#1033849 libffado2: Firewire device disconnects after a while.

Package:
libffado2
Source:
libffado2
Description:
FFADO API
Submitter:
Alexandre Lymberopoulos
Date:
2023-04-22 03:15:02 UTC
Severity:
normal
#1033849#5
Date:
2023-04-02 19:10:47 UTC
From:
To:
Dear Maintainer,

The particular device I have is a M-Audio Firewire 410 sound card
connect via a PCI Express card (Via VT6315 chipset). When I log in the
card works fine (except that I have to run ffado-mixer to have the
output to be heard from the headphone jacks - any help on this would
be appreciated too). But, after a few minutes of inactivity of the
card (meaning no sound played on it here), it plays nothing and the
leds showind output signals to its channels do not turn
on. Interesting fact is that it remais connected:

~$ ffado-test ListDevices
-----------------------------------------------
FFADO test and diagnostic utility
Part of the FFADO project -- www.ffado.org
Version: 2.4.7
(C) 2008-2021, Daniel Wagner, Pieter Palmers and others
This program comes with ABSOLUTELY NO WARRANTY.
-----------------------------------------------

=== 1394 PORT 0 ===
  Node id  GUID                  VendorId     ModelId   Vendor - Model
   0       0x000d6c01007b89ea  0x00000D6C  0x00010046   M-AUDIO - FW 410
no message buffer overruns

I don't understand how firewire connections work but it seems somehow
the card looses connection to the system. I looked for system logs and
found this on /var/log/kern.log:

2023-04-02T16:00:16.418326-03:00 desktop kernel: [13699.785450] firewire_ohci 0000:01:00.0: DMA context IT0 has stopped, error code: evt_timeout

I may provide any further information upon request, but need
instructions on how to acquire that information.

Thanks in advance and best regards,
Alexandre

#1033849#10
Date:
2023-04-22 03:10:49 UTC
From:
To:
Dear all,

I heard that VIA chipsets does not work well with those cards (mine is a
VIA 6315 on a PCIExpress card). Should I consider getting another card
with Texas Instruments chipset? Is it required to power in that card
with its MOLEX input or the power coming from the slot is enough? Should
I use the external AC adapter (this last one is what I'm doing with no
improvements)?

Any help would be largely appreciated and I'm available to help as I
can. Another incident led the following on dmesg:

[  754.550709] snd_bebob fw1.0: Detect discontinuity of cycle: 4670 4679

and after a while

[ 1561.332132] snd_bebob fw1.0: Detect discontinuity of cycle: 59487 59496
[ 1789.828800] snd_bebob fw1.0: Detect discontinuity of cycle: 31497 31505
[ 1831.919030] snd_bebob fw1.0: Detect discontinuity of cycle: 48161 48169
[ 1840.081206] snd_bebob fw1.0: Detect discontinuity of cycle: 49593 49602

Heard some clicks (probably in the timestamp above) and the car stopped
playing. Left the computer and when I got back, I went to qwgraph to
check connections and the removed an extra connection from firefox to a
channel of the 410 and the card worked normally, but not for too much:

[13364.296669] firewire_ohci 0000:01:00.0: DMA context IT0 has stopped, error code: evt_timeout
[13364.296766] firewire_ohci 0000:01:00.0: DMA context IR0 has stopped, error code: evt_descriptor_read

This is the current state as reported by ffado-test:

~# ffado-test ListDevices
-----------------------------------------------
FFADO test and diagnostic utility
Part of the FFADO project -- www.ffado.org
Version: 2.4.7
(C) 2008-2021, Daniel Wagner, Pieter Palmers and others
This program comes with ABSOLUTELY NO WARRANTY.
-----------------------------------------------

=== 1394 PORT 0 ===
  Node id  GUID                  VendorId     ModelId   Vendor - Model
   0       0x000d6c01007b89ea  0x00000D6C  0x00010046   M-AUDIO - FW 410
   1       0x0011066600000009  0x00001106  0x00000000   Linux Firewire -
no message buffer overruns

Hope to hear from you soon.

Best, Alexandre