VMWare Taken to Court Over GPL Violation
sfconservancy.org
VMWare Taken to Court Over GPL Violation
1–10 of 299 posts
Re: VMWare Taken to Court Over GPL Violation
#2Is there any information about why VMware thinks they are not bound by the GPL? Curious whether they have a technical justification, rather than simply being a blatant licensing violation.
Re: VMWare Taken to Court Over GPL Violation
#3Is there any information about why VMware thinks they are not bound by the GPL? Curious whether they have a technical justification, rather than simply being a blatant licensing violation.
This is just speculation but I suspect this is a matter of being caught with their pants down. They violated the GPL, but to comply with the license would mean that they'd have to release a lot of code that they don't want to release.
From the page, it seems like there's been ongoing negotiation, so I am guessing there's been a lot of stalling.
Re: VMWare Taken to Court Over GPL Violation
#4I'm surprised it took this long. I can't really blame them, after 15 years of ignoring the GPL, they probably figured nobody would ever come up with the financial resources to actually go after them. This is going to take years to come to a resolution and millions and millions of dollars in lawyers fees.
Re: VMWare Taken to Court Over GPL Violation
#5Is there any information about why VMware thinks they are not bound by the GPL? Curious whether they have a technical justification, rather than simply being a blatant licensing violation.
Don't know specifically, but these usually come down to hair-splitting over terms like "linking" and "derived work" and so on. I guess VMware think they can convince a judge that the way their code interacts with the Linux kernel does not cause it to fall under the GPL's source-publication requirement. On the other hand, I know Christoph enough to know how insanely smart and persistent he is. If he's confident enough in his conclusions to go forward with this, I wouldn't want to be on the other team.
Re: VMWare Taken to Court Over GPL Violation
#6I'm surprised it took this long. I can't really blame them, after 15 years of ignoring the GPL, they probably figured nobody would ever come up with the financial resources to actually go after them. This is going to take years to come to a resolution and millions and millions of dollars in lawyers fees.
I agree. If you want to help the little guy here, then pleased donate to fund the lawsuit:
http://sfconservancy.org/linux-compliance/vmware-lawsuit-app...
Re: VMWare Taken to Court Over GPL Violation
#7Is there any information about why VMware thinks they are not bound by the GPL? Curious whether they have a technical justification, rather than simply being a blatant licensing violation.
I'm not fully up-to-date on this case, but there is a bit of wiggle room on what counts as a derivative work, even in kernelspace.
My favorite example is OpenAFS. Development of the AFS kernel module started in 1983, so the code is a decade older than Linux itself. So it's hard to argue that openafs.ko is a derivative work of Linux, and since it continues to run on several kernels other than Linux (Darwin, Solaris, NT, etc.) and the Linux-specific code is just glue code to make it work, it's not really obvious that it's become a derivative work of Linux. (OpenAFS is free software under the IBM Public License, which is a GPL-incompatible free software license, and the rightsholders will not relicense. The Linux kernel infrastructure assumes that any non-GPL kernel module is proprietary, for extra amusement.)
Along the same lines is the kernel module for VMware Workstation, the desktop product that runs on Linux as well as Windows and Mac OS X. It doesn't predate Linux, so the argument is slightly less clear.
VMware's vmkernel, as I understand it, is an amalgamation of VMware's hypervisor code (that predated ESX, and originated on Workstation) plus Linux. I think it conceptually started off as a Linux server with the existing kernel module, but both have evolved quite a bit. I think the remaining Linux parts mostly exist as hardware drivers, and VMware is providing the core kernel routines like scheduling and memory management and "world" management (equivalent to process management, but it's a hypervisor).
I think that VMware's argument is that they did not start with Linux and add proprietary features to existing GPL'd code and call it proprietary; they started with their own code and linked in parts of Linux via well-defined interfaces, and they're happy to comply with the GPL for the parts that came from Linux, but they don't want to GPL the entirety of the ESXi kernel.
It's worth noting that the concept of address space isn't a legal concept, just a technical one. We generally acknowledge that it's not a GPL violation to run proprietary applications on Linux, even applications that use Linux-specific interfaces like cgroups and epoll and signalfd and all that good stuff. Is this simply because the Linux kernel copyright explicitly disclaims virality to userspace, or because there's some fundamental legal reason why userspace and kernelspace are far enough apart? Can two components in ring 0 also be "far enough apart"?
Re: VMWare Taken to Court Over GPL Violation
#8So this could be really interesting from a precedent standpoint because it could help clarify the legality of the GPL in various contexts in the court.
Whatever you think about the GPL (and I'm for open source and open licenses but I do have issues with GPL v3, but that's just me), this has been necessary, I believe, in order for GPL and OSS projects to move forward.
Re: VMWare Taken to Court Over GPL Violation
#9So this could be really interesting from a precedent standpoint because it could help clarify the legality of the GPL in various contexts in the court. Whatever you think about the GPL (and I'm for open source and open licenses but I do have issues with GPL v3, but that's just me), this has been necessary, I believe, in order for GPL and OSS projects to move forward.
I'm not sure it will be that groundbreaking; this seems to be yet another busybox violation litigated in Germany, of which we have had several already.
Re: VMWare Taken to Court Over GPL Violation
#10Is there any information about why VMware thinks they are not bound by the GPL? Curious whether they have a technical justification, rather than simply being a blatant licensing violation.
I'm not fully up-to-date on this case, but there is a bit of wiggle room on what counts as a derivative work, even in kernelspace. My favorite example is OpenAFS. Development of the AFS kernel module started in 1983, so the code is a decade older than Linux itself. So it's hard to argue that openafs.ko is a derivative work of Linux, and since it continues to run on several kernels other than Linux (Darwin, Solaris, N…
From the wording in the link it sounds like they include modified GPL code and don't publish that, not that they are expected to open-source ESXi (or whatever it is called right now) entirely:
> In 2011, Conservancy discovered that VMware had failed to provide nor offer any source code for the version of BusyBox included in VMware's ESXi products (as required by BusyBox's license, GPLv2).