Bug 125756 - [sound] [patch] cannot detect soft-modem on HDA bus
Summary: [sound] [patch] cannot detect soft-modem on HDA bus
Status: Open
Alias: None
Product: Base System
Classification: Unclassified
Component: kern (show other bugs)
Version: Unspecified
Hardware: Any Any
: Normal Affects Only Me
Assignee: freebsd-multimedia (Nobody)
Depends on:
Reported: 2008-07-18 19:00 UTC by admin
Modified: 2018-05-28 19:43 UTC (History)
0 users

See Also:

file.diff (739 bytes, patch)
2008-07-18 19:00 UTC, admin
no flags Details | Diff
manefesto.vcf (268 bytes, text/plain)
2008-07-20 08:05 UTC, manefesto
no flags Details

Note You need to log in before you can comment on or make changes to this bug.
Description admin 2008-07-18 19:00:03 UTC
cannot detect modem on HDA bus
hda - as a hub through which can be connected many devices (audio and
modem - which I know)

linux determined in all possible devices that are connected to hda

BSD made in the implementation wrong, because in many places there are
special control points to check - what this device.  And if this does not
sound - it is ignored that is - not universally

hence - Ariff only knows how to implement it hda code to add a modem
see patch for detect all HDA devices
on my machine I have sound and modem:

pcm0: HDA_DEBUG: Scanning HDA codecs [start index=0] ...
pcm0: HDA_DEBUG: Probing codec: 0
pcm0: HDA_DEBUG:    startnode=1 endnode=2
pcm0:        Vendor: 0x10ec0883
pcm0: HDA_DEBUG:    Found AFG nid=1 [startnode=1 endnode=2]
pcm0: HDA_DEBUG: Probing codec: 1
pcm0: HDA_DEBUG:    startnode=2 endnode=3
pcm0:        Vendor: 0x14f12bfa
pcm0: HDA_DEBUG:    AFG not found
pcm0: HDA_DEBUG: Probing codec: 3
pcm0: hdac_command_send_internal: TIMEOUT numcmd=1, sent=1, received=0
pcm0: hdac_command_send_internal: TIMEOUT numcmd=1, sent=1, received=0
pcm0: hdac_command_send_internal: TIMEOUT numcmd=1, sent=1, received=0
pcm0: HDA_DEBUG:    startnode=255 endnode=510
pcm0:        Vendor: 0xffffffff

Fix: rewrite HDA

Patch attached with submission follows:
How-To-Repeat: always
Comment 1 Mark Linimon freebsd_committer freebsd_triage 2008-07-18 21:06:04 UTC
Responsible Changed
From-To: freebsd-bugs->freebsd-multimedia

Over to maintainer(s).
Comment 2 manefesto 2008-07-20 08:05:45 UTC
I have a laptop with built modem.
But it is not defined at boot time.
Also, it is not visible in the list pciconf-lv.
On how much I learned modem located on the bus hda.
I would like to use the modem to work, so I ask the developer to
finalize the driver hda to work correctly with other devices except for
the sound card.
When using a patch to hda I managed to see the modem, but to use it, I
could not.
Comment 3 WATANABE Kazuhiro 2008-07-26 05:59:30 UTC

We can treat a modem codec as one of the "function group" like the
audio function group.  But almost all the structure of the modem
function group is vendor specific.  To control the modem codec, we
have to obtain the codec's specifications.

But I don't know a manufacturer which offers HDA modem codec's
datasheets.  So it is hard for us to write these codec's device

Listing a modem codec in a verbose dmesg is not so an important matter
(because we cannot use it now).

FYI, last year I tried to port the "slmodemd" daemon which supports
several number of modem codecs on Linux with the ALSA driver.


I could get some responses from the modem codec.  It generated sound
which was similar to a dialing and negotiation tone.  But I could not
use it as a modem at all.
WATANABE Kazuhiro (CQG00620@nifty.ne.jp)
Comment 4 Alexander Motin freebsd_committer 2008-12-25 01:07:29 UTC
State Changed
From-To: open->suspended

Latest snd_hda driver detects/reports all present codecs. But due 
to lack of codec specifications, modem support implementation 
looks very problematic.
Comment 5 Eitan Adler freebsd_committer freebsd_triage 2018-05-28 19:43:53 UTC
batch change:

For bugs that match the following
-  Status Is In progress 
- Untouched since 2018-01-01.
- Affects Base System OR Documentation


Reset to open status.

I did a quick pass but if you are getting this email it might be worthwhile to double check to see if this bug ought to be closed.