Skip to contentSkip to conversation

Possible GPL infringement in Broadcom-based routers

DISCUSSION Jérôme Petazzoni 6 participants 6 messages Nov 4, 2004 – Nov 5, 2004

lore
Share & tools
lore
#0 lore raw

The following routers (and they might be other models, too) :
- Us Robotics 9105 (without wireless) and 9106 (with wireless)
- Siemens SE515
- Dynalink RTA230
- Buffalo WMR-G54
- Inventel DBW-200

... are all based on the same Broadcom chipset (96345 board). They integrate a 4-ports Ethernet switch, a 802.11g wireless access point, and a DSL modem (and doing routing and/or bridging between those interfaces). The CPU runs the MIPS32 instruction set. It runs a 2.4.17 linux-mips kernel with additional patches, and is loaded with a lot of free software (busybox, uclibc, zebra...)

The vendors (probably Broadcom, in fact) had to patch the kernel to support the DSL modem, the wireless interface (which is a PCMCIA-hosted BCM4306 ; which already was subject of heated debates earlier), and also some generic stuff like the flash memory and the front leds. Miscellaneous vendors provide so-called "GPL sources", which are generally mutilated kernels, lacking all the "interesting" parts (wireless and DSL drivers for instance).

Can Broadcom and the vendors "escape" the obligations of the GPL by shipping those proprietary drivers as modules, or are they violating the GPL plain and simple by removing the related source code (and showing irrelevant code to show "proof of good will") ?

Broadcom has been contacted about this matter but hasn't answered so far, nor did US Robotics (I tried to contact USR since I own a USR router).

Any suggestions about the legal (or if it's a lost cause, technical!) ways to get support for this platform will be very welcome.

More information can be found here :
http://skaya.enix.org/wiki/GplInfringement (some extra details)
http://skaya.enix.org/wiki/BroadCom96345 (technical info that I gathered about the router, firmware and kernel formats, etc.)
http://sourceforge.net/projects/brcm6345-linux/ (sourceforge project)

Best regards,
Jérôme Petazzoni <jp at enix dot org>
PS: please be kind and cc me, as my lkml awareness is limited to KT ...

#1 lore raw
Quoted 5 lines

Can Broadcom and the vendors "escape" the obligations of the GPL by shipping those proprietary drivers as modules, or are they violating the GPL plain and simple by removing the related source code (and showing irrelevant code to show "proof of good will") ?

That is a contentious issue that has been debated on this group far too
much. In the United States, at least, the answer comes down to the complex legal question of whether the module is a "derived work" of the Linux kernel and whether the kernel as shipped with those modules is a "mere aggregation".

You can google those terms to form your own opinions. IMO, discussing and
debating these type of borderline GPL violations on the LKML is not a good idea. Save bringing up GPL violations on the LKML for cases where the violation is really obvious and all measures short of public shaming have already been tried.

DS

#2 lore raw
Quoted 6 lines · David Schwartz
Quoted 5 lines

Can Broadcom and the vendors "escape" the obligations of the GPL by shipping those proprietary drivers as modules, or are they violating the GPL plain and simple by removing the related source code (and showing irrelevant code to show "proof of good will") ?

Quoted 6 lines

That is a contentious issue that has been debated on this group far
too
much. In the United States, at least, the answer comes down to the complex legal question of whether the module is a "derived work" of the Linux kernel and whether the kernel as shipped with those modules is a "mere aggregation".

Well, from what I can see of the makefiles and sources they distribute, they _don't_ distribute it as kernel+modules, they compile their drivers directly into their kernel:

./arch/mips/brcm-boards/bcm96345/Makefile:EXTRA_CFLAGS +=
-I$(TOPDIR)/../../targets
./drivers/char/bcm96345/board/Makefile:EXTRA_CFLAGS += -I.
-I$(HPATH)/asm/bcm96345 -I$(TOPDIR)/../../targets -fno-exceptions
./Makefile:SUBDIRS +=modulesrc/drivers ../../targets
./Makefile:DRIVERS-y += modulesrc/drivers/kermods.o ../../targets/bp.o

They may call the directory "modulesrc", but it does _NOT_ appear to be linked as a kernel module, but directly into the kernel. I think that in this case their build process is too tightly integrated with the kernel to _not_ be a derivative work.

