OSDN Git Service

android-x86/bionic.git
8 years agoMerge \"Treat \"persist.sys.timezone\" as being a Java-format timezone.\"
Elliott Hughes [Thu, 21 Jul 2016 08:14:31 +0000 (08:14 +0000)]
Merge \"Treat \"persist.sys.timezone\" as being a Java-format timezone.\"
am: 91f33cd5ae

Change-Id: Iafd50c6b6b44166f2fdf1506dbcbe5c788568966

8 years agoMerge "Treat "persist.sys.timezone" as being a Java-format timezone."
Treehugger Robot [Thu, 21 Jul 2016 08:06:33 +0000 (08:06 +0000)]
Merge "Treat "persist.sys.timezone" as being a Java-format timezone."

8 years agoresolve merge conflicts of 823e32e to stage-aosp-master
Christopher Ferris [Thu, 21 Jul 2016 01:19:06 +0000 (18:19 -0700)]
resolve merge conflicts of 823e32e to stage-aosp-master

Change-Id: I93911e76d85f55d9105c1349cd919dc924fbd5e4

8 years agoTreat "persist.sys.timezone" as being a Java-format timezone.
Elliott Hughes [Thu, 21 Jul 2016 00:23:54 +0000 (17:23 -0700)]
Treat "persist.sys.timezone" as being a Java-format timezone.

POSIX and Java disagree about the sign in a timezone string like "GMT+3".
This means that if you set "persist.sys.timezone" to such a string, native
code and Java code disagree about what time it is. Resolve this by
translating to POSIX form for native code.

Bug: http://b/25463955
Change-Id: I7c08c459dd8514348a12ce419edcbfbfe8f6d327
Test: manually tested with setprop and date.

8 years agoMerge "Pass argc/argv/envp to dso constructors"
Dimitry Ivanov [Thu, 21 Jul 2016 00:08:46 +0000 (00:08 +0000)]
Merge "Pass argc/argv/envp to dso constructors"

8 years agoPass argc/argv/envp to dso constructors
Dimitry Ivanov [Wed, 20 Jul 2016 22:33:07 +0000 (15:33 -0700)]
Pass argc/argv/envp to dso constructors

The glibc implementation of the loader passes argc/argv/envp
to all elf constructors. This change makes bionic linker
behave in the same way.

Bug: http://b/30145768
Change-Id: I1c65c42aa5305a5b133c64b5748205bbde869e0e
Test: run bionic-unit-tests --gtest_filter=dl*:Dl*

8 years agoMerge \"Align-up and randomize shared libraries.\"
Evgenii Stepanov [Wed, 20 Jul 2016 20:37:26 +0000 (20:37 +0000)]
Merge \"Align-up and randomize shared libraries.\"
am: a9fbaee593

Change-Id: I5cb0caea4db3a8cb3b0e1a4dd3a4bb2d5f5069d9

8 years agoMerge "Align-up and randomize shared libraries."
Evgenii Stepanov [Wed, 20 Jul 2016 20:31:41 +0000 (20:31 +0000)]
Merge "Align-up and randomize shared libraries."

8 years agoMerge \"Move/remove android only kernel uapi headers.\"
Christopher Ferris [Wed, 20 Jul 2016 18:09:38 +0000 (18:09 +0000)]
Merge \"Move/remove android only kernel uapi headers.\"
am: fe280773ed

Change-Id: If9de2352e7522628aff751211753c11d4d453ca8

8 years agoMerge "Move/remove android only kernel uapi headers."
Christopher Ferris [Wed, 20 Jul 2016 18:03:47 +0000 (18:03 +0000)]
Merge "Move/remove android only kernel uapi headers."

8 years agoAlign-up and randomize shared libraries.
Evgenii Stepanov [Fri, 15 Jul 2016 23:31:42 +0000 (16:31 -0700)]
Align-up and randomize shared libraries.

This change implements the following property:
  Any 2**N aligned memory region on size 2**N contains no more than one DSO.

The value N can be configured, with 16 or 18 looking like a good choice.
Additionally, DSOs are loaded at random page-aligned address inside these large
regions.

This change has dual purpose:
1. Larger values of N allow a lot more compact CFI shadow implementation.
   See change I14dfea630de468eb5620e7f55f92b1397ba06217.
   For example, CFI shadow for the system_server process has the following size (RSS, KB):
   152 for N = 12, 32 for N = 16, 16 for N = 18.
2. Extra randomization is good for security.

This change does not result in extra RAM usage, because everything is still page-aligned.
It does result in a bit more VM fragmentation because of the gaps between shared libraries.
As it turns out, this fragmentation is barely noticeable because the kernel creates new mapping
at the highest possible address, and we do enough small mappings to almost completely fill the
gaps (ex. in the Zygote the gaps are filled with .ttf file mappings and thread stacks).

