24 Mar
2006
24 Mar
'06
6:11 p.m.
Hi
I have a laptop that runs debian/stable with a homemade 2.6.15.6-rt21. I
use it live with csound5.Two days ago I was at a rehearsal where I
forgot my power supply. Everything ran fine except 2 things:
1) I have a csound instrument that reads the interval between taps and
writes a percentage of that to the delay time of a global delay (code
attached). That was totally screwed up, and taps would translate to
something 2-3 times faster.
2) At the end of the rehearsal the clock of my laptop was late (I think,
maybe it was early) by something like 30-45 minutes. Normally the clock
is just fine and I set if from rdate every hour, so It had dropped over
30 minutes while running about two hours on battery.
A related issue: Today I was watching an avi-file in xine, during which
the powerchord popped out. Suddenly the audio became chopped up as if
the computer couldn't keep up. Putting the cable back in didn't change
anything. But after loggin out/in of X (and keeping the power chord in
place) everything was fine again.
What on earth...???? I'm clueless. I have disabled frequency scaling in
the kernel, and the system is quite sleek running openbox and a
minimalistic install. Any ideas as to where I should look would be
appreciated.
BTW: All behaviour is reproducable. I didn't, however reproduce the
clock strangeness from 2)...
Here are my .config and the tapping csound macro:
http://www.atte.dk/download/.config
http://www.atte.dk/download/tap_tempo.macro
--
peace, love & harmony
Atte
http://www.atte.dk
24 Mar
24 Mar
9:54 p.m.
On Fri, Mar 24, 2006 at 11:11:37PM +0100, Atte André Jensen atte.jensen(a)gmail.com wrote:
1) I have a csound instrument that reads the interval
between taps and
writes a percentage of that to the delay time of a global delay (code
attached). That was totally screwed up, and taps would translate to
something 2-3 times faster.
2) At the end of the rehearsal the clock of my laptop was late (I think,
maybe it was early) by something like 30-45 minutes. Normally the clock
is just fine and I set if from rdate every hour, so It had dropped over
30 minutes while running about two hours on battery.
This sounds supiciously like a problem I reported here not long ago,
but didn't get an answer for, especially the second problem. This
wouldn't happen to be a Pentium III, would it? After reporting the
problem on the LKML, I was given a temporary workaround by John Stultz
that involved booting with "clockwource=pit", which fixed the problem
for me. The problem also exists for the -mm kernel, and he provided
a patch for that that also worked, but none for the -rt kernel (I didn't
try the patch he provided on the -rt kernel; maybe I should).
Anyway, I don't think the fix has appeared in the -rt patches yet, but
I haven't been able to get sound to work in the newest kernels, so
I'm still trying to solve that problem before I can say for sure.
Chuck
25 Mar
25 Mar
3:52 a.m.
Chuck Martin wrote:
This sounds supiciously like a problem I reported here
not long ago,
but didn't get an answer for, especially the second problem. This
wouldn't happen to be a Pentium III, would it?
No, it's a Pentium IV...
After reporting the
problem on the LKML, I was given a temporary workaround by John Stultz
that involved booting with "clockwource=pit", which fixed the problem
for me.
Could you post a link to the discussion?
--
peace, love & harmony
Atte
http://www.atte.dk
10:06 a.m.
On Sat, Mar 25, 2006 at 08:51:29AM +0100, Atte André Jensen atte.jensen(a)gmail.com wrote:
And that should have been "clocksource=pit". Sorry for the confusion.
Unfortunately, only my original question appears in the mailing list.
John Stultz replied privately asking for more information and the
output of dmesg, suggested I try the -mm kernel to see if it also had
the problem, suggested the "clocksource=pit" workaround, and then sent
me a patch to try for the -mm kernel. When I reported that the patch
worked for -mm, he said he would "polish it up and send it to Andrew."
I asked if -mm could also be used with the -rt patch, since realtime
was important to me, but I didn't get a reply.
Chuck
Chuck Martin wrote:
>This sounds supiciously like a problem I reported here not long ago,
My keyboard appears to have been invaded by gremlins. That should have
been "suspiciously"...
but didn't
get an answer for, especially the second problem. This
wouldn't happen to be a Pentium III, would it?
No, it's a Pentium IV...
>After reporting the
>problem on the LKML, I was given a temporary workaround by John Stultz
>that involved booting with "clockwource=pit", which fixed the problem
for me.
Could you post a link to the discussion? 
26 Mar
26 Mar
11:04 a.m.
On Friday 24 March 2006 22:11, Atte André Jensen wrote:
What on earth...???? I'm clueless. I have disabled
frequency scaling in
the kernel, and the system is quite sleek running openbox and a
minimalistic install. Any ideas as to where I should look would be
appreciated.
Have you also disabled it in the BIOS? I had an opposite problem when I
accidentally enabled dynamic overclocking on my AMD-64 desktop system. The
kernel steadfastly believed the CPU was running at 2000MHz in spite of the
fact that the motherboard had ramped it up to 2200.
--
David Haggett
6819
days inactive
6821
days old
linux-audio-user@lists.linuxaudio.org
4 comments
3 participants
participants (3)
-
Atte André Jensen -
Chuck Martin -
David Haggett