| 111b0e05d130d345857ad875be484e3839ce25d4 |
|
02-Jun-2017 |
Mathieu Chartier <mathieuc@google.com> |
Revert "Change thread suspend timeout to be fatal for non-debug"
For overloaded systems, 10s might not be enough. Infinite seconds
will be though.
Bug: 37221387
Test: test-art-host
This reverts commit d69f82e155118654d989f0b436c019fb6d517341.
Change-Id: I8ac0e13e6bdae1845803682215e1794d892de8e5
|
| 1c8f4ff04062407db3d2dd834dd6bc24f6662c4d |
|
14-Apr-2017 |
Sebastien Hertz <shertz@google.com> |
Always check result of Thread::ModifySuspendCount
Ensures that we never ignore the result of ModifySuspendCount so that
we can react if the suspend count is not updated as expected.
This CL does the following:
* Adds __attribute__((warn_unused_result)) on the method to raise an
error at compilation time if the result is ignored.
* Wraps calls with DCHECK where the result used to be ignored.
Bug: 27385848
Test: make -j test-art-host
Test: art/tools/run-jdwp-tests.sh --mode=host --variant=X64
Change-Id: I2d0e1ab7158c70ec8076c8bae6e4b814aee75af6
|
| d69f82e155118654d989f0b436c019fb6d517341 |
|
21-Mar-2017 |
Mathieu Chartier <mathieuc@google.com> |
Change thread suspend timeout to be fatal for non-debug
Right now it just loops forever and doesn't give any stack traces
regarding probable deadlocks.
Bug: 36408743
Test: test-art-host
Change-Id: Ib8919f4be540291d8226dd0e7eb88c7faf7cdde9
|
| a222404a5832ab16786931576d52825d08eed3ca |
|
09-Feb-2017 |
Hiroshi Yamauchi <yamauchi@google.com> |
Fix the empty checkpoint timeout.
The problem happens when a thread is indirectly blocked on a mutex
that another thread holds and is blocked on a weak ref access.
Add a way to do a dummy wakeup on a thread that's blocked on a mutex
so that the thread will respond to the empty checkpoint request. Do
this for the mutexes that are expected to be held when a weak ref is
accessed. Add a check that detects an unexpected case.
Bug: 33006388
Bug: 12687968
Test: test-art-host.
Change-Id: Iefec69b9a21aa25a928cb31fcf4fb872f867a8c2
|
| e9429c8c5f8c5541dcbe97e03c55cba175bce182 |
|
28-Jan-2017 |
Mathieu Chartier <mathieuc@google.com> |
Add to time to suspend in FlipThreadRoots to histogram
Add the time to suspend to the suspend_all_historam_ to determine if
GC pauses are caused by GC or bad mutator activity.
Also exclude ThreadFlipBegin time being counted as paused since it
happens before threads are suspended.
Test: test-art-host
Change-Id: Ia55db0ebbdc7ba6f23d7212c0e42c7806b297c2b
|
| 3fceaf500d570722122f2d092f302d27769e3a51 |
|
22-Jan-2017 |
Mathieu Chartier <mathieuc@google.com> |
Use higher default thread suspend timeout for run tests
Increase the default for tests from 10s-50s to 500s. This is done
by adding a run-test option "--suspend-timeout" that defaults to
500000.
Example when passing a low value of 100:
test/run-test --host --suspend-timeout 100 --dev 004-ThreadStress
Results in:
dalvikvm32 F 01-22 13:33:31 12067 12118 thread_list.cc:669] Timed
out waiting for threads to suspend, waited for 100.114ms
This should fix some of the recent test flakiness.
Test: test-art-host
Change-Id: I303c4f198e560e9f38ec375e9d7a80a5e61f776e
|
| 32c8337f9fcb8e7026b7ee67dc4ff3e1ca74830e |
|
11-Jan-2017 |
Mathieu Chartier <mathieuc@google.com> |
Increase thread suspension timeout for debug builds
The current timeout may be causing some test failures. Also print how
long we waited.
Test: test-art-host -j32
Change-Id: Ib8a9c68ab4571efc8456c098b03bf35f648bfaf1
|
| 35af2ad461e9745252bdc2c1c5a0cad1386626d4 |
|
23-Dec-2016 |
rock.yeh <app1_test@weckout.com> |
Use relative futex timeout in SuspendAllInternal.
This erroneously used an absolute timeout for FUTEX_WAIT, potentially
causing it to block for MUCH longer than 10 seconds.
http://man7.org/linux/man-pages/man2/futex.2.html
Note: for FUTEX_WAIT, timeout is interpreted as a relative value.
Error case:
kernel time: 15842.476344
nsec: 31694950164731
12-22 20:25:48.020043 0 0 F [15842.476344][HeapTaskDaemon:3008]
futex_wait 00000070fcdf8980 1 1 31694950164731
Normal case:
kernel time: 15842.476344
nsec: 15842476344
Block backtrace:
"HeapTaskDaemon" sysTid=3008
#00 pc 000000000001bcac /system/lib64/libc.so (syscall+28)
#01 pc 0000000000461db0 /system/lib64/libart.so
(_ZN3art10ThreadList18SuspendAllInternalEPNS_6ThreadES2_S2_b+720)
#02 pc 000000000046249c /system/lib64/libart.so
(_ZN3art10ThreadList10SuspendAllEPKcb+532)
#03 pc 00000000001e8fc8 /system/lib64/libart.so
(_ZN3art2gc9collector9MarkSweep9RunPhasesEv+232)
#04 pc 00000000001e1694 /system/lib64/libart.so
(_ZN3art2gc9collector16GarbageCollector3RunENS0_7GcCauseEb+332)
#05 pc 0000000000211500 /system/lib64/libart.so
(_ZN3art2gc4Heap22CollectGarbageInternalENS0_9collector6GcTypeENS0_7Gc
CauseEb+3048)
#06 pc 0000000000218acc /system/lib64/libart.so
(_ZN3art2gc4Heap12ConcurrentGCEPNS_6ThreadEb+124)
#07 pc 000000000021fbf4 /system/lib64/libart.so
(_ZN3art2gc4Heap16ConcurrentGCTask3RunEPNS_6ThreadE+36)
This is a fix patch log and each timeout log interval of 10 seconds:
12-27 03:27:46.153744 1631 1713 E art : Unexpected time out during
suspend all.
12-27 03:27:56.154401 1631 1713 E art : Unexpected time out during
suspend all.
12-27 03:28:06.154956 1631 1713 E art : Unexpected time out during
suspend all.
|
| 46ee31b67d7ee1bd085fbc240502053caa3cf8fa |
|
14-Dec-2016 |
Andreas Gampe <agampe@google.com> |
ART: Move to libbase StringPrintf
Remove ART's StringPrintf implementation. Fix up clients. Add
missing includes where necessary.
Test: m test-art-host
Change-Id: I564038d5868595ac3bb88d641af1000cea940e5a
|
| 585da955bc8e5040705dcfd941b2131025ebcef8 |
|
02-Dec-2016 |
Andreas Gampe <agampe@google.com> |
ART: Add precise root visiting
Add VisitRootFlags::kVisitRootFlagPrecise to signal a
request for precise RootInfo objects.
Move VisitRootFlags to gc_root.h. Refactor VisitRoot
functions to pass flags. Add code in Thread visiting
to decode vregs in compiled code.
Bug: 31385354
Test: m test-art-host
Change-Id: I8e7280f0ab682871f729f2a1c6b18670cf2dbf82
|
| a82769c1168d38fb4d36eaecfc186aecb17d4b16 |
|
03-Dec-2016 |
Hiroshi Yamauchi <yamauchi@google.com> |
Dump more debug info for b/33006388.
Bug: 33006388
Bug: 12687968
Test: test-art-host with CC.
Change-Id: Id9d67bc603c6ff7bc8e346e181e3e09ffbda43b3
|
| fe814e89965ddf9a8b603863bd28259f8dd7be35 |
|
09-Nov-2016 |
Mathieu Chartier <mathieuc@google.com> |
Use entrypoint switching to reduce code size of GcRoot read barrier
Set the read barrier mark register entrypoints to null when the GC
is not marking. The compiler uses this to avoid needing to load the
is_gc_marking boolean.
Code size results on ritzperf CC:
arm32: 13439400 -> 13242792 (-1.5%)
arm64: 16380544 -> 16208512 (-1.05%)
Implemented for arm32 and arm64. TODO: Consider implementing on x86.
Bug: 32638713
Bug: 29516974
Test: test-art-host + run ritzperf
Change-Id: I527ca5dc4cd43950ba43b872d0ac81e1eb5791eb
|
| 3049324f4ef71b5d7a3de49bd77c75f07dbf8f3a |
|
03-Nov-2016 |
Hiroshi Yamauchi <yamauchi@google.com> |
Make empty checkpoint work while weak ref access is disabled.
Fix a potential race on PushOntoMarkStack for CC by running an empty
checkpoint (while weak ref access is disabled).
Bug: 32508093
Bug: 12687968
Test: test-art-host with CC/CMS, libartd boot with N9, Ritz EAAC.
Change-Id: I3749bb525e7734804307ee16262355f3fc730312
|
| 3f386d55f7e37eaefa828a55c8aa8c4f8788e6cf |
|
05-Oct-2016 |
Mathieu Chartier <mathieuc@google.com> |
Allow early abort logging
Print unattached threads with native stack traces for early abort.
Bug: 31855501
Test: test-art-host -j32
Test: thread_list_->Dump in Runtime::Start.
Change-Id: I5364625e798356c4fa1f88a8c646e374e59d93b0
|
| 02e7f1a46d8dbb277d045182cd1fa4b058d55162 |
|
04-Oct-2016 |
Hiroshi Yamauchi <yamauchi@google.com> |
Fix a deadlock between thread flip and suspend request.
See 31683379#9 for the deadlock scenario.
Make ModifySuspendCount(+1) retry if the thread flip function is set.
Bug: 31683379
Bug: 12687968
Test: test-art, N9 libartd boot, Ritz EAAC with CC.
Test: 129-GetThreadId with gcstress and CC.
Change-Id: Id5cdfcd90a08a2ff497f9f0e2842fa4c613549bc
|
| 0795f23920ee9aabf28e45c63cd592dcccf00216 |
|
28-Sep-2016 |
Mathieu Chartier <mathieuc@google.com> |
Clean up ScopedThreadStateChange to use ObjPtr
Also fixed inclusion of -inl.h files in .h files by adding
scoped_object_access-inl.h and scoped_fast_natvie_object_access-inl.h
Changed AddLocalReference / Decode to use ObjPtr.
Changed libartbenchmark to be debug to avoid linkage errors.
Bug: 31113334
Test: test-art-host
Change-Id: I4d2e160483a29d21e1e0e440585ed328b9811483
|
| 15af34cc172d9ae652da35a864190585cb6717a6 |
|
27-Sep-2016 |
Hiroshi Yamauchi <yamauchi@google.com> |
Fix a CC deadlock in 129-ThreadGetId.
GC should consider a thread as "waiting for GC thread flip" or
"transitioning to runnable" only if its suspend count is 1.
See 31683379#7 for the deadlock scenario.
Bug: 31683379
Bug: 12687968
Test: test-art with CC and gcstress, N9 libartd boot, Ritz EAAC.
Change-Id: Icd2548bd6d9a4f8d7b54ed20150a4801af9e26a3
|
| d6e54bbc09595465309aa7909b7a3965defe4419 |
|
26-Sep-2016 |
Andreas Gampe <agampe@google.com> |
ART: Fix nits
Follow-up to commit 3fec9ac0d5af1358d216eb2fdc2000ec0205f3f0.
Bug: 31338270
Test: m test-art-host
Change-Id: Ia50f450bef144167b946224c26242c529302996e
|
| 3fec9ac0d5af1358d216eb2fdc2000ec0205f3f0 |
|
13-Sep-2016 |
Andreas Gampe <agampe@google.com> |
ART: Use libbase logging
Move most of our logging infrastructure over to system/core/base.
Retain VLOG.