I've measured VM fragmentation as the sum of all VM gaps (unmapped regions) that are larger
than 1MB according to /proc/$PID/maps. On aosp_angler-userdebug, the numbers are (in GB):

                |   N = 12  |  N = 18
system_server   |   521.9   |  521.1
zygote64        |   522.1   |  521.3
zygote32        |   2.55    |  2.55
mediaserver     |   4.00    |  4.00

Change-Id: Ia6df840dd409c82837efd1f263be420d9723c84a

8 years agoMerge \"Expand the testing docs slightly.\"
Elliott Hughes [Tue, 19 Jul 2016 22:00:46 +0000 (22:00 +0000)]
Merge \"Expand the testing docs slightly.\"
am: d941f72e75

Change-Id: I9ad64b2f4b1af867b422af7a558338d1224823aa

8 years agoMerge \"Remove __LIBC_ABI_PUBLIC__.\"
Elliott Hughes [Tue, 19 Jul 2016 21:58:19 +0000 (21:58 +0000)]
Merge \"Remove __LIBC_ABI_PUBLIC__.\"
am: 9af76f9b48

Change-Id: I8e2e7bf6a70f978619499a0e96e424c6bdb1cdf0

8 years agoMerge "Expand the testing docs slightly."
Treehugger Robot [Tue, 19 Jul 2016 21:57:27 +0000 (21:57 +0000)]
Merge "Expand the testing docs slightly."

8 years agoMerge "Remove __LIBC_ABI_PUBLIC__."
Treehugger Robot [Tue, 19 Jul 2016 21:48:25 +0000 (21:48 +0000)]
Merge "Remove __LIBC_ABI_PUBLIC__."

8 years agoExpand the testing docs slightly.
Elliott Hughes [Tue, 19 Jul 2016 21:09:10 +0000 (14:09 -0700)]
Expand the testing docs slightly.

Change-Id: I4b172eefd8f456a62b24520a9f04f500a1d6bc85

8 years agoRemove __LIBC_ABI_PUBLIC__.
Elliott Hughes [Tue, 19 Jul 2016 20:56:46 +0000 (13:56 -0700)]
Remove __LIBC_ABI_PUBLIC__.

We have much better control over visibility now, so we don't need to
pollute the headers with cruft.

Bug: http://b/24767418
Change-Id: I349f4c3bc30102477375ad9f80926e560c7c1d8b

8 years agoMove/remove android only kernel uapi headers.
Christopher Ferris [Mon, 18 Jul 2016 23:27:22 +0000 (16:27 -0700)]
Move/remove android only kernel uapi headers.

Move all of the uapi kernel headers that are only found in the android
kernel, and are still used by the tree into their own directory.

Delete all of these files that are not used.

Also, delete binder.h from the linux directory since it already exists
in the linux/android directory.

Bug: 30072483
Change-Id: I879f16e366f7670b8be3546b0a9cf9567fce6526

8 years agoMerge \"Add a test for a known kernel bug.\"
Elliott Hughes [Tue, 19 Jul 2016 18:10:58 +0000 (18:10 +0000)]
Merge \"Add a test for a known kernel bug.\"
am: 061cf180db

Change-Id: I31af9c357077b932157eef80f3c51c96750986ab

8 years agoMerge "Add a test for a known kernel bug."
Treehugger Robot [Tue, 19 Jul 2016 17:52:03 +0000 (17:52 +0000)]
Merge "Add a test for a known kernel bug."

8 years agoMerge \"Don\'t declare SYS_* constants for unavailable __NR_* constants.\"
Elliott Hughes [Mon, 18 Jul 2016 20:48:52 +0000 (20:48 +0000)]
Merge \"Don\'t declare SYS_* constants for unavailable __NR_* constants.\"
am: 20de9ef5bd

Change-Id: Ib9e6ef7fc06df0ac42414870061da287a9692d09

8 years agoMerge "Don't declare SYS_* constants for unavailable __NR_* constants."
Treehugger Robot [Mon, 18 Jul 2016 20:39:10 +0000 (20:39 +0000)]
Merge "Don't declare SYS_* constants for unavailable __NR_* constants."

8 years agoDon't declare SYS_* constants for unavailable __NR_* constants.
Elliott Hughes [Fri, 15 Jul 2016 22:47:47 +0000 (15:47 -0700)]
Don't declare SYS_* constants for unavailable __NR_* constants.

