| b06c2974ea5e17f2fc0763d751a13b3fe660a4b9 |
|
25-May-2017 |
Nicolas Geoffray <ngeoffray@google.com> |
Flush data cache after updating the JIT table.
Otherwise, other cores could see an incomplete table when executing
the new code.
bug: 37949368
Test: test.py --jit
(cherry picked from commit 352b17af5263b27cad7a1dbf2351f9cb71d6af58)
Change-Id: Ib1e4d2386477cb537b5921bb395e4bcd0ff760b7
|
| 60889dcb3fe8e718debaddd6d2fb05a880ed87fa |
|
18-May-2017 |
Calin Juravle <calin@google.com> |
Record inline caches only for the strictly hot methods
Warm methods might have incomplete inline caches which could cause
unnecessary deoptimizations.
Test: m test-art-host
Bug: 38426301
Bug: 38412648
(cherry picked from commit a39fd980000d10ee0b3a49555cc8bcbbfacde943)
Change-Id: I12e2c068663e8bb948dcd7ba89d0c1d4c9118153
|
| d0384a7faffcf1dc52a74c1b2a30ab3f36edfb89 |
|
16-May-2017 |
Andreas Gampe <agampe@google.com> |
ART: Fix iterator invalidation
Erasing an item of a map invalidates the iterator.
(cherry picked from commit 39e673811d73ac09a822d9ee0082bba5d8e8cbad)
Bug: 38324641
Test: m test-art-host
Change-Id: I9f7499eb4a4b49e4fcd6c99d170d9b681adfd9b0
|
| 70c94bf292bfb3113b757778fd6760fad81f5f79 |
|
10-May-2017 |
Nicolas Geoffray <ngeoffray@google.com> |
Use a sentinel when clearing JIT table entries.
This can help diagnosing crashes we see.
bug: 38128052
bug: 37949368
Test: test.py --jit
(cherry picked from commit 6ca115b5fa185a2cc162a453f425efd87d8f38bd)
Change-Id: Ib048f72cef7acfdb15f515c63a2ff0ac8e3744ce
|
| 79efadfdd861584f1c47654ade975eae6c43c360 |
|
08-May-2017 |
Nicolas Geoffray <ngeoffray@google.com> |
Add runtime reasons for deopt.
Currently to help investigate. Also:
1) Log when deoptimization happens (which method and what reason)
2) Trace when deoptimization happens (to make it visible in systrace)
bug:37655083
Test: test-art-host test-art-target
(cherry picked from commit 4e92c3ce7ef354620a785553bbada554fca83a67)
Change-Id: I992398a1038ab61ea0e5106af6b6ad0a3305312e
|
| f6d4668c42933e2f85ddbc94e276c49db4e2b1dd |
|
28-Feb-2017 |
Nicolas Geoffray <ngeoffray@google.com> |
Improvements in the Inliner.
- Change from a depth limit to a total number of HInstructions
inlined limit. Remove the dex2oat depth limit argument.
- Add more stats to diagnose reasons for not inlining.
- Clean up logging to easily parse output.
Individual Ritz benchmarks improve from 3 to 10%.
No change in other heuristics. There was already an instruction budget.
Note that the instruction budget is rarely hit in the "apps" I've tried
with.
Compile-times improve from 5 to 15%.
Code size go from 4% increase (Gms) to 1% decrease (Docs).
bug:35724239
test: test-art-host test-art-target
Change-Id: I5a35c4bd826cf21fead77859709553c5b57608d6
|
| cbcedbf9382bc773713cd3552ed96f417bf1daeb |
|
13-Mar-2017 |
Mathieu Chartier <mathieuc@google.com> |
Add method info to oat files
The method info data is stored separately from the code info to
reduce oat size by improving deduplication of stack maps.
To reduce code size, this moves the invoke info and inline info
method indices to this table.
Oat size for a large app (arm64): 77746816 -> 74023552 (-4.8%)
Average oat size reduction for golem (arm64): 2%
Repurposed unused SrcMapElem deduping to be for MethodInfo.
TODO: Delete SrcMapElem in a follow up CL.
Bug: 36124906
Test: clean-oat-host && test-art-host-run-test
Change-Id: I2241362e728389030b959f42161ce817cf6e2009
|
| 13439f0c4769a0768cf1bbaa7b3f2f9ee5a317c5 |
|
21-Feb-2017 |
Calin Juravle <calin@google.com> |
Use offline inline caches during AOT compilation
Also:
- extend the testing script to understand profile when --profile is
passed
- filter inline cache types which are not loaded by the caller class
loader
Test: m test-art-host-run-test-638-checker-inline-caches
Bug: 32434870
Change-Id: Ifcc27b3cebc79b84617412aaae64a73324151b55
|
| 589e71eb71477603c88b14118b2d3a33508c3c79 |
|
04-Mar-2017 |
Calin Juravle <calin@google.com> |
Encode inline caches with missing types in the profile
Not all runtime types can be encoded in the profile. For example if the
receiver type is in a dex file which is not tracked for profiling its
type cannot be encoded.
Previously we would just skip over these types but that can lead to
encode a polymorphic inline cache when in fact it should be megamorphic.
With this CL, inline caches for which types are missing are marked in
the profile with a special bit, kIsMissingTypesEncoding.
Also, extend profman to understand text lines which specify an inline
cache with missing types.
Test: test-art-host
Bug: 35927981
Bug: 32434870
Change-Id: I34528a39c227f3133771fd4454701c1ddc234f40
|
| 3b1a7f40ad69ab5332d13ed0abf0a3a1fa7f2a9e |
|
22-Feb-2017 |
Nicolas Geoffray <ngeoffray@google.com> |
Update entrypoint directly when doing JIT code cache GC.
Otherwise, the Instrumentation::UpdateMethodsCode might access
the declaring class while the GC is concurrently processing
references.
bug:35598395
test: run-libcore-tests.sh --mode=host --variant=X64
Change-Id: I0f059b1823e81672d43cf1f7eed64482f8927232
|
| e21806f053c1a001738b121e7f7807f6cb79b97a |
|
22-Feb-2017 |
Calin Juravle <calin@google.com> |
Add more comments in GetProfiledMethods
Explain why we are looking in the method dex file for array types.
Test: m test-art-host
Change-Id: Ib94475f79844968591f72f46a2c714f9971778ba
|
| 4ca70a3102c10e2e77e97fc3b9180f00304ed29b |
|
22-Feb-2017 |
Calin Juravle <calin@google.com> |
Fix handling of array types when extracting the offline inline caches
Test: m test-art-host
regression test is being worked out
Bug: 35621933
Change-Id: If762db76cc0493db8c8babab62035e915b660da5
|
| 940eb0c00ef531dd9a0a68dbd61e377832e81eb4 |
|
31-Jan-2017 |
Calin Juravle <calin@google.com> |
Add inline caches to offline profiles
Add support for inline caches in profiles:
- extract inline caches from the jit cache when the profile saver
queries the hot methods
- bump profile version to support the new data
- add new tests
- inline caches are only supported for same-apk calls (including
multidex)
Test: m art-test-host-gtest-profile_compilation_info_test
Bug: 32434870
Change-Id: I38b4ca0a54568d2224765ff76023baef1b8fd1a2
|
| eee0bd448ec057d3f224895ddb868786758eeb5b |
|
14-Feb-2017 |
Alex Light <allight@google.com> |
Revert^6 "Make class redefinition work with native methods on stack."
We were incorrectly trying to obtain the profiling information of a
native method.
This reverts commit 02b2349b1cd2a78e86b1a7542f8330e6c3aaeb35.
Reason for revert: Fixed test failure with jit configurations
Test: ART_TEST_JIT=true mma -j40 test-art-host
Change-Id: Ic7112104aec64e597d2df80db5dc2a505d2cc2dd
|
| b93ce1853e1a28bbd6ca992c5b646aa2d2699892 |
|
15-Dec-2016 |
Kevin Brodsky <kevin.brodsky@linaro.org> |
Clarify comment on JIT cache flush issue
The original fix (#304947) is needed, but the comment was suggesting a
hardware cause and a link with A53 errata. In all likelihood neither
of these is true, and the issue is actually caused by a kernel issue
affecting many Qualcomm devices (like those using the msm tree).