Using unified Android infrastructure has two main advantages. First,
it reduces the complexity/maintenance burden in ART. Second, it
allows to detach logging for the cases where we do not want or need
a runtime, e.g., dexdump, the disassembler, etc. As a part of the
latter, libbase is also supported for all hosts (including Windows).
From a developer viewpoint, there are minor behavior changes for the
LOG statements (see above), but otherwise usage is the same. Explicit
severity enum items are in the android::base namespace now.
Bug: 31338270
Test: m test-art-host
Change-Id: I5abcb2f45f5b03d49951874c48544f72a283a91b
|
| febd0cf9b5070ecc54ba433b951b65e14a54ccde |
|
15-Sep-2016 |
Hiroshi Yamauchi <yamauchi@google.com> |
Fix a deadlock in the CC collector.
Fix a deadlock between CC GC disabling system weaks and thread attach.
See 31500969#2 for more details.
Bug: 31500969
Bug: 12687968
Test: test-art-host with CC. N9 libartd boot. Ritz EAAC.
Change-Id: Ic9a8bfb1c636643a03f4580b811fe890273576b6
|
| ee23582af60b36f982de2ad16f485a61f35ae817 |
|
20-Aug-2016 |
Hiroshi Yamauchi <yamauchi@google.com> |
Revert "Revert "Improve the thread flip.""
This reverts commit db3204f87c3f7c4de89762ce9e8502a9dc25c2d8.
Improve the thread flip.
- In addition to the threads that are suspended in FullSuspendCheck(),
prioritize the resume of threads that are blocking for the thread
flip at the JNI critical section entry and threads are about to
transition to runnable (eg. blocking at the SOA entry from JNI).
- Shorten the length of the thread flip critical
section (ThreadFlipBegin/End).
- Add some systrace scopes.
- Add a read barrier for the locked objects during the thread dump in
case the thread is in the middle of flipping.
Bug: 30980189
Bug: 29517059
Bug: 12687968
Test: test-art-host, Ritz EAAC, N9 libartd boot
Change-Id: I3a903c47c0fcc746664ec376cc31dee8af3c3ecb
|
| db3204f87c3f7c4de89762ce9e8502a9dc25c2d8 |
|
19-Aug-2016 |
Hiroshi Yamauchi <yamauchi@google.com> |
Revert "Improve the thread flip."
This reverts commit e61e6fdd404b8f04bd33fc3ea49e38e92e31ad69.
ThreadStress failing.
Bug: 29517059
Bug: 12687968
Change-Id: I0ee8ef04f77c5f9378e4bd21c5dedb4e435a2dc6
|
| e61e6fdd404b8f04bd33fc3ea49e38e92e31ad69 |
|
03-Aug-2016 |
Hiroshi Yamauchi <yamauchi@google.com> |
Improve the thread flip.
- In addition to the threads that are suspended in FullSuspendCheck(),
prioritize the resume of threads that are blocking for the thread
flip at the JNI critical section entry and threads are about to
transition to runnable (eg. blocking at the SOA entry from JNI).
- Shorten the length of the thread flip critical
section (ThreadFlipBegin/End).
- Add some systrace scopes.
Bug: 29517059
Bug: 12687968
Test: test-art-host, Ritz EAAC, N9 libartd boot.
Change-Id: Idecec9936ae432c23d5f83321ba13339852018df
|
| 5dd44d0da7999ab62f09a1bf03345280811736d4 |
|
03-Aug-2016 |
Andreas Gampe <agampe@google.com> |
ART: Factor out native stack dumping
Move DumpNativeStack and DumpKernelStack into a new file to keep
concerns separated. Clean up the file.
In preparation for modifications to how we use addr2line.
Bug: 30351473
Test: m test-art-host
Test: manual host testing with a test that times out
Change-Id: I02d5645952fe94fd8e73bcf0522721547f6cef66
|
| d778cd68c5d093c2fb95f91608deb36fea8e8bad |
|
05-Jul-2016 |
Vladimir Marko <vmarko@google.com> |
Expect spurious wake-up from futex(2)/FUTEX_WAIT call.
Test: Run standard ART test suite.
Bug: 27913630
Change-Id: I742544f7a0a7f896e47293d209a874782577e971
|
| aa45daaddee752fcc7f1e5b0e207748bc6bed0a4 |
|
20-Jun-2016 |
Nicolas Geoffray <ngeoffray@google.com> |
Release the thread_list lock while waiting for daemons.
Daemons might be in a state where they are actually waiting for
it (for example Thread.isInterrupted). If the shutdown thread
doesn't release the lock, such daemon cannot make progress.
bug:27353286
Change-Id: Ib7f4c7d1b56d73a829d01d5bfc0ab663fbb80b46
|
| d0b1dac94230c5e91639011e28160f34cc39c53e |
|
11-Jun-2016 |
Mathieu Chartier <mathieuc@google.com> |
Use ScopedObjectAccess in ThreadList::Dump
Prevent deadlocks if two different threads are calling Dump at the
same time:
Thread 1: Requests thread 2 to suspend
Thread 2: Requests thread 1 to suspend
Both threads are suspended and blocked on ScopedObjectAccess in
thread dumping.
Added to suspend all stress, reduced thread count to 4 temporarily
since this is the maximum number of simultaneous checkpoints.
Bug: 28988206
Bug: 26118154
(cherry picked from commit e99f53203904c23a26e53ca1bf6a4e45814146fe)
Change-Id: I9e4b8391ebad0bca0e42a03819c58c99ddba6b35
|
| f8a86b9ecefecae8bc0e456d540dae115249fdb6 |
|
15-Jun-2016 |
Mathieu Chartier <mathieuc@google.com> |
Visit invalid roots of only suspended threads
Since this always happens with suspended threads or self, you can
just visit these threads and do not require a suspend all. This
will not miss any roots if the caller was marking a thread root.
Fixes issues like transitioning to suspended and back blocking on a
thread suspension request from another thread. This could cause
deadlocks previously.
Bug: 29062271
Change-Id: I2fef149387aacf0cdc9a773d4f172c42fa53e4dc
|
| a3e8fc3aabc122dc4a324f15925741dccda2641c |
|
14-Jun-2016 |
Andreas Gampe <agampe@google.com> |
ART: Turn off unattached thread dumping
May be responsible for lack of ANR dumps in some situations.
Bug: 29248079
Change-Id: Ib742c063b1d77216eddfbf4392d17d024e028bf8
|
| e99f53203904c23a26e53ca1bf6a4e45814146fe |
|
11-Jun-2016 |
Mathieu Chartier <mathieuc@google.com> |
Use ScopedObjectAccess in ThreadList::Dump
Prevent deadlocks if two different threads are calling Dump at the
same time:
Thread 1: Requests thread 2 to suspend
Thread 2: Requests thread 1 to suspend
Both threads are suspended and blocked on ScopedObjectAccess in
thread dumping.
Added to suspend all stress, reduced thread count to 4 temporarily
since this is the maximum number of simultaneous checkpoints.
Bug: 28988206
Change-Id: I9e4b8391ebad0bca0e42a03819c58c99ddba6b35
|
| bb661c0f0cb72d4bbfc2e251f6ded6949a713292 |
|
04-Apr-2016 |
Bilyan Borisov <bilyan.borisov@linaro.org> |
Refactor use of __ANDROID__ macro
We use the __ANDROID__ macro, which is provided by the toolchain, in
numerous places. This patch refactors the usage of this by defining a
new macro, ART_TARGET_ANDROID, that is being passed during build to
ART_TARGET_CFLAGS in Android.common_build.mk on the same line as
ART_TARGET. The codebase currently assumes that the existence of the
__ANDROID__ macro implies that we are compiling art for an android
target device. This is because, currently, target builds are compiled
with target toolchains that provide the macro, while host toolchains
do not. With this change this assumption is still preserved. However,
in a future patch we will add the ability to compile art for a linux
target, and in that case the ART_TARGET_ANDROID macro won't be passed
anymore.
Change-Id: I1f3a811aa735c87087d812da27fc6b08f01bad51
|
| 61b3cd45c5e3a5c24f7ccce94843a718375f9b7a |
|
18-Apr-2016 |
Mathieu Chartier <mathieuc@google.com> |
Prevent holding stale Thread pointers
It is only really safe to hold non-self Thread* if you hold the
thread list lock. Changed a few places to use thread ids instead
of Thread.
Bug: 28223501
(cherry picked from commit 81c170fede9af9174aba71428334ac8f366a4b4f)
Change-Id: I15e50b699303a5c3739e4d19c153dd306e2ee504
|
| 81c170fede9af9174aba71428334ac8f366a4b4f |
|
18-Apr-2016 |
Mathieu Chartier <mathieuc@google.com> |
Prevent holding stale Thread pointers
It is only really safe to hold non-self Thread* if you hold the
thread list lock. Changed a few places to use thread ids instead
of Thread.
Bug: 28223501
Change-Id: Ie58bd755bf1dcf3c1f37da79ba0b2507f77574dd
|
| d3c59658faf9aaeb49cf63d928cead324f14bd8a |
|
17-Mar-2016 |
Nicolas Geoffray <ngeoffray@google.com> |
Check GetDumpNativeStackOnSigQuit() for daemons.
The flag is passed by run-test to avoid libunwind crash flakiness,
so we should check it when dumping threads.
Change-Id: I3a3fcfd80ab254a315d58d629c81161fea9900dc
|
| b9d4f999773754ab4e84496edf60af60b05119de |
|
01-Mar-2016 |
Andreas Gampe <agampe@google.com> |
ART: Allow unwinding unattached threads
Partial revert of commit ed8b723c5f3989d2593ec21c65c96d6d8bf25579.
Make it (constexpr) configurable whether we allow unwinding native
stacks of unattached threads.
Bug: 27449879
(cherry picked from commit 8d1594da6e97cd11580baf30ee4b75d4e7e2616a)
Change-Id: Iad3f3dde63155993bb4efb7bbbbde3e36f1379cc
|
| 8d1594da6e97cd11580baf30ee4b75d4e7e2616a |
|
01-Mar-2016 |
Andreas Gampe <agampe@google.com> |
ART: Allow unwinding unattached threads
Partial revert of commit ed8b723c5f3989d2593ec21c65c96d6d8bf25579.
Make it (constexpr) configurable whether we allow unwinding native
stacks of unattached threads.
Bug: 27449879
Change-Id: I307a80c9a0166f33fa8a41e492d7eb0d77d0e44f
|
| 47c1959f0c78eaf6ab17e1b80fa3ce9643e55849 |
|
07-Mar-2016 |
Mathieu Chartier <mathieuc@google.com> |
Increase kDumpWaitTimeout 10x for target builds
New timeout is 100s, the old one was occasionally hit by overloaded
device. This caused a confusing crash in the barrier destructor
since not all the threads had went through the barrier yet.
Bug: 27334917
(cherry picked from commit 74bdab0b14c592053ff71654ab51a0f116114a42)
Change-Id: I24cf6eb6858ad93a47446aa6bbc18942a71b191e
|
| 74bdab0b14c592053ff71654ab51a0f116114a42 |
|
07-Mar-2016 |
Mathieu Chartier <mathieuc@google.com> |
Increase kDumpWaitTimeout 10x for target builds
New timeout is 100s, the old one was occasionally hit by overloaded
device. This caused a confusing crash in the barrier destructor
since not all the threads had went through the barrier yet.
Bug: 27334917
Change-Id: I5bcdaa39cfc104f3555f62b6117b417c8bf36e7a
|
| 32ce2adefb8a3d0eda59a29f5e87c1eb43eef796 |
|
04-Mar-2016 |
Mathieu Chartier <mathieuc@google.com> |
Add more systracing everywhere
Added to:
JIT
DexFile functions
Oat file manager
Added helper ScopedTrace to prevent errors and reduce excess code.
Bug: 27502458
(cherry picked from commit dabdc0fe183d4684f3cf4d70cb09d318cff81b42)
Change-Id: Ifaeff8913d79eefc797380987d13cc00456266f8
|
| dabdc0fe183d4684f3cf4d70cb09d318cff81b42 |
|
04-Mar-2016 |
Mathieu Chartier <mathieuc@google.com> |
Add more systracing everywhere
Added to:
JIT
DexFile functions
Oat file manager
Added helper ScopedTrace to prevent errors and reduce excess code.