Bug: https://code.google.com/p/android/issues/detail?id=215853
Change-Id: Iaaa3ce888deb8d032208bf636b7badaed6a72d30

8 years agoMerge "resolve merge conflicts of cf85fd5 to stage-aosp-master" into stage-aosp-master
TreeHugger Robot [Fri, 15 Jul 2016 23:51:04 +0000 (23:51 +0000)]
Merge "resolve merge conflicts of cf85fd5 to stage-aosp-master" into stage-aosp-master

8 years agoresolve merge conflicts of cf85fd5 to stage-aosp-master
Dimitry Ivanov [Fri, 15 Jul 2016 23:49:14 +0000 (16:49 -0700)]
resolve merge conflicts of cf85fd5 to stage-aosp-master

Change-Id: Id73b6e03a224ce9a3e3838731be24a16692ef6e0

8 years agoresolve merge conflicts of cf85fd5 to stage-aosp-master
Colin Cross [Fri, 15 Jul 2016 22:49:51 +0000 (15:49 -0700)]
resolve merge conflicts of cf85fd5 to stage-aosp-master

Change-Id: I29bb554c9469109f999c6921e169625ad52cc515

8 years agoMerge "Convert linker from Android.mk to Android.bp"
Treehugger Robot [Fri, 15 Jul 2016 22:31:13 +0000 (22:31 +0000)]
Merge "Convert linker from Android.mk to Android.bp"

8 years agoMerge \"Put the legacy stdio fds back for pre-M.\"
Dan Albert [Fri, 15 Jul 2016 21:48:30 +0000 (21:48 +0000)]
Merge \"Put the legacy stdio fds back for pre-M.\"
am: 098d8ef5b2

Change-Id: I32565181dcfc9caab8212baa1c3f7a03f490ce0d

8 years agoMerge "Put the legacy stdio fds back for pre-M."
Treehugger Robot [Fri, 15 Jul 2016 21:33:00 +0000 (21:33 +0000)]
Merge "Put the legacy stdio fds back for pre-M."

8 years agoMerge \"Add an alias for _CTYPE_N.\"
Dan Albert [Fri, 15 Jul 2016 21:02:25 +0000 (21:02 +0000)]
Merge \"Add an alias for _CTYPE_N.\"
am: 730fa648c2

Change-Id: I84e7abe390146bfb2deb0987ccd997fd11c7d6aa

8 years agoMerge "Add an alias for _CTYPE_N."
Treehugger Robot [Fri, 15 Jul 2016 20:56:01 +0000 (20:56 +0000)]
Merge "Add an alias for _CTYPE_N."

8 years agoConvert linker from Android.mk to Android.bp
Colin Cross [Thu, 14 Jul 2016 23:05:46 +0000 (16:05 -0700)]
Convert linker from Android.mk to Android.bp

Change-Id: Ibf7da301defb2fd7e24fb3f4cd2e967a8f376497

8 years agoPut the legacy stdio fds back for pre-M.
Dan Albert [Fri, 15 Jul 2016 18:32:23 +0000 (11:32 -0700)]
Put the legacy stdio fds back for pre-M.

Change-Id: I2e397b04cba2969b9f9a3767b9b2fd12a289fcd2

8 years agoAdd an alias for _CTYPE_N.
Dan Albert [Fri, 15 Jul 2016 18:31:11 +0000 (11:31 -0700)]
Add an alias for _CTYPE_N.

gnustl is using this. It would be easy to change gnustl to *not* use
it for Android, but we've had this in released NDKs for years now so
we probably need this anyway.

Change-Id: I398a550664bf4ccbd9ea54f53c65428293a599a4

8 years agoMerge \"Force pthread_cond_timedwait_relative_np using CLOCK_MONOTONIC.\"
Yabin Cui [Fri, 15 Jul 2016 01:39:12 +0000 (01:39 +0000)]
Merge \"Force pthread_cond_timedwait_relative_np using CLOCK_MONOTONIC.\"
am: a24a2476d0

Change-Id: I5d1ebca87932f915fc33dd44d3b383e83b940d90

8 years agoMerge "Force pthread_cond_timedwait_relative_np using CLOCK_MONOTONIC."
Yabin Cui [Fri, 15 Jul 2016 01:29:28 +0000 (01:29 +0000)]
Merge "Force pthread_cond_timedwait_relative_np using CLOCK_MONOTONIC."

8 years agoMerge \"Fix clean mma in bionic\"
Colin Cross [Fri, 15 Jul 2016 01:00:44 +0000 (01:00 +0000)]
Merge \"Fix clean mma in bionic\"
am: 1e4a945f27

Change-Id: Ie6e69b74c7c1fa044db5e6eed58981ef0d57af28

8 years agoMerge "Fix clean mma in bionic"
Treehugger Robot [Fri, 15 Jul 2016 00:52:37 +0000 (00:52 +0000)]
Merge "Fix clean mma in bionic"

8 years agoAdd a test for a known kernel bug.
Elliott Hughes [Fri, 15 Jul 2016 00:07:17 +0000 (17:07 -0700)]
Add a test for a known kernel bug.

