kernel/linux.git/sound/core/pcm_lib.c, branch v6.8

ALSA: hda: Upgrade stream-format infrastructure

2023-11-27T16:27:41+00:00

Introduce a set of functions that ultimately facilite SDxFMT-related calculations in atomic manner: First, introduce snd_pcm_subformat_width() and snd_pcm_hw_params_bits() helpers that separate the base functionality from the HDAudio-specific one. snd_hdac_format_normalize() - format converter. S20_LE, S24_LE and their unsigned and BE friends are invalid from HDAudio perspective but still can be specified as function argument due to compatibility reasons. snd_hdac_stream_format_bits() - obtain just the bits-per-sample value. Does not ignore subformat and msbits parameters. snd_hdac_stream_format() and snd_hdac_spdif_stream_format() - obtain the SDxFMT value given the audio format parameters. The former is stripped away of spdif-related information. Useful for users that do not care about them. Acked-by: Mark Brown Signed-off-by: Cezary Rojewski Acked-by: Jaroslav Kysela Link: https://lore.kernel.org/r/20231117120610.1755254-5-cezary.rojewski@intel.com Signed-off-by: Takashi Iwai

ALSA: pcm: Fix error checks of default read/write copy ops

2023-09-02T06:11:48+00:00

copy_from/to_iter() returns the actually copied bytes, and the more correct check should be to compare with the given bytes, instead of zero-check. Fixes: cf393babb37a ("ALSA: pcm: Add copy ops with iov_iter") Reported-by: Al Viro Closes: https://lore.kernel.org/r/20230902053044.GJ3390869@ZenIV Link: https://lore.kernel.org/r/20230902061044.19366-1-tiwai@suse.de Signed-off-by: Takashi Iwai

ALSA: pcm: Drop obsoleted PCM copy_user and copy_kernel ops

2023-08-18T10:19:19+00:00

Finally all users have been converted to the new PCM copy ops, let's drop the obsoleted copy_kernel and copy_user ops completely. Link: https://lore.kernel.org/r/20230815190136.8987-26-tiwai@suse.de Signed-off-by: Takashi Iwai

ALSA: pcm: Add copy ops with iov_iter

2023-08-18T10:18:15+00:00

iov_iter is a universal interface to copy the data chunk from/to user-space and kernel in a unified manner. This API can fit for ALSA PCM copy ops, too; we had to split to copy_user and copy_kernel in the past, and those can be unified to a single ops with iov_iter. This patch adds a new PCM copy ops that passes iov_iter for copying both kernel and user-space in the same way. This patch touches only the ALSA PCM core part, and the actual users will be replaced in the following patches. The expansion of iov_iter is done in the PCM core right before calling each copy callback. It's a bit suboptimal, but I took this now as it's the most straightforward replacement. The more conversion to iov_iter in the caller side is a TODO for future. As of now, the old copy_user and copy_kernel ops are still kept. Once after all users are converted, we'll drop the old copy_user and copy_kernel ops, too. Link: https://lore.kernel.org/r/20230815190136.8987-3-tiwai@suse.de Signed-off-by: Takashi Iwai

ALSA: pcm: use exit controlled loop in snd_pcm_playback_silence()

2023-05-05T16:23:53+00:00

We already know that `frames` is greater than zero, because we just checked it. So we don't need to check the loop condition on the first iteration. Signed-off-by: Oswald Buddenhagen Link: https://lore.kernel.org/r/20230505155244.2312199-7-oswald.buddenhagen@gmx.de Signed-off-by: Takashi Iwai

ALSA: pcm: simplify top-up mode init in snd_pcm_playback_silence()

2023-05-05T16:23:52+00:00

Inline the remaining call of snd_pcm_playback_hw_avail(). This makes the top-up branch more congruent with the thresholded one, and allows simplifying the handling of the corner cases. Signed-off-by: Oswald Buddenhagen Link: https://lore.kernel.org/r/20230505155244.2312199-6-oswald.buddenhagen@gmx.de Signed-off-by: Takashi Iwai

ALSA: pcm: playback silence - move silence variable updates to separate function

2023-05-05T16:23:52+00:00

The code tracking the added samples in thresholded mode and the code tracking the just played samples in top-up mode are semantically identical, so factor it out to a common function to enhance readability. Co-developed-by: Oswald Buddenhagen Signed-off-by: Oswald Buddenhagen Signed-off-by: Jaroslav Kysela Link: https://lore.kernel.org/r/20230505155244.2312199-5-oswald.buddenhagen@gmx.de Signed-off-by: Takashi Iwai

ALSA: pcm: playback silence - remove extra code

2023-05-05T16:23:51+00:00

The removed condition handles de facto only one situation where runtime->silence_filled variable is equal to runtime->buffer_size, because this variable cannot go over the buffer size. This case is implicitly caught by the required comparison of the noise distance with the threshold. Suggested-by: Oswald Buddenhagen Signed-off-by: Jaroslav Kysela Signed-off-by: Oswald Buddenhagen Link: https://lore.kernel.org/r/20230505155244.2312199-4-oswald.buddenhagen@gmx.de Signed-off-by: Takashi Iwai

ALSA: pcm: fix playback silence - correct incremental silencing

2023-05-05T16:23:50+00:00

Commit 9a826ddba6e ("[ALSA] pcm core: fix silence_start calculations") came with exactly the right commit message, but the patch just made things broken in a different way: We'd fill at a too low address if the area was already partially zeroed, so we'd under-fill. This affected both thresholded mode (where it was somewhat less likely) and top-up mode (where it would be the case consistently). Co-developed-by: Oswald Buddenhagen Signed-off-by: Oswald Buddenhagen Signed-off-by: Jaroslav Kysela Link: https://lore.kernel.org/r/20230505155244.2312199-3-oswald.buddenhagen@gmx.de Signed-off-by: Takashi Iwai

ALSA: pcm: fix playback silence - use the actual new_hw_ptr for the threshold mode

2023-05-05T16:23:49+00:00

The snd_pcm_playback_hw_avail() function uses runtime->status->hw_ptr. Unfortunately, in case when we call this function from snd_pcm_update_hw_ptr0(), this variable contains the previous hardware pointer. Use the new_hw_ptr argument to calculate hw_avail (filled samples by the user space) to correct the threshold comparison. The new_hw_ptr argument may also be set to ULONG_MAX which means the initialization phase. In this case, use runtime->status->hw_ptr. Suggested-by: Oswald Buddenhagen Signed-off-by: Jaroslav Kysela Reviewed-by: Oswald Buddenhagen Link: https://lore.kernel.org/r/20230505155244.2312199-2-oswald.buddenhagen@gmx.de Signed-off-by: Takashi Iwai