Oracle is the #1 contributor to the core of Linux in 5.18 | Hacker News Reader

Oracle is the #1 contributor to the core of Linux in 5.18

blogs.oracle.com

125 pointsrascul81 comments

Open articleView on HN

Most active 5.18 employers [1]

- By changesets: Intel ( 11.4% ) ... Oracle ( 2.2% )

- By lines changed: AMD ( 39.9% )

[1] https://lwn.net/Articles/895800/

----------------

Most active 5.17 employers: [2]

- By changesets: Intel ( 11.6% ) ... Oracle ( 2.0%)

- By lines changed: Intel ( 12.6% )

[2] https://lwn.net/Articles/887559/

It looks like they're cheating the numbers by using the concept of "'core' of the kernel", whatever that means.

> Oracle is the #1 contributor to the "core" of the kernel, in terms of number of changesets.

While, I think most Intel and AMD contributions are very likely drivers related, it really feels like Oracle tried to find a statistic which fits their narrative.

Relevant XKCD: https://xkcd.com/1122/

To be honest I think the "core" distinction actually makes sense.

“Core of the kernel” reads like we’re getting lost in nomenclature

Nucleus of Linux-Kernel ;)

Don't you mean "the Nucleus of the Core of the Kernel"? :-D

No, the IBM Nucleus from the 60' ;)

Well, you NIP'd that in the IPL bud.

I've only used big iron in anger (as part of my company's "ways of doing things" - and all of it unexciting parts of my job, and many years ago). I was not aware of said Nucleus before you brought it up. So I guess today I learned something. Cheers!

That could actually make sense to me; drivers tend to be vendor specific, so I wouldn't feel too bad about giving credit for whoever does the most in the shared parts.

It appears that most of their contributions were to the XFS filesystem. Which means it is not used by "everyone" in a very similar sense to the way in which hardware drivers are not.

I suspect that if "core" were defined to exclude hardware _and_ filesystem independent changes Oracle would not have been at the top of the list.

IIRC AMD commits a large volume of generated data for drivers, so it's maybe not precisely what most people think of when they think of lines of code.

From your first link, Oracle is #14 down the list by changesets, behind Intel, Red Hat, Google, Linaro, Huawei, Facebook, Nvidia, SUSE, IBM, "None", "Unknown", and "Consultants"

By lines of code they don't even appear on the list.

> Counting commits or lines of code changed is a crude approximation of the activities of developers

so we came up with our own instead that just co-incidentally happens to reveal that we're #1!

And it still has this outdated mentality LoC = productivity.

I know I've said this before, but we really need to add "being utter horseshit" to the list of reasons to flag a submission.

... and therefore everyone using Linux owes us the following amounts of money...

I cannot recreate the results

> git clone https://git.kernel.org/pub/scm/linux/kernel /git/torvalds/linux.git/

> cd linux

> git log --pretty=format:"%<(60,mtrunc)%ae %h %s (%ar)" -i --no-merges v5.17..v5.18 -- fs mm kernel | grep oracle.com | wc -l

> 21

The article does not mention how the attribution to employer is done though...?

Could it be that some Oracle employed contributors are using a different email when submitting pull requests?

Going by the numbers from https://lwn.net/Articles/895800/

Intel is the top documented contributor to the overall kernel (as opposed to the "core" of the kernel, which is defined here has non-driver, non-architecture dependent code) with 11.4%, whereas Oracle has 2.2%.

7.7% are attributed to "unknown", 5.5% as "none", and 2.4% as "consultant". Almost 2/3 of that combined quantity would need to be attributed to Oracle in order to make up the gap, which seems unlikely.

So the definition of "core" seems to be the key factor here. Intel, AMD, IBM, Huawei, Linaro and Red Hat contribute substantial amounts of architecture-dependent code and/or driver code which isn't counted in this metric.

