"I just imported ix(4), a driver for the Intel 82598EB 10 Gigabit Ethernet adapters. It is based on Intel's ixgbe FreeBSD driver, with many local changes for OpenBSD.
"The driver is fully-operational and survived some long-time tests, had to work on borrowed hardware from another company since Intel didn't provide us any cards. It is currently not sure how to maintain the driver in the future without having cards in the project. John Ronciak from Intel told me told me that they don't have more cards to give away. It is purely a business case issue; a lot of customer have asked for FreeBSD, they got two cards, but nobody asked for an OpenBSD driver. He told me that they're doing the right for their customers. But of course I could still buy them on the open market."
Community
If I know the community right, and OpenBSD needs this card, they will buy it for you.
That said, I think it's sad that a big player like Intel can't "afford" to give away another card. It's not like they're asking them to write a driver or do any work for them. They'll do the job and Intel will profit. I think that's pretty shameful.
That's the price you pay for
That's the price you pay for having a project leader that treats everyone else like shit. Noone wants to help out.
BS
Where do you GET these idiotic ideas? He has helped open up more hardware documentation than anyone, has won an FSF award for it and been nominated twice. His efforts are widely recognized. Perhaps he has a low tolerance for idiots and doesn't mind saying it to their face (preferably twice), but he's not only helping his own project; he's helping everyone in the "FOSS community". That's more than can be said about other, similar, projects.
He has helped open up more
He has helped open up more hardware documentation than anyone specifically because the parent is correct in describing Theo's attitude and manners.
In the short term, it has proven to be effective. However, that approach can make people not want to work with you out of goodwill.
Also, since they started
Also, since they started with the FreeBSD driver in the first place, instead of developing their own, Intel (rightly) could have justified their stance around that, and figured that they could just continue to use the FreeBSD driver + modifications.
Testing
And how do you suppose they test their port properly without the hardware?
The same way that kernel
The same way that kernel developers of all kinds test drivers and software for which they have limited or no access: the users. At least in the minds of a very large company who seems to have been ticked off by a very small "company" with a leader like Theo.
Theo is a tool. A very intelligent tool, but still a tool. His attitude combined with the (apparent) project size of OpenBSD makes him easier to ignore.
Linus is a tool. A very intelligent tool, but still a tool. However, his project is large enough and has enough market impact that he's harder to ignore. Though his project is large enough that few outside of the developer community need to directly interact with him.
They both lack interpersonal communication skills. Which is fine for kernel development (except for those that have to work with them). They should realize this and either change their level of professionalism and tact, or not speak or interact directly with the public.
"The face you step on today may end up being the butt you have to kiss tomorrow."
Change
And when the user reports that it doesn't work properly? The developers need hardware, period.
Good heavens, who are you to tell how Linus and Theo should act? What project do you lead that's so successful that your experience and great insight should be considered invaluable?
I've not heard of a single OpenBSD developer that does not consider Theo to be friendly and caring. More than once I've read of developers meeting him for the first time, and being surprised at how nice he was.
Sure, he's a bit "outspoken" but so what? Perhaps that's just the kind of personality you need to lead a project of that magnitude.
They're not Microsoft that needs to dress up in monkey suits and pretend to be your best friend ever. They do what they care about.
> The same way that kernel
> The same way that kernel developers of all kinds test drivers and software for which they have limited or no access: the users.
That is so wrong, you cannot write or port a driver without having the hardware. Users can help to test the results to optimize the driver, but not to write something new. Do you know how many attempts it needs to get it going? Tests, reboots, etc.? You're living in a dream world where software happens on white boards.
And even with support of the users, the 82598EB NICs are still very rare in the OpenBSD community, if found at all. And I don't think that it will change very much, why should OpenBSD users buy these NICs if they know that the vendor isn't supporting the OS developers?
And even if there is a FreeBSD driver, it is a completely different operating system with many differences to OpenBSD. The OS-specific part of the driver needed to be rewritten, like DMA access, interrupt handling, etc. because the kernels have significant differences in their API, especially since FreeBSD changed all of the traditional BSD interfaces.
wrong
This post shows exactly why there is this self-perpetuating cycle of "people being tools" on the internet.
The _real_ problem is that any nobody on the internet is free to attack, insult, degrade, and waste people's time. At virtually no expense to themselves. For some reason the anonymity lifts a lot of normal inhibitions and turns them into bastards.
Now imagine you are Linus, and you have to put up with this endless shit day in day out for a decade (or even a month). It very quickly teaches you the best ways to forcefully put away the time wasters. Some people inevitably come away thinking the person is "a tool", but often they're not the people you care one slightest bit about anyway.
I have met Linus, (not Theo). He is a fine seemingly normal person. He is obviously a good leader, because Linux didn't become big enough not to ignore because people contributed to it because it was too big not to ignore, that's circular reasoning. So obviously he has actually pretty damn good communication and management skills.
One obvious difference between the projects is the strictness of the "openness". Linus is more pragmatic and tolerates being given a leg-up by closed code. Obviously the reason why Linux attracts more commercial attention than OpenBSD are far more complex than just that issue, but it is far from being that Linux is bigger.
Render unto Caesar the things which are Caesar’s
De Raadt doesn't treat everyone else like shit, but only who deserves to be treated like shit.
I've dealt with Theo de
I've dealt with Theo de Raadt a few times, and he has been nothing but polite.
Thank you
This comes to confirm my last reply.
Check this out then
http://kerneltrap.org/mailarchive/linux-kernel/2008/6/17/2153834
Two comments and we get a
Two comments and we get a theo-is-an-asshole post.
Well done.
Hey Intel
Hey Intel, give the BSD guys a card.
Quite ridiculous from a
Quite ridiculous from a global multinational corporation.
One Ethernet card from a corporation that has annual net income near to ten billion dollars?
And it is quite ridiculous that the amount of stupidity seems to grow linearly with the popularity of this so-called FOSS.
... as shown in these comments to which I am contributing.