Cheers,
Kyle Moffett

-----BEGIN GEEK CODE BLOCK-----
Version: 3.12
GCM/CS/IT/U d- s++: a17 C++++>$ UB/L/X/*++++(+)>$ P+++(++++)>$ L++++(+++) E W++(+) N+++(++) o? K? w--- O? M++ V? PS+() PE+(-) Y+ PGP+++ t+(+++) 5 X R? tv-(--) b++++(++) DI+ D+ G e->++++$ h!*()>++$ r !y?(-) ------END GEEK CODE BLOCK------

#3 lore raw
In reply to David Schwartz
Quoted 10 lines · David Schwartz
Quoted 4 lines

Can Broadcom and the vendors "escape" the obligations of the GPL by shipping those proprietary drivers as modules, or are they violating the GPL plain and simple by removing the related source code (and showing irrelevant code to show "proof of good will") ?

That is a contentious issue that has been debated on this group far too
much. In the United States, at least, the answer comes down to the complex legal question of whether the module is a "derived work" of the Linux kernel and whether the kernel as shipped with those modules is a "mere aggregation".

Not quite.

It comes down to a question of whether their firmware as a whole is a 'collective work' based in part on the Linux kernel, or whether the whole thing can be considered 'mere aggregation' despite the fundamental interdependency between the kernel and their drivers, without either of which the entire device would be useless.

"These requirements apply to the modified work as a whole. If identifiable sections of that work are not derived from the Program, and can be reasonably considered independent and separate works in themselves, then this License, and its terms, do not apply to those sections when you distribute them as separate works."

This is the bit you seem to be thinking of -- when it's not actually a derived work. But the GPL goes on and explicitly talks about your obligations with respect even to non-derived works, if distributed as part of a larger whole:

"But when you distribute the same sections as
part of a whole which is a work based on the Program, the
distribution of the whole must be on the terms of this License,
whose permissions for other licensees extend to the entire
whole, and thus to each and every part regardless of who wrote
it.
"Thus, it is not the intent of this section to claim rights or
contest your rights to work written entirely by you; rather, the
intent is to exercise the right to control the distribution of
derivative or collective works based on the Program."

Signature · 1 line
dwmw2
#4 lore raw
In reply to Jérôme Petazzoni

I work for lawyers. Let me know if you would like a referral.

Regards,
Scott Lockwood

-----Original Message-----
From: [email protected] [mailto:[email protected]] On Behalf Of Jérôme Petazzoni
Sent: Thursday, November 04, 2004 5:34 PM
To: [email protected]
Subject: Possible GPL infringement in Broadcom-based routers

The following routers (and they might be other models, too) :
- Us Robotics 9105 (without wireless) and 9106 (with wireless)
- Siemens SE515
- Dynalink RTA230
- Buffalo WMR-G54
- Inventel DBW-200

... are all based on the same Broadcom chipset (96345 board). They integrate a 4-ports Ethernet switch, a 802.11g wireless access point, and a DSL modem (and doing routing and/or bridging between those interfaces). The CPU runs the MIPS32 instruction set. It runs a 2.4.17 linux-mips kernel with additional patches, and is loaded with a lot of free software (busybox, uclibc, zebra...)

The vendors (probably Broadcom, in fact) had to patch the kernel to support the DSL modem, the wireless interface (which is a PCMCIA-hosted BCM4306 ; which already was subject of heated debates earlier), and also

some generic stuff like the flash memory and the front leds. Miscellaneous vendors provide so-called "GPL sources", which are generally mutilated kernels, lacking all the "interesting" parts (wireless and DSL drivers for instance).

Can Broadcom and the vendors "escape" the obligations of the GPL by shipping those proprietary drivers as modules, or are they violating the

GPL plain and simple by removing the related source code (and showing irrelevant code to show "proof of good will") ?

Broadcom has been contacted about this matter but hasn't answered so far, nor did US Robotics (I tried to contact USR since I own a USR router).

Any suggestions about the legal (or if it's a lost cause, technical!) ways to get support for this platform will be very welcome.

More information can be found here :
http://skaya.enix.org/wiki/GplInfringement (some extra details)
http://skaya.enix.org/wiki/BroadCom96345 (technical info that I gathered

about the router, firmware and kernel formats, etc.)
http://sourceforge.net/projects/brcm6345-linux/ (sourceforge project)

Best regards,
Jérôme Petazzoni <jp at enix dot org>
PS: please be kind and cc me, as my lkml awareness is limited to KT ... -
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [email protected] More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

BEGIN:VCARD
VERSION:2.1
N:Lockwood;Scott
FN:Scott Lockwood
TITLE:User Support Specialist
TEL;WORK;VOICE:+1 (312) 602-5140
TEL;HOME;VOICE:+1 (815) 723-2101
TEL;CELL;VOICE:+1 (312) 617-8626
TEL;VOICE:815-557-7443
TEL;WORK;FAX:+1 (312) 602-5050
EMAIL;PREF;INTERNET:[email protected]
REV:20041011T173726Z
END:VCARD