Bug: http://b/27265969
Change-Id: I981cb6a91015bd17d362987ed12c8ca638987c5b

8 years agoFix clean mma in bionic
Colin Cross [Thu, 14 Jul 2016 23:53:18 +0000 (16:53 -0700)]
Fix clean mma in bionic

Change-Id: If2410c5ff9c08372098e5ec831e33e4da0e04b30

8 years agoMerge \"Convert bionic benchmarks and tests to Android.bp\"
Colin Cross [Thu, 14 Jul 2016 21:11:23 +0000 (21:11 +0000)]
Merge \"Convert bionic benchmarks and tests to Android.bp\"
am: 55653cc33b

Change-Id: I577eb6db457bfbc520ca4528315715674173c489

8 years agoMerge "Convert bionic benchmarks and tests to Android.bp"
Treehugger Robot [Thu, 14 Jul 2016 21:05:27 +0000 (21:05 +0000)]
Merge "Convert bionic benchmarks and tests to Android.bp"

8 years agoMerge \"Fix typo in the comment\"
Dimitry Ivanov [Thu, 14 Jul 2016 20:14:51 +0000 (20:14 +0000)]
Merge \"Fix typo in the comment\"
am: d0e764f91e

Change-Id: Id4579211e39c2806debcda66c971473de7b2d2b6

8 years agoMerge "Fix typo in the comment"
Treehugger Robot [Thu, 14 Jul 2016 20:11:20 +0000 (20:11 +0000)]
Merge "Fix typo in the comment"

8 years agoMerge \"Convert malloc_debug from Android.mk to Android.bp\"
Colin Cross [Thu, 14 Jul 2016 19:47:18 +0000 (19:47 +0000)]
Merge \"Convert malloc_debug from Android.mk to Android.bp\"
am: 0393437b53

Change-Id: Iea3dbd038677e42075c6be87fceeb32e442ae152

8 years agoMerge "Convert malloc_debug from Android.mk to Android.bp"
Colin Cross [Thu, 14 Jul 2016 19:43:29 +0000 (19:43 +0000)]
Merge "Convert malloc_debug from Android.mk to Android.bp"

8 years agoMerge \"Move relocation_packer from Android.mk to Android.bp\"
Colin Cross [Thu, 14 Jul 2016 19:39:18 +0000 (19:39 +0000)]
Merge \"Move relocation_packer from Android.mk to Android.bp\"
am: 41933386c0

Change-Id: Iaac010c42c1c1d75cd36645a8b25e89a05abac06

8 years agoMerge "Move relocation_packer from Android.mk to Android.bp"
Treehugger Robot [Thu, 14 Jul 2016 19:34:47 +0000 (19:34 +0000)]
Merge "Move relocation_packer from Android.mk to Android.bp"

8 years agoConvert bionic benchmarks and tests to Android.bp
Colin Cross [Mon, 11 Jul 2016 23:20:06 +0000 (16:20 -0700)]
Convert bionic benchmarks and tests to Android.bp

The compile-time tests and a few custom libraries for dynamic linker
testing are still compiled in make.

Also converts the make rules to run tests on the host to shell scripts
in tests/run-on-host.sh and benchmarks/run-on-host.sh

Change-Id: I6f174b3a69d58c4ed74d29f4e79332d483681534

8 years agoMerge \"linker: Improve elf-file validation\"
Dimitry Ivanov [Thu, 14 Jul 2016 18:18:11 +0000 (18:18 +0000)]
Merge \"linker: Improve elf-file validation\"
am: 51f64197ae

Change-Id: I0093b7fda86f7445b5a6cbe55b7237cba3020af6

8 years agoFix typo in the comment
Dimitry Ivanov [Thu, 14 Jul 2016 18:15:44 +0000 (11:15 -0700)]
Fix typo in the comment

Addressing review comment from change I11bc2567b0cff89f48699ec74015991fee5b137b

Change-Id: I5d479ef61421c86fae06167815aa0a2681d9142d

8 years agoMerge "linker: Improve elf-file validation"
Dimitry Ivanov [Thu, 14 Jul 2016 18:12:59 +0000 (18:12 +0000)]
Merge "linker: Improve elf-file validation"

8 years agoMerge \"Support 32-bit ARM vdso.\"
Elliott Hughes [Thu, 14 Jul 2016 17:32:57 +0000 (17:32 +0000)]
Merge \"Support 32-bit ARM vdso.\"
am: 85f900d589

Change-Id: I20c259c574b519e10efc054b7f2ae13ec9f7b451

8 years agoMerge "Support 32-bit ARM vdso."
Elliott Hughes [Thu, 14 Jul 2016 16:54:23 +0000 (16:54 +0000)]
Merge "Support 32-bit ARM vdso."

