.h +35 -0
index d44a8118b2ed..a358cde61ee4 100644
@@ -236,6 +236,11 @@ struct binder_frozen_status_info {
 	__u32            async_recv;
 };
 
+struct binder_frozen_state_info {
+	binder_uintptr_t cookie;
+	__u32            is_frozen;
+};
+
 /* struct binder_extened_error - extended error information
  * @id:		identifier for the failed operation
  * @command:	command as defined by binder_driver_return_protocol
@@ -467,6 +472,17 @@ enum binder_driver_return_protocol {
 	/*
 	 * The target of the last async transaction is frozen.  No parameters.
 	 */
+
+	BR_FROZEN_BINDER = _IOR('r', 21, struct binder_frozen_state_info),
+	/*
+	 * The cookie and a boolean (is_frozen) that indicates whether the process
+	 * transitioned into a frozen or an unfrozen state.
+	 */
+
+	BR_CLEAR_FREEZE_NOTIFICATION_DONE = _IOR('r', 22, binder_uintptr_t),
+	/*
+	 * void *: cookie
+	 */
 };
 
 enum binder_driver_command_protocol {
@@ -550,6 +566,25 @@ enum binder_driver_command_protocol {
 	/*
 	 * binder_transaction_data_sg: the sent command.
 	 */
+
+	BC_REQUEST_FREEZE_NOTIFICATION =
+			_IOW('c', 19, struct binder_handle_cookie),
+	/*
+	 * int: handle
+	 * void *: cookie
+	 */
+
+	BC_CLEAR_FREEZE_NOTIFICATION = _IOW('c', 20,
+					    struct binder_handle_cookie),
+	/*
+	 * int: handle
+	 * void *: cookie
+	 */
+
+	BC_FREEZE_NOTIFICATION_DONE = _IOW('c', 21, binder_uintptr_t),
+	/*
+	 * void *: cookie
+	 */
 };
 
 #endif /* _UAPI_LINUX_BINDER_H */
Signature · 1 line
2.45.2.741.gdbec12cfda-goog
#11 lore raw

PATCH v4 1/1 binder: frozen notification

In reply to Yu-Ting Tseng
Quoted 1 line

I believe the external link would be https://r.android.com/3070045

Fixed

Quoted 1 line

BR_ERROR and binder_debug

Removed

Quoted 1 line

Are we ok modifying the ref->freeze->* space here without the

inner_lock?
Good catch. Moved freeze->* access earlier before it's assigned to ref->freeze so that's safe without locking.

Quoted 1 line

I'm not entirely sure why we attempt to queue this as thread work

Was mimicking how it's done with death notification. Removed.

Quoted 1 line

We set the work type to CLEAR regardless, might as well factor it out.

Done

Quoted 1 line

It seems the freeze->* space is protected by the proc->inner_lock right

Yes, freeze->* is protected by the proc inner lock. ref->freeze is protected by the node lock.

Quoted 1 line

I believe you should zero the 'info' before copy_to_user()

Done.

Quoted 1 line

So we traverse every single reference of every single node in this proc

looking for references subscribed to freeze, correct? That's correct.

Quoted 1 line

What if there a EAGAIN error following this below?

Fixed. Moved binder_add_freeze_work(target_proc, true) near the end of the function where we know whether freezing is successful.

Quoted 1 line

any reason why is_frozen didn't make it to the bitfields?

Fixed

Quoted 1 line

No need for a struct

Fixed

Quoted 1 line

You could still keep the original bool for bitfields too if you want.

Sure. Done.

Yu-Ting Tseng (1):
binder: frozen notification

 drivers/android/binder.c            | 300 +++++++++++++++++++++++++++-
 drivers/android/binder_internal.h   |  23 ++-
 include/uapi/linux/android/binder.h |  35 ++++
 3 files changed, 354 insertions(+), 4 deletions(-)

base-commit: 14d7c92f8df9c0964ae6f8b813c1b3ac38120825

Signature · 1 line
2.45.2.803.g4e1b14247a-goog
#12 lore raw

PATCH v4 1/1 binder: frozen notification

In reply to Yu-Ting Tseng

Frozen processes present a significant challenge in binder transactions. When a process is frozen, it cannot, by design, accept and/or respond to binder transactions. As a result, the sender needs to adjust its behavior, such as postponing transactions until the peer process unfreezes. However, there is currently no way to subscribe to these state change events, making it impossible to implement frozen-aware behaviors efficiently.

Introduce a binder API for subscribing to frozen state change events. This allows programs to react to changes in peer process state, mitigating issues related to binder transactions sent to frozen processes.

Implementation details:
For a given binder_ref, the state of frozen notification can be one of the followings:
1. Userspace doesn't want a notification. binder_ref->freeze is null.
2. Userspace wants a notification but none is in flight. list_empty(&binder_ref->freeze->work.entry) = true
3. A notification is in flight and waiting to be read by userspace. binder_ref_freeze.sent is false.
4. A notification was read by userspace and kernel is waiting for an ack. binder_ref_freeze.sent is true.

When a notification is in flight, new state change events are coalesced into the existing binder_ref_freeze struct. If userspace hasn't picked up the notification yet, the driver simply rewrites the state. Otherwise, the notification is flagged as requiring a resend, which will be performed once userspace acks the original notification that's inflight.

See https://r.android.com/3070045 for how userspace is going to use this feature.

---
 drivers/android/binder.c            | 287 +++++++++++++++++++++++++++-
 drivers/android/binder_internal.h   |  21 +-
 include/uapi/linux/android/binder.h |  36 ++++
 3 files changed, 340 insertions(+), 4 deletions(-)
drivers/android/binder.c +285 -2
index b21a7b246a0d..e1bd9810c2af 100644
@@ -3763,6 +3763,159 @@ static void binder_transaction(struct binder_proc *proc,
 	}
 }
 
+static int
+binder_request_freeze_notification(struct binder_proc *proc,
+				   struct binder_thread *thread,
+				   struct binder_handle_cookie *handle_cookie)
+{
+	struct binder_ref_freeze *freeze;
+	struct binder_ref *ref;
+	bool is_frozen;
+
+	freeze = kzalloc(sizeof(*freeze), GFP_KERNEL);
+	if (!freeze)
+		return -ENOMEM;
+	binder_proc_lock(proc);
+	ref = binder_get_ref_olocked(proc, handle_cookie->handle, false);
+	if (!ref) {
+		binder_user_error("%d:%d BC_REQUEST_FREEZE_NOTIFICATION invalid ref %d\n",
+				  proc->pid, thread->pid, handle_cookie->handle);
+		binder_proc_unlock(proc);
+		kfree(freeze);
+		return -EINVAL;
+	}
+
+	binder_node_lock(ref->node);
+
+	if (ref->freeze || !ref->node->proc) {
+		if (ref->freeze) {
+			binder_user_error("%d:%d BC_REQUEST_FREEZE_NOTIFICATION freeze notification already set\n",
+					  proc->pid, thread->pid);
+		} else {
+			binder_user_error("%d:%d BC_REQUEST_FREEZE_NOTIFICATION process is already dead\n",
+					  proc->pid, thread->pid);
+		}
+		binder_node_unlock(ref->node);
+		binder_proc_unlock(proc);
+		kfree(freeze);
+		return -EINVAL;
+	}
+	binder_inner_proc_lock(ref->node->proc);
+	is_frozen = ref->node->proc->is_frozen;
+	binder_inner_proc_unlock(ref->node->proc);
+
+	binder_stats_created(BINDER_STAT_FREEZE);
+	INIT_LIST_HEAD(&freeze->work.entry);
+	freeze->cookie = handle_cookie->cookie;
+	freeze->work.type = BINDER_WORK_FROZEN_BINDER;
+	freeze->is_frozen = ref->node->proc->is_frozen;
+
+	ref->freeze = freeze;
+
+	binder_inner_proc_lock(proc);
+	binder_enqueue_work_ilocked(&ref->freeze->work, &proc->todo);
+	binder_wakeup_proc_ilocked(proc);
+	binder_inner_proc_unlock(proc);
+
+	binder_node_unlock(ref->node);
+	binder_proc_unlock(proc);
+	return 0;
+}
+
+static int
+binder_clear_freeze_notification(struct binder_proc *proc,
+				 struct binder_thread *thread,
+				 struct binder_handle_cookie *handle_cookie)
+{
+	struct binder_ref_freeze *freeze;
+	struct binder_ref *ref;
+
+	binder_proc_lock(proc);
+	ref = binder_get_ref_olocked(proc, handle_cookie->handle, false);
+	if (!ref) {
+		binder_user_error("%d:%d BC_CLEAR_FREEZE_NOTIFICATION invalid ref %d\n",
+				  proc->pid, thread->pid, handle_cookie->handle);
+		binder_proc_unlock(proc);
+		return -EINVAL;
+	}
+
+	binder_node_lock(ref->node);
+
+	if (!ref->freeze) {
+		binder_user_error("%d:%d BC_CLEAR_FREEZE_NOTIFICATION freeze notification not active\n",
+				  proc->pid, thread->pid);
+		binder_node_unlock(ref->node);
+		binder_proc_unlock(proc);
+		return -EINVAL;
+	}
+	freeze = ref->freeze;
+	binder_inner_proc_lock(proc);
+	if (freeze->cookie != handle_cookie->cookie) {
+		binder_user_error("%d:%d BC_CLEAR_FREEZE_NOTIFICATION freeze notification cookie mismatch %016llx != %016llx\n",
+				  proc->pid, thread->pid, (u64)freeze->cookie,
+				  (u64)handle_cookie->cookie);
+		binder_inner_proc_unlock(proc);
+		binder_node_unlock(ref->node);
+		binder_proc_unlock(proc);
+		return -EINVAL;
+	}
+	ref->freeze = NULL;
+	/*
+	 * Take the existing freeze object and overwrite its work type. There are three cases here:
+	 * 1. No pending notification. In this case just add the work to the queue.
+	 * 2. An notification was sent and is pending an ack from userspace. Once an ack arrives, we
+	 *    should resend with the new work type.
+	 * 3. A notification is pending to be sent. Since the work is already in the queue, nothing
+	 *    needs to be done here.
+	 */
+	freeze->work.type = BINDER_WORK_CLEAR_FREEZE_NOTIFICATION;
+	if (list_empty(&freeze->work.entry)) {
+		binder_enqueue_work_ilocked(&freeze->work, &proc->todo);
+		binder_wakeup_proc_ilocked(proc);
+	} else if (freeze->sent) {
+		freeze->resend = true;
+	}
+	binder_inner_proc_unlock(proc);
+	binder_node_unlock(ref->node);
+	binder_proc_unlock(proc);
+	return 0;
+}
+
+static int
+binder_freeze_notification_done(struct binder_proc *proc,
+				struct binder_thread *thread,
+				binder_uintptr_t cookie)
+{
+	struct binder_ref_freeze *freeze = NULL;
+	struct binder_work *w;
+
+	binder_inner_proc_lock(proc);
+	list_for_each_entry(w, &proc->delivered_freeze, entry) {
+		struct binder_ref_freeze *tmp_freeze =
+			container_of(w, struct binder_ref_freeze, work);
+
+		if (tmp_freeze->cookie == cookie) {
+			freeze = tmp_freeze;
+			break;
+		}
+	}
+	if (!freeze) {
+		binder_user_error("%d:%d BC_FREEZE_NOTIFICATION_DONE %016llx not found\n",
+				  proc->pid, thread->pid, (u64)cookie);
+		binder_inner_proc_unlock(proc);
+		return -EINVAL;
+	}
+	binder_dequeue_work_ilocked(&freeze->work);
+	freeze->sent = false;
+	if (freeze->resend) {
+		freeze->resend = false;
+		binder_enqueue_work_ilocked(&freeze->work, &proc->todo);
+		binder_wakeup_proc_ilocked(proc);
+	}
+	binder_inner_proc_unlock(proc);
+	return 0;
+}
+
 /**
  * binder_free_buf() - free the specified buffer
  * @proc:	binder proc that owns buffer
@@ -4246,6 +4399,44 @@ static int binder_thread_write(struct binder_proc *proc,
 			binder_inner_proc_unlock(proc);
 		} break;
 
+		case BC_REQUEST_FREEZE_NOTIFICATION: {
+			struct binder_handle_cookie handle_cookie;
+			int error;
+
+			if (copy_from_user(&handle_cookie, ptr, sizeof(handle_cookie)))
+				return -EFAULT;
+			ptr += sizeof(handle_cookie);
+			error = binder_request_freeze_notification(proc, thread,
+								   &handle_cookie);
+			if (error)
+				return error;
+		} break;
+
+		case BC_CLEAR_FREEZE_NOTIFICATION: {
+			struct binder_handle_cookie handle_cookie;
+			int error;
+
+			if (copy_from_user(&handle_cookie, ptr, sizeof(handle_cookie)))
+				return -EFAULT;
+			ptr += sizeof(handle_cookie);
+			error = binder_clear_freeze_notification(proc, thread, &handle_cookie);
+			if (error)
+				return error;
+		} break;
+
+		case BC_FREEZE_NOTIFICATION_DONE: {
+			binder_uintptr_t cookie;
+			int error;
+
+			if (get_user(cookie, (binder_uintptr_t __user *)ptr))
+				return -EFAULT;
+
+			ptr += sizeof(cookie);
+			error = binder_freeze_notification_done(proc, thread, cookie);
+			if (error)
+				return error;
+		} break;
+
 		default:
 			pr_err("%d:%d unknown command %u\n",
 			       proc->pid, thread->pid, cmd);
@@ -4635,6 +4826,46 @@ static int binder_thread_read(struct binder_proc *proc,
 			if (cmd == BR_DEAD_BINDER)
 				goto done; /* DEAD_BINDER notifications can cause transactions */
 		} break;
+
+		case BINDER_WORK_FROZEN_BINDER: {
+			struct binder_ref_freeze *freeze;
+			struct binder_frozen_state_info info;
+			memset(&info, 0, sizeof(info));
+
+			freeze = container_of(w, struct binder_ref_freeze, work);
+			info.is_frozen = freeze->is_frozen;
+			info.cookie = freeze->cookie;
+			freeze->sent = true;
+			binder_enqueue_work_ilocked(w, &proc->delivered_freeze);
+			binder_inner_proc_unlock(proc);
+
+			if (put_user(BR_FROZEN_BINDER, (uint32_t __user *)ptr))
+				return -EFAULT;
+			ptr += sizeof(uint32_t);
+			if (copy_to_user(ptr, &info, sizeof(info)))
+				return -EFAULT;
+			ptr += sizeof(info);
+			binder_stat_br(proc, thread, BR_FROZEN_BINDER);
+			goto done; /* BR_FROZEN_BINDER notifications can cause transactions */
+		} break;
+
+		case BINDER_WORK_CLEAR_FREEZE_NOTIFICATION: {
+			struct binder_ref_freeze *freeze =
+			    container_of(w, struct binder_ref_freeze, work);
+			binder_uintptr_t cookie = freeze->cookie;
+
+			binder_inner_proc_unlock(proc);
+			kfree(freeze);
+			binder_stats_deleted(BINDER_STAT_FREEZE);
+			if (put_user(BR_CLEAR_FREEZE_NOTIFICATION_DONE, (uint32_t __user *)ptr))
+				return -EFAULT;
+			ptr += sizeof(uint32_t);
+			if (put_user(cookie, (binder_uintptr_t __user *)ptr))
+				return -EFAULT;
+			ptr += sizeof(binder_uintptr_t);
+			binder_stat_br(proc, thread, BR_CLEAR_FREEZE_NOTIFICATION_DONE);
+		} break;
+
 		default:
 			binder_inner_proc_unlock(proc);
 			pr_err("%d:%d: bad work type %d\n",
@@ -5242,6 +5473,48 @@ static bool binder_txns_pending_ilocked(struct binder_proc *proc)
 	return false;
 }
 
