Ingo Molnar wrote:

In other words, the changelog is useless and even borderline 
deceptive. Not a good sign if you are trying to get a patch 
accepted to the kernel.

Furthermore, no performance figures were posted along with this 
modification - it only stated that these are "performance 
improvements". Especially in cases where a change slows down the 
common case the showing of a very substantial performance 
benefit is a must-have, before a patch is considered for 
upstream merging.

I think this is mainly a misunderstanding, Stephen posted these
patches as RFT so Rick and Eric could do benchmarks, they were
not intended for merging at this time.

In fact, the submission incorrectly stated:

| This patch set is against Patrick's netfilter next tree since
| it is where it should end up.
|      
| git.kernel.org:/pub/scm/linux/kernel/git/kaber/nf-next-2.6.git

This is wrong, the "netfilter next tree" is not where the "Add 
mod_timer_noact" change should end up, and you should ask your 
contributors to submit changes to other subsystems to their 
respective maintainer trees - the timer tree in this case.

Absolutely, I wouldn't have taken it, and Dave wouldn't have taken
it from me, so no cause for alarm :)