![[RSS]](/static/img/feed-icon-14x14.2168a573d0d4.png)
Conservancy Blog
Displaying posts tagged GPL
RHEL Panel Discussion at FOSSY 2023
by on July 19, 2023
This past weekend, July 13-16th, 2023, Software Freedom Conservancy (SFC) hosted and ran a new conference, FOSSY (Free and Open Source Software Yearly) in Portland, Oregon, USA. I was glad to host the keynote panel discussion on the recent change made by Red Hat (now a subsidiary of IBM) regarding the public source code releases for Red Hat Enterprise Linux (RHEL).
The panelists included (in alphabetical order) Jeremy Alison, software engineer at CIQ (focused on Rocky Linux) and Samba co-founder, myself, Bradley M. Kuhn, policy fellow at SFC, benny Vasquez, the Chair of the AlmaLinux OS Foundation, and James (Jim) Wright, who is Oracle’s Chief Architect for Open Source Policy, Strategy, Compliance, and Alliances.
Red Hat themselves did not reply to our repeated requests to join us on this panel, but we were able to gather the key organizations impacted by Red Hat's recent decision to cease public distribution of RHEL sources. SUSE was also invited but let us know they were unable to send someone on short notice to Portland for the panel.
We're very glad to make the video available to everyone who has been following this evolving story. FOSSY is a new event, and we've hopefully shown how running a community-led FOSS event here in Portland each summer creates an environment where these kinds of important discussions can be held to explore issues impacting FOSS users around the world.
I thank our panelists again for booking last-minute travel to be with us for this exciting panel and thank all the FOSSY attendees for their excellent questions during the panel.
I hope to see all of you at next years' FOSSY!
A Comprehensive Analysis of the GPL Issues With the Red Hat Enterprise Linux (RHEL) Business Model
by on June 23, 2023
This article was originally published primarily as a response to IBM's Red Hat's change to no longer publish complete, corresponding source (CCS) for RHEL and the prior discontinuation of CentOS Linux (which are related events, as described below). We hope that this will serve as a comprehensive document that discusses the history of Red Hat's RHEL business model, the related source code provisioning, and the GPL compliance issues with RHEL.
For approximately twenty years, Red Hat (now a fully owned subsidiary of IBM) has experimented with building a business model for operating system deployment and distribution that looks, feels, and acts like a proprietary one, but nonetheless complies with the GPL and other standard copyleft terms. Software rights activists, including SFC, have spent decades talking to Red Hat and its attorneys about how the Red Hat Enterprise Linux (RHEL) business model courts disaster and is actively unfriendly to community-oriented Free and Open Source Software (FOSS). These pleadings, discussions, and encouragements have, as far as we can tell, been heard and seriously listened to by key members of Red Hat's legal and OSPO departments, and even by key C-level executives, but they have ultimately been rejected and ignored — sometimes even with a “fine, then sue us for GPL violations” attitude. Activists have found this discussion frustrating, but kept the nature and tenure of these discussions as an “open secret” until now because we all had hoped that Red Hat's behavior would improve. Recent events show that the behavior has simply gotten worse, and is likely to get even worse.
What Exactly Is the RHEL Business Model?
The most concise and pithy way to describe RHEL's business model is:
“if you exercise your rights under the GPL, your money is no good
here”. Specifically, IBM's Red Hat offers copies of RHEL to its
customers, and each copy comes with a support and automatic-update
subscription contract. As we understand it, this contract
clearly states
that the terms do not intend to contradict any rights to copy, modify,
redistribute and/or reinstall the software as many times and as many places
as the customer likes (see §1.4). Additionally, though, the contract indicates that
if the customer engages in these activities, that Red Hat reserves the
right to cancel that contract and make no further contracts with the
customer for support and update services. In essence, Red Hat requires their customers
to choose between (a) their software freedom and rights, and (b) remaining a Red Hat
customer. In some versions of these contracts that we have reviewed, Red
Hat even reserves the right to “Review” a customer (effectively a
Red Hat's lawyers clearly take the position that this business model complies with the GPL (though we aren't so sure), on grounds that that nothing in the GPL agreements requires an entity keep a business relationship with any other entity. They have further argued that such business relationships can be terminated based on any behaviors — including exercising rights guaranteed by the GPL agreements. Whether that analysis is correct is a matter of intense debate, and likely only a court case that disputed this particular issue would yield a definitive answer on whether that disagreeable behavior is permitted (or not) under the GPL agreements. Debates continue, even today, in copyleft expert circles, whether this model itself violates GPL. There is, however, no doubt that this provision is not in the spirit of the GPL agreements. The RHEL business model is unfriendly, captious, capricious, and cringe-worthy.
