OSDN Git Service
Daniel Rosenberg [Fri, 24 Jan 2020 02:09:55 +0000 (02:09 +0000)]
Merge changes from topic "casework"
* changes:
ext4.h: add EXT4_FEATURE_INCOMPAT_CASEFOLD
Allow casefolding and project quotas for f2fs
Yabin Cui [Thu, 23 Jan 2020 22:45:17 +0000 (22:45 +0000)]
Merge "simpleperf: fix according to unwindstack change."
Yabin Cui [Thu, 23 Jan 2020 20:32:26 +0000 (12:32 -0800)]
simpleperf: fix according to unwindstack change.
unwindstack::Maps uses prev_real_map instead of prev_map, which
breaks simpleperf code building up prev_map links. So switch to
use unwindstack::Maps::Sort() to build up prev_real_map links.
Also add more tests.
Bug:
148075852
Test: run simpleperf_unit_test.
Change-Id: Ica0551e2d5a84aa8a0671489821fed1d263324f4
Yabin Cui [Thu, 23 Jan 2020 18:16:25 +0000 (18:16 +0000)]
Merge "simpleperf: add --per-thread option to stat cmd."
Christopher Ferris [Thu, 23 Jan 2020 17:45:26 +0000 (17:45 +0000)]
Merge "Update for chnaged MapInfo constructor."
Christopher Ferris [Thu, 23 Jan 2020 01:22:49 +0000 (17:22 -0800)]
Update for chnaged MapInfo constructor.
Bug:
148075852
Test: Ran unit tests.
Change-Id: Ia5032ce0777437f36643a1defc82c5187ad2fa25
Oliver Nguyen [Wed, 22 Jan 2020 22:34:36 +0000 (22:34 +0000)]
Merge "Add cross-process coverage flushing for the LLVM toolchain."
Yabin Cui [Fri, 10 Jan 2020 23:33:11 +0000 (15:33 -0800)]
simpleperf: add --per-thread option to stat cmd.
It shows event count for each thread.
Bug:
146220391
Test: run simpleperf_unit_test.
Change-Id: I6346b787eb630c49631e477d0e83bf1901a9f2b3
Treehugger Robot [Fri, 17 Jan 2020 19:31:10 +0000 (19:31 +0000)]
Merge "Add warning for slow file write"
Wei Wang [Fri, 17 Jan 2020 02:26:12 +0000 (18:26 -0800)]
Add warning for slow file write
Bug:
147833210
Test: Build
Change-Id: I07a7220016c29ba290b491fbc86603f06c4b2179
Yabin Cui [Thu, 16 Jan 2020 01:00:16 +0000 (01:00 +0000)]
Merge "simpleperf: add doc to show annotated source code and disassembly."
Yabin Cui [Thu, 16 Jan 2020 00:59:55 +0000 (00:59 +0000)]
Merge "simpleperf: remove code handling cpu hotplug."
Treehugger Robot [Wed, 15 Jan 2020 19:39:21 +0000 (19:39 +0000)]
Merge "CtsKernelConfigTestCases secondary_user"
Yabin Cui [Wed, 15 Jan 2020 18:39:12 +0000 (18:39 +0000)]
Merge "simpleperf: prepare for selinux change."
Treehugger Robot [Wed, 15 Jan 2020 15:24:14 +0000 (15:24 +0000)]
Merge "Fix help string for f2fs"
Treehugger Robot [Wed, 15 Jan 2020 08:10:41 +0000 (08:10 +0000)]
Merge "fec: Add liblog as a dependency where libbase is used"
Yabin Cui [Wed, 15 Jan 2020 00:43:43 +0000 (16:43 -0800)]
simpleperf: remove code handling cpu hotplug.
cpu hotplug code was disabled in record cmd long time ago, and
becomes obsolete as pixel devices seem no long use cpu hotplug
for thermal control.
Bug: none
Test: run simpleperf_unit_test.
Change-Id: I75fc522792e33afd16dd805ee0c9946ad4a6c67a
Yabin Cui [Mon, 13 Jan 2020 22:48:01 +0000 (14:48 -0800)]
simpleperf: prepare for selinux change.
To work with controlling perf_events via selinux:
1. Check security.perf_harden property at the entry of simpleperf.
2. Record tracepoint raw data only when running as root.
Bug:
139860059
Test: run simpleperf_unit_test.
