Current cow size estimation is computed by taking the offset of next
data write. However, during COW size estimation we repeatdly increment
op_count_max. Since data section is placed right after op section,
incrementing op_count_max has the effect of moving the beginning of
data section. next_data_pos_ is never updated in this process, causing
the final estiamte to be off.
Test: th
Bug: 313962438
Change-Id: I250dff54c470c9c20d6db33d91bac898358dee31
Since currently implementation just does an bitwise or with the
storage unit, multiple calls to set_source would result in overlapping
bits. Fix by first clear the existing storage.
Test: th
Bug: 313962438
Change-Id: Iecfe8dd244c0f65ecd3cacb0404fdc39ef836d97
Allow O_DIRECT reads on source block device.
This will further cut down the Active and Inactive file pages
during partition verification.
On Pixel 6 after incremental OTA - Post OTA reboot:
Without patch With patch Delta
--------------------------------------------------------
Inactive(File): 4992MB 3887MB ~22%
Active(File): 1465MB 1014MB ~30%
Boot time however increases from 25 to 30 seconds.
This is not yet enabled. This will be behind a sysprop flag
or for low memory devices and will be enabled later.
Additionally, set the priority of worker threads to normal.
Merge threads priority is reduced. This will help low memory
devices as tested on Pixel watch.
Bug: 311233916
Test: OTA on Pixel 6
Change-Id: Icacdef08d68e28d3062611477703e7cf393a9f10
Signed-off-by: Akilesh Kailash <akailash@google.com>
Scanning of partitions post OTA leads to memory pressure. Tune
the read-ahead of source and COW block device. This is currently
set to 32KB.
This reduces Inactive(file) and Active(file) usage during entire
duration of boot post OTA.
On Pixel 6: For incremental OTA ~400M. During boot:
Without-patch With-patch Delta
--------------------------------------------
1: Peak Inactive(file): 4469MB 3118MB ~30%
2: Peak Active(file): 985MB 712MB ~27%
No regression observed on boot time.
Additionally, cut down the number of threads to verify the partitions.
Bug: 311233916
Test: Incremental OTA on Pixel 6
Change-Id: I0b842776c36fa089c39c170fa7bf0f246e16636d
Signed-off-by: Akilesh Kailash <akailash@google.com>
This does not change the default cow version. Currently v2 is still the
default. This CL only enables OEMs to set virtual_ab_cow_version to 3
for testing purposes.
Test: th
Bug: 313962438
Change-Id: I7a328fa32283560a48604ffe02edd2551ac49a83
We can shove type into source info to save 8 bits in per cow operation.
We only need 4 bits inside of source_info to enumerate all the types of
Cow Operation. Since CowOperationV3 is not used on disk(just yet) , we
can make format changes.
This CL is mechanical:
1. Remove tye .type field from CowOperation struct
2. Add a type() getter method to CowOperation struct
3. Replace all existing usage of `type` member with the new getter
No functional changes, just refactorings.
Test: th
Bug: 304602386
Bug: 313962438
Change-Id: I85d89c71fc6afede12ea299a4a3e3b2184ea2d8b
Sync writes after every 1MB instead of flushing at the end.
Bug: 299011882
Test: Boot device off snapshots
Change-Id: If91168ec92c2b2995bdf296ea1c7d4c261b12411
Signed-off-by: Akilesh Kailash <akailash@google.com>
There's a bug previsouly where we compare return value of
GetCowOpSourceInfoData() with CowOperationType. Such bugs are possible
because cow operation enums are weakly typed integers. Turn
CowOperationType into strongly typed enum to prevent such bugs.
Test: th
Bug: 304602386
Change-Id: If6941a4740c374ed066cf0aee9e52f4df05a9b38
GCC 14 starts to no longer include <algorithm> by default, resulting in
it needing to be explicitly declared.
Test: Recompiled with GCC 14 and succeeded
Change-Id: Ifc5dd58b7476ba728ae604cd2924cb68fbcab701
Signed-off-by: Christopher Fore <csfore@posteo.net>
Simplify the logic of MapScratchPartitionIfNeeded to just map the
scratch partition unconditionally. If there is a "scratch" logical
partition available, we just assume it must be needed later to
facilitate overlayfs.
Going through the fstab & fs_mgr_overlayfs_candidate_list to determine
if overlayfs is needed or not might be inaccurate as the fstab passed into
MapScratchPartitionIfNeeded is incomplete (it includes only the
"first_stage_mount" entries). So the scratch device might not get created
even though there are some "late mount" partitions that need the scratch
device to facilitate overlayfs.
