summaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAgeFilesLines
* libsensors: fix build on MZiyan2016-01-171-0/+1
|
* Update external storage configurationZiyan2016-01-171-20/+2
| | | | Change-Id: I0512a0f7a089b748c199f1a8933c6323cda86a1e
* Drop BOARD_SEPOLICY_UNIONZiyan2016-01-171-10/+0
| | | | Change-Id: Ia6b21b115fe2c4efa8c91ef00c21ec876525febe
* [TEST] audio: Set short period size to 882.Kyle Repinski2016-01-151-1/+4
| | | | Just a hunch, see the comment...
* audio: Fix buffer underrun doing short to long period transition.Kyle Repinski2016-01-151-2/+4
| | | | | This basically halves the long period size. I wish there was a different way to fix this but I don't think there is.
* audio: further tweaking to use_long_periods.Kyle Repinski2016-01-151-19/+29
| | | | | | | | | | Move the detection in start_output_stream_deep_buffer down to allow a call to pcm_set_avail_min now. Did a few micro-optimizations as well. Conflicts: audio/audio_hw.c Change-Id: I90d5663040986ffd597f37ae66334467adacea3b
* audio: Make buffer_frames dependent on OUT_RESAMPLER.Kyle Repinski2016-01-151-3/+3
| | | | (buffer_frames is only used for OUT_RESAMPLER)
* audio: Decide use_long_periods on start_output_stream.Kyle Repinski2016-01-151-2/+7
| | | | | Previously this was corrected on the first write, however that was causing a buffer underrun sometimes when a stream was started. This avoids that.
* audio: Initial work towards a variable sampling rate.Kyle Repinski2016-01-153-5/+82
| | | | | | | Conflicts: audio/policy/audio_policy.default.conf Change-Id: I4d13946e88cacbf5e4ca383d5d0756262442efd2
* audio: Use separate sampling rate for MM-UL (input stream).Kyle Repinski2016-01-151-1/+6
| | | | Fixes MM-UL not working.
* audio: Fix PLAYBACK_PERIOD_COUNT and OUT_RESAMPLER definition detection.Kyle Repinski2016-01-151-19/+19
| | | | By moving this block down, the sampling rates are properly defined for this check.
* audio: Add ability to actually force the out stream sampling rate.Kyle Repinski2016-01-152-1/+17
| | | | | This audio HAL has always required some sort of resampling. THIS. ENDS. NOW.
* audio: Determine whether or not to use out stream resampler automatically.Kyle Repinski2016-01-152-3/+1
|
* audio: Disable output resampling by default.Kyle Repinski2016-01-151-1/+1
| | | | Seems to work OK without it.
* audio: Add ability to disable built-in out stream resample.Kyle Repinski2016-01-152-0/+25
| | | | | | | | | This needs further testing before enabling by default, but so far it's been OK. Conflicts: audio/audio_hw.c Change-Id: Ic4f86440ff4d01ab4d0d9f977bdec22f10f60555
* tuna: Proper 48kHz audio support.Kyle Repinski2016-01-152-3/+3
| | | | | | This fixes a bunch of annoying junk in the logcat about fast path being denied to UI sounds due to them having a 48000Hz sample rate. Also allows playback of 96kHz audio, as Android refuses to resample to anything lower than a divide by two.
* audio: Fix aliasing violation.MWisBest2016-01-151-2/+3
|
* SGX module building: remove unused PLATFORM_VERSION parameterZiyan2016-01-081-1/+1
| | | | Change-Id: I11c251e66072cb73cf24ce520e28735bb37a0cc3
* init: increase read-ahead value to 256 kbZiyan2016-01-081-0/+3
| | | | | | | | | Apps and database file sizes have increased in recent Android versions, which makes it practical to increase our read-ahead value from 128 kb. General user experience seem to have improved a bit. Change-Id: I35788444a8552a8f376c699ead163e951ed06b30
* Revert "Switch back to CFQ as the default I/O scheduler"Ziyan2016-01-081-1/+1
| | | | | | | | | The kernels block subsystem has been updated which seem to have eliminated the lockups we were seeing. This reverts commit 0c982e56af824cfbd621e6b4e8ca3183b87ea5e9. Change-Id: I17b0a0c724fbf1bbbf9f88a54a29447d7f592832
* init: tune vm for zRAMZiyan2016-01-081-0/+6
| | | | | | | | References: https://source.android.com/devices/tech/config/low-ram.html https://01.org/android-ia/user-guides/android-memory-tuning-android-5.0-and-5.1 Change-Id: I8ec35dad3a65c213d1843d01957b308c3ab84a6e
* Increase zRAM size to 256 mbZiyan2016-01-081-1/+1
| | | | | | | With the lz4-backed zRAM we're using, it seems like we can use larger values without noticeable slowdowns. Change-Id: I6a7921d0741c6b54629b456cc5387407956d98d8
* Enable EXTENDED_FONT_FOOTPRINTZiyan2016-01-081-0/+2
| | | | | | Recent changes made enough room on /system to include this. Change-Id: I01422ff52f2ad92da777875be9a39dd9fdbced69
* Remove packages built by defaultZiyan2016-01-081-10/+0
| | | | Change-Id: Icb9afbf59715fe9039c1c03bcce9e5327e31ff57
* Remove touchscreen firmware from userspaceZiyan2016-01-084-17/+0
