Open source driver with binary HAL for BCM4313 from Broadcom

Open source driver with binary HAL for BCM4313 from Broadcom

[Tim Gardner]

John,

Here is the URL to Broadcom's web site advertising their open source
driver with support for the BCM4313.

http://www.broadcom.com/support/802.11/linux_sta.php

rtg
-- 
Tim Gardner tim.gardner@ubuntu.com

Re: Open source driver with binary HAL for BCM4313 from Broadcom

[Michael Buesch]

On Wednesday 11 June 2008 19:44:52 Tim Gardner wrote:
> John,
> 
> Here is the URL to Broadcom's web site advertising their open source
> driver with support for the BCM4313.
> 
> http://www.broadcom.com/support/802.11/linux_sta.php
> 
> rtg

<quote>
Make sure that you download the appropriate tar because the hybrid binary
file must be of the appropriate architecture type. The hybrid binary file
is agnostic to the specific version of the Linux kernel because it is
designed to perform all interactions with the operating system through
operating-system-specific files...
</quote>

-- 
Greetings Michael.

Re: Open source driver with binary HAL for BCM4313 from Broadcom

[Stephen Hemminger]

On Wed, 11 Jun 2008 18:44:52 +0100
Tim Gardner <tim.gardner@canonical.com> wrote:

> John,
> 
> Here is the URL to Broadcom's web site advertising their open source
> driver with support for the BCM4313.
> 
> http://www.broadcom.com/support/802.11/linux_sta.php
> 
> rtg

It is nice to see a driver, but their LICENSE.txt is not a standard
approved open source license (http://www.opensource.org/licenses).

Re: Open source driver with binary HAL for BCM4313 from Broadcom

[Pavel Roskin]

On Wed, 2008-06-11 at 20:47 +0200, Michael Buesch wrote:
> On Wednesday 11 June 2008 19:44:52 Tim Gardner wrote:
> > John,
> > 
> > Here is the URL to Broadcom's web site advertising their open source
> > driver with support for the BCM4313.
> > 
> > http://www.broadcom.com/support/802.11/linux_sta.php
> > 
> > rtg
> 
> <quote>
> Make sure that you download the appropriate tar because the hybrid binary
> file must be of the appropriate architecture type. The hybrid binary file
> is agnostic to the specific version of the Linux kernel because it is
> designed to perform all interactions with the operating system through
> operating-system-specific files...
> </quote>

Sure, they are trying to do the MadWifi thing, but it's too late since a
good free driver exists already, even if it doesn't support that
particular chip.

Still, the glue code and the headers are under GPL.  It may be useful
for reverse engineering efforts.

Perhaps the non-free object files could be used to extract firmware from
them.  But the license won't allow distribution of that firmware, if I
understand correctly.

What Broadcom could easily do is to release the firmware under a license
that would allow free distribution.  This way, the firmware would be on
installation CDs of the popular distributions, and there would be no
need to install b43-fwcutter, get the firmware (that's a challenge in
absence of a wired network), extract the firmware and install it into
the right directory.  That would make Broadcom cards make out-of-box,
and ultimately make them more popular.

Instead, Broadcom tries to control the code its users are running.  I
don't think it's going to work well.  But to be fair to them, at least
they don't pretend that Linux doesn't exit.

-- 
Regards,
Pavel Roskin

Re: Open source driver with binary HAL for BCM4313 from Broadcom

[Stefanik Gábor]

On Wed, Jun 11, 2008 at 8:47 PM, Michael Buesch <mb@bu3sch.de> wrote:
>
> On Wednesday 11 June 2008 19:44:52 Tim Gardner wrote:
> > John,
> >
> > Here is the URL to Broadcom's web site advertising their open source
> > driver with support for the BCM4313.
> >
> > http://www.broadcom.com/support/802.11/linux_sta.php
> >
> > rtg
>
> <quote>
> Make sure that you download the appropriate tar because the hybrid binary
> file must be of the appropriate architecture type. The hybrid binary file
> is agnostic to the specific version of the Linux kernel because it is
> designed to perform all interactions with the operating system through
> operating-system-specific files...
> </quote>
>
> --
> Greetings Michael.
> --
> To unsubscribe from this list: send the line "unsubscribe linux-wireless" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html

Also, the download links are broken. And AFAIK we do support BCM4312,
although not PCI ID 0x4315 (that would be BCM4310).
(CCing the bcm43xx mailing list.)

