Provide alternative versions that do not force callers to create
std::string objects. This patch has the intended side-effect that all
callers that pass a {string} initializer list to the 'profiles' argument
now call an std::initializer_list<> overload instead of the const
std::vector<std::string>& overload.
Additionally, add std::function<> arguments instead of calling
ExecuteForProcess() or ExecuteForTask() directly to make it easier to
write unit tests for SetTaskProfiles() and SetProcessProfiles().
Bug: 213617178
Change-Id: Ica61e944a66a17178ee43a113b8ca082f7eb834b
Signed-off-by: Bart Van Assche <bvanassche@google.com>
Do not force callers to create an std::string object. This patch
implements the following advice from go/totw/1: "Google’s preferred
option for accepting such string parameters is through a string_view."
Use std::less<> as comparison type to prevent that std::string_view
objects have to be converted into std::string objects for lookups in
std::map<>.
Bug: 213617178
Change-Id: I08125a02220a8c003d9202a7e177be776c3b9829
Signed-off-by: Bart Van Assche <bvanassche@google.com>
This change enables headers like <span>. Inside the <span> header file
file the following guard makes its functionality unavailable when
building with std=gnu++-17:
#if _LIBCPP_STD_VER > 17
[ ... ]
#endif
Bug: 213617178
Change-Id: I5c40708ea196ab112990b5ca6fae9370b75f8752
Signed-off-by: Bart Van Assche <bvanassche@google.com>
Without the directory (this happens on the very first boot),
load_persist_props can't create an initial version of
/data/property/persistent_properties (probably empty). This leads to
persisting all in-memory "persist.*" properties later when a persistent
property is set. This is regression from Android S because persistent
props from, for example, build.prop will be persisted even when there's
no process to explicitly setprop.
Bug: 242264580
Test: launch cuttlefish and verify that there's no props from build.prop
Change-Id: I5819a97750e4d5d1ee5a7c308bf944c7aeab2f90
-h: print help
-v: noisy output
-R: automatic reboot if needed
Also remove the "argc != 0" check, as this shouldn't happen most of the
time anyway. The only possible way (I can think of) for this to happen
is to explicitly call execve() with an empty argv, which is against the
calling convention, thus a footgun, to begin with.
Bug: 241688845
Test: adb shell [disable-verity|enable-verity] [-R] [-v] [-h]
Test: adb shell set-verity-state [-R] [-v] [-h] [0|1]
Change-Id: I27fc18b0958650c5926322a3f4bd1df758fd96c8
The option was added for debugging purposes in case the "test argv[0] ==
clean_scratch_files" method breaks.
Now it has no user at all and no reason to maintain it, so just remove it.
Also don't use MyLogger() when running as clean_scratch_files, as
clean_scratch_files don't have stdout/stderr.
Bug: 241179247
Test: Presubmit
Change-Id: I2d8069f59fe6b85fc84ab07bb2df6efb39d6ecaa
Reorder COW ops vector based on merge sequence. We don't
need additional vector to be stored in memory.
Memory usage for a full OTA on Pixel:
Without Patch:
RssAnon: 61020 kB
With Patch:
RssAnon: 51112 kB
Bug: 237490659
Test: OTA on Pixel
Signed-off-by: Akilesh Kailash <akailash@google.com>
Change-Id: I543dd73acfa7cf4e57379e82bc184e943072e7c8
This test has always been flaky, and is not testing something super
valuable: we know that image creation succeeds throughout the rest of
the suite, so it's not very interesting to know that it can succeed in a
low-space scenario.
The inverse test is much more valuable, since we want the correct status
code when creation fails due to low space.
Bug: 240391002
Test: vts_libsnapshot_test
Change-Id: I6235d11033d2f30efe530077b877863ba2574810
Diagnosing DM_DEV_REMOVE failures in the test harness is quite
difficult, and it's not clear if failures are spurious or not. Instead
use SnapshotManager's helper function, which can retry on failure, and
will self-diagnose issues on legitimate failures.
Bug: N/A
Test: vts_libsnapshot_test
Change-Id: Ibcaa8406e8b1e8758b99a8e9b58c58d68ed57685
These checks have historically been unreliable, and we make no
guarantees around dev_t with overlayfs.
Bug: 242240650
Test: adb-remount-test.sh
Change-Id: I19e7aabec424a22beb0b56d35b198906841178b0
Refactor set_avb_verity_state() to return tri-state result:
{.success=false}
Failed to set verity state.
{.success=true, .want_reboot=false}
Success but verity already is requested state, so no need to reboot.
{.success=true, .want_reboot=true}
Successfully changed verity state, need reboot to take effect.
Setup overlayfs only if verity is going to be or already disabled.
Teardown overlayfs regardless of enable-verity success of not.
This ensures more robust behavior of setup / teardown overlayfs.
Adjust the log messages for consistent wording.
No point logging the errno of fs_mgr_overlayfs_[setup|teardown]
because the error must already be logged by callee.
Bug: 241688845
Test: adb [disable|enable]-verity
Change-Id: I3a77fe567757ca6173e8c3142e77fef483d9b849
aosp/I212bdb0e97016dec50618962d7c24f46d35764c7 changes
`remount` to return MUST_REBOOT after initial overlay setup.
This causes DsuGsiIntegrationTest to fail because the non-zero exit code
cause the test script to think the `remount` command failed.
Change it so that we return 0 instead, to indicate that we
"successfully" setup a new overlay.
We should only return non-zero on unrecoverable error, like when we
failed to disable verity, failed to setup overlay, failed to perform
MS_REMOUNT.
Bug: 241179247
Test: DsuGsiIntegrationTest
Change-Id: I280ffa988118c59e366cdd5bd1479bb43896c278
In commit 5d72d6c, an unused code path for e2fsdroid and sload_f2fs was
removed from fastboot. Remove them from the build dependency list.
Bug: 237960487
Test: m fastboot
Change-Id: Ibee72a86eeb43fcde3a032094bfc1c19de09bc03
* changes:
remount: Remove the "backing" parameter to fs_mgr_overlayfs_setup.
remount: Prevent error spam when remounting fails.
remount: Move more stuff out of do_remount.