| | | | | | | It makes more sense to just store it in the kernel, so it doesn't have to wait for the filesystem to get mounted. Change-Id: Ib5893157edd44806018f07ad8dad32ffe3704657
* Use 2 zRAM compression streams (now that our driver can)D. Andrei Măceș2016-01-081-1/+1
| | | | Change-Id: I84f26a40e5b89d100b5a1fd2eab1ba59e5957750
* Add zram information and mount it during bootHashcode2016-01-082-0/+7
| | | | Change-Id: I2a758e866c4518a1cbbd360bd3a06f3d922161f5
* Remove lowram flagsZiyan2016-01-041-5/+0
| | | | | | | Testing shows that only disabling the atlas service results in almost identical memory savings. Change-Id: Icc6c43fae3c1682b545d0dbc217f87b01f330fff
* Enable WITH_DEXPREOPT_BOOT_IMG_ONLYZiyan2016-01-041-1/+2
| | | | | | | | Otherwise, there's not enough space on /system to install nano gapps, to which the CM wiki points to. This increases the initial boot time by ~2 mins, but saves ~45 mb. Change-Id: I680b95c9a8481cdff3af7fe367c35fdf2b9dea66
* Update extract-files scriptZiyan2016-01-034-33/+25
| | | | Change-Id: Ifda025187ddade86357f6968465718ac72bf9a59
* Switch back to CFQ as the default I/O schedulerZiyan2016-01-031-1/+1
| | | | | | | | While BFQ significantly boosts the I/O throughput, it causes random lockups. We possibly have something conflicting with it in the kernel, but until it's fixed, let's switch back to CFQ. Change-Id: I8e189c16b09f3981836f22370c5a57025c9175c4
* camera: support tunaZiyan2016-01-036-12/+158
| | | | Change-Id: Id19387b819b2e118234e415b6ea0e229e5e5ac6c
* Add camera sources in-treeZiyan2016-01-0360-1/+42407
| | | | | | From hardware/ti/omap4 revision e57f2b6. Change-Id: I2e6164fdf6c0e2a2e75aee3bba3fe58a9c470add
* Remove unused tuna productsZiyan2016-01-033-72/+0
| | | | | | What are these doing here? Tuna never was a product on it's own. Change-Id: I397467fa8696948cd09671e8fba34aa055478707
* Remove bootloader and radio updatersZiyan2016-01-036-661/+0
| | | | | | | | | We're not going to create OTA packages with these images. samsung.fs_size_fix also got removed: nobody is going to install this to a factory (never installed any factory image, bootloader or ota package), encrypted maguro device. Change-Id: Iae292f4fce48eede0d5ac34e90096c0258552b53
* Remove unused reflash-bootloaderZiyan2016-01-033-383/+0
| | | | | | I don't see why we'd ever need this. Change-Id: Ibf03f193d2d0062e14e53d2d1a0a404089ee2179
* dumpdcc: No need to override path.Kyle Repinski2016-01-021-1/+0
| | | | | | BUILD_EXECUTABLE automagically puts it in bin. Change-Id: Ie709941cd2436794e63770f0a1ac1eafc52f6770
* nfc: Build firmware lib ourselves.Kyle Repinski2016-01-023-0/+5563
| | | | | | | | | | The prebuilt libpn544_fw.so has text relocations. Because the NFC app ends up targeting the latest SDK API level, linker doesn't allow any libs referenced by it to have text relocations, including the dlopen call to libpn544_fw.so; All this lib does is contain a couple char arrays with the fw and fw version. Change-Id: Iaac980193d92cfedb87f3cd80841f724f06d4876
* Disable dex2oat watchdogZiyan2016-01-021-1/+2
| | | | Change-Id: I1c8c182778fbd0538224e56612f0845837f33601
* init: disable tf_daemon and smc_pa_wvdrm servicesZiyan2016-01-021-0/+2
| | | | | | | These are not built since Android 4.1 by Google. However, since switching to hardware/ti/omap4, they get built again, just to start and crash. Change-Id: Ia729d01a3e9bf77b5e21d6acea09eb5ffdcff591
* Grab GPS config from device/commonZiyan2016-01-022-7/+1
| | | | | | It's exactly the same as ours, so there's no point having a custom one. Change-Id: Ieb685198295bd60bbd7766a0cff4389a8a51d070
* Add missing partition sizesZiyan2016-01-021-0/+3
| | | | Change-Id: Ib541776f46da58021c50da91b0a26650572b1144
* Set the camera HAL interface to OMX-only.Kyle Repinski2016-01-021-0/+1
| | | | | | We don't use V4L2/USB at all. Change-Id: I7f94ac228ac0e0dd55e2b89c9512eb32a513ef0f
* tuna: don't reset global DEVICE_PACKAGE_OVERLAYSAndreas Blaesius2016-01-021-1/+1
| | | | Change-Id: I810d5128a0b312a189cdeec23113d0c71dcefd6d
* tuna: switch to BFQ I/O schedulerZiyan2015-09-271-0/+4
| | | | Change-Id: Ic5db1b608e8d767bd7c31f6ca6feacaeeb68a1eb
* tuna: add minfree overlayZiyan2015-09-031-0/+9
| | | | Change-Id: Icd9542b065317a34a98dc5f4160c2149b6a1b31d
* tuna: treat as as low-ram via build.propZiyan2015-08-061-0/+9
| | | | | | Thanks to Hashcode for the values. Change-Id: I166619632762da036fa6f2512d4092f7a64587a4
* tuna: minor cleanupsZiyan2015-08-061-31/+11
| | | | Change-Id: I5e5d18ad54652bdaf61ccf7ad0372439dd8eca73
* tuna: update SELinux policiesZiyan2015-08-069-17/+31
| | | | Change-Id: I0509465046ae5b22f4ab1e857db0645075d66628
* tuna: Remove deprecated ffmpeg configfuzz2015-07-211-1/+0
| | | | Change-Id: I7a0c6a12b28df732be9d1df8963ca30307b19d22
generated by cgit v1.1 at 2025-01-05 14:12:51 +0000