+static void binder_add_freeze_work(struct binder_proc *proc, bool is_frozen)
+{
+	struct rb_node *n;
+	struct binder_ref *ref;
+
+	binder_inner_proc_lock(proc);
+	for (n = rb_first(&proc->nodes); n; n = rb_next(n)) {
+		struct binder_node *node;
+
+		node = rb_entry(n, struct binder_node, rb_node);
+		binder_inner_proc_unlock(proc);
+		binder_node_lock(node);
+		hlist_for_each_entry(ref, &node->refs, node_entry) {
+			/*
+			 * Need the node lock to synchronize
+			 * with new notification requests and the
+			 * inner lock to synchronize with queued
+			 * freeze notifications.
+			 */
+			binder_inner_proc_lock(ref->proc);
+			if (!ref->freeze) {
+				binder_inner_proc_unlock(ref->proc);
+				continue;
+			}
+			ref->freeze->work.type = BINDER_WORK_FROZEN_BINDER;
+			if (list_empty(&ref->freeze->work.entry)) {
+				ref->freeze->is_frozen = is_frozen;
+				binder_enqueue_work_ilocked(&ref->freeze->work, &ref->proc->todo);
+				binder_wakeup_proc_ilocked(ref->proc);
+			} else {
+				if (ref->freeze->sent && ref->freeze->is_frozen != is_frozen)
+					ref->freeze->resend = true;
+				ref->freeze->is_frozen = is_frozen;
+			}
+			binder_inner_proc_unlock(ref->proc);
+		}
+		binder_node_unlock(node);
+		binder_inner_proc_lock(proc);
+	}
+	binder_inner_proc_unlock(proc);
+}
+
 static int binder_ioctl_freeze(struct binder_freeze_info *info,
 			       struct binder_proc *target_proc)
 {
@@ -5253,6 +5526,7 @@ static int binder_ioctl_freeze(struct binder_freeze_info *info,
 		target_proc->async_recv = false;
 		target_proc->is_frozen = false;
 		binder_inner_proc_unlock(target_proc);
+		binder_add_freeze_work(target_proc, false);
 		return 0;
 	}
 
@@ -5285,6 +5559,8 @@ static int binder_ioctl_freeze(struct binder_freeze_info *info,
 		binder_inner_proc_lock(target_proc);
 		target_proc->is_frozen = false;
 		binder_inner_proc_unlock(target_proc);
+	} else {
+		binder_add_freeze_work(target_proc, true);
 	}
 
 	return ret;
@@ -5658,6 +5934,7 @@ static int binder_open(struct inode *nodp, struct file *filp)
 	binder_stats_created(BINDER_STAT_PROC);
 	proc->pid = current->group_leader->pid;
 	INIT_LIST_HEAD(&proc->delivered_death);
+	INIT_LIST_HEAD(&proc->delivered_freeze);
 	INIT_LIST_HEAD(&proc->waiting_threads);
 	filp->private_data = proc;
 
@@ -6209,7 +6486,9 @@ static const char * const binder_return_strings[] = {
 	"BR_FAILED_REPLY",
 	"BR_FROZEN_REPLY",
 	"BR_ONEWAY_SPAM_SUSPECT",
-	"BR_TRANSACTION_PENDING_FROZEN"
+	"BR_TRANSACTION_PENDING_FROZEN",
+	"BR_FROZEN_BINDER",
+	"BR_CLEAR_FREEZE_NOTIFICATION_DONE",
 };
 
 static const char * const binder_command_strings[] = {
@@ -6232,6 +6511,9 @@ static const char * const binder_command_strings[] = {
 	"BC_DEAD_BINDER_DONE",
 	"BC_TRANSACTION_SG",
 	"BC_REPLY_SG",
+	"BC_REQUEST_FREEZE_NOTIFICATION",
+	"BC_CLEAR_FREEZE_NOTIFICATION",
+	"BC_FREEZE_NOTIFICATION_DONE",
 };
 
 static const char * const binder_objstat_strings[] = {
@@ -6241,7 +6523,8 @@ static const char * const binder_objstat_strings[] = {
 	"ref",
 	"death",
 	"transaction",
-	"transaction_complete"
+	"transaction_complete",
+	"freeze",
 };
 
 static void print_binder_stats(struct seq_file *m, const char *prefix,
drivers/android/binder_internal.h +19 -2
index 5b7c80b99ae8..3e4b35873c64 100644
@@ -129,12 +129,13 @@ enum binder_stat_types {
 	BINDER_STAT_DEATH,
 	BINDER_STAT_TRANSACTION,
 	BINDER_STAT_TRANSACTION_COMPLETE,
+	BINDER_STAT_FREEZE,
 	BINDER_STAT_COUNT
 };
 
 struct binder_stats {
-	atomic_t br[_IOC_NR(BR_TRANSACTION_PENDING_FROZEN) + 1];
-	atomic_t bc[_IOC_NR(BC_REPLY_SG) + 1];
+	atomic_t br[_IOC_NR(BR_CLEAR_FREEZE_NOTIFICATION_DONE) + 1];
+	atomic_t bc[_IOC_NR(BC_FREEZE_NOTIFICATION_DONE) + 1];
 	atomic_t obj_created[BINDER_STAT_COUNT];
 	atomic_t obj_deleted[BINDER_STAT_COUNT];
 };
@@ -159,6 +160,8 @@ struct binder_work {
 		BINDER_WORK_DEAD_BINDER,
 		BINDER_WORK_DEAD_BINDER_AND_CLEAR,
 		BINDER_WORK_CLEAR_DEATH_NOTIFICATION,
+		BINDER_WORK_FROZEN_BINDER,
+		BINDER_WORK_CLEAR_FREEZE_NOTIFICATION,
 	} type;
 };
 
@@ -275,6 +278,14 @@ struct binder_ref_death {
 	binder_uintptr_t cookie;
 };
 
+struct binder_ref_freeze {
+	struct binder_work work;
+	binder_uintptr_t cookie;
+	bool is_frozen:1;
+	bool sent:1;
+	bool resend:1;
+};
+
 /**
  * struct binder_ref_data - binder_ref counts and id
  * @debug_id:        unique ID for the ref
@@ -307,6 +318,8 @@ struct binder_ref_data {
  *               @node indicates the node must be freed
  * @death:       pointer to death notification (ref_death) if requested
  *               (protected by @node->lock)
+ * @freeze:      pointer to freeze notification (ref_freeze) if requested
+ *               (protected by @node->lock)
  *
  * Structure to track references from procA to target node (on procB). This
  * structure is unsafe to access without holding @proc->outer_lock.
@@ -323,6 +336,7 @@ struct binder_ref {
 	struct binder_proc *proc;
 	struct binder_node *node;
 	struct binder_ref_death *death;
+	struct binder_ref_freeze *freeze;
 };
 
 /**
@@ -374,6 +388,8 @@ struct binder_ref {
  *                        (atomics, no lock needed)
  * @delivered_death:      list of delivered death notification
  *                        (protected by @inner_lock)
+ * @delivered_freeze:     list of delivered freeze notification
+ *                        (protected by @inner_lock)
  * @max_threads:          cap on number of binder threads
  *                        (protected by @inner_lock)
  * @requested_threads:    number of binder threads requested but not
@@ -421,6 +437,7 @@ struct binder_proc {
 	struct list_head todo;
 	struct binder_stats stats;
 	struct list_head delivered_death;
+	struct list_head delivered_freeze;
 	u32 max_threads;
 	int requested_threads;
 	int requested_threads_started;
include/uapi/linux/android/binder.h +36 -0
index d44a8118b2ed..1fd92021a573 100644
@@ -236,6 +236,12 @@ struct binder_frozen_status_info {
 	__u32            async_recv;
 };
 
+struct binder_frozen_state_info {
+	binder_uintptr_t cookie;
+	__u32            is_frozen;
+	__u32            reserved;
+};
+
 /* struct binder_extened_error - extended error information
  * @id:		identifier for the failed operation
  * @command:	command as defined by binder_driver_return_protocol
@@ -467,6 +473,17 @@ enum binder_driver_return_protocol {
 	/*
 	 * The target of the last async transaction is frozen.  No parameters.
 	 */
+
+	BR_FROZEN_BINDER = _IOR('r', 21, struct binder_frozen_state_info),
+	/*
+	 * The cookie and a boolean (is_frozen) that indicates whether the process
+	 * transitioned into a frozen or an unfrozen state.
+	 */
+
+	BR_CLEAR_FREEZE_NOTIFICATION_DONE = _IOR('r', 22, binder_uintptr_t),
+	/*
+	 * void *: cookie
+	 */
 };
 
 enum binder_driver_command_protocol {
@@ -550,6 +567,25 @@ enum binder_driver_command_protocol {
 	/*
 	 * binder_transaction_data_sg: the sent command.
 	 */
+
+	BC_REQUEST_FREEZE_NOTIFICATION =
+			_IOW('c', 19, struct binder_handle_cookie),
+	/*
+	 * int: handle
+	 * void *: cookie
+	 */
+
+	BC_CLEAR_FREEZE_NOTIFICATION = _IOW('c', 20,
+					    struct binder_handle_cookie),
+	/*
+	 * int: handle
+	 * void *: cookie
+	 */
+
+	BC_FREEZE_NOTIFICATION_DONE = _IOW('c', 21, binder_uintptr_t),
+	/*
+	 * void *: cookie
+	 */
 };
 
 #endif /* _UAPI_LINUX_BINDER_H */
Signature · 1 line
2.45.2.803.g4e1b14247a-goog
maintainer
#13 lore raw
Quoted 68 lines · Yu-Ting Tseng

Frozen processes present a significant challenge in binder transactions. When a process is frozen, it cannot, by design, accept and/or respond to binder transactions. As a result, the sender needs to adjust its behavior, such as postponing transactions until the peer process unfreezes. However, there is currently no way to subscribe to these state change events, making it impossible to implement frozen-aware behaviors efficiently.

Introduce a binder API for subscribing to frozen state change events. This allows programs to react to changes in peer process state, mitigating issues related to binder transactions sent to frozen processes.

Implementation details:
For a given binder_ref, the state of frozen notification can be one of the followings:
1. Userspace doesn't want a notification. binder_ref->freeze is null.
2. Userspace wants a notification but none is in flight. list_empty(&binder_ref->freeze->work.entry) = true
3. A notification is in flight and waiting to be read by userspace. binder_ref_freeze.sent is false.
4. A notification was read by userspace and kernel is waiting for an ack. binder_ref_freeze.sent is true.

When a notification is in flight, new state change events are coalesced into the existing binder_ref_freeze struct. If userspace hasn't picked up the notification yet, the driver simply rewrites the state. Otherwise, the notification is flagged as requiring a resend, which will be performed once userspace acks the original notification that's inflight.

See https://r.android.com/3070045 for how userspace is going to use this feature.

---
 drivers/android/binder.c            | 287 +++++++++++++++++++++++++++-
 drivers/android/binder_internal.h   |  21 +-
 include/uapi/linux/android/binder.h |  36 ++++
 3 files changed, 340 insertions(+), 4 deletions(-)
drivers/android/binder.c +20 -0
index b21a7b246a0d..e1bd9810c2af 100644
@@ -3763,6 +3763,159 @@ static void binder_transaction(struct binder_proc *proc,
 	}
 }
 
+static int
+binder_request_freeze_notification(struct binder_proc *proc,
+				   struct binder_thread *thread,
+				   struct binder_handle_cookie *handle_cookie)
+{
+	struct binder_ref_freeze *freeze;
+	struct binder_ref *ref;
+	bool is_frozen;
+
+	freeze = kzalloc(sizeof(*freeze), GFP_KERNEL);
+	if (!freeze)
+		return -ENOMEM;
+	binder_proc_lock(proc);
+	ref = binder_get_ref_olocked(proc, handle_cookie->handle, false);
+	if (!ref) {
+		binder_user_error("%d:%d BC_REQUEST_FREEZE_NOTIFICATION invalid ref %d\n",
+				  proc->pid, thread->pid, handle_cookie->handle);
+		binder_proc_unlock(proc);
+		kfree(freeze);
+		return -EINVAL;

drivers/android/binder.c static void binder_transaction(struct binder_proc *proc, · line 3795

Quoted 10 lines
drivers/android/binder.c +10 -0
+	}
+
+	binder_node_lock(ref->node);
+
+	if (ref->freeze || !ref->node->proc) {
+		if (ref->freeze) {
+			binder_user_error("%d:%d BC_REQUEST_FREEZE_NOTIFICATION freeze notification already set\n",
+			binder_user_error("%d:%d BC_REQUEST_FREEZE_NOTIFICATION process is already dead\n",
+					  proc->pid, thread->pid);
+		}

nit: can you merge the branches into a single printk? maybe something like the following:

	binder_user_error("%d:%d invalid BC_REQUEST_FREEZE_NOTIFICATION %s\n",
			  proc->pid, thread->pid,
			  ref->freeze ? "already set" : "dead node");
Quoted 4 lines
drivers/android/binder.c +4 -0
+		binder_node_unlock(ref->node);
+		binder_proc_unlock(proc);
+		kfree(freeze);
+		return -EINVAL;

drivers/android/binder.c static void binder_transaction(struct binder_proc *proc, · line 3811

Quoted 10 lines
drivers/android/binder.c +10 -0
+	}
+	binder_inner_proc_lock(ref->node->proc);
+	is_frozen = ref->node->proc->is_frozen;
+	binder_inner_proc_unlock(ref->node->proc);
+
+	binder_stats_created(BINDER_STAT_FREEZE);
+	INIT_LIST_HEAD(&freeze->work.entry);
+	freeze->cookie = handle_cookie->cookie;
+	freeze->work.type = BINDER_WORK_FROZEN_BINDER;
+	freeze->is_frozen = ref->node->proc->is_frozen;

I believe you meant to use the 'is_frozen' here. Otherwise it doesn't seem to be used after assignment.

