| Age | Commit message (Collapse) | Author |
|---|
|
The main goal here is performance by avoiding JNI and especially
NativeAllocationRegistry overhead. We gain a factor of 10 or so
on small BigInteger arithmetic. For large computations, we gain
substantially in a few cases where OpenJDK seems to use better
algorithms. AFAIK, with this version we never lose by integral
factors relative to what we had before.
A secondary goal is to clean out open BigInteger bugs.
The base version is 8u252, which for this part of the tree is
identical to June 15 2020 ToT.
Note that this means we included the java.math part of
https://hg.openjdk.java.net/jdk8u/jdk8u/jdk/rev/a5f5d7fd9be6 (29 Oct 2019)
That appears to be a separable fix that makes no interface changes.
I re-added @NonNull annotations. We also removed some code to write
certain backwards-compatibility fields during serialization, since we
never had those.
This also adds a the <X>ValueExact Java 8 BigInteger methods that had
not been previously supported.
This removes a Harmony test for compatibility with the RI implementation
that the RI implementation itself apparently fails. And, IMO, the
observed behavior is better than what we previously tested for.
Fixed a bunch of comment spelling errors. All of these have now
been fixed in upstream ToT.
Benchmarking showed that a straight move to the RI version slowed down
large multiplications, divisions, and modular exponentiation enough to
make it problematic. Thus I reintroduced NativeBN to allow those to fall
back to boringssl, at least for sufficiently large inputs. It was moved
to a hopefully more appropriate location. The fallback is tuned for
64 bits; for 32 bits it's probably much less useful; much of the
boringssl performance advantage comes from 64-bit "digits".
The boringssl fallbacks are not completely free, since we need extra
conversions on each operation. But since we only do this for large
asymptotically expensive computations, we see at mosts tens of percents
regressions, which probably qualifies as "in the noise" here. If we
find additional performance issues, we can add more boringssl fallbacks;
the required code is now fairly straightforward.
Unlike the old version, this no longer uses NativeAllocationRegistry or
similar mechanisms at all. Native memory is only used on a short-term
basis, with explicit deallocation. We no longer use boringssl for
simple linear-time operations like addition.
Microbenchmark results for the newly added benchmark, and for a
close-to-final BigInteger version, listed as "combined":
Msecs/iteration
Digits: 5/10 50/100 500/1000 5000/10000
Inner product, 1000 terms, factors of larger indicated size:
current 2.1 2.5 6.4 168
RI 0.11 0.66 11.0 486
combined 0.12 0.67 9.1 189
Harmonic series, uses smaller indicated size
current 3.7 3.2 4.3 17.6
RI 0.16 0.34 1.4 14.3
combined 0.17 0.34 1.23 14.2
(1+1/100000)^100000, larger size
current 0.07 0.073 0.33 15.8
RI 0.011 0.049 1.553 63.6
combined 0.011 0.049 0.48 13.6
Single modPow() call, smaller size
current 0.005 0.011 1.1 583
RI 0.006 0.038 7.2 5580
combined 0.011 0.012 1.1 541
Single modInverse call, larger size
current 0.003 0.014 0.375 27.8
RI 0.003 0.003 0.026 1.6
combined 0.002 0.002 0.008 0.4
Bug: 160641142
Bug: 136887041
Bug: 119491938
Bug: 28214673
Bug: 28251030
Bug: 2950143
Test: AOSP Boots. Ran some manual Calculator tests on Cuttlefish.
Change-Id: Id577d99328013b1e3c389973dcb0195fa7f52b0b
|
|
Rename ZoneInfoDB -> ZoneInfoDb to conform with coding standards.
Bug: 148086409
Test: treehugger
Change-Id: I1054d129cc4d53587a187c32927311f7311754fe
|
|
The new annotation is android.compat.annotation.UnsupportedAppUsage.
Test: m
Bug: 145132366
Change-Id: Ie13cce08e6d0a7c15427d2c3fc7d4b4d395acedc
|
|
|
|
Bug: 147884233
Test: build
Test: atest luni/src/test/java/libcore/libcore/timezone/TimeZoneDataFilesTest.java
Change-Id: Idab0ada3bd7da46a414016acbb9d6914f18ac154
|
|
This change also adds it to the last-api.txt for the intra-core as it
will be required as part of the long-term "stable" intra-core API in
order to run mainline Conscrypt on Android 10. This API is unchanged
since Android 10 and so this is a backwards compatible change.
Note that AsynchronousCloseMonitor is implicitly part of the intra-core
API already as its native component is invoked from Conscrypt JNI code.
