On Wed, 2019-11-27 at 05:45 -0500, xiubli@xxxxxxxxxx wrote:
> From: Xiubo Li <xiubli@xxxxxxxxxx>
>
> The cap->implemented is one subset of the cap->issued, the logic
> here want to exclude the revoking caps, but the following code
> will be (~cap->implemented | cap->issued) == 0xFFFF, so it will
> make no sense when we doing the "have &= 0xFFFF".
>
> Signed-off-by: Xiubo Li <xiubli@xxxxxxxxxx>
> ---
> fs/ceph/caps.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/fs/ceph/caps.c b/fs/ceph/caps.c
> index c62e88da4fee..a9335402c2a5 100644
> --- a/fs/ceph/caps.c
> +++ b/fs/ceph/caps.c
> @@ -812,7 +812,7 @@ int __ceph_caps_issued(struct ceph_inode_info *ci, int *implemented)
> */
> if (ci->i_auth_cap) {
> cap = ci->i_auth_cap;
> - have &= ~cap->implemented | cap->issued;
> + have &= ~(cap->implemented & ~cap->issued);
> }
> return have;
> }
Nice catch. This patch looks correct to me. I'll merge it into the
testing branch and we'll see if anything breaks.
--
Jeff Layton <jlayton@xxxxxxxxxx>
