Age | Commit message (Collapse) | Author |
|
The AArch32 VIXL-based code generator has been the default
ARM code generator in ART for some time now. The old ARM
code generator does not compile anymore; retiring it.
Test: test.py
Bug: 63316036
Change-Id: Iab8fbc4ac73eac2c1a809cd7b22fec6b619755db
|
|
When generating code for ARM64, maintain the status of
Thread::Current()->GetIsGcMarking() in register X20,
dubbed MR (Marking Register), and check the value of that
register (instead of loading and checking a read barrier
marking entrypoint) in read barriers.
Test: m test-art-target
Test: m test-art-target with tree built with ART_USE_READ_BARRIER=false
Test: ARM64 device boot test
Bug: 37707231
Change-Id: Ibe9bc5c99a2176b0a0476e9e9ad7fcc9f745017b
|
|
bug: 62210114
Test: 657-branches
Change-Id: I753a9a57e404c792cd4375ea66c91839684bdee2
|
|
One should not write to something below the stack pointer, or
it could be overwritten during an interrupt.
Test: test.py
Change-Id: Ie6c997b9f7548ca5844303d6a3fc0c531f469c6e
|
|
Move iterator code from bit_utils.h into bit_utils_iterator.h. Move
Identity into stl_util_identity.h. Remove now unnecessary includes,
and fix up transitive users.
Test: m
Change-Id: Id1ce9cda66827c5d00584f39ed310b6b37629906
|
|
Test: m test-art-host-gtest
Test: testrunner.py --host
Test: testrunner.py --target
Test: Nexus 6P boots.
Test: Build aosp_mips64-userdebug.
Bug: 30627598
Change-Id: I0e54fdd2e91e983d475b7a04d40815ba89ae3d4f
|
|
|
|
In preparation for replacing the dex cache method array
with a hash-based array, get rid of one unnecessary use.
This method load kind is currently used only on mips for
irreducible loops and OSR, so this should have no impact
on x86/x86-64/arm/arm64.
Test: m test-art-host-gtest
Test: testrunner.py --host
Test: Repeat the above tests with manually changing
kDexCachePcRelative to kRuntimeCall in sharpening.cc.
(Ignore failures in 552-checker-sharpening.)
Bug: 30627598
Change-Id: Ifce42645f2dcc350bbb88c2f4642e88fc5f98152
|
|
|
|
Reverting because GenerateCalleeMethodStaticOrDirectCall()
prevents replacing kDexCacheViaMethod with kRuntimeCall
where we would not retrieve the target method at all and
leave the runtime to retrieve and call it just like for
unresolved methods.
The intrinsic should be re-implemented by loading the
flags through HLoadClass.
Note that the intrinsic was unimplemented for CC and a bit
broken for non-CC, using LDR instead of LDRB for loading
the flags.
Test: Rely on TreeHugger.
Bug: 32535355
Bug: 30627598
This reverts commit d8c052ac0aa3382c4807add33afa32580ffeecbb.
Change-Id: I81fd14dac60c94ac543e336f4f3c888259fc8bd7
|
|
The old name does not reflect the actual code anymore.
Test: testrunner.py --host
Change-Id: I2e13cf727bba9d901c4d3fc821bb526d38a775b8
|
|
|
|
In preparation for adding ArtMethod entries to the .bss
section, add direct PC-relative pointers to methods so that
the number of needed .bss entries for boot image is small.
Test: m test-art-host-gtest
Test: testrunner.py --host
Test: testrunner.py --target on Nexus 6P
Test: Nexus 6P boots.
Test: Build aosp_mips64-userdebug
Bug: 30627598
Change-Id: Ia89f5f9975b741ddac2816e1570077ba4b4c020f
|
|
constant input""
This reverts commit 3082661d260449e1d773f077e914160c7ad58de5, and
fixes the handling of HCondition with boolean inputs.
Test: m test-art-target-run-test-409-materialized-condition
Test: art/tools/run-libcore-tests.sh
Change-Id: Ib21e3a81ba41ce20c06e9a9e454c4322af1513ae
|
|
We no longer support non-PIC boot image compilation.
Also clean up some obsolete code for method patches
and make JIT correctly report itself as non-PIC.
Test: testrunner.py --host
Test: testrunner.py --target
Bug: 33192586
Change-Id: I593289c5c1b0e88b82b86a933038be97bbb15ad2
|
|
|
|
Make actual types more explicit, either by replacing "auto"
with actual type or by assigning std::pair<> elements of
an "auto" variable to typed variables. Avoid binding const
references to temporaries. Avoid copying a container.
