Re: [Alsa-devel] Re: snd-hdsp oddities

Re: [Alsa-devel] Re: snd-hdsp oddities

[Takashi Iwai]

At Thu, 01 Jul 2004 22:16:39 -0400,
Paul Davis wrote:
> 
> >I am using the latest 2.6.7 kernel (tried also 2.6.5) but with hdsp I cannot s
> >elect anything lower than 1024x2 buffer settings in jackd without having massi
> >ve xruns.
> 
> Give up on 2.6 for now. If you can't give up on it, then at least run
> all (audio) apps with the environment variable LD_ASSUME_KERNEL set to
> 2.4.19. 
> 
> When the dust settles from the kernel and NPTL, 2.6 will be more
> viable. Right now, even though it works for some people, its not a
> generally viable platform for realtime audio.

Hmm, except for some sched policy setting problems on NPTL, I don't
see bigger problem than 2.4+preempt kernel.  For some people,
2.4+preempt might work _casually_ better than 2.6, but it can happen
vice versa.

Shouting "DON'T USE 2.6" isn't a good solution.  Though, we need to
inform to "set LD_ASSUME_KERNEL as a workaround"...


Takashi

Re: [Alsa-devel] Re: snd-hdsp oddities

[Florian Schmidt]

On Fri, 02 Jul 2004 11:40:28 +0200
Takashi Iwai <tiwai-l3A5Bk7waGM@public.gmane.org> wrote:

> > When the dust settles from the kernel and NPTL, 2.6 will be more
> > viable. Right now, even though it works for some people, its not a
> > generally viable platform for realtime audio.
> 
> Hmm, except for some sched policy setting problems on NPTL, I don't
> see bigger problem than 2.4+preempt kernel.  For some people,
> 2.4+preempt might work _casually_ better than 2.6, but it can happen
> vice versa.
> 
> Shouting "DON'T USE 2.6" isn't a good solution.  Though, we need to
> inform to "set LD_ASSUME_KERNEL as a workaround"...

Yes, but i wonder: many other people use NPTL for many kinds of applications w/o problems.. The jack source code is too complex for me to really find my way through it, but maybe someone might consider summing up how jack handles threads and provide pointers into the source at the relevant spots.. Maybe NPTL isn't at fault, but jack..

Also, i think it would be very useful to have  simplified test case which shows the "erraneous" behaviour of jack/NPTL.. 

flo


-- 
Palimm Palimm!
http://affenbande.org/~tapas/wiki

Re: [Alsa-devel] Re: snd-hdsp oddities

[Takashi Iwai]

At Fri, 2 Jul 2004 13:26:44 +0200,
Florian Schmidt wrote:
> 
> On Fri, 02 Jul 2004 11:40:28 +0200
> Takashi Iwai <tiwai-l3A5Bk7waGM@public.gmane.org> wrote:
> 
> > > When the dust settles from the kernel and NPTL, 2.6 will be more
> > > viable. Right now, even though it works for some people, its not a
> > > generally viable platform for realtime audio.
> > 
> > Hmm, except for some sched policy setting problems on NPTL, I don't
> > see bigger problem than 2.4+preempt kernel.  For some people,
> > 2.4+preempt might work _casually_ better than 2.6, but it can happen
> > vice versa.
> > 
> > Shouting "DON'T USE 2.6" isn't a good solution.  Though, we need to
> > inform to "set LD_ASSUME_KERNEL as a workaround"...
> 
> Yes, but i wonder: many other people use NPTL for many kinds of
> applications w/o problems.. The jack source code is too complex for me
> to really find my way through it, but maybe someone might consider
> summing up how jack handles threads and provide pointers into the
> source at the relevant spots.. Maybe NPTL isn't at fault, but jack.. 

Well, it's hard to determine at this point.
It could be a bug of pthread_setschedparm() on NPTL.

> Also, i think it would be very useful to have  simplified test case
> which shows the "erraneous" behaviour of jack/NPTL..  

Yes, definitely.


Takashi

Re: snd-hdsp oddities

[ico-rDkNP/OoUoU]

> Shouting "DON'T USE 2.6" isn't a good solution.  Though, we need to
> inform to "set LD_ASSUME_KERNEL as a workaround"...

Pardon my ignorance but how does one do this? As a part of the config before compiling kernel or?

Also, any ideas on the odd behavior of the hdspmixer? (see my other post)

Ico

Re: Re: [linux-audio-dev] snd-hdsp oddities

[Takashi Iwai]

At Fri, 2 Jul 2004 12:56:06 +0000,
<ico@fuse.net> wrote:
> 
> > Shouting "DON'T USE 2.6" isn't a good solution.  Though, we need to
> > inform to "set LD_ASSUME_KERNEL as a workaround"...
> 
> Pardon my ignorance but how does one do this? As a part of the
> config before compiling kernel or? 

No, just set the environment variable like
	export LD_ASSUME_KERNEL=2.4.19
(better globally) and start jack.  That's all.

In this way, glibc chooses LinuxThreads instead of NPTL.

> Also, any ideas on the odd behavior of the hdspmixer? (see my other post)

Not checked yet...  Did hdspmixer work on any versions correctly on
your system?


Takashi


-------------------------------------------------------
This SF.Net email sponsored by Black Hat Briefings & Training.
Attend Black Hat Briefings & Training, Las Vegas July 24-29 - 
digital self defense, top technical experts, no vendor pitches, 
unmatched networking opportunities. Visit www.blackhat.com

Re: [Alsa-devel] Re: snd-hdsp oddities

[Florian Schmidt]

On Fri, 02 Jul 2004 15:15:46 +0200
Takashi Iwai <tiwai-l3A5Bk7waGM@public.gmane.org> wrote:

> At Fri, 2 Jul 2004 12:56:06 +0000,
> <ico-rDkNP/OoUoU@public.gmane.org> wrote:
> > 
> > > Shouting "DON'T USE 2.6" isn't a good solution.  Though, we need to
> > > inform to "set LD_ASSUME_KERNEL as a workaround"...
> > 
> > Pardon my ignorance but how does one do this? As a part of the
> > config before compiling kernel or? 
> 
> No, just set the environment variable like
> 	export LD_ASSUME_KERNEL=2.4.19
> (better globally) and start jack.  That's all.

yes, every jack client needs to be started with this env var set [additionally to the jack server], too.. otherwise it's not sufficient..

flo

P.S.: so add

export LD_ASSUME_KERNEL=2.4.19

to your ~/.bashrc

or even to your ~/.profile


-- 
Palimm Palimm!
http://affenbande.org/~tapas/wiki