LWN: Comments on "Who wrote 3.0 - from two points of view"

Microsoft, co-author of the Linux kernel

dgm — Tue, 19 Jul 2011 10:15:56 +0000

IANAL either, but Microsoft is not distributing "the program" (the Linux Kernel) here. I don't know if the set of patches can be considered "the program" by themselves.

Can I download the kernel from somewhere in microsoft.com? Is there a git tree somewhere? All those could count as "distributing".

Microsoft, co-author of the Linux kernel

etienne — Tue, 19 Jul 2011 09:12:28 +0000

IANAL, but by contributing, don't they also give patent rights to "the Program"?
(For example, if a patent license would not permit royalty-free redistribution of the Program by all those who receive copies directly or indirectly through you, then the only way you could satisfy both it and this License would be to refrain entirely from distribution of the Program. )

Microsoft, co-author of the Linux kernel

leoc — Mon, 18 Jul 2011 18:38:11 +0000

it wants its fair share of the result.

If they are playing "fair", then why do they refuse to provide a list of the patents Linux allegedly violates? Seems to me that the only share they would ever consider fair is 100%.

Microsoft, co-author of the Linux kernel

giraffedata — Mon, 18 Jul 2011 15:25:13 +0000

Do you think Microsoft is getting a fair share ?

That's a little off-topic, since we're talking about whether there is some inconsistency or irony in Microsoft contributing code to Linux while claiming that Linux infringes some Microsoft patents. The only thing that would be relevant is whether Microsoft thinks it's getting its fair share (because if it does, then I must be wrong about Microsoft's motivation in making claims of Linux patent infringement).

Microsoft, co-author of the Linux kernel

dag- — Mon, 18 Jul 2011 14:40:57 +0000

Do you think Microsoft is getting a fair share ? I think they get more than they deserve, if they deserve anything at all. And no, we don't know how much they settle for patent-infringements with those large hardware vendors, but I don't think they deserve the tiniest bit.

Personally, if patents are there to protect corporate investments in research for the benefit of the general public, I do think the same general public should get full insight into patent settlements. I do think one (patent law) contradicts the other (closed patent-infringement settlements).

How can we make laws to protect the general public, if we don't know what backroom deals are made as a result of it? I also think if those deals were forced to be made public, support for patent law, especially in software development, will soon be impossible.

Microsoft, co-author of the Linux kernel

dgm — Sun, 17 Jul 2011 09:59:40 +0000

Look up TomTom vs. Microsoft. I may be instructive for you.

Microsoft, co-author of the Linux kernel

giraffedata — Fri, 15 Jul 2011 23:20:17 +0000

Isn't it a bit weird that Microsoft is simultaneously contributing code to the Linux kernel **and** claiming that it violates their patents?

Why? Because contributing code means you want to help Linux and claiming patent violation means you don't? I don't think it's that simple.

For one thing, Microsoft so far hasn't even prosecuted any of its patent claims. The main purpose for those claims seems to be to defend against insults against Microsoft of the form that a rag-tag bunch of people working for free put out something better than what the great Microsoft machine (or any other conventional software publisher) could make. The point of the claim is simply to say that Linux developers didn't do it alone.

Even if Microsoft were to sue over use of its patents in Linux, that wouldn't mean Microsoft doesn't like helping out Linux; only that it wants its fair share of the result.

Microsoft, co-author of the Linux kernel

leoc — Fri, 15 Jul 2011 21:02:49 +0000

Isn't it a bit weird that Microsoft is simultaneously contributing code to the Linux kernel **and** claiming that it violates their patents?

Who wrote 3.0 - from two points of view

mlankhorst — Fri, 15 Jul 2011 20:07:49 +0000

With background knowledge I meant all the things you have to worry about in kernel space that might not happen in userspace. Atomics, interrupt contexts, rcu, the fact that not the whole world is a VAX^Wx86.

