usx2y 0.8.2: files different from linux-2.6.9-rc2-mm4-VP-S7
this is my latest.
if you want to try it,
cp this archive into a clean linux-2.6.9(-rc2(-mm4(-VP-S7))) tree and build alsa stuffs inside kernel.
Karsten
Hi Karsten,
Quoted 3 lines
this is my latest.
if you want to try it,
cp this archive into a clean linux-2.6.9(-rc2(-mm4(-VP-S7))) tree and
build alsa stuffs inside kernel.
Quoted 3 lines
Karsten
I have tested both ways of these 0.8.2 files. For one, a whole kernel was built, after taking the untarred archive. The other way, the earlier sent patch was applied to alsa-kernel cvs and then built from alsa-driver and installed accordingly. Into the field, both methods resulted in very similar behaviors, if not identical.
First thing to note is, as expected, the new hwdep pcm interface is not included, which makes jackd missing that nice experimental usx2y backend.
While using the alsa backend, jackd behaves as similarly as reported once before:
- jackd -R (realtime) crashes badly :( not a kernel oops however.
- jackd (non realtime) starts apparently fine, but prone to plenty and annoying XRUNs.
I noticed that the XRUN frequency decreases a lot if I disable ACPI (setting acpi=off on the boot prompt). However, these XRUNs are still annoying and quite repetitive (almost every other minute or so).
Looking to dmesg output, some pattern has been found, or kind of. For example, the following message is quite repetitive, and is coincidental to most of the XRUNs:
ALSA .../alsa-kernel/usb/usx2y/usbusx2yaudio.c:251: Sequence Error!(ep=8;wait=0,frame=65536).
Most propably some urb of usb-frame 0 is still missing. Cause could be too long delays in usb-hcd interrupt handling. This driver definitely wants 1 interrupt per ms.
To make it work for a system that can't deliver interrupts reliably at
that rate,
You'd have to make it work for NRPACKS > 1. To start, grep for NRPACKS
in this code ;-)
Other interesting, but less frequent messages are:
ALSA .../alsa-kernel/core/pcm_lib.c:158: BUG: stream = 0, pos = 0x2b11,
buffer size = 0x2b11, period size = 0xac4
ALSA .../alsa-kernel/usb/usx2y/usbusx2yaudio.c:116: should not be here
with counts=0
All these tests were mainly conducted with jackd -r 44100 -p 1024 -n 2. Using fewer frames per buffer (-p), or periods (-n), just leads to increasing the XRUN frequency, ultimately to the point of disaster. Obviously, the sound crackling due to the XRUNs is very, very noticeable.
OTHO, while taking some latency traces in attempt to match those XRUN occurrencies, nothing has popped up, even with latency thresholds of less than 200us. So I guess that astray or long kernel critical sections are not to blame here, IIUC.
Probably, attention should be given towards that "Sequence Error!" message verbosity, no?
Nuff said now.
Signature · 15 lines
rncbc aka Rui Nuno Capela [email protected] ------------------------------------------------------- This SF.net email is sponsored by: IT Product Guide on ITManagersJournal Use IT products in your business? Tell us what you think of them. Give us Your Opinions, Get Free ThinkGeek Gift Certificates! Click to find out more http://productguide.itmanagersjournal.com/guidepromo.tmpl
Am Montag 04 Oktober 2004 01:35 schrieb Rui Nuno Capela:
Quoted 7 lines
Looking to dmesg output, some pattern has been found, or kind of. For example, the following message is quite repetitive, and is coincidental to most of the XRUNs:
ALSA
.../alsa-kernel/usb/usx2y/usbusx2yaudio.c:251: Sequence Error!(ep=8;wait=0,frame=65536).
This looks like I still not got ohci's start_frame numbering scheme right. I thought it would go from 0 to 0xffff while 65536 is 0x10000 ?-) Well no. Its propably line 272 in ohci-hcd.c thats generating 0x10000 instead of 0:
urb->start_frame = ed->last_iso + ed->interval; when ed->last_iso is 0xffff. Ok, you might submit a patch request to the usb people to let that line look like: urb->start_frame = (u16)(ed->last_iso + ed->interval);
and/or you might change line 267 of usbusx2yaudio.c and line 211 of usx2yhwdeppcm.c to:
if ((0xFFFF & urb->start_frame) == usX2Y->wait_iso_frame)
The rawusb thingy is not built cause I forgot to put the right usx2y/Makefile into the tar.bz. sorry.
Here is the correct usx2yMakefile's content:
Quoted 1 line
Quoted 1 line
Quoted 1 line
snd-usb-usx2y-objs := usbusx2y.o usX2Yhwdep.o usx2yhwdeppcm.o
obj-$(CONFIG_SND_USB_USX2Y) += snd-usb-usx2y.o
<<<
Karsten
-------------------------------------------------------
This SF.net email is sponsored by: IT Product Guide on ITManagersJournal Use IT products in your business? Tell us what you think of them. Give us Your Opinions, Get Free ThinkGeek Gift Certificates! Click to find out more
http://productguide.itmanagersjournal.com/guidepromo.tmpl