[linux-audio-user] about 2.6.x woes [nptl?]

[Florian Schmidt]

Hi,

i have tried the 2.6.7 kernel now for about two weeks and i always had massive realtime probs when running jackd. xruns would occur although the jack threads were running SCHED_FIFO. all kind of weird "subgraph timed out" messages etc. even running as root didn't better the situation one bit..

i had very good rt performance in 2.4.x with ll+preempt+caps patches. i could run 2*512 for hours w/o xruns [even compile a kernel while doing ardour stuff]. now in 2.6.x i couldn't even get 4*512 or 8*512 to work reliable at all [512 is the max period size of my soundcard]. i experienced the abovementioned symptoms..

well, i stumbled across this thread on the jackit-devel mailing list about trying to set LD_ASSUME_KERNEL=2.4.22 and run all jack stuff [server and clients with this env var set] and it worked wonders.. 2*512 works very nice again..

i suspect this is nptl related (see the jackit-devel thread for details: http://sourceforge.net/mailarchive/message.php?msg_id=8629219)..
i think the LD_ASSUME_KERNEL switch somehow makes ld.so link to another libc or something w/o nptl? anyone knows details?


so i just wanted to urge all the users of 2.6.x and who experience strange weird non rt operation with jack to try it. maybe it helps...

i wrote myself a bash script called "realrealtime" and put that in my path:

LD_ASSUME_KERNEL=2.4.22 $@

and use that to start al my jack stuff :) Any ideas about why jackd behaves so weird in the first place appreciated..

Florian Schmidt

P.S.: i posted chrt output for jackd w/ and w/o realrealtime at:

http://www.affenbande.org/~tapas/realrealtimejack


-- 
Palimm Palimm!