| Age | Commit message (Collapse) | Author |
|---|
|
Add closing namespace comments.
Bug: 32619234
Test: m
Change-Id: I1f50e09dcd1038c4b540b87e5c19e319c1f592e4
|
|
Test: testrunner --host --pictest --jit
Bug: 33192586
Change-Id: I1e05874752c6a3b448ffbeff9681637a35acda2a
|
|
ART side of the change.
bug:34715556
test: test-art-host, jdwp, libcore
Change-Id: I3a73ae4af2d602431150c8ecfceaddb9ba519cee
|
|
- 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
|
|
Because we no longer support running patchoat on npic oat files, which
means the included patch information is unused .
Bug: 33192586
Test: m test-art-host
Change-Id: I9e100c4e47dc24d91cd74226c84025e961d30f67
|
|
- Don't write any quickening info.
- Don't visit methods.
Saves ~20% of compilation times.
Test: test-art-host
Change-Id: Ib18fd06c0ca42308e1d81401de0ee3e6297de0ce
|
|
Both vdex and BCI require full dequickening support. This
change implements dequickening by using the existing encoded
quickening, and adds the checkcast quickening to that encoding.
bug: 30937355
bug: 32369913
Test: dex_to_dex_decompiler_test.cc test-art-host
Change-Id: Ie95f46946d59b28157d6e47dcf4a859be032d1c3
|
|
We will need that information when taking an OTA to make sure
the same set of classes needs to be verified at runtime.
Currently, the vdex file will contain a list of unverified
classes. We could alternatively encode a bit vector of the size
of the type_id array, but the few experiments I did show that
the bit vector is actually larger. We can refine this later.
bug: 30937355
test: m test-art-host
test: verifier_deps_test.cc
Change-Id: I2670e4fd2e54ee7a148246baa705fda3a56617ff
|
|
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
|
|
For host-generated *.oat files, generate a SHA-1 build ID based on the
file content and write it to .note.gnu.build-id ELF section. This
should allow various developer tools like profilers correlate the data
captured for files like boot.oat on the device with the corresponding
known version of the file during an offline analysis.
Test: Verified that boot.oat contains the build ID section now (with
this change and https://android-review.googlesource.com/#/c/275630
applied)
Test: Added ElfWriterTest::CheckBuildIdPresent test.
Test: make test-art-host
Bug: 31292208
Change-Id: Ie5e89da2ef87e34c27c0237ab34ddc7d2dc0aa3b
|
|
Fixed oat_test to keep dex files alive. Fixed mips build.
Rewritten the .bss GC root visiting and added write barrier
to the artResolveStringFromCode().
Test: build aosp_mips-eng
Test: m ART_DEFAULT_GC_TYPE=SS test-art-target-host-gtest-oat_test
Test: Run ART test suite on host and Nexus 9.
Bug: 20323084
Bug: 30627598
This reverts commit 5f926055cb88089d8ca27243f35a9dfd89d981f0.
Change-Id: I07fa2278d82b8eb64964c9a4b66cb93726ccda6b
|
|
There are some issues with oat_test64 on host and aosp_mips-eng.
Also reverts "compiler_driver: Fix build."
Bug: 20323084
Bug: 30627598
This reverts commit 63dccbbefef3014c99c22748d18befcc7bcb3b41.
This reverts commit 04a44135ace10123f059373691594ae0f270a8a4.
Change-Id: I568ba3e58cf103987fdd63c8a21521010a9f27c4
|
|
And do some related refactorings.
Bug: 20323084
Bug: 30627598
Test: Run ART test suite including gcstress on host and Nexus 9.
Test: Run ART test suite including gcstress with baker CC on host and Nexus 9.
Test: Build aosp_mips64-eng.
Change-Id: I1b12c1570fee8e5da490b47f231050142afcbd1e
|
|
This change introduces new dex2oat switch --run-passes=. This switch
accepts path to a text file with names of passes to run.
Compiler will run optimization passes specified in the file rather
then the default ones.
There is no verification implemented on the compiler side. It is user's
responsibility to provide a list of passes that leads to successful
generation of correct code. Care should be taken to prepare a list
that satisfies all dependencies between optimizations.
We only take control of the optional optimizations. Codegen (builder),
and all passes required for register allocation will run unaffected
by this mechanism.
Change-Id: Ic3694e53515fefcc5ce6f28d9371776b5afcbb4f
|
|
Test: manually, on device.
Change-Id: If007a1657dd5769ddef03691e0a19dbbe6ba1a29
|
|
This is the intended behavior to have a good JNI transition
performance.
Bug: 28902384
(cherry picked from commit 8c185bf0c6f18a5349bc87a7e3751ba06d90f461)
Change-Id: I52767909b916ada3c619206c8838b85bff5ac316
|
|
Instead use the value of '-Xcompiler-option --compiler-filter=XXX' for
the target compiler filter to use to determine if the runtime should
try to invoke dex2oat when oat files are loaded.