I can see removing drivers but taking out architecture-dependent code seems less justifiable.

edit: Turns out most of their contributions to "the core of Linux" are to the XFS filesystem, so I take it back, this is all silly.

> Could it be that some Oracle employed contributors are using a different email when submitting pull requests?

Yes, and if you contribute from an e-mail address that Greg K-H's scripts don't understand, you get an e-mail from him asking you to disclose your employer if you're willing.

Some companies mandate you contribute from your corporate address, and it is impossible to contribute from some corporate e-mail accounts, since they don't allow SMTP access for use with git send-email. For example, my understanding is that this is the reason for the linux.ibm.com subdomain, though someone at IBM can probably elaborate.

LWN's numbers are probably pretty close to accurate, as I think most people disclose. I haven't read TFA, but I guess Oracle is counting only commits to kernel/+fs/+net/ or something like that.

kernel/+fs/+mm :

git log --pretty=format:"%<(60,mtrunc)%ae %h %s (%ar)" -i --no-merges v5.17..v5.18 -- fs mm kernel

I wonder how many of the contributors in Oracle are Sun veterans.

Sun veterans have long cashed out and headed to greener pastures. The ones that are still coding are doing what they want, for whoever they want, for such is the market today for top tier programming skill.

What an absolute dream they are living. Apart from top tier programming skill it involves being extremely lucky.

I must be missing something because even by the command in their own blogpost, I see more commits from redhat.com and google.com than oracle.com. And if I try to poke around myself for things like device-agnostic networking code (e.g. the IPv4 stack, multipart TCP, etc.) the disparity increase in Red Hat's favor.

Be picky all you like, but to a long timer like me it’s awesome to see companies contribute.

I’d be personally fine with every contributor getting their own participation reward. Let them all get whatever (silly?) title they want out of it. Like news casters even.

I wish Oracle didn't kill SPARC... I really like my Sun workstations.

I really wish Sun didn’t run itself into the ground.

Hard to believe Oracle only paid $7.4 billion

https://abcnews.go.com/amp/Technology/story?id=7395780&page=...

AMP-free link: https://abcnews.go.com/Technology/story?id=7395780&page=1

Not only that, Oracle was the only company willing to actually pay for Sun's assets.

No one else saw any value in acquiring Sun (IBM was rather quick to withdraw their offer).

I wonder if Google is looking back in hindsight saying they should have bought them given the headache and cost of Oracle vs Google Java fight and the fact that Google is also now dabbling a lot more in hardware and making and commissioning their own chips.

Most likely, thankfully they haven't done so.

Looking at the mess Android Java is, and how great they are designing programming languages.

Surprised? The company that was selling very specific hardware and at the same time created popular programming language that was multi-platform, so customers could use cheap Intel-based servers instead of that "very specific hardware" (that was also very expensive) to run this software. What could have gone wrong?

The SUN tragedy was that it moved into service/support model too late, this move could not have saved them in the time-frame that all the investors wanted to see their money.

For me, personally, it was very, very sad, as SUN people, including sales, marketing were very nice, willing to cooperate even with small companies, there was a lot of "feeling" of the company run by tech-aware people who did not care about counting every cent.

SUN had a lot of really state of the art software for what support was selling pretty good, Glassfish server, Message Broker (one of the best on the market, really hard to beat) and many others... Still not enough money.

I hoped for some other deal, like selling hardware part to Fujitsu and keep working on Software/Cloud, etc. but, as I said, greedy people wanted their money right away. Oracle had big hopes for suing Google and winning a lot cash.

Business is hard, not always the good guy wins.

Larry Ellison mentions one of the most startling things they found was that Sun's sales org was literally selling the company to bankruptcy. Those nice sales people had commission based on sales contract value, with no requirement for profitability. That meant they would just keep selling contracts that lost Sun money, hand over fist.

I have a hard time taking anything he says at face value.

