Bug: 193715461
Test: `vendor.keymint-trusty` and `vendor.gatekeeper-1-0` services up
Change-Id: If49ff7daf16ef9f1dba9eedf5979d162ac96f756
Signed-off-by: Armelle Laine <armellel@google.com>
Allows to enable Secure Storage and hardware RPMb
Bug: 192512209
Test: storageproxyd service started during boot
Change-Id: I9ceb5e8149c7279d5094c09982e0cac81c90871a
Signed-off-by: Armelle Laine <armellel@google.com>
Re-enable this hal now that OpenCL is running
Test: build, see hal, pass 54k CTS tests (1 seems to hang).
Fixes: 193679354
Change-Id: Ie88d2f6c4853145d8f8d9ed552ca33de75899157
This reverts commit 1396b4b523.
Reason for revert: Re-enable sensors as AOC boot latency is resolved.
Change-Id: I421f558f3ac05d872b8b398c6bc7f4229cf303bc
Change to sensors.gs201 to cover all gs201 devices when using fake sensors.
Signed-off-by: Pat Tjin <pattjin@google.com>
Change-Id: Ic06fee767d9e0da1cd1aa1a558e1a842a489a85a
This is to expose a common path to runtime. This is the gs201 version of the WCH change ag/15100070. It alias /dev/janeiro to /dev/edgetpu. This is a temporary measure to make /dev/edgetpu work. Once the TPU kernel driver is refactored to produce /dev/edgetpu by default, we'll remove this symlink.
Test: Manual. Verified edgetpu_logging works with modified path
Bug: 191193044
Change-Id: Icc70264c0c7d20ea7b6aa43903e109d9180dd022
Previously, we ran modprobe command for every .ko file listed in
init.insmod.*.cfg. Let's move to a different approach where we call
modprobe with the following command line.
modprobe -b --all=/vendor/lib/modules/modules.load
This will insmod all modules listed in modules.load except for those
that are mentioned in the blocklist at
/vendor/lib/modules/modules.blocklist
A common reason for a module to be on the blocklist is that it must only
be loaded under a certain condition like when a specific service is
launched.
Bug: 192241728
Change-Id: I518426308d0bfc358c8c815a69d431e27ebcf6e4
Kernel modules on the blocklist are not automatically loaded during
second stage init. Modules are often put on the blocklist if we want
them to get loaded only under certain circumstances.
Bug: 192241728
Change-Id: I05d55f8a2854619b92defcf3fb11cc2b87a8dab6
Previously, we copied all available kernel modules into the vendor_boot
ramdisk except for approx. 14 modules that we explicitly excluded.
Going forward, the kernel build will distribute the two additional files
vendor_boot.modules.load and vendor_dlkm.modules.load which define,
respectively, the lists of modules that should be loaded from
vendor_boot and vendor_dlkm.
The contents of the two *.modules.load files will be copied almost
verbatim into modules.load files in the respective images except for the
fact that the directory names of the .ko files are stripped. So, for
example, kernel/net/core/pktgen.ko becomes just pktgen.ko.
Additionally, we only copy those .ko files into the vendor_boot image
that are listed in vendor_boot.modules.load. On the other hand, the
vendor_dlkm partition image will contain all .ko files regardless of
whether they are listed in vendor_dlkm.modules.load.
Bug: 190652328
Change-Id: I6fc41a0143e7253f193cac509ba30506feb4e2d9
This allows EdgeTPU component to build different components
per target chip.
Bug: 192096682
Test: tested on prebuilt.
Change-Id: Id9ee7de22648c591f8a540c5441ad375a9ee3439
This partially reverts commit 28140b7240.
USB will hold a wakelock preventing suspend to happen
Bug: 188796118
Signed-off-by: Thierry Strudel <tstrudel@google.com>
Change-Id: Idc583a025e429a5b5f13a94b732717e9491e4e55
Give TPU device path (/dev/janeiro) the same access permission as /dev/abrolhos. This change is similar
to what we did for WCH, but for PRO.
Bug: 191185522
Test: run_tflite_test_odc passes.
https://paste.googleplex.com/5466657955774464
Change-Id: If735e699c0c9ec842f915ebc7b59d0896f4c776c
This required disabling the GPU-related PRODUCT_SOONG_NAMESPACES so that
we don't get an installed-path collision between the prebuilt gs201
libOpenCL and the gs101 libOpenCL that we build from source. Once we're
able to build both gs101 and gs201 from the same source (b/185349920)
we'll need to re-add the namespaces.
Bug: 190424017
Test: gfx-score
Change-Id: I9095e099a120494fab0c33fc6bd1ace2c69f3bc4
