summaryrefslogtreecommitdiff
path: root/audio
Commit message (Collapse)AuthorAgeFilesLines
* HAL: Align error code with HIDL wrapperKevin Rocard2017-10-061-4/+4
| | | | | | | | | | | | | | | | | | | | | Here is how the HIDL wrapper interprets status error codes: 0 => Result::OK -EINVAL => Result::INVALID_ARGUMENTS -ENODATA => Result::INVALID_STATE -ENODEV => Result::NOT_INITIALIZED -ENOSYS => Result::NOT_SUPPORTED otherwise => Result::INVALID_STATE This is consistent with most of the hal. This patch changes the few places where it was not to unify the behaviours. Test: vts-tradefed run vts --module VtsHalAudioV2_0Target Test: call/play music/record/video... Bug: 36311550 Change-Id: Id821d15ae3269cd2c9699924e2b174c7347a6d07 Signed-off-by: Kevin Rocard <krocard@google.com> Signed-off-by: André Pinela <sheffzor@gmail.com>
* Set up the mixer at adev_open()daniel_hk2016-07-031-0/+1
| | | | | | | | | | | | M doesn't select the audio device at start now. Hence the mixer is not properly set up. That's why there is no audio output after reboot. However, when the orientation sensor detected changes, the audio HAL driver will be adjusted. During the adjustment, the mixer is reset and bring the audio back. To fix this strange scenario, I set up the mixer with select_device() at the end of device open. Change-Id: I6878993da55d0ee76b81229e20c2c16f53bf7ac3
* [k3.4] audio: refine audio hal for kernel 3.4André Pinela2016-02-193-127/+157
| | | | Change-Id: Ieaa43cbc5ef62686c58d38b12671fff1171efede
* [k3.4] audio: update pcm and mixer card IDsZiyan2016-02-191-2/+2
| | | | Since the unused Intel HDA card was removed from the kernel, the card IDs changed.
* audio HAL: add parameters to open stream functionsEric Laurent2014-07-281-2/+5
| | | | | | | | Pass device address (and audio source for inputs) to open_output_stream() and open_input_stream() audio HAL functions. Bug: 14815883. Change-Id: Iec91c57f7f908e0dd278eafd95da29034c15ff54
* Enable lower latency audio captureGlenn Kasten2014-07-271-5/+20
| | | | | Bug: 16601366 Change-Id: I128db28937e2fefd4e41a924204442a72d45f1c8
* Add audio_input_flags_t to HAL open_input_streamGlenn Kasten2014-07-161-1/+2
| | | | Change-Id: Ia595260e238b3308fcf86ca4b71feea5f71c485a
* audio: deprecate audio_stream_frame_size()Eric Laurent2014-07-031-9/+9
| | | | | Bug: 15000850. Change-Id: I6eacaedc5e45457ba528c04c952ec352fa7b8c56
* audio: fixed channel count determination from channel maskEric Laurent2014-07-011-2/+2
| | | | | | | Do not use popcount() to derive channel count from channel mask. Bug: 15000850. Change-Id: I478b4c4e0076c2fdf4b9e6c007f49a9da72931ba
* Implement HAL API get_presentation_positionGlenn Kasten2013-09-291-1/+32
| | | | | Bug: 9587132 Change-Id: Ic227628c707d08e283bcee91af37a85181d94a7c
* audio: factor out audio_route into a separate librarySimon Wilson2013-01-284-537/+10
| | | | | | | | audio_route is mature enough now to belong in its own library and will help remove duplication across audio HALs. Change-Id: If3a2a0b4a6d9bbed93d206e61d22a155066a23d4
* audio: add dock supportSimon Wilson2012-10-271-2/+6
| | | | Change-Id: I03819de8e1210f14d323939a584905ab1bbe63c6
* Revert "Audio: Support audio dock."Simon Wilson2012-10-261-44/+1
| | | | This reverts commit 03729278e30ff974263c93e6df8bfb6a33fbef20.
* Revert "Audio: fix incorrect routing path."Simon Wilson2012-10-261-5/+3
| | | | This reverts commit bc5583c5f01457d36a06cfc72095cf85b2a6dea9.
* Revert "Audio: Dock: fix speaker silent issue after inserting and removing ↵Simon Wilson2012-10-263-36/+0
| | | | | | headset rapidly." This reverts commit fc3bd5eb75c3723fab25a66420379f783e0d31ba.
