OSDN Git Service
Dimitry Ivanov [Wed, 25 Jan 2017 03:56:47 +0000 (03:56 +0000)]
Do not send 0 byte string content.
am:
cafd355375
Change-Id: I8c2cc596f4f5d6e9ecc0de6130be8bdfcb19c1ec
Dimitry Ivanov [Wed, 25 Jan 2017 03:56:42 +0000 (03:56 +0000)]
Revert "Revert "Remove limit of system property name length""
am:
16b2a4de14
Change-Id: I585a55769cdca1d7f9edee479b5ec94d8821b673
Dimitry Ivanov [Tue, 24 Jan 2017 20:39:33 +0000 (12:39 -0800)]
Do not send 0 byte string content.
Trying to send even 0 bytes to closed socket leads to
broken pipe error. Sometimes property service is just
quick enough and closes the socket between send(valuelen)
and send(value) in the case where valuelen is 0.
Bug: http://b/
34670529
Test: adb reboot 20 times and make sure phone service did not fail
Test: run bionic-unit-tests --gtest_filter=prop*
Change-Id: I96f90ca6fe1790614e7efd3015bffed1ef1e9040
Evgenii Stepanov [Tue, 24 Jan 2017 21:02:40 +0000 (21:02 +0000)]
Merge "[cfi] Handle large libraries correctly."
am:
d8388cd303
Change-Id: I3032727f90e26837b69b558551ccd420ae7bc680
Dimitry Ivanov [Tue, 24 Jan 2017 20:43:29 +0000 (20:43 +0000)]
Revert "Revert "Remove limit of system property name length""
This reverts commit
489f58b5eaedd5a80635bb3a7b39e97037c585f6.
Bug: http://b/
33926793
Bug: http://b/
34670529
Test: Run bionic-unit-tests --gtest_filter=prop*
Change-Id: Id4e94652dc2310a21f5b7bd3af098bf79df3f380
Evgenii Stepanov [Tue, 24 Jan 2017 20:56:28 +0000 (20:56 +0000)]
Merge "[cfi] Handle large libraries correctly."
Dimitry Ivanov [Tue, 24 Jan 2017 20:05:38 +0000 (20:05 +0000)]
Merge "Revert "Remove limit of system property name length""
am:
40250bb683
Change-Id: Ic2565a4ad5d42146378364c06250260ca466cc98
Dimitry Ivanov [Tue, 24 Jan 2017 20:00:43 +0000 (20:00 +0000)]
Merge "Revert "Remove limit of system property name length""
Dimitry Ivanov [Tue, 24 Jan 2017 18:39:04 +0000 (18:39 +0000)]
Revert "Remove limit of system property name length"
This reverts commit
5c1ce278f31bc6fc338b9b951f1adc7c2b343d7c.
Bug: http://b/
33926793
Bug: http://b/
34670529
Change-Id: I0dc4a8ae55576c69b34b2958d8e664f7066b9c54
Vijay Venkatraman [Tue, 24 Jan 2017 17:40:50 +0000 (17:40 +0000)]
Merge "Removed include log\logger.h from log_fake.cpp"
am:
f0b42ae973
Change-Id: Iba9012e9206234e104c355443803d278a861b336
Vijay Venkatraman [Tue, 24 Jan 2017 17:36:08 +0000 (17:36 +0000)]
Merge "Removed include log\logger.h from log_fake.cpp"
Vijay Venkatraman [Tue, 24 Jan 2017 02:56:53 +0000 (18:56 -0800)]
Removed include log\logger.h from log_fake.cpp
Test: Compile malloc_debug_unit_tests
Change-Id: I635687016570ba52c571435b87d8e89f7f04d609
Dimitry Ivanov [Tue, 24 Jan 2017 02:56:24 +0000 (02:56 +0000)]
Merge "Remove limit of system property name length"
am:
440d6c7644
Change-Id: I3ff9fb98065166cb53dfe153448e059df9ca6a71
Dimitry Ivanov [Tue, 24 Jan 2017 02:54:33 +0000 (02:54 +0000)]
Merge "Remove limit of system property name length"
Dimitry Ivanov [Tue, 24 Jan 2017 00:42:31 +0000 (00:42 +0000)]
Merge "Remove libc++ dependency from prebuilt test-libs"
am:
8bb01c0719
Change-Id: I2d28e3f2be50a9726e4712bf7f731c1ecd215e47
Treehugger Robot [Tue, 24 Jan 2017 00:35:59 +0000 (00:35 +0000)]
