This helps to generate a deterministic package.
Bug: 71770360
Test: Generate a incremental package and transfers are added by file name.
Change-Id: I7562a200b97a1babbb09a77801324cc9408cc01f
In non-A/B OTA path, we've been loading the info dict for the target
build twice (once to have a peek at the OTA type, with a reload after
unzipping the target-files zip). Remove the reloading.
This CL moves the loading of the source info dict up a bit to avoid
having two copies in the A/B and non-A/B paths. It also moves up the
sanity checks of the loaded info dicts before unzipping the target-files
zips.
Test: Generate full and incremental packages for bullhead and marlin.
Change-Id: Iccb953feb0b6ecf62e6f73d6e9ceea00077d098d
GetBuildProp() and GetVendorBuildProp() have been factored into
BuildInfo class. These two functions were meant to be deleted in commit
481bab8d48.
Test: python -m unittest test_ota_from_target_files
Test: No users of these two functions.
Change-Id: Id590a46d597939973ef2339cdfea4f5163548d4d
The fallback-to-full logic doesn't look beneficial, but may hide away
real bugs, especially when it's on by default.
Test: Generate an incremental non-A/B OTA successfully.
Test: Inject bugs into script; observe that incremental OTA
generation errors out instead of silently falling back.
Test: No internal users of the flag from code search.
Change-Id: Ifc627d898552c0bd72463e8c107e36610421608e
We split large apks and generated patches for them in parallel,
resulting in nondeterminate packages between different runs. This CL
sort the split transfers by target name first; and then add them
sequentially to the final transfer list.
Also fix a side effect where we may generate a wrong sha1 for split
ranges due to synchronization error.
Bug: 71770360
Bug: 71759418
Test: Generate the package several times, compare the log and the transfer list.
Change-Id: I2a49e22594d59ffaa98b11edc776be4e3c4c561f
Prior to this CL, we have similar codes in different paths
(WriteFullOTAPackage / WriteBlockIncrementalOTAPackage /
WriteABOTAPackageWithBrilloScript). This CL factors out the common codes
that deal with OEM-specific properties into BuildInfo class, and adds
tests for the new class.
Test: python -m unittest test_ota_from_target_files
Test: Create an incremental package that uses and doesn't use OEM
properties respectively.
Test: Create a full package that uses and doesn't use OEM properties
respectively.
Change-Id: Ida914cace12803d55396fa503bafcac2db2a520e
Test: python -m unittest test_common
Test: Run sign_target_files_apks.py on a target with compressed APKs.
Change-Id: I107a8b8f2f0f82e2d1947f14c8a8b3778f633b11
* changes:
Clear resource_export_package outside conditionals
Fix adbs after move from build/tools to build/make/tools
Use proguard.jacoco.flags for EMMA_INSTRUMENT_STATIC
adbs broke when build/tools was moved to build/make/tools in
aosp/286781.
Bug: 15270043
Test: build/tools/adbs
Change-Id: I0a34b2b446f996bcc53a7dde6381402f8eed9bd0
Mostly cosmetic changes, such as moving print statement to print
function. The only functional change is to wrap the file opening in
Append() with try statement, which would dump the error message as
needed.
Test: `m dist`
Test: python -m unittest test_build_image
Change-Id: I4e6c593517cf737bfcbe51bc533b22a247b10e44
This simplifies the work to track mkdtemp()'d dirs. This CL also cleans
up the temporary file creation in build_image.py.
Test: `m dist`
Test: python -m unittest test_common
Change-Id: Ied7d86126714b9b26e5df7bcca1dfdea21cfcf60
Mostly cosmetic changes, such as replacing print statement with print
function. Also change 'import cStringIO' to optionally look for the one
in io module, to allow Python 2/3 compatibility.
Test: pylint --rcfile=pylintrc sign_target_files_apks.py
Test: Run sign_target_files_apks.py on marlin target_files.zip.
Change-Id: I4dc98b01da6f89e624114bbca5522f659901c1f2
HeapItem defines __bool__(), which contains a logical error that should
return the opposite value.
Note that the bug only manifests while using Python 3, which calls
__bool__(). With Python 2, `if x:` or bool(x) actually calls
x.__nonzero__() or x.__len__(). If a class defines neither __len__() nor
__nonzero__(), as the case in HeapItem, it always returns True.
Test: python -m unittest test_blockimgdiff
Test: python3 -m unittest test_blockimgdiff
Test: Generate an incremental non-A/B OTA package successfully.
