gr_clip was not calculating the correct values after the rotation
patch. Since gr_clip width and height instead of taking a second
pair of x,y coordinates, we cannot depend as much on the
ROTATION_X_DISP and ROTATION_Y_DISP macros. The simplest solution
was to call scissor with direct values in most cases.
Change-Id: I25f9d575ec6ddc430fca4555f29c8334441fc95d
libclang_rt.ubsan_standalone-aarch64-android.so is a library for
arm64 and is therefore not available on other architectures, causing
the build to fail because there is no rule to build it.
Change-Id: Ia4af38f42ba46a233d40e936d43e4c97c449d059
* The existence of TW_ROTATION that implements this feature at the
level of calls to libpixelflinger API closely mirrors the existence of
ro.sf.hwrotation for surfaceflinger in LineageOS.
* A brute-force approach was previously attempted via the
BOARD_HAS_FLIPPED_SCREEN makefile flag. That code iterated over the
active display surface in a double-buffered setup, and performed a
"smart" memcpy from the UI drawing surface (gr_draw) onto the display
surface. The problem was that, without heavy loop optimizations, that
code could have never scaled for 90 and 270 degree rotation.
I tried and you could literally see the for loop with the naked eye
while the display surface was updating.
* That code is now gone, but support for BOARD_HAS_FLIPPED_SCREEN := true
is still there (now means TW_ROTATION := 180).
* This patch relies on the assumption that it is impossibly difficult
and non-portable to rotate whole framebuffer display surfaces, in a
way that is not dependent upon the graphics backend (adf, fbdev, drm,
overlay etc). Therefore, it identifies the rendering primitives that
the TWRP graphics stack exposes to the GUI application above, and
implements hwrotation inside each of those calls instead:
- gr_line(), gr_fill() - 2D geometric shapes (lines, rectangles)
- gr_blit() - graphical image resources
- gr_ttf_textExWH() - font rendering
- gr_fb_width(), gr_fb_height() - framebuffer resolution
* The gist is to keep the backend and framebuffer (dimensions, row size
etc) unchanged (because making changes there is asking for trouble),
but present an altogether different reality to the calling API,
according to the compile-time constant TW_ROTATION.
* All (x, y) API coordinates and shapes are transformed before being
actually rendered as (x_disp, y_disp) display coordinates.
* With TW_ROTATION := 90 or 270 you can turn a landscape device into
a portrait one, because the GUI is fooled by the reversed dimensions
reported by gr_fb_width() and gr_fb_height() and renders the UI as
for a different device.
* For blit and text rendering operations, figuring out the transformed
coordinates in display space is not enough, as the surfaces that are
to be rendered have to be rotated themselves. This is handled by
allocating an intermediary rotated surface on each rendering
operation (not ideal), so the code with the intermediary surface
is compiled out for the TW_ROTATION := 0 case.
* This is still not as bad as rotating the whole framebuffer though, and
on a msm8976 device the performance hit is not even noticeable (for
software rendering).
* Currently there is no attempt to make a connection between the
TW_ROTATION and the { RECOVERY_TOUCHSCREEN_SWAP_XY,
RECOVERY_TOUCHSCREEN_FLIP_X, RECOVERY_TOUCHSCREEN_FLIP_Y } settings.
Change-Id: Ic8966ad5360c8a499649fdb16e242286640fd992
Signed-off-by: Vladimir Oltean <olteanv@gmail.com>
This patchset updates all the places we touch cache on AB devices
and non AB devices so that we can write to the appropriate
location. AB devices removed the cache partition and moved it to
/data/cache.
Change-Id: I485ec205a604f1f07ceb7700ae9c4e720fd55c3e
Some devices do not declare BOARD_CACHEIMAGE_PARTITION_SIZE.
This case is detected and required modules are added to LOCAL_REQUIRED_MODULES.
However, := is used instead of += and causes the loss of all previously added modules.
The end result is that twrp is not actually getting built since its module was lost.
Fix that by using += when adding modules required for cache-less devices.
Change-Id: Ib0a4db4955040ced77be7ed04d636905beadf093
Fixes broken shell prompt in Terminal and adb shell
Before: https://del.dog/7vphlkc
After: https://del.dog/vhkjgo8
Add as prebuilt with suffix "_twrp" to avoid conflict
with original mkshrc and rename via POST_INSTALL_CMD
Prebuilt version adjusts TMPDIR variable from
"/data/local/tmp" to "/tmp" for TWRP
Change-Id: I42a1b2a8c114f66dbe659c07485f1dfae8a5d314
This will add another path where looking for ats_* files
This is needed on some devices like OnePlus 6/6T
Change-Id: I9ac7cdd05ddc80f3e269efda96be2b7712d68af0
* build modified vdc_pie binary with 'checkpw' command support
if building with Android 9.0 platform. That command and others
we don't care about, are removed from Pie vdc. Our vdc_pie will
run if system sdk version is > 27, otherwise system vdc is used.
Code adapted from Android 9.0 system/vold/vdc.