8 years agoConvert malloc_debug from Android.mk to Android.bp
Colin Cross [Wed, 13 Jul 2016 18:15:21 +0000 (11:15 -0700)]
Convert malloc_debug from Android.mk to Android.bp

Change-Id: I60ad376d222a92a29263c2804dcd1bd0d3c383d0

8 years agoMove relocation_packer from Android.mk to Android.bp
Colin Cross [Wed, 13 Jul 2016 21:01:45 +0000 (14:01 -0700)]
Move relocation_packer from Android.mk to Android.bp

Change-Id: I3d24f8513fe3ca19900705d4861ab797464e0930

8 years agoMerge \"Fix dlext.ns_smoke test\"
Dimitry Ivanov [Thu, 14 Jul 2016 00:33:54 +0000 (00:33 +0000)]
Merge \"Fix dlext.ns_smoke test\"
am: 48cedda02a

Change-Id: Ib234e29c9e20b1a4133124b770c9674bd0a9dbdd

8 years agoMerge "Fix dlext.ns_smoke test"
Treehugger Robot [Thu, 14 Jul 2016 00:22:34 +0000 (00:22 +0000)]
Merge "Fix dlext.ns_smoke test"

8 years agolinker: Improve elf-file validation
Dimitry Ivanov [Thu, 14 Jul 2016 00:06:36 +0000 (17:06 -0700)]
linker: Improve elf-file validation

1. Make sure that the .dynamic section offset and size matches
   PT_DYNAMIC segment offset and filesz
2. No section offset can be 0

Bug: http://b/16548758
Bug: http://b/29637134
Change-Id: I11bc2567b0cff89f48699ec74015991fee5b137b
Test: Install and start the app from http://b/29637134

8 years agoFix dlext.ns_smoke test
Dimitry Ivanov [Wed, 13 Jul 2016 23:28:20 +0000 (16:28 -0700)]
Fix dlext.ns_smoke test

Check for updated error_message

Change-Id: I0f0dd59a6deba016d37d0ccc80113b62db86c1c2
Test: adb shell bionic-unit-tests32/64 --gtest_filter=dlext.ns*

8 years agoForce pthread_cond_timedwait_relative_np using CLOCK_MONOTONIC.
Yabin Cui [Wed, 13 Jul 2016 22:53:25 +0000 (15:53 -0700)]
Force pthread_cond_timedwait_relative_np using CLOCK_MONOTONIC.

Previous patch changed pthread_cond_timedwait_relative_np to use
CLOCK_REALTIME, which causes app compatibility problem. So change
it back to CLOCK_MONOTONIC.

Bug: 30106240
Change-Id: I8e04058e92ede098f4f9f8d133f094001921441e

8 years agoMerge \"Use trivial types to avoid calls to new for globals\"
Dimitry Ivanov [Wed, 13 Jul 2016 18:52:31 +0000 (18:52 +0000)]
Merge \"Use trivial types to avoid calls to new for globals\"
am: cc55bd8313

Change-Id: I5b594d67c33e51b1b3c6013beca618e159cb409e

8 years agoMerge "Use trivial types to avoid calls to new for globals"
Dimitry Ivanov [Wed, 13 Jul 2016 18:45:00 +0000 (18:45 +0000)]
Merge "Use trivial types to avoid calls to new for globals"

8 years agoUse trivial types to avoid calls to new for globals
Dimitry Ivanov [Wed, 13 Jul 2016 17:24:06 +0000 (10:24 -0700)]
Use trivial types to avoid calls to new for globals

To avoid initialization ordering problem avoid constructing
non-trivial global objects for linker_logger.

This patch also adds explicit check for bloc_size_ on alloc

Change-Id: I2cb3b13a10b7fac1f52a20038af77266d3e1da96

8 years agoMerge \"linker: Enable debug logging via properties\"
Dimitry Ivanov [Tue, 12 Jul 2016 22:37:55 +0000 (22:37 +0000)]
Merge \"linker: Enable debug logging via properties\"
am: c8d5d2a524

Change-Id: Ia24feca7951934046f7255fbcc88f7a4cbe0c6e8

8 years agoMerge "linker: Enable debug logging via properties"
Treehugger Robot [Tue, 12 Jul 2016 22:27:19 +0000 (22:27 +0000)]
Merge "linker: Enable debug logging via properties"

8 years agolinker: Enable debug logging via properties
Dimitry Ivanov [Tue, 12 Jul 2016 01:11:39 +0000 (18:11 -0700)]
linker: Enable debug logging via properties

This patch lets developers tune logging of dlopen/dlerror with
setting system property.

Note that for security purposes this option is disabled on user build
for non-debuggable apps.

For starters there are 3 debug options:
 dlerror - enables logging of all dlerrors
 dlopen - traces dlopen calls

To enable system-wide logging (works only for userdebug/eng builds)
use debug.ld.all property.