Merge "Remove libc++ dependency from prebuilt test-libs"
Dimitry Ivanov [Tue, 24 Jan 2017 00:27:32 +0000 (00:27 +0000)]
Merge "Mark ifunc tests failing on arm as known failures"
am:
3685e4515f
Change-Id: I1bf5509ecb820df66e28aff3b80b8c1cdb857b53
Treehugger Robot [Tue, 24 Jan 2017 00:24:40 +0000 (00:24 +0000)]
Merge "Mark ifunc tests failing on arm as known failures"
Dimitry Ivanov [Mon, 23 Jan 2017 22:48:21 +0000 (14:48 -0800)]
Mark ifunc tests failing on arm as known failures
Bug: http://b/
27930475
Test: run bionic-unit-tests on arm and arm64
Change-Id: I5084b6a5ef4b56ca9985da63a5e6896f771810f0
Dimitry Ivanov [Fri, 20 Jan 2017 23:47:03 +0000 (15:47 -0800)]
Remove libc++ dependency from prebuilt test-libs
Bug: http://b/
27549327
Test: run bionic-unit-tests --gtest_filter=dl*:Dl* on angler
Change-Id: Ia09ed2c00213bc07bf2a4c772e675419f7a30cb3
Dimitry Ivanov [Thu, 3 Dec 2015 19:26:38 +0000 (11:26 -0800)]
Remove limit of system property name length
This change introduces new __system_property_read_callback
method to use in place of deprecated __system_property_read
__system_property_set() and get() should just work but now
do not have limit on system property names.
Bug: http://b/
33926793
Test: boot device, run adb shell propget
Test: boot device with old version of init (protocol v1)
Test: run bionic-unit-tests --gtest_filter=prop*
Change-Id: I619fb5a7e27a272aac30011579665f6160888bc7
Evgenii Stepanov [Fri, 20 Jan 2017 21:47:04 +0000 (13:47 -0800)]
[cfi] Handle large libraries correctly.
Fallback to unchecked if the shadow offset overflows int16_t.
This may happen when a library's data segment is larger than 256MB.
Also updated some comments.
Bug:
22033465
Test: bionic device tests
Change-Id: I8eef42f75099f24aed566499ff1731a0bbf01ff3
Jaekyun Seok [Fri, 20 Jan 2017 04:02:14 +0000 (04:02 +0000)]
Merge "Add props files for each partition"
am:
7844b4c5db
Change-Id: I25c7789f689e3aadcca99e5fd8a6908db9bc57d2
Treehugger Robot [Fri, 20 Jan 2017 03:58:51 +0000 (03:58 +0000)]
Merge "Add props files for each partition"
Jaekyun Seok [Tue, 17 Jan 2017 05:44:31 +0000 (14:44 +0900)]
Add props files for each partition
The following files will be added.
- /vendor/default.prop for default prop overrides from vendor
partition.
- /odm/default.prop for default prop overrides from odm partition.
- /odm/build.prop for build prop overrides from odm partition.
Test: tested default/build prop files with enabling early mount, but
didn't test files of odm partition because odm partition doesn't
exist now.
Bug:
34116668
Change-Id: Ia99895a0afcdc522463d95b2f3e2841dcf2e3516
Paul Lawrence [Thu, 19 Jan 2017 23:30:28 +0000 (23:30 +0000)]
Merge "Add seccomp support library"
am:
ad3b3995e7
Change-Id: I1115c8c39de622d594447a56d2cd0893991effac
Treehugger Robot [Thu, 19 Jan 2017 23:26:09 +0000 (23:26 +0000)]
Merge "Add seccomp support library"
Paul Lawrence [Fri, 11 Nov 2016 19:33:42 +0000 (11:33 -0800)]
Add seccomp support library
Policy library which exports an autogenerated policy from SYSCALLS.TXT
blocking any other calls.
Test: Generate policy, install onto Sailfish, check boots, Chrome runs,
calls are blocked.
