Hi Tim,

On Fri, Jul 14, 2017 at 1:57 PM, Kourt, Tim A <tim.a.kourt@intel.com> wrote:

> NLMSG_OK() expects an int for the second argument.  Assuming truncation isn't an issue here, wouldn't it be better to cast to an int instead?


The uint32_t was used to keep the consistency with the other usages of NLMSG_OK() within ELL and thereafter IWD. Maybe it is time to fix all of them.

I did some digging into the NLMSG_OK() macro and found several patches and discussions related to this very same issue.  It looks like everyone does indeed cast to an unsigned integer, even though the documentation states that the type is int.  Patches to "fix" the issue with the macro have all been rejected, too. Consistency wins.  :)