Furthermore, this RHEL business model remains, to our knowledge, rather unique in the software industry. IBM's Red Hat definitely deserves credit for so carefully constructing their business model such that it has spent most of the last two decades in murky territory of “probably not violating the GPL”.
Does The RHEL Business Model Violate the GPL Agreements?
Perhaps the biggest problem with a murky business model that skirts the line of GPL compliance is that violations can and do happen — since even a minor deviation from the business model clearly violates the GPL agreements. Pre-IBM Red Hat deserves a certain amount of credit, as SFC is aware of only two documented incidents of GPL violations that have occurred since 2006 regarding the RHEL business model. We've decided to share some general details of these violations for the purpose of explaining where this business model can so easily cross the line.
In the first violation, a large Fortune 500 company (which we'll call Company A), who both used RHEL internally and also built public-facing Linux-based products, decided to create a consumer-facing product (which we'll call Product P) based primarily on CentOS Linux, but P included a few packages built from RHEL sources. Company A did not seek nor ask for support or update services for this separate Product P. Red Hat later became aware that Product P contained some part of RHEL, and Red Hat demanded royalty payments for Product P. Red Hat threatened to revoke the support and update services on Company A's internal RHEL servers if such royalties were not paid.
Since Company A was powerful and had good lawyers and savvy business development staff, they did not acquiesce. Company A ultimately continued (to our knowledge) on as a RHEL customer for their internal servers and continued selling Product P without royalty payments. Nevertheless, a demand for royalties for distribution is clearly a violation as that demand creates a “further restriction” on the permissions granted by GPL. As stated in GPLv3:
You may not impose any further restrictions on the exercise of the rights granted or affirmed under this License. For example, you may not impose a license fee, royalty, or other charge for exercise of rights granted under this License.
Red Hat tried to impose a further restriction in this situation, and therefore violated the GPL. The violation was resolved since no royalty was paid and Company A faced no consequences. SFC learned of the incident later, and informed Red Hat that the past royalty demand was a violation. Red Hat did not dispute nor agree that it was a violation, and did informally agree such demands would not be made in future.
In another violation incident, we learned that Red Hat, in a specific non-USA country, was requiring that any customer who lowered the number of RHEL machines under service contract with Red Hat sign an additional agreement. This additional agreement promised that the customer had deleted every copy of RHEL in their entire organization other than the copies of RHEL that were currently contracted for service with Red Hat. Again, this is a “further restriction”. The GPL agreements give everyone the unfettered right to make and keep as many copies of the software as they like, and a distributor of GPL'd software may not require a user to attest that they've deleted these legitimate, licensed copies of third-party-licensed software under the GPL. SFC informed Red Hat's legal department of this violation, and we were assured that this additional agreement would no longer be presented to any Red Hat customers in the future.
In both these situations, we at SFC were worried they were merely a “tip of the proverbial iceberg”. For years, we have heard from Red Hat customers who are truly confused. It's common in the industry to talk about RHEL “seat licenses”, and many software acquisition specialists in the industry are not aware of the nuances of the RHEL business model and do not understand their rights. We remain very concerned that RHEL salespeople purposely confuse customers to sell more “seat licenses”. It's often led us to ask: “If a GPL violation happens in the woods, and everyone involved doesn't hear it, how does anyone know that software rights have indeed been trampled upon in those woods?”. As we do for as many GPL violation reports as we can, we zealously pursue RHEL-related GPL violations that are reported to us, and if you're aware of one, please do email us at <compliance@sfconservancy.org> immediately. We fear that be it through incompetence or malice, many RHEL salespeople and business development professionals may regularly violate GPL and no one knows about it. That said, the business model as described by IBM's Red Hat may well comply with the GPL — it's just so murky that any tweak to the model in any direction seems to definitely violate, in our experience.
Furthermore, Red Hat exploits the classic “caveat emptor” approach — popular in many a shady business deal throughout history. While, technically speaking, a careful reader of the GPL and the RHEL agreements understands the bargain they're making, we suspect most small businesses just don't have the FOSS licensing acumen and knowledge to truly understand that deal.
Why Was an Independent CentOS So Important?