Change-Id: Ie750e6fd9eb7130982d184c22917b5c3ea4beed0
|
| 1ebe4fe1a6cb0835197ee3a848d554b79ca9d367 |
|
30-Jan-2017 |
Alex Light <allight@google.com> |
Fix creating self-recursive obsolete methods.
We were using recursive loading of the current art method which was
making us miss obsolete methods in some cases.
We could also end up checking the wrong method when walking the stack.
We also add tests for recursive obsolete methods in general.
Bug: 34815470
Test: mma -j40 test-art-host
Test: ART_TEST_JIT=true mma -j40 test-art-host
Change-Id: I522fd4cac4e3f9d35d03b128bad6d6971cfe6c4a
|
| 433b79a9130df5f2375b9d3d7518c1f3fb84a467 |
|
30-Jan-2017 |
Nicolas Geoffray <ngeoffray@google.com> |
Revert "Revert "Make --debuggable rely on JIT code.""
Fixed by https://android-review.googlesource.com/#/c/330165/
This reverts commit 81356645157af44152c7b7db383596b5cf3479b5.
Change-Id: Ifb74e1cc90ab6dea621f7f54a00b540d6ccd0cf6
|
| 81356645157af44152c7b7db383596b5cf3479b5 |
|
30-Jan-2017 |
Vladimir Marko <vmarko@google.com> |
Revert "Make --debuggable rely on JIT code."
Breaks 137-cfi with ART_TEST_RUN_TEST_DEBUGGABLE=true
This reverts commit a0619e25aacf8b8074132f4951f75fdbcfd42925.
bug: 28769520
Change-Id: Ifd7b8fc7c9b72c0a523fd57c4b1b80edd3547caa
|
| a0619e25aacf8b8074132f4951f75fdbcfd42925 |
|
20-Dec-2016 |
Nicolas Geoffray <ngeoffray@google.com> |
Make --debuggable rely on JIT code.
Removes -Xfully-deoptable in the process, which was added as a
temporary workaround until this CL.
Partial revert of https://android-review.googlesource.com/#/c/302232/
Makes things consistent with existing infrastructure:
- Parse the --debuggable from the compiler options, just like
--compiler-filter.
- Add DEBUG_JAVA_DEBUGGABLE, passed by the zygote, for debuggable apps.
- Java debuggable now solely relies on JIT for simplicity.
- Debugging under userdebug for non-java-debuggable apps is still
best effort.
Test: test-art-host, jdwp
bug: 28769520
Change-Id: Id0593aacd85b9780da97f20914a50943957c858f
|
| 3f3d4d63199e57cb1dccad62d2d3ef69fc170c06 |
|
17-Nov-2016 |
Artem Udovichenko <artem.u@samsung.com> |
Cache flush/invalidate needs RWX permission
When generating JIT code, perform cache maintenance operations before
removing page write permissions. Errata on some cores require data
flush operations to be followed by data invalidate operations
requiring write permission.
Test: ART_TEST_JIT=true test-art-target on an arm64 device.
bug: 27265969
(cherry picked from commit b18a669259aa6ba08d9ca01b5b32c1aa0417138b)
Change-Id: I53080c664f9e7cfebe25f87cf6a45cd6eb33b281
|
| dba61481035b7944173181ec9ee02aea41dd0e29 |
|
21-Dec-2016 |
Alex Light <allight@google.com> |
Revert "Revert "Revert "Revert "Basic obsolete methods support""""
A GetDeclaringClass()->GetDexCache() got inserted during the
merge/review process meaning that we would try to access incorrect
dex-cache in obsolete methods in some situations.
Also when using tracing we would loop forever (or at least until an
OOM error) in test 916 due to tracing forcing InterpretOnly mode
meaning methods would never be jitted.
Bug: 32369913
Bug: 33630159
Test: ART_TEST_TRACE=true \
ART_TEST_JIT=true \
ART_TEST_INTERPRETER=true mma -j40 test-art-host
This reverts commit f6abcda293b115a9d7d8a26376ea2dcf2d1dc510.
Change-Id: I0773bfcba52e3cd51a83be815c6a50c189558f48
|
| f6abcda293b115a9d7d8a26376ea2dcf2d1dc510 |
|
21-Dec-2016 |
Nicolas Geoffray <ngeoffray@google.com> |
Revert "Revert "Revert "Basic obsolete methods support"""
Fails in tracing mode
Bug: 32369913
Bug: 33630159
This reverts commit ce77fc0e7f60a15354bb20c356537cbf8b53b722.
Change-Id: I1bdcf6ad467f2e31f9c5d0c3c987b90a4f5efc69
|
| ce77fc0e7f60a15354bb20c356537cbf8b53b722 |
|
15-Dec-2016 |
Alex Light <allight@google.com> |
Revert "Revert "Basic obsolete methods support""
This reverts commit b81a9840b44480bfeacd74b8d9f51e06f295411d.
There were 2 issues with the original CL's test 916-obsolete-jit that
caused it to sporadically fail.
First, when checking if we had jitted the function under test in
916-obsolete-jit we failed to check to see if the function 'doCall',
which is used to work-around bugs in our deoptimization & compilation
systems, had also been jitted.
In the case where the 'sayHi' function had been jitted but the
'doCall' function had not we would (correctly) fail to redefine the
'Transform' class since we would not be able to deoptimize the 'sayHi'
function since it is under a quick_to_interpreter_bridge (runtime)
frame.
Secondly, the function Main.isInterpretedFunction was incorrect and
would always return false, regardless of the actual state of the
function. This would cause the test to fail as the
quick_to_interpreter_bridge frame prevented deoptimization of the
obsoleted function. Usually the warm-up period was enough to make sure
the methods were jitted anyway but this was not guaranteed.
Both of these problems become more likely to occur on systems with
more cpu contention such as the buildbots.
Test: stress --cpu 60 &; while ./test/run-test --host --jit 916; do ; done
Reason for revert: Fixed bug in test that was causing failures.
Original Tests:
Test: ./test/run-test --host 914
Test: ./test/run-test --host 915
Test: ./test/run-test --host 916
Test: mma -j40 test-art-host
Test: ART_TEST_JIT=true \
ART_TEST_INTERPRETER=true mma -j40 test-art-host
Test: ./art/tools/run-jdwp-tests.sh --mode=host --variant=X64
Test: ./art/tools/run-jdwp-tests.sh --mode=host --variant=X64 --no-jit
Bug: 32369913
Bug: 33630159
Change-Id: If1a92e47b90965a7dc21c5826185debe62bd1554
|
| ed015acdd0fbb1b837b3f74e8c5f624956189691 |
|
15-Dec-2016 |
Nicolas Geoffray <ngeoffray@google.com> |
Flush JIT data cache when committing code.
Otherwise, executing compiled code may see old data from
the data cache.
Test: ART_TEST_JIT=true test-art-target on arm/arm64
Test: run-libcore-tests.sh (especially jsr166) on arm/arm64
Change-Id: Id037c68897aa0e9ccacd4c5121f3743fb722c6f3
|
| b81a9840b44480bfeacd74b8d9f51e06f295411d |
|
15-Dec-2016 |
Alex Light <allight@google.com> |
Revert "Basic obsolete methods support"
This reverts commit d8936da27b792d1ca02e59c92456a1a53c7b9905.
Reason for revert: Some sort of race in JIT
Change-Id: Ibb4e520bb0721d6d7aa2c841a52eb5baff07740e
|
| d8936da27b792d1ca02e59c92456a1a53c7b9905 |
|
29-Nov-2016 |
Alex Light <allight@google.com> |
Basic obsolete methods support
Add support for executing obsolete methods following redefinitions.
This support includes methods that have been jitted. This does not add
any additional validity checks to our redefinition functions.
Note using work-arounds to ensure 916 works pending some fixes to the
compiler, deoptimizer, or both.
Test: ./test/run-test --host 914
Test: ./test/run-test --host 915
Test: ./test/run-test --host 916
Test: mma -j40 test-art-host
Test: ART_TEST_JIT=true \
ART_TEST_INTERPRETER=true mma -j40 test-art-host
Test: ./art/tools/run-jdwp-tests.sh --mode=host --variant=X64
Test: ./art/tools/run-jdwp-tests.sh --mode=host --variant=X64 --no-jit
Bug: 32369913
Bug: 33630159
Change-Id: I78ef95f484146f1fb93c37fc50f56575bdab2432
|
| 22384aeab988df7fa5ccdc48a668589c5f602c39 |
|
12-Dec-2016 |
Nicolas Geoffray <ngeoffray@google.com> |
Revert "Revert "Add kJitTableAddress for HLoadClass.""