Quoted 41 lines
drivers/android/binder.c +41 -0
+
+	ref->freeze = freeze;
+
+	binder_inner_proc_lock(proc);
+	binder_enqueue_work_ilocked(&ref->freeze->work, &proc->todo);
+	binder_wakeup_proc_ilocked(proc);
+	binder_inner_proc_unlock(proc);
+
+	binder_node_unlock(ref->node);
+	binder_proc_unlock(proc);
+	return 0;
+}
+
+static int
+binder_clear_freeze_notification(struct binder_proc *proc,
+				 struct binder_thread *thread,
+				 struct binder_handle_cookie *handle_cookie)
+{
+	struct binder_ref_freeze *freeze;
+	struct binder_ref *ref;
+
+	binder_proc_lock(proc);
+	ref = binder_get_ref_olocked(proc, handle_cookie->handle, false);
+	if (!ref) {
+		binder_user_error("%d:%d BC_CLEAR_FREEZE_NOTIFICATION invalid ref %d\n",
+				  proc->pid, thread->pid, handle_cookie->handle);
+		binder_proc_unlock(proc);
+		return -EINVAL;
+	}
+
+	binder_node_lock(ref->node);
+
+	if (!ref->freeze) {
+		binder_user_error("%d:%d BC_CLEAR_FREEZE_NOTIFICATION freeze notification not active\n",
+				  proc->pid, thread->pid);
+		binder_node_unlock(ref->node);
+		binder_proc_unlock(proc);
+		return -EINVAL;
+	}
+	freeze = ref->freeze;
+	binder_inner_proc_lock(proc);

drivers/android/binder.c static void binder_transaction(struct binder_proc *proc, · line 3862

Quoted 10 lines
drivers/android/binder.c +10 -0
+	if (freeze->cookie != handle_cookie->cookie) {
+		binder_user_error("%d:%d BC_CLEAR_FREEZE_NOTIFICATION freeze notification cookie mismatch %016llx != %016llx\n",
+				  proc->pid, thread->pid, (u64)freeze->cookie,
+				  (u64)handle_cookie->cookie);
+		binder_inner_proc_unlock(proc);
+		binder_node_unlock(ref->node);
+		binder_proc_unlock(proc);
+		return -EINVAL;
+	}
+	ref->freeze = NULL;

You detach ref->freeze here and handle the kfree(freeze) later which seems good. However, what happens when a reference is released with an active ref->freeze? It seems to me this is a memory leak. This could accumulate significantly as more references get dropped in thit way.

I think you want to add a kfree(ref->freeze) to binder_free_ref().

drivers/android/binder.c static void binder_transaction(struct binder_proc *proc, · line 3866

Quoted 4 lines
drivers/android/binder.c +4 -0
+	/*
+	 * Take the existing freeze object and overwrite its work type. There are three cases here:
+	 * 1. No pending notification. In this case just add the work to the queue.
+	 * 2. An notification was sent and is pending an ack from userspace. Once an ack arrives, we

Typo s/An/A/

Quoted 101 lines
drivers/android/binder.c +90 -0
+	 *    should resend with the new work type.
+	 * 3. A notification is pending to be sent. Since the work is already in the queue, nothing
+	 *    needs to be done here.
+	 */
+	freeze->work.type = BINDER_WORK_CLEAR_FREEZE_NOTIFICATION;
+	if (list_empty(&freeze->work.entry)) {
+		binder_enqueue_work_ilocked(&freeze->work, &proc->todo);
+		binder_wakeup_proc_ilocked(proc);
+	} else if (freeze->sent) {
+		freeze->resend = true;
+	}
+	binder_inner_proc_unlock(proc);
+	binder_node_unlock(ref->node);
+	binder_proc_unlock(proc);
+	return 0;
+}
+
+static int
+binder_freeze_notification_done(struct binder_proc *proc,
+				struct binder_thread *thread,
+				binder_uintptr_t cookie)
+{
+	struct binder_ref_freeze *freeze = NULL;
+	struct binder_work *w;
+
+	binder_inner_proc_lock(proc);
+	list_for_each_entry(w, &proc->delivered_freeze, entry) {
+		struct binder_ref_freeze *tmp_freeze =
+			container_of(w, struct binder_ref_freeze, work);
+
+		if (tmp_freeze->cookie == cookie) {
+			freeze = tmp_freeze;
+			break;
+		}
+	}
+	if (!freeze) {
+		binder_user_error("%d:%d BC_FREEZE_NOTIFICATION_DONE %016llx not found\n",
+				  proc->pid, thread->pid, (u64)cookie);
+		binder_inner_proc_unlock(proc);
+		return -EINVAL;
+	}
+	binder_dequeue_work_ilocked(&freeze->work);
+	freeze->sent = false;
+	if (freeze->resend) {
+		freeze->resend = false;
+		binder_enqueue_work_ilocked(&freeze->work, &proc->todo);
+		binder_wakeup_proc_ilocked(proc);
+	}
+	binder_inner_proc_unlock(proc);
+	return 0;
+}
+
 /**
  * binder_free_buf() - free the specified buffer
  * @proc:	binder proc that owns buffer
@@ -4246,6 +4399,44 @@ static int binder_thread_write(struct binder_proc *proc,
 			binder_inner_proc_unlock(proc);
 		} break;
 
+		case BC_REQUEST_FREEZE_NOTIFICATION: {
+			struct binder_handle_cookie handle_cookie;
+			int error;
+
+			if (copy_from_user(&handle_cookie, ptr, sizeof(handle_cookie)))
+				return -EFAULT;
+			ptr += sizeof(handle_cookie);
+			error = binder_request_freeze_notification(proc, thread,
+								   &handle_cookie);
+			if (error)
+				return error;
+		} break;
+
+		case BC_CLEAR_FREEZE_NOTIFICATION: {
+			struct binder_handle_cookie handle_cookie;
+			int error;
+
+			if (copy_from_user(&handle_cookie, ptr, sizeof(handle_cookie)))
+				return -EFAULT;
+			ptr += sizeof(handle_cookie);
+			error = binder_clear_freeze_notification(proc, thread, &handle_cookie);
+			if (error)
+				return error;
+		} break;
+
+		case BC_FREEZE_NOTIFICATION_DONE: {
+			binder_uintptr_t cookie;
+			int error;
+
+			if (get_user(cookie, (binder_uintptr_t __user *)ptr))
+				return -EFAULT;
+
+			ptr += sizeof(cookie);
+			error = binder_freeze_notification_done(proc, thread, cookie);
+			if (error)
+				return error;
+		} break;
+
 		default:
 			pr_err("%d:%d unknown command %u\n",
 			       proc->pid, thread->pid, cmd);

drivers/android/binder.c static int binder_thread_read(struct binder_proc *proc, · line 4833

Quoted 8 lines
drivers/android/binder.c +5 -0
@@ -4635,6 +4826,46 @@ static int binder_thread_read(struct binder_proc *proc,
 			if (cmd == BR_DEAD_BINDER)
 				goto done; /* DEAD_BINDER notifications can cause transactions */
 		} break;
+
+		case BINDER_WORK_FROZEN_BINDER: {
+			struct binder_ref_freeze *freeze;
+			struct binder_frozen_state_info info;
+			memset(&info, 0, sizeof(info));

The memset should have a blank line after the declarations. This should trigger a checkpatch issue I think.

Quoted 285 lines
drivers/android/binder.c +89 -2
+
+			freeze = container_of(w, struct binder_ref_freeze, work);
+			info.is_frozen = freeze->is_frozen;
+			info.cookie = freeze->cookie;
+			freeze->sent = true;
+			binder_enqueue_work_ilocked(w, &proc->delivered_freeze);
+			binder_inner_proc_unlock(proc);
+
+			if (put_user(BR_FROZEN_BINDER, (uint32_t __user *)ptr))
+				return -EFAULT;
+			ptr += sizeof(uint32_t);
+			if (copy_to_user(ptr, &info, sizeof(info)))
+				return -EFAULT;
+			ptr += sizeof(info);
+			binder_stat_br(proc, thread, BR_FROZEN_BINDER);
+			goto done; /* BR_FROZEN_BINDER notifications can cause transactions */
+		} break;
+
+		case BINDER_WORK_CLEAR_FREEZE_NOTIFICATION: {
+			struct binder_ref_freeze *freeze =
+			    container_of(w, struct binder_ref_freeze, work);
+			binder_uintptr_t cookie = freeze->cookie;
+
+			binder_inner_proc_unlock(proc);
+			kfree(freeze);
+			binder_stats_deleted(BINDER_STAT_FREEZE);
+			if (put_user(BR_CLEAR_FREEZE_NOTIFICATION_DONE, (uint32_t __user *)ptr))
+				return -EFAULT;
+			ptr += sizeof(uint32_t);
+			if (put_user(cookie, (binder_uintptr_t __user *)ptr))
+				return -EFAULT;
+			ptr += sizeof(binder_uintptr_t);
+			binder_stat_br(proc, thread, BR_CLEAR_FREEZE_NOTIFICATION_DONE);
+		} break;
+
 		default:
 			binder_inner_proc_unlock(proc);
 			pr_err("%d:%d: bad work type %d\n",
@@ -5242,6 +5473,48 @@ static bool binder_txns_pending_ilocked(struct binder_proc *proc)
 	return false;
 }
 
+static void binder_add_freeze_work(struct binder_proc *proc, bool is_frozen)
+{
+	struct rb_node *n;
+	struct binder_ref *ref;
+
+	binder_inner_proc_lock(proc);
+	for (n = rb_first(&proc->nodes); n; n = rb_next(n)) {
+		struct binder_node *node;
+
+		node = rb_entry(n, struct binder_node, rb_node);
+		binder_inner_proc_unlock(proc);
+		binder_node_lock(node);
+		hlist_for_each_entry(ref, &node->refs, node_entry) {
+			/*
+			 * Need the node lock to synchronize
+			 * with new notification requests and the
+			 * inner lock to synchronize with queued
+			 * freeze notifications.
+			 */
+			binder_inner_proc_lock(ref->proc);
+			if (!ref->freeze) {
+				binder_inner_proc_unlock(ref->proc);
+				continue;
+			}
+			ref->freeze->work.type = BINDER_WORK_FROZEN_BINDER;
+			if (list_empty(&ref->freeze->work.entry)) {
+				ref->freeze->is_frozen = is_frozen;
+				binder_enqueue_work_ilocked(&ref->freeze->work, &ref->proc->todo);
+				binder_wakeup_proc_ilocked(ref->proc);
+			} else {
+				if (ref->freeze->sent && ref->freeze->is_frozen != is_frozen)
+					ref->freeze->resend = true;
+				ref->freeze->is_frozen = is_frozen;
+			}
+			binder_inner_proc_unlock(ref->proc);
+		}
+		binder_node_unlock(node);
+		binder_inner_proc_lock(proc);
+	}
+	binder_inner_proc_unlock(proc);
+}
+
 static int binder_ioctl_freeze(struct binder_freeze_info *info,
 			       struct binder_proc *target_proc)
 {
@@ -5253,6 +5526,7 @@ static int binder_ioctl_freeze(struct binder_freeze_info *info,
 		target_proc->async_recv = false;
 		target_proc->is_frozen = false;
 		binder_inner_proc_unlock(target_proc);
+		binder_add_freeze_work(target_proc, false);
 		return 0;
 	}
 
@@ -5285,6 +5559,8 @@ static int binder_ioctl_freeze(struct binder_freeze_info *info,
 		binder_inner_proc_lock(target_proc);
 		target_proc->is_frozen = false;
 		binder_inner_proc_unlock(target_proc);
+	} else {
+		binder_add_freeze_work(target_proc, true);
 	}
 
 	return ret;
@@ -5658,6 +5934,7 @@ static int binder_open(struct inode *nodp, struct file *filp)
 	binder_stats_created(BINDER_STAT_PROC);
 	proc->pid = current->group_leader->pid;
 	INIT_LIST_HEAD(&proc->delivered_death);
+	INIT_LIST_HEAD(&proc->delivered_freeze);
 	INIT_LIST_HEAD(&proc->waiting_threads);
 	filp->private_data = proc;
 
@@ -6209,7 +6486,9 @@ static const char * const binder_return_strings[] = {
 	"BR_FAILED_REPLY",
 	"BR_FROZEN_REPLY",
 	"BR_ONEWAY_SPAM_SUSPECT",
-	"BR_TRANSACTION_PENDING_FROZEN"
+	"BR_TRANSACTION_PENDING_FROZEN",
+	"BR_FROZEN_BINDER",
+	"BR_CLEAR_FREEZE_NOTIFICATION_DONE",
 };
 
 static const char * const binder_command_strings[] = {
@@ -6232,6 +6511,9 @@ static const char * const binder_command_strings[] = {
 	"BC_DEAD_BINDER_DONE",
 	"BC_TRANSACTION_SG",
 	"BC_REPLY_SG",
+	"BC_REQUEST_FREEZE_NOTIFICATION",
+	"BC_CLEAR_FREEZE_NOTIFICATION",
+	"BC_FREEZE_NOTIFICATION_DONE",
 };
 
 static const char * const binder_objstat_strings[] = {
@@ -6241,7 +6523,8 @@ static const char * const binder_objstat_strings[] = {
 	"ref",
 	"death",
 	"transaction",
-	"transaction_complete"
+	"transaction_complete",
+	"freeze",
 };
 
 static void print_binder_stats(struct seq_file *m, const char *prefix,
drivers/android/binder_internal.h +19 -2
index 5b7c80b99ae8..3e4b35873c64 100644
@@ -129,12 +129,13 @@ enum binder_stat_types {
 	BINDER_STAT_DEATH,
 	BINDER_STAT_TRANSACTION,
 	BINDER_STAT_TRANSACTION_COMPLETE,
+	BINDER_STAT_FREEZE,
 	BINDER_STAT_COUNT
 };
 
 struct binder_stats {
-	atomic_t br[_IOC_NR(BR_TRANSACTION_PENDING_FROZEN) + 1];
-	atomic_t bc[_IOC_NR(BC_REPLY_SG) + 1];
+	atomic_t br[_IOC_NR(BR_CLEAR_FREEZE_NOTIFICATION_DONE) + 1];
+	atomic_t bc[_IOC_NR(BC_FREEZE_NOTIFICATION_DONE) + 1];
 	atomic_t obj_created[BINDER_STAT_COUNT];
 	atomic_t obj_deleted[BINDER_STAT_COUNT];
 };
@@ -159,6 +160,8 @@ struct binder_work {
 		BINDER_WORK_DEAD_BINDER,
 		BINDER_WORK_DEAD_BINDER_AND_CLEAR,
 		BINDER_WORK_CLEAR_DEATH_NOTIFICATION,
+		BINDER_WORK_FROZEN_BINDER,
+		BINDER_WORK_CLEAR_FREEZE_NOTIFICATION,
 	} type;
 };
 
@@ -275,6 +278,14 @@ struct binder_ref_death {
 	binder_uintptr_t cookie;
 };
 
+struct binder_ref_freeze {
+	struct binder_work work;
+	binder_uintptr_t cookie;
+	bool is_frozen:1;
+	bool sent:1;
+	bool resend:1;
+};
+
 /**
  * struct binder_ref_data - binder_ref counts and id
  * @debug_id:        unique ID for the ref
@@ -307,6 +318,8 @@ struct binder_ref_data {
  *               @node indicates the node must be freed
  * @death:       pointer to death notification (ref_death) if requested
  *               (protected by @node->lock)
+ * @freeze:      pointer to freeze notification (ref_freeze) if requested
+ *               (protected by @node->lock)
  *
  * Structure to track references from procA to target node (on procB). This
  * structure is unsafe to access without holding @proc->outer_lock.
@@ -323,6 +336,7 @@ struct binder_ref {
 	struct binder_proc *proc;
 	struct binder_node *node;
 	struct binder_ref_death *death;
+	struct binder_ref_freeze *freeze;
 };
 