Change-Id: Ibe8430e0b495a7d2f430cfffb716d2536ffb53d2
The code is on infeasible path since we already have assertions in
common.BlockDifference().
Also remove the dead code that checks for OPTIONS.info_dict, as we
already set that in ota_from_target_files.main(), for both of A/B and
non-A/B.
Test: Generate incremental OTAs w/ and w/o the CL, and get identical
packages.
Change-Id: Ifb8fc101e78f5ce58c60c8e49028b66ce0d20246
The CL in [1] unintentionally breaks the OEM dict loading logic in the
incremental BBOTA path. We should always require and load the OEM
property dict if _either_ of the source and target builds uses OEM
properties. Otherwise with the current "and" operator, it skips loading
the OEM property dict and thus fails to generate an OTA package that has
OEM property changes (e.g. updating from build with fingerprint to
another one using thumbprint).
The CL in [1] actually makes the right change in the file-based OTA
path, but introduces the bug in the block-based OTA path.
This CL also cleans up the line that reads recovery_mount_options.
[1] commit 7f804ba71f ("releasetools:
allow for multiple OEM property values.").
Test: Genearte an OTA that has OEM property changes successfully.
Change-Id: Idce4ad59825d432618535ce09ab22bd7ddc524f2
This change allows one to run atest <module name>. The TradeFed test
config file will be automatically created in testcases directory based
on the settings.
Bug: 69929803, 69932044
Test: m -j NotificationStressTests libminijail_test
add following lines in notificationtests/Android.mk,
before "include $(BUILD_PACKAGE)"
LOCAL_COMPATIBILITY_SUITE := general-test
Change-Id: Ib5e9cd152d957b5d5beb6ed23a899a021a7e06d6
We used to have "--gen_verify" that would generate a verify package for
non-A/B devices. Since a) we don't have active users; b) it works with
non-A/B only; c) we have better alternatives (e.g. using fastboot to
compute the partition checksum), this CL removes the support.
Test: `m dist`
Change-Id: Ib4a2c8c2a0394b54e66c106089aa52a0b900f034
We have been shelling out to 'zip -d' to delete existing ZIP entries in
add_img_to_target_files.py. This CL moves the function into common.py,
and calls that for the similar work in ota_from_target_files.py. This CL
also adds unittests for the newly added function.
Test: `m dist`
Test: python -m unittest test_common
Test: ota_from_target_files.py generates identical packages w/ and w/o
the CL (so we know the streaming property computation is intact).
Test: Run 'add_img_to_target_files.py -a' that triggers a call to
ReplaceUpdatedFiles().
Change-Id: Icaa6c3ea3ee2166023f78fa79275295f837ea842
For devices using derived fingerprint (i.e. /system/build.prop doesn't
contain ro.build.fingerprint, but has ro.build.thumbprint instead), the
current code (in android.os.Build) doesn't have a matching logic to do
the same for ro.vendor.build.fingerprint. This means we will see
ro.build.thumbprint in /system/build.prop, while there's no matching
ro.vendor.build.thumbprint in /vendor/build.prop.
From signing script point of view, it should just apply the tag
replacement (e.g. test-keys -> release-keys) for whatever it sees when
signing a target_files.zip.
This CL also adds unit tests for EditTags() and RewriteProps().
Fixes: 27950003
Test: Use 'sign_target_files_apks.py' to sign a target that uses derived
fingerprint and vendor partition. Check VENDOR/build.prop.
Test: python -m unittest test_sign_target_files_apks
Change-Id: I09019da970840cd82f54b68a32b4e94984bc1d8d
The test is mostly trivial, but it ensures the result parsing from
actual calls to mke2fs.
Test: python -m unittest test_build_image
Test: `m dist` on aosp_marlin-userdebug (w/ and w/o defining
PRODUCT_SYSTEM_HEADROOM respectively).
Change-Id: I8b9964213950e76f6d7d5518414a1bab888b4706
This is only used in file-based OTA, where we may reserve space on
/system partition. With the deprecation of file-based OTA since O, the
code has become obsolete.
