Re: [PATCH 1/2] vfstest: move sys_umount2() to cleanup path in tcore_acls()

On Thu, Jun 18, 2026 at 12:11:53PM +0200, German Maglione wrote:
> sys_umount2() in tcore_acls() only runs on the success path. If setfacl
> fails (e.g. no ACL support), the mount on DIR2 is left behind and
> test_cleanup() can't remove T_DIR1. This causes every subsequent test
> to fail with EEXIST on mkdirat(T_DIR1).
> 
> Move sys_umount2() into the out: block so it always runs.
> 
> Signed-off-by: German Maglione <[email protected]>
> ---

Makes sense to me, thanks for fixing it.

Reviewed-by: Zorro Lang <[email protected]>

>  src/vfs/idmapped-mounts.c | 5 ++---
>  1 file changed, 2 insertions(+), 3 deletions(-)
> 
> diff --git a/src/vfs/idmapped-mounts.c b/src/vfs/idmapped-mounts.c
> index ed9992f975c14b57c5d6d1634ca011b015b78bcb..d0630f5931ade713a99f5dec1cbe24bdaa75d35d 100644
> --- a/src/vfs/idmapped-mounts.c
> +++ b/src/vfs/idmapped-mounts.c
> @@ -240,12 +240,11 @@ int tcore_acls(const struct vfstest_info *info)
>  		goto out;
>  	}
>  
> -	snprintf(t_buf, sizeof(t_buf), "%s/" T_DIR1 "/" DIR2, info->t_mountpoint);
> -	sys_umount2(t_buf, MNT_DETACH);
> -
>  	fret = 0;
>  	log_debug("Ran test");
>  out:
> +	snprintf(t_buf, sizeof(t_buf), "%s/" T_DIR1 "/" DIR2, info->t_mountpoint);
> +	sys_umount2(t_buf, MNT_DETACH);
>  	safe_close(attr.userns_fd);
>  	safe_close(dir1_fd);
>  	safe_close(open_tree_fd);
> 
> -- 
> 2.49.0
>