Bug:
32313202
Change-Id: Ib590704e50122f077eeae26561eb9b0a70386551
Evgenii Stepanov [Wed, 18 Jan 2017 23:45:54 +0000 (23:45 +0000)]
Merge "Runtime support for CFI"
am:
082bec5b41
Change-Id: I47ea89f7161ff7c00a89a291ba7386c2b6f5c191
Evgenii Stepanov [Wed, 18 Jan 2017 23:39:07 +0000 (23:39 +0000)]
Merge "Runtime support for CFI"
Josh Gao [Wed, 18 Jan 2017 23:25:03 +0000 (23:25 +0000)]
Merge changes from topic 'kill_debuggerd'
am:
2d261e8e9e
Change-Id: Ibec7e7093bbe580cc126225c2f2d1f280de51a43
Josh Gao [Wed, 18 Jan 2017 23:16:47 +0000 (23:16 +0000)]
Rename libdebuggerd_client to libdebuggerd_handler.
am:
2a3b4fa6ec
Change-Id: Iac8466c0995571ce51a81f903903115442f5bc17
Josh Gao [Wed, 18 Jan 2017 23:16:39 +0000 (23:16 +0000)]
Reserve a realtime signal for debuggerd.
am:
f8a6d4745d
Change-Id: If7eced969434e50beec04625c905ef374cdcdfb1
Josh Gao [Wed, 18 Jan 2017 23:04:08 +0000 (23:04 +0000)]
Merge changes from topic 'kill_debuggerd'
* changes:
Rename libdebuggerd_client to libdebuggerd_handler.
Reserve a realtime signal for debuggerd.
Evgenii Stepanov [Wed, 6 Jul 2016 20:20:59 +0000 (13:20 -0700)]
Runtime support for CFI
Control Flow Integrity support in bionic.
General design:
http://clang.llvm.org/docs/ControlFlowIntegrityDesign.html#shared-library-support
This CL implements subsections "CFI Shadow" and "CFI_SlowPath" in the above document.
Bug:
22033465
Test: bionic device tests
Change-Id: I14dfea630de468eb5620e7f55f92b1397ba06217
Colin Cross [Wed, 18 Jan 2017 16:59:52 +0000 (16:59 +0000)]
Merge "Declare dependency on android_filesystem_config.h"
am:
2f238b32cd
Change-Id: Ia69ebf6c7d1aeefc61ff301197d3d994a009b477
Colin Cross [Wed, 18 Jan 2017 16:53:25 +0000 (16:53 +0000)]
Merge "Declare dependency on android_filesystem_config.h"
Jorim Jaggi [Wed, 18 Jan 2017 16:47:31 +0000 (16:47 +0000)]
Merge "Revert "property_contexts: split into platform and non-platform components""
am:
dbfa1bd292
Change-Id: I7279216d6987fa853ee759997e2c571fc01ce5ed
Jorim Jaggi [Wed, 18 Jan 2017 16:41:36 +0000 (16:41 +0000)]
Revert "property_contexts: split into platform and non-platform components"
am:
e2f5ea14c5
Change-Id: I9a794b93b077791ce46016cb9f8965a52c98b88e
Jorim Jaggi [Wed, 18 Jan 2017 16:33:31 +0000 (16:33 +0000)]
Merge "Revert "property_contexts: split into platform and non-platform components""
Jorim Jaggi [Wed, 18 Jan 2017 15:37:18 +0000 (15:37 +0000)]
Revert "property_contexts: split into platform and non-platform components"
This reverts commit
8213615461f7d39620e44f73980acc6107c76ac7.
Change-Id: Id2982eafadc5d1c307b60c1f8abe78df8dc8b82e
Sandeep Patil [Wed, 18 Jan 2017 06:16:50 +0000 (06:16 +0000)]
Merge "property_contexts: split into platform and non-platform components"
am:
0b6d0a8b38
Change-Id: I10c9caa9f85f21e681b8ab02cd3ea5e8ef489eb0
Sandeep Patil [Wed, 18 Jan 2017 06:12:19 +0000 (06:12 +0000)]
property_contexts: split into platform and non-platform components
am:
8213615461
Change-Id: Ifa1bb1e295cef8ecf74a4a4702114a88f3e6859e
Treehugger Robot [Wed, 18 Jan 2017 06:09:02 +0000 (06:09 +0000)]
Merge "property_contexts: split into platform and non-platform components"
Colin Cross [Wed, 18 Jan 2017 02:16:07 +0000 (18:16 -0800)]
Declare dependency on android_filesystem_config.h
android_filesystem_config.h comes from a directory outside bionic/libc
so it can't be referenced directly, add it as a source file through
a filegroup module.
