Re: [B.A.T.M.A.N.] [PATCH maint] batman-adv: Fix broken capability checks

[Sven Eckelmann <sven@narfation.org>]

[-- Attachment #1: Type: text/plain, Size: 2067 bytes --]

On Friday 03 July 2015 19:13:02 Marek Lindner wrote:
> > @@ -299,10 +299,10 @@ struct batadv_orig_node {
> >   *  (= orig node announces a tvlv of type BATADV_TVLV_MCAST)
> >   */
> >  enum batadv_orig_capabilities {
> > -	BATADV_ORIG_CAPA_HAS_DAT = BIT(0),
> > -	BATADV_ORIG_CAPA_HAS_NC = BIT(1),
> > -	BATADV_ORIG_CAPA_HAS_TT = BIT(2),
> > -	BATADV_ORIG_CAPA_HAS_MCAST = BIT(3),
> > +	BATADV_ORIG_CAPA_HAS_DAT,
> > +	BATADV_ORIG_CAPA_HAS_NC,
> > +	BATADV_ORIG_CAPA_HAS_TT,
> > +	BATADV_ORIG_CAPA_HAS_MCAST,
> >  };
> 
> Do we need an initialization of the first element or can we rely on the 
> compiler to start with 0 ?

Just discussed this with Marek on IRC. We came to the conclusion that ANSI C 
"3.5.2.2 Enumeration specifiers" is enforcing a 0 as first entry (when not 
specified otherwise):

    If the first enumerator has no = , the value of
    its enumeration constant is 0.  Each subsequent enumerator with no =
    defines its enumeration constant as the value of the constant
    expression obtained by adding 1 to the value of the previous
    enumeration constant.

But the code should maybe cleaned up to either use "= 0" everywhere or nowhere 
in the internal representations.

Places where = 0 is used:

 * net/batman-adv/bat_iv_ogm.c:enum batadv_dup_status
 * net/batman-adv/main.h:enum batadv_uev_action (maybe not internal?)
 * net/batman-adv/main.h:enum batadv_uev_type (maybe not internal?)
 * net/batman-adv/types.h:enum batadv_dhcp_recipient

Not used:

 * net/batman-adv/gateway_common.h:enum batadv_gw_modes
 * net/batman-adv/gateway_common.h:enum batadv_bandwidth_units
 * net/batman-adv/hard-interface.h:enum batadv_hard_if_state
 * net/batman-adv/hard-interface.h:enum batadv_hard_if_cleanup
 * net/batman-adv/main.h:enum batadv_mesh_state
 * net/batman-adv/multicast.h:enum batadv_forw_mode
 * net/batman-adv/types.h:enum batadv_counters

I don't count packet.h because this is clearly not internal and the packet 
definitions should not be accidentally changed just because the enum constant 
order is changed.

Kind regards,
	Sven

[-- Attachment #2: This is a digitally signed message part. --]
[-- Type: application/pgp-signature, Size: 819 bytes --]