Test: m test-art-host-gtest
Change-Id: I1a59f9ba1ee15950cacfc5853bd010c1726de603
|
|
For array accesses the element address has the following structure:
Address = CONST_OFFSET + base_addr + index << ELEM_SHIFT
Taking into account ARM64 LDR/STR addressing modes address part
(CONST_OFFSET + index << ELEM_SHIFT) can be shared across array
access with the same data type and index.
For example, for the following loop 5 accesses can share address
computation:
void foo(int[] a, int[] b, int[] c) {
for (i...) {
a[i] = a[i] + 5;
b[i] = b[i] + c[i];
}
}
Test: test-art-host, test-art-target
Change-Id: I46af3b4e4a55004336672cdba3296b7622d815ca
|
|
|
|
constant input""
|
|
Reverting to see if that change is responsible for a crash. Will share with ARM if it is.
This reverts commit b404f349d69f940ef2974915fe97c16070364efd.
Change-Id: Idd04f9109447319445ff49f3fd7dc5b069b4883f
|
|
|
|
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
Change-Id: I0c2d87b40db09e8e475cf97a7c784a034c585e97
|
|
Test: m test-art-target-gtest
Test: testrunner.py --target on Nexus 6P.
Test: testrunner.py --target on Nexus 6P with heap poisoning enabled.
Test: Repeat the above tests with ART_USE_OLD_ARM_BACKEND=true.
Bug: 29516974
Bug: 30126666
Bug: 36141117
Change-Id: I458f2ec5fe9abead4db06c7595d992945096fb68
|
|
|
|
Test: m test-art-target-run-test-409-materialized-condition
Test: m test-art-target-run-test-570-checker-select
Change-Id: Ie3418fbac1d2ecb4d7fc847d4722758e1e01bc2e
|
|
|
|
On ARM and ARM64, avoid loading the reference altogether when the
GC is not marking.
Also, extract the code logic for updating a reference field from
GenerateReferenceLoadWithBakerReadBarrier routines and move it to
new routines (UpdateReferenceFieldWithBakerReadBarrier), to make
the implementation more legible.
Test: Run ART target tests in Baker read barrier configuration.
Bug: 29516974
Change-Id: I11c53f0607e997cd02ec7911725e98ef3dc97d90
|
|
Remaining work for follow-up CLs:
- use implicit null check in field thunk,
- use 16-bit LDRs for fields and GC roots.
Test: m test-art-target-gtest
Test: testrunner.py --target on Nexus 6P.
Test: testrunner.py --target on Nexus 6P with heap poisoning enabled.
Test: Repeat the above tests with ART_USE_OLD_ARM_BACKEND=true.
Bug: 29516974
Bug: 30126666
Bug: 36141117
Change-Id: Iad5addab72d790a9d61879f61f2e75b246bcdf5a
|
|
Part one of a few upcoming CLs to optimize constructor fences.
This improves load-store-elimination; all singleton objects that are not
returned will have their associated constructor fence removed.
If the allocation is removed, so is the fence. Even if allocation is not
removed, fences can sometimes be removed.
This change is enabled by tracking the "this" object associated with the
constructor fence as an input. Fence inputs are considered weak; they do not keep
the "this" object alive; if the instructions for "this" are all deleted,
the fence can also be deleted.
Bug: 36656456
Test: art/test.py --host && art/test.py --target
Change-Id: I05659ab07e20d6e2ecd4be051b722726776f4ab1
|
|
Add explicit field initialization to default value where necessary.
Also clean up interpreter intrinsics header.
Test: m
Change-Id: I7a850ac30dcccfb523a5569fb8400b9ac892c8e5
|
|
|
|
Test: m test-art-target-run-test-009-instanceof
Test: m test-art-target-run-test-422-instanceof
Test: m test-art-target-run-test-494-checker-instanceof-tests
Test: m test-art-target-run-test-500-instanceof
Test: m test-art-target-run-test-530-instanceof-checkcast
Test: m test-art-target-run-test-603-checker-instanceof
Change-Id: Ia5e1421403605659d0f53bc794acb5e5b0af0c5e
|
|
Test: m test-art-target-run-test-570-checker-select
Change-Id: I87d2e87eb2fd30355101df07eb3754b013cedf63
|
|
Generally speaking, this optimization applies to all code
generation visitors ending with a call to Bind(), which
includes intrinsics with kNoCall CallKind. However, no
changes are done for slow paths (which frequently end with
a branch to an exit label that is bound at the end of a
visitor).
