May 2016 - aarch64-port-dev

RFR: 8157708: aarch64: StrIndexOfChar intrinsic is not implemented
by Edward Nevill 06 Jul '16

06 Jul '16

Hi, Please review the following webrev http://cr.openjdk.java.net/~enevill/8157708/webrev This implements the StrIndexOfChar on aarch64. This webrev has been contributed by ningsheng.jian(a)linaro.org Tested with fastdebug jtreg hotspot and langtools. Note: That this changeset needs to be applied after the changeset to implement StrIndexOf otherwise there will be "bad AD" errors in the jtreg test run, but this is due to missing StrIndexOf implementations. All the best, Ed.

3 5

Re: [aarch64-port-dev ] RFR: 8156943: aarch64: string compare does not support CompactStrings
by Edward Nevill 05 Jul '16

05 Jul '16

On Fri, 2016-05-20 at 17:00 +0100, Andrew Haley wrote: > On 05/20/2016 04:56 PM, Edward Nevill wrote: > > fcmpd(vtmpZ, 0.0) generates EQ iff the lower 64 bits of vtmpZ are 0 > > otherwise NE. > > > > If it is the case that fcmpd generates EQ for some bit pattern other than > > all 64 bits 0 then indeed it is broken. > > There are two zeroes, +0 and -0. > > Andrew. The following webrev fixes this by instead moving the simd registers to general registers. http://cr.openjdk.java.net/~enevill/8156943/webrev.02/ The different cases are handled as follows LL: 8 chars are loaded to each of 2 general regs UU: 4 chars are loaded to each of 2 general regs LU/UL: 4 U chars loaded to 1 general regs, 4 L chars loaded to a simd reg, zip1 used to convert L to U and then moved to a general reg. Tested fastdebug jtreg hotspot and langtools, All the best, Ed.

3 10

Critical Shenandoah fix for weak reference processing
by Roman Kennke 03 Jun '16

03 Jun '16

I discovered a critical bug in Shenandoah that was introduced recently, when I rewrote the mark-compact code (used for full-gc). It did not enable weak reference discovery, and hence would never process any weak references. This patch fixes it. Please review and apply to aarch64-por t/jdk8u-shenandoah/hotspot as soon as possible. http://cr.openjdk.java.net/~rkennke/weakrefs/webrev.00/ Thanks! /Roman

2 1

Pending approvals
by Andrew Haley 02 Jun '16

02 Jun '16

I think there are none waiting for me. Please let me know if there are. Thanks, Andrew.

3 5

RFR: Backports to aarch64/jdk8u
by Felix Yang 01 Jun '16

01 Jun '16

Hi, The following are a couple patches from jdk9 which implement some basic stub routines which I think should go into aarch64/jdk8u repo. changeset: 9657:3455d28791c8 parent: 9646:8962380eb039 user: aph date: Mon Dec 14 15:53:48 2015 +0000 files: src/cpu/aarch64/vm/stubGenerator_aarch64.cpp description: 8145320: Create unsafe_arraycopy and generic_arraycopy for AArch64 Reviewed-by: kvn changeset: 10544:0d2b390d451a user: fyang date: Mon Mar 28 21:21:41 2016 +0800 files: src/cpu/aarch64/vm/stubGenerator_aarch64.cpp description: 8152840: aarch64: improve _unsafe_arraycopy stub routine Summary: aarch64: improve StubRoutines::_unsafe_arraycopy stub routine Reviewed-by: aph changeset: 10586:d4636cc092db user: enevill date: Thu Apr 07 22:36:16 2016 +0000 files: src/cpu/aarch64/vm/aarch64.ad src/cpu/aarch64/vm/macroAssembler_aarch64.cpp src/cpu/aarch64/vm/macroAssembler_aarch64.hpp src/cpu/aarch64/vm/stubGenerator_aarch64.cpp description: 8153797: aarch64: Add Arrays.fill stub code Reviewed-by: aph Contributed-by: long.chen(a)linaro.org changeset: 11301:4b829744c5c7 user: fyang date: Fri May 27 01:02:16 2016 +0800 files: src/cpu/aarch64/vm/stubGenerator_aarch64.cpp description: 8156731: aarch64: java/util/Arrays/Correct.java fails due to _generic_arraycopy stub routine Summary: fix address calculation considering compressed oops _generic_arraycopy stub routine Reviewed-by: aph Here are the links to the original jdk9 patches: http://hg.openjdk.java.net/jdk9/hs-comp/hotspot/rev/3455d28791c8 http://hg.openjdk.java.net/jdk9/hs-comp/hotspot/rev/0d2b390d451a http://hg.openjdk.java.net/jdk9/hs-comp/hotspot/rev/d4636cc092db http://hg.openjdk.java.net/jdk9/hs-comp/hotspot/rev/4b829744c5c7 And patches for aarch64/jdk8u with minor adaptations: http://cr.openjdk.java.net/~fyang/backports/8145320-aarch64-jdk8u.patch http://cr.openjdk.java.net/~fyang/backports/8152840-aarch64-jdk8u.patch http://cr.openjdk.java.net/~fyang/backports/8153797-aarch64-jdk8u.patch http://cr.openjdk.java.net/~fyang/backports/8156731-aarch64-jdk8u.patch Spark Terasort benchmark benefit from the first two patches, about 3.5%+ performance gain tested . Tested with Jtreg using an aarch64 server build, results before and after were identical. OK to backport? Felix

