Age | Commit message (Collapse) | Author |
|
Test: Manual - Tested all audio scenarios and multiple underrun cases
Bug: 240532906
Bug: 240455340
Bug: 234755551
Bug: 243239118
Signed-off-by: Florian Muller <mullerf@google.com>
Change-Id: I5d2121a03599ab3dc9660185bca550964dfda292
|
|
Re-enable the force kernel panic button combo to get status
of all the threads in instances where userspace seems hung,
but the kernel is fine.
Enable : echo 1 > /sys/devices/platform/dss/dss_forced_panic
Disable: echo 0 > /sys/devices/platform/dss/dss_forced_panic
Bug: 244210947
Test: Verify kernel panic with button combo when enabled
Signed-off-by: Amit Bhanagay <amitbhanagay@google.com>
Change-Id: I8a2782aa5bf6d8a2ba6b375ceb521eab0924a226
|
|
Don't install dev2/proto1/proto11 overlays or modules
Bug: 225204431
Signed-off-by: Ben Fennema <fennema@google.com>
Change-Id: I3d1b9d9d180146229ec1f1a2e07d8bdb4f54c5d2
|
|
Add freq times to kernel config. This adds /proc/self/time_in_state
which is used by batterystats.
Bug: 226383071
Test: Manually verify that file exist
Change-Id: I028b45f1d1ee81ef55294060385eebaf70b80590
|
|
These have already been removed from the repo manifest, so these paths
are now invalid.
Bug: 141693040
Change-Id: I6d33e58b79db1826e010e3a6c6d3858e256d31c5
|
|
Bug: 186227605
Change-Id: I935a0e263316dfb9cb51c0f09cf2d81d0fd57e9d
|
|
Signed-off-by: Tim Joines <tjoines@google.com>
Test: manual
Bug: 189154470
Change-Id: Ie7a1d5216be2f5d8d6cdf94c862d21f7a42475f0
|
|
Build a module to register a persistent clock for better timekeeping.
Bug: 208709092
Change-Id: I423f9d83ff368ec1f0219834f709ed509af91fb0
|
|
Bug: 210543176
Signed-off-by: Ben Fennema <fennema@google.com>
Change-Id: Ib577f8a634a45360389279ab8418f96796f6e636
|
|
CONFIG_FORTIFY_SOURCE is force disabled when building with clang, due to
bugs.
Bug: 209527035
Signed-off-by: Ben Fennema <fennema@google.com>
Change-Id: I535598b0f0b7f1f3d8f81b28a855738c39165409
|
|
rename CONFIG_EXYNOS_THERMAL to CONFIG_EXYNOS_THERMAL_V2 and add
CONFIG_EXYNOS_CPU_THERMAL.
Bug: 209527035
Signed-off-by: Ben Fennema <fennema@google.com>
Change-Id: I6dd0a556318da7073bd540c3895e36fe4e496bd8
|
|
Test: Manual
Bug: 179747307
Signed-off-by: Florian Muller <mullerf@google.com>
Change-Id: Ic8b0ba3d89d37a92faed913a3076e9c92da7964b
|
|
Update config per security penetration testing findings
Bug: 194888899
Bug: 195611517
Signed-off-by: Ben Fennema <fennema@google.com>
Change-Id: I94071d975a9e12e3b6aa40abbb43f63586d95703
|
|
To pass CtsSecurityHostTestCases, the static user-mode helper must be
enabled, and it must be set to an approved path. The default path,
/sbin/usermode-helper, is not approved and does not exist on this
platform. A blank string disables user-mode helpers entirely when static
user-mode helper support is enabled, which is also acceptable.
