Binary-only firmware covered by the GPL?
A message that is not in the archive (sent off-list or not yet received): 8RnZwD.A.91B.qHYaAB@murphy
Oh, man, it seems that I *must* repeat myself one more time, at least to see if I'm not in everyone's killfile :-)
@ 30/03/2004 11:19 : wrote Pavel Machek :
Quoted 1 line
Hi!
Hi!
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
#include <
hallo.h> * David Schwartz [Thu, Mar 25 2004, 04:41:23PM]:Quoted 4 lines
Quoted 2 lines
IMHO code that can be compiled would probably be the preferred form of the work.
You are seriously arguing that the obfuscated binary of the firmware is the preferred form of the firmware for the purpose of making modifications to it?!
I don't know if that's what /he/ is arguing, but *I* am arguing that in the cases I've seen here and in debian-legal, we have the following circumstances (the qla2xxx/ql2100_fw.c canonical example):
* the file in question (and its brothers and cousins) have the following structure IIRC:
+ GPL license comment-header
+ some includes?
+ the firmware in c-blob format or unsigned char fw[] = ....
+ nothing else.
* as the file is clearly marked by the copyright holder as being _distributed under the terms of the GPL_ and no other format is given to modify the fw[], at least *legally* is MHO that any recipient/redistributor of the file _can_ and _must_ consider the file in *that* format as the preferred form for modification (pf4m) *and*, considering it the source code, follow the directions of the GPL in respect to modification and redistribution.
* the /status quo/ obtained by observation of the previous item prevails _until somebody proves_ that the fw[] = {} is *not* the source code; this, usually, can be proven only by confession, i.e., the original copyright holder *comes out and says:* "hmmm, this is not the source code". Notice that the copyright holder maintaining silence is _not_ confession.
* in this case (copyright holder confesses it's not the source code) applied to the examples in casu, i.e., firmware, the kernel people cannot distribute the binary blob *inside the kernel tree*, but can do it separately _if the copyright holder grants a license_ to.
* even so, Debian could not distribute it.
Quoted 4 lines
Quoted 4 lines
Quoted 3 lines
Yes, the driver authors PREFERS to make the changes on the C source code, he never has to modify the firmware. Exactly what the GPL requests, where is your problem?
Quoted 8 lines
Quoted 8 lines
But the firmware didn't appear out of thin air - someone wrote it
somehow. If that's using a hex editor or inside the C code
doesn't matter, but most likely they used some other language like either C or assembly (no, not all firmware is written using assembly), and there are cases where some are in fact written using a hex editor but I can't remember any that has been for the last 30 or so years but I'm sure there has been cases where there hasn't been a working assembler.
But there are cases, even if you don't know of them. And this is the case that has to be taken in account when we start *presuming* things, at least legally, IMHO.
Quoted 4 lines
If my code contains picture of human, do I have to provide his DNA, too? Pavel
(runs away)
Signature · 1 line
best regards,M
Scripsit Humberto Massa <[email protected]>
Quoted 5 lines
to modify the fw[], at least *legally* is MHO that any recipient/redistributor of the file _can_ and _must_ consider the file in *that* format as the preferred form for modification (pf4m) *and*, considering it the source code, follow the directions of the GPL in respect to modification and redistribution.
No, law does not work that way. The phrase "preferred form for modification" has a clear enough, if somewhat fuzzy, literal meaning, and one cannot *implicitly* make it mean something that directly contrast to the literal meaning. If nobody *actually* prefers the binary blob for modification, then the binary blob is *not* the preferred form for modification. That's irrespective of whether the copyright holder behaves inconsistently.
Quoted 3 lines
* the /status quo/ obtained by observation of the previous item prevails
_untilsomebodyproves_that the fw[] = {} is *not* the source code;
And Debian's approach to software freedom doesn't work that way either. We treat software as non-free and non-distributable unless and until we see good and self-consistent evidence that it is actually free and distributable. The "burden of proof", to the extent that expression applies, is always on the side that claims that the software in question is OK for Debian to distribute.
Signature · 5 lines
Henning Makholm "Nu kommer han. Kan du ikke høre knallerten?" - To unsubscribe from this list: send the line "unsubscribe linux-scsi" in the body of a message to [email protected] More majordomo info at http://vger.kernel.org/majordomo-info.html