Subject: [ecasound] Re: [linux-audio-dev] huge latency peaks with 2.4.10-smp kernel
From: Jussi Laako (jussi.laako_AT_kolumbus.fi)
Date: Fri Oct 19 2001 - 20:08:51 EEST
Paul Davis wrote:
> in the meantime, a small but focused group has provided ways to avoid
> the behaviour that Kai reported. linus has indicated that he's happy
I don't like the way it's done. Mainstream goes digging latency holes all
the time and small group of people is running after and trying to fill those
holes. I'd like to see more co-operation and less one-eyed way of doing
> with the goal of low latency, but he doesn't want "hacks" to fix
> it. despite the fact that it was "hacks" that created it (code that
That's why he should be thinking about every feature "how this affects
throughput and latency". If it increases throughput by 10% but causes
latencies to raise by order of magnitude that "improvement" should be
> any device driver will always be able to create dropouts if its badly
> written. all it has to is to block interrupts for too long, and boom!
> there is absolutely *nothing* the kernel can do about it. this is true
> on every operating system, including RTLinux, QNX and other similar
> systems. you have to be able to trust device drivers, and in general,
> you can, but alas, there are some that are not worthy of it.
Yes there is, virtual interrupts with priorities. Higher priority interrupts
go over lower priority ones and interrupts are never disabled at hardware
level. Kernel should buffer interrupts. Something like RT-Linux does for the
Linux kernel. We could also do some parallel interrupt processing in SMP
machines ala RT-Linux. Soundcard interrupts could be fed to CPU #1 while
everything else is handled by CPU #0.
Kernel could also disable badly behaving driver. Once too long in driver and
*bang* driver is unloaded and device disabled.
- Jussi Laako
-- PGP key fingerprint: 161D 6FED 6A92 39E2 EB5B 39DD A4DE 63EB C216 1E4B Available at PGP keyservers -- To unsubscribe send message 'unsubscribe' in the body of the message to <ecasound-list-request_AT_wakkanet.fi>.
This archive was generated by hypermail 2b28 : Fri Oct 19 2001 - 20:04:17 EEST