Bug: 204310674
Test: Verified that testConfigDisableUsermodehelper passes
Change-Id: Id79c7be962dfd4edb02eae4b9a8633283463016a
|
|
Bug: 206336101
Test: Pass CtsIncrementalInstallHostTestCases
Change-Id: Ic97c375f7db481fc4009ecec3d6d9692e5952ac5
|
|
Causes selinux denials:
[ 14.131050] [1: logd.auditd: 2773] type=1400 audit(1635557861.076:5): avc: denied { create } for comm="kdevtmpfs" name="dm-6" scontext=u:r:kernel:s0 tcontext=u:object_r:device:s0 tclass=blk_file permissive=0
Bug: 204938130
Signed-off-by: Ben Fennema <fennema@google.com>
Change-Id: I3cff6d43f15308df880d69331814d33ca0c23999
|
|
From kernel documentation:
In other words, when you disable this, memory resource controller has no
cares to usage of swap...a process can exhaust all of the swap
Test: manual
Bug: 190411761
Change-Id: Ic96bbea4f831f074c9cbd1e1ba80b2a0675f9393
|
|
This reverts commit cbb156e62d22e88715f14c206f36822e10e22bb6.
Reason for revert: This is no longer needed.
Change-Id: I2a40a55960c68551de1ca079fb4aeefd77450e66
|
|
Bug: 196439607
Signed-off-by: Ben Fennema <fennema@google.com>
Change-Id: Id66e4b3d74b002bf5beabbd1b9337f74547e357c
|
|
Support was removed for CLANG_TRIPLE in ASB-2020-12-05_4.19-stable.
Update CROSS_COMPILE and CROSS_COMPILE_COMPAL triple to match
build.config.aarch64
Bug: 196439607
Signed-off-by: Ben Fennema <fennema@google.com>
Change-Id: Icb10788268196ec74e545ac5547d7f5009e160c0
|
|
Bug: 196439607
Signed-off-by: Ben Fennema <fennema@google.com>
Change-Id: I8694573c7eacd686333c22afc232834b20341b40
|
|
clang-r383902 is avaliable in both android11-dev and
master-kernel-build-2021
Bug: 196439607
Signed-off-by: Ben Fennema <fennema@google.com>
Change-Id: Ib0a7c15094eb53400b0ef6f9b7bc59a541b53388
|
|
This config option is not needed
Bug: 196439607
Signed-off-by: Ben Fennema <fennema@google.com>
Change-Id: Ibe8af8d146459bbb1279bca170c73040a2a0ee2a
|
|
F2FS needs CONFIG_UNICODE for supporting casefold feature
(external_storage.casefold.enabled=1).
Bug: 199446880
Change-Id: I96c5445a79c69f45cb2396f16e66ec03a291d4e3
|
|
FAIR_SHARE works better with what we want to accomplish with multiple
trip points.
Bug: 187884905
Change-Id: Ib3c06effd6fe96647e8e33c46da6a8400f45e9d8
|
|
Bug: 184787925
Signed-off-by: Ben Fennema <fennema@google.com>
Change-Id: I56617b62009a665a9fe5c43b6949a1db6700026e
|
|
This is to allow later on in the userspace to either load cfg80211.ko
or not depending on whether we are in charger mode or not.
Test: Manual
Bug: 193050226
Signed-off-by: mullerf <mullerf@google.com>
Change-Id: Id240436c1fcc894e681da8e626986543ca72a792
|
|
Test: Manual
Bug: 179747307
Signed-off-by: mullerf <mullerf@google.com>
Change-Id: Ia4d21752b20b56eac60a572035079f49cc542f27
|
|
This change is support only, and a non-op. KSM needs to be enabled in an
init file later to make this functional.
Test: manual
Bug: 194415841
Change-Id: If5bc2c1e0f99f875286439cfeca3504bfe1517ff
|
|
Bug: 186659701
Change-Id: Ifbb126192abc05cb163870660ac4eb4cad30e6e1
|
|
Signed-off-by: Euiyoul Ryu <ryu.real@samsung.com>
Change-Id: Ice2393ea6097ecededc72583d804b3e30fc2887d
|
|
Under the v4.9 kernel, the log size was set this to 2^20 = 1 MiB but,
as part of the upgrade to v4.19, it was reduced to the 2^17 = 128 KiB
default. Increase it back up to 2^19 = 512 KiB so the entire boot up
log can be captured and saved into the persistent logcat section of
bugreports.
