Andrew Morton on kernel development

[shogun1234]

And old news interviewing Andrew Morton

reference: http://www.linuxformat.co.uk/modules.ph ... le&sid=159

>Linux Format
>
>Andrew Morton on kernel development
>
>Articles / LXF magazine
>Posted by M-Saunders on Nov 21, 2005 - 10:37 AM
>
>In issue 74 of Linux Format, on the shelves now, we have an interview with kernel 2.6 maintainer Andrew Morton. Here are a few of the questions we asked Andrew, along with his answers:
>
>LXF: Do you think it was good to have had the time with BitKeeper in kernel development, or should they have stuck with CVS?
>
>AM: Well, we never even used CVS. Before BitKeeper we basically weren't using anything - just a bunch of patches sitting on Linus' hard drive and it uploaded occasionally. We had no tracking of what had gone in the kernel at all. Which I always thought was pretty poor form; I wish Linus had used CVS from day one. He hates CVS, he has real problems with it, but for just a simple linear evolving model I think CVS would have done the job and at least we would have had some tracking history.
>
>
>I think the main thing BitKeeper brought is it got Linus using something, and even if we used one of the free revision control systems I think we would have been 90% of the way there. Obviously BitKeeper at the time was probably the best version control system available and it certainly suited our development model well. I also was very uncomfortable with using BitKeeper, obviously because of the licensing and proprietary issues, but I could live with it. It was better than having nothing. My main game is improving the kernel, not getting involved with... religious issues.
>
>I always expected it to come to tears, I always expected the wheels would fall off the BitKeeper thing. It came a little bit more abruptly than I expected, but we recovered pretty well. We didn't take a dent, really, and we're still going.
>
>LXF: In terms of kernel features that have a buzz about them, Xen being one of them, and Inotify, which has just gone into 2.6.13, which things do you find interesting?
>
>AM: Well, Xen is obviously something that a lot of people want. I tend not to think about roadmaps too much, because it's just whatever people want to send. The decision-making about what goes into the kernel tends to be distributed. Linus and I don't sit there and say, "Hey, we need x, y and z by the third quarter in 2005".
>
>Instead, if some particular group has some particular requirement in the kernel they'll put the resources behind it and they'll do the work. So it's a push rather than a pull model in that sense. But what's in the future? Filesystem and user space I guess we should get in soon. Some people would like to see that. Oracle clustered filesystem [OCFS] we'll put into 2.6.14. The whole clustering issue is the big one. We've had trouble for a number of years with getting clustering structure into the kernel.
>
>There are a whole bunch of clustering projects out there but they all seem to have gone off in totally different directions and development teams don't seem to be able to find any common ground so we just haven't been moving forward on that. But OCFS appears to be a complete solution and looks OK, so I have no problem merging that in the kernel. I'm not sure that things like Red Hat's GFS and distributed lock manager will have such an easy path.
>
>What else... I don't know, in answer to your question! Whatever people send me.
>
>LXF: Is your -mm kernel source tree quite a long way from the actual true, officially published kernel?
>
>AM: It varies. Largely it's the kernel as it will be in a few month's time, minus quite a bit. So I pull all the various development trees into my kernel for testing purposes, and also various new features I'll line up in my kernel, and probably take them down to Linus when I think they're ready to go.
>
>So at times, yes, it can diverge a lot. In terms of patch size it can, I mean it's got the Reiser 4 filesystem, that's 2MB in one hit, and various other features. They make very little impact on the kernel overall, effectively they're just an add-on on the side, but it makes the patch big.
>
>For the full interview, in which Andrew discusses current 2.6 kernel quality, bugfixing, and more, grab a copy of LXF 74.
>
>This article is from Linux Format
> http://www.linuxformat.co.uk/
>
>The URL for this story is:
> http://www.linuxformat.co.uk/modules.ph ... le&sid=159
>