Marvell 88w8335 [Libertas]

Markus Becker mab at
Thu Mar 20 04:16:34 EDT 2008

On Wed, 19 Mar 2008, Holger Schurig wrote:

>> The module loads (does not bail out in the interrupt handler)
>> and creates the devices:
> ...
>> It is not yet fully working, however. The card does not seem
>> to react on the scanning/set_channel commands:
> This mailing list is more for the 838x chips, but we don't have a
> problem with occassional 833x e-mail traffic here.
> However, my advice is the following: As long as the driver's home
> is on some "obscure" tar files, it will always suffer from
> bitrot and not-so-much contribution by third partis. So aim to
> bring the driver into the linux kernel. Talk with John W.
> Linville on the linux-wireless mailing list about this.
> should give you pointers. For drivers,
> sometimes Linux accepts even not-so-good-code, when there is a
> chance that the beast get's better via re-factoring over time.

That would also be my preferred way, as this code has taken the route 
through a few people already (Luc Saillard, Andrea Merello and David 

> When if_cs.c went into the in-kernel-libertas driver, a few
> people that I never heard of found errors or sent bugfixes.
> Second advice: I think the BSD people (FreeBSD possible) have a
> m8k driver in their kernel tree. As I have heard, this beast
> actually works. So either look there for inspiration or try
> *BSD, pepper the code with debug printks and try to understand
> the differences between their code and the current linux code.

Yes, that is the malo driver, as far as I know. The mrv8k(ng) from is using the hardware constants and 
programming methods from that malo driver as stated in the mrv8.c.

> --
> To unsubscribe from this list: send the line "unsubscribe linux-wireless" in
> the body of a message to majordomo at
> More majordomo info at

More information about the libertas-dev mailing list