Bug: 27502458
Change-Id: Ifaeff8913d79eefc797380987d13cc00456266f8
|
| 0c659394d1ee74d28157c42656b31ea48cb54f8b |
|
02-Mar-2016 |
Nicolas Geoffray <ngeoffray@google.com> |
Revert "ART: Allow unwinding unattached threads"
This reverts commit f7286721a51c772b2abb42fef5bc0d2548df2372.
Change-Id: I6c96ece063041e94d40a96bdd054f89cd7875ade
|
| f7286721a51c772b2abb42fef5bc0d2548df2372 |
|
01-Mar-2016 |
Andreas Gampe <agampe@google.com> |
ART: Allow unwinding unattached threads
Partial revert of commit ed8b723c5f3989d2593ec21c65c96d6d8bf25579.
Make it (constexpr) configurable whether we allow unwinding native
stacks of unattached threads.
Change-Id: Ibb1fd0956fb796583f78ec9c3a74521f6fbe96d9
|
| a73280df8cac1279b6dea0424722f42ef0048613 |
|
15-Feb-2016 |
Nicolas Geoffray <ngeoffray@google.com> |
Add an option to disable native stack dumping on SIGQUIT.
Some of our (stress) run-tests do ANR dumping, which end up
stressing libunwind, that has known problems. To avoid getting
flakes due to libunwind, disable native stack dumping on SIGQUIT
for our run-tests.
bug:27185632
bug:24664297
Change-Id: I69085e48db903d6240448d71666ae2dcd091922e
|
| fbc31087932a65e036a153afab3049dc5298656a |
|
24-Jan-2016 |
Mathieu Chartier <mathieuc@google.com> |
Revert "Revert "Load app images""
This reverts commit 1bc977cf2f8199311a97f2ba9431a184540e3e9c.
Bug: 22858531
Change-Id: Ide00bf3a73a02cba3bb364177204ad1b13f70295
|
| 1bc977cf2f8199311a97f2ba9431a184540e3e9c |
|
23-Jan-2016 |
Nicolas Geoffray <ngeoffray@google.com> |
Revert "Load app images"
Fails when a method is duplicated (see test 097-duplicate-method)
Bug: 22858531
This reverts commit f7fd970244f143b1abb956e29794c446e4d57f46.
Change-Id: Ib30ae5be00cc568e799290be6b3c8f29cbbe4c20
|
| f7fd970244f143b1abb956e29794c446e4d57f46 |
|
09-Nov-2015 |
Mathieu Chartier <mathieuc@google.com> |
Load app images
Support in-place patching of the app image based on boot image
location and app oat location. Only loads for art run test so far
since we do not automatically generate app images for app installs.
N5 maps launch time (~200 runs):
Before: 930ms
After: 878.18ms
After + image class table: 864.57ms
TODO:
Oatdump support.
Store class loaders as class roots in image.
Bug: 22858531
Change-Id: I9cbc645645e62ea2ed1ad8e139e91af7d88514c1
|
| 59a8c0bb74c353f879760bd6522b49cf7298666e |
|
20-Jan-2016 |
tony.ys_liu <tony.ys_liu@htc.com> |
Use current pid not tid to create a backtrace map.
The BacktraceMap::Create() function assumes that the passed in
parameter is the current pid not tid. If a tid is passed in, then
the logic that creates the local map in libunwind is not used
and can bypass all of the special handling of local maps. The result
can be crashes if this map data gets out of sync with the real maps
found in the process.
Bug: 26676205
Bug: https://code.google.com/p/android/issues/detail?id=199219
Change-Id: Ibab9f7a7b99648a2cdaa185958bf6ffc368079ce
|
| 62597d1dc4e62d36addcc2f1c452cbd6f565a420 |
|
11-Jan-2016 |
Mathieu Chartier <mathieuc@google.com> |
Add sleep in SuspendAllDaemonThreadsForShutdown
Only sleep if there are still daemon threads. Since we have already
shutdown the runtime daemons, this is a rare case. The goal of the
sleep is to leave daemon threads time to become blocked in a place
where they ideally will not touch runtime state.
Bug: 26483935
Bug: 18577101
Change-Id: Iad12209e771720523620e0a3829a75ec4647f58b
|
| ba098bafdc4bd6dc6fc85f2996fa9f2d9be051bf |
|
07-Jan-2016 |
Mathieu Chartier <mathieuc@google.com> |
Revert "Revert "Reduce sleep duration in SuspendAllDaemonThreads""
Fixed the error where I reduced the time until the warning by 10x.
Bug: 26351700
This reverts commit bc34a7ed232cb3d384c250578e2b4ede45f2b167.
Change-Id: I6b30a0711fe077421665b78c39ded88899407700
|
| 4d87df607a0b86cdf4b2c04f61d72a60d8975ce0 |
|
08-Jan-2016 |
Mathieu Chartier <mathieuc@google.com> |
Improve handling of daemon threads after runtime shutdown
The main issue comes from the fact that user daemon threads are
allowed to continue running after the runtime has shutdown. They may
still have a JNI env pointer. To prevent crashing if they call into
the env, we replace the function pointers with functions that sleep
forever.
The other issue is that user daemon threads that are blocked in an
ART condition variable may get woken up by another user daemon inside
of Monitor::Notify or by a spurious wakeup (i.e. SIGQUIT). To deal
with this issue, we check the JNI env for shutdown runtime when we
are woken up from a condition variable wait. This check fixes test
132 with --host --gdb --interpreter. Previously this test crashed
since dlclose was somehow causing a spurious futex wakeup.
TODO: Investigate adding a unit test.
Bug: 18577101
Change-Id: I479b38968ee9fbc4ee4b252ee2528787279972cc
|
| bc34a7ed232cb3d384c250578e2b4ede45f2b167 |
|
07-Jan-2016 |
Nicolas Geoffray <ngeoffray@google.com> |
Revert "Reduce sleep duration in SuspendAllDaemonThreads"
run-test 132 randomly chokes due to this change.
Bug: 26351700
This reverts commit adef5bde6c3392dca3dafa0f8d2c9428b59af3ef.
Change-Id: I18e007264e1e8b9027d7d241cb9f0053ad8d112a
|
| adef5bde6c3392dca3dafa0f8d2c9428b59af3ef |
|
07-Jan-2016 |
Mathieu Chartier <mathieuc@google.com> |
Reduce sleep duration in SuspendAllDaemonThreads
Previously we unconditionally slept for 200ms which caused runtime
shutdown to take 200ms longer than required.
Reduces runtime shutdown time for "am" from ~200ms to ~10ms.
Bug: 26351700
Change-Id: I7fd0059cb3b04dba0acc4a1754e76c89ec867f85
|
| fb6db4c841e928feb35a985923fe2d41848f8f31 |
|
07-Jan-2016 |
Mathieu Chartier <mathieuc@google.com> |
Add systrace output to runtime shutdown
Help debug why am command shutdown is slow.
Findings:
~200ms spent in deleting thread list on N5.
Bug: 26351700
Change-Id: Icb59390a5267ae8bade4ced12252743c48206bb1
|
| 9450c6c25833e8353e436bbd0491ed96cd80f34f |
|
07-Nov-2015 |
Mathieu Chartier <mathieuc@google.com> |
Remove thread suspension assertion for SuspendAllForDebugger
Fixes race condition that occasionally caused an assertion failure
for debugger tests, description in comments.
Change-Id: I18e003e01c1711dfdb0dd68ccb3d2b7dc626991f
|
| 62623401fe994ff2f2719faf3cdb3c23b92ccd96 |
|
28-Oct-2015 |
Nicolas Geoffray <ngeoffray@google.com> |
Fix deadlock with the JIT code cache.
Also remove hack done for ThreadStress.
Change-Id: Ie25c3bca08d9f2b8919706fa3fc26c5ab213f4a3
|
| b56200b014d6bcfb962dc7afdd8b5f7168043ae7 |
|
29-Oct-2015 |
Mathieu Chartier <mathieuc@google.com> |
Add comment for RunCheckpoint and clean up ThreadList a bit
Change-Id: Id512bfd15fee9a7359c77f094e16d64634943619
|
| 10d2508b105427ef1bcaf0c222873bae7acc66d3 |
|
29-Oct-2015 |
Mathieu Chartier <mathieuc@google.com> |
Change Checkpoint API to return total number of checkpoints
Fixes a race condition with SetStateUnsafe that caused some
warnings in the Barrier::~Barrier.
The race was:
RunCheckpoint sees suspended thread, runs the checkpoint. Inside the
checkpoint, the thread state had changed to runnable by
SetStateUnsafe. This occasionally caused more Barrier::Pass than
expected.
The fix is to return the total number of checkpoints instead of just
the runnable ones.
Bug: 24191051
Change-Id: If15a933ed4c8efa66a5f27cd5feaa2e5957ae804
|
| 6cff48f05894a60aa4742472983a75504536ca15 |
|
27-Jan-2014 |
Christopher Ferris <cferris@google.com> |
ART: Speed up thread stack tracing for ANRs
Share a BacktraceMap for all threads being dumped, which speeds up
ANR dumping. Results from Nexus 9 (average of five):
Before: 0.587s
After: 0.206s
Change-Id: Ia70e0dbd39049318c02de561e7b95258d4849467
|
| 4f55e22630d99ca0edd9e951e5ee96b57bb9b980 |
|
04-Sep-2015 |
Mathieu Chartier <mathieuc@google.com> |
Add and use ScopedSuspendAll
Usage replaces most SuspendAll and ResumeAll calls.
Change-Id: I355683a5365876242cea85a656dcb58455f7a294
|
| 00370827646cc21cb370c3e7e93f9c0cff4c30c2 |
|
18-Aug-2015 |
Hiroshi Yamauchi <yamauchi@google.com> |
Use thread-local is_gc_marking flags for the CC collector.
The currently global is_marking flag is used to check if the read
barrier slow path needs to be taken for GC roots access. Changing it
to a thread-local flag simplifies the fast path check and makes it
easier to do it in assembly code. It also solves the issue that we
need to avoid accessing the global flag during startup before the heap
or the collector object isn't allocated and initialized.
Bug: 12687968
Change-Id: Ibf0dca12f400bf3490188b12dfe96c7de30583e0
|
| 511683764f2580d877725e20514ca4976c91bbfb |
|
13-Aug-2015 |
Mathieu Chartier <mathieuc@google.com> |
Wait for GC to finish in ThreadList::~ThreadList
Added a DisableGC to the heap instead of using the runtime shutdown
boolean. The runtime shutting down boolean is set to true before the
non daemon threads have exited. These threads may still be doing
allocations and we don't want to throw OOME if unnecessary.
Bug: 18577101
Change-Id: Iceb7048e6bd799aa2716099459c54f8dc0fb8feb
|
| 4382f1ef6ab2d6992d330cc8b43057d9139a1243 |
|
05-Aug-2015 |
Andreas Gampe <agampe@google.com> |
Revert "Revert "ART: Use bionic TLS slot for thread-self""
This reverts commit 50832b6804d15ea5ddfe99a2753a09111b5bfe37.
Fix clearing the TLS on detach.
Change-Id: I75ed5ccccb4f31800e5f893055e96067a9695258
|
| b19ccb1d88f0bff7371c9b72f265148677c18e95 |
|
15-Jul-2015 |
Mathieu Chartier <mathieuc@google.com> |
Use SuspendAllInternal for FlipThreadRoots
Fix FlipThreadRoots to wait until all the threads are suspended by
using SuspendAllInternal. Since running threads no longer hold the
mutator lock as shared held, doing an exclusive lock will not block
if the threads are not suspended. Also implemented MarkHeapReference
which is used to preserve soft references.
Fixes CC tests.
Change-Id: I4b059238f4249cf297e21ae918becd029fe26527
|
| 19af117525f37607fc866e3375ef1450e3f5ae35 |
|
14-Jul-2015 |
Mathieu Chartier <mathieuc@google.com> |
Fix mac build
Untested.
