[alsa-devel] HDA model specification and a topic/hda bug?
Hi,
I've a couple of questions. Firstly, I've been through "MORE NOTES ON HD-AUDIO DRIVER" and was able to use codecgraph and hda-analyzer to identify non-disabled speakers on my laptop (the laptop claims four speakers plus a subwoofer, Alsa's autodetect only enables the subwoofer for output) but I'm having trouble working out how to convert this knowledge into either changes to patch_realtek.c or even a patch file to feed the driver, so that it comes up correctly by default.
I do have a printout of the Intel HDA spec's codec layout chapter, but I haven't read it yet. I'm hoping not to have to, but I suspect I will.
Anyway, I've been tracking topic/hda from the Alsa git repository, hoping it'll magically fix my sound-card detection, but I'm having a problem with 5b0cb1d850c26893b1468b3a519433a1b7a176be (ALSA: hda - add more NID->Control mapping) which I believe is part of the 1.0.22 release.
For reference, it's an MSI GT725 laptop...
Basically, snd_hda_add_nids is being called with an array of three nids, the third of which is 0.
I've been throwing snd_kprintd statements at it, and it seems that I see input_mux->num_items as 3 (which I _think_ is correct, I have a buit-in microphone, microphone jack and line-in jack) but I only have two analog aud_in nodes (0x8 and 0x9, 0x7 is a vendor node).
_From the look of it, this call to snd_hda_add_nids is wrong: err = snd_hda_add_nids(codec, kctl, i, spec->capsrc_nids, spec->input_mux->num_items); it maybe should be err = snd_hda_add_nids(codec, kctl, i, spec->capsrc_nids, spec->num_adc_nids); if I'm understanding patch_alc882 correctly, although that looks wrong in pretty much every other use I can see.
It's prolly me not understanding how the HDA stuff ties together very well, but right now the inbuilt sound fails to load with a -22 (-EINVAL) error, which is snd_hda_add_nid rejecting a nid <= 0.
I suspect http://bugs.gentoo.org/show_bug.cgi?id=297454 is the same problem, and https://bugtrack.alsa-project.org/alsa-bug/view.php?id=4846 looks like it too.
On searching, https://bugtrack.alsa-project.org/alsa-bug/view.php?id=4841 is exactly the same problem, and the same model laptop as mine.
I'm happy to do debugging and whatnot, since I hope to actually work out the correct patch overrides to get my laptop speakers properly working so any debugging I do for this problem will hopefully improve my understanding of the HDA driver.
At Wed, 23 Dec 2009 16:21:34 +1100, Paul "TBBle" Hampson wrote:
Hi,
I've a couple of questions. Firstly, I've been through "MORE NOTES ON HD-AUDIO DRIVER" and was able to use codecgraph and hda-analyzer to identify non-disabled speakers on my laptop (the laptop claims four speakers plus a subwoofer, Alsa's autodetect only enables the subwoofer for output) but I'm having trouble working out how to convert this knowledge into either changes to patch_realtek.c or even a patch file to feed the driver, so that it comes up correctly by default.
I do have a printout of the Intel HDA spec's codec layout chapter, but I haven't read it yet. I'm hoping not to have to, but I suspect I will.
Anyway, I've been tracking topic/hda from the Alsa git repository, hoping it'll magically fix my sound-card detection, but I'm having a problem with 5b0cb1d850c26893b1468b3a519433a1b7a176be (ALSA: hda - add more NID->Control mapping) which I believe is part of the 1.0.22 release.
For reference, it's an MSI GT725 laptop...
Basically, snd_hda_add_nids is being called with an array of three nids, the third of which is 0.
I've been throwing snd_kprintd statements at it, and it seems that I see input_mux->num_items as 3 (which I _think_ is correct, I have a buit-in microphone, microphone jack and line-in jack) but I only have two analog aud_in nodes (0x8 and 0x9, 0x7 is a vendor node).