To enable logging for particular app use debug.ld.app.<appname> property.

Example: Running "adb shell setprop debug.ld.all dlerror,dlopen" will log all
dlerror message as well as trace all calls to dlopen.

Bug: http://b/29458203
Change-Id: I2392c80a795509e16fe5689d0500d18b99772a64

8 years agoSupport 32-bit ARM vdso.
Elliott Hughes [Tue, 12 Jul 2016 20:55:51 +0000 (13:55 -0700)]
Support 32-bit ARM vdso.

Linux 4.1 added this.

Bug: http://b/19198045
Change-Id: I28be802ff403a61dd6733a001411b3ff05fef5a5

8 years agoMerge \"Rename Android.soong.mk to Android.mk\"
Colin Cross [Tue, 12 Jul 2016 16:20:25 +0000 (16:20 +0000)]
Merge \"Rename Android.soong.mk to Android.mk\"
am: 7fd1e75e24

Change-Id: I6ee75b0f4f53667eb820542529febcbe1751a494

8 years agoMerge "Rename Android.soong.mk to Android.mk"
Colin Cross [Tue, 12 Jul 2016 16:15:53 +0000 (16:15 +0000)]
Merge "Rename Android.soong.mk to Android.mk"

8 years agoRename Android.soong.mk to Android.mk
Colin Cross [Mon, 11 Jul 2016 23:54:20 +0000 (16:54 -0700)]
Rename Android.soong.mk to Android.mk

Soong is always on now, so we don't have to distinguish between
makefiles that should be ignored because Soong is handling them with an
Android.bp file, and makefiles that are still needed with Soong.  All
obsolete Android.mk files have been removed, rename all Android.soong.mk
files to Android.mk.

Change-Id: Ic0494e1800dec13a9f5714559cffe3a5ebe49418

8 years agoMerge \"Update the kernel header documentation.\"
Christopher Ferris [Fri, 8 Jul 2016 22:36:56 +0000 (22:36 +0000)]
Merge \"Update the kernel header documentation.\"
am: 4580ea4f16

Change-Id: I69c3d6edb6107183da96f4bf5b36d15cc50dc525

8 years agoMerge "Update the kernel header documentation."
Christopher Ferris [Fri, 8 Jul 2016 22:31:14 +0000 (22:31 +0000)]
Merge "Update the kernel header documentation."

8 years agoUpdate the kernel header documentation.
Christopher Ferris [Fri, 8 Jul 2016 22:25:21 +0000 (15:25 -0700)]
Update the kernel header documentation.

Change the references to using the android kernel source, to use
the linux stable source code repository.

Change-Id: Ibc6c90fc67c80bef235459bb3af31656f4d403ac

8 years agoMerge \"Add the record alloc option.\"
Christopher Ferris [Fri, 8 Jul 2016 22:05:00 +0000 (22:05 +0000)]
Merge \"Add the record alloc option.\"
am: 779aa5ac33

Change-Id: Icdda3305fef2804b92692b31459fcfde2c83de1c

8 years agoMerge "Add the record alloc option."
Treehugger Robot [Fri, 8 Jul 2016 21:59:18 +0000 (21:59 +0000)]
Merge "Add the record alloc option."

8 years agoAdd the record alloc option.
Christopher Ferris [Wed, 20 Apr 2016 19:30:58 +0000 (12:30 -0700)]
Add the record alloc option.

This option adds the ability to record all of the allocation requests
and dump them to a file when a signal is sent to the process.

Included in this change, redo the option processing to add a new
string option.

Bug: 27747898

Change-Id: Ida043362e38b5eb1d459c99db9c2581015dab366

8 years agoMerge "Update timezone data to 2016f" into stage-aosp-master
Neil Fuller [Thu, 7 Jul 2016 15:50:37 +0000 (15:50 +0000)]
Merge "Update timezone data to 2016f" into stage-aosp-master

8 years agoUpdate timezone data to 2016f
Neil Fuller [Wed, 6 Jul 2016 14:26:15 +0000 (15:26 +0100)]
Update timezone data to 2016f

The 2016f release of the tz code and data is available. It reflects the
following changes, which were either circulated on the tz mailing list
or are relatively minor technical or administrative changes:

   Changes affecting future time stamps

     The Egyptian government changed its mind on short notice, and
     Africa/Cairo will not introduce DST starting 2016-07-07 after all.
     (Thanks to Mina Samuel.)

     Asia/Novosibirsk switches from +06 to +07 on 2016-07-24 at 02:00.
     (Thanks to Stepan Golosunov.)

   Changes to past and future time stamps

     Asia/Novokuznetsk and Asia/Novosibirsk now use numeric time zone
     abbreviations instead of invented ones.

   Changes affecting past time stamps

     Europe/Minsk's 1992-03-29 spring-forward transition was at 02:00
not 00:00.
     (Thanks to Stepan Golosunov.)