Change-Id: I40082de0010949e54fb707f385220157a71f266b
Tom Cherry [Tue, 14 Jan 2020 22:05:42 +0000 (14:05 -0800)]
fec: Add liblog as a dependency where libbase is used
Bug:
119867234
Test: build
Change-Id: I49d2110a5469c4c58bf2587f8d7e4a5b8981caac
Yi Kong [Tue, 14 Jan 2020 10:08:39 +0000 (10:08 +0000)]
Merge "Add cache for --binary filter in inject cmd."
Yabin Cui [Tue, 14 Jan 2020 01:09:03 +0000 (01:09 +0000)]
Merge "simpleperf: remove inplace-sampler."
Yabin Cui [Tue, 14 Jan 2020 00:56:51 +0000 (16:56 -0800)]
simpleperf: add doc to show annotated source code and disassembly.
Bug:
132371337
Test: none
Change-Id: I09b8855af7d8609e7c16c8120cacda46ce83de31
jovanak [Tue, 14 Jan 2020 00:32:28 +0000 (16:32 -0800)]
CtsKernelConfigTestCases secondary_user
Marks this CTS test module as NOT supporting secondary_user, meaning that
the tests are not eligible to be run from a regular Android user other
than the system user (i.e. other than user 0).
The reason for lack of elligibility is that the module restarts before running
the tests - the device ends up in user 0 by default upon reboot.
Tradefed infrastructure does not support restarting into different users at
the moment.
Bug:
141773058
Test: module passed with secondary_user params
(--enable-optional-parameterization --module-parameter secondary_user)
Change-Id: I39fca13fe40b497cf7cb7a014d11f96e8174c953
Yi Kong [Mon, 13 Jan 2020 23:43:01 +0000 (15:43 -0800)]
Add cache for --binary filter in inject cmd.
Regex matching is slow. Cache the DSO to vastly improve performance when
the --binary filter is used.
Test: simpleperf_unit_test
Bug:
79161490
Change-Id: Idc7a8a588c8291ce59fab89ab0278aa0b3fe07f0
Yabin Cui [Mon, 13 Jan 2020 23:25:16 +0000 (15:25 -0800)]
simpleperf: remove inplace-sampler.
Remove obsolete code for inplace-sampler, which is never fully implemented.
Bug: none
Test: run simpleperf_unit_test.
Change-Id: I458c23e2d2331ec96c3c9d8462309eef544564b9
Oliver Nguyen [Tue, 10 Dec 2019 00:36:08 +0000 (16:36 -0800)]
Add cross-process coverage flushing for the LLVM toolchain.
Bug:
143977934
Test: Build with native coverage
Test: kill -37 <pid> and check for coverage in /data/misc/trace
Change-Id: Iadf93e62d92dc35cdaa35162ca4a75d26190db85
Treehugger Robot [Sat, 11 Jan 2020 01:26:05 +0000 (01:26 +0000)]
Merge "simpleperf: dump all process maps for etm system wide recording."
Treehugger Robot [Sat, 11 Jan 2020 00:05:36 +0000 (00:05 +0000)]
Merge "Dump coverage for all objects (a.out + dlopen()'d .so's)"
Yabin Cui [Sat, 11 Jan 2020 00:00:11 +0000 (16:00 -0800)]
simpleperf: dump all process maps for etm system wide recording.
If don't dump all process maps, we may miss mapping info needed
to decode etm data.
Bug:
79161490
Test: run simpleperf manually.
Test: run simpleperf_unit_test.
Change-Id: Ic98d1b461b2c11a51147c67e8f269d5be10e962b
Ray Essick [Thu, 9 Jan 2020 19:51:27 +0000 (11:51 -0800)]
Dump coverage for all objects (a.out + dlopen()'d .so's)
with code coverage enabled, the main object as well as each dlopen()ed
shared library include their own copies of the hooks for dumping gcov
data upon receiving signal GCOV_FLUSH_SIGNAL. Only the last one to
register was actually dumped.
Here, we chain the signal handlers together so that each one will call
the previously registered handler to ensure that each object's coverage
data is dumped.