Test: `m dist`
Change-Id: I2bd686b292f93b566fe28a2bcd74c564ffc75dac
'-w' is a short option for '--wipe_user_data', which happens to be the
same as the whole-file signing flag for signapk.
This CL removes the short version to make it less error-prone. Users
should use '--wipe_user_data' to generate packages with enforced data
wipe.
Bug: 70259309
Test: ota_from_target_files.py generates packages with and without
'--wipe_user_data'.
Test: Not seeing active users that uses 'ota_from_target_files -w'.
Change-Id: Ib8c138663a87aafc6a0b5b6f7c286fb74ef6134b
BuildImage can return false for lots of reasons. Log two of the more
plausible ones. You'll still need to edit this file and pass True to
RunCommand to work out *what* went wrong, but at least this logging will
tell you *where* things went wrong.
Bug: http://b/63142920
Test: debugged my actual problem
Change-Id: If84427058108dcac0a1b5e2216b8bb11eef522ca
This also updates mkf2fsuserimg.sh to call sload.f2fs.
Change-Id: I245f6b3dfcd966fbc6783329d781375879fcd065
Signed-off-by: Jaegeuk Kim <jaegeuk@google.com>
C:661, 0: Wrong hanging indentation (add 2 spaces).
OPTIONS.input_tmp, "VENDOR_IMAGES")
^ | (bad-continuation)
C:673, 0: Wrong continued indentation (add 14 spaces).
os.path.join("IMAGES", img_name))
^ | (bad-continuation)
C: 54, 0: Import "import datetime" should be placed at the top of the module (wrong-import-position)
C: 55, 0: Import "import hashlib" should be placed at the top of the module (wrong-import-position)
C: 56, 0: Import "import os" should be placed at the top of the module (wrong-import-position)
C: 57, 0: Import "import shlex" should be placed at the top of the module (wrong-import-position)
C: 58, 0: Import "import shutil" should be placed at the top of the module (wrong-import-position)
C: 59, 0: Import "import subprocess" should be placed at the top of the module (wrong-import-position)
C: 60, 0: Import "import tempfile" should be placed at the top of the module (wrong-import-position)
C: 61, 0: Import "import uuid" should be placed at the top of the module (wrong-import-position)
C: 62, 0: Import "import zipfile" should be placed at the top of the module (wrong-import-position)
C: 64, 0: Import "import build_image" should be placed at the top of the module (wrong-import-position)
C: 65, 0: Import "import common" should be placed at the top of the module (wrong-import-position)
C: 66, 0: Import "import rangelib" should be placed at the top of the module (wrong-import-position)
C: 67, 0: Import "import sparse_img" should be placed at the top of the module (wrong-import-position)
W:391,12: Redefining built-in 'dir' (redefined-builtin)
After this CL, it still gives the warning below, which is to be
addressed with follow-ups.
R:635, 2: Too many nested blocks (6/5) (too-many-nested-blocks)
Test: `m dist`
Test: pylint --rcfile=pylintrc add_img_to_target_files.py
Change-Id: I64da184b6b69e93449dbfc989a5d7f46d5223f42
This option u'required-attributes' will be removed in Pylint 2.0
This option u'ignore-iface-methods' will be removed in Pylint 2.0
Test: pylint (1.6.5) no longer gives above warnings.
Change-Id: Ib07cd5aef87407260140c1fca0b11c439cbdf390
In some non-A/B setups, recovery.img is still being used. If AVB is
enabled, we currently don't add a hash footer to recovery.img nor do
we include the hash digest in vbmeta.img. This CL fixes that.
This was tested on a build with the following settings
TARGET_NO_RECOVERY := false
BOARD_USES_RECOVERY_AS_BOOT := false
BOARD_BUILD_SYSTEM_ROOT_IMAGE := false
BOARD_RECOVERYIMAGE_PARTITION_SIZE := 33554432
BOARD_AVB_RECOVERY_ADD_HASH_FOOTER_ARGS := --prop foo:bar
and then it was verified using 'avbtool info_image' that recovery.img
has a hash footer and a 'foo' property with the value 'bar'. This was
also checked successfully for vbmeta.img.
Test: See above.
