USB Audio in Haiku - Update please


I saw a reference to this recently in a progress report, and wondered if USB audio is now feasible, and if not, when is it likely to be?

Apart from the fact that this is how I listen to music (in Linux) it does seem that the BFS is an excellent way to maintain a quickly searchable database of digital music without incurring the overhead and limitations of music playing applications.

BTW, I am hugely impressed with all the advances in Haiku recently. I now have 64 bit Haiku installed on one of my laptops, and whilst it is not yet perfect, web browsing works well most of the time, and Libre0ffice is a huge step forward. A bit more stability, a few more applications and Haiku will be able to serve the needs of most non-business computer users.

I have started making small monthly payments to the fund to show my appreciation.


I have a work-in-progress patch on Gerrit to add support to the USB 3 driver for “isochronous transfers”, which is the one missing piece the usb_audio driver needs (and while testing that, I wound up fixing most of the bugs in the usb_audio driver … I think.)

I got it to the point where it does output a tone in QEMU, but it has some strange timing issues with starting said transfers and it’s not clear the tone is actually correct. I need to do some testing on bare metal, and then attempt to resolve the remaining TODOs and timing issues. (FreeBSD and Linux’s XHCI drivers are both unbelievably convoluted when it comes to isochronous transfer timing, and I don’t understand what they or the specification is actually saying or doing re. frame calculations.)

So it is possible; I just have a bunch more stuff to try and fix here, and it’s all a bit over my head even…


Many thanks for your reply, and it’s good to know that progress is being made.

I’m sure the technicalities won’t be over your head for long!

At a slight tangent, have the powers-that-be at Haiku endeavoured to identify which applications are likely to be most wanted by potential Haiku users, and therefore where efforts should be concentrated? If not, perhaps they should.

If we are talking about Haiku being somebody’s main (or only) operating system, then I would suggest that most people would want the following (no doubt I have missed something):

Web browser (WP is coming along nicely, but not quite there yet)
Office-type program (LibreOffice. Hurrah!)
Email (Exists, still need to try it)
Picture viewer/editor (Not sure which is the best option for this)
Music server/player (This is where USB audio comes in, along with MediaPlayer)
Video player (VLC, presumably)
Plus various utilities

It’s far more important to have one good program that works than several that don’t work very well, which is the bane of Linux. Looking at Haiku Depot, I see a similar problem with a lot of rather basic programs and not many killers. LibreOffice of course is definitely a killer, and it’s the reason I have now felt able to treat Haiku as a viable OS.

Anyway, I am sure I am not the only person interested in USB music, and future reports on progress will be very warmly received. Thanks again for your efforts.


For picture viewers, the built in ShowImage is fine. For editing, try WonderBrush. The built-in MediaPlayer is also more than fine for video playback.


The “powers that be”? Who is that anyway?

The project is largely community driven, there is no central entity making decisions here. There are some developers who try to work on the OS itself. There are others working on applications. Largely each of them works first on what they need themselves.

Also, before taking care of potential users, we are trying to fit the needs of existing users. One step at a time :slight_smile:

And some of the developers are fortuantely themselves Haiku users, and they know what they need all too well. That’s the reason I’m putting a lot of effort into WebPositive, or the reason waddlesplash spent countless hours working on wifi drivers so he could get online, etc.

Also, our BeOS legacy says we should talk about “tractor apps”, not “killer apps”. We’re not here to murder anyone!


A tractor app sounds a bit like something you might find in a collective farm in the Ukraine, circa 1956, but I am happy to adopt it if killer has too many negative connotations. That said, I would be happy if you did murder a few people, economically speaking. You could start with Sativa Nutella, or whatever he is called.

The powers-that-be at Haiku are surely those who do most of the work, like you and Waddlesplash, among others.

Anyway, I hope I don’t sound too much like somebody trying to teach their grandmother how to suck eggs. Your focus on WebPositive is great because the browser is now the most important tractor app (!) and it has come on a long way in the last year or so. I do experience crashes from time to time, but I suspect a lack of RAM might be partly to blame.

The only way in which we disagree slightly is that everybody was a potential user once, and converting potential users into actual ones is surely key to the success of the project.


I think of it as related to Star Trek “tractor beams”, if that sounds better?

That matters, but not as much to me as keeping our already existing userbase (including myself). It’s just a matter of priorities: first myself (because if I can’t do my things in Haiku, why would I bother developping it), then existing users, and only then potential users. And so far the first two parts have largely managed to leep me busy.