Change-Id: I1b7803f2fdcd7f509614aa9d3dce414e50a6119d
|
| eac4424b3420c280f97ff2f815b5dedd8dac9801 |
|
29-Jun-2015 |
Yu Li <yu.l.li@intel.com> |
ART: JNI thread state transition optimization
This patch improves the JNI performance by removing the explicit acquiring and
releasing the mutator lock when a thread state transits between suspended and
runnable states.
The functions responsible for changing the state were found to be the costliest
part of the JNI. Originally, a thread needs to acquire a shared mutator lock by
a CAS instruction when entering the runnable state and also needs to release
the lock by a CAS when entering the native state from runnable. This patch
removes these CAS operations when a thread state transits between suspended and
runnable. A thread in the runnable state is considered to have shared ownership
of the mutator lock and therefore transitions in and out of the runnable state
have associated implication on the mutator lock ownership. Meanwhile, a barrier
is added to control suspending all threads from running.
JNI transition overhead was reduced by 25% on IA platform and by 17% on ARM
platform by this patch, while it has little impact on GC pause time (measured
with "suspend all histogram").
Change-Id: Icee95d8ffff1bbfc95309a41cc48836536fec689
Signed-off-by: Yu, Li <yu.l.li@intel.com>
Signed-off-by: Haitao, Feng <haitao.feng@intel.com>
Signed-off-by: Lei, Li <lei.l.li@intel.com>
|
| 0b71357fb52be9bb06d35396a3042b4381b01041 |
|
17-Jun-2015 |
Hiroshi Yamauchi <yamauchi@google.com> |
Thread-local mark stacks for the CC collector.
Thread-local mark stacks are assigned to mutators where they push
references in read barriers to reduce the (CAS) synchronization cost
in a global mark stack/queue.
We step through three mark stack modes (thread-local, shared,
GC-exclusive) and use per-thread flags to disable/enable system weak
accesses (only for the CC collector) instead of the existing global
one to safely perform the marking phase. The reasons are 1)
thread-local mark stacks for mutators need to be revoked using a
checkpoint to avoid races (incorrectly leaving a reference on mark
stacks) when terminating marking, and 2) we can’t use a checkpoint
while system weak accesses are disabled (or a deadlock would
happen). More details are described in the code comments.
Performance improvements in Ritzperf EAAC: a ~2.8% improvement
(13290->12918) in run time and a ~23% improvement (51.6s->39.8s) in
the total GC time on N5.
Bug: 12687968
Change-Id: I5d234d7e48bf115cd773d38bdb62ad24ce9116c7
|
| 3cf225386e8129dcbe32b289279ecb87ec255318 |
|
10-Jul-2015 |
Mathieu Chartier <mathieuc@google.com> |
Clean up some includes
Change-Id: Ia03a3b54b235df38d5cfe096fef1aebe2b80eb29
|
| 6ba35b50347aa7418c66c7b046cd164987e95df3 |
|
01-Jun-2015 |
Sebastien Hertz <shertz@google.com> |
JDWP: asynchronous invoke command handling
The JDWP thread used to wait for the result of a method invocation
running in an event thread. But doing that prevents the JDWP thread
from processing incoming commands from the debugger if the event
thread gets suspended by a debug event occurring in another thread.
In Android Studio (or another IDE), this leads to the debugger being
blocked (with the famous message "Waiting until last debugger command
completes" of Android Studio / IntelliJ) because it is actually
waiting for the reply of its latest command while the JDWP thread
cannot process it.
This CL changes the way invoke commands (ClassType.InvokeCommand,
ClassType.NewInstance and ObjectReference.InvokeCommand) are handled
in the ART runtime.
The JDWP thread no longer waits for the event thread to complete the
method invocation. It now simply waits for the next JDWP command to
process. This means it does not send any reply for invoke commands,
except if the information given by the debugger is wrong. In this
case, it still sends a reply with the appropriate error code.
The event thread is now responsible for sending the reply (containing
the result and the exception object of the invoked method) before
going back to the suspended state.
In other words, we add special handling for invoke commands so they
are handled asynchronously while other commands remained handled
synchronously. In the future, we may want to handle all commands
asynchronously (using a queue of reply/event for instance) to remove
the special handling code this CL is adding.
Now the JDWP thread can process commands while a thread is invoking
a method, it is possible for the debugger to detach (by sending a
VirtualMachine.Dispose command) before the invocation completes. In
that situation, we must not suspend threads again (including the
event thread that executed the method) because they would all remain
suspended forever.
Also minor cleanup of the use of JDWP constants and update comments.
Bug: 21515842
Bug: 18899981
(cherry picked from commit cbc5064ff05179b97b416f00ca579c55e38cd7d9)
Change-Id: I8d31006043468913ee8453212e6d16e11fcfe4ea
|
| cbc5064ff05179b97b416f00ca579c55e38cd7d9 |
|
01-Jun-2015 |
Sebastien Hertz <shertz@google.com> |
JDWP: asynchronous invoke command handling
The JDWP thread used to wait for the result of a method invocation
running in an event thread. But doing that prevents the JDWP thread
from processing incoming commands from the debugger if the event
thread gets suspended by a debug event occurring in another thread.
In Android Studio (or another IDE), this leads to the debugger being
blocked (with the famous message "Waiting until last debugger command
completes" of Android Studio / IntelliJ) because it is actually
waiting for the reply of its latest command while the JDWP thread
cannot process it.
This CL changes the way invoke commands (ClassType.InvokeCommand,
ClassType.NewInstance and ObjectReference.InvokeCommand) are handled
in the ART runtime.
The JDWP thread no longer waits for the event thread to complete the
method invocation. It now simply waits for the next JDWP command to
process. This means it does not send any reply for invoke commands,
except if the information given by the debugger is wrong. In this
case, it still sends a reply with the appropriate error code.
The event thread is now responsible for sending the reply (containing
the result and the exception object of the invoked method) before
going back to the suspended state.
In other words, we add special handling for invoke commands so they
are handled asynchronously while other commands remained handled
synchronously. In the future, we may want to handle all commands
asynchronously (using a queue of reply/event for instance) to remove
the special handling code this CL is adding.
Now the JDWP thread can process commands while a thread is invoking
a method, it is possible for the debugger to detach (by sending a
VirtualMachine.Dispose command) before the invocation completes. In
that situation, we must not suspend threads again (including the
event thread that executed the method) because they would all remain
suspended forever.
Also minor cleanup of the use of JDWP constants and update comments.
Bug: 21515842
Bug: 18899981
Change-Id: I15e00fb068340f3d69dc9225d8d2065246e68c58
|
| bf44d42d978861f33fd5978f49392db3321ae414 |
|
02-Jun-2015 |
Mathieu Chartier <mathieuc@google.com> |
Add a way to SuspendAll for a long duration
Hprof uses this mode to not cause thread suspend timeouts if GC tries
while the hprof dump is running.
(cherry picked from commit 77be6635f9b07a8a794924c5fb9b071949776a6d)
Bug: 21063989
Change-Id: Ic6304620afd1489719a7e0e4299f829c90fe27cc
|
| 77be6635f9b07a8a794924c5fb9b071949776a6d |
|
02-Jun-2015 |
Mathieu Chartier <mathieuc@google.com> |
Add a way to SuspendAll for a long duration
Hprof uses this mode to not cause thread suspend timeouts if GC tries
while the hprof dump is running.
Bug: 21063989
Change-Id: Ic6304620afd1489719a7e0e4299f829c90fe27cc
|
| 41b175aba41c9365a1c53b8a1afbd17129c87c14 |
|
19-May-2015 |
Vladimir Marko <vmarko@google.com> |
ART: Clean up arm64 kNumberOfXRegisters usage.
Avoid undefined behavior for arm64 stemming from 1u << 32 in
loops with upper bound kNumberOfXRegisters.
Create iterators for enumerating bits in an integer either
from high to low or from low to high and use them for
<arch>Context::FillCalleeSaves() on all architectures.
Refactor runtime/utils.{h,cc} by moving all bit-fiddling
functions to runtime/base/bit_utils.{h,cc} (together with
the new bit iterators) and all time-related functions to
runtime/base/time_utils.{h,cc}. Improve test coverage and
fix some corner cases for the bit-fiddling functions.
Bug: 13925192
(cherry picked from commit 80afd02024d20e60b197d3adfbb43cc303cf29e0)
Change-Id: I905257a21de90b5860ebe1e39563758f721eab82
|
| 80afd02024d20e60b197d3adfbb43cc303cf29e0 |
|
19-May-2015 |
Vladimir Marko <vmarko@google.com> |
ART: Clean up arm64 kNumberOfXRegisters usage.
Avoid undefined behavior for arm64 stemming from 1u << 32 in
loops with upper bound kNumberOfXRegisters.
Create iterators for enumerating bits in an integer either
from high to low or from low to high and use them for
<arch>Context::FillCalleeSaves() on all architectures.
Refactor runtime/utils.{h,cc} by moving all bit-fiddling
functions to runtime/base/bit_utils.{h,cc} (together with
the new bit iterators) and all time-related functions to
runtime/base/time_utils.{h,cc}. Improve test coverage and
fix some corner cases for the bit-fiddling functions.
Bug: 13925192
Change-Id: I704884dab15b41ecf7a1c47d397ab1c3fc7ee0f7
|
| 2cebb24bfc3247d3e9be138a3350106737455918 |
|
22-Apr-2015 |
Mathieu Chartier <mathieuc@google.com> |
Replace NULL with nullptr
Also fixed some lines that were too long, and a few other minor
details.
Change-Id: I6efba5fb6e03eb5d0a300fddb2a75bf8e2f175cb
|
| bb87e0f1a52de656bc77cb01cb887e51a0e5198b |
|
03-Apr-2015 |
Mathieu Chartier <mathieuc@google.com> |
Refactor and improve GC root handling
Changed GcRoot to use compressed references. Changed root visiting to
use virtual functions instead of function pointers. Changed root visting
interface to be an array of roots instead of a single root at a time.
Added buffered root marking helper to avoid dispatch overhead.
Root marking seems a bit faster on EvaluateAndApplyChanges due to batch
marking. Pause times unaffected.
Mips64 is untested but might work, maybe.
Before:
MarkConcurrentRoots: Sum: 67.678ms 99% C.I. 2us-664.999us Avg: 161.138us Max: 671us
After:
MarkConcurrentRoots: Sum: 54.806ms 99% C.I. 2us-499.986us Avg: 136.333us Max: 602us
Bug: 19264997
Change-Id: I0a71ebb5928f205b9b3f7945b25db6489d5657ca
|
| bf9fc581e8870faddbd320a935f9a627da724c48 |
|
14-Mar-2015 |
Mathieu Chartier <mathieuc@google.com> |
Add more info to who called SuspendAll
Helps diagnose related jank.
Change-Id: I38191cdda723c6f0355d0197c494a3dff2b6653c
|
| 1558b577907b613864e98f05862543557263e864 |
|
25-Feb-2015 |
Sebastien Hertz <shertz@google.com> |
JDWP: allocate DebugInvokeReq only when requested
Only allocates thread-local DebugInvokeReq when the debugger requests
a thread to invoke a method. The JDWP thread allocates that structure
then attaches it to the target thread. When the thread is resumed, it
executes the method. Once the invocation completes, the thread
detaches the DebugInvokeReq, signals the JDWP thread then suspends.
Finally, the JDWP thread wakes up, prepares the reply with the invoke
result (or exception) and deallocates the DebugInvokeReq.
Also ensures GC safety for object returned by the invoke. We add the
object to the JDWP object registry right after the invoke. We now
reference that object with a JDWP ObjectID instead of an Object* in
the DebugInvokeReq struct. This prevent from accessing a stale
reference if the GC runs and moves the Object*.
This CL includes the following changes:
- Move former DebugInvokeReq::ready flag to
Thread::tls_32bit_sized_values::ready_for_debug_invoke. It's needed
to know whether a thread has been suspended by an event, thus ready
to invoke a method from the debugger.
- Remove DebugInvokeReq::invoke_needed: we now test if we attached a
DebugInvokeReq* to the thread.
- Rename misleading FinishMethod function to RequestMethod.
Bug: 19142632
Bug: 18166750
Change-Id: I351fb4eb94bfe69fcafb544d21d55ff35a033000
|
| e15ea086439b41a805d164d2beb07b4ba96aaa97 |
|
10-Feb-2015 |
Hiroshi Yamauchi <yamauchi@google.com> |
Reserve bits in the lock word for read barriers.