Bug:
139313557
Test: dump gcov data via kill -37 on module with dlopen()
Change-Id: I376df2df6b65acff95ad5d135a7ba2bf0eac3695
Tom Cherry [Fri, 10 Jan 2020 16:56:33 +0000 (16:56 +0000)]
Merge "Add liblog as a dependency where libbase is used"
Treehugger Robot [Fri, 10 Jan 2020 07:19:46 +0000 (07:19 +0000)]
Merge "Allow fec to be compiled with BUILD_HOST_static."
Treehugger Robot [Fri, 10 Jan 2020 03:45:39 +0000 (03:45 +0000)]
Merge "simpleperf: Omit dwarf-callgraph tests running on ARM binary translation"
Tom Cherry [Thu, 9 Jan 2020 20:27:44 +0000 (12:27 -0800)]
Add liblog as a dependency where libbase is used
Bug:
119867234
Test: build
Change-Id: I2728559bbd3cd4f867b30baf53a62124b8abd0b6
Dario Freni [Thu, 9 Jan 2020 19:47:34 +0000 (19:47 +0000)]
Allow fec to be compiled with BUILD_HOST_static.
Bug:
144477678
Test: m fec; BUILD_HOST_static=1 m fec
Change-Id: Ic9d9fb65604e56e89a4c6ce75eb583beb34efe63
Daniel Rosenberg [Thu, 9 Jan 2020 00:10:15 +0000 (16:10 -0800)]
ext4.h: add EXT4_FEATURE_INCOMPAT_CASEFOLD
fs_mgr needs to know the value of the casefold feature flag
so it can enable it.
Bug:
138322712
Test: Compiles
Change-Id: I6acca568622174a4b15a6bfa101fcc676c12dca7
Daniel Rosenberg [Tue, 17 Dec 2019 22:46:25 +0000 (14:46 -0800)]
Allow casefolding and project quotas for f2fs
Bug:
138321217
Bug:
138322712
Test: Enable on device. Test the /data has options enabled via
tune2fs/f2fs.fsck
Change-Id: I02abdf97984c73d95632b17e866de5755f38b304
Daniel Rosenberg [Wed, 18 Dec 2019 02:35:51 +0000 (18:35 -0800)]
Fix help string for f2fs
Help showed the operands in the wrong order. The script does not appear
to allow for there to be in different order, so fixing the help message
Test: None (could confirm necessary by invoking with flags in other
order and seeing that it doesn't see the flags)
Change-Id: Ia0a5a41f4743caf03e39df7be53f717a2642b764
Nikita Ioffe [Wed, 8 Jan 2020 19:25:41 +0000 (19:25 +0000)]
Merge "Add utility to calculate digest from the root of verity tree"
Dmytro Chystiakov [Tue, 7 Jan 2020 22:52:44 +0000 (14:52 -0800)]
simpleperf: Omit dwarf-callgraph tests running on ARM binary
translation
Bug: b/
144472912
Test: Run ARM CtsSimpleperfTestCases module
Change-Id: Ia8eef58d6408532943baf5d848a77644a0adcf04
Signed-off-by: Dmytro Chystiakov <dmytro.chystiakov@intel.com>
Yi Kong [Tue, 7 Jan 2020 01:50:38 +0000 (01:50 +0000)]
Merge "simpleperf: Accept --binary option as regex in inject cmd"
Yi Kong [Mon, 6 Jan 2020 23:22:55 +0000 (15:22 -0800)]
simpleperf: Accept --binary option as regex in inject cmd
This would allow us to more efficiently generate AudoFDO profiles for
multiple binaries.
Test: simpleperf_unit_test
Bug:
79161490
Change-Id: I45b444b8cffd8eb72e4079dcf66a176da0da5b6d
Yabin Cui [Mon, 6 Jan 2020 22:57:42 +0000 (22:57 +0000)]
Merge "simpleperf: update simpleperf prebuilts to build
6087948."
Yabin Cui [Sat, 21 Dec 2019 00:20:03 +0000 (16:20 -0800)]
simpleperf: update simpleperf prebuilts to build
6087948.
Taken from branch aosp-simpleperf-release.
Bug: none
Test: run test.py.
Change-Id: I5d23c151790bb71c12d8e5a81e3038af1109522b
Nikita Ioffe [Sun, 5 Jan 2020 04:25:47 +0000 (04:25 +0000)]
Add utility to calculate digest from the root of verity tree
This function will help verifying that external hashtree wasn't
tampered. See another change in this topic for more context.