2 2

hg: aarch64-port/jdk8u/hotspot: 19 new changesets
by edward.nevill＠gmail.com 31 May '16

31 May '16

Changeset: 991edc81a90c Author: enevill Date: 2016-02-03 11:34 +0000 URL: http://hg.openjdk.java.net/aarch64-port/jdk8u/hotspot/rev/991edc81a90c 8148948: aarch64: generate_copy_longs calls align() incorrectly Summary: Fix alignments Reviewed-by: aph ! src/cpu/aarch64/vm/globals_aarch64.hpp ! src/cpu/aarch64/vm/stubGenerator_aarch64.cpp Changeset: f665f73d0a05 Author: enevill Date: 2016-02-08 14:14 +0000 URL: http://hg.openjdk.java.net/aarch64-port/jdk8u/hotspot/rev/f665f73d0a05 8149365: aarch64: memory copy does not prefetch on backwards copy Summary: Implement prefetch on backwards copies Reviewed-by: aph ! src/cpu/aarch64/vm/stubGenerator_aarch64.cpp ! src/cpu/aarch64/vm/vm_version_aarch64.cpp Changeset: 86fa562baa32 Author: hshi Date: 2016-02-05 03:55 -0800 URL: http://hg.openjdk.java.net/aarch64-port/jdk8u/hotspot/rev/86fa562baa32 8149080: AArch64: Recognise disjoint array copy in stub code Summary: Detect array copy can use fwd copy by checking (dest-src) above_same (copy_size) Reviewed-by: aph ! src/cpu/aarch64/vm/stubGenerator_aarch64.cpp Changeset: dd2bab0f8330 Author: enevill Date: 2016-02-20 15:11 +0000 URL: http://hg.openjdk.java.net/aarch64-port/jdk8u/hotspot/rev/dd2bab0f8330 8150082: aarch64: optimise small array copy Reviewed-by: aph ! src/cpu/aarch64/vm/stubGenerator_aarch64.cpp Changeset: a8c435282946 Author: enevill Date: 2016-02-20 15:15 +0000 URL: http://hg.openjdk.java.net/aarch64-port/jdk8u/hotspot/rev/a8c435282946 8150313: aarch64: optimise array copy using SIMD instructions Reviewed-by: aph ! src/cpu/aarch64/vm/globals_aarch64.hpp ! src/cpu/aarch64/vm/stubGenerator_aarch64.cpp Changeset: f0e398b95e52 Author: enevill Date: 2016-03-08 14:39 +0000 URL: http://hg.openjdk.java.net/aarch64-port/jdk8u/hotspot/rev/f0e398b95e52 8150394: aarch64: add support for 8.1 LSE CAS instructions Reviewed-by: aph Contributed-by: ananth.jasty(a)caviumnetworks.com, edward.nevill(a)linaro.org ! src/cpu/aarch64/vm/aarch64.ad ! src/cpu/aarch64/vm/assembler_aarch64.hpp ! src/cpu/aarch64/vm/c1_LIRAssembler_aarch64.cpp ! src/cpu/aarch64/vm/globals_aarch64.hpp ! src/cpu/aarch64/vm/macroAssembler_aarch64.cpp ! src/cpu/aarch64/vm/macroAssembler_aarch64.hpp ! src/cpu/aarch64/vm/register_aarch64.hpp ! src/cpu/aarch64/vm/vm_version_aarch64.cpp Changeset: 60796e32d8c2 Author: enevill Date: 2016-03-10 14:53 +0000 URL: http://hg.openjdk.java.net/aarch64-port/jdk8u/hotspot/rev/60796e32d8c2 8151502: optimize pd_disjoint_words and pd_conjoint_words Summary: optimize copy routines using inline assembler Reviewed-by: aph ! src/os_cpu/linux_aarch64/vm/copy_linux_aarch64.inline.hpp + src/os_cpu/linux_aarch64/vm/copy_linux_aarch64.s Changeset: 7fc149d14601 Author: enevill Date: 2016-03-29 10:07 +0000 URL: http://hg.openjdk.java.net/aarch64-port/jdk8u/hotspot/rev/7fc149d14601 8151775: aarch64: add support for 8.1 LSE atomic operations Reviewed-by: aph ! src/cpu/aarch64/vm/assembler_aarch64.hpp ! src/cpu/aarch64/vm/c1_LIRAssembler_aarch64.cpp ! src/cpu/aarch64/vm/macroAssembler_aarch64.cpp ! src/cpu/aarch64/vm/templateInterpreter_aarch64.cpp ! src/cpu/aarch64/vm/vm_version_aarch64.hpp ! src/os_cpu/linux_aarch64/vm/atomic_linux_aarch64.inline.hpp Changeset: 772a296872e1 Author: fyang Date: 2016-03-23 18:00 +0800 URL: http://hg.openjdk.java.net/aarch64-port/jdk8u/hotspot/rev/772a296872e1 8152537: aarch64: Make use of CBZ and CBNZ when comparing unsigned values with zero. Summary: aarch64: c2 make use of CBZ and CBNZ when comparing unsigned values with zero. Reviewed-by: aph ! src/cpu/aarch64/vm/aarch64.ad Changeset: 97204648fec5 Author: fyang Date: 2016-02-19 17:12 +0800 URL: http://hg.openjdk.java.net/aarch64-port/jdk8u/hotspot/rev/97204648fec5 8150229: aarch64: pipeline class for several instructions is not set correctly Summary: aarch64: c2 fix pipeline class for several instructions. Reviewed-by: aph Contributed-by: felix.yang(a)linaro.org ! src/cpu/aarch64/vm/aarch64.ad Changeset: fed8343c7a93 Author: fyang Date: 2016-02-18 21:53 +0800 URL: http://hg.openjdk.java.net/aarch64-port/jdk8u/hotspot/rev/fed8343c7a93 