Test: m test-art-target
Change-Id: Ie1a0c8c54ef76b01e7f0b23962c56c29ca8984a9
|
|
|
|
Test: m test-art-target-run-test-538-checker-embed-constants
Change-Id: I8e6af76b99543331e8ffec01bd8df3f09890708e
|
|
Test: mma test-art-host && mma test-art-target
Change-Id: I05051c03dbd3684c674096def84020494d28364b
|
|
|
|
* changes:
Revert "Revert "Use the holder's gray bit in Baker read barrier slow paths (ARM, ARM64).""
Revert "Revert "Use the "GC is marking" information in compiler read barriers (ARM, ARM64).""
|
|
Test: m test-art-target-run-test-566-checker-codegen-select
Test: m test-art-target-run-test-570-checker-select
Change-Id: If0140892303490701782df9a818e6d8346bf3d6c
Signed-off-by: Anton Kirilov <anton.kirilov@linaro.org>
|
|
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
|
|
(ARM, ARM64).""
This reverts commit 47b3ab2fd83aaa530b7d2c62bfc024209b8b6923.
In compiler-generated code, when deciding whether to mark
a heap reference or not in a read barrier, after checking
whether the GC is currently marking, also check (in the
slow path) whether the lock word of the reference's holder
is gray, before actually marking the reference.
This change is only for ARM and ARM64, as it does not
benefit x86 nor x86-64.
Change-Id: Ia26b07f0485e23589bfc0e65f83852f2795688c0
Test: Run ART tests in Baker read barrier configuration.
Test: Boot a device in Baker read barrier configuration.
Bug: 35780827
Bug: 29516974
|
|
barriers (ARM, ARM64).""
This reverts commit 35345a555bd7928582a7ffa6369b374b3ddc379d.
In compiler-generated code, when deciding whether to mark
a heap reference or not in a read barrier, check whether
the GC is currently marking, instead of checking the gray
bit in the reference's holder's lock word.
This change is only for ARM and ARM64, as it does not
benefit x86 nor x86-64.
Change-Id: Id3d2758c600115b2f07d345442cfa87edfc2792c
Test: Run ART tests in Baker read barrier configuration.
Test: Boot a device in Baker read barrier configuration.
Bug: 35780827
Bug: 29516974
|
|
ARM64)."
This reverts commit 1372c9f40df1e47bf775f1466bbb96f472b6b9ed.
This change (along with https://android-review.googlesource.com/#/c/342429/)
creates null pointer dereferences.
Bug: 35780827
Bug: 29516974
Change-Id: I2a9c4d0ad8d2ab870c2e0ddbff32152933c77abe
|
|
ARM64)."
This reverts commit 27b1f9cbfc1409418eee4b0e22f29f033e10b64d.
This change (along with https://android-review.googlesource.com/#/c/342428/)
creates null pointer dereferences.
Bug: 35780827
Bug: 29516974
Change-Id: If731960a405f9b89528f3daaf235da57cabc5c11
|
|
In compiler-generated code, when deciding whether to mark
a heap reference or not in a read barrier, after checking
whether the GC is currently marking, also check (in the
slow path) whether the lock word of the reference's holder
is gray, before actually marking the reference.
This change is only for ARM and ARM64, as it does not
benefit x86 nor x86-64.
Test: Run ART tests in Baker read barrier configuration.
Test: Boot a device in Baker read barrier configuration.
Bug: 29516974
Change-Id: I60595a8f4987747faeaa359ad873e9758c1ded75
|
|
In compiler-generated code, when deciding whether to mark
a heap reference or not in a read barrier, check whether
the GC is currently marking, instead of checking the gray
bit in the reference's holder's lock word.
This change is only for ARM and ARM64, as it does not
benefit x86 nor x86-64.
Test: Run ART tests in Baker read barrier configuration.
Test: Boot a device in Baker read barrier configuration.
Bug: 29516974
Change-Id: Ia5d90286bb9f753f3bbcb3a6254eb166523a2ff5
|
|
This commit mirrors the work that has already been done for ARM64.
Test: m test-art-target-run-test-551-checker-shifter-operand
Change-Id: Iec8c1563b035f40f0e18dcffde28d91dc21922f8
|
|
|