Test: atest build_verity_tree_test
Bug:
145670581
Change-Id: I10a8141a60eec142c77cde0048f0abf44b3406bb
Treehugger Robot [Thu, 2 Jan 2020 18:33:43 +0000 (18:33 +0000)]
Merge "Extend bootctl with IBootControl1.1 API"
Alessio Balsini [Thu, 19 Dec 2019 13:43:36 +0000 (13:43 +0000)]
Extend bootctl with IBootControl1.1 API
Extend bootctl to use the new API methods introduced in IBootControl1.1:
setSnapshotMergeStatus() and getSnapshotMergeStatus().
Test: adb shell
Bug: b/
144783382
Change-Id: Id7acc927561aa17b37be65fa54faa55702d73958
Signed-off-by: Alessio Balsini <balsini@google.com>
David Anderson [Fri, 27 Dec 2019 20:28:48 +0000 (20:28 +0000)]
Merge "lpdump: Add support for the DISABLED flag."
Treehugger Robot [Sat, 21 Dec 2019 01:39:34 +0000 (01:39 +0000)]
Merge "simpleperf: clean up cmd_list.cpp."
Yabin Cui [Fri, 20 Dec 2019 20:33:09 +0000 (12:33 -0800)]
simpleperf: clean up cmd_list.cpp.
1. Change return type of IsEventTypeSupported.
2. Use callback function to check if an event belongs to a type.
Bug: none
Test: run simpleperf_unit_test.
Change-Id: I2267d21c2e9501cb534a6520abb20d563de7e545
Tianjie Xu [Fri, 20 Dec 2019 21:44:11 +0000 (21:44 +0000)]
Merge "Don't padd the digest during hash computation"
Treehugger Robot [Fri, 20 Dec 2019 20:51:51 +0000 (20:51 +0000)]
Merge "simpleperf: fix pmu event tests."
Yabin Cui [Fri, 20 Dec 2019 19:01:58 +0000 (11:01 -0800)]
simpleperf: fix pmu event tests.
Bug: none
Test: run simpleperf_unit_test in shell and root.
Change-Id: I9745b9cdaf4ad2c0c0d867442d53e94d68c33806
Treehugger Robot [Fri, 20 Dec 2019 17:57:49 +0000 (17:57 +0000)]
Merge "simpleperf: Suport "pmu" type events in list command"
Treehugger Robot [Fri, 20 Dec 2019 17:57:36 +0000 (17:57 +0000)]
Merge "simpleperf: Use PMU's cpumask if available"
Treehugger Robot [Fri, 20 Dec 2019 17:57:27 +0000 (17:57 +0000)]
Merge changes I626b8146,Ibdf1180e
* changes:
simpleperf: move GetCpusFromString() to utils.cpp
simpleperf: Support event_source bus event types
Namhyung Kim [Tue, 17 Dec 2019 14:26:34 +0000 (23:26 +0900)]
simpleperf: Suport "pmu" type events in list command
Support low-level PMU counters to be listed using "pmu" argument.
Test: run simpleperf_unit_test
Change-Id: I7bc8251a4fa12bae82142b71afa70c359ad24d40
Signed-off-by: Namhyung Kim <namhyung@google.com>
Namhyung Kim [Tue, 3 Dec 2019 02:57:18 +0000 (11:57 +0900)]
simpleperf: Use PMU's cpumask if available
Some PMU events only support profiling from dedicated cpus. For those
events there are cpumask files in the /sys/bus/event_source/devices/<PMU>
directory. So override the cpu setting for them.
Test: manually tested with a PMU has the cpumask file
Change-Id: I5e50a7d6f88cb0bc80b45e774cf048116698b3de
Signed-off-by: Namhyung Kim <namhyung@google.com>
Namhyung Kim [Tue, 17 Dec 2019 08:15:02 +0000 (17:15 +0900)]
simpleperf: move GetCpusFromString() to utils.cpp
Move the GetCpusFromString() to utils.cpp so that it can be used by
event parsing code in event_type.cpp.
No functional change intended.
Test: build test only
Change-Id: I626b8146f90bf8bb161ce8574bcf70f6600c8ad3
Namhyung Kim [Wed, 23 Oct 2019 03:37:34 +0000 (12:37 +0900)]
simpleperf: Support event_source bus event types
The perf_event subsystem exports available PMUs and their events under
/sys/bus/event_source/devices directory. This patch supports those
events using 'pmu/event/' format like in the mainline perf tool.