Note: Paul Eggert changed the key used to sign the tzdata bundle and
the new public key cannot be found. The changes to the data since
2016e have been verified visually.

Test: Ran CTS/android.core.tests.libcore.package.harmony_java_util
Test: Ran CTS/android.core.tests.libcore.package.harmony_java_text
Test: Ran CTS/android.core.tests.libcore.package.libcore
Bug: 29991153
Bug: 30009483
(cherry picked from commit 516f09c963f4097e3fa0ee99a8d5537ecb56b2ab)

Change-Id: I3b1c9552ac7e1f9f436bd150d4c92916e54fc654

8 years agoMerge \"Update timezone data to 2016f\"
Neil Fuller [Thu, 7 Jul 2016 15:13:12 +0000 (15:13 +0000)]
Merge \"Update timezone data to 2016f\"
am: a0e205e3b4

Change-Id: I6f76a9c1183b601810debfba7fddeb30e543c006

8 years agoMerge "Update timezone data to 2016f"
Neil Fuller [Thu, 7 Jul 2016 12:57:02 +0000 (12:57 +0000)]
Merge "Update timezone data to 2016f"

8 years agoUpdate timezone data to 2016f
Neil Fuller [Wed, 6 Jul 2016 14:26:15 +0000 (15:26 +0100)]
Update timezone data to 2016f

The 2016f release of the tz code and data is available. It reflects the
following changes, which were either circulated on the tz mailing list
or are relatively minor technical or administrative changes:

   Changes affecting future time stamps

     The Egyptian government changed its mind on short notice, and
     Africa/Cairo will not introduce DST starting 2016-07-07 after all.
     (Thanks to Mina Samuel.)

     Asia/Novosibirsk switches from +06 to +07 on 2016-07-24 at 02:00.
     (Thanks to Stepan Golosunov.)

   Changes to past and future time stamps

     Asia/Novokuznetsk and Asia/Novosibirsk now use numeric time zone
     abbreviations instead of invented ones.

   Changes affecting past time stamps

     Europe/Minsk's 1992-03-29 spring-forward transition was at 02:00
not 00:00.
     (Thanks to Stepan Golosunov.)

Note: Paul Eggert changed the key used to sign the tzdata bundle and
the new public key cannot be found. The changes to the data since
2016e have been verified visually.

Change-Id: I11b49d9916e67313978fd893acf091e9f99f6d33
Test: Ran CTS/android.core.tests.libcore.package.harmony_java_util
Test: Ran CTS/android.core.tests.libcore.package.harmony_java_text
Test: Ran CTS/android.core.tests.libcore.package.libcore
Bug: 29991153

8 years agoMerge \"x86-64\'s va_list is not a pointer type.\"
Elliott Hughes [Thu, 30 Jun 2016 20:53:18 +0000 (20:53 +0000)]
Merge \"x86-64\'s va_list is not a pointer type.\"
am: 204c4d1b80

Change-Id: I40675ac38f8442bb4ee66c048752919f7beb099e

8 years agoMerge \"Revert \"Skip Clang\'s integrated-as for hand-coded assembly\"\"
Pirama Arumuga Nainar [Thu, 30 Jun 2016 20:49:36 +0000 (20:49 +0000)]
Merge \"Revert \"Skip Clang\'s integrated-as for hand-coded assembly\"\"
am: 1fec29e77b

Change-Id: I0e3caf46128a44f8cfdd167547817e2d8efe3b5c

8 years agoMerge "x86-64's va_list is not a pointer type."
Elliott Hughes [Thu, 30 Jun 2016 20:48:04 +0000 (20:48 +0000)]
Merge "x86-64's va_list is not a pointer type."

8 years agox86-64's va_list is not a pointer type.
Elliott Hughes [Thu, 30 Jun 2016 20:47:16 +0000 (13:47 -0700)]
x86-64's va_list is not a pointer type.

Change-Id: I680feeb641d5411ac481daaae6fa1580e74c5563

8 years agoMerge "Revert "Skip Clang's integrated-as for hand-coded assembly""
Treehugger Robot [Thu, 30 Jun 2016 20:33:41 +0000 (20:33 +0000)]
Merge "Revert "Skip Clang's integrated-as for hand-coded assembly""

8 years agoMerge \"Apparently mips and x86\'s va_list is a pointer type.\"
Elliott Hughes [Thu, 30 Jun 2016 19:55:44 +0000 (19:55 +0000)]
Merge \"Apparently mips and x86\'s va_list is a pointer type.\"
am: 60be638ae6

Change-Id: I72cf962f53e703b2b52529f43cc2cf8b1ad106f3

8 years agoMerge "Apparently mips and x86's va_list is a pointer type."
Elliott Hughes [Thu, 30 Jun 2016 19:48:19 +0000 (19:48 +0000)]
Merge "Apparently mips and x86's va_list is a pointer type."

