Linux Laptop

[Ward Vandewege]

On Thu, May 04, 2006 at 03:38:59PM +0100, Chris Ball wrote:
> >> On Thu, 4 May 2006 08:27:16, Ward Vandewege <ward at pong.be> said:
> 
>    > Now get this: the HP bios checks the pci id of the mini-pci card,
>    > and *REFUSES TO BOOT* if you don't have an HP-branded card in
>    > there. It says 'illegal mini-pci card detected' or something along
>    > those lines.
> 
> It's possible to reprogram the BIOS to contain different allowed PCI
> IDs, or to disable the check completely.  
> 
> For ThinkPads:
>    http://www.srcf.ucam.org/~mjg59/thinkpad/wireless.html
> 
> For HPs:
>    http://www.richud.com/HP-Pavilion-104-Bios-Fix/

Yeah, I think my colleague saw those. He said that it won't work for the
NX6125, because there seems to be a BIOS checksum that prevents 'tampering'
like this. He did find those pci ids in the bios rom though.

> This behaviour isn't limited to IBM and HP; now Intel ipw3945 chipset
> wireless cards require a binary userspace daemon (running as root!) to
> monitor the channel/power output of the card, again to comply with the
> FCC.

! Can you say 'security vulnerability'? Great. Closed-source crap running as
root.... 

> A userspace daemon is slightly less obnoxious than using a
> proprietary kernel module, since it will be easy to reverse-engineer the
> way that the userspace daemon talks to the (open-source) kernel module
> and use an open-source driver instead.

Quite.

>    > Nonsense about FCC requirements is just that - the third-party
>    > cards that are sold (e.g. the MSI one) _are FCC approved_. And if
>    > Dell can get away with not DRM'ing their machines, why on earth can
>    > HP and IBM not do this?
> 
> Given that now Intel are involved too, I do at least believe that each
> of the companies are doing what their lawyers believe they need to.

I'm sorry, I still don't get it. Does that mean that HP and IBM have more
conservative lawyers than Dell does? And to be honest Intel is hardly a
reference for being a 'nice guy'. Their whole EFI thing is *designed* to
enforce DRM in the bios. Ever looked into the EFI 'sandbox' model? That's
what that is for - DRM.

> I don't think this is a case of lock-in as you're describing it to be.

So.... if that is the case, why will the NX6125 take the HP-branded Intel
card, but NOT the non-HP branded Intel card? Exact same card, different pci
id, different price (+$40). The firmware is _identical_, we verified that
with md5sum.

>    > Besides; ndiswrapper is a workaround. Ralink releases GPL'd
>    > drivers. They work. Their hardware works. It's cheap.
> 
> I still have a treasured old silver Orinoco PC card lying around in case
> this situation gets any worse.  :)

I just got a new MSI (ralink) mini-pci card off newegg ($20) to replace the
piece of broadcom/ndiswrapper crap in my 2-year-old Dell. It took it happily,
no BIOS locks here. I'm voting with my $.

It's a shame, really. I'm in the market for a new laptop - or will be in the
next 6 months or so. The NX6125 looked like the perfect candidate - AMD64,
cheap, nice specs, great screen resolution. But I'm not going to buy it. Not
with DRM in the bios.

Ward.

-- 
Pong.be         -( "Free Software as in free speech, not free beer. Think  )-
Virtual hosting -(       of freedom, not price." -- Richard Stallman       )-
http://pong.be  -(                                                         )-
GnuPG public key: http://gpg.dtype.org