Adding formatting here, as we will start a push for people
using this, if they run into any linkToDeath issue, or if
they want to test this.
Bugs: me
Test: N/A
Change-Id: I72510d61926ce694671d897519463f37892a8d12
Add post-fs-data-checkpointed trigger when vold completes checkpointing
after an OTA update.
Bug: 350362101
Test: ABTD
Change-Id: I647a73a942174015b46c5f40bd8f8d3347977ecd
This commit introduces a function to allow ARC-specific logic during
first-stage mount by checking for the existence of an indicator file
at the path "/is_arcvm".
ARC uses the virtio-console device `hvc1` to pass byte data to Android
before second-stage init. Ensure that /dev/hvc1 is initialized before
ARCVM continues booting, but allow other devices to bypass this device
initialization.
Bug: 325538592
Test: boot ARC, ensure /dev/hvc1 can be read during PropertyInit
Change-Id: Ic258b7b004b59da462f4990131a5c11fc94eca62
This change allows init to ensure that a specified virtio-console
device file (`/dev/hvc*`) is available before `ueventd` coldboot.
Times out if device path is not encountered within 10 seconds.
Bug: 325538592
Test: build bertha_x86_64 and bertha_arm64
Change-Id: Ia1512e69ea607bf4d235595caa53668e2dac500c
This is a follow-up on I37380c19232f2c497bdf492a83cdc16616f0ae8d.
Bug: 338160898
Bug: 345110999
Test: Microdroid boots even with BOARD_USES_RECOVERY_AS_BOOT
Change-Id: I41c1e40aeaffd5499fb6bd25e80b5be83470bc6b
In a GRF config, if Vendor partition is updated from Android 12; post
OTA reboot, first stage init will communicate to daemon to check if the
daemon can support socket handoff. If that succeeds, then it is a signal
that the vendor has been updated from Android 12. Use a marker in
/metadata to signal that the vendor was updated. If the marker is present,
then post OTA reboot, userspace snapshot will be used.
Bug: 333854394
Test: OTA
Android U (system) + S (vendor) -> Android V (system) + V (Vendor)
Change-Id: Ie38c4379010789a84e5b44529b407f9f82135271
Signed-off-by: Akilesh Kailash <akailash@google.com>
The `select` syntax rewrite makes it more concise and easier to
understand.
Bug: 347605145
Test: m init_vendor
Change-Id: I866bbe9360fdbdf69cac3c6a24bbe37306227755
`init_first_stage` is a dependency of `init_vendor` only when
`BOARD_USES_RECOVERY_AS_BOOT` is false.
Since `BOARD_USES_RECOVERY_AS_BOOT` is already defined in
`build/make/core/android_soong_config_vars.mk` within a
soong_namespace, we can use the `soong_config_module_type` to easily
convert this to Android.bp.
Bug: 347600829
Test: m init_vendor
Change-Id: I1ddcd5fb62983b01e51452c9b7367750e03e7f48
For visibility.
We could make this only for new API levels, but it isn't
currently exposed at build time, and visibility is good
on upgrades.
Bug: 340953047
Test: build, on device passing and failing requirements
Change-Id: I3a0ea47560c65114bc1b8685954d1fb7687cb8df
This changes user root to use the year date format
that vendor API uses.
Previously this still applied to V devices, which switched
to using API level 202404 after 35.
They recommend the year/date form, and there is no
constant for this.
Bug: 340953047
Test: service_test
Change-Id: I5ba8d1251c877a785e41f0cfcc35d7cb0776f8c7
This link will be used to change the sysfs attributes of the zoned block
device from an .rc file.
Bug: 335708738
Change-Id: I99f74c121e7d9da404c0564860c03ac1efe6c6d8
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
Make all member functions private that are not called from outside
DeviceHandlermember functions.
Bug: 335708738
Change-Id: Ibe2e5f6b8b45a79b5be59665627e0a46de4e439a
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
Slightly reduce the time that is required to start the init process by
changing two static std::string variables into std::string_view variables.
Change-Id: Ib683a603da5f3d8968b26453b12544fed8a6397f
Signed-off-by: Bart Van Assche <bvanassche@google.com>
Linux 6.9+ supports 33 bits with Sv48 and Sv57. Increase the
maximum and update the comment.
Bug: https://github.com/google/android-riscv64/issues/1
Test: Treehugger
Change-Id: Ia2731143ad30c5febe2058d35a381c01b14d8c3c
We'll make it a build time check in V after this.
Due to being a runtime flog, it was hard to spot and
debug, so delaying.
Bug: 340953047
Test: boot, CtsInitTestCases
Change-Id: Ica1a1c780e230c88615dad7937404d2b0fe9c68f
When someone uses the 16kb dev option, they switch
out the kernel, but the same build is used. Make
sure this is visible in the fingerprint to help
split up any analytics or reports.
Bug: 332406754
Test: no change on normal device, on 16k dev option device:
:) adb shell getprop ro.product.build.16k_page.enabled
true
:) adb shell getconf PAGE_SIZE
16384
:) adb shell getprop ro.build.fingerprint
generic/aosp_cf_x86_64_phone_pgagnostic_16kb/vsoc_x86_64_pgagnostic:VanillaIceCream/MAIN.b7aab5e4/eng.smorel.20240514.203625:userdebug/test-keys
^^^^^^
----------------------------/
cheers
Change-Id: I9632237cef4a3b077761022eda8cf42784a41753
Turns out we didn't need the syscall because the kernel checks
for async MTE fault on exit.
Test: atest mte_upgrade_test with and without MTE
Change-Id: Iebdf181d67b4586926499c7549c72e73efc13c56
init ignored the rc file because of incorrect permissions
also replace deprecated "push" with "push-file"
Test: atest mte_upgrade_test
Bug: 335232356
Change-Id: I4c17311dbb1d001567f4cdda65036f16aec2b5c1
So far, we have used `instalable: false` to avoid collision with the
other modules that are installed to the same path. A typical example was
<foo> and <foo>.microdroid. The latter is a modified version of the
former for the inclusion of the microdroid image. They however both have
the same instalation path (ex: system/bin) and stem (ex: foo) so that we
can reference them using the same path regardless of whether we are in
Android or microdroid.
However, the use of `installable: false` for the purpose is actually
incorrect, because `installable: false` also means, obviously, "this
module shouldn't be installed". The only reason this incorrect way has
worked is simply because packaging modules (ex: android_filesystem)
didn't respect the property when gathering the modules.
As packaging modules are now fixed to respect `installable: false`, we
need a correct way of avoiding the collision. `no_full_install: true` is
it.
If a module has this property set to true, it is never installed to the
full instal path like out/target/product/<partition>/... It can be
installed only via packaging modules.
Bug: 338160898
Test: m
Change-Id: I37380c19232f2c497bdf492a83cdc16616f0ae8d
