OSDN Git Service
Steven Moreland [Fri, 23 Jun 2017 17:23:41 +0000 (10:23 -0700)]
IWYU RegionHelper += <limits>
Test: compiles when transitive dependency is removed
Bug:
35363681
Change-Id: Ia405e41102e6ded07720e424167df7e4d080581e
Nicolas Geoffray [Tue, 6 Jun 2017 08:13:27 +0000 (08:13 +0000)]
Merge "Revert "Disable in-place vdex update""
Nicolas Geoffray [Mon, 5 Jun 2017 12:38:10 +0000 (12:38 +0000)]
Revert "Disable in-place vdex update"
Reverting to only apply this when doing dexlayout.
PS1 is a pure revert. PS2 contains the fix.
This reverts commit
e446fede9258359b61d4c8753d7e96477fa29388.
Change-Id: I04a7dbde142d59e8dcdf4e93ab0456c4aa3d02ab
Treehugger Robot [Fri, 2 Jun 2017 19:40:44 +0000 (19:40 +0000)]
Merge "correct value of TYPE_MAGNIFICATION_OVERLAY"
kenzyun_chen [Fri, 2 Jun 2017 05:36:21 +0000 (13:36 +0800)]
correct value of TYPE_MAGNIFICATION_OVERLAY
It should be sync to the same value in WindowManager.java
or the InputWindowInfo::isTrustedOverlay() function will fail
Change-Id: I11bb518049724252f2795a6400efc17ddd545fbb
Treehugger Robot [Thu, 1 Jun 2017 21:49:25 +0000 (21:49 +0000)]
Merge "Installd: deleteOdex's outputPath can be null"
Andreas Gampe [Wed, 31 May 2017 23:39:58 +0000 (16:39 -0700)]
Installd: deleteOdex's outputPath can be null
Add @nullable, as the outputPath can be null for system apps.
(cherry picked from commit
c523409f0fd03aa498cedc486f85e9a4b7257f3a)
Bug:
62236497
Test: m
Test: Fill up /data, apply OTA
Change-Id: I61d86721b485f85195b153d76ce25412dcb68be6
Treehugger Robot [Wed, 31 May 2017 22:00:43 +0000 (22:00 +0000)]
Merge "Disable in-place vdex update"
Narayan Kamath [Wed, 31 May 2017 11:55:47 +0000 (11:55 +0000)]
Merge "dumpstate: Track debuggerd API changes."
Narayan Kamath [Wed, 31 May 2017 09:33:28 +0000 (10:33 +0100)]
dumpstate: Track debuggerd API changes.
dump_backtrace_to_file_timeout now needs an additional argument
that specifies the dump type.
Bug:
36024548
Test: make; manual bug report capture
Change-Id: I6ce3c172dd571e296d9ced9c576b88ea487b2fbb
David Sehr [Fri, 26 May 2017 23:32:55 +0000 (16:32 -0700)]
Disable in-place vdex update
Vdex files can be corrupted by dexlayout or by previously dexlayout-ed
files being recompiled if they are used as both input and output.
Disable this feature.
Bug:
38113240
Test: OTA and dexopt
Change-Id: Iee2d9a6c1ba8cddda25290f68f50649d5288d588
Treehugger Robot [Fri, 26 May 2017 20:21:26 +0000 (20:21 +0000)]
Merge "Add pids for processes that dlopens implementations"
Yifan Hong [Thu, 25 May 2017 02:45:02 +0000 (19:45 -0700)]
Add pids for processes that dlopens implementations
Test: lshal -itrpc
Test: lshal
Test: lshal --init-vintf
Test: lshal_test
Bug:
37639822
Change-Id: Ia415e5f8dbcd10f24e9a5cfdd591eb829a1da89e
Nicolas Geoffray [Fri, 26 May 2017 10:21:25 +0000 (10:21 +0000)]
Merge "Delete vdex file in delete_odex."
Nicolas Geoffray [Thu, 25 May 2017 12:58:06 +0000 (13:58 +0100)]
Delete vdex file in delete_odex.
bug:
38513395
Test: unfortunately no unit test here, so after discussing
with agampe@ just relied on build.
