On 2/1/21 8:25 PM, Dan Carpenter wrote:
> The error handling in this function frees "reg" but it is still on the
> "o2hb_all_regions" list so it will lead to a use after free.  The fix
> for this is to only add it to the list after everything has succeeded.
> 
Seems we have to clear the bitmap as well in error case.
So how about add a new error label and handle them both?

Thanks,
Joseph

> Fixes: 1cf257f51191 ("ocfs2: fix memory leak")
> Signed-off-by: Dan Carpenter <dan.carpen...@oracle.com>
> ---
> This is from static analysis and hasn't been tested.
> 
>  fs/ocfs2/cluster/heartbeat.c | 5 ++++-
>  1 file changed, 4 insertions(+), 1 deletion(-)
> 
> diff --git a/fs/ocfs2/cluster/heartbeat.c b/fs/ocfs2/cluster/heartbeat.c
> index 0179a73a3fa2..92af4dc813e7 100644
> --- a/fs/ocfs2/cluster/heartbeat.c
> +++ b/fs/ocfs2/cluster/heartbeat.c
> @@ -2025,7 +2025,6 @@ static struct config_item 
> *o2hb_heartbeat_group_make_item(struct config_group *g
>               }
>               set_bit(reg->hr_region_num, o2hb_region_bitmap);
>       }
> -     list_add_tail(&reg->hr_all_item, &o2hb_all_regions);
>       spin_unlock(&o2hb_live_lock);
>  
>       config_item_init_type_name(&reg->hr_item, name, &o2hb_region_type);
> @@ -2053,6 +2052,10 @@ static struct config_item 
> *o2hb_heartbeat_group_make_item(struct config_group *g
>  
>       o2hb_debug_region_init(reg, o2hb_debug_dir);
>  
> +     spin_lock(&o2hb_live_lock);
> +     list_add_tail(&reg->hr_all_item, &o2hb_all_regions);
> +     spin_unlock(&o2hb_live_lock);
> +
>       return &reg->hr_item;
>  
>  unregister_handler:
> 

Reply via email to