Until Red Hat's “aquisition” of CentOS in early 2014, CentOS provided an excellent counterbalance to the problems with the RHEL business model. Specifically, CentOS was a community-driven project, with many volunteers, supported by some involvement from small businesses, to re-create RHEL releases using the CCS releases made for RHEL. Our pre-2014 view was that CentOS was the “canary in the murky coalmine” of the RHEL business. If CentOS seemed vibrant, usable, and a viable alternative to RHEL for those who didn't want to purchase Red Hat's updates and services, the community could rest easy. Even if there were GPL violations by Red Hat on RHEL, CentOS' vibrancy assured that such violations were having only a minor negative impact on the FOSS community around RHEL's codebase.
Red Hat, however, apparently knew that this vibrant community was cutting into their profits. Starting in 2013, Red Hat engaged in a series of actions that increased their grip. First, they “acquired” CentOS. This was initially couched as a cooperation agreement, but Red Hat systematically made job offers that key CentOS volunteers couldn't refuse, acquired the small businesses who might ultimately build CentOS into a product, and otherwise integrated CentOS into Red Hat's own operations.
After IBM acquired Red Hat, the situation got worse. Having gotten rights to the CentOS brand as part of the “aquisition”, Red Hat slowly began to change what CentOS was. CentOS Linux quickly ceased to be a check-and-balance on RHEL, and just became a testing ground for RHEL. Then, in 2020, when most of us were distracted by the worst of the COVID-19 pandemic, Red Hat unilaterally terminated all CentOS Linux development. Later (during the Delta variant portion of the pandemic in late 2021) Red Hat ended CentOS Linux entirely. IBM's Red Hat then used the name “CentOS Stream” to refer to experimental source packages related to RHEL. These were (and are) not actually the RHEL source releases — rather, they appear to be primarily a testing ground for what might appear in RHEL later.
Finally, Red Hat announced two days ago that RHEL CCS will no longer be publicly available in any way. Now, to be clear, the GPL agreements did not obligate Red Hat to make its CCS publicly available to everyone. This is a common misconception about GPL's requirements. While the details of CCS provisioning vary in the different versions of the GPL agreements, the general principle is that CCS need to be provided either (a) along with the binary distributions to those who receive, or (b) to those who request pursuant to a written offer for source. In a normal situation, with no mitigating factors, the fact that a company moved from distributing CCS publicly to everyone to only giving it to customers who received the binaries already would not raise concerns.
In this situation, however, this completes what appears to be a decade-long plan by Red Hat to maximize the level of difficulty of those in the community who wish to “trust but verify” that RHEL complies with the GPL agreements. Namely, Red Hat has badly thwarted efforts by entities such as Rocky Linux and Alma Linux. These entities are de-facto the intellectual successors to CentOS Linux project that Red Hat carefully dismantled over the last decade. These organizations sought to build Linux-based distributions that mirrored RHEL releases, and it is now unclear if they can do that effectively, since Red Hat will undoubtedly capriciously refuse to sell them exactly-one RHEL service and update “seat license” at a reasonable price. It appears that, as of this week, one must have at least that to get timely access to RHEL CCS.
What Should Those Who Care About Software Rights Do About RHEL?
Due to this ongoing bad behavior by IBM's Red Hat, the situation has become increasingly complex and difficult to face. No third party can effectively monitor RHEL compliance with the GPL agreements, since customers live in fear of losing their much-needed service contracts. Red Hat's legal department has systematically refused SFC's requests in recent years to set up some form of monitoring by SFC. (For example, we asked to review the training materials and documents that RHEL salespeople are given to convince customers to buy RHEL, and Red Hat has not been willing to share these materials with us.) Nevertheless, since SFC serves as the global watchdog for GPL compliance, we welcome reports of RHEL-related violations.
We finally express our sadness that this long road has led the FOSS community to such a disappointing place. I personally remember standing with Erik Troan in a Red Hat booth at a USENIX conference in the late 1990s, and meeting Bob Young around the same time. Both expressed how much they wanted to build a company that respected, collaborated with, engaged with, and most of all treated as equals the wide spectrum of individuals, hobbyists, and small businesses that make the plurality of the FOSS community. We hope that the modern Red Hat can find their way back to this mission under IBM's control.
(Software) Repair info on EnergyGuide labels: Conservancy replies to FTC's request
by on December 21, 2022
Software Freedom Conservancy has today submitted its reply to the FTC's request for comments on how repair information should be displayed on EnergyGuide labels. In particular, SFC has recommended that the FTC mandate a "Software Repair Instructions" section on the EnergyGuide labels that are already required on a variety of home appliances, including televisions, refrigerators, clothes washers, and dishwashers. This would not be a new notice requirement for most manufacturers, since it (currently) only requires manufacturers to provide the notice when they already had obligations under copyleft licenses to offer source code already. This merely changes the prominence of such notices, so that users can more easily see which products contain copylefted software (and thus software repair instructions) or not. This is important because many manufacturers make efforts to deemphasize or obscure their offers (if they have them at all), which prevents consumers from learning that they have rights with respect to their software.
