svn commit: r286955 - in head/sys: net netinet netinet6

[Alexander V. Chernikov]

20.08.2015, 16:29, "Bjoern A. Zeeb" <bz at FreeBSD.org>:
>>  On 20 Aug 2015, at 13:13 , George Neville-Neil <gnn at neville-neil.com> wrote:
>>
>>  Why was this work not in Phabricator? This is a large change that has not been reviewed,
>>  as far as I can tell, by anyone else on the project. I am tempted to ask that this be backed out
>>  and reviewed BEFORE it goes into the tree.
>
> I probably wouldn’t go as far as backing out.
>
> Given this seems part of a larger change that seem to be streamed into HEAD and I have probably missed a posting or heads up or something on net@ [if someone could point me at that it would be very helpful], I appreciate things being broken into smaller reviewable bits. I have no clear idea of what the end goal is or will be or look like, which makes it hard to comment on design decisions.
This is really the part of large change. The posting itself is here: https://wiki.freebsd.org/ProjectsRoutingProposal
It was actually implemented (in slightly different way) in projects/routing branch. Unfortunately, I had to postpone that until recently. HEAD changed significantly since then, so merging is a bit complex.
I've got routing part merged (and I'll update projects/routing branch within a few days). However, there are bunch of "general" lle changes, related to code "cleanup". So, my plan was to merge these, update projects/routing branch so anyone can clearly see the difference on fresh HEAD and create a bunch of reviews with proposed locking / timer changes for lle (and do the same for routing changes). But currently we're here..

>
> I however, as I have stated in the past, would really really appreciate each change in PB, with the proposed commit message, and allow a few days of review or for people to comment on. I have spend two years on fixing bugs on the previous rewrite of that code. I am not exactly looking forward to see history repeat itself.
I see no problem in sending changes there for design changes and I actually do prefer PB for that. I actually not against using PB for everything, but here we come very close to exact definition of "a few days".
My intentions on the last commits were actually to ease the review of upcoming stuff, not to push-everything-without-review.

>
> Thanks,
> Bjoern