At Tue, 20 Sep 2011 08:15:09 +0800, Raymond Yau wrote:
2011/9/18 Takashi Iwai tiwai@suse.de:
Well, I can't reproduce it in hda-emu, so I don't know now. Try to get a stacktrace to see what's calling there actually.
set stream active : active stream 1 ALSA hda_intel.c:1757: azx_pcm_prepare: bufsize=0xd750, format=0x4111 ALSA hda_codec.c:1412: hda_codec_setup_stream: NID=0x4, stream=0x5, channel=0, format=0x4111 ALSA hda_codec.c:1412: hda_codec_setup_stream: NID=0x3, stream=0x5, channel=0, format=0x4111 ALSA hda_codec.c:1412: hda_codec_setup_stream: NID=0x6, stream=0x5, channel=0, format=0x4111 ALSA hda_codec.c:1412: hda_codec_setup_stream: NID=0x5, stream=0x5, channel=0, format=0x4111 ALSA hda_codec.c:1412: hda_codec_setup_stream: NID=0xa, stream=0x5, channel=0, format=0x4111 hda-intel: IRQ timing workaround is activated for card #1. Suggest a bigger bdl_pos_adj. ALSA hda_codec.c:1475: hda_codec_cleanup_stream: NID=0x4 ------------[ cut here ]------------
(snip)
Thanks, so it's called from hw_free. This is a call just to be sure, and should be harmless. If you mind, the verbs are suppressed by a patch like below.
Takashi
--- diff --git a/sound/pci/hda/hda_codec.c b/sound/pci/hda/hda_codec.c index 6b611d5..e3db196 100644 --- a/sound/pci/hda/hda_codec.c +++ b/sound/pci/hda/hda_codec.c @@ -1491,8 +1491,11 @@ static void really_cleanup_stream(struct hda_codec *codec, struct hda_cvt_setup *q) { hda_nid_t nid = q->nid; - snd_hda_codec_write(codec, nid, 0, AC_VERB_SET_CHANNEL_STREAMID, 0); - snd_hda_codec_write(codec, nid, 0, AC_VERB_SET_STREAM_FORMAT, 0); + if (q->stream_tag || q->channel_id) + snd_hda_codec_write(codec, nid, 0, AC_VERB_SET_CHANNEL_STREAMID, 0); + if (q->format_id) + snd_hda_codec_write(codec, nid, 0, AC_VERB_SET_STREAM_FORMAT, 0 +); memset(q, 0, sizeof(*q)); q->nid = nid; }