Bug: None
Change-Id: I98124d5661ea768411416fa8d2a2ae6cc664fdc8
Switch attribute error to attribute unavailable, which apparently works
since it then broke the build :-/
Suppress unused parameter warnings. They aren't that useful for windows
functions that are just #ifndef WIN32. This is marked as a global cflag,
but something is overriding that with the clang+mingw changes, and it's
not actually that hard to suppress or fix where necessary.
Bug: 69933068
Test: mmma build/tools/libhost
Change-Id: I09adda4544bc9b56452b4fe182dba3d0b701dcbb
Imgdiff expects the input files to be valid zip archives. During the
split of large apks, imgdiff may fail when there's a hole in the input
file; potentially due to the blocks allocation of mke2fs. This CL solves
the issue by falling back to normal split in case of the imgdiff split
failure. The split transfers will then use bsdiff instead.
Bug: 69624507
Test: generate the incremental package for the failed targets and check the
transfers.
Change-Id: I4882452378123e60ad3434053b57f33e53ac4b82
... to include system/update_engine/scripts if ANDROID_BUILD_TOP is set
(e.g. after lunching a target).
Test: ./build/make/tools/releasetools/check_ota_package_signature.py
works without manually setting PYTHONPATH.
Change-Id: I0c5101fab2dd69cb6a598b909aa04bc4d5f45284
Bug: 68224784
Test: Enable AVB on aosp_bullhead-userdebug. `m dist`. Note that the
prebuilt vendor.img needs to be AVB-signed first.
Test: `m dist` with aosp_walleye-userdebug.
Change-Id: I18235e4d0dde6af71a96e46ec434480cc9a22bef
AddImagesToTargetFiles() takes either a zip file, or a zip root as
input. We used to create IMAGES/ directory only when working with zip
root input. Commit 262bf3f0b5 has changed
to also stage boot / recovery images there when working with a zip file.
This CL makes sure the directory is always available under both modes.
Bug: 63456822
Test: zip -d target_files.zip IMAGES/\*;
add_img_to_target_files.py target_files.zip
Test: sign_target_files_apks.py target_files.zip signed-target_files.zip
Change-Id: Iea91d0403cdec1b16bb93bb71d3ed06856b8f7c3
Most logtags files only have entries with hardcoded numbers.
Allow these files to be converted to java without needing the
merged logtags file.
Test: m checkbuild
Change-Id: I4cd38aa502fc9b8e50af92ff8296d3ad6abfb56e
If only vbmeta.img is missing from the target_files.zip,
'add_img_to_target_files.py -a' will fail to re-create vbmeta.img.
++++ vbmeta ++++
Traceback (most recent call last):
File "./build/make/tools/releasetools/add_img_to_target_files.py", line 693, in <module>
main(sys.argv[1:])
File "./build/make/tools/releasetools/add_img_to_target_files.py", line 687, in main
AddImagesToTargetFiles(args[0])
File "./build/make/tools/releasetools/add_img_to_target_files.py", line 584, in AddImagesToTargetFiles
boot_contents = boot_image.WriteToTemp()
AttributeError: 'NoneType' object has no attribute 'WriteToTemp'
Because it has skipped loading the boot.img, which leads to the above error.
Bug: 63456822
Test: As follows:
$ zip -d target_files.zip vbmeta.img
$ add_img_to_target_files.py -a target_files.zip
Test: `m dist` on aosp_arm64-userdebug
Change-Id: Ieba5f0b6848c3ec5d8bcfd24d48cc5af8b39f06c
When generating full OTAs for AVB-enabled targets, it fails the size
assertion for boot.img.
ERROR: boot size (31457280) is 100.00% of limit (31457280)
The size assumption is no longer true for AVB-signed images, because
the actual image size should be always identical to the limit.
Bug: 67841177
Test: Enable AVB on bullhead. `m dist`.
Change-Id: I9bd432ccd88094628eb71308554839549eccb6d9
With the new implementation of handling large apks, we need to call
imgdiff with block-limit to split the apk and generate the patch at
the same time. The call to imgdiff would significantly increase the
time consumption of the "FindTransfers" function which we used to
execute sequentially. This cl addresses this issue and speeds up the
process by making the imgdiff call parallel.
Bug: 34220646
Test: Create and sideload an incremental package for angler
Change-Id: Id62e348418fc1d22e32ea6c8ac16d9ab3ec92d7b
With the update of brotli to version 1.0.1 the new binary tool is now
called "brotli". This patch updates the scripts to use the new name.