8149907: aarch64: use load/store pair instructions in call_stub Summary: aarch64: make use of load/store pair instructions in call_stub to save space Reviewed-by: aph Contributed-by: felix.yang(a)linaro.org ! src/cpu/aarch64/vm/stubGenerator_aarch64.cpp Changeset: 52da1053c974 Author: fyang Date: 2016-02-17 20:19 +0800 URL: http://hg.openjdk.java.net/aarch64-port/jdk8u/hotspot/rev/52da1053c974 8150038: aarch64: make use of CBZ and CBNZ when comparing narrow pointer with zero Summary: aarch64: c2 make use of CBZ and CBNZ when comparing narrow pointer with zero Reviewed-by: aph Contributed-by: felix.yang(a)linaro.org ! src/cpu/aarch64/vm/aarch64.ad Changeset: a828298c0a52 Author: fyang Date: 2016-01-27 12:20 +0800 URL: http://hg.openjdk.java.net/aarch64-port/jdk8u/hotspot/rev/a828298c0a52 8148328: aarch64: redundant lsr instructions in stub code. Summary: avoid redundant lsr instructions in jbyte_arraycopy and jbyte_disjoint_arraycopy. Reviewed-by: aph Contributed-by: felix.yang(a)linaro.org ! src/cpu/aarch64/vm/stubGenerator_aarch64.cpp Changeset: 6822ab128969 Author: enevill Date: 2016-04-07 22:36 +0000 URL: http://hg.openjdk.java.net/aarch64-port/jdk8u/hotspot/rev/6822ab128969 8153797: aarch64: Add Arrays.fill stub code Reviewed-by: aph Contributed-by: long.chen(a)linaro.org ! src/cpu/aarch64/vm/aarch64.ad ! src/cpu/aarch64/vm/macroAssembler_aarch64.cpp ! src/cpu/aarch64/vm/macroAssembler_aarch64.hpp ! src/cpu/aarch64/vm/stubGenerator_aarch64.cpp Changeset: 125cb0e9349e Author: fyang Date: 2016-03-05 22:22 +0800 URL: http://hg.openjdk.java.net/aarch64-port/jdk8u/hotspot/rev/125cb0e9349e 8151340: aarch64: prefetch the destination word for write prior to ldxr/stxr loops. Summary: aarch64: add prefetch for write prior to ldxr/stxr loops. Reviewed-by: aph Contributed-by: felix.yang(a)linaro.org ! src/cpu/aarch64/vm/aarch64.ad ! src/cpu/aarch64/vm/c1_LIRAssembler_aarch64.cpp ! src/cpu/aarch64/vm/macroAssembler_aarch64.cpp Changeset: ab888f53c092 Author: fyang Date: 2016-04-12 11:53 +0800 URL: http://hg.openjdk.java.net/aarch64-port/jdk8u/hotspot/rev/ab888f53c092 8153713: aarch64: improve short array clearing using store pair Summary: aarch64: generate store pair instruction to clear short arrays Reviewed-by: aph ! src/cpu/aarch64/vm/aarch64.ad ! src/cpu/aarch64/vm/macroAssembler_aarch64.cpp ! src/cpu/aarch64/vm/macroAssembler_aarch64.hpp Changeset: 531c3bc671ef Author: enevill Date: 2016-04-28 13:26 +0000 URL: http://hg.openjdk.java.net/aarch64-port/jdk8u/hotspot/rev/531c3bc671ef 8155617: aarch64: ClearArray does not use DC ZVA Summary: Implement block zero using DC ZVA Reviewed-by: aph Contributed-by: long.chen(a)linaro.org, edward.nevill(a)gmail.com ! src/cpu/aarch64/vm/aarch64.ad ! src/cpu/aarch64/vm/assembler_aarch64.hpp ! src/cpu/aarch64/vm/globals_aarch64.hpp ! src/cpu/aarch64/vm/macroAssembler_aarch64.cpp ! src/cpu/aarch64/vm/macroAssembler_aarch64.hpp ! src/cpu/aarch64/vm/stubGenerator_aarch64.cpp ! src/cpu/aarch64/vm/stubRoutines_aarch64.cpp ! src/cpu/aarch64/vm/stubRoutines_aarch64.hpp ! src/cpu/aarch64/vm/vm_version_aarch64.cpp ! src/cpu/aarch64/vm/vm_version_aarch64.hpp Changeset: cb418416cec2 Author: enevill Date: 2016-05-25 13:30 +0000 URL: http://hg.openjdk.java.net/aarch64-port/jdk8u/hotspot/rev/cb418416cec2 8157841: aarch64: prefetch ignores cache line size Summary: fix prefetch to take account of cache line size Reviewed-by: aph Contributed-by: stuart.monteith(a)linaro.org, edward.nevill(a)linaro.org ! src/cpu/aarch64/vm/aarch64.ad ! src/cpu/aarch64/vm/c1_LIRAssembler_aarch64.cpp ! src/cpu/aarch64/vm/macroAssembler_aarch64.cpp ! src/cpu/aarch64/vm/macroAssembler_aarch64.hpp ! src/cpu/aarch64/vm/templateInterpreter_aarch64.cpp ! src/cpu/aarch64/vm/vm_version_aarch64.cpp ! src/cpu/aarch64/vm/vm_version_aarch64.hpp Changeset: c24d5e044039 Author: fyang Date: 2016-05-27 20:38 +0800 URL: http://hg.openjdk.java.net/aarch64-port/jdk8u/hotspot/rev/c24d5e044039 8157906: aarch64: some more integer rotate instructions are never emitted Summary: fix wrong definition of source operand of left rotate instructions Reviewed-by: aph Contributed-by: teng.lu(a)linaro.org ! src/cpu/aarch64/vm/aarch64.ad