This reverts commit d2d5262c8370309e1f2a009f00aafc24f1cf00a0.
Change-Id: I6149d5c7d5df0b0fc5cb646a802a2eea8d01ac08
|
| d2d5262c8370309e1f2a009f00aafc24f1cf00a0 |
|
12-Dec-2016 |
Nicolas Geoffray <ngeoffray@google.com> |
Revert "Add kJitTableAddress for HLoadClass."
One test failure after merge.
This reverts commit 5b12f7973636bfea29da3956a9baa7a6bbe2b666.
Change-Id: I120c49e53274471fc1c82a10d52e99c83f5f85cc
|
| 5b12f7973636bfea29da3956a9baa7a6bbe2b666 |
|
09-Dec-2016 |
Nicolas Geoffray <ngeoffray@google.com> |
Add kJitTableAddress for HLoadClass.
This new kind loads classes from the root table associated with
JIT compiled code.
Also remove kDexCacheAddress, which is replaced by kJitTableAddress.
test: ART_TEST_JIT=true test-art-host-jit test-art-target-jit
Change-Id: Ia23029688d1a60c178bf2ffa7463927c5d5de4d0
|
| f4b94427072a42ee6814aa6e1abc04ced5235355 |
|
05-Dec-2016 |
Nicolas Geoffray <ngeoffray@google.com> |
Set the root table size when allocating.
In order to deallocate it at any point.
test: ART_TEST_JIT=true m test-art-host-run-test-jit
Change-Id: I36574fd1a34c9d67800a164fa801db7a17c827bf
|
| 063fc772b5b8aed7d769cd7cccb6ddc7619326ee |
|
02-Aug-2016 |
Mingyao Yang <mingyao@google.com> |
Class Hierarchy Analysis (CHA)
The class linker now tracks whether a method has a single implementation
and if so, the JIT compiler will try to devirtualize a virtual call for
the method into a direct call. If the single-implementation assumption
is violated due to additional class linking, compiled code that makes the
assumption is invalidated. Deoptimization is triggered for compiled code
live on stack. Instead of patching return pc's on stack, a CHA guard is
added which checks a hidden should_deoptimize flag for deoptimization.
This approach limits the number of deoptimization points.
This CL does not devirtualize abstract/interface method invocation.
Slides on CHA:
https://docs.google.com/a/google.com/presentation/d/1Ax6cabP1vM44aLOaJU3B26n5fTE9w5YU-1CRevIDsBc/edit?usp=sharing
Change-Id: I18bf716a601b6413b46312e925a6ad9e4008efa4
Test: ART_TEST_JIT=true m test-art-host/target-run-test test-art-host-gtest
|
| b84defb1b09803d5a47035e8010ac43984da1d4c |
|
30-Nov-2016 |
Nicolas Geoffray <ngeoffray@google.com> |
Do not call IsMarked on null.
The concurrent collector checks it.
Test: ART_TEST_JIT=true test-art-host-run-test-jit
Change-Id: I02c2462ce1d8f3092f0b2f52f8b6dede8ff4a339
|
| e51ca8bfa8a193b64901ad214842f213adca92eb |
|
22-Nov-2016 |
Nicolas Geoffray <ngeoffray@google.com> |
Move inline caches GC handling in JitCodeCache.
Make the classes they hold weak references and visit
them during SweepJitRoots.
This fixes the newly introduced deadlock:
Thread1:
1) Lock JitCodeCache lock to create Profiling info for
ArtMethod m.
2) m is a copied method, we need to track the actual holder,
needing to decode a weak reference.
3) Weak references are not accessible due to GC.
GC Thread:
- Disallow weak reference access.
- Wait for checkpoint.
Thread2:
- Try to lock JitCodeCache lock
- Deadlock, as Thread1 owns the JitCodeCache lock.
Test: test-art-host
bug: 31289185
bug: 33198826
Change-Id: I7ee17631015450ace8d2a0264415a81c5a902bb8
|
| b18a669259aa6ba08d9ca01b5b32c1aa0417138b |
|
17-Nov-2016 |
Artem Udovichenko <artem.u@samsung.com> |
Cache flush/invalidate needs RWX permission
When generating JIT code, perform cache maintenance operations before
removing page write permissions. Errata on some cores require data
flush operations to be followed by data invalidate operations
requiring write permission.
Test: ART_TEST_JIT=true test-art-target on an arm64 device.
bug: 27265969
Change-Id: I70cbb0dc2ea37e42f2011d23333e386ae3f639c7
|
| 7a704bea1b1ff1e7deb903513997e0f35040287b |
|
22-Nov-2016 |
Mathieu Chartier <mathieuc@google.com> |
Always fill in root table length
Otherwise, there may be DCHECK failure in JitCodeCache::ClearData.
Fixes gcstress JIT tests.
Test: test/run-test --host --relocate --prebuild --64 --jit --gcstress --dev 021-string2
Change-Id: I757ccd1e5a015c5e686254417a2a3f68b650da1c
|
| f46501c6ad11025843682267c10f221323a206b1 |
|
22-Nov-2016 |
Nicolas Geoffray <ngeoffray@google.com> |
Clear the right data pointer.
Internally, stack_map_data is an offset of the roots_data allocation.
Pass both to the ClearData method, which will use the right pointer
to deallocate.
Test: test-art-host in debug mode
Change-Id: Ibfe38e7f1853870076048427f2efe7121c11c136
|
| 132d8363bf8cb043d910836672192ec8c36649b6 |
|
16-Nov-2016 |
Nicolas Geoffray <ngeoffray@google.com> |
Revert "Revert "Revert "Revert "JIT root tables.""""
Test: 626-set-resolved-string, test-art-host, test-art-target
Test: run-libcore-tests.sh
Test: phone boots and runs
This reverts commit 3395fbc20bcd20948bec8958db91b304c17cacd8.
Change-Id: I104b73d093e3eb6a271d564cfdb9ab09c1c8cf24
|
| 3395fbc20bcd20948bec8958db91b304c17cacd8 |
|
14-Nov-2016 |
Nicolas Geoffray <ngeoffray@google.com> |
Revert "Revert "Revert "JIT root tables."""
libcore failures:
dalvikvm32 F 11-14 03:04:06 14870 14870 jit_code_cache.cc:310] Check failed: new_string != nullptr
This reverts commit 75afcdd3503a8a8518e5b23d21b6e73306ce39ce.
Change-Id: I5a6b6b48aa79a763d1ff1ba4d85d63811254787d
|
| 75afcdd3503a8a8518e5b23d21b6e73306ce39ce |
|
10-Nov-2016 |
Nicolas Geoffray <ngeoffray@google.com> |
Revert "Revert "JIT root tables.""
Also contains Revert "Support kJitTableAddress in x86/arm/arm64."
This reverts commit 4acd03638fcdb4e5d1666f8eec7eb3bf6d6be035.
This reverts commit 997d1217830c0a18b70faeabd53c04700a87d7d9.
Test: ART_USE_READ_BARRIER=true/false test-art-host test-art-target
Change-Id: I77cb1e9bf8f1b4c58b72d3cf5ca31ced2aaa1ea3
|
| 4acd03638fcdb4e5d1666f8eec7eb3bf6d6be035 |
|
09-Nov-2016 |
Nicolas Geoffray <ngeoffray@google.com> |
Revert "JIT root tables."
May be the offender for jit-gcstress failure of 902.
This reverts commit ac3ebc3150760425ed00abd56da48f9a6e0666bc.
Change-Id: I9ea6c9236fd1729fed7d1868dd8a111172932308
|
| ac3ebc3150760425ed00abd56da48f9a6e0666bc |
|
05-Oct-2016 |
Nicolas Geoffray <ngeoffray@google.com> |
JIT root tables.
Implement root tables for the JIT. Each JIT compiled method
gets a table allocated before the stack maps.
The table gets visited through Runtime::SweepSystemWeaks.
Implement String roots for x86_64 as an example.
