drm/panthor: Prevent potential UAF in group creation

This commit prevents the possibility of a use after free issue in the
GROUP_CREATE ioctl function, which arose as pointer to the group is
accessed in that ioctl function after storing it in the Xarray.
A malicious userspace can second guess the handle of a group and try
to call GROUP_DESTROY ioctl from another thread around the same time
as GROUP_CREATE ioctl.

To prevent the use after free exploit, this commit uses a mark on an
entry of group pool Xarray which is added just before returning from
the GROUP_CREATE ioctl function. The mark is checked for all ioctls
that specify the group handle and so userspace won't be abe to delete
a group that isn't marked yet.

v2: Add R-bs and fixes tags

Fixes: de85488138247 ("drm/panthor: Add the scheduler logical block")
Co-developed-by: Boris Brezillon <[e-mail]>
Signed-off-by: Boris Brezillon <[e-mail]>
Signed-off-by: Akash Goel <[e-mail]>
Reviewed-by: Boris Brezillon <[e-mail]>
Reviewed-by: Steven Price <[e-mail]>
Reviewed-by: Chia-I Wu <[e-mail]>
Link: https://patch.msgid.link/[e-mail]

drivers/gpu/drm/panthor/panthor_sched.c

···
  */
 #define MAX_GROUPS_PER_POOL 128
 
+/*
+ * Mark added on an entry of group pool Xarray to identify if the group has
+ * been fully initialized and can be accessed elsewhere in the driver code.
+ */
+#define GROUP_REGISTERED XA_MARK_1
+
 /**
  * struct panthor_group_pool - Group pool
  *
···
 		return;
 
 	xa_lock(&gpool->xa);
-	xa_for_each(&gpool->xa, i, group) {
+	xa_for_each_marked(&gpool->xa, i, group, GROUP_REGISTERED) {
 		guard(spinlock)(&group->fdinfo.lock);
 		pfile->stats.cycles += group->fdinfo.data.cycles;
 		pfile->stats.time += group->fdinfo.data.time;
···
 
 	group_init_task_info(group);
 
+	xa_set_mark(&gpool->xa, gid, GROUP_REGISTERED);
+
 	return gid;
 
 err_erase_gid:
···
 	struct panthor_device *ptdev = pfile->ptdev;
 	struct panthor_scheduler *sched = ptdev->scheduler;
 	struct panthor_group *group;
+
+	if (!xa_get_mark(&gpool->xa, group_handle, GROUP_REGISTERED))
+		return -EINVAL;
 
 	group = xa_erase(&gpool->xa, group_handle);
 	if (!group)
···
 }
 
 static struct panthor_group *group_from_handle(struct panthor_group_pool *pool,
-					       u32 group_handle)
+					       unsigned long group_handle)
 {
 	struct panthor_group *group;
 
 	xa_lock(&pool->xa);
-	group = group_get(xa_load(&pool->xa, group_handle));
+	group = group_get(xa_find(&pool->xa, &group_handle, group_handle, GROUP_REGISTERED));
 	xa_unlock(&pool->xa);
 
 	return group;
···
 		return;
 
 	xa_lock(&gpool->xa);
-	xa_for_each(&gpool->xa, i, group) {
+	xa_for_each_marked(&gpool->xa, i, group, GROUP_REGISTERED) {
 		stats->resident += group->fdinfo.kbo_sizes;
 		if (group->csg_id >= 0)
 			stats->active += group->fdinfo.kbo_sizes;