In two-step OTAs, we write recovery image to /boot as the first step so
that we can reboot from there and install a new recovery image to
/recovery. However, bootloader will show "Your device is corrupt"
message when booting /boot with the recovery image. Because the recovery
image encodes the path of "/recovery" as part of the signature metadata,
which fails the verified boot.
This CL generates a special "recovery-two-step.img" in addition to the
regular recovery.img. This image encodes "/boot" when being signed,
which will be flashed to /boot at stage 1/3 in a two-step OTA.
Here are the desired changes:
- 'IMAGES/recovery-two-step.img' exists in target_files.zip for non-A/B
targets (e.g. bullhead). The image should not exist for targets that
don't have a recovery partition (e.g. A/B devices like sailfish).
- <device>-img.zip should not contain 'recovery-two-step.img'.
- Nothing should change when building non-two-step OTAs. For two-step
OTAs, 'recovery-two-step.img' should be included in the OTA package;
'updater-script' should flash this image to /boot at stage 1/3.
- When building a two-step OTA with an input TF.zip that doesn't have
IMAGES/recovery-two-step.img, it should use the existing
IMAGES/recovery.img instead.
Bug: 32986477
Test: Tested the steps above on bullhead and sailfish.
Change-Id: I34e6c599bcf2011d4cd5c926999418b3975d6d0f
The 'system_img_path' parameter was introduced in commit
d995f4b04d, but became obsolete since
commit 2ce63edab7.
Test: m dist
Change-Id: Iffd496d929db5cc3dfc955a48bfc1b1317bd012f
We are investigating replacing make_ext4fs with the upstream tool mke2fs.
To mitigate the trouble that may arise if the new tool behave differently
compared to the old one, there will be a transition period.
Devices that want to use the new way of creating ext4 images can set the
variable "TARGET_USES_MKE2FS" to true in their BoardConfig.mk
By default, the build system will choose the old tool 'make_ext4fs'.
Test: m otapackage with TARGET_USES_MKE2FS={,false,true}
Change-Id: I282bcb9efe335a86c53986283090ca947d65c7f8
Prior to this CL, it was calling the hard-coded "java" although it was
accepting a "--java_path" option.
Also switch OPTIONS.java_args from string to list. Otherwise it won't
work when providing multiple args.
Bug: 32737832
Test: Specify "--java_path=" and "--java_args" when invoking
sign_target_files_apks.py with "-v". Check the commands being
called.
Change-Id: Id7ef98e778646d532027434de7fba9b7a104dbd0
set() doesn't keep elements according to the order of insertion. So
Transfers managed with set() in intermediate steps may not appear in the
same order across runs. This leads to slightly different output packages
when generating the same incremental OTA.
This CL fixes the issue by replacing set() with OrderedDict() in
blockimgdiff.GenerateDigraph() and blockimgdiff.FindVertexSequence().
It also adds a testcase that ensures blockimgdiff.GenerateDigraph()
preserves the insertion order for Transfer.goes_after set.
Bug: 32220816
Test: ota_from_target_files.py gives identical package when running
multiple times.
Change-Id: I56d551e5ca926993ab46896e33c80e0ce42e506a
Currently, whether contains patch or verbatim, compute with file size
and patch size.
But ota file must be compressed with zip, so it should be better with
compressed size than uncompressed.
Test: aosp_shamu-user build without proprietary blobs between MOB30P and NRD90S
$ du -k ota_shamu_old.zip ota_shamu_new.zip
217252 ota_shamu_old.zip
216520 ota_shamu_new.zip
Change-Id: If68cb1fbe2f7815067451915a0dcfe93ea5ba8d6
Signed-off-by: YOUNG HO CHA <ganadist@gmail.com>
system/extras/verity/build_verity_metadata.py now accepts
"--signer_args" to specify verity signer args.
Also remove the duplicate "--verity_signer_args" in
add_img_to_target_files.py, as we already have that in common.py.
Bug: 31500665
Test: Building and signing work w/ and w/o --signer_args.
Change-Id: I02f59c50a1ebf15c5505e9fffd5b9bbbbaa785be
While the system.img images currently built with AVB support verify
correctly, mounting the filesystem content fails. This is because
'avbtool add_hashtree_footer' used to claim some of the unused /
DONT_CARE space for stashing the verity tables and this resulting in the
mapped device ending up being smaller causing the mount failure.
Fix this by leaving enough room for AVB hashtree and metadata before
building the image. This is achieved by moving the AVB hashtree support
into build_image.py and using a just added '--calc_max_image_size'
option to 'avbtool add_hashtree_footer' to figure out how much space to
leave out.
This depends on https://android-review.googlesource.com/#/c/281821/
Bug: 31264226
Test: Mounting dm-verity set up from system.img now works.