Bug: 147870945
Test: TreeHugger
Change-Id: I61c52fc1499860db887c4cce38817913536cc98a
|
|
Use ThreadPrioritySetter registered in Android runtime to set thread
priority.
Bug: 139521784
Test: CtsLibcoreTestCases, device boot to home, thread priority set
as expected.
Change-Id: I46093c9539be3acca9ca580a1e88676bdfd2de54
|
|
Revert "Implement the ThreadPrioritySetter interface"
Revert submission 1194951-ThreadPrioritySetter2
Reason for revert: need to update new API
Reverted Changes:
Iee33a26c0: Implement the ThreadPrioritySetter interface
I71c5052be: Use ThreadPrioritySetter to set thread priority
Change-Id: I98cb32d5d216a83634d38406f574ee76e77daa97
|
|
Use ThreadPrioritySetter registered in Android runtime to set thread
priority.
Bug: 139521784
Test: CtsLibcoreTestCases, device boot to home, thread priority set
as expected.
Change-Id: I71c5052beb171f966eba97401159306b87e397b7
|
|
Test: m + flash
Bug: 142650523
Change-Id: I4c50bebb9fbfc3df89d948004c6c16dbb60ce8c7
|
|
|
|
Add core platform APIs / impl code needed to access telephonylookup.xml
on device.
There are now two XML files related to time zones: the common code
between them has been extracted into the package-private XmlUtils class.
Bug: 143451294
Test: atest libcore.libcore.timezone
Change-Id: I7da534f92681ca6d567b09355eebb9c88db74ad2
|
|
usage of ICU"
|
|
- The moved code will be part of i18n module in the future
- Some new @IntraCoreApi, e.g. EmptyArray, is unnecessary, but
the cleanup will be done in a separate CL.
Bug: 138994281
Test: atest --test-mapping external/icu/android_icu4j/src/main/java/com/android/icu/charset:presubmit
Change-Id: Ic69725584e4a485595bfc25fc72546abb11a6904
|
|
|
|
Assume it does not break app because
no method is marked with @UnsupportedAppUsage
Test: m droid cts
Change-Id: Ia18f795ac598d5d8fcc583641a91130c8dd20c26
|
|
Test: m checkbuild
Change-Id: I7e5b200906512b11f2f5a0bcbe1ee8704a511bac
|
|
Avoids splitting packages that are part of java.base.
Removes patch_module: "java.base" that was used to work around this
issue.
Bug: 143057012
Test: m checkbuild
Change-Id: Iad65343fd3fde55f2d3b0e68c4bbcd51377d85d9
|
|
Adds an art.module.api.annotations that frameworks/base can use
instead of the unsupportedappusage_annotation_files filegroup.
Also, does some internal cleanup.
* Moves the definition of the filegroup(s) into
non_openjdk_java_files.bp alongside the rest.
* Stops duplicating paths to the various files.
* Renames the internal target to make it clearer what it contains.
Bug: 142113521
Test: m java
Change-Id: I059652d4bb3085da0b975fa8cbd05dd16e16654f
|
|
MimeMap is not specific to Networking (URLConnection.getFileNameMap()
is not the only client) so it didn't make sense for the class to live
in the old packet. The new packet mirrors frameworks' android.content.type
which holds the logic constructing the default instance.
The new packet is specific to this purpose (libcore.content did not
previously exist) and makes for a more sensible CorePlatformApi that
we are more likely to be able to support long term.
Bug: 124232146
Bug: 136256059
Test: atest CtsMimeMapTestCases CtsLibcoreTestCases
Change-Id: I2a6576cf6f5f06bf159a975ad40a1c09a615e9b5
|
|
This is the second attempt to submit this CL. The first attempt
regressed on app startup because RuntimeInit installed the
custom MimeMap from commonInit() which runs post-fork of the zygote,
but that was fixed by installing it pre-fork.
This CL topic moves the default MimeMap implementation to frameworks.
Libcore starts with a minimal implementation sufficient to pass
CtsLibcoreTestCases, but frameworks can inject the real implementation.
Before this CL topic, the data files and logic (MimeMapImpl) were part of
core-*.jar on device; after this CL, they instead live in framework.jar.
Tests from MimeMapTest that check behavior of that default
implementation also move to a non-libcore CTS test.
Planned work for follow-up CL:
1. Make CTS more opinionated, with a plan to assert that all of
the default mappings are present. How exactly the expectated
mapping will be bundled in CTS is still TBD.
2. Add a vendor.mime.types file (defaults to empty) where vendors
can add additional mappings; I plan to make it such that mappings
in that file are parsed last but never override any earlier
mappings, as if each mime type / file extension was prefixed
with '?'.