That's a pretty one-sided view of Sun in their prime. They were also horrendously overpriced, and when Linux came along their overpriced hardware and software was obsolete. They were much too late to wake up to that, and climb down from their castle.

I liked their hardware, and some of their software was amazing (DTrace, ZFS), but there was just no way to justify buying Sun hardware when any old PC running Linux would blow it out of the water.

Sun, IBM and the other Unix companies were in a bitter fight with MS, especially at the mid-tier.

Java, notably Enterprise Java, became the teeth of that fight. With Java, it let the hardware makers do what they do best -- make hardware. It let the OS designers do what they do best -- design OSes.

While the Unix "shell" experience was commoditized, the kernels certainly were not. The management of the systems still had proprietary magic pixie dust to help distinguish the products.

The fact that Java per se couldn't necessarily directly take advantage of these OS developments isn't germane since the bulk of the work, especially at the time, was in the back end databases. And those COULD take advantage of the platforms. But all of them could leverage the management capabilities of the platform.

Java was just the high level, portable business logic layer. Java was the safe solution that didn't require total commitment to a platform. Java was the rising tide to lift all boats that let the hardware folks sell and compete at the hardware level, especially against a company with monopolist, "lock in" tendencies such as Microsoft.

Write your software in Java, pick all of these platforms to deploy on, each shining in their own special way. By developing and promoting Java, which was not just Sun, IBM and others did a lot of Java work as well, it expanded the overall market for their hardware.

Sure, it opened the door to Intel et al, but Intel wasn't able to compete at the high end where Sun and IBM were able to compete.

Later, of course, as CPUs improved, there was less and less need for larger computers. Especially when "the network is the computer", allowing diverse, smaller computers to compete with larger monolithic systems.

Where Sun struggled was turning around from a hardware company to a services company. Oracle and IBM are both service companies. One happens to make a popular database, the other happens to make some, today, niche hardware. Sun, with its engineering core, wasn't really able to make that leap. HP tired to do the same thing, perhaps with less success.

But Java wasn't what killed Sun. Java helped them all rally and fight against MS, and thrive.

GlassFish server was a reference implementation for Java EE. It was quite slow. I’d say the best to come out of Sun was Java and ZFS.

I was a fan of the NetScape Directory Server. And the physical design language of the Sparc 5/10/20 & lunchbox IPC/IPX workstations.

cheap x86 hardware + Linux beat Sun.

And, yet, now we realize the way to differentiate in cloud is through different architectures. Instead of ARM, it could be SPARC.

Only when they provide a specific benefit, or are better cost/performance.

Different architectures will be best fits for different use cases. I bet even IBM LinuxONE machines would be competitive with generic x86 depending on the workload.

It seems especially unwise now in the light of Apples M1 M2 arm shining a new light on a old x64x86 world.

Whoever made that decision is going to appear in economic textbooks in the DONT Exampels.

Also, SPARC was one of the rare processors which had an integrated Ethernet interface, which allowed it to interface itself with outside worlds with negligible latency.

First generation Niagara swept the floor with Intel chips on server and XML tasks without breaking a sweat. With massive number of cores and Unstallable Pipelines (extremely fast context switching for its time IIRC), it was a drool-worthy architecture to read and indulge into.

It was good.

Those machines with like a 100 cores were miserable, they were all super slow. So it’s only useful if you have c code or something fast. Any kind of webserving code ran faster on a dual core amd machine

They were designed to be front end machines, and XML parsers. They were meant to handle the incoming workload and handle a lot of light threads. DB and other CRUD stuff was meant to be offloaded to backend servers with beefier processors.

They had shared FPUs, so anything ALU oriented fell flat on them, even Sun openly admitted that.

So they were not meant to be SMP supercomputer nodes, but to reduce number of front end server counts and allowed datacenter floor to be available for more backend servers and storage.

In the Ultrasparc T2 (Niagra 2), each core had its own floating point hardware.

The wiki says that the T2 was over 10x faster than the T1 for floating point arithmetic.

