At Thu, 05 Apr 2007 10:47:52 +0200, Dominique Dumont wrote:
Takashi Iwai tiwai@suse.de writes:
The patch is a bit dangerous because it unconditionally resets the DIGI_CONVERT verb for each widget.
Agreed. We could also test the NID before switching on and off DigEn in the DIGI_CONVERT.
OTOH the switch would occur for *all* HDA chips which may not be what you want.
I'm not 100% sure that this particular procedure is necessary for all codecs. So, I prefer having it in the hda-codec part rather in hda-intel part.
The below is my modified version. It doesn't turn off before the controller side reset as you did. So, if my version doesn't work, it implies that the SPDIF reset has to be done before the controller reset. Please give it a try.
It works also ! I've tested AC3 and 44KHz. I cannot test 96KHz or 196Khz (my amp is too old).
Good to hear.
- /* turn off SPDIF once; otherwise the IEC958 bits won't be updated */
I had more the impression that the SPDIF stream is messed up if you change the stream format while the digital converter is working...
- if (codec->spdif_ctls & 1)
Shouldn't you use AC_DIG1_ENABLE for a proper patch ?
if (codec->spdif_ctls & AC_DIG1_ENABLE)
snd_hda_codec_write(codec, nid, 0, AC_VERB_SET_DIGI_CONVERT_1,codec->spdif_ctls & 0xfe);likewise: codec->spdif_ctls & 0xff & ~AC_DIG1_ENABLE ?
- snd_hda_codec_setup_stream(codec, nid, stream_tag, 0, format);
- /* turn on again (if needed) */
- if (codec->spdif_ctls & 1)
if (codec->spdif_ctls & AC_DIG1_ENABLE) ?
Yeah, that'd be better. I'll fix and commit to the HG tree later.
thanks,
Takashi