11 Oct
2004
11 Oct
'04
10:12 p.m.
Lee Revell wrote:
One of the JACK developers reported this problem on
LKML and the reply
was very interesting, this could explain a lot of the weird latency
issues that laptop users are seeing.
http://lkml.org/lkml/2004/10/11/182
The basic issue seems to be if your laptop has a broken BIOS which
implements ACPI using SMM you are out of luck. Here is some more
information on the problem (near the end of the page):
http://www.infoworld.com/cgi-bin/displayNew.pl?/glass/bg112596.htm
If ACPI causes you massive xrun/latency problems then you might have
such a broken BIOS. If you are having weird latency problems with a
laptop, even with ACPI disabled, try to see if it corresponds to the fan
turning on or changing speed. If so then you might be screwed :-(
There does not seem to be a lot of hope if you do have one of these
broken machines. But we can at least identify the problematic laptops,
complain to the manufacturers, and warn people not to buy them.
My laptop, a Compaq Presario 2516EA, probably suffers from this ACPI disease.
However, this has been mitigated by just disabling it; just booting with
acpi=off the xrun rate is almost negligible (e.g.running 2.6.9-rc3-mm3-T3
UP).
Don't know if the SMM issue is deadly hidden and commanded somewhere from
BIOS, then making life a pain, even though apci is off. How can I check
this?
For the record, I'm quite used for acpi being a real pita, but the lack of
it doesn't seem to be much important to me. OK, with acpi=off, I miss
battery status, temparature monitoring, whatever, just to name a few
goodies, but I buy that for cheap, just about when I get to do low-latency
audio.
For example, I can do along with jackd -p128 -n2 using the onboard audio
(ali5451) without much trouble, iif acpi=off, otherwise a xrun breakbeat
is featured ;)
Take care.
--
rncbc aka Rui Nuno Capela
rncbc(a)rncbc.org