Vladimir Gurevich

It is absolutely a valid type for the `select()`. Moreover, we the `if()` statements in the parser are transformed into very similar selects already. The only thing is that one...

Given the current situation on the ground, I think it would make sense to re-think this. As far as I am aware, most, if not all, practical implementations do things...

@apinski-cavium -- I wonder what exact model do you plan to implement. Specifically, is there any example where you would **not** want to implement atomic behavior and where you can't...

@jafingerhut -- you are absolutely correct, the current language spec allows this to happen. I simply stated that no current implementation works this way as far as I know and...

I agree that if the target wants to allow something like what @jafingerhut just mentioned, then we need to have a special, first-class way to mark those critical sections, which...

I think that a clear indication of architecture for which the program has been written (instead of relying on a specific `#include` file) might be even more useful. Adding the...

@apinski-cavium -- I am all for that. I would not make this mandatory (e.g. someone wants to see those fields in the tools or whatever), but if we can extend...

@blp -- one example is a header that one would prepend to recirculated packet. In many cases these headers need to carry all kinds of fields that are not necessarily...

@blp -- I agree. For example in the Intel Connectivity Academy course that I teach, I specifically point out the difference in how one should construct internal-only vs. externally visible...

@apinski-cavium -- @hanw beat me to it and he is totally correct: requesting that reading from padding will return 0 defeats the purpose of the field, especially in such cases...