From the look of it, this call to snd_hda_add_nids is wrong:
err = snd_hda_add_nids(codec, kctl, i, spec->capsrc_nids, spec->input_mux->num_items); it maybe should be err = snd_hda_add_nids(codec, kctl, i, spec->capsrc_nids, spec->num_adc_nids); if I'm understanding patch_alc882 correctly, although that looks wrong in pretty much every other use I can see.
Thanks for your report.
Well, your fix isn't really correct, but the code in the place is indeed broken (also others too).
The below is a fix patch. It's already applied to sound git tree and the snapshot tarball.
thanks,
Takashi
--- From 21949f00a022e090a7e8bc9a01dfca88273c6146 Mon Sep 17 00:00:00 2001 From: Takashi Iwai tiwai@suse.de Date: Wed, 23 Dec 2009 08:31:59 +0100 Subject: [PATCH] ALSA: hda - Fix NID association for capture mixers
Fix the wrong implementation of NID <-> kctl mapping for capture mixers introduced by the ocmmit 5b0cb1d850c26893b1468b3a519433a1b7a176be. So far, the driver returns an error at probe.
Signed-off-by: Takashi Iwai tiwai@suse.de --- sound/pci/hda/hda_codec.c | 26 -------------------------- sound/pci/hda/hda_local.h | 2 -- sound/pci/hda/patch_analog.c | 3 +-- sound/pci/hda/patch_cirrus.c | 12 ++++++++---- sound/pci/hda/patch_cmedia.c | 3 +-- sound/pci/hda/patch_realtek.c | 3 +-- sound/pci/hda/patch_via.c | 3 +-- 7 files changed, 12 insertions(+), 40 deletions(-)
diff --git a/sound/pci/hda/hda_codec.c b/sound/pci/hda/hda_codec.c index c848ec0..29c90d7 100644 --- a/sound/pci/hda/hda_codec.c +++ b/sound/pci/hda/hda_codec.c @@ -3537,32 +3537,6 @@ int snd_hda_add_new_ctls(struct hda_codec *codec, struct snd_kcontrol_new *knew) } EXPORT_SYMBOL_HDA(snd_hda_add_new_ctls);
-/** - * snd_hda_add_nids - assign nids to controls from the array - * @codec: the HDA codec - * @kctl: struct snd_kcontrol - * @index: index to kctl - * @nids: the array of hda_nid_t - * @size: count of hda_nid_t items - * - * This helper function assigns NIDs in the given array to a control element. - * - * Returns 0 if successful, or a negative error code. - */ -int snd_hda_add_nids(struct hda_codec *codec, struct snd_kcontrol *kctl, - unsigned int index, hda_nid_t *nids, unsigned int size) -{ - int err; - - for ( ; size > 0; size--, nids++) { - err = snd_hda_add_nid(codec, kctl, index, *nids); - if (err < 0) - return err; - } - return 0; -} -EXPORT_SYMBOL_HDA(snd_hda_add_nids); - #ifdef CONFIG_SND_HDA_POWER_SAVE static void hda_set_power_state(struct hda_codec *codec, hda_nid_t fg, unsigned int power_state); diff --git a/sound/pci/hda/hda_local.h b/sound/pci/hda/hda_local.h index d505d05..7cee364 100644 --- a/sound/pci/hda/hda_local.h +++ b/sound/pci/hda/hda_local.h @@ -343,8 +343,6 @@ int snd_hda_check_board_codec_sid_config(struct hda_codec *codec, const struct snd_pci_quirk *tbl); int snd_hda_add_new_ctls(struct hda_codec *codec, struct snd_kcontrol_new *knew); -int snd_hda_add_nids(struct hda_codec *codec, struct snd_kcontrol *kctl, - unsigned int index, hda_nid_t *nids, unsigned int size);