Change-Id: I305ddaadc813673fb55473f5e8ca357f39cf98f3
Treehugger Robot [Thu, 25 May 2017 22:42:24 +0000 (22:42 +0000)]
Merge "Add thread usage output to lshal."
Treehugger Robot [Thu, 25 May 2017 00:59:58 +0000 (00:59 +0000)]
Merge "SF: Call makeCurrent after presenting each display"
Steven Moreland [Wed, 24 May 2017 18:23:08 +0000 (11:23 -0700)]
Add thread usage output to lshal.
This output shows how many threads are in use at a give time.
Test: lshal, manual
Bug:
35099601
Change-Id: I3a22bb131b828cdd77e73e7810229d9c68a496d3
Season Li [Thu, 4 Aug 2016 21:32:44 +0000 (14:32 -0700)]
SF: Call makeCurrent after presenting each display
HWC can change EGLContext in each present. Restore to
default EGLContext after commiting each display.
Test: see https://android-review.googlesource.com/#/c/372703/
Change-Id: I4abc84fb3d4b74964fabcad5a2568ff01a6c5dbb
Treehugger Robot [Wed, 24 May 2017 16:10:16 +0000 (16:10 +0000)]
Merge "Fix a anr bug caused by sendFinishedSignal logical error"
gaoshang [Wed, 17 May 2017 06:36:46 +0000 (14:36 +0800)]
Fix a anr bug caused by sendFinishedSignal logical error
Because of eliminate multiple benign overflow conditions see:
https://android-review.googlesource.com/#/c/172237/
Changed the do while loop resulting in a logical difference.
When the chainIndex-- to 0, the loop is not run.
It is the most important cycle, it will push head seq to mSeqChains.
If not run, will lead to a batch of seq can not corrected finish.
Eventually leading to the occurrence of anr.
Signed-off-by: gaoshang <gaoshang@xiaomi.com>
Test: Input dispatche process
Bug:
38366215
Change-Id: I87e609dfcb00ac7b8e82c6de789df094e7c25efd
Howard Chen [Thu, 18 May 2017 02:50:48 +0000 (02:50 +0000)]
Merge "Support halt the trace when schd-dbg hit deadline"
Treehugger Robot [Wed, 17 May 2017 07:28:41 +0000 (07:28 +0000)]
Merge "Make the binder_lock category optional"
Howard Chen [Wed, 10 May 2017 08:32:11 +0000 (16:32 +0800)]
Make the binder_lock category optional
The binder_lock event type has been deprecated after binder
driver uses fine grained locking. This patch makes "binder_lock"
an optional category for atrace to support system w/ w/o FGL.
Bug:
37559846
Test: pixel phone with "atrace --async_start binder_lock"
Merged-In: I691efd355e903a26629f896a56a664502e5807f7
Change-Id: I691efd355e903a26629f896a56a664502e5807f7
Treehugger Robot [Mon, 15 May 2017 20:05:51 +0000 (20:05 +0000)]
Merge "Add lshal --neat option"
Yifan Hong [Fri, 12 May 2017 23:56:43 +0000 (16:56 -0700)]
Add lshal --neat option
... that doesn't output explanatory text and title line.
Test: lshal --neat
Test: lshal
Test: lshal --neat -itrpc
Change-Id: I269a5d284903dc0e9bc3be96379ed41aecea8ddb
Fixes:
38211518
Treehugger Robot [Sat, 13 May 2017 01:25:34 +0000 (01:25 +0000)]
Merge "Actually remove dead processes from lshal"
Yifan Hong [Fri, 12 May 2017 23:33:57 +0000 (16:33 -0700)]
Actually remove dead processes from lshal
Test: run lshal with and without the change
Change-Id: Id8c52a5be3253f5ff9c1cb851459a16e01ca5dda
Fixes:
35765788
Yifan Hong [Fri, 12 May 2017 22:21:25 +0000 (22:21 +0000)]
Merge changes from topic 'lshal_debug'
* changes:
Add tests to lshal.