Merged-In: I4c5de1004c1059f8c582e76b3b8517d427aa1a87
Change-Id: I945a5f1f6782791736cd319f216cfa6b448fb04d
Remove the obsolete workaround for API 24. Also make it pylint clean.
Test: 1. Sign a target_files.zip and get identical results.
2. `pylint --rcfile=pylintrc sign_target_files_apks.py` gives 10.00/10.
Change-Id: I21736f959c5182486fd8ccebea9bbc594edef9fb
sign_target_files_apks.py calls common.GetBootableImage() but without
calling 'OPTIONS = common.OPTIONS' first. In common.GetBootableImage(),
we should use the local info_dict parameter instead of OPTIONS.info_dict.
Test: sign_target_files_apks.py generates signed-TF.zip successfully.
Change-Id: Ia3d32b88691c26e5fb98feea709e3e3c3eb70fdb
This updates the build system for the new Android Verified Boot
codebase. As this is based on Brillo Verified Boot, this change replaces
the existing BVB support.
Android Verified Boot is enabled by the BOARD_AVB_ENABLE variable
BOARD_AVB_ENABLE := true
This will make the build system create vbmeta.img which will contain a
hash descriptor for boot.img, a hashtree descriptor for system.img, a
kernel-cmdline descriptor for setting up dm-verity for system.img and
append a hash-tree to system.img.
Additionally, the descriptors are left in boot.img and system.img so a
third party can create their own vbmeta.img file linking - using the
option --chain_partition - to these images. If this is not needed
footers can be erased using the 'avbtool erase_footer' command. It's
also harmless to just leave them in the images.
By default, the algorithm SHA256_RSA4096 is used with a test key from
the AVB source directory. This can be overriden by the
BOARD_AVB_ALGORITHM and BOARD_AVB_KEY_PATH variables to use e.g. a
4096-bit RSA key and SHA-512:
BOARD_AVB_ALGORITHM := SHA512_RSA4096
BOARD_AVB_KEY_PATH := /path/to/rsa_key_4096bits.pem
To prevent rollback attacks, the rollback index should be increased on a
regular basis. The rollback index can be set with the
BOARD_AVB_ROLLBACK_INDEX variable:
BOARD_AVB_ROLLBACK_INDEX := 5
If this is not set, the rollback index defaults to 0.
The variable BOARD_AVB_MAKE_VBMETA_IMAGE_ARGS can be used to specify
additional options passed to 'avbtool make_vbmeta_image'. Typical
options to be used here include '--prop', '--prop_from_file', and
'--chain_partition'.
The variable BOARD_AVBTOOL_BOOT_ADD_HASH_FOOTER_ARGS can be used to
specify additional options passed to 'avbtool add_hash_footer' for
boot.img. Typical options to be used here include '--hash_algorithm' and
'--salt'.
The variable BOARD_AVBTOOL_SYSTEM_ADD_HASHTREE_FOOTER_ARGS can be used
to specify additional options passed to 'avbtool add_hashtree_footer'
for systems.img. Typical options to be used here include
'--hash_algorithm', '--salt', and '--block_size'.
BUG=31264226
TEST=Manually tested on edison-eng by inspecting {boot, system,
vbmeta}.img in out/ directory as well as their counterparts in
the IMAGES/ directory of edision-target_files-eng.zeuthen.zip
Merged-In: Ic9a61cfc65c148b12996e57f04da5432eef6b982
Change-Id: I97042655bca15e7eac899f12c5bada2f6184d307
In BBOTA, we generate patches based on _all_ the blocks of a pair of
input files (src and tgt). For security incremental OTAs, one common
pattern is that only a few blocks are changed in odex files (e.g.
headers). We don't really need to stash/patch the unchanged blocks.
This CL analyzes the unchanged blocks in odex files and computes the
diff for the changed blocks only. It reduces the OTA install time by
about 25% to 40% in our experiments, by paying an increase of 5% to 30%
OTA generation time cost.
Bug: 31570716
Test: Generate an incremental and apply on device.
Change-Id: If842c1afeff6894a3d27eb60b7e8f65a179b7977
Currently it supports verifying packages signed with RSA algorithms
(v1-v4 as in bootable/recovery/verifier.cpp). No support for ECDSA (v5)
signed packages yet.
$ ./build/tools/releasetools/check_ota_package_signature.py \
bootable/recovery/tests/testdata/testkey_v1.x509.pem \
bootable/recovery/tests/testdata/otasigned_v1.zip
Package: bootable/recovery/tests/testdata/otasigned_v1.zip
Certificate: bootable/recovery/tests/testdata/testkey_v1.x509.pem
Comment length: 1738
Signed data length: 2269
Use SHA-256: False
Digest: 115e688ec3b77743070b743453e2fc6ce8754484
VERIFIED
Bug: 31523193
Test: Used the tool to verify existing packages (like above).