Bug:
34283327
Test: builds
Test: libc rebuilds after touch system/core/include/private/android_filesystem_config.h
Change-Id: I90f6b7b25b70842b8619d558074449f13e6e6b03
Josh Gao [Thu, 27 Oct 2016 00:55:49 +0000 (17:55 -0700)]
Rename libdebuggerd_client to libdebuggerd_handler.
Bug: http://b/
30705528
Test: debuggerd_test
Change-Id: I625a9a2e7304d172873ba9344b74efa57ad7bdad
Josh Gao [Wed, 26 Oct 2016 20:27:19 +0000 (13:27 -0700)]
Reserve a realtime signal for debuggerd.
Allocate __SIGRTMIN + 3 for triggering native stack dumps (like SIGQUIT
for Java processes).
Bug: http://b/
30705528
Test: debuggerd_test
Change-Id: I81d622fba32a651555268a8a2ba6721c61c93a58
Elliott Hughes [Sat, 14 Jan 2017 18:35:50 +0000 (18:35 +0000)]
Merge "linker: convert 'mov pc' instruction to 'bx'"
am:
8d0e0d4952
Change-Id: I8664911d4c068c9c719f85e44b7a5ac99670581f
Elliott Hughes [Sat, 14 Jan 2017 18:30:09 +0000 (18:30 +0000)]
Merge "linker: convert 'mov pc' instruction to 'bx'"
Park Ju Hyung [Thu, 12 Jan 2017 20:52:06 +0000 (05:52 +0900)]
linker: convert 'mov pc' instruction to 'bx'
From Linux 3.17
6ebbf2ce437b33022d30badd49dc94d33ecfa498:
ARMv6 and greater introduced a new instruction ("bx") which can be used
to return from function calls. Recent CPUs perform better when the
"bx lr" instruction is used rather than the "mov pc, lr" instruction,
and this sequence is strongly recommended to be used by the ARM
architecture manual (section A.4.1.1).
Signed-off-by: Park Ju Hyung <qkrwngud825@gmail.com>
Test: No regressions detected
Test: Passes full CTS run
Change-Id: Ie268f9893e3df0f68fbfe82a13f3c7cc5c5909d8
Signed-off-by: Alex Naidis <alex.naidis@linux.com>
Yabin Cui [Sat, 14 Jan 2017 02:24:35 +0000 (02:24 +0000)]
Merge "Switch bionic-unit-tests to libc++_static."
am:
37f1d11482
Change-Id: Ie427d277e5f1997a170e90cf98846cf1244d7a84
Yabin Cui [Sat, 14 Jan 2017 02:20:19 +0000 (02:20 +0000)]
Merge "Switch bionic-unit-tests to libc++_static."
Elliott Hughes [Fri, 13 Jan 2017 21:16:00 +0000 (21:16 +0000)]
Merge "Describe our dynamic linker compatibility policy."
am:
98d200899b
Change-Id: I247f8d16edb201e251670e699a3a16244a4e510e
Elliott Hughes [Fri, 13 Jan 2017 21:12:14 +0000 (21:12 +0000)]
Merge "Describe our dynamic linker compatibility policy."
Yabin Cui [Fri, 13 Jan 2017 20:31:59 +0000 (12:31 -0800)]
Switch bionic-unit-tests to libc++_static.
Bug: http://b/
27549327
Test: run bionic-unit-tests, bionic-unit-tests-glibc.
Change-Id: I719899624dec83728a3d5b8169079e5c91c63c04
Elliott Hughes [Fri, 13 Jan 2017 19:07:00 +0000 (11:07 -0800)]
Describe our dynamic linker compatibility policy.
Bug: N/A
Test: N/A
Change-Id: I6f472d1ab22636de74fd103e9a480fb1ab51e314
Pavel Labath [Fri, 13 Jan 2017 10:24:46 +0000 (10:24 +0000)]
Merge "Beef up process_vm_readv/writev tests"
am:
a970ac50d1
Change-Id: Ic1628705232788fdd871c0e91afe933707a15169
Pavel Labath [Fri, 13 Jan 2017 10:21:31 +0000 (10:21 +0000)]
Merge "Beef up process_vm_readv/writev tests"
Pavel Labath [Thu, 12 Jan 2017 11:56:39 +0000 (11:56 +0000)]
Beef up process_vm_readv/writev tests
LLDB uses process_vm_readv quite extensively (it's an order of
magnitude faster than PTRACE_PEEKDATA). Add a bit more tests of
the function to make sure it stays working.