We are very happy to see the FTC requesting comments on how repair information for home appliances can be better provided to purchasers of these products. While the FTC's EnergyGuide labeling program started out as a way for purchasers to better assess how much energy each appliance would likely use, and approximately how much that would cost them, the FTC has been taking a more holistic view of how appliance purchases impact the world, not just in terms of how much energy they consume while operating, but also how much energy is required to manufacture them and, consequently, how we can reduce the number of appliances going into landfills, reducing the number of new appliances that need to be manufactured. Free and open source software provides many answers to these repair and longevity questions, and we hope that appliance purchasers will be made more aware of this through the FTC's updated labeling requirements.
By making a lot more people aware that software repair information is available for a device, the chance of a repair community forming for that class of devices increases dramatically. And these communities are immensely helpful to device owners, both for fixing problems that may arise in the software (which can be shared quickly and easily after one person makes them to anyone with that device, regardless of their level of technical expertise), but also for maintaining that software long after the manufacturer has stopped supporting it, meaning they can keep that device operating safely for years to come rather than having to dispose of it, which increases landfill usage and needless new device purchases. We already have several examples of such communities, including SamyGO for older Samsung TVs, LineageOS for most Android phones, and OpenWrt for wireless routers. SFC has fought extensively to protect the right to install your own firmware on your devices. By showing people that software repair information is available to them, we can build many many more communities like these, keeping more devices lasting longer (and better serving their users' needs), and fewer devices in our landfills.
We recommend those interested in this issue read our submission to the FTC, and consider whether to make their own submission in support of this or similar (especially hardware) repair information requirements. While we hope our own submission carries weight and is deemed relatively easy to implement given that it requires no new information to be provided by most manufacturers, it would help for others to provide their own experiences with lack of easily-accessible software repair information to the FTC so they are aware of the extent of the problem. The comment period is open until December 27 (likely to be extended until January 31, 2023) and you can see more details about the FTC's request for submissions and submit your own comment here.
For those that do read our submission, note that the FTC has trimmed some of its attachments from the website. You can find the attachments here instead:
You may notice that SFC has suggested the FTC require manufacturers to provide a URL to their source code distribution website, while not mentioning other ways of fulfilling an offer for source code, which we normally request that manufacturers provide (such as offering the source code on a durable physical medium, e.g. a USB stick or optical disc). Our main reason for this usual request that manufacturers provide source code on a durable physical medium is that not everyone in the world has a reliable or fast Internet connection. As a result, if a manufacturer only provides source code over the Internet, the most disadvantaged people are further disadvantaged by not being able to download the source code for their device (most source releases are hundreds of megabytes, if not more).
With our reply to the FTC, we were trying to make the best argument based on current practices and the least amount of additional work for manufacturers (to improve the chance of our suggestion being adopted, and reduce the chance that a company could make any credible argument against it), while also keeping in mind the jurisdiction this ruling applies to (USA) and its Internet connectivity standards. Though not complete yet, the National Broadband Plan in the USA does have this aim: "Every American should have affordable access to robust broadband service". Given the balance of people in the USA already connected to broadband, and the strong intent to connect the rest, we felt it was practical to make the recommendation include only web-accessible source code as the labeling requirement applies only in the USA. Note that we still request manufacturers make source code available on a durable physical medium, and would advise the FTC to make this part of their labeling requirements as well if they felt it feasible to include.
Although we have much work to do to ensure that people purchasing free and open source software (as part of appliances and other devices they may buy) know that they can repair, maintain, and modify this software, steps like this from the FTC will bring us closer. We are looking forward to the FTC's decision on our recommendation, and hope to help more people access the information they need to make their devices work for them, for as long as they choose to keep them. Together we can improve our own lives, but also the lives of others, and our planet.
Trademark Was Made to Prevent Attack of the “Clones” Problem in App Stores
by on July 11, 2022
Suppose you go to your weekly MyTown market. The market runs Saturday and Sunday, and vendors set up booths to sell locally made products and locally grown and produced food. On Saturday, you buy some delicious almond milk from a local vendor — called Al's Awesome Almond Milk. You realize that Al's Awesome would make an excellent frozen dessert, so you make your new frozen dessert, which you name Betty's Best Almond Frozen Dessert. You get a booth for Sunday for yourself, and you sell some, but not as much as you'd like.