Bug: 27641809
Change-Id: I1856e0e37df91835b81105567c70d8a285a88f62
|
|
This reverts commit 845e5064580bd37ad5014f7aa0d078be7265464d.
Add an option to change what OatFileManager considers up-to-date.
In our tests we're allowed to write to the dalvik-cache, so it
cannot be kSpeed.
Bug: 27689078
Change-Id: I0c578705a9921114ed1fb00d360cc7448addc93a
|
|
Bots are red. Tentative reverting as this is likely the offender.
Bug: 27689078
This reverts commit a62d2f04a6ecf804f8a78e722a6ca8ccb2dfa931.
Change-Id: I3ec6947a5a4be878ff81f26f17dc36a209734e2a
|
|
Record the compiler filter in the oat header. Use that to determine
when the oat file is up-to-date with respect to a target compiler
filter level.
New xxx-profile filter levels are added to specify if a profile should
be used instead of testing for the presence of a profile file.
This change should allow for different compiler-filters to be set for
different package manager use cases.
Bug: 27689078
Change-Id: Id6706d0ed91b45f307142692ea4316aa9713b023
|
|
So long, old friend.
Change-Id: I0241c798a34b92bf994fed83888da67d6e7f1891
|
|
Only verifies and dex2dex compiles classes in the profile. Goal
is to reduce application launch time.
~2x faster than interpret-only for Facebook.
Bug: 27688727
(cherry picked from commit a079e3aa62cceb76c1c1811e6e09bcaf75e20289)
Change-Id: Iad5aa1adee3aa6c2408820e8cbbab2d4412021b8
|
|
Check if both --debuggable and --generate-debug-info are set instead.
History: I have recently added the --native-debuggable flag to control
whether extra stack maps are generated in order to produce accurate
native line number mapping of all generated code. I didn't want to
include it into --debuggable since we compile boot.oat as debuggable
and so it would be affected.
On second thought, it would have been reasonable to generate the extra
stackmaps only if both --debuggable and --generate-debug-info are set,
instead of introducing another compiler flag. This means we do not
affect boot.oat but we still get the extra stack maps if DWARF debug
information is explicitly requested.
Change-Id: I7e5e557e4850a88b3b6f86178d2cb645fb1e1110
|
|
Expose some compiler options so we can run the image test with
the optimizing compiler without running into a stack overflow.
Also allow a variable amount of threads in CreateCompilerDriver.
Use 16 as a middle ground on the host to speed up the otherwise
now slowed-down test.
Bug: 27240085
Bug: 27552475
Change-Id: I8db5055d32ae722c8f430903244faa9166cc4886
|
|
Oat files compiled with --compiler-filter=verify-at-runtime contain
no compiled code and therefore are independent of the boot image.
This patch stores an ExtractOnly flag in the oat header and skips
the image checksum test if the flag is set, rendering the oat file
up to date even after OTAs.
Bug: 26813999
Change-Id: I25291d5b49d9e9d0018844e957a2dc88ef6bdc27
|
|
|
|
To ensure reliable stable generation of a boot image, add a flag
for forced determinism, trading compile time for a deterministic
output.
We have to run certain passes in the compiler-driver single-threaded.
It is also necessary to try to make the heap layout deterministic.
Switch to nonconcurrent GC, use the free-list implementation for LOS,
and try to allocate the main space at a known location. This is best
effort at the moment.
To allow the compiler phase to be parallelized, const-strings need
to be created beforehand.
The identity hashcode seed needs to be pinned.
Besides the Dex object we also need to null the DexFile pointer in
dex caches.
For classes, we need to remove the clinit thread ID.
Fix oatdump alignment padding accounting.
Bug: 26687569
Change-Id: Ia82120e8f715bb3691d861817b12778ac677355a
|
|
|
|
This will allow tests to specify arbitrary dex files from
which we should not inline, in addition to core-oj, so that
we can test the related functionality. Additionally, should
the core-oj.jar core grow beyond a single classes.dex and
require a multi-dex .jar, this change makes sure that we
prevent inlining also from the extra classes<N>.dex files.
Change-Id: I74da4839bf9bb405dd62ad80563bf646a7a65dd9
|
|
This reverts commit 8546cc9aeb05e866e1fb6a9e4130d53ea330baa8.
Change-Id: I676fdf9af27fa3b16fa8921778ff8832ab8c437d
|
|
This reverts commit 5fdcc3c931b70204fd8c491afa66f57f8428490f.
Change-Id: I9c1f5aad6933a46af6717e3a90a51f76111f9c8a
|
|
Add flag --generate-mini-debug-info which generates
LZMA compressed .symtab and .debug_frame, which are
sufficient to print java backtraces in libunwind.
