This CL makes the following changes to RangeSet:
- Uses std::pair<size_t, size_t> to represent a Range;
- Uses std::vector<Range> to represent a RangeSet;
- Provides const iterators (forward and reverse);
- Provides const accessor;
- 'blocks()' returns the number of blocks (formerly 'size');
- 'size()' returns the number of Range's (formerly 'count').
Test: recovery_unit_test
Test: Apply an incremental update with the new updater.
Change-Id: Ia1fbb343370a152e1f7aa050cf914c2da09b1396
We don't need to take raw pointers out of the parsed arguments.
std::unique_ptr handles the dereferencing automatically.
Test: mmma bootable/recovery
Change-Id: I1beabf6e04dc350bdad7b36cee5fb345c82b28f2
Also move RangeSet into a header file to make it testable, and add unit
tests.
In RangeSet::Parse() (the former parse_range()), use libbase logging to
do assertions. This has the same effect as the previous
exit(EXIT_FAILURE) to terminate the updater process and abort an update.
The difference lies in the exit status code (i.e. WEXITSTATUS(status) in
install.cpp), which changes from 1 (i.e. EXIT_FAILURE) to 0.
Test: recovery_unit_test
Test: Apply an incremental update with the new updater.
Change-Id: Ie8393c78b0d8ae0fd5f0ca0646d871308d71fff0
LOG(INFO) already appends a newline. Don't print redundant newline.
Test: No extra blank lines when calling ui_print(). And on-screen UI
shows the same.
Change-Id: I74e9a8504a7146a6cb3dae02fe2406d0dd54069b
Then rename RangeSinkState to RangeSinkWriter. RangeSinkWriter reads
data from the given FD, and writes them to the desination RangeSet.
Test: Apply an incremental with the new updater.
Change-Id: I5e3ab6fc082efa1726562c55b56e2d418fe4acaf
Mostly for applypatch family APIs like ApplyBSDiffPatch() and
ApplyImagePatch(). Changing to size_t doesn't indicate they would
necessarily work with very large size_t (e.g. > ssize_t), just
similar to write(2). But otherwise accepting negative length doesn't
make much sense.
Also change the return type of SinkFn from ssize_t to size_t. Callers
tell a successful sink by comparing the number of written bytes against
the desired value. Negative return values like -1 are not needed. This
also makes it consistent with bsdiff::bspatch interface.
Test: recovery_component_test
Test: Apply an incremental with the new updater.
Change-Id: I7ff1615203a5c9854134f75d019e266f4ea6e714
Rename to LoadSourceBlocks() by moving the target blocks parsing part
into the caller. This allows detecting whether the target blocks have
already had the expected data before loading the source blocks. It
doesn't affect anything when applying an update package for the first
time, but it skips loading the unneeded source blocks when resuming an
update. It additionally avoids unnecessarily dumping the "corrupt"
source/stash blocks when resuming an update.
Bug: 33694730
Test: Apply an incremental update with the new updater.
Test: Resume an incremental update with the new updater.
Change-Id: I794fd0d1045be7b3b7f8619285dc0dade01398d0
Clean up a few functions that take CommandParameters& as the first
parameter. We don't need to take duplicate arguments if they always come
from CommandParameters. This redundancy came from the point we replaced
strtok()s (commit baad2d454d).
Test: Apply an incremental update with the new updater.
Change-Id: I2912b8ce6bc7580bf7f566e125f12270e679e155
The script support for BBOTA v1 and v2 has been dropped in commit
8fad03e7712082eb880ffaaffb69eb13252ce220 (platform/build).
Bug: 33694730
Test: Apply an incremental with the new updater.
Test: recovery_component_test
Change-Id: I038b1bf8d10f030cab8ec0aa6ee565c5a9545dfd
As of C++ specification size_type erase( const key_type& key );
removes the element (if one exists). There is no need to perform
the check twice.
Change-Id: I4b057c08526abc7c2a483a60f9e166e4d8f56a74
It's valid to provide only 1 argument to apply_patch_check(). We
shouldn't fail the argument parsing.
Bug: 36541737
Test: recovery_component_test passes.
Test: recovery_component_test captures the failure without the fix.
Test: The previously failed update applies successfully.
Change-Id: Iee4c54ed33b877fc4885945b085341ec5c64f663
And switch them to std::vector & std::unique_ptr
Bug: 32117870
Test: recovery tests passed on sailfish
Change-Id: I5a45951c4bdf895be311d6d760e52e7a1b0798c3
Our updater created the stashes with root permission. This causes an
access denial when the RecoverySystem service tries to clean up these
blocks after a failing update. As a result, the subsequent OTA updates
may fail due to insufficient cache space.