Change-Id: I64e17e6d56842f118a9485e3a18f42ca5e1b7577
Test: run the new test
Elliott Hughes [Thu, 12 Jan 2017 23:23:34 +0000 (23:23 +0000)]
Merge "localtime_r(3) should act as if it calls tzset(3)."
am:
1b2975d54b
Change-Id: I85c99ed325efbd119b5a7513b4cd997f9dc6b5aa
Elliott Hughes [Thu, 12 Jan 2017 23:19:50 +0000 (23:19 +0000)]
Merge "localtime_r(3) should act as if it calls tzset(3)."
Josh Gao [Thu, 12 Jan 2017 00:09:24 +0000 (00:09 +0000)]
Merge "Improve fmemopen tests."
am:
27fa754287
Change-Id: Ie62d2d2850fc4dee566317d5e0fcde54693694a1
Josh Gao [Wed, 11 Jan 2017 23:59:01 +0000 (23:59 +0000)]
Merge "Improve fmemopen tests."
Elliott Hughes [Wed, 11 Jan 2017 22:34:16 +0000 (14:34 -0800)]
localtime_r(3) should act as if it calls tzset(3).
See code comment.
Bug: http://b/
31339449
Test: ran tests & benchmarks
Change-Id: I6b6a63750ef41664dc4698207e6a53e77cc28cdf
Josh Gao [Wed, 11 Jan 2017 22:38:08 +0000 (22:38 +0000)]
Merge "Add support for expected failures to gtest runner."
am:
94ea19895c
Change-Id: I91e4f3673d6715326a2f1a43937343e59e15048c
Treehugger Robot [Wed, 11 Jan 2017 22:32:17 +0000 (22:32 +0000)]
Merge "Add support for expected failures to gtest runner."
Josh Gao [Tue, 10 Jan 2017 21:00:37 +0000 (13:00 -0800)]
Improve fmemopen tests.
Tag fmemopen_NULL as a known failure, and make it reliably fail by
memsetting the buffer we read to.
Bug: http://b/
33251022
Test: bionic-unit-tests/bionic-unit-tests64 on bullhead
Test: bionic-unit-tests-glibc --gtest_filter="*memopen*"
Change-Id: I381783282359851c9de47146dafbb5a291960c2a
Dimitry Ivanov [Wed, 11 Jan 2017 02:03:20 +0000 (02:03 +0000)]
Merge "Hide __system_property_add/update, deprecate find_nth"
am:
907fcc98b2
Change-Id: I68bbcf32432de9dba91116faae981701e122dbaa
Treehugger Robot [Wed, 11 Jan 2017 01:57:56 +0000 (01:57 +0000)]
Merge "Hide __system_property_add/update, deprecate find_nth"
Dimitry Ivanov [Mon, 9 Jan 2017 19:05:52 +0000 (11:05 -0800)]
Hide __system_property_add/update, deprecate find_nth
These functions are supposed to be used only by the
property service.
__system_property_find_nth is deprecated and no longer part
of NDK. Call to this function will result in abort for apps
targeting Android O.
Bug: http://b/
34114501
Test: bionic-unit-tests --gtest_filter=prop*
Change-Id: I9846965bf248e2ddf45cd7b293618245bbd87145
Josh Gao [Tue, 10 Jan 2017 00:43:33 +0000 (16:43 -0800)]
Add support for expected failures to gtest runner.
Bug: http://b/
33251022
Test: manually added tests that unexpectedly failed
Test: bionic-unit-tests/bionic-unit-tests64 on bullhead
Change-Id: Ie9aef80014b9fd6512b68f0e42abc298732a9b6e
Christopher Ferris [Tue, 10 Jan 2017 18:47:45 +0000 (18:47 +0000)]
Merge "Fix streaming(memcpy) performance on Cortex-A7"
am:
0625beb270
Change-Id: I24340a1b7d9e93505abf517e4ecc24175c58a164
Christopher Ferris [Tue, 10 Jan 2017 18:41:14 +0000 (18:41 +0000)]
Merge "Fix streaming(memcpy) performance on Cortex-A7"
Elliott Hughes [Tue, 10 Jan 2017 01:23:34 +0000 (01:23 +0000)]
Merge "Revert "Revert "move android_ids into bionic"""
am:
940d460f59
Change-Id: Id5e500e39cd544fab4398318ff48289764af1e59
Elliott Hughes [Tue, 10 Jan 2017 01:23:26 +0000 (01:23 +0000)]
Revert "Revert "move android_ids into bionic""
am:
3f6eee9793
Change-Id: I6a65710f926a76e48705fa4594f479f5d48df068
Elliott Hughes [Tue, 10 Jan 2017 01:15:29 +0000 (01:15 +0000)]
Merge "Revert "Revert "move android_ids into bionic"""
Elliott Hughes [Tue, 10 Jan 2017 00:48:20 +0000 (00:48 +0000)]
Merge "Dynamic linker errors should go to stderr."