Add `lshal debug` command.
Lshal: support commands.
Yifan Hong [Wed, 10 May 2017 21:33:05 +0000 (14:33 -0700)]
Add tests to lshal.
Test: lshal_test
Bug:
37954458
Change-Id: I1914e6274974ed5eb0ce2d655f1333d2344b49f5
Yifan Hong [Wed, 10 May 2017 02:33:08 +0000 (19:33 -0700)]
Add `lshal debug` command.
Supported command:
lshal debug android.hardware.foo@1.0::IFoo option option
Test: adb unroot && lshal --debug ; echo $?
Test: adb unroot && lshal debug android.hardware.nfc@1.0::INfc ; echo $?
Test: adb root && lshal --debug ; echo $?
Test: adb root && lshal debug android.hardware.nfc@1.0::INfc ; echo $?
Bug:
37954458
Change-Id: Ia2f4c9c0d3fb0a7bb26e76f01d02f49dc426e7f8
Yifan Hong [Wed, 10 May 2017 01:49:45 +0000 (18:49 -0700)]
Lshal: support commands.
Support the following commands:
lshal list
lshal list -itrpc
lshal help
lshal help list
lshal list -h
lshal list --help
Test: run these commands
Bug:
37725279
Change-Id: I970fbc8d250d43f57e92f783229e0645d7e8df4e
Nicolas Geoffray [Thu, 11 May 2017 09:42:23 +0000 (09:42 +0000)]
Merge "Remove obsolete DEXOPT_SAFEMODE."
Yifan Hong [Thu, 11 May 2017 00:23:58 +0000 (00:23 +0000)]
Merge "lshal: remove "// static" comments"
Treehugger Robot [Wed, 10 May 2017 23:35:33 +0000 (23:35 +0000)]
Merge "Add explicit default copy constructor"
Steven Moreland [Tue, 9 May 2017 23:43:46 +0000 (16:43 -0700)]
lshal: remove "// static" comments
Test: pass
Change-Id: I4d710c8aaad87cfcd7896800a5ab24f812775703
Yi Kong [Wed, 10 May 2017 21:27:39 +0000 (14:27 -0700)]
Add explicit default copy constructor
Definition of implicit copy constructor is deprecated if it has a
user-declared destructor. This triggers -Wdeprecated compiler warning
and breaks the build in latest Clang update.
Test: Build
Bug:
37752547
Change-Id: I467d5ff3b4cca4644c7ec8c99ae35556029ce509
Alex Klyubin [Wed, 10 May 2017 19:25:30 +0000 (19:25 +0000)]
Merge "Switch from libcrypto_static to libcrypto target"
Nicolas Geoffray [Fri, 5 May 2017 13:32:51 +0000 (14:32 +0100)]
Remove obsolete DEXOPT_SAFEMODE.
installd side of the change.
bug:
37929796
Test: build
(cherry picked from commit
bad623a225d0b3d94f30f8654bd4684969e90d4c)
Change-Id: Ice855d20184443a641e081efa4ec943326ccefdd
Nicolas Geoffray [Wed, 10 May 2017 08:52:29 +0000 (08:52 +0000)]
Merge "Always unlink the out_vdex_file."
Nicolas Geoffray [Tue, 9 May 2017 13:20:09 +0000 (14:20 +0100)]
Always unlink the out_vdex_file.
Even if we pass the same vdex file as input and output to
dex2oat, make sure we unlink it if dex2oat failed.
bug:
38113240
Test: manual, killing dex2oat invocation
Change-Id: I406014f97a2c3b2bd646e13b5d86dfa0eaa2692f
Alex Klyubin [Mon, 8 May 2017 21:12:57 +0000 (14:12 -0700)]
Switch from libcrypto_static to libcrypto target
libcrypto target offers the same static library as libcrypto_static.
Thus, there's no need to use libcrypto_static.
Test: make dumpstate
dumpstate binary unchanged by this commit
Bug:
26160319
Change-Id: I62ca775870482386f0bb7ce1e2f8c94cf6f0342a
Treehugger Robot [Sat, 6 May 2017 00:11:31 +0000 (00:11 +0000)]
Merge "Fix lshal --init-vintf may have empty <instance> tags."