1 0

RFR: Backports to jdk8u
by Edward Nevill 31 May '16

31 May '16

Hi, Please review the following webrev:- http://cr.openjdk.java.net/~enevill/jdk8u_backports_1605 This backports the changesets listed below from jdk9 to jdk8u. Tested with jtreg hotspot & langtools. Results the same before and after. Hotspot: Test results: passed: 682; failed: 17; error: 2 Langtools: Test results: passed: 3,094 No shared code was harmed in the preparation of this webrev! OK to push? Ed. --- CUT HERE --- changeset: 8958:ae4066bd624d tag: tip user: fyang date: Fri May 27 20:38:38 2016 +0800 files: src/cpu/aarch64/vm/aarch64.ad description: 8157906: aarch64: some more integer rotate instructions are never emitted Summary: fix wrong definition of source operand of left rotate instructions Reviewed-by: aph Contributed-by: teng.lu(a)linaro.org changeset: 8957:8613e5f2c21f user: enevill date: Wed May 25 13:30:07 2016 +0000 files: src/cpu/aarch64/vm/aarch64.ad src/cpu/aarch64/vm/c1_LIRAssembler_aarch64.cpp src/cpu/aarch64/vm/mac roAssembler_aarch64.cpp src/cpu/aarch64/vm/macroAssembler_aarch64.hpp src/cpu/aarch64/vm/templateInterpreter_aar ch64.cpp src/cpu/aarch64/vm/vm_version_aarch64.cpp src/cpu/aarch64/vm/vm_version_aarch64.hpp description: 8157841: aarch64: prefetch ignores cache line size Summary: fix prefetch to take account of cache line size Reviewed-by: aph Contributed-by: stuart.monteith(a)linaro.org, edward.nevill(a)linaro.org changeset: 8956:9546e60bb6ec user: enevill date: Thu Apr 28 13:26:29 2016 +0000 files: src/cpu/aarch64/vm/aarch64.ad src/cpu/aarch64/vm/assembler_aarch64.hpp src/cpu/aarch64/vm/globals_a arch64.hpp src/cpu/aarch64/vm/macroAssembler_aarch64.cpp src/cpu/aarch64/vm/macroAssembler_aarch64.hpp src/cpu/a arch64/vm/stubGenerator_aarch64.cpp src/cpu/aarch64/vm/stubRoutines_aarch64.cpp src/cpu/aarch64/vm/stubRoutines_ aarch64.hpp src/cpu/aarch64/vm/vm_version_aarch64.cpp src/cpu/aarch64/vm/vm_version_aarch64.hpp description: 8155617: aarch64: ClearArray does not use DC ZVA Summary: Implement block zero using DC ZVA Reviewed-by: aph Contributed-by: long.chen(a)linaro.org, edward.nevill(a)gmail.com changeset: 8955:7875fdf99b33 user: fyang date: Tue Apr 12 11:53:44 2016 +0800 files: src/cpu/aarch64/vm/aarch64.ad src/cpu/aarch64/vm/macroAssembler_aarch64.cpp src/cpu/aarch64/vm/macr oAssembler_aarch64.hpp description: 8153713: aarch64: improve short array clearing using store pair Summary: aarch64: generate store pair instruction to clear short arrays Reviewed-by: aph changeset: 8954:06f81cfc8a51 user: fyang date: Sat Mar 05 22:22:37 2016 +0800 files: src/cpu/aarch64/vm/aarch64.ad src/cpu/aarch64/vm/c1_LIRAssembler_aarch64.cpp src/cpu/aarch64/vm/mac roAssembler_aarch64.cpp description: 8151340: aarch64: prefetch the destination word for write prior to ldxr/stxr loops. Summary: aarch64: add prefetch for write prior to ldxr/stxr loops. Reviewed-by: aph Contributed-by: felix.yang(a)linaro.org changeset: 8953:c24fe9ae7b70 user: enevill date: Thu Apr 07 22:36:16 2016 +0000 files: src/cpu/aarch64/vm/aarch64.ad src/cpu/aarch64/vm/macroAssembler_aarch64.cpp src/cpu/aarch64/vm/macr oAssembler_aarch64.hpp src/cpu/aarch64/vm/stubGenerator_aarch64.cpp description: 8153797: aarch64: Add Arrays.fill stub code Reviewed-by: aph Contributed-by: long.chen(a)linaro.org changeset: 8952:bbf661770808 user: fyang date: Wed Jan 27 12:20:53 2016 +0800 files: src/cpu/aarch64/vm/stubGenerator_aarch64.cpp description: 8148328: aarch64: redundant lsr instructions in stub code. Summary: avoid redundant lsr instructions in jbyte_arraycopy and jbyte_disjoint_arraycopy. Reviewed-by: aph Contributed-by: felix.yang(a)linaro.org changeset: 8951:fa4527eac528 user: fyang date: Wed Feb 17 20:19:24 2016 +0800 files: src/cpu/aarch64/vm/aarch64.ad description: 8150038: aarch64: make use of CBZ and CBNZ when comparing narrow pointer with zero Summary: aarch64: c2 make use of CBZ and CBNZ when comparing narrow pointer with zero Reviewed-by: aph Contributed-by: felix.yang(a)linaro.org changeset: 