Bug: 36457133
Test: stashed blocks cleaned successfully after reboot
Change-Id: If0ca99638cdfa1033646f29d9cc92b5ff1bacac1
Since this was putting the intermediate file in obj/PACKAGING, every
installclean was removing it and triggering updater to rebuild. Instead,
use the standard generated-sources-dir. The dep file can also be removed
now that ninja will re-run the generator if the command line changes.
Test: m -j updater; m installclean; m -j updater
Test: Only change to aosp_fugu updater before/after is the debug info
Change-Id: I20928bd2049d4a3d4e21f83fd64d16cfdc541958
This reverts commit 90eff6a340.
Also fix the bug where stashed blocks are not freed.
Bug: 21124445
Test: Previous failed update succeeded on bullhead
Change-Id: I23d232331a2beb51b6dcc82c957c87bc247d0268
Test: Apply an incremental BBOTA package with the new updater.
Test: Resume an interrupted BBOTA (so it cleans up the partial stash).
Change-Id: I620cc57ee6366845bcffbc19210f7a01e2196052
This reverts commit bb0cd75a0e.
Broke the 'free' command that deletes a stash.
Bug: 36242722
Test: The previously failed incremental applies successfully.
Change-Id: I1237cb0a33adfbeea57e0465b629704862ba13aa
It will be helpful for debug if we know which blocks are corrupted after
a verification failure. This CL prints the SHA-1 for each source block
in a transfer command if these blocks don't have an expected hash. And
along with the correct SHA-1, we will catch the corrupted blocks.
Bug: 21124445
Test: Printed the mismatched SHA-1 for bullhead during an update.
Change-Id: I683d4bdaf9a335035045b3f532b3a265b2fcbbfc
This CL removes the updater support for delete(), symlink(), rename(),
set_metadata() and set_metadata_recursive(). Such functions have been
removed from the generation script in commit
f388104eaacd05cfa075d6478369e1d0df5ddbf3 (platform/build).
Note: This CL also removes delete_recursive() which seems to have never
been supported in generation script.
Bug: 35853185
Test: recovery_component_test passes.
Change-Id: I51e1ec946fa73761118fa1eaa082423df6d588e9
-1 is not a valid exit status.
Also replace a few exit(1) with exit(EXIT_FAILURE).
Test: mmma bootable/recovery
Change-Id: I4596c8328b770bf95acccc06a4401bd5cabd4bfd
Now ApplyBSDiffPatch() will stream the output to sink as we go instead
of sinking everything at the end.
Test: recovery_host_test
Bug: 26982501
Change-Id: I05b6ed40d45e4b1b19ae72784cf705b731b976e3
Currently the ui_print command between the recovery and updater doesn't
append newline. Updater has to send an extra "ui_print" command without
any argument to get the line break. This looks unnecessary. And not all
the callers (including the ones in bootable/recovery) are following this
protocol when sending the ui_print command.
This CL simplifies the protocol to always print with a newline for
ui_print command. When updating from an old recovery with the new
updater, all the ui_print'd strings would appear in one line as a side
effect. But a) it would only affect the text-mode UI, which won't be
shown to users; b) log files won't be affected.
Bug: 32305035
Test: Apply an update with the new updater on top of an old and new
recovery image respectively.
Change-Id: I305a0ffc6f180daf60919cf99d24d1495d68749b
We could inject I/O faults during an OTA update for test purpose. But we
should skip the injection if the update is an retry. Otherwise the
update test will simply keeps failing.
Bug: 34159970
Test: Apply the same package on angler and the update succeeds on the 2nd try.
Change-Id: Id274e5475e3bc8d25d50a8cf61a77d2e32c569d6
Returning the parsed RangeSet directly (as opposed to using some pointer
parameter) to make the code cleaner.
Test: Apply an incremental with the new updater.
Change-Id: I8c99e701f189eb6a3eacc0d647e5a3a85fbeb3eb
Change the stash size computation from int to size_t.
Test: Apply an incremental BBOTA with the new updater.
Change-Id: Ib45b71b826fec6aa0ffafc67c17735825634eae0
We should always use unique_fd or unique_file to hold the FD or FILE*
pointer when opening via ota_(f)open functions.
This CL avoids accidentally closing raw FDs or FILE* pointers that are
managed by unique_fd/unique_file.
Test: recovery_component_test passes.
