Hi Gavin,
>>> diff --git a/net/bluetooth/af_bluetooth.c b/net/bluetooth/af_bluetooth.c
>>> index ee319779781e6..69374321130e4 100644
>>> --- a/net/bluetooth/af_bluetooth.c
>>> +++ b/net/bluetooth/af_bluetooth.c
>>> @@ -568,7 +568,7 @@ int bt_sock_wait_state(struct sock *sk, int state, unsigned long timeo)
>>> EXPORT_SYMBOL(bt_sock_wait_state);
>>>
>>> /* This function expects the sk lock to be held when called */
>>> -int bt_sock_wait_ready(struct sock *sk, unsigned long flags)
>>> +int bt_sock_wait_ready(struct sock *sk, unsigned int flags)
>>
>> can we then also do s/flags/msg_flags/ then.
> I prefer keeping it as flags because all other net code also uses
> flags, msg_flags only appears
> in msg->msg_flags.
while that might be true, I find it a lot clearer if the variable is msg_flags.
>>> @@ -576,7 +576,7 @@ int bt_sock_wait_ready(struct sock *sk, unsigned long flags)
>>>
>>> BT_DBG("sk %p", sk);
>>>
>>> - timeo = sock_sndtimeo(sk, flags & O_NONBLOCK);
>>> + timeo = sock_sndtimeo(sk, flags & MSG_DONTWAIT);
>>
>> Since sock_sndtimeo() is taking a bool. This might be better !!(flags & MSG_DONTWAIT).
> It appears to be well-known in the net code that sock_sndtimeo takes a
> bool, since no other
> uses of it do the "!!" conversion.
>
> Let me know what you think. I can make the changes if needed but I was
> just trying my best
> to match the currently existing convention.
And other code in the kernel makes sure to clearly turn something into a bool. You get 0x00 and 0x40.
Regards
Marcel