8950:f2a2314324d1 user: fyang date: Thu Feb 18 21:53:24 2016 +0800 files: src/cpu/aarch64/vm/stubGenerator_aarch64.cpp description: 8149907: aarch64: use load/store pair instructions in call_stub Summary: aarch64: make use of load/store pair instructions in call_stub to save space Reviewed-by: aph Contributed-by: felix.yang(a)linaro.org changeset: 8949:bdd5013d22ac user: fyang date: Fri Feb 19 17:12:14 2016 +0800 files: src/cpu/aarch64/vm/aarch64.ad description: 8150229: aarch64: pipeline class for several instructions is not set correctly Summary: aarch64: c2 fix pipeline class for several instructions. Reviewed-by: aph Contributed-by: felix.yang(a)linaro.org changeset: 8948:0bcef370b8f0 user: fyang date: Wed Mar 23 18:00:46 2016 +0800 files: src/cpu/aarch64/vm/aarch64.ad description: 8152537: aarch64: Make use of CBZ and CBNZ when comparing unsigned values with zero. Summary: aarch64: c2 make use of CBZ and CBNZ when comparing unsigned values with zero. Reviewed-by: aph changeset: 8947:3977fdb30508 user: enevill date: Tue Mar 29 10:07:54 2016 +0000 files: src/cpu/aarch64/vm/assembler_aarch64.hpp src/cpu/aarch64/vm/c1_LIRAssembler_aarch64.cpp src/cpu/aar ch64/vm/macroAssembler_aarch64.cpp src/cpu/aarch64/vm/templateInterpreter_aarch64.cpp src/cpu/aarch64/vm/vm_vers ion_aarch64.hpp src/os_cpu/linux_aarch64/vm/atomic_linux_aarch64.inline.hpp description: 8151775: aarch64: add support for 8.1 LSE atomic operations Reviewed-by: aph changeset: 8946:8c3bd2b4dfa1 user: enevill date: Thu Mar 10 14:53:09 2016 +0000 files: src/os_cpu/linux_aarch64/vm/copy_linux_aarch64.inline.hpp src/os_cpu/linux_aarch64/vm/copy_linux_aa rch64.s description: 8151502: optimize pd_disjoint_words and pd_conjoint_words Summary: optimize copy routines using inline assembler Reviewed-by: aph changeset: 8945:3ae4a888109c user: enevill date: Tue Mar 08 14:39:50 2016 +0000 files: src/cpu/aarch64/vm/aarch64.ad src/cpu/aarch64/vm/assembler_aarch64.hpp src/cpu/aarch64/vm/c1_LIRAss embler_aarch64.cpp src/cpu/aarch64/vm/globals_aarch64.hpp src/cpu/aarch64/vm/macroAssembler_aarch64.cpp src/cpu/ aarch64/vm/macroAssembler_aarch64.hpp src/cpu/aarch64/vm/register_aarch64.hpp src/cpu/aarch64/vm/vm_version_aarc h64.cpp description: 8150394: aarch64: add support for 8.1 LSE CAS instructions Reviewed-by: aph Contributed-by: ananth.jasty(a)caviumnetworks.com, edward.nevill(a)linaro.org changeset: 8944:8612528a4488 user: enevill date: Sat Feb 20 15:15:35 2016 +0000 files: src/cpu/aarch64/vm/globals_aarch64.hpp src/cpu/aarch64/vm/stubGenerator_aarch64.cpp description: 8150313: aarch64: optimise array copy using SIMD instructions Reviewed-by: aph changeset: 8943:31d7f80d1e61 user: enevill date: Sat Feb 20 15:11:42 2016 +0000 files: src/cpu/aarch64/vm/stubGenerator_aarch64.cpp description: 8150082: aarch64: optimise small array copy Reviewed-by: aph changeset: 8942:f8b0c08a84f6 user: hshi date: Fri Feb 05 03:55:51 2016 -0800 files: src/cpu/aarch64/vm/stubGenerator_aarch64.cpp description: 8149080: AArch64: Recognise disjoint array copy in stub code Summary: Detect array copy can use fwd copy by checking (dest-src) above_same (copy_size) Reviewed-by: aph changeset: 8941:a636a276f009 user: enevill date: Mon Feb 08 14:14:35 2016 +0000 files: src/cpu/aarch64/vm/stubGenerator_aarch64.cpp src/cpu/aarch64/vm/vm_version_aarch64.cpp description: 8149365: aarch64: memory copy does not prefetch on backwards copy Summary: Implement prefetch on backwards copies Reviewed-by: aph changeset: 8940:c8a3c4b20148 user: enevill date: Thu Feb 04 16:24:28 2016 +0000 files: src/cpu/aarch64/vm/macroAssembler_aarch64.cpp description: 8148783: aarch64: SEGV running SpecJBB2013 Summary: Fix calculation of offset for adrp Reviewed-by: aph changeset: 8939:28b78aca07ca user: enevill date: Wed Feb 03 11:34:12 2016 +0000 files: src/cpu/aarch64/vm/globals_aarch64.hpp src/cpu/aarch64/vm/stubGenerator_aarch64.cpp description: 8148948: aarch64: generate_copy_longs calls align() incorrectly Summary: Fix alignments Reviewed-by: aph --- CUT HERE ---