This prepares for the CC collector to use the standard object header
model by storing the read barrier state in the lock word.
Bug: 19355854
Bug: 12687968
Change-Id: Ia7585662dd2cebf0479a3e74f734afe5059fb70f
|
| 91e56692c6bd9fa1d41951ee7dc311f19461f4be |
|
03-Mar-2015 |
Mathieu Chartier <mathieuc@google.com> |
Wait for threads to finish unregistering
There was a race where Thread::join would return before the thread was
unregistered. This caused a problem with Daemons.stop since the thread
list could get deleted before the daemon thread was removed from list_.
This caused occasional "Request to unregister unattached thread"
errors and warnings.
The fix is to wait until threads finish registering before destroying
the thread list. The only threads which can be unregistering at this
point are the daemons we stopped earlier during the runtime shutdown
process. The issue is that thread join finishes before we remove the
thread from the thread list.
Also some cleanup.
Bug: 18713034
Change-Id: I8921122fe8462643a6b814b5f00632481e3831fb
|
| 9db831acffa509b2e46c1d63fe25650d3bedd415 |
|
25-Feb-2015 |
Mathieu Chartier <mathieuc@google.com> |
Print thread name for attempting to unregister unattached thread
Bug: 18713034
Change-Id: Ic0654c6569277e5bbd6d2582140595faaa36cded
|
| 794ad76e8d5b5b9132819d5b08a0570e27615644 |
|
23-Feb-2015 |
Andreas Gampe <agampe@google.com> |
ART: Introduce NO_RETURN, Mark DoLongJump noreturn
Add NO_RETURN macro that adds C++11 noreturn attribute. Mark
DoLongJump methods as noreturn.
Change-Id: Ifde4318e370493237050d4c1349285a0382df23f
|
| c4448f4866b10f021c38947aac656746e62e5fa3 |
|
04-Feb-2015 |
Mathieu Chartier <mathieuc@google.com> |
Fix incorrect static const
Bug: 19265183
(cherry picked from commit 3a958aac75ec491a71fba4a9ec345cdd1817708c)
Change-Id: Iabb1abd81b022469b3b8aabd2d2a1c29f3b10ea0
|
| 3a958aac75ec491a71fba4a9ec345cdd1817708c |
|
04-Feb-2015 |
Mathieu Chartier <mathieuc@google.com> |
Fix incorrect static const
Bug: 19265183
Change-Id: I573df60fe90952246d2126ced010f2bdac006bca
|
| 991438608d5223128a199282b85418f08c471b35 |
|
03-Feb-2015 |
Mathieu Chartier <mathieuc@google.com> |
Fix occasional long ThreadSuspendSleep
Since usleep can wait much longer than the specified us, we use
sched_yield instead for more predictable behavior.
Bug: 18379850
Change-Id: I729212251db1e86e470240b7a042bbf2d4e981cd
|
| dd9943d4466b052ef6c5ee5b32187adb48cbce74 |
|
02-Feb-2015 |
Lei Li <lei.l.li@intel.com> |
ART: checkpoint mechanism optimization
GC thread and trim thread are both using checkpoint mechanism. GC thread
will request java threads to mark their thread roots by themselves. Trim
thread will request java threads to trim their jni local reference
tables by themselves.
The checkpint mechanism semantics is that the runnable java threads will
run checkpoint function itself at safepoint, and finally the java
threads and gc thread or trim thread is synchronized via barrier. If the
java threads are not runnable, gc thread or trim thread will suspend
them and then run their checkpoint functions one by one on behalf of
them. If all the java threads are not runnable, then gc thread or trim
thread will do all the work itself. In this case, there is no need
synchronization. This will save unnecessary synchronization and thread
state transitions.
Change-Id: If55940946cb3f8b1af42c7237c334f09c8ec7a9f
|
| 2cd334ae2d4287216523882f0d298cf3901b7ab1 |
|
09-Jan-2015 |
Hiroshi Yamauchi <yamauchi@google.com> |
More of the concurrent copying collector.
Bug: 12687968
Change-Id: I62f70274d47df6d6cab714df95c518b750ce3105
|
| e34fa1df67fbe0173b4ea9abddcc3ae3d0537037 |
|
14-Jan-2015 |
Mathieu Chartier <mathieuc@google.com> |
Print more info in MarkSweep::VerifyRoot
Refactored old root callback to use a new class called RootInfo.
RootInfo contains all the relevant info related to the root
associated with the callback. The MarkSweep::VerifyRoot function
now uses this info to print the StackVisitor's described location
if the GC root is of the type kRootJavaFrame.
Some other cleanup.
Example output:
E/art (12167): Tried to mark 0x123 not contained by any spaces
E/art (12167): Attempting see if it's a bad root
E/art (12167): Found invalid root: 0x123 with type RootJavaFrame
E/art (12167): Location=Visiting method
'void java.lang.Runtime.gc()' at dex PC 0xffffffff (native PC 0x0)
vreg=0
(cherry picked from commit 12f7423a2bb4bfab76700d84eb6d4338d211983a)
Bug: 18588862
Change-Id: Ic5a2781f704e931265ffb3621c2eab4b2e25f60f
|
| 12f7423a2bb4bfab76700d84eb6d4338d211983a |
|
14-Jan-2015 |
Mathieu Chartier <mathieuc@google.com> |
Print more info in MarkSweep::VerifyRoot
Refactored old root callback to use a new class called RootInfo.
RootInfo contains all the relevant info related to the root
associated with the callback. The MarkSweep::VerifyRoot function
now uses this info to print the StackVisitor's described location
if the GC root is of the type kRootJavaFrame.
Some other cleanup.
Example output:
E/art (12167): Tried to mark 0x123 not contained by any spaces
E/art (12167): Attempting see if it's a bad root
E/art (12167): Found invalid root: 0x123 with type RootJavaFrame
E/art (12167): Location=Visiting method
'void java.lang.Runtime.gc()' at dex PC 0xffffffff (native PC 0x0)
vreg=0
Bug: 18588862
Change-Id: Ic5a2781f704e931265ffb3621c2eab4b2e25f60f
|
| 1e4b0ca3bde8ea792aad83aefc891628ff6ec806 |
|
14-Jan-2015 |
Andreas Gampe <agampe@google.com> |
ART: Change ternary location for timeout
Make the decision for timeout length in the constant, not at the
use site.
Bug: 18964791
Change-Id: I73e6ef5e1c23fd515eaa61ba68236698d3e5c81a
|
| 66516ac4547fe40f26e4e117b3ee595904607b95 |
|
09-Jan-2015 |
Sebastien Hertz <shertz@google.com> |
JDWP: allow VirtualMachine.Resume on partial suspension
Allows to resume multiple threads with a single VirtualMachine.Resume
command. When some threads are suspended for an event by the debugger
with a suspend count of 1 (other threads are running), a debugger can
resume them all this way.
Bug: 18924933
(cherry picked from commit f9d233dae8ca66ed5a2a474155d6bee0d74c355b)
Change-Id: I948c3b50d43693e86ae31e0928a33a467328583f
|
| 4a3d19bfc16e3f5bfb796fb7ed24404fa77007a4 |
|
10-Jan-2015 |
Andreas Gampe <agampe@google.com> |
ART: Increase dump checkpoint timeout on the host
Increase the dump checkpoint timeout on the host, as addr2line is
used and may put more strain onto the system.
Bug: 18964791
Change-Id: I9c1e9c1289fddfce25854e4425194882628bcf21
|
| f9d233dae8ca66ed5a2a474155d6bee0d74c355b |
|
09-Jan-2015 |
Sebastien Hertz <shertz@google.com> |
JDWP: allow VirtualMachine.Resume on partial suspension
Allows to resume multiple threads with a single VirtualMachine.Resume
command. When some threads are suspended for an event by the debugger
with a suspend count of 1 (other threads are running), a debugger can
resume them all this way.
Bug: 18924933
Change-Id: I81543df8228d56a4cf201e59885015880650f202
|
| 23f6e69e374ce4a7fd0d45f633a3c9efcc8e8a07 |
|
19-Dec-2014 |
Mathieu Chartier <mathieuc@google.com> |
Only dump suspend all histogram if we have samples
Helps fix gcstress ThreadStress.
Bug: 18576985
Change-Id: Ibebc0fed8a21e7b149f4b6e5154fb86deddcc68f
|
| 70a596d61f8cf5b6447326c46c3386e0fbd5bfb5 |
|
17-Dec-2014 |
Mathieu Chartier <mathieuc@google.com> |
Add thread suspend histogram
Helps measure time to suspend.
Example output (maps after a few seconds):
suspend all histogram: Sum: 2.806ms 99% C.I. 2us-1090.560us Avg: 43.843us Max: 1126us
Change-Id: I7bd9dd3b401fb3e3059e8718556d60910e541611
|
| f0dc8b5519102b3d3e738aed846975ae4239421e |
|
17-Dec-2014 |
Mathieu Chartier <mathieuc@google.com> |
Add systrace events to monitors and thread suspension
Added systrace events to SuspendThreadByPeer, SuspendThreadByThreadId
and monitor contention.
Bug: 18379850
Bug: 17470248
Change-Id: I29eaf97c67361a6c6b437940b8729600d78696df
|
| 82a800d1408937951155f546a3671b83b7379ab5 |
|
16-Dec-2014 |
Mathieu Chartier <mathieuc@google.com> |
Decrement suspend count if thread is shutting down
Prevents deadlock caused by incrementing suspend count in
SuspendThreadByPeer, then getting a cleared nativePeer field. This
resulted in us not decrementing the suspend count which caused a
deadlock in WaitForOtherNonDaemonThreadsToExit.
Bug: 18739541
Change-Id: I4a63f1823993a0f99f32025cd479072be49ba8d5
|
| db978719dbcb73fc6acfd193561445c4462786b8 |
|
09-Dec-2014 |
Nicolas Geoffray <ngeoffray@google.com> |
Revert "Tidy gAborting."
Creates infinite loop: b/18674776.
This reverts commit 015b137efb434528173779bc3ec8d72494456254.
Change-Id: I67fe310d2e95ee2ec37bec842be06fb1123b6f4e
|
| 015b137efb434528173779bc3ec8d72494456254 |
|
05-Dec-2014 |
Ian Rogers <irogers@google.com> |
Tidy gAborting.
Reduce scope to Runtime::Abort and short-cut recursive case earlier. gAborting
remains global to avoid two fatal errors in thread and the verifier.
Change-Id: Ibc893f891ffee9a763c65cde9507d99083d47b3f
|
| 7443c9abb22b10fc45895b2171fc0f67866863b7 |
|
26-Nov-2014 |
Andreas Gampe <agampe@google.com> |
ART: Avoid a recursive abort
When waiting for checkpoints on abort, don't abort on timeout.
Change-Id: Ic87c25d466021d391563f870642251b69dd62af5
|
| 4ad5cd3e7d519484559ef778d96fb3f0be8919fa |
|
12-Nov-2014 |
Ian Rogers <irogers@google.com> |
Modify the behavior of thread suspend shootouts.
The thread doing the suspension doesn't attempt to suspend the other thread
unless it knows another thread isn't trying to suspend it. Use the suspend
count, and its lock, for this purpose.
Re-enable ThreadStress test.
Bug: 15446488
Change-Id: Idd34410c7b89d8abd6973e5699a15ca699472c78
|
| 277ccbd200ea43590dfc06a93ae184a765327ad0 |
|
04-Nov-2014 |
Andreas Gampe <agampe@google.com> |
ART: More warnings
Enable -Wno-conversion-null, -Wredundant-decls and -Wshadow in general,
and -Wunused-but-set-parameter for GCC builds.
Change-Id: I81bbdd762213444673c65d85edae594a523836e5
|
| 8409ec440079020bbe4ad066cf18a5fadfba67d2 |
|
05-Nov-2014 |
Ian Rogers <irogers@google.com> |
Make thread dumping more tolerant of broken invariants during abort.