 /**
@@ -374,6 +388,8 @@ struct binder_ref {
  *                        (atomics, no lock needed)
  * @delivered_death:      list of delivered death notification
  *                        (protected by @inner_lock)
+ * @delivered_freeze:     list of delivered freeze notification
+ *                        (protected by @inner_lock)
  * @max_threads:          cap on number of binder threads
  *                        (protected by @inner_lock)
  * @requested_threads:    number of binder threads requested but not
@@ -421,6 +437,7 @@ struct binder_proc {
 	struct list_head todo;
 	struct binder_stats stats;
 	struct list_head delivered_death;
+	struct list_head delivered_freeze;
 	u32 max_threads;
 	int requested_threads;
 	int requested_threads_started;
include/uapi/linux/android/binder.h +36 -0
index d44a8118b2ed..1fd92021a573 100644
@@ -236,6 +236,12 @@ struct binder_frozen_status_info {
 	__u32            async_recv;
 };
 
+struct binder_frozen_state_info {
+	binder_uintptr_t cookie;
+	__u32            is_frozen;
+	__u32            reserved;
+};
+
 /* struct binder_extened_error - extended error information
  * @id:		identifier for the failed operation
  * @command:	command as defined by binder_driver_return_protocol
@@ -467,6 +473,17 @@ enum binder_driver_return_protocol {
 	/*
 	 * The target of the last async transaction is frozen.  No parameters.
 	 */
+
+	BR_FROZEN_BINDER = _IOR('r', 21, struct binder_frozen_state_info),
+	/*
+	 * The cookie and a boolean (is_frozen) that indicates whether the process
+	 * transitioned into a frozen or an unfrozen state.
+	 */
+
+	BR_CLEAR_FREEZE_NOTIFICATION_DONE = _IOR('r', 22, binder_uintptr_t),
+	/*
+	 * void *: cookie
+	 */
 };
 
 enum binder_driver_command_protocol {
@@ -550,6 +567,25 @@ enum binder_driver_command_protocol {
 	/*
 	 * binder_transaction_data_sg: the sent command.
 	 */
+
+	BC_REQUEST_FREEZE_NOTIFICATION =
+			_IOW('c', 19, struct binder_handle_cookie),
+	/*
+	 * int: handle
+	 * void *: cookie
+	 */
+
+	BC_CLEAR_FREEZE_NOTIFICATION = _IOW('c', 20,
+					    struct binder_handle_cookie),
+	/*
+	 * int: handle
+	 * void *: cookie
+	 */
+
+	BC_FREEZE_NOTIFICATION_DONE = _IOW('c', 21, binder_uintptr_t),
+	/*
+	 * void *: cookie
+	 */
 };
 
 #endif /* _UAPI_LINUX_BINDER_H */
Signature · 1 line
2.45.2.803.g4e1b14247a-goog
#14 lore raw

PATCH v5 1/1 binder: frozen notification

In reply to Yu-Ting Tseng
Quoted 1 line

can you merge the branches into a single printk?

Done

Quoted 1 line

I believe you meant to use the 'is_frozen' here

Fixed

Quoted 1 line

I think you want to add a kfree(ref->freeze) to binder_free_ref().

Done

Quoted 1 line

Typo s/An/A/

Done

Quoted 1 line

The memset should have a blank line after the declarations. This should trigger a checkpatch issue I think.

Fixed. Reran checkpatch and there is no other issues flagged now. (except for line-too-long which doesn't really have a fix)

Also merged the binder_features change into this patch as requested.

Yu-Ting Tseng (1):
binder: frozen notification

 drivers/android/binder.c                      | 284 +++++++++++++++++-
 drivers/android/binder_internal.h             |  21 +-
 drivers/android/binderfs.c                    |   8 +
 include/uapi/linux/android/binder.h           |  36 +++
 .../filesystems/binderfs/binderfs_test.c      |   1 +
 5 files changed, 346 insertions(+), 4 deletions(-)

base-commit: 14d7c92f8df9c0964ae6f8b813c1b3ac38120825

Signature · 1 line
2.45.2.803.g4e1b14247a-goog
#15 lore raw

PATCH v5 1/1 binder: frozen notification

In reply to Yu-Ting Tseng

Frozen processes present a significant challenge in binder transactions. When a process is frozen, it cannot, by design, accept and/or respond to binder transactions. As a result, the sender needs to adjust its behavior, such as postponing transactions until the peer process unfreezes. However, there is currently no way to subscribe to these state change events, making it impossible to implement frozen-aware behaviors efficiently.

Introduce a binder API for subscribing to frozen state change events. This allows programs to react to changes in peer process state, mitigating issues related to binder transactions sent to frozen processes.

Implementation details:
For a given binder_ref, the state of frozen notification can be one of the followings:
1. Userspace doesn't want a notification. binder_ref->freeze is null.
2. Userspace wants a notification but none is in flight. list_empty(&binder_ref->freeze->work.entry) = true
3. A notification is in flight and waiting to be read by userspace. binder_ref_freeze.sent is false.
4. A notification was read by userspace and kernel is waiting for an ack. binder_ref_freeze.sent is true.

When a notification is in flight, new state change events are coalesced into the existing binder_ref_freeze struct. If userspace hasn't picked up the notification yet, the driver simply rewrites the state. Otherwise, the notification is flagged as requiring a resend, which will be performed once userspace acks the original notification that's inflight.

See https://r.android.com/3070045 for how userspace is going to use this feature.

---
 drivers/android/binder.c                      | 284 +++++++++++++++++-
 drivers/android/binder_internal.h             |  21 +-
 drivers/android/binderfs.c                    |   8 +
 include/uapi/linux/android/binder.h           |  36 +++
 .../filesystems/binderfs/binderfs_test.c      |   1 +
 5 files changed, 346 insertions(+), 4 deletions(-)
drivers/android/binder.c +282 -2
index b21a7b246a0d..77f318127a6e 100644
@@ -1295,6 +1295,7 @@ static void binder_free_ref(struct binder_ref *ref)
 	if (ref->node)
 		binder_free_node(ref->node);
 	kfree(ref->death);
+	kfree(ref->freeze);
 	kfree(ref);
 }
 
@@ -3763,6 +3764,155 @@ static void binder_transaction(struct binder_proc *proc,
 	}
 }
 
+static int
+binder_request_freeze_notification(struct binder_proc *proc,
+				   struct binder_thread *thread,
+				   struct binder_handle_cookie *handle_cookie)
+{
+	struct binder_ref_freeze *freeze;
+	struct binder_ref *ref;
+	bool is_frozen;
+
+	freeze = kzalloc(sizeof(*freeze), GFP_KERNEL);
+	if (!freeze)
+		return -ENOMEM;
+	binder_proc_lock(proc);
+	ref = binder_get_ref_olocked(proc, handle_cookie->handle, false);
+	if (!ref) {
+		binder_user_error("%d:%d BC_REQUEST_FREEZE_NOTIFICATION invalid ref %d\n",
+				  proc->pid, thread->pid, handle_cookie->handle);
+		binder_proc_unlock(proc);
+		kfree(freeze);
+		return -EINVAL;
+	}
+
+	binder_node_lock(ref->node);
+
+	if (ref->freeze || !ref->node->proc) {
+		binder_user_error("%d:%d invalid BC_REQUEST_FREEZE_NOTIFICATION %s\n",
+				  proc->pid, thread->pid,
+				  ref->freeze ? "already set" : "dead node");
+		binder_node_unlock(ref->node);
+		binder_proc_unlock(proc);
+		kfree(freeze);
+		return -EINVAL;
+	}
+	binder_inner_proc_lock(ref->node->proc);
+	is_frozen = ref->node->proc->is_frozen;
+	binder_inner_proc_unlock(ref->node->proc);
+
+	binder_stats_created(BINDER_STAT_FREEZE);
+	INIT_LIST_HEAD(&freeze->work.entry);
+	freeze->cookie = handle_cookie->cookie;
+	freeze->work.type = BINDER_WORK_FROZEN_BINDER;
+	freeze->is_frozen = is_frozen;
+
+	ref->freeze = freeze;
+
+	binder_inner_proc_lock(proc);
+	binder_enqueue_work_ilocked(&ref->freeze->work, &proc->todo);
+	binder_wakeup_proc_ilocked(proc);
+	binder_inner_proc_unlock(proc);
+
+	binder_node_unlock(ref->node);
+	binder_proc_unlock(proc);
+	return 0;
+}
+
+static int
+binder_clear_freeze_notification(struct binder_proc *proc,
+				 struct binder_thread *thread,
+				 struct binder_handle_cookie *handle_cookie)
+{
+	struct binder_ref_freeze *freeze;
+	struct binder_ref *ref;
+
+	binder_proc_lock(proc);
+	ref = binder_get_ref_olocked(proc, handle_cookie->handle, false);
+	if (!ref) {
+		binder_user_error("%d:%d BC_CLEAR_FREEZE_NOTIFICATION invalid ref %d\n",
+				  proc->pid, thread->pid, handle_cookie->handle);
+		binder_proc_unlock(proc);
+		return -EINVAL;
+	}
+
+	binder_node_lock(ref->node);
+
+	if (!ref->freeze) {
+		binder_user_error("%d:%d BC_CLEAR_FREEZE_NOTIFICATION freeze notification not active\n",
+				  proc->pid, thread->pid);
+		binder_node_unlock(ref->node);
+		binder_proc_unlock(proc);
+		return -EINVAL;
+	}
+	freeze = ref->freeze;
+	binder_inner_proc_lock(proc);
+	if (freeze->cookie != handle_cookie->cookie) {
+		binder_user_error("%d:%d BC_CLEAR_FREEZE_NOTIFICATION freeze notification cookie mismatch %016llx != %016llx\n",
+				  proc->pid, thread->pid, (u64)freeze->cookie,
+				  (u64)handle_cookie->cookie);
+		binder_inner_proc_unlock(proc);
+		binder_node_unlock(ref->node);
+		binder_proc_unlock(proc);
+		return -EINVAL;
+	}
+	ref->freeze = NULL;
+	/*
+	 * Take the existing freeze object and overwrite its work type. There are three cases here:
+	 * 1. No pending notification. In this case just add the work to the queue.
+	 * 2. A notification was sent and is pending an ack from userspace. Once an ack arrives, we
+	 *    should resend with the new work type.
+	 * 3. A notification is pending to be sent. Since the work is already in the queue, nothing
+	 *    needs to be done here.
+	 */
+	freeze->work.type = BINDER_WORK_CLEAR_FREEZE_NOTIFICATION;
+	if (list_empty(&freeze->work.entry)) {
+		binder_enqueue_work_ilocked(&freeze->work, &proc->todo);
+		binder_wakeup_proc_ilocked(proc);
+	} else if (freeze->sent) {
+		freeze->resend = true;
+	}
+	binder_inner_proc_unlock(proc);
+	binder_node_unlock(ref->node);
+	binder_proc_unlock(proc);
+	return 0;
+}
+
+static int
+binder_freeze_notification_done(struct binder_proc *proc,
+				struct binder_thread *thread,
+				binder_uintptr_t cookie)
+{
+	struct binder_ref_freeze *freeze = NULL;
+	struct binder_work *w;
+
+	binder_inner_proc_lock(proc);
+	list_for_each_entry(w, &proc->delivered_freeze, entry) {
+		struct binder_ref_freeze *tmp_freeze =
+			container_of(w, struct binder_ref_freeze, work);
+
+		if (tmp_freeze->cookie == cookie) {
+			freeze = tmp_freeze;
+			break;
+		}
+	}
+	if (!freeze) {
+		binder_user_error("%d:%d BC_FREEZE_NOTIFICATION_DONE %016llx not found\n",
+				  proc->pid, thread->pid, (u64)cookie);
+		binder_inner_proc_unlock(proc);
+		return -EINVAL;
+	}
+	binder_dequeue_work_ilocked(&freeze->work);
+	freeze->sent = false;
+	if (freeze->resend) {
+		freeze->resend = false;
+		binder_enqueue_work_ilocked(&freeze->work, &proc->todo);
+		binder_wakeup_proc_ilocked(proc);
+	}
+	binder_inner_proc_unlock(proc);
+	return 0;
+}
+
 /**
  * binder_free_buf() - free the specified buffer
  * @proc:	binder proc that owns buffer
@@ -4246,6 +4396,44 @@ static int binder_thread_write(struct binder_proc *proc,
 			binder_inner_proc_unlock(proc);
 		} break;
 
+		case BC_REQUEST_FREEZE_NOTIFICATION: {
+			struct binder_handle_cookie handle_cookie;
+			int error;
+
+			if (copy_from_user(&handle_cookie, ptr, sizeof(handle_cookie)))
+				return -EFAULT;
+			ptr += sizeof(handle_cookie);
+			error = binder_request_freeze_notification(proc, thread,
+								   &handle_cookie);
+			if (error)
+				return error;
+		} break;
+
+		case BC_CLEAR_FREEZE_NOTIFICATION: {
+			struct binder_handle_cookie handle_cookie;
+			int error;
+
+			if (copy_from_user(&handle_cookie, ptr, sizeof(handle_cookie)))
+				return -EFAULT;
+			ptr += sizeof(handle_cookie);
+			error = binder_clear_freeze_notification(proc, thread, &handle_cookie);
+			if (error)
+				return error;
+		} break;
+
+		case BC_FREEZE_NOTIFICATION_DONE: {
+			binder_uintptr_t cookie;
+			int error;
+
+			if (get_user(cookie, (binder_uintptr_t __user *)ptr))
+				return -EFAULT;
+
+			ptr += sizeof(cookie);
+			error = binder_freeze_notification_done(proc, thread, cookie);
+			if (error)
+				return error;
+		} break;
+
 		default:
 			pr_err("%d:%d unknown command %u\n",
 			       proc->pid, thread->pid, cmd);
@@ -4635,6 +4823,46 @@ static int binder_thread_read(struct binder_proc *proc,
 			if (cmd == BR_DEAD_BINDER)
 				goto done; /* DEAD_BINDER notifications can cause transactions */
 		} break;
+
+		case BINDER_WORK_FROZEN_BINDER: {
+			struct binder_ref_freeze *freeze;
+			struct binder_frozen_state_info info;
+
+			memset(&info, 0, sizeof(info));
+			freeze = container_of(w, struct binder_ref_freeze, work);
+			info.is_frozen = freeze->is_frozen;
+			info.cookie = freeze->cookie;
+			freeze->sent = true;
+			binder_enqueue_work_ilocked(w, &proc->delivered_freeze);
+			binder_inner_proc_unlock(proc);
+
+			if (put_user(BR_FROZEN_BINDER, (uint32_t __user *)ptr))
+				return -EFAULT;
+			ptr += sizeof(uint32_t);
+			if (copy_to_user(ptr, &info, sizeof(info)))
+				return -EFAULT;
+			ptr += sizeof(info);
+			binder_stat_br(proc, thread, BR_FROZEN_BINDER);
+			goto done; /* BR_FROZEN_BINDER notifications can cause transactions */
+		} break;
+
+		case BINDER_WORK_CLEAR_FREEZE_NOTIFICATION: {
+			struct binder_ref_freeze *freeze =
+			    container_of(w, struct binder_ref_freeze, work);
+			binder_uintptr_t cookie = freeze->cookie;
+
+			binder_inner_proc_unlock(proc);
+			kfree(freeze);
+			binder_stats_deleted(BINDER_STAT_FREEZE);
+			if (put_user(BR_CLEAR_FREEZE_NOTIFICATION_DONE, (uint32_t __user *)ptr))
+				return -EFAULT;
+			ptr += sizeof(uint32_t);
+			if (put_user(cookie, (binder_uintptr_t __user *)ptr))
+				return -EFAULT;
+			ptr += sizeof(binder_uintptr_t);
+			binder_stat_br(proc, thread, BR_CLEAR_FREEZE_NOTIFICATION_DONE);
+		} break;
+
 		default:
 			binder_inner_proc_unlock(proc);
 			pr_err("%d:%d: bad work type %d\n",
@@ -5242,6 +5470,48 @@ static bool binder_txns_pending_ilocked(struct binder_proc *proc)
 	return false;
 }
 