2 1

hg: aarch64-port/jdk8u/hotspot: 8148783: aarch64: SEGV running SpecJBB2013
by felix.yang＠linaro.org 31 May '16

31 May '16

Changeset: 1a5fa9c42359 Author: enevill Date: 2016-02-04 16:24 +0000 URL: http://hg.openjdk.java.net/aarch64-port/jdk8u/hotspot/rev/1a5fa9c42359 8148783: aarch64: SEGV running SpecJBB2013 Summary: Fix calculation of offset for adrp Reviewed-by: aph ! src/cpu/aarch64/vm/macroAssembler_aarch64.cpp

1 0

RFR: u92
by Andrew Hughes 27 May '16

27 May '16

Hi, This update synchronises aarch64/jdk8u with jdk8u92-b14 upstream and tags the result as aarch64-jdk8u92-b14: http://cr.openjdk.java.net/~andrew/aarch64-8/u92/ The webrevs are again rather hard to read. The new changes are: - S6869327: Add new C2 flag to keep safepoints in counted loops. - S8022865: [TESTBUG] Compressed Oops testing needs to be revised - S8029630: Thread id should be displayed as a hex number in error report - S8029726: On OS X some dtrace probe names are mismatched with Solaris - S8029727: On OS X dtrace probes Call<type>MethodA/Call<type>MethodV are not fired. - S8029728: On OS X dtrace probes SetStaticBooleanField are not fired - S8038184: XMLSignature throws StringIndexOutOfBoundsException if ID attribute value is empty String - S8038349: Signing XML with DSA throws Exception when key is larger than 1024 bits - S8041501: ImageIO reader is not capable of reading JPEGs without JFIF header - S8041900: [macosx] Java forces the use of discrete GPU - S8044363: Remove special build options for unpack200 executable - S8046471: Use OPENJDK_TARGET_CPU_ARCH instead of legacy value for hotspot ARCH - S8046611: Build errors with gcc on sparc/fastdebug - S8047763: Recognize sparc64 as a sparc platform - S8048232: Fix for 8046471 breaks PPC64 build - S8052396: Catch exceptions resulting from missing font cmap - S8058563: InstanceKlass::_dependencies list isn't cleared from empty nmethodBucket entries - S8061624: [TESTBUG] Some tests cannot be ran under compact profiles and therefore shall be excluded - S8062901: Iterators is spelled incorrectly in the Javadoc for Spliterator - S8064330: Remove SHA224 from the default support list if SunMSCAPI enabled - S8065579: WB method to start G1 concurrent mark cycle should be introduced - S8065986: Compiler fails to NullPointerException when calling super with Object<>() - S8066974: Compiler doesn't infer method's generic type information in lambda body - S8067800: Clarify java.time.chrono.Chronology.isLeapYear for out of range years - S8068033: JNI exception pending in jdk/src/share/bin/java.c - S8068042: Check jdk/src/share/native/sun/misc/URLClassPath.c for JNI pending - S8068162: jvmtiRedefineClasses.cpp: guarantee(false) failed: OLD and/or OBSOLETE method(s) found - S8068254: Method reference uses wrong qualifying type - S8074696: Remote debugging session hangs for several minutes when calling findBootType - S8074935: jdk8 keytool doesn't validate pem files for RFC 1421 correctness, as jdk7 did - S8078423: [TESTBUG] javax/print/PrintSEUmlauts/PrintSEUmlauts.java relies on system locale - S8080492: [Parfait] Uninitialised variable in jdk/src/java/desktop/windows/native/libawt/ - S8080650: Enable stubs to use frame pointers correctly - S8087120: [GCC5] java.lang.StackOverflowError on Zero JVM initialization on non x86 platforms. - S8122944: perfdata used is seen as too high on sparc zone with jdk1.9 and causes a test failure - S8129348: Debugger hangs in trace mode with TRACE_SENDS - S8129847: Compiling methods generated by Nashorn triggers high memory usage in C2 - S8130506: javac AssertionError when invoking MethodHandle.invoke with lambda parameter - S8130910: hsperfdata file is created in wrong directory and not cleaned up if /tmp/hsperfdata_<username> has wrong permissions - S8131129: Attempt to define a duplicate BMH$Species class - S8131665: Bad exception message in