3. Perhaps enforce that public APIs android.webkit.MimeTypeMap
and java.net.URLConnection.getFileNameMap() behave consistently
with MimeMap.getDefault().
Test: atest CtsLibcoreTestCases
Test: atest CtsMimeMapTestCases
Test: Checked that CtsLibcoreTestCases still passes on a build that
is missing the MimeMap.setDefault() call from RuntimeInit.java.
Test: Checked that app startup time does not regress as part of this
CL topic - see http://b/136256059#comment17
Bug: 136256059
Change-Id: I54103f0e1b83bdca5e083b1968ccb50421b94748
|
|
This is unused and @hide.
The last test in libcore.libcore.net.http is being deleted by this
change, so the TEST_MAPPING which references it is removed as well.
Test: make
Change-Id: Ief0a803db2137d9ab5b3ffc8d9bd555862a9edcd
|
|
Use filegroup to specify sources files with base directory via 'path'
property. This allows the build system to filter Java source files by
their packages.
Bug: 140764681
Test: m
Change-Id: I52760b3cbe3c99f5c48473f667322d6102cb7f53
|
|
am: d60d516bcc
Change-Id: I34fa7b76f95c16293abb8a7fc7dd71f6c8699467
|
|
This reverts commit d9e06a7351c25bf3275cffc382e3e7f8c87e6079.
Reason for revert: Caused slower app startup (I don't know why).
Change-Id: I7865dd97ee202e26236498505af3d746143f74e4
|
|
am: 14090d8c85
Change-Id: Ic73f2cfec61853af937bf5eb6635826ee3d9a91d
|
|
This CL topic moves the default MimeMap implementation to frameworks.
Libcore starts with a minimal implementation sufficient to pass
CtsLibcoreTestCases, but frameworks can inject the real implementation.
Before this CL topic, the data files and logic (MimeMapImpl) were part of
core-*.jar on device; after this CL, they instead live in framework.jar.
Tests from MimeMapTest that check behavior of that default
implementation also move to a non-libcore CTS test.
Planned work for follow-up CL:
1. Make CTS more opinionated, with a plan to assert that all of
the default mappings are present. How exactly the expectated
mapping will be bundled in CTS is still TBD.
2. Add a vendor.mime.types file (defaults to empty) where vendors
can add additional mappings; I plan to make it such that mappings
in that file are parsed last but never override any earlier
mappings, as if each mime type / file extension was prefixed
with '?'.
3. Perhaps enforce that public APIs android.webkit.MimeTypeMap
and java.net.URLConnection.getFileNameMap() behave consistently
with MimeMap.getDefault().
Test: atest CtsLibcoreTestCases
Test: atest CtsMimeMapTestCases
Test: Checked that CtsLibcoreTestCases still passes on a build that
is missing the MimeMap.setDefault() call from RuntimeInit.java.
Bug: 136256059
Change-Id: I1621bd2907d5923a19541e3e859238c46187f8df
|
|
am: f54cb9109c
Change-Id: I9d03803ae3e39bd52afa7d6ee3bd84e8dbff0034
|
|
Move Half FP16 implementations to libcore, to allow ART compiler to
optimize them with intrinsic implementations.
For example, on ARM64 with ARMv8.2 FP16 half registers and instructions:
- Half toFloat/toHalf can be implemented with FCVT;
- Half floor/ceil/round can be implmented with FRINT(pna);
- Half max/min can be implmented with FMIN/FMAX.
Such fast Half FP16 intrinsics can help accelerate ColorLong ARGB
encoding/decoding in Android framework.
Change-Id: I10b835fa6a3a33c9ec19df5de4949681c97b4667
|
|
am: 3598a59db2
Change-Id: I70a14de700c8ddb136096cdb180818bf4390a1ed
|
|
Takes advantage of a new mechanism for specifying the default visibility
of all modules in a package to restrict the visibility of libcore
modules as much as possible:
package {default_visibility: ["//visibility:private"]}
Bug: 133290645
Bug: 134379140
Test: m droid
Change-Id: I2ba1a6131993d218fd65cd4a8f0ad29a7a83df14
|
|
7e391faa2f
am: 5a71074adb
Change-Id: Ic9a3fb9778b163ea4b8ff533845315d8b9b8d3e6
|
|
am: 6a19cb82fe
Change-Id: I33a879389aa2d60a32a976a199a37136fc3f884a
|
|
|
|
|
|
This commit changes StructIcmpHdr to the byte[] factory that it
effectively already is. StructIcmpHdr is not referenced by anything in
android.system.Os so is not the same as other Struct classes.