* Audio: Dock: fix speaker silent issue after inserting and removing headset ↵able_liao2012-10-233-0/+36
| | | | | | | | | | | | | | | | rapidly. 1. Design "AUX" and "speaker" dapm mixer control for dock feature in HAL instead of kernel. The previous kernel design could cause speaker silent after inserting and removing headphone rapidly, or even just normally insert and remove headphone during system suspend. 2. For dock feature, original speaker output scenario must be separated into dock path and speaker path. Thus we have to update the current state of speaker and dock respectively to make sure speaker and dock output paths can be enabled/disabled as expected. Bug: 7388692 Change-Id: I6e75cf63cdd8134e34ab22fb818a17ec8c4bdce5 Signed-off-by: able_liao <able_liao@asus.com>
* Audio: fix incorrect routing path.Ramanan Rajeswaran2012-09-101-3/+5
| | | | | | | Check output device is speaker first and then check if dock is plugged in. Change-Id: I7723f6de90bf0548320423c465e8950593df9e46
* Nakasi 512 frames per HAL output bufferGlenn Kasten2012-09-091-1/+1
| | | | | Bug: 6542224 Change-Id: I026622956d1e299603f9f26ddcaa2d542729a18b
* audio: changes for new audio device enumsEric Laurent2012-09-061-35/+19
| | | | | | | | | | | | Modifications for new audio device enums: - Separated input and output device fields as output and input device values are now on 32 bits. - changed audio device API version to 2.0 Also removed get_supported_devices() function not needed if audio_policy.conf file is present. Change-Id: Ieb32abfa3c3aac1fb9c16dc73350d5cc231e41a4
* audio: fix crash when starting captureEric Laurent2012-08-301-10/+12
| | | | | | | | | | | | | When starting capture, we check if a playback stream is active with an incompatible sampling rate and if yes force this stream in standby. The output stream lock is acquired while requesting standby. The problem is that this stream is referenced by adev->active_out which is cleared by do_output_standby() and a crash occurs when releasing the lock. The same problem is present when starting a playback stream and a capture stream is active. Change-Id: I72ce058308371199655e3b2e1cfc0529328555ec
* audio: fix capture problemEric Laurent2012-08-221-0/+1
| | | | | | | Reset input buffer frame count when starting capture to avoid reading stale data from previous capture. Change-Id: Ied07eda9b2094b8a36f07871d6e071ac96f5892d
* Audio: Support audio dock.sam_chen2012-08-101-1/+44
| | | | Change-Id: I313e31be8205b7cbbbf5102c76469764b271f791
* Revert "audio: add watchdog on buffer full condition"Dave Burke2012-06-281-33/+8
| | | This reverts commit 8e94790ebb933b039e8e20b635f53e4186af9499
* audio: add watchdog on buffer full conditionEric Laurent2012-06-281-8/+33
| | | | | | | | | | | | | | | | There have been several reports of devices losing audio completely until rebooted. The logs showed that the kernel buffer was steadily full as if playback did not start or stopped and did not restart. This change adds a watchdog to restart media server in case such a condition persists. Also add more log in case of errors reported by tinyalsa. Bug 6699340. Change-Id: Ie3e1565e1c1e6fd8581e6969343d481a63d5c880
* audio: add 24000 Hz and 32000 Hz capture ratesEric Laurent2012-06-131-3/+12
| | | | | | | | | Also modified adev_get_input_buffer_size() to return a size consistent with in_get_buffer_size(). Bug 6615379. Change-Id: Id2f7a2de7e393d057b0987866507f2f1f699f944
* audio: fix latency reported when screen is onEric Laurent2012-06-131-7/+41
| | | | | | | | | | | | | | | The change for variable buffer size made that the latency reported is always the largest value corresponding to screen off buffer size. This causes an A/V sync problem as the audio delay compensation is based on the latency reported by audio HAL. Also added a maximum sleep time when waiting for buffer filling status to go below write threshold. Also added progressive write threshold increase when recovering from low buffer filling status. Change-Id: Ia29c00e7c32b66faefd2817dca2b4b54cc0d8c7b
* audio: implement variable buffer size.Eric Laurent2012-06-061-22/+111
| | | | | | | | | | | Use larger buffers on audio output when screen is OFF to minimize risks of underruns when the CPU clock is at the lowest. Maximum buffer size is currently limited to 32kB by the driver. Bug 6605215. Change-Id: I1725a9eb284a87eabe63caedec330eba47a76c53
* audio: swap L/R channels of speaker when only that is activeSimon Wilson2012-05-311-2/+2
| | | | | | | | | | | | | When only the speaker is active, swap the L/R channels. If the headphone is active, do not swap the channels. In order to do this properly, the mixer controls in the headphone path must be applied in preference to the speaker path. Fixes bug: 6558120 Change-Id: Ica8d1fd31b3b1a98392c57db08c8374a4bf631e5
* Initial commit of grouper files.Jean-Baptiste Queru2012-05-294-0/+1710
Change-Id: I855f4685d11d3cd16e4efbeb35ae6389045cb200 Kernel: 8dd2eab Audio: Disable UART debug cable detection to support active
generated by cgit v1.2.3 (git 2.25.1) at 2025-12-20 00:48:40 +0000