Hi P-G,
On Wed, Aug 15, 2012, Par-Gunnar HJALMDAHL wrote:
> Basically I tried to keep the patch to a minimum since it was quite a
> simple issue to fix.
>
> I have no problems to use the UUID in lib/uuid.c. I could add a 2 new
> exported functions for retrieving base UUID, one for little endian and
> one for big endian, since uuid.c endian depends on CPU endian while
> PAN endian is always big endian. So:
> void get_le_bt_base_uuid(uint128_t *uuid);
> void get_be_bt_base_uuid(uint128_t *uuid);
Note that the uuid.{c,h} name space is bt_uuid_*
So far uuid.c has been fully hiding any endianness issues in its public
API by only talking host-endianness. So, if this was to be really done
in a clean way I'd say the calling code should look like:
uint128_t u128;
bt_uuid_t uuid;
u128 = bt_get_be128(dest);
bt_uuid128_create(&uuid, u128);
if (!bt_uuid_is_bluetooth(&uuid))
return BNEP_CONN_INVALID_DST;
/* same for source */
However, as this would require at least two new functions
(bt_get_be128() in lib/bluetooth.h and bt_uuid_is_bluetooth() in
lib/uuid.c) and like you say the use case is quite trivial, I think this
might be a bit overkill.
So, since you were fine with my alteration of your patch I think I'll
just go ahead and push it upstream. If you still want to add the
necessary APIs for the "truly clean" approach, feel free, but I won't be
bugging you about it ;)
Johan
--
To unsubscribe from this list: send the line "unsubscribe linux-bluetooth" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