The "event" part above might have multiple terms if supported by PMU.
In that case PMU would have "format" directory for available terms.
They would be seperated by comma, so the enclosing "/"s are needed to
group those terms for an event.
But currently all PMUs seem to support a single term only so I left
the parser to split events by comma and match the whole string
(including "/"s).
For example you can use cpu-cycles events like below:
# simpleperf stat -a -e armv8_pmuv3/cpu_cycles/ sleep 1
Performance counter statistics:
986,481,766 armv8_pmuv3/cpu_cycles/ # (100%)
Total test time: 1.007272 seconds.
Test: build and run simpleperf_unit_test
Change-Id: Ibdf1180e58d190f8108cdc8d8a096e6323b2a6f3
Tianjie Xu [Fri, 20 Dec 2019 02:59:24 +0000 (18:59 -0800)]
Don't padd the digest during hash computation
There's no need to do that. Also check_block_hash() only allocates a
buffer of digest_length_.
Bug:
144388532
Test: unit tests pass
Change-Id: Iec3209d260b51cb41905cc84ae86572bcb1cc130
Yabin Cui [Thu, 19 Dec 2019 22:34:40 +0000 (22:34 +0000)]
Merge "simpleperf: move to llvm binutils."
Yabin Cui [Thu, 19 Dec 2019 22:22:28 +0000 (14:22 -0800)]
simpleperf: move to llvm binutils.
Bug: none
Test: run test.py TestTools*.
Change-Id: I9fcbb559ea4cb52e1ac79f3a7aa801a1d7e94dce
Tianjie Xu [Thu, 19 Dec 2019 20:42:03 +0000 (20:42 +0000)]
Merge changes Ie4aef1bd,I1bfc9323
* changes:
Add sha1 support for hashtree
Add test to parse verity data
David Anderson [Wed, 18 Dec 2019 23:11:45 +0000 (15:11 -0800)]
lpdump: Add support for the DISABLED flag.
Bug:
134949511
Test: manual test
Change-Id: I53d35167bc27d3d8d2516e890697514d64635ad7
Tianjie Xu [Mon, 2 Dec 2019 20:55:33 +0000 (12:55 -0800)]
Add sha1 support for hashtree
AVB format uses sha1 as the hash algorithm when computing the hash of
each block. This cl adds the support to use either sha1 or sha256 in
the hashtree_info struct. It also moves some helper functions as the member
function of the struct.
Bug:
144388532
Test: unit tests pass
Change-Id: Ie4aef1bdd8fa89240f7c6f83fcb2e33f47a468cb
Tianjie Xu [Thu, 21 Nov 2019 18:25:17 +0000 (10:25 -0800)]
Add test to parse verity data
Add a unittest to load and verify a VB1.0 image.
Bug:
144388532
Test: unittest pass
Change-Id: I1bfc93239a6696c58e9220864009759a2468e136
David Anderson [Wed, 18 Dec 2019 15:42:44 +0000 (15:42 +0000)]
Merge "partition_tools: Add support for the VIRTUAL_AB_DEVICE flag."
Eric Biggers [Tue, 17 Dec 2019 17:35:39 +0000 (17:35 +0000)]
Merge "libfscrypt: Use <linux/fscrypt.h> from Bionic"
Treehugger Robot [Tue, 17 Dec 2019 17:34:27 +0000 (17:34 +0000)]
Merge "lpdump: Use unsigned value for shift"
Greg Kaiser [Tue, 17 Dec 2019 13:43:47 +0000 (05:43 -0800)]
lpdump: Use unsigned value for shift
Shifting a signed 32-bit value by 31 bits is implementation-defined
behavior. While our toolchains are most probably doing the
expected thing, we make this an unsigned value out of paranoia.
Test: TreeHugger
Change-Id: I551fcf1ac7a27ebd7c1365cd6a9df7267f569c44
Eric Biggers [Tue, 17 Dec 2019 00:02:23 +0000 (16:02 -0800)]
libfscrypt: Use <linux/fscrypt.h> from Bionic
aosp/
1184798 has updated the kernel headers to android-mainline, so we
no longer need to manually declare the declarations for v2 policies, nor
do we need to manually declare FSCRYPT_POLICY_FLAG_IV_INO_LBLK_64.