Nicolas Geoffray [Fri, 5 May 2017 08:37:16 +0000 (08:37 +0000)]
Merge "Handle safe mode in PackageManager."
Yi Kong [Thu, 4 May 2017 22:13:17 +0000 (22:13 +0000)]
Merge "Add 'override' specifier to overridden destructor methods"
Yi Kong [Tue, 2 May 2017 23:55:24 +0000 (16:55 -0700)]
Add 'override' specifier to overridden destructor methods
Test: Build
Bug:
37752547
Change-Id: I157e1c2aa5d9d08fcab6055a3ca48c6ed8a050fc
Yi Kong [Thu, 4 May 2017 00:14:49 +0000 (00:14 +0000)]
Merge "Add const specifier to SortLayersByZ comparator"
Tom Cherry [Wed, 3 May 2017 23:42:58 +0000 (23:42 +0000)]
Merge "installd create_data_user_ce_path uses dir instead of symlink"
Yifan Hong [Wed, 3 May 2017 20:13:23 +0000 (13:13 -0700)]
Fix lshal --init-vintf may have empty <instance> tags.
Test: on bullhead, lshal --init-vintf | grep "<instance></instance>"
Change-Id: Ic6ab977adcdc5293015e9daba506b625629e5352
Fixes:
37311367
Yi Kong [Wed, 3 May 2017 20:26:48 +0000 (13:26 -0700)]
Add const specifier to SortLayersByZ comparator
No functionality change.
Test: build
Bug:
37752547
Change-Id: I33ffccbd7bc67031680e3fe7e9263de5402a9cde
Treehugger Robot [Wed, 3 May 2017 16:15:21 +0000 (16:15 +0000)]
Merge "Lshal: note debug in help"
Nicolas Geoffray [Wed, 3 May 2017 12:21:37 +0000 (13:21 +0100)]
Handle safe mode in PackageManager.
installd side of the change.
bug:
37929796
Test: manual OTA with a safemode app.
Change-Id: I1cdb034b9e27b72ad606a4935172d0040364c380
Treehugger Robot [Wed, 3 May 2017 02:00:25 +0000 (02:00 +0000)]
Merge "Remove warnings in Rect.h"
Steven Moreland [Wed, 3 May 2017 00:27:52 +0000 (17:27 -0700)]
Lshal: note debug in help
Test: pass
Bug:
37725279
Change-Id: I2f43628e42f2cb334083dcb4f7d174100021a745
Colin Cross [Tue, 2 May 2017 22:59:14 +0000 (15:59 -0700)]
Remove warnings in Rect.h
Fixes:
frameworks/native/include/ui/Rect.h:53:69: error: format specifies type 'unsigned int' but the argument has type 'unsigned long' [-Werror,-Wformat]
Bug:
31751828
Test: m -j checkbuild
Change-Id: Ic8a2ce741f550651f943805382adc035a8da1096
Treehugger Robot [Tue, 2 May 2017 13:23:17 +0000 (13:23 +0000)]
Merge "Revert "Disable input-vdex when doing speed-profile.""
Treehugger Robot [Mon, 1 May 2017 23:19:24 +0000 (23:19 +0000)]
Merge "Remove redundant dependencies"
Steven Moreland [Mon, 1 May 2017 17:19:17 +0000 (10:19 -0700)]
Remove redundant dependencies
android.hidl.base@1.0 and android.hidlmanager@1.0 are built into libhidltransport.
Test: links
Bug:
33276472
Change-Id: Ia22e134d3882f2e503e540dda332fc17213b6bb0
Nicolas Geoffray [Thu, 27 Apr 2017 11:40:05 +0000 (11:40 +0000)]
Merge "Update installd to new compiler filters."
Nicolas Geoffray [Mon, 24 Apr 2017 13:34:59 +0000 (13:34 +0000)]
Revert "Disable input-vdex when doing speed-profile."
https://android-review.googlesource.com/#/c/377754/ is fixing vdex and delayout.
bug:
37558732
Test: device boots, vdex works on OTA, input-vdex is passed
for speed-profile
This reverts commit
a2dbefc7a51482acb8d5d139c9e2f01306ef9ceb.