+static void binder_add_freeze_work(struct binder_proc *proc, bool is_frozen)
+{
+	struct rb_node *n;
+	struct binder_ref *ref;
+
+	binder_inner_proc_lock(proc);
+	for (n = rb_first(&proc->nodes); n; n = rb_next(n)) {
+		struct binder_node *node;
+
+		node = rb_entry(n, struct binder_node, rb_node);
+		binder_inner_proc_unlock(proc);
+		binder_node_lock(node);
+		hlist_for_each_entry(ref, &node->refs, node_entry) {
+			/*
+			 * Need the node lock to synchronize
+			 * with new notification requests and the
+			 * inner lock to synchronize with queued
+			 * freeze notifications.
+			 */
+			binder_inner_proc_lock(ref->proc);
+			if (!ref->freeze) {
+				binder_inner_proc_unlock(ref->proc);
+				continue;
+			}
+			ref->freeze->work.type = BINDER_WORK_FROZEN_BINDER;
+			if (list_empty(&ref->freeze->work.entry)) {
+				ref->freeze->is_frozen = is_frozen;
+				binder_enqueue_work_ilocked(&ref->freeze->work, &ref->proc->todo);
+				binder_wakeup_proc_ilocked(ref->proc);
+			} else {
+				if (ref->freeze->sent && ref->freeze->is_frozen != is_frozen)
+					ref->freeze->resend = true;
+				ref->freeze->is_frozen = is_frozen;
+			}
+			binder_inner_proc_unlock(ref->proc);
+		}
+		binder_node_unlock(node);
+		binder_inner_proc_lock(proc);
+	}
+	binder_inner_proc_unlock(proc);
+}
+
 static int binder_ioctl_freeze(struct binder_freeze_info *info,
 			       struct binder_proc *target_proc)
 {
@@ -5253,6 +5523,7 @@ static int binder_ioctl_freeze(struct binder_freeze_info *info,
 		target_proc->async_recv = false;
 		target_proc->is_frozen = false;
 		binder_inner_proc_unlock(target_proc);
+		binder_add_freeze_work(target_proc, false);
 		return 0;
 	}
 
@@ -5285,6 +5556,8 @@ static int binder_ioctl_freeze(struct binder_freeze_info *info,
 		binder_inner_proc_lock(target_proc);
 		target_proc->is_frozen = false;
 		binder_inner_proc_unlock(target_proc);
+	} else {
+		binder_add_freeze_work(target_proc, true);
 	}
 
 	return ret;
@@ -5658,6 +5931,7 @@ static int binder_open(struct inode *nodp, struct file *filp)
 	binder_stats_created(BINDER_STAT_PROC);
 	proc->pid = current->group_leader->pid;
 	INIT_LIST_HEAD(&proc->delivered_death);
+	INIT_LIST_HEAD(&proc->delivered_freeze);
 	INIT_LIST_HEAD(&proc->waiting_threads);
 	filp->private_data = proc;
 
@@ -6209,7 +6483,9 @@ static const char * const binder_return_strings[] = {
 	"BR_FAILED_REPLY",
 	"BR_FROZEN_REPLY",
 	"BR_ONEWAY_SPAM_SUSPECT",
-	"BR_TRANSACTION_PENDING_FROZEN"
+	"BR_TRANSACTION_PENDING_FROZEN",
+	"BR_FROZEN_BINDER",
+	"BR_CLEAR_FREEZE_NOTIFICATION_DONE",
 };
 
 static const char * const binder_command_strings[] = {
@@ -6232,6 +6508,9 @@ static const char * const binder_command_strings[] = {
 	"BC_DEAD_BINDER_DONE",
 	"BC_TRANSACTION_SG",
 	"BC_REPLY_SG",
+	"BC_REQUEST_FREEZE_NOTIFICATION",
+	"BC_CLEAR_FREEZE_NOTIFICATION",
+	"BC_FREEZE_NOTIFICATION_DONE",
 };
 
 static const char * const binder_objstat_strings[] = {
@@ -6241,7 +6520,8 @@ static const char * const binder_objstat_strings[] = {
 	"ref",
 	"death",
 	"transaction",
-	"transaction_complete"
+	"transaction_complete",
+	"freeze",
 };
 
 static void print_binder_stats(struct seq_file *m, const char *prefix,
drivers/android/binder_internal.h +19 -2
index 5b7c80b99ae8..3e4b35873c64 100644
@@ -129,12 +129,13 @@ enum binder_stat_types {
 	BINDER_STAT_DEATH,
 	BINDER_STAT_TRANSACTION,
 	BINDER_STAT_TRANSACTION_COMPLETE,
+	BINDER_STAT_FREEZE,
 	BINDER_STAT_COUNT
 };
 
 struct binder_stats {
-	atomic_t br[_IOC_NR(BR_TRANSACTION_PENDING_FROZEN) + 1];
-	atomic_t bc[_IOC_NR(BC_REPLY_SG) + 1];
+	atomic_t br[_IOC_NR(BR_CLEAR_FREEZE_NOTIFICATION_DONE) + 1];
+	atomic_t bc[_IOC_NR(BC_FREEZE_NOTIFICATION_DONE) + 1];
 	atomic_t obj_created[BINDER_STAT_COUNT];
 	atomic_t obj_deleted[BINDER_STAT_COUNT];
 };
@@ -159,6 +160,8 @@ struct binder_work {
 		BINDER_WORK_DEAD_BINDER,
 		BINDER_WORK_DEAD_BINDER_AND_CLEAR,
 		BINDER_WORK_CLEAR_DEATH_NOTIFICATION,
+		BINDER_WORK_FROZEN_BINDER,
+		BINDER_WORK_CLEAR_FREEZE_NOTIFICATION,
 	} type;
 };
 
@@ -275,6 +278,14 @@ struct binder_ref_death {
 	binder_uintptr_t cookie;
 };
 
+struct binder_ref_freeze {
+	struct binder_work work;
+	binder_uintptr_t cookie;
+	bool is_frozen:1;
+	bool sent:1;
+	bool resend:1;
+};
+
 /**
  * struct binder_ref_data - binder_ref counts and id
  * @debug_id:        unique ID for the ref
@@ -307,6 +318,8 @@ struct binder_ref_data {
  *               @node indicates the node must be freed
  * @death:       pointer to death notification (ref_death) if requested
  *               (protected by @node->lock)
+ * @freeze:      pointer to freeze notification (ref_freeze) if requested
+ *               (protected by @node->lock)
  *
  * Structure to track references from procA to target node (on procB). This
  * structure is unsafe to access without holding @proc->outer_lock.
@@ -323,6 +336,7 @@ struct binder_ref {
 	struct binder_proc *proc;
 	struct binder_node *node;
 	struct binder_ref_death *death;
+	struct binder_ref_freeze *freeze;
 };
 
 /**
@@ -374,6 +388,8 @@ struct binder_ref {
  *                        (atomics, no lock needed)
  * @delivered_death:      list of delivered death notification
  *                        (protected by @inner_lock)
+ * @delivered_freeze:     list of delivered freeze notification
+ *                        (protected by @inner_lock)
  * @max_threads:          cap on number of binder threads
  *                        (protected by @inner_lock)
  * @requested_threads:    number of binder threads requested but not
@@ -421,6 +437,7 @@ struct binder_proc {
 	struct list_head todo;
 	struct binder_stats stats;
 	struct list_head delivered_death;
+	struct list_head delivered_freeze;
 	u32 max_threads;
 	int requested_threads;
 	int requested_threads_started;
drivers/android/binderfs.c +8 -0
index 3001d754ac36..ad1fa7abc323 100644
@@ -58,6 +58,7 @@ enum binderfs_stats_mode {
 struct binder_features {
 	bool oneway_spam_detection;
 	bool extended_error;
+	bool freeze_notification;
 };
 
 static const struct constant_table binderfs_param_stats[] = {
@@ -74,6 +75,7 @@ static const struct fs_parameter_spec binderfs_fs_parameters[] = {
 static struct binder_features binder_features = {
 	.oneway_spam_detection = true,
 	.extended_error = true,
+	.freeze_notification = true,
 };
 
 static inline struct binderfs_info *BINDERFS_SB(const struct super_block *sb)
@@ -608,6 +610,12 @@ static int init_binder_features(struct super_block *sb)
 	if (IS_ERR(dentry))
 		return PTR_ERR(dentry);
 
+	dentry = binderfs_create_file(dir, "freeze_notification",
+				      &binder_features_fops,
+				      &binder_features.freeze_notification);
+	if (IS_ERR(dentry))
+		return PTR_ERR(dentry);
+
 	return 0;
 }
 
include/uapi/linux/android/binder.h +36 -0
index d44a8118b2ed..1fd92021a573 100644
@@ -236,6 +236,12 @@ struct binder_frozen_status_info {
 	__u32            async_recv;
 };
 
+struct binder_frozen_state_info {
+	binder_uintptr_t cookie;
+	__u32            is_frozen;
+	__u32            reserved;
+};
+
 /* struct binder_extened_error - extended error information
  * @id:		identifier for the failed operation
  * @command:	command as defined by binder_driver_return_protocol
@@ -467,6 +473,17 @@ enum binder_driver_return_protocol {
 	/*
 	 * The target of the last async transaction is frozen.  No parameters.
 	 */
+
+	BR_FROZEN_BINDER = _IOR('r', 21, struct binder_frozen_state_info),
+	/*
+	 * The cookie and a boolean (is_frozen) that indicates whether the process
+	 * transitioned into a frozen or an unfrozen state.
+	 */
+
+	BR_CLEAR_FREEZE_NOTIFICATION_DONE = _IOR('r', 22, binder_uintptr_t),
+	/*
+	 * void *: cookie
+	 */
 };
 
 enum binder_driver_command_protocol {
@@ -550,6 +567,25 @@ enum binder_driver_command_protocol {
 	/*
 	 * binder_transaction_data_sg: the sent command.
 	 */
+
+	BC_REQUEST_FREEZE_NOTIFICATION =
+			_IOW('c', 19, struct binder_handle_cookie),
+	/*
+	 * int: handle
+	 * void *: cookie
+	 */
+
+	BC_CLEAR_FREEZE_NOTIFICATION = _IOW('c', 20,
+					    struct binder_handle_cookie),
+	/*
+	 * int: handle
+	 * void *: cookie
+	 */
+
+	BC_FREEZE_NOTIFICATION_DONE = _IOW('c', 21, binder_uintptr_t),
+	/*
+	 * void *: cookie
+	 */
 };
 
 #endif /* _UAPI_LINUX_BINDER_H */
tools/testing/selftests/filesystems/binderfs/binderfs_test.c +1 -0
index 5f362c0fd890..319567f0fae1 100644
@@ -65,6 +65,7 @@ static int __do_binderfs_test(struct __test_metadata *_metadata)
 	static const char * const binder_features[] = {
 		"oneway_spam_detection",
 		"extended_error",
+		"freeze_notification",
 	};
 
 	change_mountns(_metadata);
Signature · 1 line
2.45.2.803.g4e1b14247a-goog
#16 lore raw

v6 cover binder: frozen notification

In reply to Yu-Ting Tseng

Separated the binder_features change into its own patch.

Yu-Ting Tseng (2):
binder: frozen notification
binder: frozen notification binder_features flag

 drivers/android/binder.c                      | 284 +++++++++++++++++-
 drivers/android/binder_internal.h             |  21 +-
 drivers/android/binderfs.c                    |   8 +
 include/uapi/linux/android/binder.h           |  36 +++
 .../filesystems/binderfs/binderfs_test.c      |   1 +
 5 files changed, 346 insertions(+), 4 deletions(-)

base-commit: 14d7c92f8df9c0964ae6f8b813c1b3ac38120825

Signature · 1 line
2.45.2.803.g4e1b14247a-goog
maintainer
#17 lore raw
Quoted 14 lines · Yu-Ting Tseng

Separated the binder_features change into its own patch.

Yu-Ting Tseng (2):
binder: frozen notification
binder: frozen notification binder_features flag

 drivers/android/binder.c                      | 284 +++++++++++++++++-
 drivers/android/binder_internal.h             |  21 +-
 drivers/android/binderfs.c                    |   8 +
 include/uapi/linux/android/binder.h           |  36 +++
 .../filesystems/binderfs/binderfs_test.c      |   1 +
 5 files changed, 346 insertions(+), 4 deletions(-)
Quoted 1 line

base-commit: 14d7c92f8df9c0964ae6f8b813c1b3ac38120825

This email thread is crazy, sorry, look at how it shows up in sane email clients:

	https://lore.kernel.org/all/[email protected]/

Please just send a new series, as it's own thread, so we can figure out what to do here.

And for some reason I have 2 different copies of the v6 series, one sent to the list, and one not, which makes no sense either. For that reason alone I can't take this :(

Please, realize what this looks like from a reviewer's point of view, you want to make it simple and easy for us to know what to do...

Please fix up and send a v7.

thanks,

greg k-h

#18 lore raw

PATCH v6 1/2 binder: frozen notification

In reply to Yu-Ting Tseng

Frozen processes present a significant challenge in binder transactions. When a process is frozen, it cannot, by design, accept and/or respond to binder transactions. As a result, the sender needs to adjust its behavior, such as postponing transactions until the peer process unfreezes. However, there is currently no way to subscribe to these state change events, making it impossible to implement frozen-aware behaviors efficiently.

Introduce a binder API for subscribing to frozen state change events. This allows programs to react to changes in peer process state, mitigating issues related to binder transactions sent to frozen processes.

Implementation details:
For a given binder_ref, the state of frozen notification can be one of the followings:
1. Userspace doesn't want a notification. binder_ref->freeze is null.
2. Userspace wants a notification but none is in flight. list_empty(&binder_ref->freeze->work.entry) = true
3. A notification is in flight and waiting to be read by userspace. binder_ref_freeze.sent is false.
4. A notification was read by userspace and kernel is waiting for an ack. binder_ref_freeze.sent is true.

When a notification is in flight, new state change events are coalesced into the existing binder_ref_freeze struct. If userspace hasn't picked up the notification yet, the driver simply rewrites the state. Otherwise, the notification is flagged as requiring a resend, which will be performed once userspace acks the original notification that's inflight.

See https://r.android.com/3070045 for how userspace is going to use this feature.

