On Mon, 15 Nov 2010 18:15:23 +0530
Suresh Jayaraman <sjayaraman@xxxxxxx> wrote:
> cifs_get_tcon() could return any of the following errors:
> -ENOMEM/-ENODEV/EREMOTEIO. We should follow the DFS referral code path only
> when we get -EREMOTEIO (STATUS_PATH_NOT_COVERED) from the server and not
> otherwise.
>
> Compile-tested only.
>
> Signed-off-by: Suresh Jayaraman <sjayaraman@xxxxxxx>
> ---
> fs/cifs/connect.c | 5 ++++-
> 1 files changed, 4 insertions(+), 1 deletions(-)
>
> diff --git a/fs/cifs/connect.c b/fs/cifs/connect.c
> index 251a17c..c3a2323 100644
> --- a/fs/cifs/connect.c
> +++ b/fs/cifs/connect.c
> @@ -2780,7 +2780,10 @@ try_mount_again:
> if (IS_ERR(tcon)) {
> rc = PTR_ERR(tcon);
> tcon = NULL;
> - goto remote_path_check;
> + if (rc == -EREMOTEIO)
> + goto remote_path_check;
> + else
> + goto mount_fail_check;
> }
>
> /* do not care if following two calls succeed - informational */
Don't you mean "EREMOTE" here? I've always interpreted "EREMOTEIO" to
mean that the server had the equivalent of an I/O error, whereas
"EREMOTE" means "Object is remote".
I don't see how this patch materially changes anything. Pseudocode,
assuming that rc == -EREMOTE:
remote_path_check:
/* check if a whole path (including prepath) is not remote */
if (!rc && cifs_sb->prepathlen && tcon) {
/* stuff we won't hit because rc is non-zero */
}
if (rc == -EREMOTE) {
/* handle a DFS referral */
}
if (rc)
goto mount_fail_check;
I'll grant that cifs_mount is still a mess, but I'm not sure this
really improves anything.
--
Jeff Layton <jlayton@xxxxxxxxxx>
--
To unsubscribe from this list: send the line "unsubscribe linux-cifs" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