Change-Id: Ie1bc20debe72f2ea2bf4c1be50cd9877c823670a
|
| 6a3c1fcb4ba42ad4d5d142c17a3712a6ddd3866f |
|
31-Oct-2014 |
Ian Rogers <irogers@google.com> |
Remove -Wno-unused-parameter and -Wno-sign-promo from base cflags.
Fix associated errors about unused paramenters and implict sign conversions.
For sign conversion this was largely in the area of enums, so add ostream
operators for the effected enums and fix tools/generate-operator-out.py.
Tidy arena allocation code and arena allocated data types, rather than fixing
new and delete operators.
Remove dead code.
Change-Id: I5b433e722d2f75baacfacae4d32aef4a828bfe1b
|
| 4eec026ec4d86869baa27ad0afd9564cca8856c9 |
|
14-Oct-2014 |
Sebastien Hertz <shertz@google.com> |
Fix JDWP Virtualmachine.Resume command
We used to call ThreadList::UndoDebuggerSuspensions but this is not
correct. Indeed Virtualmachine.Resume command should be performed
as many times as all threads have been suspended (either by the
Virtualmachine.Suspend command or the application of suspend policy
ALL for event).
This CL fixes this issue by adding ThreadList::ResumeAllForDebugger.
It decrements suspend count and debug suspend count by 1 (globally
and for each thread).
It also handles the case where Virtualmachine.Resume is called while
threads have not been suspended (global debug suspend count is 0). In
this case we just print a warning and do nothing.
Bug: 17974047
(cherry picked from commit 253fa555d9424a56c52879449eabe73d5c96cf61)
Change-Id: I38b38240f84f141670506e88fa5c2480f856fdd6
|
| c7dd295a4e0cc1d15c0c96088e55a85389bade74 |
|
22-Oct-2014 |
Ian Rogers <irogers@google.com> |
Tidy up logging.
Move gVerboseMethods to CompilerOptions. Now "--verbose-methods=" option to
dex2oat rather than runtime argument "-verbose-methods:".
Move ToStr and Dumpable out of logging.h, move LogMessageData into logging.cc
except for a forward declaration.
Remove ConstDumpable as Dump methods are all const (and make this so if not
currently true).
Make LogSeverity an enum and improve compile time assertions and type checking.
Remove log_severity.h that's only used in logging.h.
With system headers gone from logging.h, go add to .cc files missing system
header includes.
Also, make operator new in ValueObject private for compile time instantiation
checking.
Change-Id: I3228f614500ccc9b14b49c72b9821c8b0db3d641
|
| 253fa555d9424a56c52879449eabe73d5c96cf61 |
|
14-Oct-2014 |
Sebastien Hertz <shertz@google.com> |
Fix JDWP Virtualmachine.Resume command
We used to call ThreadList::UndoDebuggerSuspensions but this is not
correct. Indeed Virtualmachine.Resume command should be performed
as many times as all threads have been suspended (either by the
Virtualmachine.Suspend command or the application of suspend policy
ALL for event).
This CL fixes this issue by adding ThreadList::ResumeAllForDebugger.
It decrements suspend count and debug suspend count by 1 (globally
and for each thread).
It also handles the case where Virtualmachine.Resume is called while
threads have not been suspended (global debug suspend count is 0). In
this case we just print a warning and do nothing.
Bug: 17974047
Change-Id: I25af22b1936760292d2a8c286e0867f2ae864108
|
| e094b87c6f6ea9ebf83aa56a3114ac59556aaf9f |
|
14-Oct-2014 |
Jeff Hao <jeffhao@google.com> |
Store exiting thread ids and names while tracing to output later.
Bug: 17909204
(cherry picked from commit fdcbc5c4c7c67bba06e038ac96a2e8bc49b91f84)
Change-Id: Iea087560ba8b983412a6dde2ec166c9e7214f3a1
|
| fdcbc5c4c7c67bba06e038ac96a2e8bc49b91f84 |
|
14-Oct-2014 |
Jeff Hao <jeffhao@google.com> |
Store exiting thread ids and names while tracing to output later.
Bug: 17909204
Change-Id: I4ce027af7c51fb310d6cdbdf067ae340e4c932ff
|
| fec72f4a625122ab9972708b1f44f24b53a8b734 |
|
09-Oct-2014 |
Mathieu Chartier <mathieuc@google.com> |
Fix race in thread list destructor
The ThreadList::Contains wasn't being guarded by the thread list
lock, which could cause a corrupted thread list if another thread
unregistered itself while std::find was searching the thread list.
Bug: 17896374
(cherry picked from commit b90132cb5132eaeb4dbfca1e63d79b4005dacec5)
Change-Id: I8f6f979f365ca00ac0655e04eb26020d2ad7d6ee
|
| b90132cb5132eaeb4dbfca1e63d79b4005dacec5 |
|
09-Oct-2014 |
Mathieu Chartier <mathieuc@google.com> |
Fix race in thread list destructor
The ThreadList::Contains wasn't being guarded by the thread list
lock, which could cause a corrupted thread list if another thread
unregistered itself while std::find was searching the thread list.
Bug: 17896374
Change-Id: I629b1eac3473e3cf7355ebde9080dfd7c58af73a
|
| 5869a2c27ee0dbd7b420614c76ff14a706f0c5fb |
|
08-Oct-2014 |
Mathieu Chartier <mathieuc@google.com> |
Dump mutator lock for thread suspend timeout
Should help us see if thread suspend timeouts are due to someone
holding on the mutator lock for too long or if it is a missed
futex wake.
(cherry picked from commit f924d2381a5ea53967ba1e279766d601b9be05ea)
Bug: 17837911
Change-Id: I3417156a78f574076d5da17c86836c5a7c76084a
|
| f924d2381a5ea53967ba1e279766d601b9be05ea |
|
08-Oct-2014 |
Mathieu Chartier <mathieuc@google.com> |
Dump mutator lock for thread suspend timeout
Should help us see if thread suspend timeouts are due to someone
holding on the mutator lock for too long or if it is a missed
futex wake.
Bug: 17837911
Change-Id: I3417156a78f574076d5da17c86836c5a7c76084a
|
| dcfd6cad9670ca3ae460a4af0bbe7d88cae21a3c |
|
18-Sep-2014 |
Sebastien Hertz <shertz@google.com> |
Move spammy logs to JDWP verbose mode
We are spammed by warning messages when debugging, especially each time we
suspend/resume all threads (to update instrumentation or collect monitor info).
It's common to get into the cases where these warnings are logged so they
shouldn't be warning but debug messages.
This CL moves these LOG(WARNING) to VLOG(jdwp) to not disturb developers when
debugging their app (especially when looking for specific messages in logcat).
We keep them in JDWP verbose mode because they help knowing when we initiate
these sequences of "suspend/resume all threads".
Also adds debug suspend count in the log message for more context.
Bug: 17524544
Bug: 17170697
(cherry picked from commit f272af4b9dcd39cdd50fa6655601a26e837eaea9)
Change-Id: I61df70ace1475bf10d83202b6bb774f7036354cb
|
| f272af4b9dcd39cdd50fa6655601a26e837eaea9 |
|
18-Sep-2014 |
Sebastien Hertz <shertz@google.com> |
Move spammy logs to JDWP verbose mode
We are spammed by warning messages when debugging, especially each time we
suspend/resume all threads (to update instrumentation or collect monitor info).
It's common to get into the cases where these warnings are logged so they
shouldn't be warning but debug messages.
This CL moves these LOG(WARNING) to VLOG(jdwp) to not disturb developers when
debugging their app (especially when looking for specific messages in logcat).
We keep them in JDWP verbose mode because they help knowing when we initiate
these sequences of "suspend/resume all threads".
Also adds debug suspend count in the log message for more context.
Bug: 17524544
Bug: 17170697
Change-Id: Ic87985ac6913151d15fd89849e41bde61092c3dd
|
| a2af5c7e3262bb7165e9320426bd6925f1a3c90d |
|
16-Sep-2014 |
Ian Rogers <irogers@google.com> |
Only log an error if an unattached thread is unregistered.
Turn a fatal check into a diagnostic error. It looks like this issue arises
when a runtime shutsdown without properly being started.
Bug: 17011539
Change-Id: I2983c8332e83769e9480e8f30a46ca3b80a2e90e
(cherry picked from commit e2bcf579b0bc2f53db68c38755d69543198c7b00)
|
| e2bcf579b0bc2f53db68c38755d69543198c7b00 |
|
16-Sep-2014 |
Ian Rogers <irogers@google.com> |
Only log an error if an unattached thread is unregistered.
Turn a fatal check into a diagnostic error. It looks like this issue arises
when a runtime shutsdown without properly being started.
Bug: 17011539
Change-Id: I2983c8332e83769e9480e8f30a46ca3b80a2e90e
|
| 2156ff1ff09166ad8270f27d758e913f120c3640 |
|
14-Sep-2014 |
Ian Rogers <irogers@google.com> |
Add time out to dump check point.
Change-Id: I2263c092c9957585ae90cf846e77dc8861f8a14b
|
| 7b078e8c04f3e1451dbdd18543c8b9692b5b067e |
|
10-Sep-2014 |
Ian Rogers <irogers@google.com> |
Compile time performance improvements focusing on interpret-only.
Reduce virtual method dispatch in the method verifier and make more code
inline-able.
Add a StringPiece with const char* equality operator to avoid redundant
StringPieces and strlens.
Remove back link from register line to verifier and pass as argument to reduce
size of RegisterLine.
Remove instruction length from instruction flags and compute from the
instruction, again to reduce size.
Add suspend checks to resolve and verify to allow for more easy monitor
inflation and reduce contention on Locks::thread_list_suspend_thread_lock_.
Change ThrowEarlierClassFailure to throw pre-allocated exception.
Avoid calls to Thread::Current() by passing self.
Template specialize IsValidClassName.
Make ANR reporting with SIGQUIT run using checkpoints rather than suspending
all threads. This makes the stack/lock analysis less lock error prone.
Extra Barrier assertions and condition variable time out is now returned as a
boolean both from Barrier and ConditionVariable::Wait.
2 threaded host x86-64 interpret-only numbers from 341 samples:
Before change: Avg 176.137ms 99% CI 3.468ms to 1060.770ms
After change: Avg 139.163% 99% CI 3.027ms to 838.257ms
Reduction in average compile time after change is 20.9%.
Slow-down without change is 26.5%.
Bug: 17471626 - Fix bug where RegTypeCache::JavaLangObject/String/Class/Throwable
could return unresolved type when class loading is disabled.
Bug: 17398101
Change-Id: Id59ce3cc520701c6ecf612f7152498107bc40684
|
| ed2be1725fb79075892b1a9103487c9d9a95b350 |
|
19-Aug-2014 |
Sebastien Hertz <shertz@google.com> |
Reduce lock contention when debugging
Uses a ReaderWriterMutex for the breakpoint lock to reduce contention during
debugging session.
Also adds missing thread safety annotations on fields and methods related to
instrumentation and debugging.
Bug: 16814665
Bug: 11667502
(cherry picked from commit 59d9d668d4f4286813afe2b4e7c6db839222ce96)
Change-Id: I5f1156da8c6cc8316d6db16e6cfb2470fe289ad3
|
| ba32de47e32f436d7c11cb4a2e78bdd4ad4dc5d2 |
|
28-Aug-2014 |
Brian Carlstrom <bdc@google.com> |
Fix issue with Thread.setName hanging after Thread.start
When suspending a thread by peer the invariant that only attached threads are
suspended must be maintained. Add a ThreadList::Contains check which requires
making this method non-static.
Add some extra thread logging.
Bug: 17302037
(cherry picked from commit 37c16453a92bbf1a47f042000318a1b60381017d)
Change-Id: I51832785d4b4b431e035318e75635f442e89a1fb
|
| 37c16453a92bbf1a47f042000318a1b60381017d |
|
28-Aug-2014 |
Brian Carlstrom <bdc@google.com> |
Fix issue with Thread.setName hanging after Thread.start
When suspending a thread by peer the invariant that only attached threads are
suspended must be maintained. Add a ThreadList::Contains check which requires
making this method non-static.
Add some extra thread logging.