Also replace the FS_* constants with their new FSCRYPT_* names. This
doesn't change the numerical values; it just changes the names.
Test: build and 'atest libfscrypt_unit_test'
Bug: None
Change-Id: I03ce177923bfa9e0fecbbdbf1718fbf1c17176d9
David Anderson [Mon, 16 Dec 2019 20:31:06 +0000 (20:31 +0000)]
Merge "lpmake: Use an enum class for option codes."
David Anderson [Mon, 16 Dec 2019 17:31:57 +0000 (17:31 +0000)]
Merge "lpdump: Add support for the metadata flags field."
David Anderson [Fri, 13 Dec 2019 23:34:54 +0000 (15:34 -0800)]
partition_tools: Add support for the VIRTUAL_AB_DEVICE flag.
Bug:
134949511
Test: manual test
Change-Id: If31234d9da0b15a3dbde7dcc835b2161c5dbba61
David Anderson [Fri, 13 Dec 2019 23:31:10 +0000 (15:31 -0800)]
lpmake: Use an enum class for option codes.
There's no reason to have short-form option codes for everything, this
will make that more clear.
Bug:
134949511
Test: super.img builds
Change-Id: If60f6648ac311b77d68768bb39825d01c663c500
David Anderson [Fri, 13 Dec 2019 01:12:27 +0000 (17:12 -0800)]
lpdump: Add support for the metadata flags field.
Bug:
134949511
Test: manual test
Change-Id: Iafc5193c71315b14f192fcc0ae42188a51b0d2cb
Yabin Cui [Thu, 12 Dec 2019 05:25:14 +0000 (05:25 +0000)]
Merge "simpleperf: fix reading jit debug descriptors."
Yabin Cui [Thu, 12 Dec 2019 00:50:24 +0000 (16:50 -0800)]
simpleperf: fix reading jit debug descriptors.
Clang may generate elf files with segments not starting from
page-aligned addresses. For these segments, dynamic linker
creates maps mapping to (segment.p_offset & PAGE_MASK). So
map.start is mapped to file vaddr (segment.p_vaddr & PAGE_MASK)
instead of segment.p_vaddr.
So add PAGE_MASK when calculating memory address of descriptors,
using below formula:
descriptor.vaddr_in_file - (segment.p_vaddr & PAGE_MASK) + map.start.
Bug:
145920143
Bug:
145930241
Test: run simpleperf_unit_test.
Change-Id: Id185a6d6b766ecaa30b5067664687e556f92ef27
Christopher Ferris [Tue, 10 Dec 2019 22:50:18 +0000 (22:50 +0000)]
Merge "Add no decay variants of benchmarks."
Christopher Ferris [Tue, 10 Dec 2019 21:19:43 +0000 (13:19 -0800)]
Add no decay variants of benchmarks.
This will help to tell what the performance impact is of turning
off the decay timer.
Test: Ran benchmarks and verified that times are slower for no decay.
Change-Id: Ia0d5aa879e92fb0790194c2efada6c2e407427a9
Tianjie Xu [Wed, 4 Dec 2019 23:07:23 +0000 (23:07 +0000)]
Merge "Factor out the hashtree info from the verity_info"
Alistair Delva [Mon, 2 Dec 2019 22:06:45 +0000 (14:06 -0800)]
Merge "mkf2fsuserimg.sh: Make sparse mode optional"
am:
8b724d9075
Change-Id: Ic040444e05471954247abd688054c79544ac427c
Alistair Delva [Mon, 2 Dec 2019 21:54:20 +0000 (21:54 +0000)]
Merge "mkf2fsuserimg.sh: Make sparse mode optional"
Tianjie Xu [Wed, 20 Nov 2019 20:18:40 +0000 (12:18 -0800)]
Factor out the hashtree info from the verity_info
Even though the format for the VB2.0(AVB) image is different from the VB1.0
image, the majority of the content is the same. In specific, they both
have a hashtree and ecc data. By factoring the hashtree info, fec_read
can later use only the hashtree for error correction and doesn't need to
have the knowledge of verity format.