If enabled, it increases the size of boot.oat by about 3.5%.
Change-Id: Ic3c2ef7704c05fa328720c6781ca2a9b8e3935a3
|
|
It simplifies passing the option to the JIT.
Change-Id: Iee1b722362899e5809ef97be90961e3dda1e16cc
|
|
The option is disabled by default unless explicitly enabled.
In particular, it is no longer enabled in debug builds,
and the --debuggable option does not affect it.
I want to use this flag to control the debug data generated
by the JIT as well. Since this takes run-time memory,
I want to avoid enabling it unless explicitly requested.
Change-Id: I2e2afa2f56bb0a113e92cc2e26e00dceac1689ca
|
|
This reverts commit b17d1ccff0ac26fc22df671907ba2b4f4c656ce4.
Change-Id: I26f6f8702a448c3da12662cbc6bc0f6e562bc40b
|
|
Multiple changes to dex2oat and the runtime to support a --multi-image
option. This generates a separate oat file and image file output for
each dex file input.
Change-Id: Ie1d6f0b8afa8aed5790065b8c2eb177990c60129
|
|
Generate extra stack map at the start of each java statement.
The stack maps are later translated to DWARF which allows
LLDB to set breakpoints and view local variables.
Change-Id: If00ab875513308e4a1399d1e12e0fe8934a6f0c3
|
|
This reverts commit fcb7613d3aaa9a6802800b6e957aaad51cedf6dc.
Change-Id: Idc0df6a2f68e8b5aa740bb1259f19c2953811510
|
|
go/lem driven:
Performance:
Richards +41%
CaffeineMethod +43%
ReversiBench: +52%
Towers: +73%
Tak: +85%
Memory use: 7% less memory
CompileTime: 14% increase
CodeSize: 8% increase
Last three measures are now more acceptable given we JIT.
Change-Id: Ic4aa6535d2b76cf3545ef00e9b2ae32330f10745
|
|
Currently only the CompilerOptions class. We should also
do it for the CompilerDriver options.
This will fix the flakiness seen on jdwp testing, as the
debuggable flag was not passed to the compiler.
Change-Id: I4218dd5928c9f4fe2d6191ab11e5598e7cf84bcf
|
|
This change increases the maximum code unit size, and fold
parameters in the inlinee in the hope to reduce the overall
size of the graph. We then make sure we don't inline methods
that have more than N HInstructions.
Also, remove the kAccDontInline flag on ArtMethod. The compiler
does not need it anymore.
Change-Id: I4cd3da40e551f30ba83b8b274728b87e67f6812e
|
|
Bug: 21868508
Change-Id: Ic83813a966cef18e59447083926bf033aa587154
|
|
Bug: 21868508
Change-Id: Ice7f1604ed65e3d4ed2a010ee431272b7d000cdb
|
|
Previously the kEverything filter failed to compile class
initializers. Now it will.
Bug: 19576257
Change-Id: I189d2b5b379aee112c4e95f8d3e6c32abab6ed41
|
|
Replace the flags --include-debug-symbols and --include-cfi
with single flag called --generate-debug-info (with alias -g).
The name "symbol" was not ideal, since depending on context it
may be interpreted as "ELF symbols", or "debugging information".
This CL also means that we have only the options to include
either all debugging information or none. This should be fine,
since we can use standard tools to strip anything we do not want.
Change-Id: I721fded56d755d7eed0ef36aa84e841a1f5747f8
|
|
Decouple generation of CFI from the rest of debug symbols.
This makes it possible to generate oat with CFI but without
the rest of debug symbols.
This is in line with intention of the .eh_frame section.
The section does not have the .debug_ prefix because it
is considered somewhat different to the rest of debug symbols.
Change-Id: I32816ecd4f30ac4e0dc69d69a4993e349c737f96
|
|
It is never used. We generally use --include-debug-symbols instead.
Change-Id: I933495cd85ab4f7dfcf3ef3cd2d943d4b870ffeb
|
|
Move the relative patcher classes to compiler/linker/ and
compiler/linker/<arch>/ . Refactor them to avoid OatWriter
dependency so that they can be unit tested. Add tests for
x86 and x86-64.
Change-Id: I1b42baa9fc431378e4cce1399bec590c5b5a409f
|
|
Verifies at runtime only, instead of at compilation time.
AOSP HH boot time after clean-oat: ~30s instead of ~35s if enabled.
Also helps install time if enabled there.
TODO: See if there is any possible deadlocks that can result from
this.
Bug: 19762303
Change-Id: Ibfba77148da9039e8d7d7497c05486bc044eefe7
|
|
Add an option that aborts dex2oat when a hard verifier failure occurs.
Bug: 19606409
Change-Id: I53195284e22fe6207274101e85745af763c06271
|