Test: test-art-host test-art-target
Change-Id: Id3d5bc67479e08b52dd4b253e970201203a0f0d2
|
| 709b070044354d9f47641f273edacaeeb0240ab7 |
|
13-Oct-2016 |
David Sehr <sehr@google.com> |
Remove mirror:: and ArtMethod deps in utils.{h,cc}
The latest chapter in the ongoing saga of attempting to dump a DEX
file without having to start a whole runtime instance. This episode
finds us removing references to ArtMethod/ArtField/mirror.
One aspect of this change that I would like to call out specfically
is that the utils versions of the "Pretty*" functions all were written
to accept nullptr as an argument. I have split these functions up as
follows:
1) an instance method, such as PrettyClass that obviously requires
this != nullptr.
2) a static method, that behaves the same way as the util method, but
calls the instance method if p != nullptr.
This requires using a full class qualifier for the static methods,
which isn't exactly beautiful. I have tried to remove as many cases
as possible where it was clear p != nullptr.
Bug: 22322814
Test: test-art-host
Change-Id: I21adee3614aa697aa580cd1b86b72d9206e1cb24
|
| 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
|
| bdf7f1c3ab65ccb70f62db5ab31dba060632d458 |
|
31-Aug-2016 |
Andreas Gampe <agampe@google.com> |
ART: SHARED_REQUIRES to REQUIRES_SHARED
This coincides with the actual attribute name and upstream usage.
Preparation for deferring to libbase.
Test: m
Test: m test-art-host
Change-Id: Ia8986b5dfd926ba772bf00b0a35eaf83596d8518
|
| 99c959f3868512bb95432cf02a1f0ad3971698bf |
|
30-Jul-2016 |
Nicolas Geoffray <ngeoffray@google.com> |
Put the deletion of profiling info under a GC critical section.
Otherwise the GC could see dangling pointers.
bug:30033802
(cherry picked from commit cf48fa030780c3185f225d558b704c396f7713cc)
Change-Id: I2c43e973878f50dc147aa0af81551ecc942a790d
|
| 542451cc546779f5c67840e105c51205a1b0a8fd |
|
26-Jul-2016 |
Andreas Gampe <agampe@google.com> |
ART: Convert pointer size to enum
Move away from size_t to dedicated enum (class).
Bug: 30373134
Bug: 30419309
Test: m test-art-host
Change-Id: Id453c330f1065012e7d4f9fc24ac477cc9bb9269
|
| cf48fa030780c3185f225d558b704c396f7713cc |
|
30-Jul-2016 |
Nicolas Geoffray <ngeoffray@google.com> |
Put the deletion of profiling info under a GC critical section.
Otherwise the GC could see dangling pointers.
bug:30033802
Change-Id: I839c4f6e537aba98769e99e4d75ff869c9ff13eb
|
| 9962962eb26f906e8f92cde1d6854dadbbc5e8ad |
|
19-Apr-2016 |
Calin Juravle <calin@google.com> |
ProfileSaver: query profiling_infos instead of the code_map
This is the first step in being able to record profiles without jit
being active.
Bug: 27916886
(cherry picked from commit e2d066d0337b7c81d47e4806e6025b70d83fcd56)
Change-Id: Ie36c4b4a11d3b6e5de959121a1acea22a9e25121
|
| c8a4f2c632afd44fc558aa37c98b9cf042b4bb5b |
|
18-Apr-2016 |
Nicolas Geoffray <ngeoffray@google.com> |
Fix interaction between instrumentation and jit lock.
The jit lock needs to have higher priority than the
deoptimized methods lock.
bug:28236735
(cherry picked from commit 480d5108fa62f28cbc2e7072610f974953ff73a8)
Change-Id: I6e5f8597c4ab88993f6a32a204383b766ff03e33
|
| e2d066d0337b7c81d47e4806e6025b70d83fcd56 |
|
19-Apr-2016 |
Calin Juravle <calin@google.com> |
ProfileSaver: query profiling_infos instead of the code_map
This is the first step in being able to record profiles without jit
being active.
Bug: 27916886
Change-Id: I2fcbd03560e109e6866ad6d82c79b902a889e620
|
| 480d5108fa62f28cbc2e7072610f974953ff73a8 |
|
18-Apr-2016 |
Nicolas Geoffray <ngeoffray@google.com> |
Fix interaction between instrumentation and jit lock.
The jit lock needs to have higher priority than the
deoptimized methods lock.
bug:28236735
Change-Id: I82862b8bfc82a5641156290926c04c80b1371534
|
| f3f9cf6b65c4bcf9ea44253188d8d910b7cf7e64 |
|
14-Apr-2016 |
Nicolas Geoffray <ngeoffray@google.com> |
Add weight to compiled/interpreter transitions.
Also:
- Cleanup logging.
- Check ArtMethod status before adding compilation requests.
- Don't request osr compilation if we know AddSamples does not come
from a back edge.
Bug: 27865109
(cherry picked from commit 71cd50fb67fa48667b0ab59aa436a582c04ba43d)
Change-Id: Icbe89fe6cc495b113616391a8f257758d34b4b60
|
| 71cd50fb67fa48667b0ab59aa436a582c04ba43d |
|
14-Apr-2016 |
Nicolas Geoffray <ngeoffray@google.com> |
Add weight to compiled/interpreter transitions.
Also:
- Cleanup logging.
- Check ArtMethod status before adding compilation requests.
- Don't request osr compilation if we know AddSamples does not come
from a back edge.
Bug: 27865109
Change-Id: I84512f7d957b61ce2458360ed430adb151830278
|
| 454b3b6774fe07765273b917ccc97da2eba776ac |
|
07-Apr-2016 |
buzbee <buzbee@google.com> |
ART: Fix race in on-stack replacement
The expected sequence of events for on-stack replacement is:
1. Method goes warm, triggering enhanced profiling
2. Method goes hot, triggering method compilation
3. Method goes really hot, triggering an osr method compilation.
4. Interpreter polls for the existence of an osr entry point,
and transitons to compiled code if found.
We have a race problem if #2 and #3 happen closely together, and
the osr method compilation begins before the regular method
compilation. In that case, the jit sees that the method is
already being compiled (the osr method - but it does not
distinguish the two), and discards the normal compilation request.
So, the osr version is compiled and the normal version is discarded.
In #4, the MaybeDoOnStackReplacement() check assumes that a normal
version of the compiled method must exist before doing an on-stack
replacement, so it keeps returning false.
This is why we were seeing sporadic timeout failures of
570-checker-osr when the mterp fast branch profiling was
introduced. The branch profiling performance enhancements
greatly reduced the time between #2 and #3, increasing the liklihood
of losing the race. Further, the new code clamped hotness to avoid
wrap-around. The race existed (and likely occurred) in the previous
version, but because hotness counters were allowed to overflow and
wrap around you'd eventually hit the threshold a second time and
try again - masking the problem.
Tip 'o the hat to Serguei Katkov for identifying the problem.
A possible solution (taken in this CL) is to differentiate osr
compilations from normal method compilations.
Bug: 27939339
Change-Id: I71044516b35dc69de9fc2d2a445e33809ac650ed
|
| 9d07e3d128ccfa0ef7670feadd424a825e447d1d |
|
31-Mar-2016 |
Vladimir Marko <vmarko@google.com> |
Clean up OatQuickMethodHeader after Quick removal.
This reduces the size of the pre-header by 8 bytes, reducing
oat file size and mmapped .text section size. The memory
needed to store a CompiledMethod by dex2oat is also reduced,
for 32-bit dex2oat by 8B and for 64-bit dex2oat by 16B. The
aosp_flounder-userdebug 32-bit and 64-bit boot.oat are each
about 1.1MiB smaller.
Disable the broken StubTest.IMT, b/27991555 .
Change-Id: I05fe45c28c8ffb7a0fa8b1117b969786748b1039
|
| b9a639df2f21dd67b9f4379f31b7ccc5dd48ebd6 |
|
22-Mar-2016 |
Nicolas Geoffray <ngeoffray@google.com> |
Fix braino in JitCodeCache.
A OatQuickMethodHeader can never have its entry point be the
interpreter entrypoint.
Clear hotness counter of methods in more relevant places.