--
Vista: [V]iruses, [I]ntruders, [S]pyware, [T]rojans and [A]dware. :-)

Re: Open source driver with binary HAL for BCM4313 from Broadcom

[Larry Finger]

Stefanik G=E1bor wrote:
> On Wed, Jun 11, 2008 at 8:47 PM, Michael Buesch <mb@bu3sch.de> wrote:
>> On Wednesday 11 June 2008 19:44:52 Tim Gardner wrote:
>>> John,
>>>
>>> Here is the URL to Broadcom's web site advertising their open sourc=
e
>>> driver with support for the BCM4313.
>>>
>>> http://www.broadcom.com/support/802.11/linux_sta.php
>>>
>>> rtg
>> <quote>
>> Make sure that you download the appropriate tar because the hybrid b=
inary
>> file must be of the appropriate architecture type. The hybrid binary=
 file
>> is agnostic to the specific version of the Linux kernel because it i=
s
>> designed to perform all interactions with the operating system throu=
gh
>> operating-system-specific files...
>> </quote>
>>
>> --
>> Greetings Michael.
>> --
>> To unsubscribe from this list: send the line "unsubscribe linux-wire=
less" in
>> the body of a message to majordomo@vger.kernel.org
>> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>=20
> Also, the download links are broken. And AFAIK we do support BCM4312,
> although not PCI ID 0x4315 (that would be BCM4310).
> (CCing the bcm43xx mailing list.)

Watch out for their tarballs as they will dump all over the directory=20
that you are in when you invoke tar.

Larry
--
To unsubscribe from this list: send the line "unsubscribe linux-wireles=
s" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: Open source driver with binary HAL for BCM4313 from Broadcom

[Stefanik Gábor]

On Wed, Jun 11, 2008 at 9:47 PM, Pavel Roskin <proski@gnu.org> wrote:
> On Wed, 2008-06-11 at 20:47 +0200, Michael Buesch wrote:
>> On Wednesday 11 June 2008 19:44:52 Tim Gardner wrote:
>> > John,
>> >
>> > Here is the URL to Broadcom's web site advertising their open source
>> > driver with support for the BCM4313.
>> >
>> > http://www.broadcom.com/support/802.11/linux_sta.php
>> >
>> > rtg
>>
>> <quote>
>> Make sure that you download the appropriate tar because the hybrid binary
>> file must be of the appropriate architecture type. The hybrid binary file
>> is agnostic to the specific version of the Linux kernel because it is
>> designed to perform all interactions with the operating system through
>> operating-system-specific files...
>> </quote>
>
> Sure, they are trying to do the MadWifi thing, but it's too late since a
> good free driver exists already, even if it doesn't support that
> particular chip.

I really don't understand why - they should do it the iwlwifi way.

>
> Still, the glue code and the headers are under GPL.  It may be useful
> for reverse engineering efforts.
>
> Perhaps the non-free object files could be used to extract firmware from
> them.  But the license won't allow distribution of that firmware, if I
> understand correctly.
>
> What Broadcom could easily do is to release the firmware under a license
> that would allow free distribution.  This way, the firmware would be on
> installation CDs of the popular distributions, and there would be no
> need to install b43-fwcutter, get the firmware (that's a challenge in
> absence of a wired network), extract the firmware and install it into
> the right directory.  That would make Broadcom cards make out-of-box,
> and ultimately make them more popular.
>
> Instead, Broadcom tries to control the code its users are running.  I
> don't think it's going to work well.  But to be fair to them, at least
> they don't pretend that Linux doesn't exit.

And that's quite an achievement...
Anyway, what are the actual, working download links? The ones on the
site just 404 out for me.

>
> --
> Regards,
> Pavel Roskin
> --
> To unsubscribe from this list: send the line "unsubscribe linux-wireless" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>



-- 
Vista: [V]iruses, [I]ntruders, [S]pyware, [T]rojans and [A]dware. :-)

Re: Open source driver with binary HAL for BCM4313 from Broadcom

[Larry Finger]

Stefanik G=E1bor wrote:
> On Wed, Jun 11, 2008 at 9:47 PM, Pavel Roskin <proski@gnu.org> wrote:
>> Instead, Broadcom tries to control the code its users are running.  =
I
>> don't think it's going to work well.  But to be fair to them, at lea=
st
>> they don't pretend that Linux doesn't exit.
>=20
> And that's quite an achievement...
> Anyway, what are the actual, working download links? The ones on the
> site just 404 out for me.

