ipv6 steaming pile of poo...

3 messages, 2 authors, 2011-01-24 · open the first message on its own page

ipv6 steaming pile of poo...

From: David Miller <davem@davemloft.net>
Date: 2011-01-23 20:37:19

Thanks to the fact that I let a change causing known regressions live
in the tree for almost a year (don't worry I've learned my lesson):
From dc2b99f71ef477a31020511876ab4403fb7c4420 Mon Sep 17 00:00:00 2001
From: stephen hemminger <redacted>
Date: Mon, 8 Feb 2010 19:48:05 +0000
Subject: [PATCH] IPv6: keep permanent addresses on admin down

We have over a year of follow-up changes to revert if we want to
get rid of this turd.

This means getting rid of all the changes that tried to fix bugs in
the new ipv6 semantics made by Lorenzo Colitti, and the hlist
conversions, and all changes dependent upon those.

But no matter, this has to be done and I don't care if we lost lots
of cleanups and other work as a result of doing this, it just has to
be done.

And this is _why_ we must revert changes causing known regressions
as soon as they are discovered, not months or a year later.  Because
after that much time, undoing the damage becomes that much harder.

So if you folks wonder what I'm working on the next couple of days,
this is it.

Re: ipv6 steaming pile of poo...

From: Stephen Hemminger <hidden>
Date: 2011-01-23 23:20:49

On Sun, 23 Jan 2011 12:37:54 -0800 (PST)
David Miller [off-list ref] wrote:
Thanks to the fact that I let a change causing known regressions live
in the tree for almost a year (don't worry I've learned my lesson):

From dc2b99f71ef477a31020511876ab4403fb7c4420 Mon Sep 17 00:00:00 2001
From: stephen hemminger <redacted>
Date: Mon, 8 Feb 2010 19:48:05 +0000
Subject: [PATCH] IPv6: keep permanent addresses on admin down

We have over a year of follow-up changes to revert if we want to
get rid of this turd.

This means getting rid of all the changes that tried to fix bugs in
the new ipv6 semantics made by Lorenzo Colitti, and the hlist
conversions, and all changes dependent upon those.

But no matter, this has to be done and I don't care if we lost lots
of cleanups and other work as a result of doing this, it just has to
be done.

And this is _why_ we must revert changes causing known regressions
as soon as they are discovered, not months or a year later.  Because
after that much time, undoing the damage becomes that much harder.

So if you folks wonder what I'm working on the next couple of days,
this is it.
--
Please post the full set of diffs, to allow for putting the bug fixes
back.

Re: ipv6 steaming pile of poo...

From: David Miller <davem@davemloft.net>
Date: 2011-01-24 04:02:50

From: Stephen Hemminger <redacted>
Date: Mon, 24 Jan 2011 09:20:43 +1000
Please post the full set of diffs, to allow for putting the bug fixes
back.
I'm going to do one commit, with the list of commits I'm reverting
in the commit message so people can go back and go through the
history.
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help