Hi, Gregoire Favre wrote: > maybe my hardware is not optimal : running a simple bash command : Which three cards do you have? card 0: card 1: card 2: > for j in `seq 0 2`;do for i in `seq 1 12`; do szap -a $j -x -n $i ; date > ;done;done > TUNING.log > > And then grep FE_HAS_LOCK TUNING.log |wc > 28 448 2212 > > Which show I miss some lock ??? If I didn't drop a line while analysing the output, it seems that card 0 locks always, card 1 misses some locks and card 2 misses more locks. To test whether this is a switch or card issue, would you mind permuting the cables on your cards? For example: plug the cable on card 0 into card 2 and vice versa and run the test again. If card 2 locks always now, then it looks like a switch issue. With the original configuration, what happens if you run the inner szap loop in parallel on the three cards? > I got lots of "cx88[1]/2-mpeg: cx8802_timeout" in syslogd. Hhm, a quick look into cx88-mpeg.c shows that a DMA transfer timed out in this case. Don't know what this means. Maybe a driver developer can tell you more. BTW: what has changed in your setup recently, as I you didn't report such issues in the past? Bye. -- Dipl.-Inform. (FH) Reinhard Nissl mailto:rnissl@xxxxxx