http://www.broadcom.com/docs/linux_sta/hybrid-portsrc-x86_64_5_10_18.ta=
r.gz

for the 64-bit version, and

http://www.broadcom.com/docs/linux_sta/hybrid-portsrc_5_10_18.tar.gz

for the 32-bit version. The binary blob is for x86 architecture only.=20
Other platforms are out of luck. They also have a README at

http://www.broadcom.com/docs/linux_sta/README.txt

As noted earlier, the BCM4310 has a PCI ID of 0x4315. I have one of=20
those cards and could test their blob pile. ATM, I'm using b43 on a=20
BCM4312 (a/b/g) card, but it has an ID of 0x4312.

I've only had a quick look at their code, but I don't think it will be=20
much help for the RE project. Most of the interesting stuff seems to=20
be in the binary blob.

Larry
--
To unsubscribe from this list: send the line "unsubscribe linux-wireles=
s" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: Open source driver with binary HAL for BCM4313 from Broadcom

[Pavel Roskin]

On Wed, 2008-06-11 at 22:26 +0200, Stefanik G=E1bor wrote:
> >> > http://www.broadcom.com/support/802.11/linux_sta.php

> And that's quite an achievement...
> Anyway, what are the actual, working download links? The ones on the
> site just 404 out for me.

Click on text ("32-bit driver", "64-bit driver"), not on the icons.
Links with underscores in the version are correct:

http://www.broadcom.com/docs/linux_sta/hybrid-portsrc_5_10_18.tar.gz
http://www.broadcom.com/docs/linux_sta/hybrid-portsrc-x86_64_5_10_18.ta=
r.gz

--=20
Regards,
Pavel Roskin
--
To unsubscribe from this list: send the line "unsubscribe linux-wireles=
s" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: Open source driver with binary HAL for BCM4313 from Broadcom

[Hauke Mehrtens]

[-- Attachment #1: Type: text/plain, Size: 961 bytes --]

Some time ago I found some source code for the wireless device in my
Broadcom brcm63xx router in the source code package for it.
It contains ~3 MB of source code for the wireless device (Broadcom
BCM43XX 802.11abg), but there are no make files and the build process
takes an pre-compiled version. It looks like someone forgot to delete
some files.

There is a text like this in all files:
This is UNPUBLISHED PROPRIETARY SOURCE CODE of Broadcom Corporation;

The package can be downloaded at:
http://broadband.eip.siemens.ch/public/GPL_source/GPL_source_CL_SL_SLI_series_consumer_release.tar

The code is in the directory broadcom/net/wl/impl1

If you know Broadcom had published it, forget about this e-mail.

--
Hauke

Tim Gardner schrieb:
> John,
> 
> Here is the URL to Broadcom's web site advertising their open source
> driver with support for the BCM4313.
> 
> http://www.broadcom.com/support/802.11/linux_sta.php
> 
> rtg


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 890 bytes --]

Re: Open source driver with binary HAL for BCM4313 from Broadcom

[Johannes Berg]

[-- Attachment #1: Type: text/plain, Size: 377 bytes --]


> > Here is the URL to Broadcom's web site advertising their open source
> > driver with support for the BCM4313.

> It is nice to see a driver, but their LICENSE.txt is not a standard
> approved open source license (http://www.opensource.org/licenses).

It's not a "binary HAL" either, it's a huge binary driver blob with a
few lines of linux glue code.

johannes

[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 836 bytes --]

Broadcom's Hybrid Driver

[Larry Finger]

I have two ideas about what Broadcom is doing.

The b/g device with an ID of 0x4315 and an LP-PHY is the one that HP 
has been shipping in many laptops for the past few months. Perhaps HP 
wants to offer Linux on those computers, but doesn't want to use 
ndiswrapper for wireless access. HP might have enough clout to get 
Broadcom to budge a little on the subject of Linux.

It is also possible that they want to stifle our RE efforts. The 
license accompanying the binary blob states:

"2.6.    No Other Rights Granted; Restrictions.  Apart from the 
license rights expressly set forth in this Agreement, Broadcom does 
not grant and Licensee does not receive any ownership right, title or 
interest nor any security interest or other interest in any 
intellectual property rights relating to the Software, nor in any copy 
of any part of the foregoing.

