'monotonic' has been non-optional since [1] (L-MR1). Fix the comment in
RangeSet.parse(), as well as the use in blockimgdiff.py.
[1] commit 8b72aefb5a.
Test: Generate an incremental BBOTA package.
Change-Id: I7f95231683473b4f0f07f9c83fccc0e36a1340cb
This CL detects incomplete block ranges (e.g. due to the holes in
mke2fs created images). Such block ranges will be tagged, so we won't
attempt to imgdiff those files. Note that the change to blockimgdiff.py,
which uses the tag info, will come in a separate CL.
An 'extra' attribute is added to RangeSet class, which defaults to an
empty dict. An 'incomplete' tag will be added into the dict by the
caller of the class. Not adding this tag as an immediate attribute,
because it is not a property regarding the ranges being represented, but
rather some storage space for the caller.
This CL also refactors GetSparseImage and RoundUpTo4K into common.py, so
the same code can be called from both of ota_from_target_files.py and
validate_target_files.py. Not able to add unittests for
GetSparseImage(), as SparseImage requires data in specific format.
Bug: 68016761
Test: Run validate_target_files.py on target-files.zip. It skips
validating files with missing holes as before.
Test: Run ota_from_target_files.py on angler target-files.zip. It gives
identical packages w/ and w/o the CL.
Test: pylint on changed files. There're warnings with common.py, but
unrelated to this change.
Change-Id: I126ccfea13c0d5ebcc8c1b4ff1a4f9200e97423a
$ 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
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
After applying update_target_files_incr_ext4.sh, some files may end up
occupying unsorted block fragments. In one example, an apk file has the
block range [258768-259211,196604]. The monotonic flag in rangelib sets
incorrectly for this example and leads to a bad input file for imgdiff.
After fixing the flag, bsdiff is called instead of imgdiff and the
incremental OTA package generates successfully.
Bug:28053885
Change-Id: Ib841bf449ff6a29314fc4a1b8fba941a6dc532ac
RangeSet("2-10").monotonic gives a wrong result of "False". Fix the bug
and add more tests into test_rangelib.py.
Change-Id: I04780571b45ecafd34040f405756b9745a9e21a5
For BBOTA v3, we need to stash source blocks to support resumable
feature. However, with the growth of file size and the shrink of the
cache size, source blocks that represent a file are too large to be
stashed as a whole. CL in [1] solves the issue by replacing the diff
command with a "new" command. However, it may increase the generated
package size substantially (e.g. from ~100MB to ~400MB).
With this CL, if a file spans too many blocks, we split it into smaller
pieces by generating multiple commands. For the same case above, it
reduces the package size to ~150MB.
One potential downside is that after splitting, files like .jar,
.apk and .zip can no longer use imgdiff. We may lose the potential
benefit of using imgdiff for patch size reduction.
[1] commit 82c47981bd
Bug: 22430577
Change-Id: I5684fab0ac41c8d94fe18f52a34cef77e06f741c
dm-verity may touch some blocks that are not in the care_map due to
block device read-ahead. It will fail if such blocks contain
non-zeroes. As a workaround, we mark them as extended blocks and
zero out explicitly to avoid dm-verity failures.
Bug: 20881595
Change-Id: Id097138bfd065c84eac088b3ad49758010b2927b
(cherry picked from commit 2fd2c9bfd6)
This caught a few bugs/syntax errors (a few character classes were not
escaped properly in regex patterns, some indentation was illegal,
etc).
Change-Id: I50637607524e68c4fb9cad7167f58a46b8d26b2c
Generate version 2 of the block_image_update transfer list format.
This improves patch size by a different strategy for dealing with
out-of-order transfers. If transfer A must be done before transfer B
due to B overwriting A's source but we want to do B before A, we
resolve the conflict by:
- before B is executed, we save ("stash") the overlapping region (ie
the blocks B will overwrite that A wants to read)
- when A is executed, it will read those parts of source data from
the stash rather than from the image.
This reverses the ordering constraint; with these additions now B
*must* go before A. The implementation of the stash is left up to the
code that executes the transfer list to apply the patch; it could hold
stashed data in RAM or on a scratch disk such as /cache, if available.
The code retains the ability to build a version 1 block image patch;
it's needed for processing older target-files.
Change-Id: Ia9aa0bd45d5dc3ef7c5835e483b1b2ead10135fe
Replace the xdelta/xz-based block OTA generation with a new system
based on the existing bsdiff/imgdiff tools.
Bug: 16984795
Change-Id: Ia9732516ffdfc12be86260b2cc4b1dd2d210e886