8 years agoApparently mips and x86's va_list is a pointer type.
Elliott Hughes [Thu, 30 Jun 2016 19:46:21 +0000 (12:46 -0700)]
Apparently mips and x86's va_list is a pointer type.

And clang won't let you have a function declaration where some arguments
have nullability specifiers and others don't.

Change-Id: I450b0221a3f7f068d5fe971dfbc0ba91d25710e8

8 years agoMerge \"Remove __nonnull (in favor of _Nonnull).\"
Elliott Hughes [Thu, 30 Jun 2016 18:18:38 +0000 (18:18 +0000)]
Merge \"Remove __nonnull (in favor of _Nonnull).\"
am: 3b4e11faa8

Change-Id: I5b844da400d21cc64730065f2f9a2111a89ea560

8 years agoMerge "Remove __nonnull (in favor of _Nonnull)."
Treehugger Robot [Thu, 30 Jun 2016 18:04:38 +0000 (18:04 +0000)]
Merge "Remove __nonnull (in favor of _Nonnull)."

8 years agoRemove __nonnull (in favor of _Nonnull).
Elliott Hughes [Thu, 30 Jun 2016 16:12:40 +0000 (09:12 -0700)]
Remove __nonnull (in favor of _Nonnull).

It's a dangerous pain in the ass, and we're only using it in one place, and
that's a mistake too.

Change-Id: Ida1d1286aaa1343bdee95fa092fcb45ff3258f77

8 years agoRevert "Skip Clang's integrated-as for hand-coded assembly"
Pirama Arumuga Nainar [Thu, 30 Jun 2016 00:16:54 +0000 (00:16 +0000)]
Revert "Skip Clang's integrated-as for hand-coded assembly"

This reverts commit 8d414dd2b1d9503f6d4b3d9310796be50a824ede.

Reverting this since https://android-review.googlesource.com/#/c/244800/ adds -fno-integrated-as globally.

Change-Id: I29f53a385c16f44ead5dc744b0f2dd18a4af1731

8 years agoMerge changes I98c9329f,I40936f7b,I09d20ff1
Josh Gao [Wed, 29 Jun 2016 21:56:02 +0000 (21:56 +0000)]
Merge changes I98c9329f,I40936f7b,I09d20ff1
am: 1159c196db

Change-Id: Id118e7961eaeb3fa3f7c40abcbf158e45bee32f6

8 years agoMerge changes I98c9329f,I40936f7b,I09d20ff1
Treehugger Robot [Wed, 29 Jun 2016 21:45:34 +0000 (21:45 +0000)]
Merge changes I98c9329f,I40936f7b,I09d20ff1

* changes:
  Defer registration of the arc4random fork-detect handler.
  Make getpid work before the main thread is initialized.
  Take the arc4random lock before forking.

8 years agoDefer registration of the arc4random fork-detect handler.
Josh Gao [Fri, 24 Jun 2016 23:18:21 +0000 (16:18 -0700)]
Defer registration of the arc4random fork-detect handler.

Previously, arc4random would register a fork-detecting pthread_atfork
handler to not have to call getpid() after a fork. pthread_atfork uses
pthread_mutex_lock, which requires the current thread to be initialized,
preventing the use of arc4random for initializing the global stack guard,
which needs to happen before the main thread has been initialized.

Extract the arc4random fork-detection flag and use the existing
arc4random fork handler to set it.

Bug: http://b/29622562
Change-Id: I98c9329fa0e489c3f78cad52747eaaf2f5226b80

8 years agoMake getpid work before the main thread is initialized.
Josh Gao [Fri, 24 Jun 2016 23:13:33 +0000 (16:13 -0700)]
Make getpid work before the main thread is initialized.

Bug: http://b/29622562
Test: code dependent on this change no longer crashes
Change-Id: I40936f7b35d9e58182aeb2e34e52f54088700825

8 years agoTake the arc4random lock before forking.
Josh Gao [Fri, 24 Jun 2016 20:04:09 +0000 (13:04 -0700)]
Take the arc4random lock before forking.

Bug: http://b/24675038
Test: stepped through a fork call in gdb
Change-Id: I09d20ff1d103d0c005f2a0cdd9b0a8710ab2392c

8 years agoMerge \\"Skip Clang\\'s integrated-as for hand-coded assembly\\" am: 9158386758
Pirama Arumuga Nainar [Mon, 27 Jun 2016 22:51:14 +0000 (22:51 +0000)]
Merge \\"Skip Clang\\'s integrated-as for hand-coded assembly\\" am: 9158386758
am: 476b301eb1

Change-Id: Ia8c325990aa409f3d89ecd12689b0e0f8ef52778