Change-Id: I8fcab30ec617b39769fbf7e5818e5d503d647fb4
Dan Albert [Wed, 26 Apr 2017 18:34:14 +0000 (18:34 +0000)]
Merge "Add missing stdbool.h to sensor.h."
Dan Albert [Tue, 25 Apr 2017 19:24:28 +0000 (12:24 -0700)]
Add missing stdbool.h to sensor.h.
This got hacked into the NDK headers ages ago, but never got
upstreamed to the frameworks.
Test: make checkbuild
Bug: https://github.com/android-ndk/ndk/issues/370
Merged-In: Id38cbc96633c1c61de36996c2dd9e6c2331bd6e7
Change-Id: Id38cbc96633c1c61de36996c2dd9e6c2331bd6e7
Nicolas Geoffray [Tue, 25 Apr 2017 11:28:28 +0000 (12:28 +0100)]
Update installd to new compiler filters.
bug:
34715556
Test: build and boot aosp bullhead
Change-Id: If10fb8cdfe34f109d4caa1df56a49304025da553
Steven Moreland [Fri, 21 Apr 2017 01:57:34 +0000 (01:57 +0000)]
Merge "Revert "(lib)?sensorservice: Android.mk -> Android.bp""
Steven Moreland [Fri, 21 Apr 2017 01:57:19 +0000 (01:57 +0000)]
Revert "(lib)?sensorservice: Android.mk -> Android.bp"
This reverts commit
f1b07d51ed6a11a0446115934fb75c22a1badb87.
Bad internal merge and other errors.
Change-Id: If9b812916223773ee98354d0805626cf759355fd
Merged-In: Ib3a3f249c278be582de04a583e18de3c0e502d5e
Steven Moreland [Fri, 21 Apr 2017 00:03:10 +0000 (00:03 +0000)]
Merge "(lib)?sensorservice: Android.mk -> Android.bp"
Steven Moreland [Thu, 20 Apr 2017 22:56:38 +0000 (15:56 -0700)]
(lib)?sensorservice: Android.mk -> Android.bp
Test: links
Change-Id: I2ddab32c12bf07044416f983b5cf0ae19cc12ba0
Treehugger Robot [Thu, 20 Apr 2017 17:59:45 +0000 (17:59 +0000)]
Merge "Convert inputflinger to Android.bp"
Colin Cross [Wed, 19 Apr 2017 23:56:46 +0000 (16:56 -0700)]
Convert inputflinger to Android.bp
See build/soong/README.md for more information.
Test: m -j checkbuild
Change-Id: Iebdc3f4d2d560b0fdf5447dde0ac06df64826bb5
Merged-In: Iebdc3f4d2d560b0fdf5447dde0ac06df64826bb5
(cherry picked from commit
45a1b9c95de619785b68ce51b1a3e271abc4ffd9)
Treehugger Robot [Thu, 20 Apr 2017 01:39:21 +0000 (01:39 +0000)]
Merge "libbinder: vendor_available"
Yifan Hong [Thu, 20 Apr 2017 00:51:48 +0000 (00:51 +0000)]
Merge "Fix lshal not to squash HAL entries incorrectly."
Treehugger Robot [Wed, 19 Apr 2017 22:22:32 +0000 (22:22 +0000)]
Merge "Fix passing --classpath-dir to dex2oat for split apks."
Steven Moreland [Fri, 14 Apr 2017 04:19:48 +0000 (21:19 -0700)]
libbinder: vendor_available
By setting vendor_available, the following may become true:
* a prebuilt library from this release may be used at runtime by
in a later releasse (by vendor code compiled against this release).
so this library shouldn't depend on runtime state that may change
in the future.
* this library may be loaded twice into a single process (potentially
an old version and a newer version). The symbols will be isolated
using linker namespaces, but this may break assumptions about 1
library in 1 process (your singletons will run twice).