When writing a bugfix for EFI I had to deal with early memory allocators and the transition to the full page allocator. For fun I looked up what allocators were used in the kernel. I came up with brk, e820, memblock, bootmem, xen's special brew, and the kernel's full page allocator. This list is probably incomplete. ;)

Who wrote 3.0 - from two points of view

Julie — Fri, 15 Jul 2011 19:28:10 +0000

the thing that probably is lacking is the background knowledge

I'm a bit confused by your comment, I had taken Jon's remark about it becoming difficult for new volunteer developers to participate as meaning the same as the above (newcomers don't _have_ that knowledge). Along with a couple of other potential difficulties:

1. Volunteers necessarily have less time than full-timers and therefore anything they swot up on at the time is likely to be dated when they come to try to put it into practice due to the current fast rate of development

2. The kernel's pretty much saturated with regards to experienced (sometimes full-time, paid) devs (at least in the areas of core work; I know Greg K-H is always looking for more contributors to staging).

What did you mean?

Following lwn for a few years helps a lot to learn about that. ;)

Well, I can't agree with you here - I think that following LWN for a few years has become pretty _essential_ for learning about that ;-)

Microsoft, co-author of the Linux kernel

patrick_g — Fri, 15 Jul 2011 06:21:36 +0000

If I understand correctly Steve Friedl's technical discussion, it was only about the 200+ initial patchs to clean the MS driver before entering -staging.
What about the 300+ patchs during this 3.0 cycle? It can't be coding style work because that was done previously. Perhaps he need to write another article and change his previous conclusion.

Microsoft, co-author of the Linux kernel

pr1268 — Fri, 15 Jul 2011 02:47:08 +0000

Somewhat amusing, but very enlightening. Thanks for the blog entry. I also found Steve Friedl's technical discussion you linked fun—it's reassuring that the Linux kernel source isn't polluted with DWORDs or HANDLEs or all other types in ALL_CAPS. Phew!

Microsoft, co-author of the Linux kernel

david.a.wheeler — Fri, 15 Jul 2011 00:01:57 +0000

These stats are so amazing that I just posted "Microsoft, co-author of the Linux kernel", to comment on this. You might find that article amusing.

Who wrote 3.0 - from two points of view

mlankhorst — Thu, 14 Jul 2011 23:48:20 +0000

I respectfully disagree with the assertion that linux kernel development is hard for newcomers. However the thing that probably is lacking is the background knowledge every kernel hacker takes for granted. Following lwn for a few years helps a lot to learn about that. ;)

MS only slightly behind IBM, really?

dgm — Thu, 14 Jul 2011 14:14:51 +0000

Wow. Should I take my skates in my next visit to hell?

MS only slightly behind IBM, really?

patrick_g — Thu, 14 Jul 2011 10:38:25 +0000

See the messages posted on LKML with mail addresses ending with "microsoft.com".

Who wrote 3.0 - from two points of view

neilbrown — Thu, 14 Jul 2011 08:35:48 +0000

Partly 2.6.1 had a lot fewer commits to subsequent releases. Commit counts for the first 10 are:

2.6.1 680
2.6.2 1252
2.6.3 1431
2.6.4 1407
2.6.5 1571
2.6.6 2020
2.6.7 2796
2.6.8 2650
2.6.9 4148
2.6.10 4510

and partly because Andrew Morton was getting a lot of patches attributed that didn't really belong to him. The counts for akpm patches in that first 10 are:

2.6.1 201
2.6.2 477
2.6.3 294
2.6.4 415
2.6.5 345
2.6.6 618
2.6.7 764
2.6.8 191
2.6.9 102
2.6.10 143

So about 25% of 2.6.1 and 2.6.2, and much less in the later releases.
(A quick look shows that the first few releases have several patches with
>>>> From: NeilBrown <neilb@cse.unsw.edu.au>
in the comment, but:
>>>> Author: Andrew Morton <akpm@osdl.org>
so I figure I should be in the exclusive "41-club" too!!)

MS only slightly behind IBM, really?

kragilkragil2 — Thu, 14 Jul 2011 08:07:40 +0000

But K. Y. Srinivasan is a Novell employee, right?
This could become a big headline so it is probably better to be safe.

Who wrote 3.0 - from two points of view

dlang — Thu, 14 Jul 2011 06:25:48 +0000

a combination of tracking problems and the fact that at the time many people were still using 2.4.x with little or no intention of _ever_ migrating to this scary new 2.6 thing.

Who wrote 3.0 - from two points of view

jamesmrh2 — Thu, 14 Jul 2011 05:35:26 +0000

Why was 2.6.1 was so commonly missed?