---
 drivers/android/binder.c            | 284 +++++++++++++++++++++++++++-
 drivers/android/binder_internal.h   |  21 +-
 include/uapi/linux/android/binder.h |  36 ++++
 3 files changed, 337 insertions(+), 4 deletions(-)
drivers/android/binder.c +282 -2
index b21a7b246a0d..77f318127a6e 100644
@@ -1295,6 +1295,7 @@ static void binder_free_ref(struct binder_ref *ref)
 	if (ref->node)
 		binder_free_node(ref->node);
 	kfree(ref->death);
+	kfree(ref->freeze);
 	kfree(ref);
 }
 
@@ -3763,6 +3764,155 @@ static void binder_transaction(struct binder_proc *proc,
 	}
 }
 
+static int
+binder_request_freeze_notification(struct binder_proc *proc,
+				   struct binder_thread *thread,
+				   struct binder_handle_cookie *handle_cookie)
+{
+	struct binder_ref_freeze *freeze;
+	struct binder_ref *ref;
+	bool is_frozen;
+
+	freeze = kzalloc(sizeof(*freeze), GFP_KERNEL);
+	if (!freeze)
+		return -ENOMEM;
+	binder_proc_lock(proc);
+	ref = binder_get_ref_olocked(proc, handle_cookie->handle, false);
+	if (!ref) {
+		binder_user_error("%d:%d BC_REQUEST_FREEZE_NOTIFICATION invalid ref %d\n",
+				  proc->pid, thread->pid, handle_cookie->handle);
+		binder_proc_unlock(proc);
+		kfree(freeze);
+		return -EINVAL;
+	}
+
+	binder_node_lock(ref->node);
+
+	if (ref->freeze || !ref->node->proc) {
+		binder_user_error("%d:%d invalid BC_REQUEST_FREEZE_NOTIFICATION %s\n",
+				  proc->pid, thread->pid,
+				  ref->freeze ? "already set" : "dead node");
+		binder_node_unlock(ref->node);
+		binder_proc_unlock(proc);
+		kfree(freeze);
+		return -EINVAL;
+	}
+	binder_inner_proc_lock(ref->node->proc);
+	is_frozen = ref->node->proc->is_frozen;
+	binder_inner_proc_unlock(ref->node->proc);
+
+	binder_stats_created(BINDER_STAT_FREEZE);
+	INIT_LIST_HEAD(&freeze->work.entry);
+	freeze->cookie = handle_cookie->cookie;
+	freeze->work.type = BINDER_WORK_FROZEN_BINDER;
+	freeze->is_frozen = is_frozen;
+
+	ref->freeze = freeze;
+
+	binder_inner_proc_lock(proc);
+	binder_enqueue_work_ilocked(&ref->freeze->work, &proc->todo);
+	binder_wakeup_proc_ilocked(proc);
+	binder_inner_proc_unlock(proc);
+
+	binder_node_unlock(ref->node);
+	binder_proc_unlock(proc);
+	return 0;
+}
+
+static int
+binder_clear_freeze_notification(struct binder_proc *proc,
+				 struct binder_thread *thread,
+				 struct binder_handle_cookie *handle_cookie)
+{
+	struct binder_ref_freeze *freeze;
+	struct binder_ref *ref;
+
+	binder_proc_lock(proc);
+	ref = binder_get_ref_olocked(proc, handle_cookie->handle, false);
+	if (!ref) {
+		binder_user_error("%d:%d BC_CLEAR_FREEZE_NOTIFICATION invalid ref %d\n",
+				  proc->pid, thread->pid, handle_cookie->handle);
+		binder_proc_unlock(proc);
+		return -EINVAL;
+	}
+
+	binder_node_lock(ref->node);
+
+	if (!ref->freeze) {
+		binder_user_error("%d:%d BC_CLEAR_FREEZE_NOTIFICATION freeze notification not active\n",
+				  proc->pid, thread->pid);
+		binder_node_unlock(ref->node);
+		binder_proc_unlock(proc);
+		return -EINVAL;
+	}
+	freeze = ref->freeze;
+	binder_inner_proc_lock(proc);
+	if (freeze->cookie != handle_cookie->cookie) {
+		binder_user_error("%d:%d BC_CLEAR_FREEZE_NOTIFICATION freeze notification cookie mismatch %016llx != %016llx\n",
+				  proc->pid, thread->pid, (u64)freeze->cookie,
+				  (u64)handle_cookie->cookie);
+		binder_inner_proc_unlock(proc);
+		binder_node_unlock(ref->node);
+		binder_proc_unlock(proc);
+		return -EINVAL;
+	}
+	ref->freeze = NULL;
+	/*
+	 * Take the existing freeze object and overwrite its work type. There are three cases here:
+	 * 1. No pending notification. In this case just add the work to the queue.
+	 * 2. A notification was sent and is pending an ack from userspace. Once an ack arrives, we
+	 *    should resend with the new work type.
+	 * 3. A notification is pending to be sent. Since the work is already in the queue, nothing
+	 *    needs to be done here.
+	 */
+	freeze->work.type = BINDER_WORK_CLEAR_FREEZE_NOTIFICATION;
+	if (list_empty(&freeze->work.entry)) {
+		binder_enqueue_work_ilocked(&freeze->work, &proc->todo);
+		binder_wakeup_proc_ilocked(proc);
+	} else if (freeze->sent) {
+		freeze->resend = true;
+	}
+	binder_inner_proc_unlock(proc);
+	binder_node_unlock(ref->node);
+	binder_proc_unlock(proc);
+	return 0;
+}
+
+static int
+binder_freeze_notification_done(struct binder_proc *proc,
+				struct binder_thread *thread,
+				binder_uintptr_t cookie)
+{
+	struct binder_ref_freeze *freeze = NULL;
+	struct binder_work *w;
+
+	binder_inner_proc_lock(proc);
+	list_for_each_entry(w, &proc->delivered_freeze, entry) {
+		struct binder_ref_freeze *tmp_freeze =
+			container_of(w, struct binder_ref_freeze, work);
+
+		if (tmp_freeze->cookie == cookie) {
+			freeze = tmp_freeze;
+			break;
+		}
+	}
+	if (!freeze) {
+		binder_user_error("%d:%d BC_FREEZE_NOTIFICATION_DONE %016llx not found\n",
+				  proc->pid, thread->pid, (u64)cookie);
+		binder_inner_proc_unlock(proc);
+		return -EINVAL;
+	}
+	binder_dequeue_work_ilocked(&freeze->work);
+	freeze->sent = false;
+	if (freeze->resend) {
+		freeze->resend = false;
+		binder_enqueue_work_ilocked(&freeze->work, &proc->todo);
+		binder_wakeup_proc_ilocked(proc);
+	}
+	binder_inner_proc_unlock(proc);
+	return 0;
+}
+
 /**
  * binder_free_buf() - free the specified buffer
  * @proc:	binder proc that owns buffer
@@ -4246,6 +4396,44 @@ static int binder_thread_write(struct binder_proc *proc,
 			binder_inner_proc_unlock(proc);
 		} break;
 
+		case BC_REQUEST_FREEZE_NOTIFICATION: {
+			struct binder_handle_cookie handle_cookie;
+			int error;
+
+			if (copy_from_user(&handle_cookie, ptr, sizeof(handle_cookie)))
+				return -EFAULT;
+			ptr += sizeof(handle_cookie);
+			error = binder_request_freeze_notification(proc, thread,
+								   &handle_cookie);
+			if (error)
+				return error;
+		} break;
+
+		case BC_CLEAR_FREEZE_NOTIFICATION: {
+			struct binder_handle_cookie handle_cookie;
+			int error;
+
+			if (copy_from_user(&handle_cookie, ptr, sizeof(handle_cookie)))
+				return -EFAULT;
+			ptr += sizeof(handle_cookie);
+			error = binder_clear_freeze_notification(proc, thread, &handle_cookie);
+			if (error)
+				return error;
+		} break;
+
+		case BC_FREEZE_NOTIFICATION_DONE: {
+			binder_uintptr_t cookie;
+			int error;
+
+			if (get_user(cookie, (binder_uintptr_t __user *)ptr))
+				return -EFAULT;
+
+			ptr += sizeof(cookie);
+			error = binder_freeze_notification_done(proc, thread, cookie);
+			if (error)
+				return error;
+		} break;
+
 		default:
 			pr_err("%d:%d unknown command %u\n",
 			       proc->pid, thread->pid, cmd);
@@ -4635,6 +4823,46 @@ static int binder_thread_read(struct binder_proc *proc,
 			if (cmd == BR_DEAD_BINDER)
 				goto done; /* DEAD_BINDER notifications can cause transactions */
 		} break;
+
+		case BINDER_WORK_FROZEN_BINDER: {
+			struct binder_ref_freeze *freeze;
+			struct binder_frozen_state_info info;
+
+			memset(&info, 0, sizeof(info));
+			freeze = container_of(w, struct binder_ref_freeze, work);
+			info.is_frozen = freeze->is_frozen;
+			info.cookie = freeze->cookie;
+			freeze->sent = true;
+			binder_enqueue_work_ilocked(w, &proc->delivered_freeze);
+			binder_inner_proc_unlock(proc);
+
+			if (put_user(BR_FROZEN_BINDER, (uint32_t __user *)ptr))
+				return -EFAULT;
+			ptr += sizeof(uint32_t);
+			if (copy_to_user(ptr, &info, sizeof(info)))
+				return -EFAULT;
+			ptr += sizeof(info);
+			binder_stat_br(proc, thread, BR_FROZEN_BINDER);
+			goto done; /* BR_FROZEN_BINDER notifications can cause transactions */
+		} break;
+
+		case BINDER_WORK_CLEAR_FREEZE_NOTIFICATION: {
+			struct binder_ref_freeze *freeze =
+			    container_of(w, struct binder_ref_freeze, work);
+			binder_uintptr_t cookie = freeze->cookie;
+
+			binder_inner_proc_unlock(proc);
+			kfree(freeze);
+			binder_stats_deleted(BINDER_STAT_FREEZE);
+			if (put_user(BR_CLEAR_FREEZE_NOTIFICATION_DONE, (uint32_t __user *)ptr))
+				return -EFAULT;
+			ptr += sizeof(uint32_t);
+			if (put_user(cookie, (binder_uintptr_t __user *)ptr))
+				return -EFAULT;
+			ptr += sizeof(binder_uintptr_t);
+			binder_stat_br(proc, thread, BR_CLEAR_FREEZE_NOTIFICATION_DONE);
+		} break;
+
 		default:
 			binder_inner_proc_unlock(proc);
 			pr_err("%d:%d: bad work type %d\n",
@@ -5242,6 +5470,48 @@ static bool binder_txns_pending_ilocked(struct binder_proc *proc)
 	return false;
 }
 
+static void binder_add_freeze_work(struct binder_proc *proc, bool is_frozen)
+{
+	struct rb_node *n;
+	struct binder_ref *ref;
+
+	binder_inner_proc_lock(proc);
+	for (n = rb_first(&proc->nodes); n; n = rb_next(n)) {
+		struct binder_node *node;
+
+		node = rb_entry(n, struct binder_node, rb_node);
+		binder_inner_proc_unlock(proc);
+		binder_node_lock(node);
+		hlist_for_each_entry(ref, &node->refs, node_entry) {
+			/*
+			 * Need the node lock to synchronize
+			 * with new notification requests and the
+			 * inner lock to synchronize with queued
+			 * freeze notifications.
+			 */
+			binder_inner_proc_lock(ref->proc);
+			if (!ref->freeze) {
+				binder_inner_proc_unlock(ref->proc);
+				continue;
+			}
+			ref->freeze->work.type = BINDER_WORK_FROZEN_BINDER;
+			if (list_empty(&ref->freeze->work.entry)) {
+				ref->freeze->is_frozen = is_frozen;
+				binder_enqueue_work_ilocked(&ref->freeze->work, &ref->proc->todo);
+				binder_wakeup_proc_ilocked(ref->proc);
+			} else {
+				if (ref->freeze->sent && ref->freeze->is_frozen != is_frozen)
+					ref->freeze->resend = true;
+				ref->freeze->is_frozen = is_frozen;
+			}
+			binder_inner_proc_unlock(ref->proc);
+		}
+		binder_node_unlock(node);
+		binder_inner_proc_lock(proc);
+	}
+	binder_inner_proc_unlock(proc);
+}
+
 static int binder_ioctl_freeze(struct binder_freeze_info *info,
 			       struct binder_proc *target_proc)
 {
@@ -5253,6 +5523,7 @@ static int binder_ioctl_freeze(struct binder_freeze_info *info,
 		target_proc->async_recv = false;
 		target_proc->is_frozen = false;
 		binder_inner_proc_unlock(target_proc);
+		binder_add_freeze_work(target_proc, false);
 		return 0;
 	}
 
@@ -5285,6 +5556,8 @@ static int binder_ioctl_freeze(struct binder_freeze_info *info,
 		binder_inner_proc_lock(target_proc);
 		target_proc->is_frozen = false;
 		binder_inner_proc_unlock(target_proc);
+	} else {
+		binder_add_freeze_work(target_proc, true);
 	}
 
 	return ret;
@@ -5658,6 +5931,7 @@ static int binder_open(struct inode *nodp, struct file *filp)
 	binder_stats_created(BINDER_STAT_PROC);
 	proc->pid = current->group_leader->pid;
 	INIT_LIST_HEAD(&proc->delivered_death);
+	INIT_LIST_HEAD(&proc->delivered_freeze);
 	INIT_LIST_HEAD(&proc->waiting_threads);
 	filp->private_data = proc;
 
@@ -6209,7 +6483,9 @@ static const char * const binder_return_strings[] = {
 	"BR_FAILED_REPLY",
 	"BR_FROZEN_REPLY",
 	"BR_ONEWAY_SPAM_SUSPECT",
-	"BR_TRANSACTION_PENDING_FROZEN"
+	"BR_TRANSACTION_PENDING_FROZEN",
+	"BR_FROZEN_BINDER",
+	"BR_CLEAR_FREEZE_NOTIFICATION_DONE",
 };
 
 static const char * const binder_command_strings[] = {
@@ -6232,6 +6508,9 @@ static const char * const binder_command_strings[] = {
 	"BC_DEAD_BINDER_DONE",
 	"BC_TRANSACTION_SG",
 	"BC_REPLY_SG",
+	"BC_REQUEST_FREEZE_NOTIFICATION",
+	"BC_CLEAR_FREEZE_NOTIFICATION",
+	"BC_FREEZE_NOTIFICATION_DONE",
 };
 
 static const char * const binder_objstat_strings[] = {
@@ -6241,7 +6520,8 @@ static const char * const binder_objstat_strings[] = {
 	"ref",
 	"death",
 	"transaction",
-	"transaction_complete"
+	"transaction_complete",
+	"freeze",
 };
 
 static void print_binder_stats(struct seq_file *m, const char *prefix,
drivers/android/binder_internal.h +19 -2
index 5b7c80b99ae8..3e4b35873c64 100644
@@ -129,12 +129,13 @@ enum binder_stat_types {
 	BINDER_STAT_DEATH,
 	BINDER_STAT_TRANSACTION,
 	BINDER_STAT_TRANSACTION_COMPLETE,
+	BINDER_STAT_FREEZE,
 	BINDER_STAT_COUNT
 };
 
 struct binder_stats {
-	atomic_t br[_IOC_NR(BR_TRANSACTION_PENDING_FROZEN) + 1];
-	atomic_t bc[_IOC_NR(BC_REPLY_SG) + 1];
+	atomic_t br[_IOC_NR(BR_CLEAR_FREEZE_NOTIFICATION_DONE) + 1];
+	atomic_t bc[_IOC_NR(BC_FREEZE_NOTIFICATION_DONE) + 1];
 	atomic_t obj_created[BINDER_STAT_COUNT];
 	atomic_t obj_deleted[BINDER_STAT_COUNT];
 };
@@ -159,6 +160,8 @@ struct binder_work {
 		BINDER_WORK_DEAD_BINDER,
 		BINDER_WORK_DEAD_BINDER_AND_CLEAR,
 		BINDER_WORK_CLEAR_DEATH_NOTIFICATION,
+		BINDER_WORK_FROZEN_BINDER,
+		BINDER_WORK_CLEAR_FREEZE_NOTIFICATION,
 	} type;
 };
 