Background:
This means that these modules may be built and installed twice --
once for the system partition and once for the vendor partition. The
system version will build just like today, and will be used by the
framework components on /system. The vendor version will build
against a reduced set of exports and libraries -- similar to, but
separate from, the NDK. This means that all your dependencies must
also mark vendor_available.
At runtime, /system binaries will load libraries from /system/lib*,
while /vendor binaries will load libraries from /vendor/lib*. There
are some exceptions in both directions -- bionic(libc,etc) and liblog
are always loaded from /system. And SP-HALs (OpenGL, etc) may load
/vendor code into /system processes, but the dependencies of those
libraries will load from /vendor until it reaches a library that's
always on /system. In the SP-HAL case, if both framework and vendor
libraries depend on a library of the same name, both versions will be
loaded, but they will be isolated from each other.
It's possible to compile differently -- reducing your source files,
exporting different include directories, etc. For details see:
https://android-review.googlesource.com/368372
None of this is enabled unless the device opts into the system/vendor
split with BOARD_VNDK_VERSION := current.
Bug:
36426473
Bug:
36079834
Test: m -j libbinder
Test: attempt to compile with BOARD_VNDK_VERSION := current
(cherry picked from commit
3aa761697b93dab3e4b1acf90d14502a79a49219)
Merged-In: Iec94a66adbc852f7f23352307cf36f49a79ee3a6
Change-Id: Iec94a66adbc852f7f23352307cf36f49a79ee3a6
Jeff Hao [Wed, 19 Apr 2017 20:50:15 +0000 (13:50 -0700)]
Fix passing --classpath-dir to dex2oat for split apks.
Merged wrong version of CL in by mistake. CL should have been
cherry-picked from
cc5c4e32637601432bc6359fb7ff34e4cadb6502.
Bug:
34169257
Test: cts-tradefed run singleCommand cts -d --module
CtsAppSecurityHostTestCases -t android.appsecurity.cts.SplitTests
Change-Id: Ibebcdce1c9ab0de3254727c8dfc3d0de18932f1c
Treehugger Robot [Wed, 19 Apr 2017 17:13:11 +0000 (17:13 +0000)]
Merge "Run dex2oat from the staged apk directory."
Yifan Hong [Tue, 18 Apr 2017 21:19:40 +0000 (14:19 -0700)]
Fix lshal not to squash HAL entries incorrectly.
Bug:
37311367
Test: lshal --init-vintf
Change-Id: I693b53f5f342c7600e4031d020792fd347e3ae83
Jeff Hao [Thu, 6 Apr 2017 00:11:39 +0000 (17:11 -0700)]
Run dex2oat from the staged apk directory.
The frameworks/base change makes the PackageManager pass relative paths
of previous code paths as shared library paths. This change passes the
base directory for these relative paths to dex2oat using the new switch
--base-dir.
Part of a multi-project change.
Bug:
34169257
Test: cts-tradefed run singleCommand cts -d --module
CtsAppSecurityHostTestCases -t android.appsecurity.cts.SplitTests
Merged-In: Ic20aa0b6a80104d0759ac84df3eb89570604e53f
(cherry-picked from commit
cc5c4e32637601432bc6359fb7ff34e4cadb6502)
Change-Id: I29e6e7365654cc1a538272591356e2253d6ec0e4
cjbao [Tue, 11 Apr 2017 16:12:24 +0000 (00:12 +0800)]
installd create_data_user_ce_path uses dir instead of symlink
Select whichever is real dir instead of symbolic link from either
/data/data or /data/user/0. This is to minimize path walking overhead
in kernel.
Test: Manual test
Change-Id: Ie06640624746ff40a2b5b47ead50a7f10eb2db3e
Signed-off-by: cjbao <cathy.bao@intel.com>
Treehugger Robot [Mon, 17 Apr 2017 20:26:40 +0000 (20:26 +0000)]
Merge "Fix error BufferQueue attaching flow."
Yifan Hong [Mon, 17 Apr 2017 18:12:27 +0000 (18:12 +0000)]
Merge "Update to allow multiple HALs in a manifest."