Licensee shall not (i) use, license, sell or otherwise distribute the 
Software except as provided in this Agreement, (ii) attempt to reverse 
engineer, decompile or disassemble any portion of the Software; or 
(iii) use the Software or other material in violation of any 
applicable law or regulation, including but not limited to any 
regulatory agency, such as FCC, rules."

Given some of the other parts of the license relating to penalties, 
etc., anyone trying to RE this code could be in a heap of trouble.

Larry

Re: Broadcom's Hybrid Driver

[Stefanik Gábor]

On Wed, Jun 11, 2008 at 11:31 PM, Larry Finger
<Larry.Finger@lwfinger.net> wrote:
> I have two ideas about what Broadcom is doing.
>
> The b/g device with an ID of 0x4315 and an LP-PHY is the one that HP has
> been shipping in many laptops for the past few months. Perhaps HP wants to
> offer Linux on those computers, but doesn't want to use ndiswrapper for
> wireless access. HP might have enough clout to get Broadcom to budge a
> little on the subject of Linux.
>
> It is also possible that they want to stifle our RE efforts. The license
> accompanying the binary blob states:
>
> "2.6.    No Other Rights Granted; Restrictions.  Apart from the license
> rights expressly set forth in this Agreement, Broadcom does not grant and
> Licensee does not receive any ownership right, title or interest nor any
> security interest or other interest in any intellectual property rights
> relating to the Software, nor in any copy of any part of the foregoing.
>
> Licensee shall not (i) use, license, sell or otherwise distribute the
> Software except as provided in this Agreement, (ii) attempt to reverse
> engineer, decompile or disassemble any portion of the Software; or (iii) use
> the Software or other material in violation of any applicable law or
> regulation, including but not limited to any regulatory agency, such as FCC,
> rules."
>
> Given some of the other parts of the license relating to penalties, etc.,
> anyone trying to RE this code could be in a heap of trouble.
>
> Larry
>

(Resend of previous message with proper CC list. I clicked "Reply"
instead of "Reply to all".)
Hmm... IANAL and I don't live in the US, but AFAIK clean-room
reverse-engineering for the purposes of making documentation and other
non-controversial purposes is allowed, and they can't force you to
forfeit this right of yours. Otherwise a patent holder would have no
way to know if the software in question infringes their patents. I can
also see no copy-protection mechanism attached to the driver, which
prevents DMCA attacks (which could override the right to clean-room
reverse-engineer the software). To me, this text looks like a response
to FCC's robustness requirements. (It's possible that it's only legal
to reverse-engineer this blob if you are outside the US.)

--
Vista: [V]iruses, [I]ntruders, [S]pyware, [T]rojans and [A]dware. :-)

Re: Broadcom's Hybrid Driver

[Pavel Roskin]

On Wed, 2008-06-11 at 16:31 -0500, Larry Finger wrote:
> I have two ideas about what Broadcom is doing.
> 
> The b/g device with an ID of 0x4315 and an LP-PHY is the one that HP 
> has been shipping in many laptops for the past few months. Perhaps HP 
> wants to offer Linux on those computers, but doesn't want to use 
> ndiswrapper for wireless access. HP might have enough clout to get 
> Broadcom to budge a little on the subject of Linux.

That's sounds plausible.

> It is also possible that they want to stifle our RE efforts.

By releasing the driver?  I don't think so.  They don't think we are
such idiots to take their code and get in trouble, do they?

Besides, Broadcom contributes to the kernel.  There are files with their
copyright, even in the SSB code.  They just don't want to help with the
wireless driver.

> Given some of the other parts of the license relating to penalties, 
> etc., anyone trying to RE this code could be in a heap of trouble.

Reverse engineering efforts should always be careful about the legal
aspects.  Yet I see two possible loopholes.

1) The recipient is not necessarily bound by the license without having
signed it (in some jurisdictions).

2) Reverse engineering to the purpose of compatibility may not be
restricted (in some other jurisdictions).

Even if neither applies, it's not like we are losing anything.  We
should just be careful, as always.

-- 
Regards,
Pavel Roskin

Re: Broadcom's Hybrid Driver

[Johannes Berg]

[-- Attachment #1: Type: text/plain, Size: 376 bytes --]


> > Given some of the other parts of the license relating to penalties, 
> > etc., anyone trying to RE this code could be in a heap of trouble.
> 
> Reverse engineering efforts should always be careful about the legal
> aspects.  Yet I see two possible loopholes.

[...]

This this is completely useless for RE anyway, we have much better
information.

johannes

[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 836 bytes --]