"One floating point unit per core, up from just one FPU for the entire chip... Better floating-point throughput (>10x improvement)... Better floating-point single-thread performance (>5x improvement)"

https://en.wikipedia.org/wiki/UltraSPARC_T2

The T1 and T2 were both open-sourced. I would love to have a dual-core T2 implementation on a Raspberry PI, with completely open hardware.

https://www.oracle.com/servers/technologies/opensparc-t2-pag...

> an integrated Ethernet interface, which allowed it to interface itself with outside worlds with negligible latency

Yes, the MPI ping-pong latency I measured between x2200s was the lowest I ever saw for 1GbE. However, the motherboards in them were a disaster area (unlike any other Sun kit I ever saw).

Unfortunately we were never able to procure SPARC based systems since we are running general purpose clusters, however we were able to get some Sun ZFS 74xx series storage.

We had its smaller sibling 7320 for testing for some time. I was lucky to make the benchmarks and tests on that, and it was unbelievable for its size, so we got its bigger sibling with a pretty hefty configuration.

It had a dtrace based zero overhead real-time monitoring dashboard and, the insights and optimization possibilities were unbelievable for a device of that era. It was basically a proper enterprise storage with all the interfaces, bells and whistles with the practicality of a home desktop NAS device.

Select LUN, select capacity & disk config, select share strategy, select interfaces and it's done. The fun part? It was able to saturate any interface on itself, incl. Infiniband.

It made sense at the time. It was too big for Sun to handle alone (well, with Fujitsu, IIRC). Sun had a lot of things on their plate at the time. Sadly, chip architecture was something they could drop without much loss.

Also remember, this was the era when Apple needed to switch from the PPC to Intel too. There was a major shift in the server market from older architectures (Sparc, Power, Alpha) where everyone was standardizing on x86. It’s only been recently where ARM has been performance competitive. Only then could the power savings from ARM be a killer feature from a $/unit work point of view.

And ARM gets to leverage many different users/fabs for chips. It’s a very different ecosystem than having a single company drive an entire architecture, regardless of how great the tech was. (And support software, porting, etc.)

Oracle isn't a hardware company like Apple and Sun. And while RISC has had a renaissance there's no guarantee that would have applied to SPARC given that SPARC's licensing is very different to ARM (and RISC-V).

So I wouldn't say it was a bad management decision to kill SPARC -- putting aside my personal love for the architecture and Sun's technologies in general.

Oracle does do hardware [1] but it is mainly used just to house backend Oracle DB better than much of commodity servers.

https://www.oracle.com/engineered-systems/exadata/

Yeah but isn't Exadata largely just off the shell components? At it's heart it's a high end Intel server running Linux. I wouldn't say this product makes Oracle a hardware company in the same way that Apple and Sun were. This just proves Oracle are a software company like Microsoft (ie they do dabble in hardware but it's the software that makes their margins).

I guess another way of putting it is this:

- do they build the software to sell their hardware,

- or build their hardware to sell their software.

In the case of Apple and Sun, it was the former. In the case of Oracle, it's the latter.

Exadata at least used to use Infiniband hardware inherited from Sun and not on anyone else's shelf.

Sparc != ARM

They are both RISC-ish but ARM has always had a focus on low power, SPARC was marketed for servers

There is no magic in any of them. ARM focused on low power because that was its niche. SPARC focused on performance and/or throughput because that's what Sun was building. And, well, x86 is focused on still being able to run MS-DOS.

I think it is different.

The reason ARM is doing well now, is because of massive economies of scale due to being the dominant architecture in phones.

Same with x86 Intel during that time. Because it was the dominant desktop architecture, Intel had a massive advantage in economies of scale and eventually gained such an advantage in fabs, that it made x86 chips the price/performance champs, despite any disadvantages of the actual architecture.

