During userspace reboot FscryptInstallKeyring will be called again, this
CL will make it second call a no-op, which IMHO is better than having a
special logic in init to conditionally call FscryptInstallKeyring
depending on whenever it's normal boot, or userspace reboot.
Test: adb reboot userspace
Test: checked in kernel logs that new keyring is not created
Bug: 135984674
Change-Id: I4ad5aee6887b7318fb1cd02bf1c7be8da6ece599
We're never going to fix this warning and it's not causing any active
issues that vendors are using this header. It is however causing
active issues that we have extraneous warning messages in the log.
Bug: 123758136
Test: no build warnings from <cutils/log.h>
Change-Id: Iccf702bf2536ba106a00ac04467d54401cbe2a50
The code was not properly getting the variable addresses and using
the offset and address fields of the .data section.
Fix all of that, and update the tests.
Bug: 145162678
Test: Unit tests pass.
Test: ./art/test/run-test --dex2oat-jobs 4 --host --prebuild --compact-dex-level fast --jit --no-relocate --runtime-option -Xcheck:jni 137-cfi
Test: ./art/test/testrunner/testrunner.py -t 137 --host
Change-Id: Ic61c4487334fd2273cda9c56eb1a3b525a03edb7
Helper function to compare the source and destination extents of a
SOURCE_COPY InstallOperation.
The function returns true iff source and destination are identical with
the use of std::equal().
Bug: 141207436
Test: build
Change-Id: I146aeba1c8ede35f21cfef8e21d4af62274bda84
Signed-off-by: Alessio Balsini <balsini@google.com>
dm-verity and error correction use part of the partitions using these
features to store their data. Their data may be modified during an
update, so the COW device for the dm-snapshot must reserve some extra
space for them.
This patch extends the PartitionCowCreator data structure with the
extra_extents field that will store the (optional) additional extents,
e.g., the hash table extent for dm-verity or the error correction
extent.
Test: incremental OTA apply
Test: libsnapshot_test
Bug: 145180464
Change-Id: I387a6cc8438507ad41a85cc3400241ecaf627b8f
Signed-off-by: Alessio Balsini <balsini@google.com>
Logcat dup-es STDOUT and used the dup-ed file descriptor to write logging
information. Because of this writes happen on that fd and not on stdout.
This in turn causes no output in the LogcatService in the emulator.
The qemu-pipe forwarder now dup-es the same fd as logcat is using so
that we receive data again.
Bug: b/145214141
Test: Logcat works again in emulator 29.2.10
Change-Id: Ia6e2430f1d7a0476a5b438d315a9ea337c13806d
Currently, any netlink error reported in ifc_act_on_address is
reported as EINVAL, because the receive buffer is too short and
the NLMSG_OK check fails on it.
Make the buffer long enough so the function can return the
correct netlink error to the caller.
Test: new test coverage in other CL in this topic
Change-Id: Ica752db834a0fd614312b800d88721826be08d56
The test requires to load and verify vbmeta struct from partitions,
and thus needs adb root.
Bug: None
Test: atest libfs_avb_device_test
Change-Id: I4924d6e41edc78898d9ef9c3d7f52c9066f750b1
Currently, passing an invalid IP address to interfaceAddAddress
and friends results in an invalid error message:
12-02 13:00:42.675 interfaceAddAddress("testtap2355", "2001:db8::1/64", 64) -> ServiceSpecificException(-8, "InterfaceController error: Unknown error -8") <1.63ms>
This is due to confusion between getaddrinfo error codes and
errno values.
Test: new unit test added to IpClientIntegrationTest
Change-Id: Ifdaa4281a9bcf3998e3216472c5c1df0f5285214
Try to not allocate as many blocks on the heap, and reuse
memory instead of copying it
Get rid of unique_ptr and shared_ptr where possible, move
the Block objects themselves
Overall this reduces the time spent in memcpy() from 30% to
15% of the whole 'adb push' command, and gets rid of about 5%
of the time spent in the malloc/free calls
Test: builds
Change-Id: I8995115274b6f08a4df13c58183c928ef384a767
