Re: Do we need /dev/sequencer anymore?
- Go to: [ bottom of page ] [ top of archives ] [ this month ]
Date: Sun, 08 Jun 2025 15:12:37 UTC
Bertrand Petit wrote: > On Sun, Jun 08, 2025 at 04:57:52PM +0300, Rozhuk Ivan wrote: > > > > I see no reasons to keep in kernel sequencer code, since there is user space > > implementation: > > https://github.com/rozhuk-im/virtual_midi > > You see no reason to keep the ability to communicate with external > instruments? I don't see any gain in that move. Well, it may be time to drop > FreeBSD and switch to Haiku. Can you elaborate? For one, I do not understand how having this in the kernel is beneficial. Even Haiku that you mention doesn't seem to be doing in-kernel sequencing. In what way was communication with external instruments possible with _specifically_ /dev/sequencer before, that it won't be now? Is there any existing workflow/software that you know of that relies on /dev/sequencer _and_ works? /dev/sequencer works only with MIDI devices connected through actual MIDI ports, which is not the case for most people nowadays, as they use USB ones. Ivan says Wine uses it, but I'm wondering whether it works at all. There are multiple forum posts saying it doesn't work, and people instead revert to userland sequencers like audio/alsa-seq-server. So what is the point of keeping a most likely broken sequencer, in the kernel, when most applications nowadays use userland ones anyway? What Ivan proposed about a userland OSS sequencer makes more sense IMHO. Christos