Bug: 34220646
Test: `make dist`; ota_from_target_files on non-A/B device.
Change-Id: Ie8da4333388d029d5960a723a55efd875d8ee820
When signing a target_files.zip, the OTA certificate specified by
default_system_dev_certificate could be replaced with a mapped key. When
that happens, we must explicitly specify --package_key when generating
OTA packages with ota_from_target_files.py. Otherwise the OTA package
will be signed with the wrong key, which leads to verification failures.
This CL updates the default_system_dev_certificate value in
misc_info.txt accordingly.
Test: Sign a target_files.zip and replace the OTA key. Check
META/misc_info.txt in the generated target_files.zip.
$ ./build/make/tools/releasetools/sign_target_files_apks.py -v \
--replace_ota_keys \
-k build/target/product/security/testkey=build/target/product/security/platform \
out/dist/aosp_marlin-target_files-eng.tbao.zip \
signed-marlin-target_files-test.zip
Change-Id: I093234b5add3e27c5b3887cefeffd74e6f0a3e98
test_common constructs a few 2GiB strings in memory, which leads to huge
memory footprint (18GiB). This CL moves away from in-memory strings to
generators, which reduces the memory use down to 41MiB. It also reduces
the time cost from 294s to 139s as an extra benefit for free.
The CL addresses some trivial pylint warnings as well.
* Before
$ /usr/bin/time -v python -m unittest -v test_common
...
----------------------------------------------------------------------
Ran 11 tests in 294.986s
OK
Command being timed: "python -m unittest -v test_common"
User time (seconds): 110.51
System time (seconds): 109.34
Percent of CPU this job got: 74%
Elapsed (wall clock) time (h:mm:ss or m:ss): 4:55.06
Average shared text size (kbytes): 0
Average unshared data size (kbytes): 0
Average stack size (kbytes): 0
Average total size (kbytes): 0
Maximum resident set size (kbytes): 18894172
Average resident set size (kbytes): 0
Major (requiring I/O) page faults: 1
Minor (reclaiming a frame) page faults: 20774908
Voluntary context switches: 48
Involuntary context switches: 3241
Swaps: 0
File system inputs: 184
File system outputs: 8406424
Socket messages sent: 0
Socket messages received: 0
Signals delivered: 0
Page size (bytes): 4096
Exit status: 0
* After
$ /usr/bin/time -v python -m unittest -v test_common
...
----------------------------------------------------------------------
Ran 11 tests in 139.100s
OK
Command being timed: "python -m unittest -v test_common"
User time (seconds): 59.00
System time (seconds): 4.73
Percent of CPU this job got: 45%
Elapsed (wall clock) time (h:mm:ss or m:ss): 2:19.17
Average shared text size (kbytes): 0
Average unshared data size (kbytes): 0
Average stack size (kbytes): 0
Average total size (kbytes): 0
Maximum resident set size (kbytes): 41252
Average resident set size (kbytes): 0
Major (requiring I/O) page faults: 0
Minor (reclaiming a frame) page faults: 106569
Voluntary context switches: 44
Involuntary context switches: 103
Swaps: 0
File system inputs: 8
File system outputs: 8422808
Socket messages sent: 0
Socket messages received: 0
Signals delivered: 0
Page size (bytes): 4096
Exit status: 0
Fixes: 68988396
Test: See above.
Change-Id: I00f16603a4ee59fb085b189c6f5b5ee9d2378690
imgdiff fails on plain strings after we put more checks the input
data. This leads to failure when running unittest
"test_recovery_from_boot". The CL fixes the issue by serving gzipped
dummy images as the test input.
Bug: 68988329
Test: InstallRecoveryScriptFormatTest passes
Change-Id: I1952b311681ac4f39c0e51aef1de1177aebeac43
$ python rangelib.py
**********************************************************************
File "rangelib.py", line 323, in __main__.RangeSet.next_item
Failed example:
list(rangelib.RangeSet("10-19 3 5 7").next_item())
Exception raised:
Traceback (most recent call last):
File "/usr/lib/python2.7/doctest.py", line 1315, in __run
compileflags, 1) in test.globs
File "<doctest __main__.RangeSet.next_item[2]>", line 1, in <module>
list(rangelib.RangeSet("10-19 3 5 7").next_item())
NameError: name 'rangelib' is not defined
**********************************************************************
1 items had failures:
1 of 3 in __main__.RangeSet.next_item
***Test Failed*** 1 failures.
Test: python rangelib.py
Change-Id: I6c791ff4453f40d90b34f97b881460da9878fe64
"ro.vendor." prefix will be used only for vendor-specific properties
except for ro.vendor.build.date, ro.vendor.build.date.utc and
ro.vendor.build.fingerprint.
