Interview: Eben Moglen
Posted Aug 11, 2005 16:25 UTC (Thu) by iabervon
Parent article: Interview: Eben Moglen
I think he's minorly wrong about the linking behavior of the kernel; it at least used to be the case that a kernel module binary object couldn't be staticly linked into a kernel image, because it had the wrong entry points and such. Furthermore, the common model is that the end user links the module into the kernel, creating a work that can't be legally distributed (but it can't really be moved out of RAM anyway). I don't see the runtime legal concerns as substantially legally different from userspace code.
As for legal concerns over the creation of binary modules, it seems poor to me that the set of functions available to proprietary modules is determined by marking in the places where those functions are implemented, not by marking in the places where they are declared. It seems like a published API would need to list what parts of the API are available. And a special dispensation is clearly needed for functions which are intended to be part of the API, but are inlines.
to post comments)