Change-Id: If58eb8b5c5da507563f85efd5d56276472a1c957
Add read_bootloader_message_from() and write_bootloader_message_to() to
allow specifying the BCB device (/misc).
Also add testcases for set_stage() and get_stage().
Test: recovery_component_test passes.
Test: Build a recovery image and apply a two-step OTA package.
Change-Id: If5ab06a1aaaea168d2a9e5dd63c07c0a3190e4ae
write_value(value, filename) writes 'value' to 'filename'. It can be
used to tune device settings when applying an OTA package. For example,
write_value("960000", "/sys/devices/system/cpu/cpu0/cpufreq/scaling_max_freq").
Bug: 32463933
Test: recovery_component_test passes.
Test: Apply an OTA package that contains a call to write_value(), and
check the result.
Change-Id: Ib009ecb8a45a94353f10c59e2383fe1f49796e35
'bool success = ExtractEntryToFile()' gives opposite result. Fix the
issue and add testcases.
Change the one-argument version of package_extract_file() to explicitly
abort for non-existent zip entry. Note that this is NOT changing the
behavior. Prior to this CL, it aborts from Evaluate() function, by
giving a general cause code. Now it returns kPackageExtractFileFailure.
BUg: 32903624
Test: recovery_component_test works.
Change-Id: I7a273e9c0d9aaaf8c472b2c778f7b8d90362c24f
Clean up SymlinkFn() a bit. Also clean up the temp files created when
running the tests; otherwise non-empty TemporaryDir won't be removed.
Test: recovery_component_test passes.
Change-Id: Id3844abebd168c40125c4dcec54e6ef680a83c3a
ReadArgs will switch to using std::string and std::unique_ptr. Also
cleanup the callers.
Test: mma & component test passed.
Change-Id: I4724406ae6c0c134a27bbd1cdd24ad5d343b2a3b
Clean up the duplicated codes that handle the zip files in
bootable/recovery; and rename the library of the remaining
utility functions to libotautil.
Test: Update package installed successfully on angler.
Bug: 19472796
Change-Id: Iea8962fcf3004473cb0322b6bb3a9ea3ca7f679e
Changing the field of 'Value' in edify to std::string from char*.
Meanwhile cleaning up the users of 'Value' and switching them to
cpp style.
Test: compontent tests passed.
Bug: 31713288
Change-Id: Iec5a7d601b1e4ca40935bf1c70d325dafecec235
- Remove dead declarations in expr.h: SetError(), GetError(),
ClearError().
- Remove the declaration of Build() out of expr.h.
- Use std::unordered_map to implement RegisterFunction() and
FindFunction(); kill FinishRegistration().
- Add a testcase for calling unknown functions.
Test: mmma bootable/recovery; recovery_component_test passes.
Change-Id: I9af6825ae677f92b22d716a4a5682f58522af03b
Also add a testcase for sha1_check().
Test: mmma bootable/recovery; recovery_component_test passes.
Change-Id: I4d06d551a771aec84e460148544f68b247a7e721
So that we can write native tests for updater functions. This CL adds a
testcase for getprop() function.
Test: mmma bootable/recovery; Run recovery_component_test on device.
Change-Id: Iff4c1ff63c5c71aded2f9686fed6b71cc298c228
This way we kill a few strdup() and free() calls.
Test: 1. recovery_component_test still passes;
2. Applying an update with the new updater works;
3. The error code in a script with abort("E310: xyz") is recorded into
last_install correctly.
Change-Id: Ibda4da5937346e058a0d7cc81764d6f02920010a
Also remove the 0xff comparison when validating the bootloader
message fields. As the fields won't be erased to 0xff after we
remove the MTD support.
Bug: 28202046
Test: The recovery folder compiles for aosp_x86-eng
Change-Id: Ibb30ea1b2b28676fb08c7e92a1e5f7b6ef3247ab
Clean up the recovery image and switch to libbase logging.
Bug: 28191554
Change-Id: Icd999c3cc832f0639f204b5c36cea8afe303ad35
Merged-In: Icd999c3cc832f0639f204b5c36cea8afe303ad35
We might end up in an infinite loop if read(2) reached EOF unexpectedly.
The problematic code in uncrypt mentioned in the bug has been fixed
by switching to libbase ReadFully(). So I grepped through the recovery
code and fixed some other occurences of the issue.
Bug: 31073201
Change-Id: Ib867029158ba23363b8f85d61c25058a635c5a6b
Check the results from applypatch in PerformCommandDiff; and abort the
update on failure.
