Per-application memory.stat files are not available when per-application
memcgs are not used (per_app_memcg=false). Disable its usage based on
ro.config.per_app_memcg property.
minchan:
* correct indentation of memory_stat_parse
* move per_app_memcg check into memory_stat_parse inside
* change low_ram_device to per_app_memcg
Bug: 110384555
Test: manual test to see lkmd log message with memory hogger
Merged-In: Ib6dd7586d3ef1c64cb04d16e2d2b21fa9c8e6a3a
Change-Id: Ib6dd7586d3ef1c64cb04d16e2d2b21fa9c8e6a3a
Signed-off-by: Minchan Kim <minchan@google.com>
Signed-off-by: Suren Baghdasaryan <surenb@google.com>
This code was unreachable, since all of the callers were calling
register_socket_transport with foo.c_str() as the serial. Lift this
assumption into the type system by switching from char* to std::string
for the argument type.
Bug: http://b/112147760
Bug: https://www.viva64.com/en/b/0579/
Test: mma
Change-Id: I5a6ee265feee6b83bc933a64d895eed39fce68e7
An accidental side-effect of the change to unique_fd was that we
started closing the reboot service socket before pausing and waiting
for our death. This results in the client immediately returning, which
means they might get a chance to run other commands before we actually
manage to reboot.
Bug: http://b/112104037
Test: adb reboot; adb devices
Change-Id: Id61321da0c08ebd1ab57f1ce926bafd9109dd9c2
DM_TABLE_LOAD will reject dm-linear entries if their size is not a
multiple of the backing device's logical block size. For example, a
partition of 10GiB+512 bytes will fail to map in device-mapper if the
logical block size is 4096 bytes. To address this, this patch adds a
few changes to liblp:
The block size given to lpmake is now recorded in LpGeometryMetadata.
The block size must be a multiple of the sector size. In addition,
partiton sizes are now aligned to the block size, and the super
partition must have enough free space to allocate at least one block (in
addition to storing metadata).
GrowPartition now has multiple checks that the block-size invariant is not
violated, to ensure that no invalid partition tables will be created.
Bug: 79173901
Test: liblp_test gtest
Change-Id: I484aac1f9b90ebd92dc1c89ce1e09cd89bbb441e
The partition resize algorithm duplicates a lot of logic because it
handles the final free interval separately from other free intervals.
This is unnecessary and makes it harder to change the actual algorithm.
This change makes GrowPartition() treat the final free space region the
same as free gaps in between partitions. It does this by converting the
extent list into a gap list, and then adds a final gap for the remainder
of the free space. The resize function no longer has to treat the end of
the disk separately.
This patch does not change the way partitions are allocated, it is
purely a refactoring.
Bug: 79173901
Test: liblp_test gtest
Change-Id: I4780f20b23fe021eac62de874b061857712c04fe
Add android_log_parser_reset and android_log_reset to reset reader/writer
log context. This allows to generate multiple log messages of the same
format without context reallocation.
Change-Id: Icb0f15855378aa4e66fc671ffefd15d9df6fee1e
This change is to assist with implementing the fastbootd "resize-partition"
command. The GrowPartition and ShrinkPartition functions are now
private.
Bug: 78793464
Test: N/A
Change-Id: Ic66a3052359e2558663272ff6e014704206b197e
Some devices have modules.alias and modules.dep for modprobe and other
purposes but do not want to opt into ueventd auto loading their
modules. Therefore we add a flag that can be added to ueventd
configuration files to opt into this behavior.
Bug: 111916071
Bug: 112048758
Test: check that modules are loaded with this opt-in
Test: check that modules are not loaded without this opt-in
Change-Id: Ifb281b273059b4671eea1ca5bc726c9e79f3adfb
* changes:
adb: disable ReconnectHandler in adbd.
adb: fix error message for `adb {forward,reverse}`
adb: more immediately try to reconnect connections.
adb: don't pass time_point::max to condition_variable::wait_until.
adb: move list-forward, kill-forward back into handle_forward_request.
Previously, when a TCP connection was disconnected from adbd, we were
registering it with ReconnectHandler, which led to the transport
sticking around after the socket was closed. Due to the naming of
TCP transports in adbd (host-<fd number>), this results in incoming
connections being immediately closed if their file descriptor number
ends up being the same as a TCP transport that had previously
disconnected.
Guard all of the reconnect logic with ADB_HOST, to fix this.
Bug: http://b/112054041
Test: while true; do adb connect <device>; adb connect <device>; adb shell true; done
Change-Id: Ib55d304d7e07d6d744e8321d34671bb6d4b91afe
This CL is largely and adaptation of
Ifd119650dd3316508870df0dfc770099e95ae1d1
Bug: 80741439
Test: Successfully built product-services.img with one module in it, and
flashed on device. Also successfully built image with
/system/product-services and no dedicated /product-services partition.
