firmware: arm_ffa: Move memory allocation outside the mutex locking

JIRA: https://issues.redhat.com/browse/RHEL-102692

commit 27e850c
Author: Sudeep Holla <sudeep.holla@arm.com>
Date: Wed, 28 May 2025 09:49:42 +0100

    The notifier callback node allocation is currently done while holding
    the notify_lock mutex. While this is safe even if memory allocation may
    sleep, we need to move the allocation outside the locked region in
    preparation to move from using muxtes to rwlocks.

    Move the memory allocation to avoid potential sleeping in atomic context
    once the locks are moved from mutex to rwlocks.

    Fixes: e057344 ("firmware: arm_ffa: Add interfaces to request notification callbacks")
    Message-Id: <20250528-ffa_notif_fix-v1-2-5ed7bc7f8437@arm.com>
    Reviewed-by: Jens Wiklander <jens.wiklander@linaro.org>
    Signed-off-by: Sudeep Holla <sudeep.holla@arm.com>

Signed-off-by: Marcin Juszkiewicz <mjuszkiewicz@redhat.com>
Message-ID: <2e157ea409eff06998e85e5400ac462b33012bab.1752653152.git.mjuszkiewicz@redhat.com>
Message-ID: <db7df8ed141e0d3ae7d6c6a48c77b8e4c654a4cc.1752659027.git.mjuszkiewicz@redhat.com>

Author: hrw

drivers/firmware/arm_ffa/driver.c

@@ -1250,13 +1250,12 @@ notifier_hnode_get_by_type(u16 notify_id, enum notify_type type)
 	return NULL;
 }
 
-static int
-update_notifier_cb(struct ffa_device *dev, int notify_id, void *cb,
-		   void *cb_data, bool is_registration, bool is_framework)
+static int update_notifier_cb(struct ffa_device *dev, int notify_id,
+			      struct notifier_cb_info *cb, bool is_framework)
 {
 	struct notifier_cb_info *cb_info = NULL;
 	enum notify_type type = ffa_notify_type_get(dev->vm_id);
-	bool cb_found;
+	bool cb_found, is_registration = !!cb;
 
 	if (is_framework)
 		cb_info = notifier_hnode_get_by_vmid_uuid(notify_id, dev->vm_id,
@@ -1270,18 +1269,7 @@ update_notifier_cb(struct ffa_device *dev, int notify_id, void *cb,
 		return -EINVAL;
 
 	if (is_registration) {
-		cb_info = kzalloc(sizeof(*cb_info), GFP_KERNEL);
-		if (!cb_info)
-			return -ENOMEM;
-
-		cb_info->dev = dev;
-		cb_info->cb_data = cb_data;
-		if (is_framework)
-			cb_info->fwk_cb = cb;
-		else
-			cb_info->cb = cb;
-
-		hash_add(drv_info->notifier_hash, &cb_info->hnode, notify_id);
+		hash_add(drv_info->notifier_hash, &cb->hnode, notify_id);
 	} else {
 		hash_del(&cb_info->hnode);
 		kfree(cb_info);
@@ -1303,8 +1291,7 @@ static int __ffa_notify_relinquish(struct ffa_device *dev, int notify_id,
 
 	mutex_lock(&drv_info->notify_lock);
 
-	rc = update_notifier_cb(dev, notify_id, NULL, NULL, false,
-				is_framework);
+	rc = update_notifier_cb(dev, notify_id, NULL, is_framework);
 	if (rc) {
 		pr_err("Could not unregister notification callback\n");
 		mutex_unlock(&drv_info->notify_lock);
@@ -1335,35 +1322,48 @@ static int __ffa_notify_request(struct ffa_device *dev, bool is_per_vcpu,
 {
 	int rc;
 	u32 flags = 0;
+	struct notifier_cb_info *cb_info = NULL;
 
 	if (ffa_notifications_disabled())
 		return -EOPNOTSUPP;
 
 	if (notify_id >= FFA_MAX_NOTIFICATIONS)
 		return -EINVAL;
 
+	cb_info = kzalloc(sizeof(*cb_info), GFP_KERNEL);
+	if (!cb_info)
+		return -ENOMEM;
+
+	cb_info->dev = dev;
+	cb_info->cb_data = cb_data;
+	if (is_framework)
+		cb_info->fwk_cb = cb;
+	else
+		cb_info->cb = cb;
+
 	mutex_lock(&drv_info->notify_lock);
 
 	if (!is_framework) {
 		if (is_per_vcpu)
 			flags = PER_VCPU_NOTIFICATION_FLAG;
 
 		rc = ffa_notification_bind(dev->vm_id, BIT(notify_id), flags);
-		if (rc) {
-			mutex_unlock(&drv_info->notify_lock);
-			return rc;
-		}
+		if (rc)
+			goto out_unlock_free;
 	}
 
-	rc = update_notifier_cb(dev, notify_id, cb, cb_data, true,
-				is_framework);
+	rc = update_notifier_cb(dev, notify_id, cb_info, is_framework);
 	if (rc) {
 		pr_err("Failed to register callback for %d - %d\n",
 		       notify_id, rc);
 		if (!is_framework)
 			ffa_notification_unbind(dev->vm_id, BIT(notify_id));
 	}
+
+out_unlock_free:
 	mutex_unlock(&drv_info->notify_lock);
+	if (rc)
+		kfree(cb_info);
 
 	return rc;
 }