Bug: 306124139
Test: adb-remount-test
Change-Id: I8590750e822219dec3d7d3c8648e9c9a5a32f68c
Right now there is a bug in second-stage-init that screws up the
overlayfs overrides. This happens because:
1. second-stage-init mount_all might be executed in the "bootstrap"
mount namespace.
2. In order to move (MS_MOVE) submounts in fs_mgr_overlayfs_mount_all(),
we change the mount propagation type of overridden filesystems to
MS_PRIVATE.
3. This means that the "default" mount namespace would not receive the
mount events of the overlayfs overrides.
4. After /data is mounted, init would switch to the "default" namespace.
This means any new processes spawned after this period would not be
able to see the overlayfs overrides.
We fix this by changing the mount order of second-stage-init mount_all
to mount the overlayfs override of a partition immediately after the
partition is mounted. This way we don't need to move any submounts as
there can't be any, thus we don't need to set any mountpoint to
MS_PRIVATE so the mount event of the overlayfs would be propagated to
the "default" mount namespace, too.
Bug: 309407859
Bug: 306124139
Test: adb-remount-test
Test: verify that overlayfs tookover successfully from second-stage-init
Change-Id: If2ff4852516292ccbc7abdeebe0e9a7c1c7ba606
Need to support exfat fs type point mount in recovery mode
Bug:312674538
Test: SIU (Apply update from SD card) in recovery mode
Change-Id: I2735e5068c47b183a25d2305bca89f08de1e32c7
v3 writer to write sequence data. Sequence data will be written after
the scratch space and before the resume space. Since this is just a list
of integers, writing and reading should be trivial
Test: cow_api_test
Change-Id: If3b6b1cfa155aeb65bf693263fc373154ba8e81d
Op count should be set before we sync the header. This way subsequence
writers can initialize with the correct op buffer size
Test: cow_api_test
Change-Id: I56a0d747b3f2a1d9d582d8f9d643b81cbdd9b8d7
After we write emit a label, we need to update the number of resume
points + sequence data and op_count. Realistically we could just call
Finalize, but maybe synching these specific fields could prevent
unexpected outcomes.
Test: cow_api_test
Change-Id: I1585601a134221689ce8d5675a2a3e32f1e8a0e6
This passes the block size to mkfs. Currently f2fs must have page size
equal to block size. If this changes, we may want to revisit this.
Change-Id: I1ecdffcb2a271aa80c0bd07fd095941720fb48d0
Signed-off-by: Daniel Rosenberg <drosen@google.com>
Bug: 279820706
Test: Boot 16K device
Currently, f2fs block size must be equal to page size.
If this is not the case in the future, this code should be revisited.
Bug: 279820706
Test: adb remount on 16k f2fs device
Change-Id: Ib0d1ea81f921be99471d0433c8499ae077208db7
Signed-off-by: Daniel Rosenberg <drosen@google.com>
Framework or applications can try to find userdata dm target from "/proc/mounts"
which gives "/dev/block/mapper/userdata".
That causes breaking the assumption in all the places.
Bug: 311165039
Change-Id: I9ea1b4589cdd52021d9807f7240c2e4b6d6d05ef
Signed-off-by: Jaegeuk Kim <jaegeuk@google.com>
No need partition verification when device boots on snapshot without
slot switch.
This also saves couple of seconds of boot time.
Bug: 299011882
Test: Boot device on snapshot, OTA on Pixel
Change-Id: I5b781de7e0f745bbfe9646f88ca912139b2d853e
Signed-off-by: Akilesh Kailash <akailash@google.com>
Since these functions are used across both parser and writer, updating
it as inline functions in cow_format.
Test: cow_api_test
Change-Id: I9824684e3b9b48947accce935335d4019d745ae0
If any of the read-only partitions are mounted off dm-user
then certainly update is in-progress.
Bug: 308900853
Test: OTA on Pixel, reboot during OTA.
Change-Id: I36121e1d99ec7c1f1110a65fc67996190875af18
Signed-off-by: Akilesh Kailash <akailash@google.com>
We want to add a resume_point_count in the header to represent how many
resume points we've written. In the case that we've written less than
resume_buffer_size, we only want to read the valid resume points.
without these changes incremental OTA runs into segfault or have faulty
data when trying to FindResumeOp() as our resume points contain invalid
entries
Test: full ota followed by inc ota on cuttlefish
Change-Id: I0a8971955439639f2d0f39d9d518c1145ae15c3d
Update FindResumeOp to take in a uint64_t to match the value of the
caller function
Test: ota with following CL
Change-Id: I19568d119b7ebd75ea9e98970b311ae7da92ff0e
Cow size should just be wherever the last data position is written. In
v3 we no longer have a footer, so this calculation is simple. This
function is used by cow estimator
Test: cow_api_test
Change-Id: I08204e2d560b120450019a529baa41de9b8e66d5
