Re: libceph: rework behavior for not recognized tag codes

Hi Slava,

The hardening itself looks reasonable to me, but I have a couple of small
comments before adding my Reviewed-by.

The Fixes tag looks like it points at the msgr1 field split rather than the
commit that introduced the WARN_ON() behavior. If the original commit is hard
to identify, I think it would be better to drop the Fixes tag and describe this
as hardening.

Also, could you include the peer identity/address in the new connect_seq
mismatch message? The nearby protocol feature mismatch errors include that
context, and it would make this useful when multiple connections are active.

Splitting the connect_seq change and the bad_tag ratelimit into separate
patches would also be cleaner, but I don't feel strongly about that.

-- 
Alex Markuze