That's at least a very simple patch.
-
-
Replying to @vathpela
Actually, I don't think channel swaps are a thing in the ALSA USB driver. Seems that's handled in userspace... which means only some userspace will handle it properly. Probably needs a separate PulseAudio config etc...
1 reply 0 retweets 0 likes -
Replying to @marcan42
Okay, but it's an easy pa module *or* an easy kernel patch to add that entirely.
1 reply 0 retweets 0 likes -
Forgive me, I've been guess-and-test debugging a terrible USB sound chip.
1 reply 0 retweets 0 likes -
Replying to @vathpela
I think the kernel guys don't want audio format conversions in kernel space, so it's gotta be userspace.
1 reply 0 retweets 0 likes -
Replying to @marcan42
Fair enough, I guess, though it's pretty arbitrarily given the existing quirk tables.
1 reply 0 retweets 0 likes -
Replying to @vathpela
I think the point is the kernel shouldn't be doing format conversions, which have a high CPU cost. All the existing quirks are about metadata/control, not formats.
1 reply 0 retweets 0 likes -
Replying to @marcan42
Hmm. Maybe. Still not seeing how it's less expensive to do the operation in userland. Outcome for the user is certainly better if we do it in kernel, just from less effort involved. If you want to backdoor this with hype, just add an ebpf hook with quirkable defaults.
1 reply 0 retweets 0 likes -
Replying to @vathpela
Doing such work in kernelspace often makes it non-interruptible, may have latency implications for other stuf, etc. Also userspace is often doing conversions *anyway* which means you can amortize channel swaps, often for free.
1 reply 0 retweets 0 likes -
Replying to @marcan42
No? Just no. We've had deferrable work queues since like 2.5? And you can amortize that on either side, but putting it in userland means you have to get two pieces of software right.
2 replies 0 retweets 0 likes
Sure, but it still introduces a layer between userspace and the device, whereas otherwise you can often mmap the hardware buffer directly from userspace. Might be less of an issue with USB which is higher overhead anyway, but still.
-
-
Replying to @marcan42
Yeah, I'm not seeing where mmap is necessarily an optimization when you're mapping a bounce buffer.
0 replies 0 retweets 0 likesThanks. Twitter will use this to make your timeline better. UndoUndo
-
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.