It also add toString() to structs / containers where missing and makes
a field final.
Bug: 135660232
Test: build / boot
Test: atest CtsLibcoreTestCases:libcore.android.system.OsTest
Change-Id: I84985c817234aadc2380a08a6afa71400366a6f6
|
|
am: 4e714431df
Change-Id: I6f093919870ab8004167c676640410dc17550f9f
|
|
Os.fcntlFlock() is not used and is untested. There seems to be
no reason to keep it.
Test: build only
Bug: 135660232
Change-Id: I55ed799690fd3f273a4aa9e88f567428baf6bbcf
|
|
The annotations form the basis of the new framework, allowing
compatibility change IDs to be defined, and targeted to apps based on
their targetSdkVersion.
No real implementation of the API is provided here, but callbacks are
defined to allow an implementation to be provided by higher levels of
the stack.
Further CLs will implement the APIs added here, add annotation processors
to process the break IDs and enforce some policy, and more. An internal
NDK API will also be added to allow use from C++ code.
The annotations and API are added in libcore to allow their use in all
necessary places. Compatibility changes may be implemented in libcore, in
frameworks/base or in other places. Putting the basic into libcore allows
their usage in all of these places.
Bug: 135010838
Test: m
Change-Id: Iac22820b6d83679d70ff6615790a2cbf783fc307
|
|
#2)." am: 287d9b5a37
am: c8fbf4ca68
Change-Id: I33db15b141063a3469f0ac6d10de4a24779d1508
|
|
This is attempt #2 to submit this CL. The first attempt failed because
some usages of MimeUtils had remained in the downstream -plus-aosp
branch. These are now cleaned up by cherry-picking the corresponding
internal master changes to that branch.
Delete unused class MimeUtils.
Some MimeUtils calls used to be on the graylist and used by some apps
through reflection; such apps can instead use this public API, for now:
android.webkit.MimeTypeMap.getSingleton().getMimeTypeFromExtension(String)
android.webkit.MimeTypeMap.getSingleton().getExtensionFromMimeType(String)
Test: Treehugger
Change-Id: I8bec5aebbbcf0a0a4a8e5dc317d188830f219a1e
|
|
am: 0e9ea2bcbf
am: 2d85b26917
Change-Id: I6c5a410c425f4931a622d1d6781946b7f336d4fa
|
|
am: 0e9ea2bcbf
Change-Id: I6ef357c98ab027ebea792aac4a215a3ad7773394
|
|
This reverts commit b23d88cba391ac2b0757bf01f7cc7fe0311d3da8.
Reason for revert: Caused a build failure in a downstream -plus-aosp branch that still had references to MimeUtils remaining in the sources.
Change-Id: I71ab7c03fba233ae60b729c0d2172e909d420307
|
|
40eea593a0
am: 8996619c33
Change-Id: Ib7effb43e0996561bbd335f2def0be4821831998
|
|
am: 40eea593a0
Change-Id: I36def99acc1d7e328af4c51b991ff4a9ebdc6d82
|
|
Delete unused class MimeUtils.
Some MimeUtils calls used to be on the graylist and used by some apps
through reflection; such apps can instead use this public API, for now:
android.webkit.MimeTypeMap.getSingleton().getMimeTypeFromExtension(String)
android.webkit.MimeTypeMap.getSingleton().getExtensionFromMimeType(String)
Test: Treehugger
Change-Id: I409342f1c95d7f2c49b158efa0e43d32081abc76
|
|
b152621257
am: 8f97234ef3
Change-Id: Id611bb49a234a70238b1e505c884ef6616554d14
|
|
am: b152621257
Change-Id: Ied0adb311cfc201e7b6259e9105752a75ee37f1c
|
|
This CL factors out MimeMap{,Impl} from MimeUtils. MimeMap supports
customizing the implementation instance, but this is currently an
internal implementation detail; this CL does not change the behavior
of any API surface.
In future CLs, I plan to do the following:
1.) Expose MimeMap, including {get,compareAndSet}Default, via a new
@CorePlatformApi.
2.) Convert existing users of MimeUtils to use MimeMap.getDefault() instead.
3.) Delete MimeUtils.
Bug: 120171727
Test: Checked that this CL does not change behavior of MimeUtils by
checking that the test logic attached to http://b/122831291#comment7
passes both before and after this CL.
Test: atest CtsLibcoreTestCases:libcore.libcore.net.Mime{Map,Utils}Test
Change-Id: Idc27c29826a6eb2e4c9c54be058eb0577d525aec
|
