OpenBSD Ahead of Linux for Wi-Fi Drivers

algae writes "It looks like some kernel developers have noticed that the OpenBSD project is including reverse-engineered drivers for wireless ethernet cards while Linux is still using binary blobs. A large part of the issue is that much OpenBSD development takes place abroad, where having to do clean-room reverse-engineering isn't as important." From the article: "Christoph Hellwig took another stance, 'please don't let this reverse engineering idiocy hinder wireless driver adoption, we're already falling far behind openbsd who are very successfully reverse engineering lots of wireless chipsets.'"

Comments

Take Action

[neonprimetime]

Far too many people have a careless 'U.S.A. laws suck, merge it anyway' attitude

Sometimes, like at this point, it's the right attitude. They better take action soon, or openbsd will make them look like a joke.

You can help end this argument

[Bruce Perens]

BLOBs are bad, and their legality in the kernel is questionable. Of course really free drivers that let us extend devices are better.

Leaving BLOBs in the kernel might just mean you have different plaintiffs than if you used a reverse-engineered driver.

However, full clean-room reverse-engineering a free driver with full source code, rather than one that you have to disassemble and figure out, is a reasonably easy task. And, we have to write a Linux driver anyway. So, find one friend to work with and get started.

One person must not write any kernel code concerned with the driver. That person must read the existing driver, document the hardware, and publish the document. The document should not reproduce algorithms in the existing driver unless they are integral to driving the device and there isn't another way to do it.

A second person must not look at the existing driver. This person reads the document and writes a new driver.

Keep notes about the entire process. You could someday have to testify that you did it the right way.

Bruce

Re:You can help end this argument

[herodiade42]

No, the problem he is talking about was not with Stallman.
In fact there were two differents campaigns goals, about two very different kinds of "blobs":

• For obtaining the right to redistribute firmwares with the system. With recent wireless chipsets, the firmware is often loaded in the hardware at runtime, so it should be provided separetly (a scale economy compared to providing a flash memory onboard). So even BSD or MIT licences were acceptables here: the firmware is to be loaded and executed on the chipset, not in the main system memory nor in the kernel. It's not ideal, but we're used to accept this with eg. PC Bios, or any other drivers. I understand that it could be non orthodox for Stallman, but this is not the vital part of what Theo asked for (it was rather secondary).
• For obtaining proper chipsets documentations and specifications so that any free OS can write his proper driver. And NOT a driver (mostly binary, but even free) from the vendor, nor specs under NDA. This was the crucial part.

But by no way the vendors provided any BSD drivers replacement for any blob (really, they don't care about OpenBSD). Some (Ralink, Atmel) agreed to redistribute the firmwares under MIT like licences anyway.
And from Theo words, Stallman was found very supportive on this work (but not the Linux distros makers & users).

The problem was with the second point, obtaining the hardware specs and not a binary "blob" driver. Theo de Raadt was replied by some vendors that (citing from memory) "you just have to sign our agreement and accept our binary blob drivers, look, many distributions (namely Mandriva, Ubuntu and Suse) agreed so it's an acceptable proposal, and you've no other choice". Theo was chocked by those replies, and chocked that only the FSF helped on his lobbying: the Linux distros makers wheren't supportive, rather the opposite, and the users of those distros didn't react to such a betrayal. It's somewhat similar to the recent Sun's Java redistribution licence for Linux (except that it's about kernel code, more sensitive).

If you're interested on this story, I can find propers links, but most of it was covered on www.kerneltrap.org and www.undeadly.org, so you'll easily find the full story.

Blob is bad!

[grub]

I really hope the OpenBSD group's steadfast stance on "blob" is maintained. The Linux guys, who overall don't seem to mind blob, sound like they're starting to see the light. In the end it can only be good for all open source, not just OpenBSD.

Open Secrets

[Doc Ruby]

If those OpenBSD drivers are under the BSD license, doesn't that mean anyone (except the very few under some kind of other legal constraint for some other reason) with the chops can port them to Linux? And those chops don't have to be as tight as the original BSD coders. Shouldn't the lead be very short-lived?

Re:This seems bogus

[Just Some Guy]

Yeah, because all the BSDs and Linux have identical kernel interfaces, PCI subsystems, DMA handlers, etc. A simple ./configure; make install is all that separates OpenBSD's kernel from 2.6.15, after all.

In reality, on the other hand, the reverse engineered drivers can serve as excellent documentation for how the same logic can be implemented in another OS, but that's about as close as it's likely to get.

Re:You can help end this argument-Buy foreign

[Bruce Perens]

AC wrote: What would end the argument, Bruce is open-source hardware.

Yes, that would be excellent. How do we get there? OpenCores.org has the start. However, all of the gate-arrays that they have to work with have a proprietary bitstream format and thus they require proprietary tools to program them. We need an Open Source gate-array to facilitate Open Source hardware. Initial full-custom full-wafer mass fabrication cost is about $1 Million. At that point, you can get the parts down to a reasonable price. You can do small runs in MOSIS (or whatever they have these days) to make sure the design works before you go that far.

I figure this is at least $2 Million to get done. We need good hardware designers and people to help write the grant. If I can hook up with such people, I'll do whatever I can to help. I don't have the hardware expertise to lead this, or I'd already be started. Any volunteers? I'm quite serious.

Bruce

Re:in other news...

[mike_the_kid]

Of course, OpenBSD is a much smaller market than the Linux market, which is probably part of why binary blobs just aren't availble for it at all -- the hardware developers just ignore OpenBSD entirely rather than throwing them a bone in the form of a binary blob.

OpenBSD doesn't have any blobs because the project's leaders will not consider using them. What's the point of having an open source, audited, secure operating system if you allow arbitrary blocks of binary code into the kernel?

The reason OpenBSD doesn't have blobs is not because of their size -- they could port FreeBSD blobs in easilly. The reason is that the project is focused on quality. Their view is that quality and openness go hand in hand. Can't have one without the other. See this interview with Damien Bergamini, who implemented a driver for the Intel 3945 802.11abg NIC without any of Intel's blobs. The OpenBSD driver is considerably fewer lines of code than Intel's. Because its simpler, its easier to audit, and easier to find bugs in. Of course, you can't find any bugs in Intel's driver because you can't see the source code. Not because the Intel driver is bug free.