* Choreographer VSync has some performance impact for CPU rendered apps
as the app has to wait for VSync to start a new frame draw. When the
frame draw time exceeds one VSync, this has a direct impact on the FPS.
debug.cpurend.vsync property controls the VSync for CPU rendered apps.
Disable VSync for CPU rendered app, by setting this property to false.
Signed-off-by: ImPrashantt <prashant33968@gmail.com>
Signed-off-by: pix106 <sbordenave@gmail.com>
Enable the persist.sys.fuse.passthrough.enable flag for the P21 device
configuration to enable the FUSE passthrough feature.
This feature has been enabled on Cuttlefish, Wembley and Redbull devices for
months and no issues have been detected yet.
Bug: 168023149
Test: 'adb shell getprop | grep persist.sys.fuse.passthrough.enable',
ScopedStorageTest
Signed-off-by: Alessio Balsini <balsini@google.com>
Change-Id: I3f39b02af436f7e508bb70efc702565a667a051c
Signed-off-by: Reinazhard <reinazhard@gmail.com>
Signed-off-by: OmarAlCoptan <omarcoptan9@gmail.com>
Signed-off-by: pix106 <sbordenave@gmail.com>
* Since there are no need for us to do any specific changes here
Signed-off-by: Ratoriku <a1063021545@gmail.com>
Signed-off-by: raiyanbinmohsin <raiyanbinmohsinshishir@gmail.com>
Signed-off-by: pix106 <sbordenave@gmail.com>
* Some banking apps read the file /proc/net/unix to
find out whether things like Magisk are installed/running.
To prevent that, chmod it 440.
This file isn't needed by any other process when boot is finished.
I've tested this and banking apps that did not work before
and recognized Magisk being installed now don't recognize
that and work just fine.
Signed-off-by: anandhan07 <anandhasayanan007jb@gmail.com>
Signed-off-by: rk134 <rahul.kantrapally@gmail.com>
Signed-off-by: Rahul K <rahul-k@bigdi.cc>
Signed-off-by: rk134 <rahul-k@bigdi.cc>
Signed-off-by: OmarAlCoptan <omarcoptan9@gmail.com>
Signed-off-by: pix106 <sbordenave@gmail.com>
Extracted from RD1A.200810.020.
These values appear to be the same across all Pixel devices, including
taimen, walleye, and redfin.
Change-Id: I253e4a1b98d1b777dcae3c69c40b132397ba807d
Signed-off-by: Vitorgl2003 <vitorgl.2003@gmail.com>
Signed-off-by: pix106 <sbordenave@gmail.com>
- This is for Fixing Lagging if Scroll to Menu Application and QS Scroll Down.
Change-Id: I75959ac082f3509560b0280da217baad38a617c5
Signed-off-by: PrajjuS <theprajjus@gmail.com>
Signed-off-by: Rahul K <rahul-k@bigdi.cc>
Signed-off-by: rk134 <rahul-k@bigdi.cc>
Signed-off-by: ImPrashantt <prashant33968@gmail.com>
Signed-off-by: pix106 <sbordenave@gmail.com>
* Needs an external patch to BoardConfigQcom.mk to ensure the platform is
included as a supported QSSI target
- Requires: Allow targets to upgrade to UM 4.19 family in vendor/{rom}
Signed-off-by: Albert I <kras@raphielgang.org>
Change-Id: Ia5c0084914314288b819971e30557839e375f091
Signed-off-by: clarencelol <clarencekuiek@icloud.com>
Signed-off-by: pix106 <sbordenave@gmail.com>
The blobs link against this interface library because of a fallback
mechanism when mapper@3.0 isn't made available (all our devices should
always provide that, never 2.0). Consequently this results in a silent
failure (not even a linker error appears) when trying to open the GPU
through EGL:
I AdrenoGLES-0: Driver Path : /odm/lib64/egl/libGLESv2_adreno.so
W libEGL : eglInitialize(0xb40000743497a2d0) failed (EGL_BAD_ALLOC)
E libEGL : call to OpenGL ES API with no current context (logged once per thread)
F RenderEngine: eglQueryStringImplementationANDROID(EGL_VERSION) failed
Adding this library - which is not explicitly linked/used by any other
bits in the OSS part of the stack - resolves that error.
Note that while the composer has a similar fallback mechanism to 2.0 it
only uses (and links against) the AOSP interface
(`android.hardware.graphics.mapper@2.0`), not the QTI one at
`vendor.qti.hardware.display.mapper@2.0`.
Note^2: All these hidl interfaces are defined as `system_ext`, which is
the desired place for packages (in this case interface
updates/extensions to the AOSP variant) may they ever land upstream.
We have not configured such a partition currently (on super) but instead
rely on it being in `/system/system_ext`. This is not compatible with
GSIs; we just request the vendor variant to be put on `/vendor`
(`.vendor` postfix on the package name).
Signed-off-by: Marijn Suijten <marijns95@gmail.com>
Signed-off-by: Jabiyeff <cebiyevanar@gmail.com>
Signed-off-by: clarencelol <clarencekuiek@icloud.com>
Signed-off-by: pix106 <sbordenave@gmail.com>
* Thanks to kunmun for the blobs
* Did not update Graphics blobs
* Reverts "sdm660-common: Decommmonize cnss-daemon (WiFi)"
Signed-off-by: clarencelol <clarencekuiek@icloud.com>
Signed-off-by: pix106 <sbordenave@gmail.com>
* Fully using ginkgo radio props as we are using their blobs as well
Signed-off-by: clarencelol <clarencekuiek@icloud.com>
Signed-off-by: pix106 <sbordenave@gmail.com>
Bug: 187912470
Test: make -j128
Test: deep buffer and compress offload via speaker, USB headset, BT
earphone
Test: phone call via handset
Test: hotword detection
Signed-off-by: pix106 <sbordenave@gmail.com>
Kanged from CAF product-board
This also fixed triple frame buffer regression
Nuke triple frame buffer from prop as it causing regression
Signed-off-by: Reinazhard <reinazhard@gmail.com>
Signed-off-by: clarencelol <clarencekuiek@icloud.com>
Signed-off-by: pix106 <sbordenave@gmail.com>