Bug:29339536
Change-Id: I5087d79ba532b54250f4c17560524255c8a4fabc
If the update is a retry, ioctl(BLKDISCARD) the destination blocks before
writing to these blocks.
Bug: 28990135
Change-Id: I1e703808e68ebb1292cd66afd76be8fd6946ee59
Write error code, cause code, and retry count into last_install. So we
can have more information about the reason of a failed OTA.
Example of new last_install:
@/cache/recovery/block.map package name
0 install result
retry: 1 retry count (new)
error: 30 error code (new)
cause: 12 error cause (new)
Details in:
go/android-ota-errorcode
Bug: 28471955
Change-Id: I00e7153c821e7355c1be81a86c7f228108f3dc37
am: dd874b1c87
* commit 'dd874b1c87eb04f28db0db2629df0adde568a74c':
Add time and I/O info to last_install
Change-Id: I02aa858d5ce488d3acbf5400811e2565cf7d9c75
Remove O_SYNC from mzExtractRecursive() and PackageExtractFileFn().
These functions deal with extracting whole files from the update
package onto a filesystem. If run on ext4 on a rotating disk, for
example, the O_SYNC flag will cause serious performance problems
and the extraction proecss can take over 30 minutes, with no
obvious benefits.
This API function already calls fsync(fd) after each file is
extracted to ensure data and metadata is written to the underlying
block device, so the O_SYNC calls should be superfluous and safely
removable.
This change does not affect the OTA patch paths or any modification
of the bootloader partition or writes to other 'emmc' partitions.
Signed-off-by: Alistair Strachan <alistair.strachan@imgtec.com>
Change-Id: I9cbb98a98e6278bf5c0d7efaae340773d1fbfcd2
[1] switched a few things to android::base::unique_fd including
CommandParameters.fd. However, we were using memset(3) to zero out the
struct, which effectively assigned unique_fd(0) to fd. When it called
fd.reset(), file descriptor 0 was unintentionally closed. When FD 0 was
later reassigned via open(2), it led to lseek(2) errors: "Bad file
descriptor".
This CL switches to using braced-init (i.e. '= {}') instead, so that the
default constructor unique_fd(-1) would be called.
[1]: commit bcabd09293
Bug: 28391985
Change-Id: If1f99932b15552714c399e65c8b80550344b758a
Currently block_image_verify() stashes source blocks to /cache and
in some case triggers I/O errors. To avoid this risk, We create
a map from the hash value to the source blocks' range_set. When
executing stash command in verify mode, source range is saved but block
contents aren't stashed. And load_stash could get its value from
either the stashed file from the previous update, or the contents on
the source partition specified by the saved range.
Bug: 27584487
Bug: 25633753
Change-Id: I775baf4bee55762b6e7b204f8294afc597afd996
(cherry picked from commit 0188935d55)
Currently block_image_verify() stashes source blocks to /cache and
in some case triggers I/O errors. To avoid this risk, We create
a map from the hash value to the source blocks' range_set. When
executing stash command in verify mode, source range is saved but block
contents aren't stashed. And load_stash could get its value from
either the stashed file from the previous update, or the contents on
the source partition specified by the saved range.
Bug: 27584487
Bug: 25633753
Change-Id: I775baf4bee55762b6e7b204f8294afc597afd996
Cherry pick this patch because it fixes the problem that
a newed Value is released by free().
Bug: 26906416
Change-Id: Ib53b445cd415a1ed5e95733fbc4073f9ef4dbc43
(cherry picked from commit d6c93afcc2)
When I/O error happens, reboot and retry installation two times
before we abort this OTA update.
Bug: 25633753
Change-Id: Iba6d4203a343a725aa625a41d237606980d62f69
(cherry picked from commit 3c62b67faf)
If two libraries both use LOCAL_WHOLE_STATIC_LIBRARIES and include a same
library, there would be linking errors when generating a shared library
(or executable) that depends on the two libraries both.
Also clean up Android.mk files.
Remove the "LOCAL_MODULE_TAGS := eng" line for the updater module. The
module will then default to "optional" which won't be built until needed.
Change-Id: I3ec227109b8aa744b7568e7f82f575aae3fe0e6f
When I/O error happens, reboot and retry installation two times
before we abort this OTA update.
Bug: 25633753
Change-Id: Iba6d4203a343a725aa625a41d237606980d62f69
Add and register a function to check if the device has been remounted
since last update during incremental OTA. This function reads block 0
and executes before partition recovery for version >= 4.
Bug: 21124327
Change-Id: I8b915b9f1d4736b3609daa9d16bd123225be357f
(cherry picked from commit 30bf476559)
Add and register a function to check if the device has been remounted
since last update during incremental OTA. This function reads block 0
and executes before partition recovery for version >= 4.