Change-Id: I0563a57eace2e0e4fe2abc76f5e293ddf47b25e3
|
| 6300fd77033817664aee3056c4ff5fab85698d2b |
|
18-Mar-2016 |
Nicolas Geoffray <ngeoffray@google.com> |
Code cleanup between debugger and jit.
This is to allow secondary dex files to be fully compiled.
- No need to do full deopt anymore for breakpoints: code in boot.oat
is deoptimized as soon as the debugger attaches (and we decided that
existing frames could not be debugged), and application being
debugged is compiled debuggable.
- jit should only call UpdateMethodsCode if exit stubs are not
installed due to lock violation otherwise (the lock level of the
jit code cache is lower than the deoptimization lock).
This part needs an overall cleanup beyond the scope of this change.
Change-Id: I38d85dcb270db746c1d6b0ceb7893a1aad8c9655
|
| 933330a192d43d7c7304120999907857a89c2f96 |
|
16-Mar-2016 |
Nicolas Geoffray <ngeoffray@google.com> |
Add more statistics for memory use of the JIT.
Collect data for stack maps, profiling info, and compiled code.
bug:27520994
Change-Id: Ic87361230c96ce0090027a37d750e948d806c597
|
| 07e3ca9a46801990d662c48d0fddedc63c4c053c |
|
11-Mar-2016 |
Nicolas Geoffray <ngeoffray@google.com> |
Fix thread race when fetching the ProfilingInfo object.
Problem is:
1) Compiler fetches the ProfilingInfo of A, it's null.
2) Mutator creates the ProfilingInfo.
3) Compiler notifies it's not using A anymore, calls
ProfilingInfo::DecrementInlineUse -> Crash as we expected
ProfilingInfo::IncrementUse to be called before.
Also update some namings to better reflect what is going on.
Change-Id: I55ea4c5d81988131467095e18a0d13a8be9d0ef7
|
| b6e20ae17d0881a66c22532e4152ce6779454a92 |
|
07-Mar-2016 |
Nicolas Geoffray <ngeoffray@google.com> |
Clear inline caches at each full GC.
This fixes occasional failures of 141-class-unload.
Also fix a bug where clearing inline caches also cleared the dex
pc associated with it.
bug:26846185
bug:23128949
Change-Id: I77bf1dee229d7764c3cc21440829c7fba7b37001
|
| 33fbf37feb959293b10d0b6e26d5e8ee7dff786c |
|
07-Mar-2016 |
Mathieu Chartier <mathieuc@google.com> |
Add systrace logging to ScopedCodeCacheWrite
Shows if we are taking a long time to mprotect.
Bug: 27502458
(cherry picked from commit 462cd2cfa8c5fbb6548419eb8f27b448a12aca11)
Change-Id: Ic0f24641f5856b13f3b949dcd53dc01ec952d171
|
| 462cd2cfa8c5fbb6548419eb8f27b448a12aca11 |
|
07-Mar-2016 |
Mathieu Chartier <mathieuc@google.com> |
Add systrace logging to ScopedCodeCacheWrite
Shows if we are taking a long time to mprotect.
Bug: 27502458
Change-Id: Ib83aed3d15b252baaf8f0cffab6bf58d69471d64
|
| 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
|
| 9abb2978f09643227664ab70c0677744b8f6e946 |
|
04-Mar-2016 |
Nicolas Geoffray <ngeoffray@google.com> |
Fix very rare bug around JIT code cache collection.
The bug is the following:
1) JIT thread: We start a code cache collection.
2) JIT thread: We mark all code that is in the call stack of all
threads.
3) Mutator thread: after marking its stack, resumes and does call
that pushes JIT compiled code to the call stack.
4) Mutator thread: deoptimizes compiled code of ArtMethod Foo,
and therefore updates the entry point of Foo through
JitCodeCache::InvalidateCompiledCodeFor.
(Note that updating the entrypoint could also be done through
instrumentation).
5) JIT thread: Call JitCodeCache::RemoveUnusedAndUnmarkedCode.
The method used to remove entries that were not entrypoints.
It sees the compiled code for Foo but that is not an entrypoint
anymore, so deletes it.
6) Mutator thread problem: it now has compiled code in its call
stack that is deleted.
If it's only one mutator thread, we only hit a DCHECK when walking
the stack, as we are now seeing an invalid pc. The deoptimization
will longjmp to the caller of that invalid entry anyway.
However, if multiple mutator threads are involved, one thread
might invalidate the compiled code while the other is still
running it. And we end up deleting code that is in the call
stack of a thread, and we will crash.
The fix is to mark entrypoints before marking call stacks,
so that anything a thread might jump to is marked and kept.
bug:27424509
bug:23128949
bug:26846185
Change-Id: I07cd08cedd96b9900629f7535e95404f622104ea
|
| bcd94c8ea9bde4e075c25fbdfb3a2ef6858eed7b |
|
03-Mar-2016 |
Nicolas Geoffray <ngeoffray@google.com> |
Refine statistics around the JIT.
- Better namings.
- Now also time the code cache collection time.
- Random cleanups.
bug:23128949
bug:27445008
bug:27442890
Change-Id: I1dd52544bea678af868e7c47907f7a0fc9a146c3
|
| 67e4f81b232150605cbba03d0818313737b904d3 |
|
02-Mar-2016 |
Nicolas Geoffray <ngeoffray@google.com> |
Clear inline caches if a ProfilingInfo gets revived.
This avoids stalled class references.
Bug: 27398183
Bug: 23128949
Bug: 26846185
(cherry picked from commit 511e41b4ec2d378a5c434598ebd0b3d6136fff9d)
Change-Id: I5e33496859e65456d185cc09ca12736dd6414d77
|
| af3f5e0136277bd0c01e7e2b7c4797c63ecddca6 |
|
02-Mar-2016 |
Nicolas Geoffray <ngeoffray@google.com> |
Revert "Revert "Use the interpreter as a heartbeat for the JIT.""
Bug: 27398183
Bug: 23128949
Bug: 26846185
This reverts commit a96917a6983a5abbe973255a3846fda549fb1657.
(cherry picked from commit 35122443e5f8606cc5a660ac32745a06aefb341b)
Change-Id: Ib02401c76bf28e1aac6bbce01f97ddd0373c1ba5
|
| 511e41b4ec2d378a5c434598ebd0b3d6136fff9d |
|
02-Mar-2016 |
Nicolas Geoffray <ngeoffray@google.com> |
Clear inline caches if a ProfilingInfo gets revived.
This avoids stalled class references.
Bug: 27398183
Bug: 23128949
Bug: 26846185
Change-Id: I9539215241708e26fef887e02201ce1feabc2d1a
|
| 35122443e5f8606cc5a660ac32745a06aefb341b |
|
02-Mar-2016 |
Nicolas Geoffray <ngeoffray@google.com> |
Revert "Revert "Use the interpreter as a heartbeat for the JIT.""
Bug: 27398183
Bug: 23128949
Bug: 26846185
This reverts commit a96917a6983a5abbe973255a3846fda549fb1657.
Change-Id: I5c4f0d87d3293a6a7ab56a33396670704b66a347
|
| a96917a6983a5abbe973255a3846fda549fb1657 |
|
01-Mar-2016 |
Nicolas Geoffray <ngeoffray@google.com> |
Revert "Use the interpreter as a heartbeat for the JIT."
Hits a DCHECK testing libcore.
art F 11973 12675 art/runtime/jit/jit_code_cache.cc:644] Check failed: it.second->GetProfilingInfo(sizeof(void*)) != nullptr
Bug: 27398183
Bug: 23128949
Bug: 26846185
This reverts commit 7273a5d045d3ceb3ff011ad65765356b69b155e8.
Change-Id: I6614a82e775ea71aa16f041313f67546db41eac8
|
| 034eb1a6ae870d53f9b3774e416ebbbd9859fdda |
|
01-Mar-2016 |
Nicolas Geoffray <ngeoffray@google.com> |
Revert "Fix bogus interaction between code cache and instrumentation."
hits a DCHECK when testing libcore:
art F 11973 12675 art/runtime/jit/jit_code_cache.cc:644] Check failed: it.second->GetProfilingInfo(sizeof(void*)) != nullptr
This reverts commit e9924b7d4fcafe7261233d2c16ebb73d7bb5e8e8.
