[alsa-devel] [PATCH]Support MCP89 and GT21x hdmi audio

Takashi Iwai tiwai at suse.de
Wed Mar 10 10:44:59 CET 2010


At Wed, 10 Mar 2010 09:17:16 +0100 (CET),
Jaroslav Kysela wrote:
> 
> On Wed, 10 Mar 2010, VDR User wrote:
> 
> > On Tue, Mar 9, 2010 at 11:58 PM, Jaroslav Kysela <perex at perex.cz> wrote:
> >> On Tue, 9 Mar 2010, VDR User wrote:
> >>
> >>> On Sun, Mar 7, 2010 at 8:46 PM, Wu Fengguang <fengguang.wu at intel.com>
> >>> wrote:
> >>>>
> >>>> CC Jaroslav, maybe he has some idea on
> >>>>
> >>>>        pcm_hw.c: snd_pcm_hw_delay() SNDRV_PCM_IOCTL_DELAY failed.
> >>>>
> >>>> On Mon, Mar 08, 2010 at 12:35:37PM +0800, Wei Ni wrote:
> >>>>>
> >>>>> Hi,
> >>>>> Because I don't have GT220 card on hand, I didn't test it yet.
> >>>>> So I didn't add this card to the patch file. I think you could add it
> >>>>> after
> >>>>> you test it.
> >>>>>
> >>>>> About the warning messages, I don't know why.
> >>>>> May be some others can answer.
> >>>>>
> >>>>> Thanks
> >>>>> Wei.
> >>>>> nvpublic
> >>>>>
> >>>>> -----Original Message-----
> >>>>> From: VDR User [mailto:user.vdr at gmail.com]
> >>>>> Sent: Monday, March 08, 2010 3:22 AM
> >>>>> To: Wei Ni
> >>>>> Cc: Takashi Iwai; akpm; alsa-devel; Wu Fengguang; linux-kernel; Pavel
> >>>>> Hofman
> >>>>> Subject: Re: [alsa-devel] [PATCH]Support MCP89 and GT21x hdmi audio
> >>>>>
> >>>>> I think this may be relevant to this subject, if not then my apologies.
> >>>>>
> >>>>> I've just installed a GT220 card and was able to get audio-over-hdmi
> >>>>> working by compiling the snapshot drivers from 20100307 which contain
> >>>>> Wei's new code.  I further had to patch the driver to recognize my
> >>>>> card with:
> >>>>>
> >>>>> --- alsa-driver-orig/alsa-kernel/pci/hda/patch_nvhdmi.c 2010-03-06
> >>>>> 18:00:12.000000000 -0800
> >>>>> +++ alsa-driver/alsa-kernel/pci/hda/patch_nvhdmi.c      2010-03-06
> >>>>> 18:02:00.000000000 -0800
> >>>>> @@ -1335,6 +1335,8 @@ static struct hda_codec_preset snd_hda_p
> >>>>>           .patch = patch_nvhdmi_8ch_89 },
> >>>>>         { .id = 0x10de000b, .name = "GT21x HDMI",
> >>>>>           .patch = patch_nvhdmi_8ch_89 },
> >>>>> +       { .id = 0x10de000a, .name = "GT220 HDMI",
> >>>>> +         .patch = patch_nvhdmi_8ch_89 },
> >>>>>         { .id = 0x10de000d, .name = "GT240 HDMI",
> >>>>>           .patch = patch_nvhdmi_8ch_89 },
> >>>>>         {} /* terminator */
> >>>>> @@ -1347,6 +1349,7 @@ MODULE_ALIAS("snd-hda-codec-id:10de0006"
> >>>>>  MODULE_ALIAS("snd-hda-codec-id:10de0007");
> >>>>>  MODULE_ALIAS("snd-hda-codec-id:10de0067");
> >>>>>  MODULE_ALIAS("snd-hda-codec-id:10de8001");
> >>>>> +MODULE_ALIAS("snd-hda-codec-id:10de000a");
> >>>>>  MODULE_ALIAS("snd-hda-codec-id:10de000c");
> >>>>>  MODULE_ALIAS("snd-hda-codec-id:10de000b");
> >>>>>  MODULE_ALIAS("snd-hda-codec-id:10de000d");
> >>>>>
> >>>>> This seems to work fine when running VDR but if I do anything else
> >>>>> like upgrade packages, compile something, or maybe create a bz2, then
> >>>>> I start to see _a lot_ of the following in my xine log:
> >>>>>
> >>>>> pcm_hw.c: snd_pcm_hw_delay() SNDRV_PCM_IOCTL_DELAY failed.
> >>>>>
> >>>>> With the old code this didn't happen.  However, my setup was an 8400gs
> >>>>> using SPDIF OUT from my mainboard to SPDIF IN on the card.  Then using
> >>>>> a dvi->hdmi cable into the tv.
> >>>>>
> >>>>> Again, I apologize if I'm posting this in the wrong place.
> >>>>>
> >>>>> Best regards.
> >>>
> >>> So nobody has any ideas at all?  This is turning out to be a big
> >>> problem and I've discovered it will occur even without anything else
> >>> happening on the system.  (originally I thought you needed to do some
> >>> other activity.)  I've checked with some other devs I know and
> >>> everyone has said the problem is with alsa so it would seem there is a
> >>> serious bug somewhere since eventually those errors will happen and
> >>> the audio will start skipping like crazy.
> >>>
> >>> Please let me know if there's anything I can do within my ability to help.
> >>
> >> Could you do strace for ALSA ioctls? At least having the error code might
> >> help to identify this issue.
> >
> > I don't know how but I'm willing to try if you want to tell me how to
> > do it.  Please keep in mind, I'm just an end-user -- not a developer
> > or even c coder.  The most I've done is run xine in gdb to help supply
> > bug info to those devs.
> 
> strace -e trace=open,ioctl <xine>

Better to add -f, too, for a program like xine.


Takashi


More information about the Alsa-devel mailing list