Change-Id: Id3f6a03c8dda3c1a51ac08b208c4a9fe0886a7af
|
|
Bug: 190677341
Change-Id: I30e47e711902645817a769b6df60f6477081b1a5
|
|
Bug: 188003517
Signed-off-by: Ben Fennema <fennema@google.com>
Change-Id: I174eca9335269bd12ea5fbc0424392d37636cccb
|
|
The nanohub module needs to be built before other modules which depend
upon its symbols or else there will be depmod warnings. Fortunately,
the sound is after sensorhub alphabetically so just order the modules
that way.
Bug: 188087487
Change-Id: I552f857b4a6c70fdb376cc1bc348f1e27b923c84
|
|
Signed-off-by: Matthew Bouyack <mbouyack@google.com>
Bug: 185396981
Change-Id: I6fa516b0c8602f8a826fe7e904e03bfd38cf0954
|
|
(with the patch decon0 takes a lot of CPU so needs investigation)
This patch adds a compat compiler that enables
the VDSO generation for 32bit
VDSO improves the time functions by up to 30x
Pre patch performance:
130|r11:/data/local # ./bionic-benchmarks --benchmark_filter=BM_time_clock*
2021-05-04 18:34:37
Running ./bionic-benchmarks
Run on (2 X 1144 MHz CPU s)
***WARNING*** CPU scaling is enabled, the benchmark real time measurements may be noisy and will incur extra overhead.
---------------------------------------------------------------------------------
Benchmark Time CPU Iterations
---------------------------------------------------------------------------------
BM_time_clock_getres 1434 ns 1433 ns 488235
BM_time_clock_getres_BOOTTIME 1435 ns 1434 ns 487981
BM_time_clock_getres_MONOTONIC_COARSE 1459 ns 1458 ns 480207
BM_time_clock_getres_MONOTONIC_RAW 1448 ns 1441 ns 488421
BM_time_clock_getres_REALTIME 1436 ns 1434 ns 488180
BM_time_clock_getres_REALTIME_COARSE 1458 ns 1458 ns 480123
BM_time_clock_getres_syscall 1413 ns 1413 ns 495407
BM_time_clock_gettime 1736 ns 1736 ns 403663
BM_time_clock_gettime_BOOTTIME 1746 ns 1745 ns 401205
BM_time_clock_gettime_MONOTONIC_COARSE 1485 ns 1484 ns 471822
BM_time_clock_gettime_MONOTONIC_RAW 1715 ns 1714 ns 408407
BM_time_clock_gettime_REALTIME 1715 ns 1714 ns 408397
BM_time_clock_gettime_REALTIME_COARSE 1459 ns 1459 ns 479789
BM_time_clock_gettime_syscall 1711 ns 1711 ns 407826
post patch:
---------------------------------------------------------------------------------
Benchmark Time CPU Iterations
---------------------------------------------------------------------------------
BM_time_clock_getres 41.4 ns 39.9 ns 17566596
BM_time_clock_getres_BOOTTIME 39.9 ns 39.8 ns 17665970
BM_time_clock_getres_MONOTONIC_COARSE 40.6 ns 40.6 ns 17216056
BM_time_clock_getres_MONOTONIC_RAW 40.3 ns 39.8 ns 17607225
BM_time_clock_getres_REALTIME 39.7 ns 39.6 ns 17548755
BM_time_clock_getres_REALTIME_COARSE 40.5 ns 40.5 ns 17288572
BM_time_clock_getres_syscall 1417 ns 1415 ns 495227
BM_time_clock_gettime 1765 ns 1763 ns 395992
BM_time_clock_gettime_BOOTTIME 1775 ns 1774 ns 393764