HandshakeHash.getFinishedHash - S8131782: C1 Class.cast optimization breaks when Class is loaded from static final - S8132503: [macosx] Chinese full stop symbol cannot be entered with Pinyin IM on OS X - S8132890: Text Overlapping on Dot Matrix Printers - S8133207: ParallelProbes.java test fails after changes for JDK-8080115 - S8133924: NPE may be thrown when xsltc select a non-existing node after JDK-8062518 - S8134007: Improve string folding - S8134759: jdb: Incorrect stepping inside finally block - S8134963: [Newtest] New stress test for changing the coarseness level of G1 remembered set - S8136442: Don't tie Certificate signature algorithms to ciphersuites - S8137106: EUDC (End User Defined Characters) are not displayed on Windows with Java 8u60+ - S8138745: Implement ExitOnOutOfMemory and CrashOnOutOfMemory in HotSpot - S8138764: In some cases the usage of TreeLock can be replaced by other synchronization - S8138966: Intermittent SEGV running ParallelGC - S8139258: PPC64LE: argument passing problem when passing 15 floats in native call - S8139373: [TEST_BUG] java/net/MulticastSocket/MultiDead.java failed with timeout - S8139421: PPC64LE: MacroAssembler::bxx64_patchable kill register R12 - S8139424: SIGSEGV, Problematic frame: # V [libjvm.so+0xd0c0cc] void InstanceKlass::oop_oop_iterate_oop_maps_specialized<true,oopDesc*,MarkAndPushClosure> - S8139436: sun.security.mscapi.KeyStore might load incomplete data - S8139751: Javac crash with -XDallowStringFolding=false - S8139863: [TESTBUG] Need to port tests for JDK-8134903 to 8u-dev - S8139985: JNI exception pending in jdk/src/jdk/hprof/agent/share/native/libhprof - S8140031: SA: Searching for a value in Threads does not work - S8140244: Port fix of JDK-8075773 to AIX and possibly MacOSX - S8140249: JVM Crashing During startUp If Flight Recording is enabled - S8140344: add support for 3 digit update release numbers - S8140483: Atomic*FieldUpdaters final fields should be trusted - S8140587: Atomic*FieldUpdaters should use Class.isInstance instead of direct class check - S8141260: isReachable crash in windows xp - S8143397: It looks like InetAddress.isReachable(timeout) works incorrectly - S8143855: Bad printf formatting in frame_zero.cpp - S8143963: improve ClassLoader::trace_class_path to accept an additional outputStream* arg - S8144483: One long Safepoint pause directly after each GC log rotation - S8144487: PhaseIdealLoop::build_and_optimize() must restore major_progress flag if skip_loop_opts is true - S8144885: agent/src/os/linux/libproc.h needs to support Linux/SPARC builds - S8144935: C2: safepoint is pruned from a non-counted loop - S8144937: [TEST_BUG] testlibrary_tests should be excluded for compact1 and compact2 execution - S8145017: Add support for 3 digit hotspot minor version numbers - S8145099: Better error message when SA can't attach to a process - S8145442: Add the facility to verify remembered sets for G1 - S8145466: javac: No line numbers in compilation error - S8145539: (coll) AbstractMap.keySet and .values should not be volatile - S8145722: NullPointerException in javadoc - S8145754: PhaseIdealLoop::is_scaled_iv_plus_offset() does not match AddI - S8146566: OpenJDK build can't handle commas in LDFLAGS - S8146725: Issues with SignatureAndHashAlgorithm.getSupportedAlgorithms - S8146979: Backport of 8046471 breaks ppc64 build in jdk8u because 8072383 was badly backported before - S8147087: Race when reusing PerRegionTable bitmaps may result in dropped remembered set entries - S8147857: RMIConnector logs attribute names incorrectly - S8148353: [linux-sparc] Crash in libawt.so on Linux SPARC - S8150791: 8u76 L10n resource file translation update Ok to push? -- Andrew :) Senior Free Java Software Engineer Red Hat, Inc. (http://www.redhat.com) PGP Key: ed25519/35964222 (hkp://keys.gnupg.net) Fingerprint = 5132 579D D154 0ED2 3E04 C5A0 CFDA 0F9B 3596 4222