Bug: 21124327
Change-Id: I8b915b9f1d4736b3609daa9d16bd123225be357f
We are already using O_SYNC and fsync() for the recursive case
(package_extract_dir()). Make it consistent for the single-file case.
Bug: 20625549
Change-Id: I487736fe5a0647dd4a2428845e76bf642e0f0dff
Output messages in log when recovery is attempted or succeeded during
incremental OTA update.
Change-Id: I4033df7ae3aaecbc61921d5337eda26f79164fda
(cherry picked from commit b686ba2114)
To accommodate new changes in N release, such as error correction [1]
and other potential changes to the updater.
[1]: commit 0a7b47397d
Change-Id: I4dd44417d07dd0a31729894628635a0aa1659008
Add block_image_recover function to rewrite corrupted blocks on the
partition. This can be attempted if block_image_verify fails.
Note that we cannot use libfec during block_image_update as it may
overwrite blocks required for error correction. A separate recovery
pass in case the image is corrupted is the only viable option.
Bug: 21893453
Change-Id: I6ff25648fff68d5f50b41a601c95c509d1cc5bce
We have the last line being empty as a result of
android::base::Split("a\nb\n"), which leads to "missing command"
warnings in the update. Just skip all the empty lines.
Bug: 24373789
Change-Id: I5827e4600bd5cf0418d95477e4592fec47bbd3a9
When processing ui_print commands in the updater, it misses a line break
when printing to the recovery log.
Also clean up uiPrintf() and UIPrintFn() with std::string's.
Change-Id: Ie5dbbfbc40b024929887d3c3ccd3a334249a8c9d
Replace C-string with std::string, pointers with references, and
variable-size arrays in struct with std::vector.
Change-Id: I57f361a0e58286cbcd113e9be225981da56721b2
There is an integer overflow when the size of system goes beyond the
signed int limits. Hence changing pos to size_t.
Change-Id: I6e5e1b2f0e72030b30a6df09a01642f4c82abc79
Currently the fsync() inside write_all() may be called multiple times
when performing a command. Move that to the outer loop and call it
only after completing the command.
Also remove the O_SYNC flag when writing a stash.
Change-Id: I71e51d76051a2f7f504eef1aa585d2cb7a000d80
This reverts commit b65f0272c8.
It slows down the update too much on some devices (e.g. increased
from 8 mins to 40 mins to take a full OTA update).
Bug: 22129621
Change-Id: I016e3b47313e3113f01bb4f8eb3c14856bdc35e5
(cherry picked from commit 7125f9594d)
A RangeSet has half-closed half-open bounds. For example, "3,5" contains
blocks 3 and 4. So "3,5" and "5,7" are actually not overlapped.
Bug: 22098085
Change-Id: I75e54a6506f2a20255d782ee710e889fad2eaf29
Due to observed BLKDISCARD flakiness, overwrite blocks that we want
to discard with zeros first to avoid later issues with dm-verity if
BLKDISCARD is not successful.
Bug: 20614277
Bug: 20881595
Change-Id: I4f6f2db39db990879ff10468c9db41606497bd6f
Due to observed BLKDISCARD flakiness, overwrite blocks that we want
to discard with zeros first to avoid later issues with dm-verity if
BLKDISCARD is not successful.
Bug: 20614277
Bug: 20881595
Change-Id: I0280fe115b020dcab35f49041fb55b7f8e793da3
Although stdout and stderr are both redirected to log file with no
buffering, we are seeing some outputs are mixed in random order.
This is because ui_print commands from the updater are passed to the
recovery binary via a pipe, which may interleave with other outputs
that go to stderr directly.
In recovery, adding ui::PrintOnScreenOnly() function to handle
ui_print command, which skips printing to stdout. Meanwhile, updater
prints the contents to stderr in addition to piping them to recovery.
Change-Id: Idda93ea940d2e23a0276bb8ead4aa70a3cb97700
In the block updater, if BLKDISCARD fails, the error is silently
ignored and some of the blocks may not be erased. This means the
target partition will have inconsistent contents.
If the ioctl fails, return an error and abort the update.
Bug: 20614277
Change-Id: I33867ba9337c514de8ffae59f28584b285324067
These are already getting libc++, so it isn't necessary. If any of the
other static libraries (such as adb) use new or delete from libc++,
there will be symbol collisions.
Change-Id: I55e43ec60006d3c2403122fa1174bde06f18e09f