@@ -275,6 +278,14 @@ struct binder_ref_death {
 	binder_uintptr_t cookie;
 };
 
+struct binder_ref_freeze {
+	struct binder_work work;
+	binder_uintptr_t cookie;
+	bool is_frozen:1;
+	bool sent:1;
+	bool resend:1;
+};
+
 /**
  * struct binder_ref_data - binder_ref counts and id
  * @debug_id:        unique ID for the ref
@@ -307,6 +318,8 @@ struct binder_ref_data {
  *               @node indicates the node must be freed
  * @death:       pointer to death notification (ref_death) if requested
  *               (protected by @node->lock)
+ * @freeze:      pointer to freeze notification (ref_freeze) if requested
+ *               (protected by @node->lock)
  *
  * Structure to track references from procA to target node (on procB). This
  * structure is unsafe to access without holding @proc->outer_lock.
@@ -323,6 +336,7 @@ struct binder_ref {
 	struct binder_proc *proc;
 	struct binder_node *node;
 	struct binder_ref_death *death;
+	struct binder_ref_freeze *freeze;
 };
 
 /**
@@ -374,6 +388,8 @@ struct binder_ref {
  *                        (atomics, no lock needed)
  * @delivered_death:      list of delivered death notification
  *                        (protected by @inner_lock)
+ * @delivered_freeze:     list of delivered freeze notification
+ *                        (protected by @inner_lock)
  * @max_threads:          cap on number of binder threads
  *                        (protected by @inner_lock)
  * @requested_threads:    number of binder threads requested but not
@@ -421,6 +437,7 @@ struct binder_proc {
 	struct list_head todo;
 	struct binder_stats stats;
 	struct list_head delivered_death;
+	struct list_head delivered_freeze;
 	u32 max_threads;
 	int requested_threads;
 	int requested_threads_started;
include/uapi/linux/android/binder.h +36 -0
index d44a8118b2ed..1fd92021a573 100644
@@ -236,6 +236,12 @@ struct binder_frozen_status_info {
 	__u32            async_recv;
 };
 
+struct binder_frozen_state_info {
+	binder_uintptr_t cookie;
+	__u32            is_frozen;
+	__u32            reserved;
+};
+
 /* struct binder_extened_error - extended error information
  * @id:		identifier for the failed operation
  * @command:	command as defined by binder_driver_return_protocol
@@ -467,6 +473,17 @@ enum binder_driver_return_protocol {
 	/*
 	 * The target of the last async transaction is frozen.  No parameters.
 	 */
+
+	BR_FROZEN_BINDER = _IOR('r', 21, struct binder_frozen_state_info),
+	/*
+	 * The cookie and a boolean (is_frozen) that indicates whether the process
+	 * transitioned into a frozen or an unfrozen state.
+	 */
+
+	BR_CLEAR_FREEZE_NOTIFICATION_DONE = _IOR('r', 22, binder_uintptr_t),
+	/*
+	 * void *: cookie
+	 */
 };
 
 enum binder_driver_command_protocol {
@@ -550,6 +567,25 @@ enum binder_driver_command_protocol {
 	/*
 	 * binder_transaction_data_sg: the sent command.
 	 */
+
+	BC_REQUEST_FREEZE_NOTIFICATION =
+			_IOW('c', 19, struct binder_handle_cookie),
+	/*
+	 * int: handle
+	 * void *: cookie
+	 */
+
+	BC_CLEAR_FREEZE_NOTIFICATION = _IOW('c', 20,
+					    struct binder_handle_cookie),
+	/*
+	 * int: handle
+	 * void *: cookie
+	 */
+
+	BC_FREEZE_NOTIFICATION_DONE = _IOW('c', 21, binder_uintptr_t),
+	/*
+	 * void *: cookie
+	 */
 };
 
 #endif /* _UAPI_LINUX_BINDER_H */
Signature · 1 line
2.45.2.803.g4e1b14247a-goog
#19 lore raw

PATCH v6 2/2 binder: frozen notification binder_features flag

In reply to Yu-Ting Tseng

Add a flag to binder_features to indicate that the freeze notification feature is available.

---
 drivers/android/binderfs.c                                | 8 ++++++++
 .../selftests/filesystems/binderfs/binderfs_test.c        | 1 +
 2 files changed, 9 insertions(+)
drivers/android/binderfs.c +8 -0
index 3001d754ac36..ad1fa7abc323 100644
@@ -58,6 +58,7 @@ enum binderfs_stats_mode {
 struct binder_features {
 	bool oneway_spam_detection;
 	bool extended_error;
+	bool freeze_notification;
 };
 
 static const struct constant_table binderfs_param_stats[] = {
@@ -74,6 +75,7 @@ static const struct fs_parameter_spec binderfs_fs_parameters[] = {
 static struct binder_features binder_features = {
 	.oneway_spam_detection = true,
 	.extended_error = true,
+	.freeze_notification = true,
 };
 
 static inline struct binderfs_info *BINDERFS_SB(const struct super_block *sb)
@@ -608,6 +610,12 @@ static int init_binder_features(struct super_block *sb)
 	if (IS_ERR(dentry))
 		return PTR_ERR(dentry);
 
+	dentry = binderfs_create_file(dir, "freeze_notification",
+				      &binder_features_fops,
+				      &binder_features.freeze_notification);
+	if (IS_ERR(dentry))
+		return PTR_ERR(dentry);
+
 	return 0;
 }
 
tools/testing/selftests/filesystems/binderfs/binderfs_test.c +1 -0
index 5f362c0fd890..319567f0fae1 100644
@@ -65,6 +65,7 @@ static int __do_binderfs_test(struct __test_metadata *_metadata)
 	static const char * const binder_features[] = {
 		"oneway_spam_detection",
 		"extended_error",
+		"freeze_notification",
 	};
 
 	change_mountns(_metadata);
Signature · 1 line
2.45.2.803.g4e1b14247a-goog
maintainer
#20 lore raw
In reply to Yu-Ting Tseng
Quoted 62 lines · Yu-Ting Tseng

Frozen processes present a significant challenge in binder transactions. When a process is frozen, it cannot, by design, accept and/or respond to binder transactions. As a result, the sender needs to adjust its behavior, such as postponing transactions until the peer process unfreezes. However, there is currently no way to subscribe to these state change events, making it impossible to implement frozen-aware behaviors efficiently.

Introduce a binder API for subscribing to frozen state change events. This allows programs to react to changes in peer process state, mitigating issues related to binder transactions sent to frozen processes.

Implementation details:
For a given binder_ref, the state of frozen notification can be one of the followings:
1. Userspace doesn't want a notification. binder_ref->freeze is null.
2. Userspace wants a notification but none is in flight. list_empty(&binder_ref->freeze->work.entry) = true
3. A notification is in flight and waiting to be read by userspace. binder_ref_freeze.sent is false.
4. A notification was read by userspace and kernel is waiting for an ack. binder_ref_freeze.sent is true.

When a notification is in flight, new state change events are coalesced into the existing binder_ref_freeze struct. If userspace hasn't picked up the notification yet, the driver simply rewrites the state. Otherwise, the notification is flagged as requiring a resend, which will be performed once userspace acks the original notification that's inflight.

See https://android-review.googlesource.com/c/platform/frameworks/native/+/3070045
for how userspace is going to use this feature.

---
V1 -> V2: addressed review comments

 drivers/android/binder.c            | 294 +++++++++++++++++++++++++++-
 drivers/android/binder_internal.h   |  19 +-
 include/uapi/linux/android/binder.h |  35 ++++
 3 files changed, 344 insertions(+), 4 deletions(-)
drivers/android/binder.c +12 -0
index b21a7b246a0d..048f76e5dd35 100644
@@ -3763,6 +3763,168 @@ static void binder_transaction(struct binder_proc *proc,
 	}
 }
 
+static int
+binder_request_freeze_notification(struct binder_proc *proc,
+				   struct binder_thread *thread,
+				   struct binder_handle_cookie *handle_cookie)
+{
+	struct binder_ref_freeze *freeze;
+	struct binder_ref *ref;
+
+	/*
+	 * Allocate memory for freeze notification before taking lock.
+	 */
+	freeze = kzalloc(sizeof(*freeze), GFP_KERNEL);

drivers/android/binder.c static void binder_transaction(struct binder_proc *proc, · line 3787

Quoted 10 lines
drivers/android/binder.c +10 -0
+	if (!freeze) {
+		thread->return_error.cmd = BR_ERROR;
+		binder_enqueue_thread_work(thread, &thread->return_error.work);
+		binder_debug(BINDER_DEBUG_FAILED_TRANSACTION,
+			     "%d:%d BC_REQUEST_FREEZE_NOTIFICATION failed\n",
+			     proc->pid, thread->pid);
+		return -ENOMEM;
+	}
+	freeze->sent = false;
+	freeze->resend = false;

nit: freeze was allocated with kzalloc(), you could drop the "= false".

Quoted 16 lines
drivers/android/binder.c +16 -0
+
+	binder_proc_lock(proc);
+	ref = binder_get_ref_olocked(proc, handle_cookie->handle, false);
+	if (!ref) {
+		binder_user_error("%d:%d BC_REQUEST_FREEZE_NOTIFICATION invalid ref %d\n",
+				  proc->pid, thread->pid, handle_cookie->handle);
+		binder_proc_unlock(proc);
+		kfree(freeze);
+		return -EINVAL;
+	}
+
+	binder_node_lock(ref->node);
+
+	if (ref->freeze) {
+		binder_user_error("%d:%d BC_REQUEST_FREEZE_NOTIFICATION freeze notification already set\n",
+				  proc->pid, thread->pid);

drivers/android/binder.c static void binder_transaction(struct binder_proc *proc, · line 3813

Quoted 10 lines
drivers/android/binder.c +10 -0
+		binder_node_unlock(ref->node);
+		binder_proc_unlock(proc);
+		kfree(freeze);
+		return -EINVAL;
+	}
+	binder_stats_created(BINDER_STAT_FREEZE);
+	INIT_LIST_HEAD(&freeze->work.entry);
+	freeze->cookie = handle_cookie->cookie;
+	ref->freeze = freeze;
+	if (ref->node->proc) {

If !node->proc then process is dead. Do we really need to continue?

drivers/android/binder.c static void binder_transaction(struct binder_proc *proc, · line 3815

Quoted 2 lines
drivers/android/binder.c +2 -0
+		ref->freeze->work.type = BINDER_WORK_FROZEN_BINDER;
+		ref->freeze->is_frozen = ref->node->proc->is_frozen;

This access to node->proc->* doesn't seem safe. I don't see it being locked.

drivers/android/binder.c static void binder_transaction(struct binder_proc *proc, · line 3817

Quoted 2 lines
drivers/android/binder.c +2 -0
+		binder_inner_proc_lock(proc);
+		binder_enqueue_work_ilocked(&ref->freeze->work, &proc->todo);

Why do we queue this notification? Is it to get the current state back to userspace?

Quoted 118 lines
drivers/android/binder.c +111 -0
+		binder_wakeup_proc_ilocked(proc);
+		binder_inner_proc_unlock(proc);
+	}
+	binder_node_unlock(ref->node);
+	binder_proc_unlock(proc);
+	return 0;
+}
+
+static int
+binder_clear_freeze_notification(struct binder_proc *proc,
+				 struct binder_thread *thread,
+				 struct binder_handle_cookie *handle_cookie)
+{
+	struct binder_ref_freeze *freeze;
+	struct binder_ref *ref;
+
+	binder_proc_lock(proc);
+	ref = binder_get_ref_olocked(proc, handle_cookie->handle, false);
+	if (!ref) {
+		binder_user_error("%d:%d BC_CLEAR_FREEZE_NOTIFICATION invalid ref %d\n",
+				  proc->pid, thread->pid, handle_cookie->handle);
+		binder_proc_unlock(proc);
+		return -EINVAL;
+	}
+
+	binder_node_lock(ref->node);
+
+	if (!ref->freeze) {
+		binder_user_error("%d:%d BC_CLEAR_FREEZE_NOTIFICATION freeze notification not active\n",
+				  proc->pid, thread->pid);
+		binder_node_unlock(ref->node);
+		binder_proc_unlock(proc);
+		return -EINVAL;
+	}
+	freeze = ref->freeze;
+	if (freeze->cookie != handle_cookie->cookie) {
+		binder_user_error("%d:%d BC_CLEAR_FREEZE_NOTIFICATION freeze notification cookie mismatch %016llx != %016llx\n",
+				  proc->pid, thread->pid, (u64)freeze->cookie,
+				  (u64)handle_cookie->cookie);
+		binder_node_unlock(ref->node);
+		binder_proc_unlock(proc);
+		return -EINVAL;
+	}
+	ref->freeze = NULL;
+	binder_inner_proc_lock(proc);
+	if (list_empty(&freeze->work.entry)) {
+		freeze->work.type = BINDER_WORK_CLEAR_FREEZE_NOTIFICATION;
+		if (thread->looper &
+			(BINDER_LOOPER_STATE_REGISTERED | BINDER_LOOPER_STATE_ENTERED)) {
+			binder_enqueue_thread_work_ilocked(thread, &freeze->work);
+		} else {
+			binder_enqueue_work_ilocked(&freeze->work, &proc->todo);
+			binder_wakeup_proc_ilocked(proc);
+		}
+	} else {
+		/*
+		 * There is already a freeze notification. Take it over and rewrite
+		 * the work type. If it was already sent, flag it for re-sending;
+		 * Otherwise it's pending and will be sent soon.
+		 */
+		freeze->work.type = BINDER_WORK_CLEAR_DEATH_NOTIFICATION;
+		if (freeze->sent)
+			freeze->resend = true;
+	}
+	binder_inner_proc_unlock(proc);
+	binder_node_unlock(ref->node);
+	binder_proc_unlock(proc);
+	return 0;
+}
+
+static int
+binder_freeze_notification_done(struct binder_proc *proc,
+				struct binder_thread *thread,
+				binder_uintptr_t cookie)
+{
+	struct binder_ref_freeze *freeze = NULL;
+	struct binder_work *w;
+
+	binder_inner_proc_lock(proc);
+	list_for_each_entry(w, &proc->delivered_freeze, entry) {
+		struct binder_ref_freeze *tmp_freeze =
+			container_of(w, struct binder_ref_freeze, work);
+
+		if (tmp_freeze->cookie == cookie) {
+			freeze = tmp_freeze;
+			break;
+		}
+	}
+	if (!freeze) {
+		binder_user_error("%d:%d BC_FREEZE_NOTIFICATION_DONE %016llx not found\n",
+				  proc->pid, thread->pid, (u64)cookie);
+		binder_inner_proc_unlock(proc);
+		return -EINVAL;
+	}
+	binder_dequeue_work_ilocked(&freeze->work);
+	freeze->sent = false;
+	if (freeze->resend) {
+		freeze->resend = false;
+		if (thread->looper &
+			(BINDER_LOOPER_STATE_REGISTERED | BINDER_LOOPER_STATE_ENTERED)) {
+			binder_enqueue_thread_work_ilocked(thread, &freeze->work);
+		} else {
+			binder_enqueue_work_ilocked(&freeze->work, &proc->todo);
+			binder_wakeup_proc_ilocked(proc);
+		}
+	}
+	binder_inner_proc_unlock(proc);
+	return 0;
+}
+
 /**
  * binder_free_buf() - free the specified buffer
  * @proc:	binder proc that owns buffer
@@ -4246,6 +4408,44 @@ static int binder_thread_write(struct binder_proc *proc,
 			binder_inner_proc_unlock(proc);
 		} break;
 
+		case BC_REQUEST_FREEZE_NOTIFICATION: {

drivers/android/binder.c static int binder_thread_write(struct binder_proc *proc, · line 4421

Quoted 10 lines
drivers/android/binder.c +10 -0
+			struct binder_handle_cookie handle_cookie;
+			int error;
+
+			if (copy_from_user(&handle_cookie, ptr, sizeof(handle_cookie)))
+				return -EFAULT;
+			ptr += sizeof(handle_cookie);
+			error = binder_request_freeze_notification(proc, thread,
+								   &handle_cookie);
+			if (error)
+				return error;

I'm looking at the death notification code and it seems it only queues a BR_ERROR after failing to allocate a "death" and that other errors are silently ignored? I'm not sure if that was intentional design. I wonder if we are trying to cover up for some potential races, e.g. multiple threads trying to request a notification of the same binder at the same time. I'll double check this, maybe I'm mistaken.

