On 06/24/10 21:27, Shawn Emery wrote:
> On 06/24/10 04:44 PM, Glenn Barry wrote:
>>
>> looks good in general, a few comments:
>>
>>    Abstracted interfaces
>>     =====================
>>
>> <add descriptive blurb of func here>
>>     krb5_error_code __krb5_kt_add_ad_entries(krb5_context ctx,
>>         char **svc_princs, krb5_kvno kvno, uint_t flags, char *password)
>> **** maybe add blurb for each func like for this one cuz parts of it 
>> like
>>      '_ad_' may not be obvious:
>>          Add keys for service principals to krb5.keytab for an Active 
>> Directory client:
>
> Done.
>
>> krb5_error_code __krb5_kt_remove_by_svcprinc(krb5_context ctx,
>>         char *svc_princ)
>>
>>     where ctx is the pointer passed back from krb5_init_context
>>     where svc_princ is the service principal name that is matched for 
>> any
>>     keytab entries to be removed
>> **** partial matches and/or wildcards supported?
>> **** pro/con using char * for svc_princ vs krb5_principal?
>
> Clarified in the beginning text.  See diffs below.
>
>>     krb5_error_code __krb5_kt_validate(krb5_context ctx,
>>         char *svc_princ, uint_t flags, boolean_t *valid)
>> **** krb5_boolean?
>
> I thought of this, but I didn't like how they were typically assigned 
> in the mech, with 0 or 1, instead of TRUE and FALSE.  In any case, 
> I'll change them.
>
>> **** n/n-1 Note:  this w/respect to kvno?  (make that clear)
>
> Done.
>
> @@ -40,10 +40,13 @@
>      =====================
>
>      krb5_error_code __krb5_kt_add_ad_entries(krb5_context ctx,
>          char **svc_princs, krb5_kvno kvno, uint_t flags, char *password)
>
> +    Adds keys to the keytab file for a default set of service principals
> +    (refer to PSARC/2007/401) in an Active Directory environment.
> +
>      where ctx is the pointer passed back from krb5_init_context
>      where svc_princs is an array of service principal names to be added
>          to the keytab file, terminated by a NULL pointer
>      where kvno is the key version number of the set of service principal
>          keys to be added
> @@ -69,10 +72,13 @@
>
>      _________
>
>      krb5_error_code __krb5_kt_remove_by_realm(krb5_context ctx, char 
> *realm)
>
> +    Removes all key entries in the keytab file that match the exact 
> realm
> +    name specified.
> +
>      where ctx is the pointer passed back from krb5_init_context
>      where realm is the realm name that is matched for any keytab entries
>      to be removed
>      Note: if there are no entries matching realm then 0 (success) is
>      returned
> @@ -80,10 +86,13 @@
>      _________
>
>      krb5_error_code __krb5_kt_remove_by_svcprinc(krb5_context ctx,
>          char *svc_princ)
>
> +    Removes all key entries in the keytab file that match the exact 
> service
> +    principal name specified.
> +
>      where ctx is the pointer passed back from krb5_init_context
>      where svc_princ is the service principal name that is matched for 
> any
>      keytab entries to be removed
>
>      Note: if there are no entries matching svc_princ then 0 (success) is
> @@ -90,12 +99,16 @@
>      returned
>
>      _________
>
>      krb5_error_code __krb5_kt_validate(krb5_context ctx,
> -        char *svc_princ, uint_t flags, boolean_t *valid)
> +        char *svc_princ, uint_t flags, krb5_boolean *valid)
>
> +    The validate function determines that the correct set of service
> +    principals (refer to PSARC/2007/401) exist and that this set has a
> +    valid set of encryption types.
> +
>      where ctx is the pointer passed back from krb5_init_context
>      where svc_princ is the principal to be validated in the keytab file
>      where flags is the set of conditions that affects the key table 
> entries
>      that the function considers valid
>          current set of defined flags:
> @@ -106,13 +119,13 @@
>
>      where valid is a boolean that is set if the svc_princ is correctly
>      populated in the keytab file based on the flags set else valid is 
> unset.
>
>      Note: The validate function does not distinguish between having n 
> and
> -    n-1 key sets in the keytab file.  This is from the fact that AD
> -    environments will refresh old credential caches if n-1 keys are not
> -    present in the keytab file.
> +    n-1 key version numbers (kvno) in the keytab file.  This is from the
> +    fact that AD environments will refresh old credential caches if 
> kvno,
> +    n-1, are not present in the keytab file.
>
>          Stability Level
>      ===============
>
>      Contract project private for both primitive and abstracted 
> interfaces.
>
> Shawn.
> -- 
>> On 06/21/10 09:57, Shawn Emery wrote:
>>> On 06/21/10 08:30 AM, Mark Phalan wrote:
>>>> On Mon, 2010-06-21 at 09:41 -0400, Wyllys Ingersoll wrote:
>>>>> I filed the case today - PSARC 2010/229
>>>>>
>>>> Shawn:
>>>> Was this sent out to the i-team alias for review? I don't remember
>>>> seeing it here first but it's possible I missed it.
>>>>
>>>> If it wasn't I think it should have been.
>>>
>>> I apologize for not sending this out to the i-team alias for review 
>>> earlier, however the design is still malleable and I still expect 
>>> review comments from this and other groups.
>>>
>>> Shawn.
>>> -- 
>>
>>
>

_______________________________________________
opensolaris-arc mailing list
[email protected]

Reply via email to