Change-Id: I37d05991b9506a53a7c76738a7e4a6def1010958
|
| e9924b7d4fcafe7261233d2c16ebb73d7bb5e8e8 |
|
01-Mar-2016 |
Nicolas Geoffray <ngeoffray@google.com> |
Fix bogus interaction between code cache and instrumentation.
Change-Id: Icc314089ec1a05248b8994476d478a543e629e4d
|
| 1e7da9bd04bdd2664a4196f1d7e285c010f8881f |
|
01-Mar-2016 |
Nicolas Geoffray <ngeoffray@google.com> |
Do a TryLock when allocating a ProfilingInfo from the interpreter.
This removes some thread contentions just for allocating ProfilingInfo.
bug:23128949
Change-Id: I9ff7d44c4b0ee272425cf4c6248d3065f67958f3
|
| 7273a5d045d3ceb3ff011ad65765356b69b155e8 |
|
29-Feb-2016 |
Nicolas Geoffray <ngeoffray@google.com> |
Use the interpreter as a heartbeat for the JIT.
When doing a partial code cache collection, update all entrypoints to
interpreter, so that the next full collection will remove code
that wasn't executed during that window.
bug:27398183
bug:23128949
bug:26846185
Change-Id: I4423f5c4810dac183dc8973078bf218818745e80
|
| 932282595045739d624cbbf5b3a3d337fe3b6797 |
|
24-Feb-2016 |
Nicolas Geoffray <ngeoffray@google.com> |
Forgot to initialize JIT memory use fields.
bug:23128949
(cherry picked from commit b0d220835f2ddc03510f17fcc4fb4c64b426a0f5)
Change-Id: Iea81925ebd0fd60fe58a972269313a98f93fdac5
|
| 5b3fbd26a9f8b8c15e8e85db2403b3545536cd87 |
|
25-Feb-2016 |
Nicolas Geoffray <ngeoffray@google.com> |
Dump the number of OSR compiled code.
bug:23128949
(cherry picked from commit fcdd72941810f03460a8efca0c6255439be80c35)
Change-Id: I7f6c428511ee7fe346c665be8f02880bcc317cae
|
| 82de410ad37ea90f43e61fb5dc44ab420a159b54 |
|
23-Feb-2016 |
Nicolas Geoffray <ngeoffray@google.com> |
Introduce partial code cache collection.
It will collect code that is known unused (because it deoptimized),
and osr code.
bug:26846185
(cherry picked from commit 8d37250f7b9a7839b11488f45a1842b025026f94)
Change-Id: Iffe87e356066a3f0bd28cb3df41c17bf2ac9ea0f
|
| fcdd72941810f03460a8efca0c6255439be80c35 |
|
25-Feb-2016 |
Nicolas Geoffray <ngeoffray@google.com> |
Dump the number of OSR compiled code.
Change-Id: I20efc80e8556da8220dab92c3a7947f883d48cf8
|
| b0d220835f2ddc03510f17fcc4fb4c64b426a0f5 |
|
24-Feb-2016 |
Nicolas Geoffray <ngeoffray@google.com> |
Forgot to initialize JIT memory use fields.
Change-Id: I82ae97adffd8e633de048f044cdbdb953b2f9763
|
| 8d37250f7b9a7839b11488f45a1842b025026f94 |
|
23-Feb-2016 |
Nicolas Geoffray <ngeoffray@google.com> |
Introduce partial code cache collection.
It will collect code that is known unused (because it deoptimized),
and osr code.
bug:26846185
Change-Id: Ic27dfeb944efb2ca464039007ba365c1e0d4a040
|
| 65b83d875e4f9c021bc8380024455c9f1e658b14 |
|
22-Feb-2016 |
Nicolas Geoffray <ngeoffray@google.com> |
Add a new kReservedCapacity to decide when to start GC code.
Change-Id: Ib9d0ad9bc86c66ac29975f7f907f79f842342500
|
| 38ea9bd50debc63ceb77ec1e65036d8e2070ec7d |
|
19-Feb-2016 |
Nicolas Geoffray <ngeoffray@google.com> |
Track memory usage of code cache.
Needed once we start having a strategy based on how full the code
cache is.
Change-Id: I181ad05e1c216afabf106f91ac54bbf144d7406a
|
| b88d59ef4fe611fe47e50a6a19785e03bbd5f93b |
|
17-Feb-2016 |
Nicolas Geoffray <ngeoffray@google.com> |
Be a bit smarter with JIT code triggering deoptimization.
Do not re-use an OSR method that triggered deoptimization.
Also add a stack overflow check before doing OSR.
bug:27094810
Change-Id: I6ff6a7fb9b3df9b7c0ff37e3610595efa70ad067
|
| a9b91313cee629d81d4784798d03419fc7fa9e11 |
|
17-Feb-2016 |
Nicolas Geoffray <ngeoffray@google.com> |
We also need to delete osr entries when deleting ArtMethod.
In the unfortunate event an ArtMethod gets allocated at the
same location as an old (deleted) ArtMethod, the osr_code_map_
lookup will succeed and return garbage. So we need to delete
entries in the osr_code_map_ when an ArtMethod gets deleted.
This should finally fix:
dalvik.system.DexClassLoaderTest#test_twoJar_diff_getResourceAsStream
Change-Id: I7c8b775c3376a6cfcb907f09b783e393967ad82d
|
| d9994f069dfeaa32ba929ca78816b5b83e2a4134 |
|
11-Feb-2016 |
Nicolas Geoffray <ngeoffray@google.com> |
Re-enable OSR.
Fixes two bugs:
- Dealing with proxy methods, which the compiler and code cache
does not handle.
- Dealing with phi types, that may have been speculatively optimized
but do not hold once jumping to the compiled code.
Change-Id: I7dcd9976ef7b12128fff95d2b7ed3e69cc42e90a
|
| b331febbab8e916680faba722cc84b66b84218a3 |
|
05-Feb-2016 |
Nicolas Geoffray <ngeoffray@google.com> |
Revert "Revert "Implement on-stack replacement for arm/arm64/x86/x86_64.""
This reverts commit bd89a5c556324062b7d841843b039392e84cfaf4.
Change-Id: I08d190431520baa7fcec8fbdb444519f25ac8d44
|
| bd89a5c556324062b7d841843b039392e84cfaf4 |
|
05-Feb-2016 |
David Brazdil <dbrazdil@google.com> |
Revert "Implement on-stack replacement for arm/arm64/x86/x86_64."
DCHECK whether loop headers are covered fails.
This reverts commit 891bc286963892ed96134ca1adb7822737af9710.
Change-Id: I0f9a90630b014b16d20ba1dfba31ce63e6648021
|
| 891bc286963892ed96134ca1adb7822737af9710 |
|
29-Jan-2016 |
Nicolas Geoffray <ngeoffray@google.com> |
Implement on-stack replacement for arm/arm64/x86/x86_64.
High-level overview:
- osr_method_threshold is used to know when to compile a method
in osr mode (-> treat all loops as irreducible).
- branch instructions in the compiler query whether they can
jump to an osr method.
- An osr entry point is found through the stack maps: if a stack
map is duplicated in the CodeInfo, it is an osr entry point.
Change-Id: Ifb39338cd281e2c7eccce67f4e18d46428be71e4
|
| 324302696c0a93511fef1310d309a54ebfb25a46 |
|
01-Feb-2016 |
Roland Levillain <rpl@google.com> |
Ensure instruction cache is flushed in compiler tests with Clang.
Invoke __builtin___clear_cache with Clang in
art::CommonCompilerTest::MakeExecutable. This fixes
"Illegal instruction" failures in codegen_test on angler.
Factor this logic in a function and use it in
art::jit::JitCodeCache::CommitCodeInternal too.
Bug: 26786447
Change-Id: I81fa056c8d7c6bf2168c9340deeab82231cb1084
|
| c26f1288aaf9141ac4b2150869f0c594d52dbf36 |
|
29-Jan-2016 |
Nicolas Geoffray <ngeoffray@google.com> |
Re-enable test 566-polymorphic-inlining.
- Can not rely on debug builds.
- Need to wait for the method to be compiled.
Change-Id: I26ce89075075da8555fd59ade56bd04bec23f4ce
|
| a42363f79832a6e14f348514664dc6dc3edf9da2 |
|
17-Dec-2015 |
Nicolas Geoffray <ngeoffray@google.com> |
Implement first kind of polymorphic inlining.