In retrospect, it was a huge mistake for Intel to ignore the mobile sector with x86. Even if it did not have the margins, the mobile sector allowed ARM to gain the economy of scale that it needed to compete with Intel on the desktop/laptop and server

As you mentioned ARM processors sell at lower margin than Intel processors. ARM as a company is not a highly profitable business itself. The company had always struggled as a public company (low valuation) and was finally privatized. The IP is great, but the actual business was not very good.

Look at the BOM cost of a typical Android phone. https://www.notebookcheck.net/The-Samsung-Galaxy-Note-20-Ult... The CPU (AP) is even cheaper than the storage, baseband, and display components. An Intel i5 sells at ~$200 at midrange while a high end ARM chip sells at $57.

If Intel diverted fab production to ARM chips, they would have made less money per chip. This is why Intel sold it’s ARM business in the 2000s. Apple does not care if ARM components are profitable, because they are a vertical and the sum of components is worth more than the individual component costs. Intel is a component maker so they are sensitive to component margin.

Intel attempted an x86 based mobile phone processor in the 2010s, but ultimately failed due to selling the chips at a loss. Making components for the mobile market is cut throat and the margins are too low.

Semiconductors business is extremely tough.

I've considered buying an IPX or IPC case and putting a few raspberry pis in it. That was such an awesome form factor.

Actually it was Oracle that introduced ADI, making Solaris SPARC the first UNIX with proper hardware memory tagging, while everyone else is still catching up.

The hate for Oracle apparently has blinded folks of the features that were introduced after they rescued Sun from being completly destroyed from debt collectors.

No-one thinks Oracle's technical execution is poor (other than when integrating acquisitions, which is impossible to do well). It's their corporate behaviour.

Their corporate behaviour is how they managed to have enough money in the bank to buy Sun's assets.

It's the way they chose to do it. Doesn't mean it was the only way.

Yes. It may sound cheesy, but there are definitely emotions one can feel for computer hardware/software or brands. It seems like i'd love to understand psychologically why that is. It seems to me that people like the feeling of using something that isn't the default. Maybe it feels bad to be too normal.

Really? It seems like it was one of the least compelling RISC instruction set outside of the fact that it was less proprietary.

What do Oracle employees tend to contribute to the Linux kernel?

From the article:

"A large part of this burst in contributions came from Matthew Wilcox's large patchset establishing Page Folios; also arriving in 5.18 were a number of XFS and mm fixes, and we're contining to work towards building online filesystem check and repair for XFS."

Drivers don't count but alternative filesystems do? XFS online check is "core of Linux"?

XFS is the default filesystem in RHEL

And x86-64 is the only supported desktop architecture but apparently code written to support it doesn't count...

They included "fs mm kernel". No idea how representative that is for the "core of Linux", but it does (presumably) include all filesystems, and not just XFS.

This is one of the more abysmal attempts by Oracle to market using the most obfuscated, obtuse, oblique bit of orifice-hosing imaginable.

Even on the off-chance they were the biggest contribs by lines of code, that is a stupid KPI to fixate on.

Oracle, IBM, Microsoft... I'm starting to trust linux less.

Fortunately they're all in competition with each other so they can't subvert the kernel to their own ends.

Before everyone pisses all over this because it’s Oracle, I think the better first of action is to challenge other companies to step up and try to step up.

Back in the day everyone used to whine incessantly about Sun Microsystems, despite being one of the most charitable contributers.

Before everyone pisses all over this because it’s Oracle

Did you miss the part where Oracle is twisting the truth by redefining what "core" means? You're right in that Oracle being a bunch of fucktrumpets doesn't mean we should dismiss everything they say out of hand. It just means it doesn't take long to figure out that, yep, they're still fucktrumpets.

Oh, I don't know. They've been like this for decades, at some point buying anything from Oracle, or believing anything they say, is just a laughable idea.

Oracle trying to not look evil.

This is really scary...

Reply on news.ycombinator.com