In the iommu_group_alloc, when the kobject_init_and_add
failed, the group->kobj is associate with iommu_group_ktype,
thus its release function iommu_group_release will be called
by the following kobject_put. The iommu_group_release calls
ida_free with the group->id, so we do not need to do it before
kobject_put.

Signed-off-by: Yuan Can <yuancan@huawei.com>
---
 drivers/iommu/iommu.c | 1 -
 1 file changed, 1 deletion(-)

diff --git a/drivers/iommu/iommu.c b/drivers/iommu/iommu.c
index 780fb7071577..74bc9111c7be 100644
--- a/drivers/iommu/iommu.c
+++ b/drivers/iommu/iommu.c
@@ -654,7 +654,6 @@ struct iommu_group *iommu_group_alloc(void)
 	ret = kobject_init_and_add(&group->kobj, &iommu_group_ktype,
 				   NULL, "%d", group->id);
 	if (ret) {
-		ida_free(&iommu_group_ida, group->id);
 		kobject_put(&group->kobj);
 		return ERR_PTR(ret);
 	}
-- 
2.17.1

On Mon, Aug 15, 2022 at 03:14:23AM +0000, Yuan Can wrote:
>  drivers/iommu/iommu.c | 1 -
>  1 file changed, 1 deletion(-)

Applied, thanks.