Change-Id: I71d3569e858c729cb64825c5c7688ededc397aa8
For some partition sizes, we currently build an image that's 1-2
blocks smaller than the actual partition, which causes fs_mgr to
not find metadata. This change adds padding to FEC metadata that
correctly positions the metadata header at the end.
Bug: 28865197
Change-Id: Ie0e044715a9c5ae8ba395e7d2ff9fbd7cffc0b4c
The userdata.img and cache.img entries are not useful in signed builds;
because fastboot doesn't look at these two entries in the *img.zip when
flashing a device. And they aren't used elsewhere. Therefore, skip
building the image files for them when signing the target files with
sign_target_files_apks. Also, add an option "--is_signing" to avoid
adding these two images when we call add_img_to_target_files.
Change-Id: I39ba91a86d9a856d7d01771f6d1403dbf21f2011
Test: Run sign_target_files_apks on a target file and userdata/cache.img doesn't not generate.
Bug: 30642470
add_img_to_target_files.py fails when the target_files.zip is over 4GiB
when adding IMAGES/ folder. Specify the flag to allow creating
target_files.zip with ZIP64 extension.
Other zip artifacts (-img.zip, -ota.zip etc) remain in non-ZIP64 format.
zip2zip is not affected, which still creates non-ZIP64 zips even when
copying from target_files in ZIP64.
Bug: 30961841
Test: "make dist" with large system image and check the artifacts.
Change-Id: I0568745f01ef8f0239081f783eac92288d4fdd84
Do not copy the "META/care_map.txt" from the source zipfile when
signing the target files with sign_target_files_apks. Because we'll
generate a new care_map after rebuilding the system/vendor images;
and we'll write the new "META/care_map.txt" to the signed-target-file.
Change-Id: I6919cfdf8314a4084b5f612a9c89469f391486a4
Test: Run sign_target_files_apks locally, and the entry is updated.
Bug: 30812253
On A/B devices (i.e. system_root_image="true"), /default.prop is
packaged at ROOT/default.prop (as opposed to BOOT/RAMDISK/default.prop
for non-A/B devices). Update the path so that we handle properties like
ro.bootimage.build.fingerprint properly.
The one for recovery is not affected, which stays at
BOOT/RAMDISK/default.prop for A/B devices and gets updated correctly.
Bug: 30811237
Test: Verify the property in the generated signed-TF.zip.
Change-Id: Id201a042d7ea988a64f89c6d04f43326a9851e27
The update-payload-key is used by update_engine_sideload from recovery
to verify an update payload.
Bug: 27178350
Change-Id: I7a0a307ae565e5e9cbf2c9b58fbcc055e87771ce
We were using the package name as the key to index APKs. APKs from the
same package got messed up and gave wrong signature summary. Switch to
using the package filename as the key, which is identical in a given build.
Also fix the trailing space when printing the signature summary.
Bug: 30418268
Test: Run with a target_files.zip that has multiple APKs from the same package.
Change-Id: I6317e8c05e987c5690915e05c294153d10e2f0ab
We should disable using imgdiff if *any* of the source and target
partitions uses squashfs.
Bug: 30004734
Test: Create an incremental with two builds with one of them uses squashfs.
Change-Id: I826cd13d7b852c548e4b45e61f5ae00f6407cac3
(cherry picked from commit f8acad1480)
We use imgdiff to handle files in zip format (e.g. jar/zip/apk) for
higher compression ratio.
For system/vendor in squashfs, a) all files are compressed in LZ4
format; b) we use 4096-byte block size in their sparse images, but the
files in squashfs may not be laid out as 4K-aligned. So the blocks for
a given file as listed in block map may not form a valid zip file, which
may fail the patch generation with imgdiff.
Disable using imgdiff for squashfs images, and use bsdiff instead.
Bug: 22322817
Change-Id: Ie76aa4cece5c9d38cb1d1a34c505a4a8f37512d3
(cherry picked from commit 293fd135c7)
update_engine expects the extracted public key instead of the
certificate.
Bug: 28701652
Change-Id: I292d39da9e039f96d01a4214226aeb46f8cb881d
(cherry picked from commit afaf295cb8)
add_img_to_target_files.py has an option of "-a" to add missing
images only. Under this option, the script should skip copying
the radio images for A/B devices when given image exists already
under "IMAGES/".
Test: Run the command on an A/B device, the existing radio images under "IMAGES/" don't get overwritten; and missing images are added correctly.
Bug: 29608905
Change-Id: Ie034b85a5d777d53e367f99470cea4d19cb1aaaf