Change-Id: I204a831e052d018018ab124d70ad1ff1610ee007
Most disconnects we're likely to encounter are cases where either we
notice immediately and can start reconnecting almost immediately (adbd
restarting because of `adb root`, etc.), or where we won't notice for a
while anyway, so a 10 second sleep is somewhat meaningless.
Test: adb root; time adb wait-for-device shell
Change-Id: I18e9213dc4e84d735e9240118a368dcb38f21c78
libstdc++ implements wait_until by calculating the offset between its
default clock and the clock that it's given by calling now() on each
and subtracting, and then adds that offset to the time_point argument.
When time_point::max is used, this overflows, resulting in the
reconnection thread spinning.
Test: wine adb.exe server nodaemon
Change-Id: Ife58f0aad14bc44c0804483d3ff2351c28b3d576
The daemon-side reverse functions depended on handle_forward_request:
move them back instead of duplicating the logic we had in
handle_host_request. Accomplish what we originally wanted to do in this
change by changing the transport argument of handle_forward_request to a
std::function that acquires a transport, either via
acquire_one_transport or immediately returning a value that we already
have.
As a side effect, fix a bug where we would emit spurious errors for host
service requests.
Bug: http://b/112009742
Test: echo "001chost:connect:127.0.0.1:5555" | nc localhost 5037
Test: python test_device.py
Test: python test_adb.py
Change-Id: Iccc555575df6dbd7de10382854c4ea2c6f4beeaa
commandline.cpp includes commandline.h which defines
StandardStreamsCallbackInterface which calls fwrite(). In the
compilation unit for commandline.cpp, fwrite is getting remapped to
adb_fwrite (good), but the compilation unit for bugreport.cpp includes
sysdeps.h pretty late, which prevents fwrite from getting remapped to
adb_fwrite.
Apparently when linking, the version of StandardStreamsCallbackInterface
that gets used is the one from bugreport.cpp's compilation unit, which
doesn't call adb_fwrite().
And it's necessary to call adb_fwrite() to get Unicode console output on
Windows.
The fix is to #include "sysdeps.h" earlier in bugreport.cpp. I searched
the other object files for other unremapped calls to printf/fwrite/etc.
and didn't find any.
Bug: https://issuetracker.google.com/issues/111972753
Test: mma, manual test on Windows 10 x64
Change-Id: I322dff75a878397f5e10227e746e77b0024129d0
Signed-off-by: Spencer Low <CompareAndSwap@gmail.com>
The fs_has_shared_blocks function had two problems. One, it called
statfs() on the mount point, which will return information about the
file system the mount point is on, rather than the filesystem mounted at
that location. Second, the check for EXT4_SUPER_MAGIC was inverted.
Bug: N/A
Test: adb remount -R works on an ext4 deduplicated device
Change-Id: I2e5ef895ea274cb7cc3c35295120da90a026d0d2
When host fastboot sends sparse blocks to the device, it tries to only
send blocks in multiples of 1024 bytes. If a block is not aligned to this
size, the excess bytes are prepended to the next write operation. This
is implemented by doing the write in two steps: first the previous
excess from the last write (plus new data up to alignment), then a
second write for the aligned remainder of the new data.
This logic has a bug if the final block plus the previous excess data
contains >= 1024 but < 2048 bytes. In this case the first write will
drain 1024 bytes from the data, and the second write will not have 1024
bytes to write. Instead of retaining this data for the next write, it
tries to write 0 chunks (and thus 0 bytes), which hangs the ioctl() call.
Bug: N/A
Test: "fastboot flash super super.img" where super.img is generated by
lpmake, containing system and product_services partitions and
images.
Change-Id: I9e8523c976ec84d5a57b36a28f4b1ca800edb7e7
This change adds a `chmod 0555 /acct` just after mounting the cgroup fs
on it. This makes it such that even on systems where the rootdir+system
partitions are shared, the permissions allow for the system user to open
files within /acct, which in turn makes libprocessgroup not complain
that it can never do anything with those files.
Bug: 111996377
Test: adb shell 'ls -ldZ /acct' # Before and after, on an aosp_sailfish
dr-xr-xr-x 48 root root u:object_r:cgroup:s0 0 1971-12-29 17:41 /acct
Change-Id: Iee6531126c6e81aa7794e44500555bd3c1cdf7b8
Log an error if minui init fails, and avoid subsequent calls to minui
in this case (which are likely to crash).
Warn if no system text font available. This is now expected on most
devices, and the fallback text is not expected to be needed, but just
in case.
Avoid the attempt to print text if no system font available, log
warnings instead.
Bug: 80249440
Test: manual: charger mode with no system font
Change-Id: Ib0c761d37e41a893b2a192438eebcf48c4827049
Merged-In: Ib0c761d37e41a893b2a192438eebcf48c4827049
(cherry picked from commit 8211515fbe)
Original code is unblank screen for all frames of the first
cycle, unblank is needed only before the first frame of the
first cycle.
Bug: 74771887
Change-Id: I51d8390e359335a0b1cb127b096db1a8b889b169
Signed-off-by: Thierry Strudel <tstrudel@google.com>
