Re: [PATCH] for_each_zone / for_each_pgdat

[Mike Fedyk <mfedyk@matchmail.com>]

On Tue, Apr 16, 2002 at 10:00:36AM -0400, Bill Davidsen wrote:
> On Mon, 15 Apr 2002, Mike Fedyk wrote:
> 
> > No matter how much someone can go through their own code and say "it's
> > ready" there's always a good chance there is some bug that will trigger
> > under testing.  Also, Andrew found a problem with your locking changes when
> > he split up your patch, and at the time you were saying it is ready and
> > there were no bug reports against in...
> 
> If you are going to reject code from people who send in code which turns
> out to have bugs you are going to have a VERY small set of submitters.

No that's not what I was saying.

> It's good to have someone else read the code, for breakup or whatever, but
> to avoid cleanup in a stable kernel seems long term the wrong direction.
> 

Exactly.  I'm just saying that you will get more eyes on the code and less
possible detrimental impact (if any, which I doubt) if the patches don't all
go into one set of -pre patches but spread out over a few releases
(2.4.19,20 and possibly 21).  The -pre kernels get testing, but not nearly
as much as the releases do.  Test the -pre and -rc kernels as much as
possible, but also know that something might be flushed out by some people
that only use the released kernels (non -pre or -rc).

Mike