Add HClassTableGet to fetch an ArtMethod from the vtable or imt,
and compare it to the only method the profiling saw.
Change-Id: I76afd3689178f10e3be048aa3ac9a97c6f63295d
|
| b4eddd21c364c51b6fc5c439cda6958ae255dcd5 |
|
14-Jan-2016 |
Calin Juravle <calin@google.com> |
Add profile support for apps with shared runtime.
- extend ProfileSaver to track different dex locations to different
profile files.
- allow repeated calls to Runtime::RegisterAppInfo() which will
register a new pair <profile_file, locations_to_track> with the profile
saver.
Bug: 26080105
Change-Id: I58c4587d5f462c557411ad7126a00c29cd7db3a6
|
| 0a52223d9173315aeaca05217b8c792b3088ea7d |
|
19-Jan-2016 |
Nicolas Geoffray <ngeoffray@google.com> |
Keep track of number of JIT compilations.
Change-Id: Ic436bc3f939f6236190ddd840955cb72f76e8f85
|
| 3fdb3fec5c36127d8568b2f89698906ba4b68576 |
|
14-Jan-2016 |
Mathieu Chartier <mathieuc@google.com> |
Don't allow thread suspension in ProfilingInfo::Create
AddSamples calls ProfilingInfo::Create which would occasionally
transition to suspended and wait for code cache GC to complete.
This CL removes the thread state change and
WaitForPotentialCollectionToComplete which caused thread suspension.
The thread suspension caused occasional stale object references in
InvokeVirtualOrInterface since moving GC could occur.
If this_object became stale, we would put it's stale class into the
profile info, resulting in a GC crash.
Bug: 26086970
Change-Id: I5a86561098d72b7be80e8a3bcf3d8401403a0b00
|
| c3fec4c4ca01055d4d5920f363620b5726ebbbef |
|
14-Jan-2016 |
Nicolas Geoffray <ngeoffray@google.com> |
Fix braino in parameters passed.
Change-Id: I4683586184617bb8ac4f377b873263a339d1d960
|
| a25dce9b452ba17ef7cef768926c884177a3025e |
|
12-Jan-2016 |
Nicolas Geoffray <ngeoffray@google.com> |
Support --perf with jitted code.
Change-Id: I76f29d815234e9506efc59d4865780e52b2381a1
|
| 5cc349f3dd578e974f78314c50b6a0267c23e591 |
|
18-Dec-2015 |
David Srbecky <dsrbecky@google.com> |
Report DWARF debug information for JITed code.
Change-Id: Ia5b2133c54386932c76c22774cf3d2ae61e0925f
|
| 07f3564d73550495a937bc667d03721106315400 |
|
04-Jan-2016 |
Nicolas Geoffray <ngeoffray@google.com> |
Add a fence to ensure other threads see the correct data.
Other threads may see the ProfilingInfo pointer in the ArtMethod,
but the data they have for it is bogus due to a prior load in their
cache.
bug:26221227
Change-Id: I1fa8f088a3144df39d07a986ed57d6fc20edba18
|
| 4d77b6a511659f26fdc711e23825ffa6e7feed7a |
|
01-Dec-2015 |
Calin Juravle <calin@google.com> |
Save profile information in a separate thread.
Previously we would save the profiling information only when the app
was sent to background. This missed on an important number of updates
on the jit code cache and it didn't work for background processes.
Bug: 26080105
Change-Id: I84075629870e69b3ed372f00f4806af1e9391e0f
|
| 66f55237679db90cb0a0a265043a787932b466f8 |
|
08-Dec-2015 |
Calin Juravle <calin@google.com> |
Pass down to the runtime the application code paths.
When registering the app with the runtime the framework needs to pass
down the list of application code paths. This will be used by JIT to
know what profile info to persist.
This fixes the reliance on OatFileManager::GetPrimaryOatFile which may
produce inconsistent results based on external factors (i.e. class path
order or failing to compile the first dex file from the class path)
Bug: 26080105
Change-Id: Iadcebd2684fcd48569e8f76ef21bd4d117fedc05
|
| 73be1e8f8609708f6624bb297c9628de44fd8b6f |
|
17-Sep-2015 |
Nicolas Geoffray <ngeoffray@google.com> |
Inline monomorphic calls.
Change-Id: If38171c2dc7d4a4378df5d050afc4fff4499c98f
|
| 31f2c155975c5794d481df03eb0947cb48d2c6b5 |
|
23-Oct-2015 |
Calin Juravle <calin@google.com> |
Save jit profiling info to file.
Currently saves only the hot method references and omits the inline
cache.
Change-Id: I79e91f855f6b52dc19a9a86662604f2eee16613c
|
| 0a3be1620a3560253cfa789cb9819013293c5654 |
|
18-Nov-2015 |
Nicolas Geoffray <ngeoffray@google.com> |
Increase code cache after 1 full collection.
Also add a max capacity option.
Change-Id: Icd442b72e9be0c6b091b588b4c4473c69b7cde10
|
| 8ccbd279b91c96aa0f7bb649e399b12b31b8bda1 |
|
09-Nov-2015 |
Nicolas Geoffray <ngeoffray@google.com> |
Remove unused variable.
Change-Id: I29d80004eba897b72304ec83fe89ed1071a24024
|
| a5891e81a2fb833307cf7c7e7267070dc0223dc8 |
|
06-Nov-2015 |
Nicolas Geoffray <ngeoffray@google.com> |
Fix interaction between JIT and instrumentation.
- The JIT needs to go through the instrumentation to update
entry points.
- The instrumention needs to know if a method got JITted
to know if needs to deoptimize.
bug:25438583
Change-Id: I4b186a1da9f4a3fb329efd052a774d5502a902a1
|
| d28b969c273ab777ca9b147b87fcef671b4f695f |
|
04-Nov-2015 |
Nicolas Geoffray <ngeoffray@google.com> |
Code cleanup to avoid CompilerDriver abstractions in JIT.
Avoids allocating a CompiledMethod.
Change-Id: I35b4aa0d7c74daba68e827a01e71c300fce3b3bf
|
| 5a23d2ea6e0d89112ff11ec765e676c03818b7c2 |
|
03-Nov-2015 |
Nicolas Geoffray <ngeoffray@google.com> |
Fix TODO on instrumentation and add some more DCHECKs.
bug:25343683
bug:25438583
Change-Id: I232deb1b6761466b514c687ce304f61928755cdc
|
| 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
|
| 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
|
| 4e915fbc2be5fc43c782389bbbb7306cd76a523a |
|
28-Oct-2015 |
Nicolas Geoffray <ngeoffray@google.com> |
Tune heuristics a bit.
First version. Seems to be more in line with what we're seeing.
Change-Id: I5117352c13cb2e95d16f0c9b05b1faf37c05d710
|
| 26705e2b1245b65989a0341b24c5dbf2658d4bb6 |
|
28-Oct-2015 |
Nicolas Geoffray <ngeoffray@google.com> |
Add support for collection ProfilingInfo objects.
Change-Id: I8bb6069530253a7372acdf2b5aee71e1de644822
|
| 1dad3f68b7f5a4a4cb2b281413357adc2309a8fd |
|
23-Oct-2015 |
Nicolas Geoffray <ngeoffray@google.com> |
Support garbage collection of JITted code.
Change-Id: I9afc544460ae4fb31149644b6196ac7f5182c784
|
| 1e7de6cfcabc87ebd36bf6f2c9ed466152d21d4e |
|
21-Oct-2015 |
Nicolas Geoffray <ngeoffray@google.com> |
Remove the magic 32 constant and ensure alignment.
Change-Id: I383315bf3cd5f0b8634e2982da55d5d864009a91
|
| 0c3c2668ef44fdbd18d97f9134a85d1a7d561aa4 |
|
15-Oct-2015 |
Nicolas Geoffray <ngeoffray@google.com> |
Use DlMallocSpace for the JIT code cache.
- Also tidy up some code in the JIT compiler.
- And mprotect code space to be writable only when allocating.
Change-Id: I46ea5c029aec489f2af63452de31db3736aebc20
|
| 5550ca8bcc742b109d77e62f3a0877c667d894d3 |
|
21-Aug-2015 |
Nicolas Geoffray <ngeoffray@google.com> |
Record profiling information before Jitting.