Yifan Hong [Thu, 13 Apr 2017 08:00:12 +0000 (01:00 -0700)]
Update to allow multiple HALs in a manifest.
Test: libvintf_test
Test: lshal --init_vintf
Bug:
37294247
Change-Id: I202787d7856db6e06e595bd52c644f6a0eefcb66
Josh Gao [Wed, 12 Apr 2017 19:43:13 +0000 (19:43 +0000)]
Merge "atrace: don't use sendfile."
Treehugger Robot [Wed, 12 Apr 2017 19:21:34 +0000 (19:21 +0000)]
Merge "Mark as vendor_available"
Josh Gao [Tue, 11 Apr 2017 22:21:13 +0000 (15:21 -0700)]
atrace: don't use sendfile.
sendfile appears to have horrendous performance when used with the
ftrace output file, being up to 100x slower than the naive read/write
loop.
Switch to doing that instead, speeding up atrace from:
$ atrace --async_start sched freq; sleep 1; time atrace --async_stop > dummy
c apturing trace... 0m08.93s real 0m00.08s user 0m07.98s system
to:
$ atrace --async_start sched freq; sleep 1; time atrace --async_stop > dummy
capturing trace... 0m00.78s real 0m00.07s user 0m00.21s system
Bug: http://b/
37164190
Test: atrace --async_start sched freq; sleep 1; time atrace --async_stop > dummy
Change-Id: I22fe1871e263867f9ac54c8f5b474df824b4bc69
Treehugger Robot [Tue, 11 Apr 2017 16:19:14 +0000 (16:19 +0000)]
Merge "Remove .(ll)ndk suffix from (ll)ndk_library"
Dan Willemsen [Fri, 7 Apr 2017 22:48:39 +0000 (15:48 -0700)]
Remove .(ll)ndk suffix from (ll)ndk_library
Soong handles these automatically now.
Bug:
33241851
Test: Android-aosp_arm.mk is the same before/after
Test: build.ninja is the same before/after
Test: build-aosp_arm.ninja is the same before/after
Merged-In: Ie1cd267564095c51e536b7a84d53e7aed9dca288
Change-Id: Ie1cd267564095c51e536b7a84d53e7aed9dca288
Dan Willemsen [Fri, 7 Apr 2017 21:14:08 +0000 (14:14 -0700)]
Mark as vendor_available
By setting vendor_available, the following may become true:
* a prebuilt library from this release may be used at runtime by
in a later releasse (by vendor code compiled against this release).
so this library shouldn't depend on runtime state that may change
in the future.
* this library may be loaded twice into a single process (potentially
an old version and a newer version). The symbols will be isolated
using linker namespaces, but this may break assumptions about 1
library in 1 process (your singletons will run twice).
Background:
This means that these modules may be built and installed twice --
once for the system partition and once for the vendor partition. The
system version will build just like today, and will be used by the
framework components on /system. The vendor version will build
against a reduced set of exports and libraries -- similar to, but
separate from, the NDK. This means that all your dependencies must
also mark vendor_available.
At runtime, /system binaries will load libraries from /system/lib*,
while /vendor binaries will load libraries from /vendor/lib*. There
are some exceptions in both directions -- bionic(libc,etc) and liblog
are always loaded from /system. And SP-HALs (OpenGL, etc) may load
/vendor code into /system processes, but the dependencies of those
libraries will load from /vendor until it reaches a library that's
always on /system. In the SP-HAL case, if both framework and vendor
libraries depend on a library of the same name, both versions will be
loaded, but they will be isolated from each other.
It's possible to compile differently -- reducing your source files,
exporting different include directories, etc. For details see:
https://android-review.googlesource.com/368372
None of this is enabled unless the device opts into the system/vendor
split with BOARD_VNDK_VERSION := current.
Bug:
36426473
Bug:
36079834
Test: Android-aosp_arm.mk is the same before/after
Test: build.ninja is the same before/after
Test: build-aosp_arm.ninja is the same before/after
Test: attempt to compile with BOARD_VNDK_VERSION := current
Change-Id: I875a75d785aa13b97751e1c145183d87add01845
Treehugger Robot [Thu, 6 Apr 2017 06:34:45 +0000 (06:34 +0000)]
Merge "lshal --init-vintf stripped <impl> tag."