VB 1.0:
Filesystem
hash_tree
metadata <- assume 8 blocks
ecc_data
VB 2.0:
Filesystem
hash_tree
ecc_data
vbmeta
avb footer
Also switch the internal struct to use std containers so that we don't
need to worry about the memory management.
Bug:
144388532
Test: load a verity image, unit tests pass
Change-Id: I83c55464c3ac75ac41be9f507b674886ad298350
David Anderson [Mon, 25 Nov 2019 20:24:12 +0000 (12:24 -0800)]
Merge "partition_tools: Add an "lpadd" tool for appending images to super.img."
am:
cf9b34eccd
Change-Id: I4fc759fba2d489aee3bdf2870434d904f614e1e8
David Anderson [Mon, 25 Nov 2019 20:17:19 +0000 (20:17 +0000)]
Merge "partition_tools: Add an "lpadd" tool for appending images to super.img."
Yabin Cui [Fri, 22 Nov 2019 20:23:41 +0000 (12:23 -0800)]
Merge "simpleperf: fix flaky test TestApiProfiler."
am:
184821b4f0
Change-Id: Id953ebb3cfbebeaff09e837c1c084565820ad0fb
Yabin Cui [Fri, 22 Nov 2019 20:23:12 +0000 (12:23 -0800)]
Merge "simpleperf: update simpleperf prebuilts to build
6021660."
am:
b17755fb7d
Change-Id: Ib5f29fe15f21a8dadde3da69d2e6a2c680e6f151
Yabin Cui [Fri, 22 Nov 2019 20:14:19 +0000 (20:14 +0000)]
Merge "simpleperf: fix flaky test TestApiProfiler."
Yabin Cui [Fri, 22 Nov 2019 20:13:41 +0000 (20:13 +0000)]
Merge "simpleperf: update simpleperf prebuilts to build
6021660."
Yabin Cui [Fri, 22 Nov 2019 00:06:24 +0000 (16:06 -0800)]
simpleperf: fix flaky test TestApiProfiler.
Bug: none
Test: run test.py TestApiProfiler* --repeat 10.
Change-Id: I5829aa9b2be8057b64130a6fb4b076958456c7ca
Yabin Cui [Wed, 20 Nov 2019 23:26:06 +0000 (15:26 -0800)]
simpleperf: update simpleperf prebuilts to build
6021660.
Taken from branch aosp-simpleperf-release.
Bug: none
Test: run test.py on devices with different android versions.
Change-Id: I39c33e6f7c1d3776288db78554feafea0bff959f
David Anderson [Thu, 21 Nov 2019 21:48:53 +0000 (13:48 -0800)]
Merge "lpdumpd: Do not use stack to store arguments."
am:
46038c2002
Change-Id: If7616927d6c970851458139906e0cbe9f9784dbd
David Anderson [Thu, 21 Nov 2019 21:39:28 +0000 (21:39 +0000)]
Merge "lpdumpd: Do not use stack to store arguments."
David Anderson [Thu, 21 Nov 2019 20:32:23 +0000 (12:32 -0800)]
Merge "lpdump: When using -h, print usage to stdout so lpdumpd prints it."
am:
2e61a7758c
Change-Id: I9698754f5f0379db8310fb7d2ad2fca6d07e83b7
David Anderson [Thu, 21 Nov 2019 20:18:05 +0000 (20:18 +0000)]
Merge "lpdump: When using -h, print usage to stdout so lpdumpd prints it."
Cody Caldwell [Wed, 20 Nov 2019 20:56:16 +0000 (12:56 -0800)]
Merge "Revert "lpmake: Don't assume a device is retrofit if it has one block device for super.""
am:
cafb6e9140
Change-Id: Iae536a2eaef715a899bd97ccf1645541a520376f
Cody Caldwell [Wed, 20 Nov 2019 20:40:30 +0000 (20:40 +0000)]
Merge "Revert "lpmake: Don't assume a device is retrofit if it has one block device for super.""
Cody Caldwell [Wed, 20 Nov 2019 20:36:12 +0000 (20:36 +0000)]
Revert "lpmake: Don't assume a device is retrofit if it has one block device for super."
This reverts commit
13945230d36282a0dd2ad7c240f2b88ffc506205.
Reason for revert: Potential culprit for Bug b/
144809373 - verifying through Forrest before revert submission
Change-Id: Ic4ed750a5386f6f45a871ac7112142a6991dfd2c