-c

This option is deprecated and is currently ignored.

--no-env-probe|-E

Normally, cifs.upcall will probe the environment variable space of the process that initiated the upcall in order to fetch the value of $KRB5CCNAME. This can assist the program with finding credential caches in non-default locations. If this option is set, then the program won't do this and will rely on finding credcaches in the default locations specified in krb5.conf. Note that this is never performed when the uid is 0. The default credcache location is always used when the uid is 0, regardless of the environment variable setting in the process.

--krb5conf|-k=/path/to/krb5.conf

This option allows administrators to set an alternate location for the krb5.conf file that cifs.upcall will use.

--keytab=|-K=/path/to/keytab

This option allows administrators to specify a keytab file to be used. When a user has no credential cache already established, cifs.upcall will attempt to use this keytab to acquire them. The default is the system-wide keytab /etc/krb5.keytab.

--trust-dns|-t

With krb5 upcalls, the name used as the host portion of the service principal defaults to the hostname portion of the UNC. This option allows the upcall program to reverse resolve the network address of the server in order to get the hostname.

This is less secure than not trusting DNS. When using this option, it's possible that an attacker could get control of DNS and trick the client into mounting a different server altogether. It's preferable to instead add server principals to the KDC for every possible hostname, but this option exists for cases where that isn't possible. The default is to not trust reverse hostname lookups in this fashion.

--legacy-uid|-l

Traditionally, the kernel has sent only a single uid= parameter to the upcall for the SPNEGO upcall that's used to determine what user's credential cache to use. This parameter is affected by the uid= mount option, which also governs the ownership of files on the mount.

Newer kernels send a creduid= option as well, which contains what uid it thinks actually owns the credentials that it's looking for. At mount time, this is generally set to the real uid of the user doing the mount. For multisession mounts, it's set to the fsuid of the mount user. Set this option if you want cifs.upcall to use the older uid= parameter instead of the creduid= parameter.

--expire|-e

Override default timeout value (600 seconds) for dns_resolver key.

--version|-v

Print version number and exit.

GSS_USE_PROXY="yes"

Enable usage of gssproxy for credential retrieval. This includes keytab based client initiation as well as (Resource Based) Constrained Delegation. See gssproxy-mech(8).

cifs.spnego

This keytype is for retrieving kerberos session keys

dns_resolver

This key type is for resolving hostnames into IP addresses. Support for this key type may eventually be deprecated (see below).

To make this program useful for CIFS, you'll need to set up entries for them in request-key.conf(5). Here's an example of an entry for each key type:

#OPERATION  TYPE           D C PROGRAM ARG1 ARG2...
#=========  =============  = = ================================
create      cifs.spnego    * * /usr/bin/cifs.upcall %k
create      dns_resolver   * * /usr/bin/cifs.upcall %k

See request-key.conf(5) for more info on each field.

The keyutils package has also started including a dns_resolver handling program as well that is preferred over the one in cifs.upcall. If you are using a keyutils version equal to or greater than 1.5, you should use key.dns_resolver to handle the dns_resolver keytype instead of cifs.upcall. See key.dns_resolver(8) for more info.