[Bug 3488] [IcedTea8] Add latest updates from AArch64 port for 3.6.0

bugzilla-daemon at icedtea.classpath.org bugzilla-daemon at icedtea.classpath.org
Wed Nov 1 06:40:21 UTC 2017


https://icedtea.classpath.org/bugzilla/show_bug.cgi?id=3488

--- Comment #2 from hg commits <mercurial at icedtea.classpath.org> ---
details:
http://icedtea.classpath.org//hg/icedtea8?cmd=changeset;node=f71e75cc8eed
author: Andrew John Hughes <gnu_andrew at member.fsf.org>
date: Wed Nov 01 05:26:05 2017 +0000

    Bump shenandoah to aarch64-shenandoah-jdk8u151-b13.

    Upstream changes:
      - Add missing UseShenandoahGC checks to C2
      - [backport] Add JVMTI notifications to Shenandoah GC pauses.
      - [backport] After Evac verification should run consistently
      - [backport] All definitions should start with Shenandoah*
      - [backport] Allocation latency tracing
      - [backport] Allow allocations in pinned regions
      - [backport] Assorted monitoring support fixes
      - [backport] Avoid Full STW GC on System.gc() + related fixes
      - [backport] BrooksPointer tracing overwhelms -Xlog:gc=trace
      - [backport] Cannot do more than 1000 Full GCs
      - [backport] Cap heap size for TestRegionSizeArgs test
      - [backport] Cleanup "dirty" mentions
      - [backport] Cleanup unused methods and statements + Trivial cleanup:
removed unused field, etc.
      - [backport] Common pause marker to capture everything before/after pause
      - [backport] Consistent print_on and tty handling
      - [backport] "continuous" heuristics
      - [backport] Disable biased locking by default
      - [backport] Fix build error: avoid loops with empty bodies
      - [backport] Fix build error: switches over enums should take all enums
      - [backport] Fix build error: verifier liveness should not be implicitly
casted to size_t
      - [backport] Fixed assertion failures when printing heap region to trace
output
      - [backport] Fixed C calling convention of shenandoah_wb() on Windows
      - [backport] LotsOfCycles test always degrades to Full GC
      - [backport] Made ShenandoahPrinter debug only
      - [backport] Make sure different Verifier levels work
      - [backport] Make sure we have at least one memory pool per memory
manager (JMX) + JMX double-counts heap used size
      - [backport] Mark heuristics diagnostic/experimental
      - [backport] Move Verifier "start" message under (gc,start)
      - [backport] On-demand commit as heap resizing strategy
      - [backport] Periodic GC
      - [backport] PhiNode::has_only_data_users() needs to apply to shenandoah
barrier only
      - [backport] Pinning humongous regions should be allowed
      - [backport] Reclaimed humongous regions should count towards immediate
garbage
      - [backport] Refactor region flags into finite state machine
      - [backport] Refactor ShConcThread dispatch
      - [backport] Refactor ShenandoahFreeSet + Fast-forward over humongous
regions to keep "current" non-humongous
      - [backport] Refactor ShenandoahHeapLock
      - [backport] Refactor ShenandoahHeapRegionSet
      - [backport] Region (byte|word) shifts as the replacement for divisions
      - [backport] Rehash -XX:-UseTLAB in tests + Rehash allocation tests
      - [backport] Rename inline guards
      - [backport] Selectable humongous threshold + Humongous top() should be
correct for iteration
      - [backport] Shortcut concurrent cycle when enough immediate garbage is
reclaimed
      - [backport] Templatize and improve inlining of arraycopy and clone
barriers.
      - [backport] TestRegionSampling test
      - [backport] TestSmallHeap test for Shenandoah
      - [backport] Uncommit heap regions after given delay
      - [backport] Underflow in adaptive free_threshold calculation
      - [backport] Unlock more GC-specific tests for Shenandoah
      - [backport] Update counters on slow-path more rarely
      - [backport] Verifier should avoid pushing on stack when walking objects
past TAMS
      - [backport] Verifier should walk cset and humongous regions
      - [backport] Verify humongous regions liveness
      - [backport] Verify liveness data
      - Correct way to fix Windows call convention issue
      - Fix build error in release config.
      - Fixed Fixed message logging
      - Handle Java heap initialization and expansion failures
      - Make sure -verbose:gc, PrintGC, PrintGCDetails work consistently
      - Missing barriers on constant oops + acmp rework + cas fix + write
barrier on constant oop fix
      - Missing UseShenandoahGC check in LibraryCallKit::inline_multiplyToLen()
      - Missing UseShenandoahGC check to C2
      - OOME in SurrogateLockerThread deadlocks the GC cycle
      - Properly unlock ShenandoahVerify
      - Remove unused memory_for, fixing the build
      - Remove useless code following acmp rework
      - Revert accidental G1 closure rename
      - S8161190: AArch64: Fix overflow in immediate cmp instruction
      - S8173207: Upgrade compression library
      - S8173770: Image conversion improvements
      - S8174164: SafePointNode::_replaced_nodes breaks with irreducible loops
      - S8175097: [TESTBUG] 8174164 fix missed the test
      - S8179084: HotSpot VM fails to start when AggressiveHeap is set
      - S8180711: Better invokespecial checks
      - S8181420: PPC: Image conversion improvements
      - S8181946: JDK 8 THIRD_PARTY_README - Minor Cleanup
      - S8184235: Backout JDK-8173207 from 8u141, 7u151 and higher updates
source base
      - S8184682: Upgrade compression library
      - S8187224: aarch64: some inconsistency between aarch64_ad.m4 and
aarch64.ad
      - Test bug: test library and flags in TestHeapAlloc
      - UnlockDiagnosticVMOptions flag is needed for ShenandoahVerify
      - Write barrier pin and expand cleanup

    ChangeLog:

    2017-10-31  Andrew John Hughes  <gnu_andrew at member.fsf.org>

        Bump shenandoah to aarch64-shenandoah-jdk8u151-b13.
        * patches/hotspot/shenandoah/8161190-pr3488.patch,
        * patches/hotspot/shenandoah/8187224-pr3488.patch:
        AArch64 fixes removed as applied in upstream Shenandoah.
        * patches/hotspot/shenandoah/8179084-pr3409-rh1455694.patch,
        * patches/hotspot/shenandoah/8180711.patch,
        * patches/hotspot/shenandoah/8181946.patch,
        * patches/hotspot/shenandoah/8184682.patch:
        Moved to aarch32 patches tree as applied in upstream
        Shenandoah.
        * Makefile.am:
        (ICEDTEA_PATCHES): Drop 8187224 and 8161190 as AArch64
        only and applied upstream in Shenandoah now. Move
        8179084, 8181946, 8180711 and 8184682 to AArch32 only.
        * NEWS: Updated.
        * hotspot.map.in: Bump shenandoah to
        aarch64-shenandoah-jdk8u151-b13.
        * patches/hotspot/aarch32/8179084-pr3409-rh1455694.patch,
        * patches/hotspot/aarch32/8180711.patch,
        * patches/hotspot/aarch32/8181946.patch,
        * patches/hotspot/aarch32/8184682.patch:
        Replace symlinks with full versions from Shenandoah
        patches tree.

-- 
You are receiving this mail because:
You are on the CC list for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.openjdk.java.net/pipermail/distro-pkg-dev/attachments/20171101/d0589942/attachment.html>


More information about the distro-pkg-dev mailing list