- Add a new instrumentation kind to record dynamic invokes.
- Use the JNI entry point field to store the profiling data.
- Record seen receivers for every dynamic invoke.
Change-Id: I2c1738ab2a72052d45964d055dc16b44b906e54c
|
| 3d21bdf8894e780d349c481e5c9e29fe1556051c |
|
22-Apr-2015 |
Mathieu Chartier <mathieuc@google.com> |
Move mirror::ArtMethod to native
Optimizing + quick tests are passing, devices boot.
TODO: Test and fix bugs in mips64.
Saves 16 bytes per most ArtMethod, 7.5MB reduction in system PSS.
Some of the savings are from removal of virtual methods and direct
methods object arrays.
Bug: 19264997
(cherry picked from commit e401d146407d61eeb99f8d6176b2ac13c4df1e33)
Change-Id: I622469a0cfa0e7082a2119f3d6a9491eb61e3f3d
Fix some ArtMethod related bugs
Added root visiting for runtime methods, not currently required
since the GcRoots in these methods are null.
Added missing GetInterfaceMethodIfProxy in GetMethodLine, fixes
--trace run-tests 005, 044.
Fixed optimizing compiler bug where we used a normal stack location
instead of double on ARM64, this fixes the debuggable tests.
TODO: Fix JDWP tests.
Bug: 19264997
Change-Id: I7c55f69c61d1b45351fd0dc7185ffe5efad82bd3
ART: Fix casts for 64-bit pointers on 32-bit compiler.
Bug: 19264997
Change-Id: Ief45cdd4bae5a43fc8bfdfa7cf744e2c57529457
Fix JDWP tests after ArtMethod change
Fixes Throwable::GetStackDepth for exception event detection after
internal stack trace representation change.
Adds missing ArtMethod::GetInterfaceMethodIfProxy call in case of
proxy method.
Bug: 19264997
Change-Id: I363e293796848c3ec491c963813f62d868da44d2
Fix accidental IMT and root marking regression
Was always using the conflict trampoline. Also included fix for
regression in GC time caused by extra roots. Most of the regression
was IMT.
Fixed bug in DumpGcPerformanceInfo where we would get SIGABRT due to
detached thread.
EvaluateAndApplyChanges:
From ~2500 -> ~1980
GC time: 8.2s -> 7.2s due to 1s less of MarkConcurrentRoots
Bug: 19264997
Change-Id: I4333e80a8268c2ed1284f87f25b9f113d4f2c7e0
Fix bogus image test assert
Previously we were comparing the size of the non moving space to
size of the image file.
Now we properly compare the size of the image space against the size
of the image file.
Bug: 19264997
Change-Id: I7359f1f73ae3df60c5147245935a24431c04808a
[MIPS64] Fix art_quick_invoke_stub argument offsets.
ArtMethod reference's size got bigger, so we need to move other args
and leave enough space for ArtMethod* and 'this' pointer.
This fixes mips64 boot.
Bug: 19264997
Change-Id: I47198d5f39a4caab30b3b77479d5eedaad5006ab
|
| e401d146407d61eeb99f8d6176b2ac13c4df1e33 |
|
22-Apr-2015 |
Mathieu Chartier <mathieuc@google.com> |
Move mirror::ArtMethod to native
Optimizing + quick tests are passing, devices boot.
TODO: Test and fix bugs in mips64.
Saves 16 bytes per most ArtMethod, 7.5MB reduction in system PSS.
Some of the savings are from removal of virtual methods and direct
methods object arrays.
Bug: 19264997
Change-Id: I622469a0cfa0e7082a2119f3d6a9491eb61e3f3d
|
| a4885cbaafd35fe9c60eb6cd95e41e2c86f54f66 |
|
09-Mar-2015 |
Mathieu Chartier <mathieuc@google.com> |
Add way to print JIT related info at shutdown
Added a runtime option called -XX:DumpJITInfoOnShutdown with prints
various interesting JIT statistics during the shutdown of the runtime
if enabled.
Example of running EvaluateAndApplyChanges with the option:
I art : Code cache size=427KB data cache size=84KB num methods=1598
I art : Start Dumping histograms for 1598 iterations for JIT timings
I art : Compiling: Sum: 727.850ms 99% C.I. 135us-4402us Avg: 455.475us Max: 15567us
I art : Verifying: Sum: 191.074ms 99% C.I. 19us-1450.500us Avg: 119.570us Max: 5142us
I art : TrimMaps: Sum: 62.053ms 99% C.I. 13us-199.374us Avg: 38.831us Max: 10058us
I art : MakeExecutable: Sum: 11.153ms 99% C.I. 4us-47us Avg: 6.979us Max: 47us
I art : Initializing: Sum: 7.991ms 99% C.I. 1us-49.781us Avg: 5us Max: 4852us
I art : Done Dumping histograms
Bug: 17950037
Change-Id: I23538b24f947e5f6934f55579b8baa9d806eb3d5
|
| 5c42c29b89286e5efa4a4613132b09051ce5945b |
|
25-Feb-2015 |
Vladimir Marko <vmarko@google.com> |
Add support for .bss section in oat files.
Change-Id: I779b80b8139d9afdc28373f8c68edff5df7726ce
|
| e5f13e57ff8fa36342beb33830b3ec5942a61cca |
|
24-Feb-2015 |
Mathieu Chartier <mathieuc@google.com> |
Revert "Revert "Add JIT""
Added missing EntryPointToCodePointer.
This reverts commit a5ca888d715cd0c6c421313211caa1928be3e399.
Change-Id: Ia74df0ef3a7babbdcb0466fd24da28e304e3f5af
|
| a5ca888d715cd0c6c421313211caa1928be3e399 |
|
24-Feb-2015 |
Nicolas Geoffray <ngeoffray@google.com> |
Revert "Add JIT"
Sorry, run-test crashes on target:
0-05 12:15:51.633 I/DEBUG (27995): Abort message: 'art/runtime/mirror/art_method.cc:349] Check failed: PcIsWithinQuickCode(reinterpret_cast<uintptr_t>(code), pc) java.lang.Throwable java.lang.Throwable.fillInStackTrace() pc=71e3366b code=0x71e3362d size=ad000000'
10-05 12:15:51.633 I/DEBUG (27995): r0 00000000 r1 0000542b r2 00000006 r3 00000000
10-05 12:15:51.633 I/DEBUG (27995): r4 00000006 r5 b6f9addc r6 00000002 r7 0000010c
10-05 12:15:51.633 I/DEBUG (27995): r8 b63fe1e8 r9 be8e1418 sl b6427400 fp b63fcce0
10-05 12:15:51.633 I/DEBUG (27995): ip 0000542b sp be8e1358 lr b6e9a27b pc b6e9c280 cpsr 40070010
10-05 12:15:51.633 I/DEBUG (27995):
Bug: 17950037
This reverts commit 2535abe7d1fcdd0e6aca782b1f1932a703ed50a4.
Change-Id: I6f88849bc6f2befed0c0aaa0b7b2a08c967a83c3
|
| 2535abe7d1fcdd0e6aca782b1f1932a703ed50a4 |
|
17-Feb-2015 |
Mathieu Chartier <mathieuc@google.com> |
Add JIT
Currently disabled by default unless -Xjit is passed in.
The proposed JIT is a method JIT which works by utilizing interpreter
instrumentation to request compilation of hot methods async during
runtime.
JIT options:
-Xjit / -Xnojit
-Xjitcodecachesize:N
-Xjitthreshold:integervalue
The JIT has a shared copy of a compiler driver which is accessed
by worker threads to compile individual methods.
Added JIT code cache and data cache, currently sized at 2 MB
capacity by default. Most apps will only fill a small fraction of
this cache however.
Added support to the compiler for compiling interpreter quickened
byte codes.
Added test target ART_TEST_JIT=TRUE and --jit for run-test.
TODO:
Clean up code cache.
Delete compiled methods after they are added to code cache.
Add more optimizations related to runtime checks e.g. direct pointers
for invokes.
Add method recompilation.
Move instrumentation to DexFile to improve performance and reduce
memory usage.
Bug: 17950037
Change-Id: Ifa5b2684a2d5059ec5a5210733900aafa3c51bca
|