Quoted 210 lines
drivers/android/binder.c +119 -2
+		} break;
+
+		case BC_CLEAR_FREEZE_NOTIFICATION: {
+			struct binder_handle_cookie handle_cookie;
+			int error;
+
+			if (copy_from_user(&handle_cookie, ptr, sizeof(handle_cookie)))
+				return -EFAULT;
+			ptr += sizeof(handle_cookie);
+			error = binder_clear_freeze_notification(proc, thread, &handle_cookie);
+			if (error)
+				return error;
+		} break;
+
+		case BC_FREEZE_NOTIFICATION_DONE: {
+			binder_uintptr_t cookie;
+			int error;
+
+			if (get_user(cookie, (binder_uintptr_t __user *)ptr))
+				return -EFAULT;
+
+			ptr += sizeof(cookie);
+			error = binder_freeze_notification_done(proc, thread, cookie);
+			if (error)
+				return error;
+		} break;
+
 		default:
 			pr_err("%d:%d unknown command %u\n",
 			       proc->pid, thread->pid, cmd);
@@ -4635,6 +4835,45 @@ static int binder_thread_read(struct binder_proc *proc,
 			if (cmd == BR_DEAD_BINDER)
 				goto done; /* DEAD_BINDER notifications can cause transactions */
 		} break;
+
+		case BINDER_WORK_FROZEN_BINDER: {
+			struct binder_ref_freeze *freeze;
+			struct binder_frozen_state_info info;
+
+			freeze = container_of(w, struct binder_ref_freeze, work);
+			info.is_frozen = freeze->is_frozen;
+			info.cookie = freeze->cookie;
+			freeze->sent = true;
+			binder_enqueue_work_ilocked(w, &proc->delivered_freeze);
+			binder_inner_proc_unlock(proc);
+
+			if (put_user(BR_FROZEN_BINDER, (uint32_t __user *)ptr))
+				return -EFAULT;
+			ptr += sizeof(uint32_t);
+			if (copy_to_user(ptr, &info, sizeof(info)))
+				return -EFAULT;
+			ptr += sizeof(info);
+			binder_stat_br(proc, thread, BR_FROZEN_BINDER);
+			goto done; /* BR_FROZEN_BINDER notifications can cause transactions */
+		} break;
+
+		case BINDER_WORK_CLEAR_FREEZE_NOTIFICATION: {
+			struct binder_ref_freeze *freeze =
+			    container_of(w, struct binder_ref_freeze, work);
+			binder_uintptr_t cookie = freeze->cookie;
+
+			binder_inner_proc_unlock(proc);
+			kfree(freeze);
+			binder_stats_deleted(BINDER_STAT_FREEZE);
+			if (put_user(BR_CLEAR_FREEZE_NOTIFICATION_DONE, (uint32_t __user *)ptr))
+				return -EFAULT;
+			ptr += sizeof(uint32_t);
+			if (put_user(cookie, (binder_uintptr_t __user *)ptr))
+				return -EFAULT;
+			ptr += sizeof(binder_uintptr_t);
+			binder_stat_br(proc, thread, BR_CLEAR_FREEZE_NOTIFICATION_DONE);
+		} break;
+
 		default:
 			binder_inner_proc_unlock(proc);
 			pr_err("%d:%d: bad work type %d\n",
@@ -5242,6 +5481,48 @@ static bool binder_txns_pending_ilocked(struct binder_proc *proc)
 	return false;
 }
 
+static void binder_add_freeze_work(struct binder_proc *proc, bool is_frozen)
+{
+	struct rb_node *n;
+	struct binder_ref *ref;
+
+	binder_inner_proc_lock(proc);
+	for (n = rb_first(&proc->nodes); n; n = rb_next(n)) {
+		struct binder_node *node;
+
+		node = rb_entry(n, struct binder_node, rb_node);
+		binder_inner_proc_unlock(proc);
+		binder_node_lock(node);
+		hlist_for_each_entry(ref, &node->refs, node_entry) {
+			/*
+			 * Need the node lock to synchronize
+			 * with new notification requests and the
+			 * inner lock to synchronize with queued
+			 * freeze notifications.
+			 */
+			binder_inner_proc_lock(ref->proc);
+			if (!ref->freeze) {
+				binder_inner_proc_unlock(ref->proc);
+				continue;
+			}
+			ref->freeze->work.type = BINDER_WORK_FROZEN_BINDER;
+			if (list_empty(&ref->freeze->work.entry)) {
+				ref->freeze->is_frozen = is_frozen;
+				binder_enqueue_work_ilocked(&ref->freeze->work, &ref->proc->todo);
+				binder_wakeup_proc_ilocked(ref->proc);
+			} else {
+				if (ref->freeze->sent && ref->freeze->is_frozen != is_frozen)
+					ref->freeze->resend = true;
+				ref->freeze->is_frozen = is_frozen;
+			}
+			binder_inner_proc_unlock(ref->proc);
+		}
+		binder_node_unlock(node);
+		binder_inner_proc_lock(proc);
+	}
+	binder_inner_proc_unlock(proc);
+}
+
 static int binder_ioctl_freeze(struct binder_freeze_info *info,
 			       struct binder_proc *target_proc)
 {
@@ -5253,6 +5534,7 @@ static int binder_ioctl_freeze(struct binder_freeze_info *info,
 		target_proc->async_recv = false;
 		target_proc->is_frozen = false;
 		binder_inner_proc_unlock(target_proc);
+		binder_add_freeze_work(target_proc, false);
 		return 0;
 	}
 
@@ -5266,6 +5548,7 @@ static int binder_ioctl_freeze(struct binder_freeze_info *info,
 	target_proc->async_recv = false;
 	target_proc->is_frozen = true;
 	binder_inner_proc_unlock(target_proc);
+	binder_add_freeze_work(target_proc, true);
 
 	if (info->timeout_ms > 0)
 		ret = wait_event_interruptible_timeout(
@@ -5658,6 +5941,7 @@ static int binder_open(struct inode *nodp, struct file *filp)
 	binder_stats_created(BINDER_STAT_PROC);
 	proc->pid = current->group_leader->pid;
 	INIT_LIST_HEAD(&proc->delivered_death);
+	INIT_LIST_HEAD(&proc->delivered_freeze);
 	INIT_LIST_HEAD(&proc->waiting_threads);
 	filp->private_data = proc;
 
@@ -6209,7 +6493,9 @@ static const char * const binder_return_strings[] = {
 	"BR_FAILED_REPLY",
 	"BR_FROZEN_REPLY",
 	"BR_ONEWAY_SPAM_SUSPECT",
-	"BR_TRANSACTION_PENDING_FROZEN"
+	"BR_TRANSACTION_PENDING_FROZEN",
+	"BR_FROZEN_BINDER",
+	"BR_CLEAR_FREEZE_NOTIFICATION_DONE",
 };
 
 static const char * const binder_command_strings[] = {
@@ -6232,6 +6518,9 @@ static const char * const binder_command_strings[] = {
 	"BC_DEAD_BINDER_DONE",
 	"BC_TRANSACTION_SG",
 	"BC_REPLY_SG",
+	"BC_REQUEST_FREEZE_NOTIFICATION",
+	"BC_CLEAR_FREEZE_NOTIFICATION",
+	"BC_FREEZE_NOTIFICATION_DONE",
 };
 
 static const char * const binder_objstat_strings[] = {
@@ -6241,7 +6530,8 @@ static const char * const binder_objstat_strings[] = {
 	"ref",
 	"death",
 	"transaction",
-	"transaction_complete"
+	"transaction_complete",
+	"freeze",
 };
 
 static void print_binder_stats(struct seq_file *m, const char *prefix,
drivers/android/binder_internal.h +5 -2
index 5b7c80b99ae8..21a4b3c48255 100644
@@ -129,12 +129,13 @@ enum binder_stat_types {
 	BINDER_STAT_DEATH,
 	BINDER_STAT_TRANSACTION,
 	BINDER_STAT_TRANSACTION_COMPLETE,
+	BINDER_STAT_FREEZE,
 	BINDER_STAT_COUNT
 };
 
 struct binder_stats {
-	atomic_t br[_IOC_NR(BR_TRANSACTION_PENDING_FROZEN) + 1];
-	atomic_t bc[_IOC_NR(BC_REPLY_SG) + 1];
+	atomic_t br[_IOC_NR(BR_CLEAR_FREEZE_NOTIFICATION_DONE) + 1];
+	atomic_t bc[_IOC_NR(BC_FREEZE_NOTIFICATION_DONE) + 1];
 	atomic_t obj_created[BINDER_STAT_COUNT];
 	atomic_t obj_deleted[BINDER_STAT_COUNT];
 };
@@ -159,6 +160,8 @@ struct binder_work {
 		BINDER_WORK_DEAD_BINDER,
 		BINDER_WORK_DEAD_BINDER_AND_CLEAR,
 		BINDER_WORK_CLEAR_DEATH_NOTIFICATION,
+		BINDER_WORK_FROZEN_BINDER,
+		BINDER_WORK_CLEAR_FREEZE_NOTIFICATION,
 	} type;
 };
 

drivers/android/binder_internal.h struct binder_ref_death { · line 286

Quoted 9 lines
drivers/android/binder_internal.h +6 -0
@@ -275,6 +278,14 @@ struct binder_ref_death {
 	binder_uintptr_t cookie;
 };
 
+struct binder_ref_freeze {
+	struct binder_work work;
+	binder_uintptr_t cookie;
+	bool is_frozen;
+	bool sent;
+	bool resend;

hmm, these could be just bitfields.

drivers/android/binder_internal.h struct binder_ref_data { · line 321

Quoted 10 lines
drivers/android/binder_internal.h +3 -0
+};
+
 /**
  * struct binder_ref_data - binder_ref counts and id
  * @debug_id:        unique ID for the ref
@@ -307,6 +318,7 @@ struct binder_ref_data {
  *               @node indicates the node must be freed
  * @death:       pointer to death notification (ref_death) if requested
  *               (protected by @node->lock)
+ * @freeze:      pointer to freeze notification (ref_freeze) if requested (protected by @node->lock)

nit: please enter the "protected by ..." in a new line as with the other members.

Quoted 90 lines
drivers/android/binder_internal.h +3 -0
  *
  * Structure to track references from procA to target node (on procB). This
  * structure is unsafe to access without holding @proc->outer_lock.
@@ -323,6 +335,7 @@ struct binder_ref {
 	struct binder_proc *proc;
 	struct binder_node *node;
 	struct binder_ref_death *death;
+	struct binder_ref_freeze *freeze;
 };
 
 /**
@@ -374,6 +387,7 @@ struct binder_ref {
  *                        (atomics, no lock needed)
  * @delivered_death:      list of delivered death notification
  *                        (protected by @inner_lock)
+ * @delivered_freeze:     list of delivered freeze notification (protected by @inner_lock)
  * @max_threads:          cap on number of binder threads
  *                        (protected by @inner_lock)
  * @requested_threads:    number of binder threads requested but not
@@ -421,6 +435,7 @@ struct binder_proc {
 	struct list_head todo;
 	struct binder_stats stats;
 	struct list_head delivered_death;
+	struct list_head delivered_freeze;
 	u32 max_threads;
 	int requested_threads;
 	int requested_threads_started;
include/uapi/linux/android/binder.h +35 -0
index d44a8118b2ed..a358cde61ee4 100644
@@ -236,6 +236,11 @@ struct binder_frozen_status_info {
 	__u32            async_recv;
 };
 
+struct binder_frozen_state_info {
+	binder_uintptr_t cookie;
+	__u32            is_frozen;
+};
+
 /* struct binder_extened_error - extended error information
  * @id:		identifier for the failed operation
  * @command:	command as defined by binder_driver_return_protocol
@@ -467,6 +472,17 @@ enum binder_driver_return_protocol {
 	/*
 	 * The target of the last async transaction is frozen.  No parameters.
 	 */
+
+	BR_FROZEN_BINDER = _IOR('r', 21, struct binder_frozen_state_info),
+	/*
+	 * The cookie and a boolean (is_frozen) that indicates whether the process
+	 * transitioned into a frozen or an unfrozen state.
+	 */
+
+	BR_CLEAR_FREEZE_NOTIFICATION_DONE = _IOR('r', 22, binder_uintptr_t),
+	/*
+	 * void *: cookie
+	 */
 };
 
 enum binder_driver_command_protocol {
@@ -550,6 +566,25 @@ enum binder_driver_command_protocol {
 	/*
 	 * binder_transaction_data_sg: the sent command.
 	 */
+
+	BC_REQUEST_FREEZE_NOTIFICATION =
+			_IOW('c', 19, struct binder_handle_cookie),
+	/*
+	 * int: handle
+	 * void *: cookie
+	 */
+
+	BC_CLEAR_FREEZE_NOTIFICATION = _IOW('c', 20,
+					    struct binder_handle_cookie),
+	/*
+	 * int: handle
+	 * void *: cookie
+	 */
+
+	BC_FREEZE_NOTIFICATION_DONE = _IOW('c', 21, binder_uintptr_t),
+	/*
+	 * void *: cookie
+	 */
 };
 
 #endif /* _UAPI_LINUX_BINDER_H */
Signature · 1 line
2.45.2.741.gdbec12cfda-goog