>> This elimininates and the big reformatting done before eliminates the follow
>> up patch with indentation changes.
>>
>> The original patch already replaces some instances of option_permission_mask
>> with CLIENT_CONNECT_OPT_MASK. The V3 version does this more consequently.
>>
> 
> Did you mean "more consistently" ?

Hm both work I think. Can change it to cosistently.


>>          options_server_import(&mi->context.options,
>>                                dc_file,
>>                                D_IMPORT_ERRORS|M_OPTERR,
>> -                              option_permissions_mask,
>> +                              CLIENT_CONNECT_OPT_MASK,
> 
> every invocation of options_server_import() gets CLIENT_CONNECT_OPT_MASK
> as parameter. Do we really need it to be a parameter?


The call in init.c does gets OPT_P_DEFAULT & ~OPT_P_PLUGIN instead.

               mi->context.c2.es);
>>              }
>> @@ -1716,21 +1714,19 @@ multi_client_connect_post_plugin(struct 
>> multi_context *m,
>>  static void
>>  multi_client_connect_mda(struct multi_context *m,
>>                           struct multi_instance *mi,
>> -                         const struct buffer_list *config,
>> -                         unsigned int option_permissions_mask,
>>                           unsigned int *option_types_found)
>>  {
>> -    if (config)
>> +    if (mi->cc_config)
> 
> Also getting rid of config and using mi->cc_config looks like something
> that could go in its own patch, as it is quite unrelated.
> 

Okay. That will be a tiny patch.



>> -
>> +#define CLIENT_CONNECT_OPT_MASK (OPT_P_INSTANCE | OPT_P_INHERIT   \
>> +                                 |OPT_P_PUSH | OPT_P_TIMER | OPT_P_CONFIG   
>> \
>> +                                 |OPT_P_ECHO | OPT_P_COMP | OPT_P_SOCKFLAGS)
>>  
>>  static inline bool
>>  multi_process_outgoing_link_dowork(struct multi_context *m, struct 
>> multi_instance *mi, const unsigned int mpp_flags)
>>
> 
> 
> other than those few comments the change "looks" reasonable. However, do
> we have a way to test the code being changed? I mean all the code being
> changed (not just the part used by the deferred connect logic)
> 
> Because even though it all looks sane, we might be breaking something else.
>

I am not aware of a test suite other than manually testing the changes
what I did with the whole patch set but not with individual commits.

Arne

Attachment: signature.asc
Description: OpenPGP digital signature

_______________________________________________
Openvpn-devel mailing list
Openvpn-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/openvpn-devel

Reply via email to