* include prebuilt vdc_pie(arm, arm64) binary if building with lower
than Android 9.0 platform - vdc_pie cannot be build from source
with those platforms without additional imports from Android 9.0
* skip vdc "getpwtype" command for Pie - vds communicates with vold
directly, no need for connection retries first
* add /system/bin/servicemanager to required services
* mount per-devive additional partitions needed for decryption
listed with device BoardConfig.mk TW_CRYPTO_SYSTEM_VOLD_MOUNT
flag like(space separated):
TW_CRYPTO_SYSTEM_VOLD_MOUNT := vendor cust odm
* add function to backup crypto footer before running vdc commands
and restore it after - on Xiaomi Mi Max 3 both Oreo and Pie stock
roms vold alters cripto footer when decrypting data in recovery
which causes system to ask for crypto password at next reboot
although password stays unchanged. Crypto footer backup/restore
added as workaround for systems whit ro.build.version.sdk > 25.
Also to preserve crypto footer integrity decryption attempts are
skipped if footer backup fails to ensure no data loss.
Code adapted from
https://gerrit.omnirom.org/#/c/android_bootable_recovery/+/31206/
Change-Id: I0a383f3843578fa55595cfea3b7c9c4431646a1a
-switch to full screen when user is not suppose to do
anything but wait for current operation to finish
Change-Id: Ib54afc380a5cfe70e011e782cf3d7fe4c80958a4
This reverts commit 1ccdefab79.
Reason for revert: Causes delay during boot resulting in failed decryption on FDE devices. Recommend having device maintainers specify this variable manually in rc file and include actual ld.config.txt (or placeholder) in ramdisk directly.
Change-Id: Ib2410675396b38522d690c6057057d164cc0cff2
Currently when building for SDK 27 with TW_USE_TOOLBOX,
no getprop symlink is produced. This minor change creates
a symlink for getprop on SDK 27.
Change-Id: I71bd3e0d0525bd0ae51634f52a5749ba0ff58444
Fix rules for adf graphics to detect their presence when ninja
make files are used.
Change pixelflinger graphics mode so that colors display properly
on the Pixel 3
Change-Id: Ie7181be6c5e1e9a007b58fd798b2cef787ce4a01
external/libdrm was fully converted to Android.bp on Pie, causing
the wildcard to match nothing, even though libdrm is present in
the build tree. Change it to cover both build files.
Change-Id: I4b092b29ac021cc1aa3bcf7346b225b79fd99f4b
This type of vibrator is found on newer kernel versions (4.9+) and
registers with LED class framework (located at /sys/class/leds/vibrator).
Change-Id: I85e93fdac17b3f4b6f2ae689bbbd490806b5c29b
Respect TARGET_RECOVERY_PIXEL_FORMAT config in screenshot, graphics.
This fixes other wrong theme color in minuitwrp on Xiaomi MIX 2S.
Change-Id: Ieb8480c411e2f0c72cc50ffca66943ab025e2b7e
Rather than using hard-coded system, use ANDROID_ROOT
environment variable to allow AB devices to mount
system_root at a custom path. This allows the /system
path to be bind mounted from $ANDROID_ROOT/system
so that the vold_decrypt paths can remain unchanged.
Change-Id: I9a7b13385e43f169f1df4c75b2a003fc6913952c
* Following error while running mke2fs in TWRP compiled in 8.1+ tree:
libc: CANNOT LINK EXECUTABLE "mke2fs": library "libext2_misc.so" not found
Change-Id: I786f325229167dbd0d1d5f2a1ea95e9fe15a227e
Signed-off-by: Vladimir Oltean <olteanv@gmail.com>
On some devices (e.g. ASUS Zenfone) the serial number of the device
is stored in a file in the /factory partition. Consequently, the only
way to load it is using a user space application during the init
process, and to update the ro.serialno system property.
In most places, TWRP already uses the system property to get the
serial number of the device. However, when generating the device ID
used for the backup folder name it checks the boot command line
directly. On these devices, the serialno is not included on the
command line.
Add a TW_USE_SERIALNO_PROPERTY_FOR_DEVICE_ID option that enables
usage of ro.serialno for generating the device id. It is disabled
by default to avoid changing the device id on existing ports.
Change-Id: I4a2eeca883f38fdaeb1209507fd8ebe44b1b04a0
Building in 9.0 may require you to add a flag to your twrp fstab
with the fileencryption details like:
fileencryption=ice:aes-256-heh
Verify this against your device's stock fstab of course.
Change-Id: If9286f5d5787280814daca9fbc8f5191ff26a839
instead of hard-coded /system path
I updated most of the references I found,
but there might be more
For devices that have to mount system at /system_root,
this allows system to be bind mounted to /system
and detected properly by TWRP
Change-Id: I9f142fd8cec392f5b88e95476258dab9c21a9aac
Ethan Yonker
bigbiff bigbiff
lambdadroid
Vladimir Oltean
oshmoun
Captain Throwback
Mauronofrio Matarrese
Erfan Abdi
nijel8
Andreas Schneider
Simon Shi
notsyncing
Captain Throwback
Simon Shields
Sean hoyt