Howard Chen [Tue, 28 Mar 2017 09:05:11 +0000 (17:05 +0800)]
Support halt the trace when schd-dbg hit deadline
Usage:
start atrace as usual:
>atrace --async_start sched freq ...
then use schd-dbg with -trace arguments
./schd-dbg -trace -deadline_us 2500
This makes schd-dbg to stop trace once it detects a transaction
duration over the deadline. By writing '0' to
/sys/kernel/debug/tracing and halt the process. The tracelog is
then available on /sys/kernel/debug/trace
Bug: b/
36669012
Test: sailfish with prebuilt kernel
Change-Id: I9f81109cd22705f070f3f78ddf3539b8fa1f9973
Treehugger Robot [Thu, 6 Apr 2017 01:39:35 +0000 (01:39 +0000)]
Merge "This test is used to check if the kernel scheduler"
Yifan Hong [Tue, 4 Apr 2017 22:33:59 +0000 (15:33 -0700)]
lshal --init-vintf stripped <impl> tag.
<impl> tag is deprecated from the manifest.
Test: lshal --init-vintf
Change-Id: Ifa9336e3dfe7f1091ee61aa6a10df9629ad98b49
Howard Chen [Sat, 25 Mar 2017 09:12:59 +0000 (17:12 +0800)]
This test is used to check if the kernel scheduler
1. supports priority inheritance.
2. utilize the sync flag.
It also make statistic on average, best and worst latency
as well as number of deadline miss with a configurable
deadline specified in us.
The output is in json format hopefully to make it easy
to integrated with present test framework.
usage:
./schd-dbg # test and output json
./schd-dbg -i 1 # number of iterations
./schd-dbg -pair 4 # number of process pairs
./schd-dbg -deadline_us 2500 # deadline in us
./schd-dbg -v # debug
Test: pixel
Change-Id: Idccf0d402dcd7b81c639020364b5ee32075e18cb
Treehugger Robot [Tue, 4 Apr 2017 23:08:13 +0000 (23:08 +0000)]
Merge "vulkan: update Vulkan headers to 1.0.46"
Jesse Hall [Mon, 20 Mar 2017 20:09:19 +0000 (13:09 -0700)]
vulkan: update Vulkan headers to 1.0.46
- Add VK_EXT_swapchain_colorspace
- Add VK_KHR_incremental_present
Test: build aosp_arm64-eng
Change-Id: I6cd81baa2870bf436abab6191a05ebfd60ccbc46
Merged-In: I70ffd56a63da39301cd9706b1b91cf3643aab916
Dan Shi [Fri, 31 Mar 2017 17:13:30 +0000 (17:13 +0000)]
Merge "Add test config to dumpsys_test"
am:
cf617abcb2
Change-Id: I6257e48eeaf89e26ff4c46e69dd20941dae58392
Dan Shi [Fri, 31 Mar 2017 17:13:17 +0000 (17:13 +0000)]
Merge "Add test config to libgui_test"
am:
7f1e92dc64
Change-Id: Iaf39fdc5a6a9fb45cd0fb44504144d6eee66b909
Treehugger Robot [Fri, 31 Mar 2017 17:07:15 +0000 (17:07 +0000)]
Merge "Add test config to dumpsys_test"
Treehugger Robot [Fri, 31 Mar 2017 17:06:53 +0000 (17:06 +0000)]
Merge "Add test config to libgui_test"
Richard Uhler [Fri, 31 Mar 2017 07:50:15 +0000 (07:50 +0000)]
Merge "Remove obsolete and unused PATCHOAT_FOR_RELOCATION."
am:
804b145356
Change-Id: I0c9a38e9be249bcb81f7b04b436a00d60b779fc5
Richard Uhler [Fri, 31 Mar 2017 07:46:45 +0000 (07:46 +0000)]
Merge "Remove obsolete and unused PATCHOAT_FOR_RELOCATION."