The next week, you realize you might sell more if you call it Al's Awesome Almond Frozen Dessert instead of your own name. Folks at the market know Al, but not you. So you change the name. Is this a morally and legally acceptable thing to do?
This is a question primarily regarding trademarks. We spend a lot of time in the Free and Open Source Software (FOSS) community talking about copyrights and patents, but another common area of legal issues that face FOSS projects (in addition to copyright and patent) is trademark.
In fact, FOSS projects probably don't spend enough time thinking about their trademark. Nearly ten years ago, Pam Chestek — a lawyer and expert in trademark law as it relates to FOSS and board member of OSI — gave an excellent talk at FOSDEM (2013), wherein she explored how FOSS projects can use trademarks better and to ensure rights of consumers — particularly when dealing with bad actors. Our own Executive Director, Karen Sandler, had also spoken about this issue as well. These older talks, in turn, spawned an ongoing conversation that continues to this day in FOSS policy circles.
Specifically, last week, we learned that the Microsoft Store was changing their policies, ostensibly to deal with folks (probably some of whom are unscrupulous) rebuilding binaries for well-known FOSS projects and uploading them to the Microsoft Store. Yet, this is a longstanding issue in FOSS policy. FOSS experts in this area would have been happy to share what's been learned over the last ten years of studying this issue.
The problem Microsoft faces here is the same problem that the MyTown market folks face if you show up trying to sell Al's Awesome Almond Frozen Dessert. The store/market can set rules that you will no longer be able to sell if you are found to infringe the trademark of another seller. The market could simply require the trademark holder to take trademark action themselves, or it could offer some form of assistance, arbitration, or other-extra-legal resolution mechanism0.
There is often temptation in FOSS to give special status to maintainers, or the original developer, or the copyright holder, or some other entity that is considered “official”. In FOSS, though, the only mechanism of officialness is the trademark — the name of the upstream project (or the fork). The entire point of FOSS is that for the code itself, everyone should have equal rights to the original developers, to the maintainers, or to any other entity.
We have faced this with our member project, Inkscape. While the Inkscape Project Leadership Committee has chosen not to charge for the version of Inkscape that they upload on Microsoft Store, we did see this very problem for many years before these app stores even existed. Namely, it was common for third-parties to sell Windows binaries on CD's for Inkscape in an effort to make a quick buck. We did trademark enforcement in these cases — not forbidding these vendors from selling — but simply requiring the vendors to clearly say that the product was a modified version of Inkscape. Or, if it was unmodified redistribution of Inkscape's own binaries, we required the vendor to note that the Inkscape project's website was the official source for these binaries.
I have often written to complain about copyright and patent law. I have my complaints about trademark law (and I've seen trademark grossly abused, even), but trademark laws tenets are really reasonable and solid: to ensure consumers know the source and quality of the products they receive.
The problem of concern here is one well handled by trademark. It doesn't need excessive app store rules; we don't need FOSS licenses to be usurped or superseded by Draconian policy. And, this solution to this particular problem has been long-known by FOSS. Pam's talk in 2013 explained it quite well!
The MyTown Market doesn't need to create a policy that forbids you from buying Al's Awesome Almond Milk on Saturday and reselling a product based on it on Sunday. They just need to let Al know his rights under trademark, and maybe offer a lightweight provisional suspension of your booth if the trademark complaint seems primia facie valid. But, most importantly, before it announces new rules with a 30 day clock, MyTown's leadership really should discuss it with the citizens first to find a policy that takes into account concerns of the people. Even if they fail to do that, there are MyTown's elected officials whose actions are accountable to the people. App store companies are accountable only to their shareholders, not the authors of the apps. Companies could benefit by learning that the FOSS community prioritizes respecting authors, protecting consumers' and users' rights, and by understanding that the line between user and contributor should blur. The FOSS marketplace functions because the community works.
Footnotes
0 I hesitate to even suggest that an app store should create an extra-legal process regarding trademark enforcement beyond the typical governmental mechanisms — lest they decide they have to do it. A major problem with app stores is that they create rules for software distribution that are capricious, and arbitrary. We all do want FOSS available on Microsoft, Apple, and Google-based platforms — and as such are forced to negotiate (or, rather, try to negotiate) for FOSS-friendly terms. Ultimately, though, the story of major vendor-controlled app stores is always the story of “just barely” being able to put FOSS on them, because the goal of these entities is to profit themselves, not serve the community. We prefer app stores like F-Droid that are community-organized and are not run for-profit.
