On Fri, Jan 27, 2006 at 07:33:05AM -0500, John W. Linville wrote: > On Fri, Jan 27, 2006 at 09:17:50AM +0000, Christoph Hellwig wrote: > > > Folks, please stop these stupid ideas. There's a free driver, let's improve > > and merge it. That's a thousand times better than any half-free driver > > with buggy binary blobs. > > I presume you mean the ath-driver.org stuff? I'm perfectly > open to using that code, but some had raised issues about the > reverse-engineering process used to create it. Has that been resolved?
It is unclear. I'm in contact with lawyers from the institute for low and open source software here in Germany, who conducted an analysis on reverse engineering. The result of their analysis leaves more open questions than it answers. The first problem is that there appears to be no precedent whatsoever on the reverse engineering interoperability exceptions. Also, the wording of those exceptions seem to assume software/software interoperability, not software/hardware. Furthermore, the .de implementation on the EU directive on 'reverse engineering' explicitly states 'decompilation' is forbidden. Technically speaking, that's different from disassembly. But what would a court rule? Also, the interoperability exceptions seem to be written with closed-source software in mind, since the result of such conditionally allowed re-engineering is not allowed to be published. So for now, I personally would ignore the concerns. Please tell me how many Linux device drivers we would have, if nobody had ever looked into some disassembled proprietary driver? Don't get me wrong, I'm very much against any kind of 'carbon copying' code from a proprietary driver into a free one. However, merely looking at existing proprietary code in order to figure out which bit of a register resembles what function, and then implementing a driver from scratch using that information seems pretty safe to me. Actually, running the proprietary code in a simulator and just looking at the inputs/outputs of that simulator seems to be the safest way. Also, who will be able to prove that you actually looked at some assembly instructions rather than observing the proprietary program running in a simulator? -- - Harald Welte <[EMAIL PROTECTED]> http://gnumonks.org/ ============================================================================ "Privacy in residential applications is a desirable marketing option." (ETSI EN 300 175-7 Ch. A6)
pgpY8sDn6yEHs.pgp
Description: PGP signature