am:
74e1fe1cc0
Change-Id: I045e0116e73a69989e80d8bd281d38ee6c5112b9
Elliott Hughes [Tue, 10 Jan 2017 00:43:38 +0000 (00:43 +0000)]
Merge "Dynamic linker errors should go to stderr."
Sandeep Patil [Wed, 28 Dec 2016 01:37:44 +0000 (17:37 -0800)]
property_contexts: split into platform and non-platform components
Bug:
33746484
Test: Successfully boot with original service and property contexts.
Test: Successfully boot with split serivce and property contexts.
Change-Id: Id1acb8c4273c6f93396e8cf7a063571a2bf08786
Signed-off-by: Sandeep Patil <sspatil@google.com>
Elliott Hughes [Mon, 9 Jan 2017 22:39:44 +0000 (14:39 -0800)]
Dynamic linker errors should go to stderr.
Also add the missing \n to the end of the error.
Addresses review comment.
Bug: N/A
Test: manual
Change-Id: I7c997a71e6e02381d21ac762595ba90370db9d05
Elliott Hughes [Mon, 9 Jan 2017 22:33:13 +0000 (22:33 +0000)]
Merge "Exit rather than abort if asked to run a non-PIE executable."
am:
71952ef625
Change-Id: Ibb2940eef382fabce0c2c5252fe962c145f5be95
Elliott Hughes [Mon, 9 Jan 2017 22:26:34 +0000 (22:26 +0000)]
Merge "Exit rather than abort if asked to run a non-PIE executable."
Elliott Hughes [Mon, 9 Jan 2017 21:30:38 +0000 (21:30 +0000)]
Merge "Move to .md files for even trivial documentation."
am:
39b6d59049
Change-Id: Ia1950ac9cf0eaa9ca19904ae54d2903471e77626
Elliott Hughes [Mon, 9 Jan 2017 21:23:00 +0000 (21:23 +0000)]
Merge "Move to .md files for even trivial documentation."
Elliott Hughes [Sat, 7 Jan 2017 20:47:28 +0000 (12:47 -0800)]
Move to .md files for even trivial documentation.
So it's automatically displayed for folks browsing the source.
Bug: N/A
Test: N/A
Change-Id: I80e823415f1fb12ad8ce5348e926519b6e29580f
Elliott Hughes [Sat, 7 Jan 2017 18:38:20 +0000 (10:38 -0800)]
Exit rather than abort if asked to run a non-PIE executable.
Each release we're asked to investigate tombstones from code that hasn't
been allowed to run on Android since L. This is just wasting our time,
and clearly the "obviousness" of aborting rather than exiting hasn't ensured
that all app developers rebuild their old binaries. In some cases it seems
like they run them "just in case" and don't care if they fail.
Bug: http://b/
34112178
Test: ran libsupervisor.so from com.ss.android.article.news
Change-Id: I8a3f196c4755601a3888281566fbb7b817f01dca
Dimitry Ivanov [Sat, 7 Jan 2017 04:03:15 +0000 (04:03 +0000)]
Merge "Fix bionic-unit-tests-static crash"
am:
500b9be411
Change-Id: I98d7587228db28d920354f925b0d20e3734838df
Treehugger Robot [Sat, 7 Jan 2017 03:59:14 +0000 (03:59 +0000)]
Merge "Fix bionic-unit-tests-static crash"
Dimitry Ivanov [Fri, 6 Jan 2017 22:49:57 +0000 (14:49 -0800)]
Fix bionic-unit-tests-static crash
The crash happens because for static executables call to dlopen
results in crash. This change moves dlopen() == nullptr check
from static variable to a functions so that the dlopen crash only
affects tests calling dlopen but not the whole executable.
