Patch August 2020

patch@alsa-project.org

3 participants
5 discussions

Re: [ALSA patch] [PATCH] ALSA: hda: Refactor calculating SDnFMT according to specification
by Takashi Iwai 02 Sep '20

02 Sep '20

On Mon, 24 Aug 2020 14:16:26 +0200, Kai Vehmanen wrote: > > Hey, > > On Mon, 24 Aug 2020, Pawel Harlozinski wrote: > > > Set SDnFMT depending on which format was given, as maxbps only describes container size. > > hmm, I'm not entirely sure that is correct. Usage may be a bit varied, but > most places in existing code, "maxbps" is treated as number of significant > bits, not the container size. E.g. in hdac_hda.c: > > » if (substream->stream == SNDRV_PCM_STREAM_PLAYBACK) > » » maxbps = dai->driver->playback.sig_bits; > » else > » » maxbps = dai->driver->capture.sig_bits; > > It would seem "maxbps" is a bit superfluous given the same information can > be relayed in "format" as well. But currently it's still used. E.g. if you > look at snd_hdac_query_supported_pcm(), if codec reports 24bit support, > format is always set to SNDRV_PCM_FMTBIT_S32_LE, even if only 24bit are > valid. So snd_pcm_format_width() will not return the expected significant > bits info, but you have to use "maxbps". So original code seems correct > (or at least you'd need to update both places). Hm, we need to check the call pattern, then. The maxbps passed to this function was supposed to be the value obtained from snd_hdac_query_supported_pcm(), i.e. the codec capability. But, basically this patch wouldn't change any practical behavior. In the current code, snd_pcm_format_width() can be never 20 or 24, because the 24 and 24bit supports are also with SNDRV_PCM_FMT_S32_LE. That is, the cases 20 and 24 there are superfluous from the beginning (although the checks of maxbps are still needed). Instead, what we could improve is: - Set up the proper msbits hw_constraint to reflect the maxbps value - Choose the right AC_FMT_BITS_* depending on the hw_params msbits We may change the query function not to return a single maxbps value but rather storing the raw PCM parameter value (AC_SUPPCM_*), and pass it at re-encoding the format value, too, if we want to make things perfect. thanks, Takashi

2 2

[PATCH] ALSA: hda: Refactor calculating SDnFMT according to specification
by Pawel Harlozinski 24 Aug '20

24 Aug '20

Fix setting SDnFMT based on High Definition Audio Specification Rev. 1.0a page 48. Bits per Sample (BITS): 000 = 8 bits. The data will be packed in memory in 8-bit containers on 16-bit boundaries. 001 = 16 bits. The data will be packed in memory in 16-bit containers on 16-bit boundaries. 010 = 20 bits. The data will be packed in memory in 32-bit containers on 32-bit boundaries. 011 = 24 bits. The data will be packed in memory in 32-bit containers on 32-bit boundaries. 100 = 32 bits. The data will be packed in memory in 32-bit containers on 32-bit boundaries. 101-111 = Reserved Set SDnFMT depending on which format was given, as maxbps only describes container size. Henceforth split cases for formats 20, 24, 32 bits. For format SNDRV_PCM_FORMAT_FLOAT_LE width is equal 32 so as previously it will end up with mask for 32 bits. Signed-off-by: Pawel Harlozinski <pawel.harlozinski(a)linux.intel.com> --- sound/hda/hdac_device.c | 11 +++++------ 1 file changed, 5 insertions(+), 6 deletions(-) diff --git a/sound/hda/hdac_device.c b/sound/hda/hdac_device.c index 333220f0f8af..2ccf79866f99 100644 --- a/sound/hda/hdac_device.c +++ b/sound/hda/hdac_device.c @@ -762,14 +762,13 @@ unsigned int snd_hdac_calc_stream_format(unsigned int rate, val |= AC_FMT_BITS_16; break; case 20: + val |= AC_FMT_BITS_20; + break; case 24: + val |= AC_FMT_BITS_24; + break; case 32: - if (maxbps >= 32 || format == SNDRV_PCM_FORMAT_FLOAT_LE) - val |= AC_FMT_BITS_32; - else if (maxbps >= 24) - val |= AC_FMT_BITS_24; - else - val |= AC_FMT_BITS_20; + val |= AC_FMT_BITS_32; break; default: return 0; -- 2.17.1

1 0

[PATCH] ALSA: usb-audio: ignore broken processing/extension unit
by tom.ty89＠gmail.com 17 Aug '20

