Subject: [linux-audio-dev] Re: And kernel 2.4 ? Was: Re: RFD: Audio Needs for Kernel 2.5+
From: Benno Senoner (sbenno_AT_gardena.net)
Date: Wed Jun 21 2000 - 19:40:20 EEST
Paul , I agree with you, on almost all points.
It's not my intention to pressure anyone.
Ithe petition was only a potential idea to see how many people would be
interested/need in these features.
Anyway as for the broken implementation:
as Ingo said in some cases you can decompose the code, but other code sections
have long execution paths by design and you need this preemption "hack".
see old linux-kernel messages.
Plus this feature is already used in some sections of the kernel where the
developers know that the kernel might stay in kernel space for very long time.
Since I am unable to write that kind of code myself, the only option remaining
is waiting.
It would be nice if at some point this feature would be added to 2.4 kernels.
But the usual kernel policy seems to prohibit that.
Mentioning the chicken-egg problem, no vendor would be motivated to port
their flagship-audio software to Linux, because it sucks (plain linux) big
when doing realtime work. (100ms latency).
So perhaps there will be second possibility to solve the problem:
on the LAD site, put a kernel-rpm online which is already patched with the
latest lowlatency patches.
So that joe-sixpack can install it with a single rpm -i ourcustomkernel.rpm
Of course it would be much cooler if the distro would allow this at install
time.
( like asking if the user wants to install a lowlatency kernel, in the case
he plans to run realtime software)
Mandrake could be a potential supporter of this strategy since they are
always including cutting-edge before all other distros.
For example they allow to install the secure-linux patch (defeats stack-overflow
exploits).
comments ?
Benno.
On Tue, 20 Jun 2000, Paul Barton-Davis wrote:
> >I am not a kernel programmer but looking at the patch,
> >I think that Ingo's modifications do not make a radical change within
> >the kernel.
> >There are still issues with SMP in the 2.2.x patch, but sometime
> >ago Ingo said for 2.3.x it would be much easier to solve.
>
> >My proposal would be that the linux-audio community (we) should write
> >a letter to the linux-kernel folks ( CCing Linus,Alan,Ingo,SCT,ecc)
> >and explain the audio issue in detail asking for official statements.
>
> Linus has made his position fairly clear. He isn't against low latency
> patches, but he has a very clear idea of what he considers the "right"
> kind of solution to be. There were some places where Ingo's patches
> were considered OK, but many more where Linus felt that they were
> fundamentally the wrong approach to take. Ingo's stuff consisted of
> inserting conditional reschedules in the middle of otherwise blocking
> operations. Linus wants those operations decomposed so that they don't
> block in the same way. Its a fair point.
>
> >Another proposal would be to make a sort of "low-latency/audio" petition
> >and submit it to the leading linux news sites.
> >That could generate some buzz and bring us some supporters.
>
> Gack. The implicit (and sometimes explicit) motto on l-k is "write the
> code". Putting pressure on people to write code for you is a bad way
> to go about getting this done.
>
> Linus has made it clear what has to be done to get his approval for LL
> patches in future versions - they have to be done "right". Either
> start writing the code, or start paying someone to write the code, or
> convince someone else to pay someone to write the code, or wait patiently.
>
> :)
>
> --p
This archive was generated by hypermail 2b28 : Wed Jun 21 2000 - 20:27:18 EEST