At the second update for Virtual A/B device,
PartitionCowCreator thinks the COW partitions in
the existing super metadata as occupied. In fact, these
partitions aren't used after the merge.
Now, unmap these partitions from the device mapper. If
something bad happens and the previous update has not been
merged yet, the unmap will fail. Then, delete these old
COW partitions from the device.
Test: Virtual A/B update twice
Bug: 135752105
Change-Id: Iab867ded19755089e6e0800e553a10fbcddbb931
- Add TestDeviceInfo(string, string) to avoid holding
the 'rebooted' TestDeviceInfo variable after ownership is
transferred
- OTA client calls BeginUpdate after unmapping partitions.
Test: run libsnapshot_test
Change-Id: Id37e34b421d728eb320c0e0906abef3b61897e7c
There are no uevent fils in these dirs, so there's no need to
traverse these dirs.
Test: manual
Change-Id: I57b82846ad863952d8717e6c3ab80c1e243ee4f8
Signed-off-by: lijiazi <lijiazi@xiaomi.com>
... which handles merge when boot completed. It also
dumps debug information when requested.
Bug: 135752105
Test: adb shell su 0 snapshotctl dump
Test: call snapshotctl merge at different stage of OTA:
- before OTA (exit normally)
- during OTA is applied (exit with error)
- after OTA is applied but before reboot (exit with error)
- after reboot (finish merge and exit normally)
- Manually stop its execution during merge and run again
(finish merge and exit normally)
Change-Id: Idcc6aea8d7bbeb9a1a288c966b8f5e14b3f6a3e7
Instead, the bug_map entries will now include the b/ prefix. This will
allow arbitrary URLs to be added as well to SELinux denial messages
Bug: 141014771
Test: Generated a denial, verified that the bug id in the dmesg logs
remains unchanged.
Change-Id: I4679117abdc2e13af81336b0b7fd8d69def80147
Execute mkdir -p /mnt/pass_through/0/emulated
and bootstrap the emulated volume for user 0.
The 'pass_through' paths allow us bind mount the lower
filesystem directly into /storage, bypassing any sdcardfs
of FUSE mounts.
This change is part of enabling upcoming platform changes that are
described in the bug linked below.
Bug: 135341433
Test: builds, boots and pass through directories are created
Change-Id: I46ce207d06a1ec550b8bacac259387371fc0b841
There is still some flakiness, so increase the timeout values.
Also remove the TEMP_FAILURE_RETRY macro usage in TIMEOUT calls.
That macro disables the ability of the alarm code to interrupt
the system call.
Bug: 141045754
Test: Unit tests pass.
Change-Id: Ia3c95dccc3076a3fd5ef6432097a57e4ccee4df3
libcutils has trace implemented for host. In order to increase code
portability, opening that up for libutils Trace.h usage as well.
Bug: 124524556
Test: use Trace.h on host glinux
Change-Id: Ia873f88e7436a18f6c0f761000bf697c68ffea17
Shared VNDK should be used instead for vendor modules.
Bug: 137178339
Test: m
Test: run property_info_checker on host
Change-Id: I40672d0f6251148915f9ae2e2a14f291d9c2b6c5
There might be partitions like product or system_ext in addition
to the system partition. Those partitions can contain dependencies
required by the system so we need to make the init to recognize
these partitions and mount them accordingly.
Bug: 140092208
Test: gsi_tool install & reboot
Change-Id: Ie5ef063292c3656e79cce9f887f23a8faa1ba2be
There is a race condition with Service::Start and socket creation.
Since socket creation currently happens after the fork(), it's
possible that init can continue executing other commands before the
socket is created. If init starts another service that relies on that
socket, it isn't guaranteed to be available.
Particularly, we've seen this with hwservicemanager starting after
logd, but hwservicemanager's logs sometimes not showing up.
Bug: 140810300
Test: boot and logging functions correctly
Change-Id: Ib2932e836d345830cd38f3b556598508fd953058
The Android framework, notably android.view.Choreographer, assumes
that System.nanoTime and SystemClock.uptimeMills return consistent
values. This was true on device, but not on host.
This commit makes those values consistent on host linux. The necessary
support should be in place in kernels 2.6.39 and newer, which have been
available since 2012.
Test: m -j libutils
Change-Id: I833a89a810ae9fb3e8c01f6095ee2aca893c284f
If this check fails and an OTA or mainline module update has recently
happened, we want to rollback the recent change. The easiest way to
handle this is to reboot, which will trigger the fallback mechanisms
that are already in place.
Bug: 141082587
Test: device reboots if self test fails
Test: device rolls back a recently applied OTA with failing self test
Test: device rolls back a recently applied conscrypt apex update with
failing self test
Change-Id: Iff879deff09d347262dc7a2acadb9164a5029d4a
All of the logic in reboot.cpp is meant to safely shutdown services,
safely unmount emulated RW file systems, then finally unmount the
remaining RW file systems, particularly /data. If /data hasn't been
mounted, then none of this logic is required.
Running this logic caused a lock up when shutting down blueline from
early-init. Vold, or potentially a related HAL, locked up during the
ShutdownVold() calls. debuggerd separately locked up in the watchdog
thread.
Therefore, this change immediately reboots if /data is not mounted.
It also removes the lines to call into debuggerd. debuggerd will not
run due to SELinux in any case, so it can only be used when hands-on
debugging a device.
Bug: 141082587
Test: shutdown with /data mounted continues as normal
Test: shutdown from early-init immediately shuts the device down
Change-Id: I79c72346b17c7dfe57e955d9739bcaf559badc14
No longer required for hwasan builds, since b/140790209 is fixed.
This reverts change-id I27069b20f7c7068b931340f548b284ce1676466c.
This also reverts the identified change for the memory regression in
b/140648539.
Test: Build & boot on taimen_hwasan-userdebug
Test: Build & boot on crosshatch_hwasan-userdebug
Bug: 140790209
Bug: 139408016
Bug: 140648539
Change-Id: Ib61c53571a4de0970a86ebc391a2ce780247943a
Any used files need to be in the dependencies list for rules. In this
case, this is more than what's actually read in the script, but the
older ones shouldn't change as much, and replicating the logic is
non-trivial.
Bug: 130111713
Test: run with RBE, no longer see error about file not found
Change-Id: I88baf541ce6250a5dbf7b8a7d6b8005ed7cf5cc6
This commit constructs the super_vbmeta image to eliminate the
need of adding /vbmeta_system and/or /vbmeta_vendor when AVB
chain partition is used with Android Dynamic Partition.
See BOARD_AVB_VBMETA_SYSTEM under the link:
https://android.googlesource.com/platform/external/avb/#build-system-integration
The structure of super_vbmeta :
| VBMeta Table | (fixed-length 2KiB)
| Backup VBMeta Table | (fixed-length 2KiB)
| VBMeta Images | (fixed-length 64KiB each)
The structure of VBMeta Table :
| Super VBMeta Header | (fixed-length 128B)
| VBMeta Descriptors | (variable-length)
The VBMeta Table records the slot number of each
vbmeta image within the /super_vbmeta partition.
Bug: 137054296
Test: m libvbmeta_test
Test: ./out/host/linux-x86/nativetest/libvbmeta_test/libvbmeta_test
Change-Id: I01aeadd850750ae87d9125484c1b1f570bb84756