BM_time_clock_gettime_MONOTONIC_COARSE 75.7 ns 75.6 ns 9251004
BM_time_clock_gettime_MONOTONIC_RAW 1802 ns 1801 ns 388419
BM_time_clock_gettime_REALTIME 1744 ns 1744 ns 401373
BM_time_clock_gettime_REALTIME_COARSE 75.6 ns 75.6 ns 9254838
BM_time_clock_gettime_syscall 1705 ns 1705 ns 409374
Still need to investigate the gettime_BOOTTIME/REALTIME/MONOTONIC_RAW to
see if it's expected to not gain performance
Bug: 187240443
Test: local on device
Change-Id: I11b6b1be0ccc02a5d3b34fd0904b8367858cf43f
|
|
Test: Manual
Bug: 182553618
Signed-off-by: mullerf <mullerf@google.com>
Change-Id: Ic927eb37f6ad23fd34c5c5d412fea1989abbbabf
|
|
Bug: 185487453
Signed-off-by: Seungbin Lee <seungbin.lee@samsung.com>
Change-Id: Ifbf66236be77d4bdc827333f6a05189d977d309d
|
|
- Disable debug logging for the emmc driver
to both improve perf (~5%) and reduce mem usage (tiny)
Test: local build
Bug: 184900231
Change-Id: Id538b66c2d61abb46df55f421fa1edd2975d1087
|
|
- Reduce the amount of log lines
saved by the debug snapshot
Bug: 184900231
Change-Id: I42f9608c84fa98767ab9786432d782b88df1e92a
|
|
Bug: 183993769
Test: Run r11 and check /sys/block/zram0/backing_dev is populated.
Change-Id: I107a11865311052a189b8697b27ef41924f5362c
|
|
Build it as a module instead
Bug: 183969928
Signed-off-by: Ben Fennema <fennema@google.com>
Change-Id: I2ba689dceff6d4e754f8dea9bdc1eac067273fa5
|
|
CDT entry 1 2 1 1 1 2 is dev2 btwifi BOE (see go/r11-cdt)
Test: Manual
Bug: 183670289
Signed-off-by: mullerf <mullerf@google.com>
Change-Id: I25db3cb295831229ec6f746131aa16578d0d47ae
|
|
Bug: 182918181
Test: local boot of dev2
Change-Id: Id5025932c30188b2f2fd3d0b85c9d0b2c0308e9c
|
|
Bug: 180940282
Signed-off-by: Ben Fennema <fennema@google.com>
Change-Id: Ibfab4c64223c73a9c9c92eb947d7811b708784e2
|
|
The dev board sent to Lassen with alternate audio routing has AUO
display.
Bug: 181819710
Change-Id: Ie2dc0a2e42c26edef2705cf973d1486b5931fdbb
|
|
On Android R, CONFIG_IP6_NF_NAT and subsequently the config flags that
rely on it are not allowed, causing build errors.
Test: `adb shell downloader` is still working in both "WiFi" and "BT
proxy only" cases
Bug: 160151746
Signed-off-by: mullerf <mullerf@google.com>
Change-Id: Ie7a08a013e4db63830533f458b89cb0e7a6b2446
|
|
Bug: 177103103
Change-Id: Ifc7c202345cc25cdf8c15f41916af15934671155
Signed-off-by: Payam Moradshahi <payamm@google.com>
|
|
Android expects CPUSETS to be supported by the kernel and will attempt
to mount its virtual filesystem at /dev/cpusets and then configure it.
If the kernel doesn't support CPUSETS then Android ends up writing to
the /dev tmpfs which means only files explicitly written are present
which confuses systrace as it expects certain files to always exist in
each cpusets subdirectory.
Bug: 180046865
Change-Id: I9858580420d3a3f71eca81e62db561bd1f795c51
|