On 01/04/10 16:54, Dan S wrote:
Hi -
I don't have firewire kit to test, I'm afraid, but maybe you could try
running the "latencytop" program (not installed by default, but
available in the repos) to detect the source of the xruns?
It could be a question of IRQ priorities I imagine.
Dan
hi Dan
thanks for the tip
here is the output of latencytop while running sc (135 xruns after a few
minutes with cpu at 35%...):
############################################################################
Cause Maximum Percentage
Executing raw SCSI command 15.9 msec 4.1 %
Scheduler: waiting for cpu 15.1 msec 69.4 %
SCSI cdrom ioctl 14.0 msec 0.6 %
Waiting for event (select) 4.9 msec 15.7 %
Waiting for event (poll) 4.5 msec 8.8 %
Reading from a pipe 3.0 msec 0.2 %
opening cdrom device 2.2 msec 0.8 %
Userspace lock contention 1.0 msec 0.2 %
Receiving TCP/IP data 0.5 msec 0.0 %
Process hald-addon-stor (1555) Total: 111.3 msec
############################################################################
I'm not sure to understand this I'm afraid...
_y
ps: I had only jack, scel and xterm running
pps: is there still Leek and Potato on a server somewhere so I could do
tests with it also ?
---
[email protected]
http://identi.ca/group/puredyne
irc://irc.goto10.org/puredyne