Bug: 17302037
Change-Id: I89af9840e398c2ceadeb736c431dfe3be8f3ba06
|
| 59d9d668d4f4286813afe2b4e7c6db839222ce96 |
|
19-Aug-2014 |
Sebastien Hertz <shertz@google.com> |
Reduce lock contention when debugging
Uses a ReaderWriterMutex for the breakpoint lock to reduce contention during
debugging session.
Also adds missing thread safety annotations on fields and methods related to
instrumentation and debugging.
Bug: 16814665
Bug: 11667502
Change-Id: I056cdafa91109e0c83806c8d8df75c37ade0a354
|
| 68d8b42ddec39ec0174162d90d4abaa004d1983e |
|
17-Jul-2014 |
Ian Rogers <irogers@google.com> |
Wire up check JNI force copy mode.
Increase check JNI checks.
Break apart jni_internal.h in to jni_env_ext.h and java_vm_ext.h.
Fix the abuse of ScopedObjectAccess/annotalysis by ScopedCheck in the case
of VM routines.
Make class loader override and shared library class loader JNI global
references rather than mirror pointers.
Clean-ups to native bridge.
Change-Id: If7c6110b5aade7a402bfb67534af86a7b2cdeb55
|
| c7c6d18a90079b67a7592d3f2cb3d4c6f0081512 |
|
29-Jul-2014 |
Jeff Hao <jeffhao@google.com> |
Stop thread from reattaching during runtime shutdown while tracing.
Bug: 16024763
(cherry picked from commit c5d824a20c225763761a6dff43294b229ff35469)
Change-Id: Ideab9adedabf3c58f5162e43e79c72af0642bdfa
|
| 82558acbca3eae5b7e47748bcfb8070855266676 |
|
30-Jul-2014 |
Jeff Hao <jeffhao@google.com> |
Submitted wrong changelist!
Revert "Stop thread from reattaching during runtime shutdown while tracing."
This reverts commit c96657c4243e04888d5948fa023d78cc9213800b.
Change-Id: I859347f93704a94d23f72ef7d483298b08bbed0d
|
| c96657c4243e04888d5948fa023d78cc9213800b |
|
29-Jul-2014 |
Jeff Hao <jeffhao@google.com> |
Stop thread from reattaching during runtime shutdown while tracing.
Bug: 16024763
(cherry picked from commit c5d824a20c225763761a6dff43294b229ff35469)
Change-Id: Id5dcd4c8a131b93de00023c3ae9f518ea182614a
|
| c5d824a20c225763761a6dff43294b229ff35469 |
|
29-Jul-2014 |
Jeff Hao <jeffhao@google.com> |
Stop thread from reattaching during runtime shutdown while tracing.
Bug: 16024763
Change-Id: Iad5ba180241ff74b15baf5c3a15ed2d2ed60fcf0
|
| f3d874c60ee3ada19ce26a5c4e532312b6f3a9e9 |
|
18-Jul-2014 |
Ian Rogers <irogers@google.com> |
Avoid race in single thread suspension.
Don't allow more than one concurrent single thread suspension to avoid
potential cycles and deadlocks where threads try to suspend each other.
Bug: 16364458, 16354227
Change-Id: I907f1d5591a6aa5c241d37d6b4a34f968f98df77
|
| 6093a5c277e54bcd949dd6fac7b3856e5f371d06 |
|
18-Jul-2014 |
Ian Rogers <irogers@google.com> |
Avoid race in single thread suspension.
Don't allow more than one concurrent single thread suspension to avoid
potential cycles and deadlocks where threads try to suspend each other.
Bug: 16364458, 16354227
(cherry picked from commit f3d874c60ee3ada19ce26a5c4e532312b6f3a9e9)
Change-Id: I907f1d5591a6aa5c241d37d6b4a34f968f98df77
|
| b960cd7342cd68a614bb0c54d0d676cc2b8caf29 |
|
16-Jul-2014 |
Mathieu Chartier <mathieuc@google.com> |
Use sched_yield in Monitor::MonitorEnter.
Previously we used NanoSleep(1000), but this was unreliable. It could
result in waiting for >= 40ms instead of 1us. Since this was in a loop
it was especially bad if the GC was trying to suspend all the
threads when we were sleeping. This resulted in thread suspension
occasionally taking longer than a second.
Results on the provided picasso-sample app on Nexus 5:
Longest GC pause before: ~1.5s.
Longest GC pause after: <5ms.
Also added a warning if thread suspension takes longer than a
threshold (currently 5ms).
Bug: 16307460
External bug: https://code.google.com/p/android-developer-preview/issues/detail?id=367
(cherry picked from commit 251755cd511463260e60be98bf138b6aa1c14bf3)
Change-Id: I491c0799cc23ee6e68abc1bf77d1f0f4263e8ef5
|
| 251755cd511463260e60be98bf138b6aa1c14bf3 |
|
16-Jul-2014 |
Mathieu Chartier <mathieuc@google.com> |
Use sched_yield in Monitor::MonitorEnter.
Previously we used NanoSleep(1000), but this was unreliable. It could
result in waiting for >= 40ms instead of 1us. Since this was in a loop
it was especially bad if the GC was trying to suspend all the
threads when we were sleeping. This resulted in thread suspension
occasionally taking longer than a second.
Results on the provided picasso-sample app on Nexus 5:
Longest GC pause before: ~1.5s.
Longest GC pause after: <5ms.
Also added a warning if thread suspension takes longer than a
threshold (currently 5ms).
Bug: 16307460
External bug: https://code.google.com/p/android-developer-preview/issues/detail?id=367
Change-Id: I3c2a9636357e255f38634615101eff8ca84e632f
|
| bae182cbc6adc8796154162a87fc54ae804e0469 |
|
17-Dec-2013 |
Sebastien Hertz <shertz@google.com> |
Fix method tracing from command-line
Transitions current thread to the new kWaitingForMethodTracingStart thread
state when starting method tracing.
Ensures there is a current thread when method tracing is stopped due to runtime
shutdown. If the current thread has been detached, we now re-attach it.
Note: we only do this if method tracing has been activated from command-line.
Fixes instrumentation when forcing interpreter mode (-Xint) with method tracing
enabled.
Removes unused parameter from UnsafeLogFatalForThreadSuspendAllTimeout.
Bug: https://code.google.com/p/android/issues/detail?id=72094
Bug: 11683397
Change-Id: I70f000fb46ddd95d6ad51ea0a8eee77697a045e9
|
| 9db911eb76976fc1c73c8e563beb3620f1cadf99 |
|
04-Jun-2014 |
Mathieu Chartier <mathieuc@google.com> |
Print mutexes held in Thread::Dump.
Should help finding and fixing thread suspend timeout issues.
Bug: 15328839
Change-Id: I30a10529cec0716c7571a0318e9f66be54734fd8
|
| ffddfdf6fec0b9d98a692e27242eecb15af5ead2 |
|
03-Jun-2014 |
Tim Murray <timmurray@google.com> |
DO NOT MERGE
Merge ART from AOSP to lmp-preview-dev.
Change-Id: I0f578733a4b8756fd780d4a052ad69b746f687a9
|
| 9e36931cc79ca665908db9575126881d1cfdea5a |
|
21-May-2014 |
Chao-ying Fu <chao-ying.fu@intel.com> |
Move modify_ldt_lock into global lock order.
Mutex modify_ldt_lock was being removed during runtime shutdown while
daemons thread may still detach. Avoid this by placing in global lock
order.
This fixes cts dalvik vm-tests-tf that hang on some x86 devices.
By irogers: also, tidy global locks to agree with enum constants and
add extra verification that the global annotalysis order agrees with
the LockLevel order. Bumped the oat version and moved the locks as
LockLevel additions previously caused entrypoints to be moved. Make
unattached lock not handle the default mutex level case by moving the
allocated thread ids lock into the global order.
Change-Id: I9d03f19d44ea254accf0ceae8022563c77f7a02f
Signed-off-by: Chao-ying Fu <chao-ying.fu@intel.com>
|
| 5a9ba014516fcdeab617d1c006fdf2936865f178 |
|
19-May-2014 |
Ian Rogers <irogers@google.com> |
Build fix.
Change-Id: Ie5a3abb9a729a4a724e247b127c200a7be4d900d
|
| 4d466a8e4587422c989705dce3b2a19e7f0137f5 |
|
09-May-2014 |
Brian Carlstrom <bdc@google.com> |
Remove all LOG(DEBUG) which is not intended to be checked in
Bug: 14632493
Change-Id: Id9da8b87798af0a1b2bd2c178133e5f1ba47d43d
|
| 6f365cc033654a5a3b45eaa1379d4b5f156b0cee |
|
23-Apr-2014 |
Mathieu Chartier <mathieuc@google.com> |
Enable concurrent sweeping for non-concurrent GC.
Refactored the GarbageCollector to let all of the phases be run by
the collector's RunPhases virtual method. This lets the GC decide
which phases should be concurrent and reduces how much baked in GC
logic resides in GarbageCollector.
Enabled concurrent sweeping in the semi space and non concurrent
mark sweep GCs. Changed the semi-space collector to have a swap semi
spaces boolean which can be changed with a setter.
Fixed tests to pass with GSS collector, there was an error related to
the large object space limit.
Before (EvaluateAndApplyChanges):
GSS paused GC time 7.81s/7.81s, score: 3920
After (EvaluateAndApplyChanges):
GSS paused GC time 6.94s/7.71s, score: 3900
Benchmark score doesn't go up since the GC happens in the allocating
thread. There is a slight reduction in pause times experienced by
other threads (0.8s total).
Added options for pre sweeping GC heap verification and pre sweeping
rosalloc verification.
Bug: 14226004
Bug: 14250892
Bug: 14386356
Change-Id: Ib557d0590c1ed82a639d0f0281ba67cf8cae938c
|
| a2cee18799d7e61adb037a4219adb23fef7d022b |
|
17-Apr-2014 |
Christopher Ferris <cferris@google.com> |
Modify unwind to comply with stack parser tools.
Bug: 14081592
Change-Id: I6906b2575c74d64f1c3ba7602779b3a789de1c69
|
| bb43b433c845e253a11346f0f03706d58a44f23c |
|
14-Apr-2014 |
Sebastien Hertz <shertz@google.com> |
Clear unused roots from debugger
SingleStepControl and DebugInvokeReq hold roots which are only used during a
limited amount of time. Clear these roots once we finish using them.
Also cleans some code around single-step.
Change-Id: Ica2f65e04a4d60272199490cd40ad1803d3b9f02
|
| 7bf9f190cd33a7e2f8584299eb889e9df66e0323 |
|
04-Apr-2014 |
Mathieu Chartier <mathieuc@google.com> |
Improve invalid root dumping.
The invalid root dumping now attempts to print the root type.
Change-Id: Ie821296d569f34909ba6e2705f5c347cd2143a3a
|
| 2966e13d504a72d55c62bf864e183ec80703c699 |
|
02-Apr-2014 |
Ian Rogers <irogers@google.com> |
Avoid debug check violation in thread registration.
Bug: 13757855
Change-Id: Ied83a7fc43e2b9a89a8b99aa4f74069ef8a369b1
|
| dd7624d2b9e599d57762d12031b10b89defc9807 |
|
15-Mar-2014 |
Ian Rogers <irogers@google.com> |
Allow mixing of thread offsets between 32 and 64bit architectures.
Begin a more full implementation x86-64 REX prefixes.
Doesn't implement 64bit thread offset support for the JNI compiler.
Change-Id: If9af2f08a1833c21ddb4b4077f9b03add1a05147
|
| b373e091eac39b1a79c11f2dcbd610af01e9e8a9 |
|
21-Feb-2014 |
Dave Allison <dallison@google.com> |
Implicit null/suspend checks (oat version bump)
This adds the ability to use SEGV signals
to throw NullPointerException exceptions from Java code rather
than having the compiler generate explicit comparisons and
branches. It does this by using sigaction to trap SIGSEGV and when triggered
makes sure it's in compiled code and if so, sets the return
address to the entry point to throw the exception.
