get syslog message "rtc: lost 1 interrupts" every 10mn

Hello,

I already posted this on the jackit-devel list. But I can try my luck here as well :

My system runs very nicely overall, I get no xruns, while running jackd at 64 frames/period with my Delta1010-LT, 48000 sample rate. I can record many tracks at once in ardour, do some networking at the same time, etc. Very happy about my system … except for one thing, which is :
tail /var/log/syslog :
Aug 5 19:20:26 studio100 kernel: [13808.066497] rtc: lost 1 interrupts
Aug 5 19:30:26 studio100 kernel: [14408.059890] rtc: lost 1 interrupts
Aug 5 19:40:26 studio100 kernel: [15008.052740] rtc: lost 1 interrupts
Aug 5 19:50:27 studio100 kernel: [15608.046750] rtc: lost 1 interrupts
Aug 5 20:00:27 studio100 kernel: [16208.040591] rtc: lost 1 interrupts
Aug 5 20:10:26 studio100 kernel: [16808.033612] rtc: lost 1 interrupts
Aug 5 20:20:26 studio100 kernel: [17408.026723] rtc: lost 1 interrupts
Aug 5 20:30:26 studio100 kernel: [18008.019861] rtc: lost 1 interrupts

which translates into jackd messages :
15:50:27.199 XRUN callback (1).
delay of 2619.000 usecs exceeds estimated spare time of 1260.000; restart …
16:10:27.199 XRUN callback (2).
delay of 2620.000 usecs exceeds estimated spare time of 1267.000; restart …
16:20:27.200 XRUN callback (3).
delay of 2620.000 usecs exceeds estimated spare time of 1259.000; restart …
16:30:27.199 XRUN callback (4).
delay of 2618.000 usecs exceeds estimated spare time of 1263.000; restart …
16:40:27.200 XRUN callback (5).
delay of 2618.000 usecs exceeds estimated spare time of 1261.000; restart …
16:50:27.199 XRUN callback (6).
delay of 2620.000 usecs exceeds estimated spare time of 1257.000; restart …
17:00:27.200 XRUN callback (7).
**** alsa_pcm: xrun of at least 0.011 msecs
17:10:27.199 XRUN callback (8).
delay of 2618.000 usecs exceeds estimated spare time of 1258.000; restart …
17:20:27.199 XRUN callback (9).
delay of 1267.000 usecs exceeds estimated spare time of 1262.000; restart …
**** alsa_pcm: xrun of at least 0.067 msecs
17:30:27.200 XRUN callback (10).
17:30:27.247 XRUN callback (1 skipped).
delay of 2616.000 usecs exceeds estimated spare time of 1261.000; restart …
17:40:27.199 XRUN callback (12).
**** alsa_pcm: xrun of at least 0.008 msecs
19:00:27.198 XRUN callback (13).
delay of 2642.000 usecs exceeds estimated spare time of 1260.000; restart …
19:20:27.199 XRUN callback (14).
**** alsa_pcm: xrun of at least 0.010 msecs
20:00:27.199 XRUN callback (15).
delay of 2638.000 usecs exceeds estimated spare time of 1272.000; restart …
20:10:27.200 XRUN callback (16).

look at the timestamps, every 10mn or multiple of 10mn, something happens (interrup lost), which causes a tiny xrun. I thought it had first to do with the fact that the sound card was sharing IRQ with some USB device. I managed to get the card on its own IRQ but that does not change much the symptom. I also thought it had to do with real time clock precision (rtc vs hpet). But honestly, these considerations are a bit obscure to me.

Did anyone observe such symptom before ? or does anyone have a clue as to what could trigger the loss of interrupt every 10mn ?

I’ve just tried 2.6.23-rc2 patched with IM’s rt patch patch-2.6.23-rc2-rt2. So far so good, I’ve been recording for about 1 hour, no lost interrupt, xrun counter still at 0 :slight_smile: let’s cross fingers that it will continue like that.