On Feb 17, 2008 9:30 AM, William Boshuck <[EMAIL PROTECTED]> wrote: > On Sun, Feb 17, 2008 at 05:33:12PM +0530, Mayuresh Kathe wrote: > > Hi, > > > > NOTE: No intention to behave like a troll. > > > > I've been following the "multi-threaded ssh/scp" thread and read Ted's > > comment that he's stopped working on the kernel threads code because > > he doesn't have the time for it nor does he need it any more. > > Also that multi-threaded ssh/scp would weaken security features within > the OS. > > I think it is to meet the developers' needs, and to provide them > with a highly enjoyable (irrespective of remuneration) vocation. > The rest of us are along for the ride, and we chip in to keep the > ride going, because we like the ride.
I think that's key to the understanding. There are a limited number of developers, all with their own interests. They are working on a number of things that interest them, for example, henning was working on improvements in tcpip/pf that resulted in almost doubling of performance. The OpenBSD developers have made calls in the past for high performance, 10G equipment. So, to conclude that OpenBSD *needs* a high performance tcpip stack without keeping this in mind is... naive. Then, there's the fact that there's a limited number of developers, all with their own interests (yes, I'm repeating, but it is important). This is not linux, where there're tons of paid developers. Most folks who work on OpenBSD are volunteers or paid way less than the market rate. Obviously they will work on the pieces that they need and interest them the most. If you *need* threaded kernel, *you* can work on it, or you can *fund* someone to work on it. Else, wait for it to become interesting for a developer. Mayuresh started out by saying he is not a C developer. So? If something interests you, pick it up. One very important thing to keep in mind is that the OpenBSD environment is the code then release. Not pre-announce, then look for code. There's no "FireEngine" project, because "fireengine" type code is expected. And if they're working on improvements, they don't declare it to the world "I'm going to try to double the tcpip stack's performance!", then spend time trying to do that. No, they look at things that can be improved, improve it, then measure it. The first we'll hear about it is when the developers say "hey can you guys help test this out?". For example - softraid. ccd and raidframe has been limiting for years, and raidframe has been bitrotting too. Softraid showed up in 4.2, mirroring only. 4.3 appears to have softraid with raid 0. I'm sure work continues on softraid. But they're building it ontop of bioctl and all the other things. But there was no grand pronouncement of "I'm building a FireHouse" on "FireStreet" to handle all the "FireTraffic". This does have one unfortunate side effect - us, mere users, do not know what some of the new things that are coming, are. However, it also does not set up for false or unrealistic expectations. Maybe softraid with raid5 will not show up in 4.4. Maybe raid 0 won't even show up in 4.3. Point is - the developers work on what they want, and release it when they can. You may want a FireEngine. Maybe they want a Lego BrainStorm, and are working on it? -- http://www.glumbert.com/media/shift http://www.youtube.com/watch?v=tGvHNNOLnCk "This officer's men seem to follow him merely out of idle curiosity." -- Sandhurst officer cadet evaluation. "Securing an environment of Windows platforms from abuse - external or internal - is akin to trying to install sprinklers in a fireworks factory where smoking on the job is permitted." -- Gene Spafford learn french: http://www.youtube.com/watch?v=j1G-3laJJP0&feature=related