It also uses this signal mechanism to determine whether to check
for thread suspension. Instead of the compiler generating calls
to a function to check for threads being suspended, the compiler
will now load indirect via an address in the TLS area. To trigger
a suspend, the contents of this address are changed from something
valid to 0. A SIGSEGV will occur and the handler will check
for a valid instruction pattern before invoking the thread
suspension check code.
If a user program taps SIGSEGV it will prevent our signal handler
working. This will cause a failure in the runtime.
There are two signal handlers at present. You can control them
individually using the flags -implicit-checks: on the runtime
command line. This takes a string parameter, a comma
separated set of strings. Each can be one of:
none switch off
null null pointer checks
suspend suspend checks
all all checks
So to switch only suspend checks on, pass:
-implicit-checks:suspend
There is also -explicit-checks to provide the reverse once
we change the default.
For dalvikvm, pass --runtime-arg -implicit-checks:foo,bar
The default is -implicit-checks:none
There is also a property 'dalvik.vm.implicit_checks' whose value is the same
string as the command option. The default is 'none'. For example to switch on
null checks using the option:
setprop dalvik.vm.implicit_checks null
It only works for ARM right now.
Bumps OAT version number due to change to Thread offsets.
Bug: 13121132
Change-Id: If743849138162f3c7c44a523247e413785677370
|
| 39c3bfbd03d85c63cfbe69f17ce5800ccc7d6c13 |
|
29-Jan-2014 |
Dave Allison <dallison@google.com> |
Make use of profiling information for dex2oat
If the profile file exists, the compiler driver will read it
and store the data in an internal map. Then, when we want to work
out whether to compile a method or not, the map is consulted and if
the method shows up with a high enough percentage of use we compile it.
The profile file itself is created by installd and is writeable by the
app. The file is in /data/dalvik-cache/profiles and is named by
the package name.
This also modifies the profiler itself to:
1. Only count runnable threads (not suspended threads) in the profile
2. Use system properties to allow tuning of the profile parameters
3. Merge profiles from multiple processes using file locking.
Bug: 12877748
Change-Id: Iab2f3a327a2860db2a80d5724277d6c626227f2b
Conflicts:
compiler/dex/frontend.cc
compiler/dex/mir_analysis.cc
compiler/dex/verification_results.cc
compiler/driver/compiler_driver.cc
dex2oat/dex2oat.cc
runtime/class_linker.cc
runtime/runtime.cc
runtime/runtime.h
|
| 719d1a33f6569864f529e5a3fff59e7bca97aad0 |
|
06-Mar-2014 |
Ian Rogers <irogers@google.com> |
Enable annotalysis on clang ART builds.
Fix clang build errors aswell as restructure locking/mutex code for correct
thread safety analysis support.
Reorder make dependencies so that host builds build first as they should
provide better compilation errors than target.
Remove host's use of -fno-omit-frame-pointer as it has no value with correct
use of CFI, which we should have.
Change-Id: I72cea8da9a3757b1a0b3acb4081feccb7c6cef90
|
| 6dda898d47b3e8931e4404330e81b7110108e34f |
|
06-Mar-2014 |
Mathieu Chartier <mathieuc@google.com> |
Disable compaction for jni workarounds.
Compaction can't work when jni workarounds is enabled. Also some
other refactoring.
Change-Id: Ia7b0f2b39c79f5a0a5f50874d823b950ab02a0c3
|
| 6449c62e40ef3a9bb75f664f922555affb532ee4 |
|
11-Feb-2014 |
Brian Carlstrom <bdc@google.com> |
Create CompilerOptions
Package up most compiler related options in CompilerOptions. Details include:
- Includes compiler filter, method thresholds, SEA IR mode.
- Excludes those needed during Runtime::Init such as CompilerCallbacks and VerificationResults.
- Pass CompilerOptions to CompilerDriver.
- Remove CompilerOptions from Runtime.
- Add ability to pass options for app and image dex2oat to runtime via
-Xcompiler-option and -Ximage-compiler-option respectively.
Other
- Replace 2x CompilerCallbacks implementations with one.
- Factor out execv code for use by both image and oat generation.
- More OatFile error_msg reporting.
- DCHECK for SuspendAll found trying to run valgrind.
Change-Id: Iecb57da907be0c856d00c3cd634b5042a229e620
|
| 21e729c7859ce7909e85d135f6c9e32bbba38860 |
|
18-Feb-2014 |
Sebastien Hertz <shertz@google.com> |
Fix failing check during JDWP invoke.
Fixes a race where the JDWP thread expects the invoke thread can be suspended
(suspend_count != 0) before invoke thread actually updates its suspend count.
It happens after the invoke thread signals the JDWP thread to notify invoke is
completed but before incrementing its suspend count. In the meantime, the JDWP
thread wakes up and checks whether invoke thread has updated its supend count.
But there is no way to prevent the JDWP thread to do so before the invoke
thread updates its suspend count.
We now move the invoke completion code after the suspend count update. Then the
JDWP thread wakes up at the right time.
Bug: 11247837
Bug: 12578041
Change-Id: Ib6079c6e330671b34217838e26b1e758706d3da6
|
| 815873ecc312b1d231acce71e1a16f42cdaf09f2 |
|
14-Feb-2014 |
Mathieu Chartier <mathieuc@google.com> |
Change root visitor to use Object**.
Simplifies code and improves the performance of root visiting since
we usually don't need to check to see if the object moved.
Change-Id: Iba998f5a15ae1fa1b53ca5226dd2168a411196cf
|
| 83c8ee000d525017ead8753fce6bc1020249b96a |
|
28-Jan-2014 |
Mathieu Chartier <mathieuc@google.com> |
Add root types and thread id to root visiting.
Enables us to pass the root type and thread id to hprof.
Bug: 12680863
Change-Id: I6a0f1f9e3aa8f9b4033d695818ae7ca3460d67cb
|
| 0aded089f565008ba5908e395e5914ca4f91f2de |
|
07-Nov-2013 |
Dave Allison <dallison@google.com> |
Add sampling profiler
This adds a sampling profiler thread that runs every so often, gathering
profiling data and writing the results to a file in /data/data (specific to
app running). The intention is to use these files as input to the compiler
so that it can determine the best methods to compile.
Bug: 11539952
Change-Id: I0bfbb4146fb7966673c792f017ffac8107b6272d
|
| 5f51d4b80058236759fea1d932470a57f348c199 |
|
03-Dec-2013 |
Mathieu Chartier <mathieuc@google.com> |
Fix races in thread list Unregister.
First race:
We were releasing the thin_lock_id in Unregister before the thread
was not suspended. This could cause problems in
SuspendThreadByThreadId since there was a small window of time where
two threads could share the same thread id. This race caused an
occasional check failure in SuspendThreadByThreadId.
Second race:
We were setting the thin_lock_thread_id_ to 0 in Unregister before
waiting to not be suspended. This caused another race in
SuspendThreadByThreadId where we modified the thread suspend count,
busy waited, but didn't find the thread the next iteration. This
meant that we were returning null even though we had modified the
suspend count. This caused the suspend count to not get decremented
since the caller didn't know that the suspend count had been
increased. Removing the self->thin_lock_thread_id_ = 0 in
ThreadList::UnRegister fixes this race.
Added a bit of additional checks and logging to prevent these issues
from resurfacing, other misc cleanup.
Added thread names to threads in ThreadStress.
Bug: 11319866
Change-Id: I48e3a0700193b72079e450be1e924a2f88cf52e2
|
| 590fee9e8972f872301c2d16a575d579ee564bee |
|
13-Sep-2013 |
Mathieu Chartier <mathieuc@google.com> |
Compacting collector.
The compacting collector is currently similar to semispace. It works by
copying objects back and forth between two bump pointer spaces. There
are types of objects which are "non-movable" due to current runtime
limitations. These are Classes, Methods, and Fields.
Bump pointer spaces are a new type of continuous alloc space which have
no lock in the allocation code path. When you allocate from these it uses
atomic operations to increase an index. Traversing the objects in the bump
pointer space relies on Object::SizeOf matching the allocated size exactly.
Runtime changes:
JNI::GetArrayElements returns copies objects if you attempt to get the
backing data of a movable array. For GetArrayElementsCritical, we return
direct backing storage for any types of arrays, but temporarily disable
the GC until the critical region is completed.
Added a new runtime call called VisitObjects, this is used in place of
the old pattern which was flushing the allocation stack and walking
the bitmaps.
Changed image writer to be compaction safe and use object monitor word
for forwarding addresses.
Added a bunch of added SIRTs to ClassLinker, MethodLinker, etc..
TODO: Enable switching allocators, compacting on background, etc..
Bug: 8981901
Change-Id: I3c886fd322a6eef2b99388d19a765042ec26ab99
|
| d9c4fc94fa618617f94e1de9af5f034549100753 |
|
02-Oct-2013 |
Ian Rogers <irogers@google.com> |
Inflate contended lock word by suspending owner.
Bug 6961405.
Don't inflate monitors for Notify and NotifyAll.
Tidy lock word, handle recursive lock case alongside unlocked case and move
assembly out of line (except for ARM quick). Also handle null in out-of-line
assembly as the test is quick and the enter/exit code is already a safepoint.
To gain ownership of a monitor on behalf of another thread, monitor contenders
must not hold the monitor_lock_, so they wait on a condition variable.
Reduce size of per mutex contention log.
Be consistent in calling thin lock thread ids just thread ids.
Fix potential thread death races caused by the use of FindThreadByThreadId,
make it invariant that returned threads are either self or suspended now.
Code size reduction on ARM boot.oat 0.2%.
Old nexus 7 speedup 0.25%, new nexus 7 speedup 1.4%, nexus 10 speedup 2.24%,
nexus 4 speedup 2.09% on DeltaBlue.
Change-Id: Id52558b914f160d9c8578fdd7fc8199a9598576a
|
| 423d2a3dcbb260b020efb5da59f784c9f02accbf |
|
13-Sep-2013 |
Mathieu Chartier <mathieuc@google.com> |
Add support for changing roots through the root visitor callback.
Needed for copying collectors.
Change-Id: Icc4a342a57e0cfb79587edb02ef8c85e08808877
|
| 967a0adf8b93a23d2a8fef82e06bd913db94ac19 |
|
11-Sep-2013 |
Hiroshi Yamauchi <yamauchi@google.com> |
More allocation code optimizations.
- Inline Class::AllocObject() and Array::Alloc().
- Inline some short Mutex functions and add LIKELY/UNLIKELY to some
Mutex functions.
- This change improves the Ritz MemAllocTest by ~6% on Nexus 4 and
~10% on host.
Bug: 9986565
Change-Id: I1606c74ddb21676cbc1de1a40e9b076fc23eaea4
|
| 02e25119b15a6f619f17db99f5d05124a5807ff3 |
|
15-Aug-2013 |
Mathieu Chartier <mathieuc@google.com> |
Fix up TODO: c++0x, update cpplint.
Needed to update cpplint to handle const auto.
Fixed a few cpplint errors that were being missed before.
Replaced most of the TODO c++0x with ranged based loops. Loops which
do not have a descriptive container name have a concrete type instead
of auto.
Change-Id: Id7cc0f27030f56057c544e94277300b3f298c9c5
|
| 7934ac288acfb2552bb0b06ec1f61e5820d924a4 |
|
26-Jul-2013 |
Brian Carlstrom <bdc@google.com> |
Fix cpplint whitespace/comments issues
Change-Id: Iae286862c85fb8fd8901eae1204cd6d271d69496
|
| df62950e7a32031b82360c407d46a37b94188fbb |
|
18-Jul-2013 |
Brian Carlstrom <bdc@google.com> |
Fix cpplint whitespace/parens issues
Change-Id: Ifc678d59a8bed24ffddde5a0e543620b17b0aba9
|
| 7940e44f4517de5e2634a7e07d58d0fb26160513 |
|
12-Jul-2013 |
Brian Carlstrom <bdc@google.com> |
Create separate Android.mk for main build targets
The runtime, compiler, dex2oat, and oatdump now are in seperate trees
to prevent dependency creep. They can now be individually built
without rebuilding the rest of the art projects. dalvikvm and jdwpspy
were already this way. Builds in the art directory should behave as
before, building everything including tests.
Change-Id: Ic6b1151e5ed0f823c3dd301afd2b13eb2d8feb81
|