/* * unsolicited event handler diff --git a/sound/pci/hda/patch_analog.c b/sound/pci/hda/patch_analog.c index 92b72d4..45ee352 100644 --- a/sound/pci/hda/patch_analog.c +++ b/sound/pci/hda/patch_analog.c @@ -244,8 +244,7 @@ static int ad198x_build_controls(struct hda_codec *codec) if (!kctl) kctl = snd_hda_find_mixer_ctl(codec, "Input Source"); for (i = 0; kctl && i < kctl->count; i++) { - err = snd_hda_add_nids(codec, kctl, i, spec->capsrc_nids, - spec->input_mux->num_items); + err = snd_hda_add_nid(codec, kctl, i, spec->capsrc_nids[i]); if (err < 0) return err; } diff --git a/sound/pci/hda/patch_cirrus.c b/sound/pci/hda/patch_cirrus.c index 093cfbb..7de782a 100644 --- a/sound/pci/hda/patch_cirrus.c +++ b/sound/pci/hda/patch_cirrus.c @@ -753,6 +753,7 @@ static int build_input(struct hda_codec *codec) spec->capture_bind[1] = make_bind_capture(codec, &snd_hda_bind_vol); for (i = 0; i < 2; i++) { struct snd_kcontrol *kctl; + int n; if (!spec->capture_bind[i]) return -ENOMEM; kctl = snd_ctl_new1(&cs_capture_ctls[i], codec); @@ -762,10 +763,13 @@ static int build_input(struct hda_codec *codec) err = snd_hda_ctl_add(codec, 0, kctl); if (err < 0) return err; - err = snd_hda_add_nids(codec, kctl, 0, spec->adc_nid, - spec->num_inputs); - if (err < 0) - return err; + for (n = 0; n < AUTO_PIN_LAST; n++) { + if (!spec->adc_nid[n]) + continue; + err = snd_hda_add_nid(codec, kctl, 0, spec->adc_nid[i]); + if (err < 0) + return err; + } } if (spec->num_inputs > 1 && !spec->mic_detect) { diff --git a/sound/pci/hda/patch_cmedia.c b/sound/pci/hda/patch_cmedia.c index cc1c223..ff60908 100644 --- a/sound/pci/hda/patch_cmedia.c +++ b/sound/pci/hda/patch_cmedia.c @@ -345,8 +345,7 @@ static int cmi9880_build_controls(struct hda_codec *codec) /* assign Capture Source enums to NID */ kctl = snd_hda_find_mixer_ctl(codec, "Capture Source"); for (i = 0; kctl && i < kctl->count; i++) { - err = snd_hda_add_nids(codec, kctl, i, spec->adc_nids, - spec->input_mux->num_items); + err = snd_hda_add_nid(codec, kctl, i, spec->adc_nids[i]); if (err < 0) return err; } diff --git a/sound/pci/hda/patch_realtek.c b/sound/pci/hda/patch_realtek.c index e7cdc6a..a451990 100644 --- a/sound/pci/hda/patch_realtek.c +++ b/sound/pci/hda/patch_realtek.c @@ -2551,8 +2551,7 @@ static int alc_build_controls(struct hda_codec *codec) hda_nid_t *nids = spec->capsrc_nids; if (!nids) nids = spec->adc_nids; - err = snd_hda_add_nids(codec, kctl, i, nids, - spec->input_mux->num_items); + err = snd_hda_add_nid(codec, kctl, i, nids[i]); if (err < 0) return err; } diff --git a/sound/pci/hda/patch_via.c b/sound/pci/hda/patch_via.c index de4839e..9ddc373 100644 --- a/sound/pci/hda/patch_via.c +++ b/sound/pci/hda/patch_via.c @@ -1907,8 +1907,7 @@ static int via_build_controls(struct hda_codec *codec) /* assign Capture Source enums to NID */ kctl = snd_hda_find_mixer_ctl(codec, "Input Source"); for (i = 0; kctl && i < kctl->count; i++) { - err = snd_hda_add_nids(codec, kctl, i, spec->mux_nids, - spec->input_mux->num_items); + err = snd_hda_add_nid(codec, kctl, i, spec->mux_nids[i]); if (err < 0) return err; }
participants (2)
-
Paul "TBBle" Hampson
-
Takashi Iwai