Also make static tests report crashes to debugerd
Test: run bionic-unit-tests-static (for arm 32 and 64)
Bug: http://b/
34129417
Change-Id: I7c4d8caf2a43250234fe24496b1c95eab572769f
Nick Kralevich [Fri, 6 Jan 2017 21:59:19 +0000 (21:59 +0000)]
Merge "Make /dev/__properties__ 0711"
am:
11798bcc35
Change-Id: I4c418dd0567013b003541977f5b2f97a35ac29e3
Treehugger Robot [Fri, 6 Jan 2017 21:53:05 +0000 (21:53 +0000)]
Merge "Make /dev/__properties__ 0711"
Nick Kralevich [Fri, 6 Jan 2017 20:04:29 +0000 (12:04 -0800)]
Make /dev/__properties__ 0711
Don't allow processes to read the contents of the directory
/dev/__properties__. This is an implementation detail of the properties
system that processes shouldn't be concerned with.
Test: Device boots and no problems reading individual properties.
Test: ls -la /dev/__properties__ fails
Change-Id: I00130fe4529525935654bff91e3cc59253b86e26
Josh Gao [Fri, 6 Jan 2017 00:50:46 +0000 (00:50 +0000)]
Merge "Add declaration of tgkill to signal.h."
am:
d26267b8b9
Change-Id: I398886a42d4d3d72b430159ac9c6dcaee5e26d09
Treehugger Robot [Fri, 6 Jan 2017 00:44:17 +0000 (00:44 +0000)]
Merge "Add declaration of tgkill to signal.h."
Josh Gao [Thu, 5 Jan 2017 21:22:26 +0000 (13:22 -0800)]
Add declaration of tgkill to signal.h.
Expose a useful function that we've had since Jelly Bean.
Bug: http://b/
34111810
Test: TreeHugger
Change-Id: Iaf3097f224c09b533f36050cf21394ba148007ad
Dimitry Ivanov [Tue, 3 Jan 2017 18:12:14 +0000 (18:12 +0000)]
Merge "Downgrade dynamic section checks to warning"
am:
fb07c36bc0
Change-Id: I84efe43a93ae0689492ddf229c52e6710e00bc4a
Dimitry Ivanov [Tue, 3 Jan 2017 18:03:04 +0000 (18:03 +0000)]
Merge "Downgrade dynamic section checks to warning"
Dimitry Ivanov [Thu, 29 Dec 2016 00:21:49 +0000 (16:21 -0800)]
Downgrade dynamic section checks to warning
For apps targeting pre-O releases print warn about
invalid dynamic section.
Bug: http://b/
33842263
Test: Start an app in question and check logs for warning.
Test: Run bionic-unit-test --gtest_filter=dl*:Dl*
Change-Id: I80bd51f2133b956fc04e20a6538cccd50118c1b8
Chitti Babu Theegala [Thu, 15 Dec 2016 20:43:28 +0000 (02:13 +0530)]
Fix streaming(memcpy) performance on Cortex-A7
Stream-mode detection for L1 in A7-core is failing for
non cache-line-size (non 64 byte) aligned addresses.
This leads to destination data getting cached unnecessarily.
This A7 issue is confirmed by ARM
This issue is solved by aligning destination address to 64 byte before
entering the loop in memcpy routine.
Though we get lower score for micro_bench memcpy when L1 cache is bypassed,
it is desirable since it avoids unnecessary eviction of other process data
from L1 which is good for overall system performance.
Higher micro_bench memcpy numbers for < 64byte alignment shows good numbers
but this is at the cost of L1 cache pollution. During memcpy/memset,
unnecessary data is filled in L1 cache, this causes eviction of other
process data from L1.
For example during msmset(0), L1 cache gets filled with 0s which should be
avoided.
Additionally, there is another issue with cortex A7 that impacts performance
for all alignments / all Android Wear versions:
Store Buffer on A7 is 32 byte which limits the 32-byte back to back stores.
In the current implementation back to back 32bytes writes is causing CPU stalls.
This issue can be solved by interleaved Loads and Stores.
This helps in avoiding CPU stalls during memcpy by utilizing efficiently the
A7 internal load and store buffers.
Change-Id: Ie5f12f2bb5d86f627686730416279057e4f5f6d0
Josh Gao [Fri, 16 Dec 2016 22:25:27 +0000 (22:25 +0000)]
Merge "versioner: start using C++17."
am:
c816e9fa03
Change-Id: Ie148482169ab9cfae841a94b95b8266c520a45e6