Those exceptions have been used since Android L. So we can't rename them
to support the system-only upgrade from old Android.
Bug: 36796459
Test: confirmed that ro.vendor.product.* were changed to
ro.product.vendor.* in vendor/build.prop
Merged-In: Ibc107b00264a2dc194261d1968b27a9b17b5d798
Change-Id: Ibc107b00264a2dc194261d1968b27a9b17b5d798
(cherry picked from commit 8cf6443264)
It has been long deprecated since commit
fdd8e69c42 (Donut).
Test: `m dist` with aosp_marlin-userdebug and aosp_angler-userdebug.
Test: No user of "--board_config" in code search.
Change-Id: Ic0336c22cac32de5de88161748d09b579f892e99
Commit b937ead5d9 added the fallback to
bsdiff on imgdiff failures. However, it missed setting the transfer
style accordingly, which led to patch header mismatch.
Bug: 68659848
Test: Generate an incremental that has the fallback from imgdiff to
bsdiff. Examine the generated transfer list and verify that it has
"bsdiff" for the fallback command.
Change-Id: I55e46879d590a8af82ea796b9d98ffdb30360408
Following the new API contract, this effectively add extra padding
before central dir to make it 4KB aligned.
Test: build succeeded
Bug: 30972906
Change-Id: I7cac9d2c4371b473c88df867b3b2ae906443db10
Use zip_archive directly. Note that this codepath is used only
when recompressing archives with zopfli during the alignment step.
It's unclear whether this is in use at all, but I verified that the
results are identical (note the usage of the "-z" flag in the test
below).
Test: make && out/host/linux-x86/bin/zipalign -v -f -z 8 \
out/target/product/marlin/system/app/Email/Email.apk ./out.zip
Bug: 35246701
Change-Id: I641cdb6d409cc07974d49d42c9f9e6d4f905e472
"ro.vendor." prefix will be used only for vendor-specific properties
except for ro.vendor.build.date, ro.vendor.build.date.utc and
ro.vendor.build.fingerprint.
Those exceptions have been used since Android L. So we can't rename them
to support the system-only upgrade from old Android.
Bug: 36796459
Test: confirmed that ro.vendor.product.* were changed to
ro.product.vendor.* in vendor/build.prop
Change-Id: Ibc107b00264a2dc194261d1968b27a9b17b5d798
Duplicate essential product sysprops for manufacturer, model, brand, name, and device
from system on vendor so that mixed builds (vendor.img + GSI system.img) have correct
product information in their CTS/VTS reports.
Test: make
Bug: 64458205
Merged-In: Ib63b37772be493b9e035d9b7e8c5e2cf66f2fa8c
Change-Id: Ib63b37772be493b9e035d9b7e8c5e2cf66f2fa8c
(cherry picked from commit aea0f92ff3)
Bug: 36232423
Test: Create OTA packages (full, incremental and downgrade) with
ota_from_target_files.py; check the parameters passed down to
brillo_update_payload.
Change-Id: I3af397f05fe16f4a9f0eb73e8272bac6afdc002a
(cherry picked from commit ed08bc4a22)
When generating block based OTAs, we read files from the sparse image
directly with the help of block map file. However, the block map info
might not be accurate if the image is created with mke2fs. Because
mke2fs may skip allocating actual blocks if they contain all zeros.
ota_from_target_files.py consequently passes incomplete APK files to
imgdiff, which fails to generate patches.
This CL works around the issue by falling back from imgdiff to bsdiff on
failures. We should figure out a better way in b/68016761 to remove the
workaround, which would otherwise hide other issues in imgdiff.
Bug: 67824829
Bug: 68016761
Test: ota_from_target_files.py passes on previously failing TF zips.
Change-Id: Ib24c5b5f89812b97a0c87c6bf0dc147ae39bc92f
