Hi Johan,
> On Fri, Nov 7, 2014 at 4:20 AM, Johan Hedberg <johan.hedberg@xxxxxxxxx> wrote:
> Hi Arman,
>
> On Fri, Nov 07, 2014, Arman Uguray wrote:
>> This patch addresses warnings that arise when compiled with
>> -W=cast-align.
>> ---
>> src/shared/gatt-client.c | 9 +++++----
>> 1 file changed, 5 insertions(+), 4 deletions(-)
>>
>> diff --git a/src/shared/gatt-client.c b/src/shared/gatt-client.c
>> index b4f28b2..2e321b2 100644
>> --- a/src/shared/gatt-client.c
>> +++ b/src/shared/gatt-client.c
>> @@ -1647,7 +1647,7 @@ bool bt_gatt_characteristic_iter_init(struct bt_gatt_characteristic_iter *iter,
>> return false;
>>
>> memset(iter, 0, sizeof(*iter));
>> - iter->service = (struct service_list *) service;
>> + iter->service = (void *) service;
>>
>> return true;
>> }
>> @@ -1677,7 +1677,7 @@ bool bt_gatt_include_service_iter_init(struct bt_gatt_incl_service_iter *iter,
>> return false;
>>
>> memset(iter, 0, sizeof(*iter));
>> - iter->service = (struct service_list *) service;
>> + iter->service = (void *) service;
>>
>> return true;
>> }
>> @@ -2387,7 +2387,7 @@ bool bt_gatt_client_register_notify(struct bt_gatt_client *client,
>> struct chrc_data *chrc = NULL;
>> struct bt_gatt_service_iter iter;
>> const bt_gatt_service_t *service;
>> - size_t i;
>> + size_t i, offset;
>>
>> if (!client || !chrc_value_handle || !callback)
>> return false;
>> @@ -2402,7 +2402,8 @@ bool bt_gatt_client_register_notify(struct bt_gatt_client *client,
>> while (bt_gatt_service_iter_next(&iter, &service)) {
>> if (chrc_value_handle >= service->start_handle &&
>> chrc_value_handle <= service->end_handle) {
>> - svc_data = (struct service_list *) service;
>> + offset = offsetof(struct service_list, service);
>> + svc_data = (void *) (service - offset);
>> break;
>
> Are you sure a (void *) case isn't enough for the last one too? After
> all offsetof() should be returning 0 which in the end results in a void *
> cast of the service pointer anyway.
>
Yeah, just a (void *) cast should be enough. I figured it's a bit more
"correct" to do the pointer math in the code, but yeah that offset
should be 0 (and afaik is guaranteed to be 0). I'll just remove the
math I guess.
> Btw, if we at any point find the need to really do this kind of pointer
> math I'd suggest adding a container_of() helper macro like the kernel
> uses.
I actually thought of adding a container_of macro then I decided that
would be a bit of an overkill for just this case.
>
> Johan
Cheers,
Arman
--
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