2 1

RFR: 8157906: aarch64: some more integer rotate instructions are never emitted
by Felix Yang 26 May '16

26 May '16

Hi, Please review webrev: http://cr.openjdk.java.net/~fyang/8157906/webrev.00 JIRA issue: https://bugs.openjdk.java.net/browse/JDK-8157906 This patch has been contributed by teng.lu(a)linaro.org The left rotate instructions have the same issue with https://bugs.openjdk.java.net/browse/JDK-8154537. (src << shift) | src >>> -shift) with src an int have some support in the aarch64.ad ad file: rolI_rReg_Var_C_32 and rolI_rReg_Var_C0, but their definitions is broken and never match any ideal graph subtree. Tested with fastdebug jtreg hotspot and langtools, OK to push? Simple test case: import java.lang.Integer.*; public class test { public static int foo (int x) { return Integer.rotateLeft(x, 1); } public static void main(String[] args) { int x = 0x1234; for (int i = 0; i < 50000; i++) { x = foo(x); } } } JIT code snippet before change: 0x0000007f74d13454: neg w11, w2 0x0000007f74d13458: lsr w12, w1, w11 0x0000007f74d1345c: lsl w10, w1, w2 0x0000007f74d13460: orr w0, w10, w12 ;*ior {reexecute=0 rethrow=0 return_oop=0} ; - java.lang.Integer::rotateLeft@7 (line 1751) JIT code snippet after change: 0x0000007f6cd135d4: neg w8, w2 0x0000007f6cd135d8: ror w0, w1, w8 ;*ior {reexecute=0 rethrow=0 return_oop=0} ; - java.lang.Integer::rotateLeft@7 (line 1751)

2 1