17 Aug '20

From: Tom Yan <tom.ty89(a)gmail.com> Some devices have broken extension unit where getting current value doesn't work. Attempt that once when creating mixer control for it. If it fails, just ignore it, so that it won't cripple the device entirely (and/or make the error floods). Signed-off-by: Tom Yan <tom.ty89(a)gmail.com> --- sound/usb/mixer.c | 8 +++++++- 1 file changed, 7 insertions(+), 1 deletion(-) diff --git a/sound/usb/mixer.c b/sound/usb/mixer.c index eab0fd4fd7c3..e0b7174c1043 100644 --- a/sound/usb/mixer.c +++ b/sound/usb/mixer.c @@ -2367,7 +2367,7 @@ static int build_audio_procunit(struct mixer_build *state, int unitid, int num_ins; struct usb_mixer_elem_info *cval; struct snd_kcontrol *kctl; - int i, err, nameid, type, len; + int i, err, nameid, type, len, val; const struct procunit_info *info; const struct procunit_value_info *valinfo; const struct usbmix_name_map *map; @@ -2470,6 +2470,12 @@ static int build_audio_procunit(struct mixer_build *state, int unitid, break; } + err = get_cur_ctl_value(cval, cval->control << 8, &val); + if (err < 0) { + usb_mixer_elem_info_free(cval); + return -EINVAL; + } + kctl = snd_ctl_new1(&mixer_procunit_ctl, cval); if (!kctl) { usb_mixer_elem_info_free(cval); -- 2.28.0

2 1

[PATCH] ALSA: usb-audio: ignore broken processing/extension unit
by tom.ty89＠gmail.com 16 Aug '20

16 Aug '20

From: Tom Yan <tom.ty89(a)gmail.com> Some devices have broken extension unit where getting current value doesn't work. Attempt that once when creating mixer control for it. If it fails, just ignore it, so that it won't cripple the device entirely (and/or make the error floods). --- sound/usb/mixer.c | 8 +++++++- 1 file changed, 7 insertions(+), 1 deletion(-) diff --git a/sound/usb/mixer.c b/sound/usb/mixer.c index eab0fd4fd7c3..e0b7174c1043 100644 --- a/sound/usb/mixer.c +++ b/sound/usb/mixer.c @@ -2367,7 +2367,7 @@ static int build_audio_procunit(struct mixer_build *state, int unitid, int num_ins; struct usb_mixer_elem_info *cval; struct snd_kcontrol *kctl; - int i, err, nameid, type, len; + int i, err, nameid, type, len, val; const struct procunit_info *info; const struct procunit_value_info *valinfo; const struct usbmix_name_map *map; @@ -2470,6 +2470,12 @@ static int build_audio_procunit(struct mixer_build *state, int unitid, break; } + err = get_cur_ctl_value(cval, cval->control << 8, &val); + if (err < 0) { + usb_mixer_elem_info_free(cval); + return -EINVAL; + } + kctl = snd_ctl_new1(&mixer_procunit_ctl, cval); if (!kctl) { usb_mixer_elem_info_free(cval); -- 2.28.0

1 0

Re: [ALSA patch] [PATCH] pcm: dmix: fix access to sum-buffer in non-interleaved mixing mode
by Takashi Iwai 13 Aug '20

13 Aug '20

On Wed, 12 Aug 2020 08:18:25 +0200, Vijay Palaniswamy (RBEI/ECF3) wrote: > > Hi All, > > When dmix uses non-interleaved mixing mode the offset and step width to > sum_buffer was calculated by using the dmix channels instead of the slave > channels. This leads to audio distortions due to frame corruption. > > Issue happens while doing channel wise mixing of src and destination buffers. > > Attached the patch to fix the offset and step width calculation based on slave > channels for sum buffer. So mixed data will be saved in respective channels of > sum buffer. > > Steps to reproduce the issue: > > With below configuratio, Do aplay on both device in parallel for audio > distortion > > pcm.dmix_2_channels { > > type dmix > > ipc_key 5678293 > > ipc_perm 0660 > > ipc_gid audio > > bindings [0 1] > > slave { > > pcm "hardware" > > channels 2 > > periods 4 > > period_time 40000 > > } > > } > > pcm.dmix_1_channels { > > type dmix > > ipc_key 5678293 > > ipc_perm 0660 > > ipc_gid audio > > bindings [0] > > slave { > > pcm "hardware" > > channels 1 > > periods 4 > > period_time 40000 > > } > > } > > pcm.hardware { > > type hw > > card 0 > > channels 2 > > rate 48000 > > format S16_LE > > } > > Best regards > > Palaniswamy Vijay Applied now. Thanks. Takashi

1 0

2023

2022

2021

2020

2019

Patch August 2020