[PATCH] ceph: ensure we return an error when parsing corrupt mdsmap

[PATCH] ceph: ensure we return an error when parsing corrupt mdsmap

[Jeff Layton]

Commit ba5e57de7b20 (ceph: reconnect to the export targets on new
mdsmaps) changed ceph_mdsmap_decode to "goto corrupt" when given a
bogus mds rank in the export targets. It did not set the err variable
however, which caused that function to return a NULL pointer instead of
a proper ERR_PTR value for the error.

Fix this by setting err before doing the "goto corrupt".

URL: https://tracker.ceph.com/issues/52436
Fixes: ba5e57de7b20 ("ceph: reconnect to the export targets on new mdsmaps")
Signed-off-by: Jeff Layton <jlayton@kernel.org>
---
 fs/ceph/mdsmap.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

I'll plan to fold this change into the original patch since it hasn't
been merged yet. Let me know if you have objections.

diff --git a/fs/ceph/mdsmap.c b/fs/ceph/mdsmap.c
index d995cb02d30c..61d67cbcb367 100644
--- a/fs/ceph/mdsmap.c
+++ b/fs/ceph/mdsmap.c
@@ -263,8 +263,10 @@ struct ceph_mdsmap *ceph_mdsmap_decode(void **p, void *end, bool msgr2)
 				goto nomem;
 			for (j = 0; j < num_export_targets; j++) {
 				target = ceph_decode_32(&pexport_targets);
-				if (target >= m->possible_max_rank)
+				if (target >= m->possible_max_rank) {
+					err = -EIO;
 					goto corrupt;
+				}
 				info->export_targets[j] = target;
 			}
 		} else {
-- 
2.31.1

Re: [PATCH] ceph: ensure we return an error when parsing corrupt mdsmap

[Xiubo Li]

On 8/30/21 8:10 PM, Jeff Layton wrote:
> Commit ba5e57de7b20 (ceph: reconnect to the export targets on new
> mdsmaps) changed ceph_mdsmap_decode to "goto corrupt" when given a
> bogus mds rank in the export targets. It did not set the err variable
> however, which caused that function to return a NULL pointer instead of
> a proper ERR_PTR value for the error.
>
> Fix this by setting err before doing the "goto corrupt".
>
> URL: https://tracker.ceph.com/issues/52436
> Fixes: ba5e57de7b20 ("ceph: reconnect to the export targets on new mdsmaps")
> Signed-off-by: Jeff Layton <jlayton@kernel.org>
> ---
>   fs/ceph/mdsmap.c | 4 +++-
>   1 file changed, 3 insertions(+), 1 deletion(-)
>
> I'll plan to fold this change into the original patch since it hasn't
> been merged yet. Let me know if you have objections.
>
> diff --git a/fs/ceph/mdsmap.c b/fs/ceph/mdsmap.c
> index d995cb02d30c..61d67cbcb367 100644
> --- a/fs/ceph/mdsmap.c
> +++ b/fs/ceph/mdsmap.c
> @@ -263,8 +263,10 @@ struct ceph_mdsmap *ceph_mdsmap_decode(void **p, void *end, bool msgr2)
>   				goto nomem;
>   			for (j = 0; j < num_export_targets; j++) {
>   				target = ceph_decode_32(&pexport_targets);
> -				if (target >= m->possible_max_rank)
> +				if (target >= m->possible_max_rank) {
> +					err = -EIO;
>   					goto corrupt;
> +				}
>   				info->export_targets[j] = target;
>   			}
>   		} else {

Make sense and LGTM.