July 2017 - aarch64-port-dev

Re: [aarch64-port-dev ] RFR(S): 8154537: AArch64: some integer rotate instructions are never emitted
by Andrew Haley 05 Sep '17

05 Sep '17

I'm looking at the webrev in http://cr.openjdk.java.net/~roland/8154537/webrev.00/ and I see the the changes were made to aarch64.ad but not to ad_aarch64.m4. This is problematic because some .m4 files are used to generate the .ad file, and if anyone regenerates the .ad file the bug will regress. I think this is the change we need to make. It won't affect generated code at all, but it is something of a ticking bomb. diff -r 214a94e9366c src/cpu/aarch64/vm/aarch64_ad.m4 --- a/src/cpu/aarch64/vm/aarch64_ad.m4 Mon Jul 17 12:11:32 2017 +0000 +++ b/src/cpu/aarch64/vm/aarch64_ad.m4 Fri Jul 28 18:57:25 2017 +0100 @@ -268,21 +268,21 @@ ins_pipe(ialu_reg_reg_vshift); %}')dnl define(ROL_INSN, ` -instruct $3$1_rReg_Var_C$2(iRegLNoSp dst, iRegL src, iRegI shift, immI$2 c$2, rFlagsReg cr) +instruct $3$1_rReg_Var_C$2(iReg$1NoSp dst, iReg$1 src, iRegI shift, immI$2 c$2, rFlagsReg cr) %{ match(Set dst (Or$1 (LShift$1 src shift) (URShift$1 src (SubI c$2 shift)))); expand %{ - $3L_rReg(dst, src, shift, cr); + $3$1_rReg(dst, src, shift, cr); %} %}')dnl define(ROR_INSN, ` -instruct $3$1_rReg_Var_C$2(iRegLNoSp dst, iRegL src, iRegI shift, immI$2 c$2, rFlagsReg cr) +instruct $3$1_rReg_Var_C$2(iReg$1NoSp dst, iReg$1 src, iRegI shift, immI$2 c$2, rFlagsReg cr) %{ match(Set dst (Or$1 (URShift$1 src shift) (LShift$1 src (SubI c$2 shift)))); expand %{ - $3L_rReg(dst, src, shift, cr); + $3$1_rReg(dst, src, shift, cr); %} %}')dnl ROL_EXPAND(L, rol, rorv) -- Andrew Haley Java Platform Lead Engineer Red Hat UK Ltd. <https://www.redhat.com> EAC8 43EB D3EF DB98 CC77 2FAD A5CD 6035 332F A671

2 4

[aarch64-port-dev] [10] RFR: 8184049 : Matching rule for ubfiz
by Daniel Stewart 28 Aug '17

28 Aug '17

Please see the webrev [1] for adding some matching rules to ubfiz [2]. [1]: http://cr.openjdk.java.net/~njian/8184049/webrev.00/ [2]: https://bugs.openjdk.java.net/browse/JDK-8184049 Thanks, Daniel Stewart -- Daniel Stewart

5 25

[OpenJDK8u] java.nio.Bits.unaligned() doesn't handle aarch64
by White, Derek 14 Aug '17

14 Aug '17

In aarch64-port/jdk8u, java.nio.Bits.unaligned() only returns true for x86(s). This ends up being important because Apache Spark uses Bits.unaligned() to decide if it can use off-heap memory or not. In mainline jdk8u, this was updated to include PPC as well (See https://bugs.openjdk.java.net/browse/JDK-8165231) JDK9 has a much better fix that supports aarch64 already. Tested on Spark and off-heap memory was being used. Note: The PPC fix also patched sun/security/provider/ByteArrayAccess.java, but I didn't include this in the patch below (not sure if it matters, or how to test it). This is my first attempt at an aarch64-jdk8u patch. Please let me know if I'm not following protocol correctly. Thanks, - Derek Patch follows (in jdk/src). --- old/src/share/classes/java/nio/Bits.java 2017-05-11 13:10:31.916219496 -0400 +++ new/src/share/classes/java/nio/Bits.java 2017-05-11 13:10:31.832221728 -0400 @@ -614,7 +614,9 @@ String arch = AccessController.doPrivileged( new sun.security.action.GetPropertyAction("os.arch")); unaligned = arch.equals("i386") || arch.equals("x86") - || arch.equals("amd64") || arch.equals("x86_64"); + || arch.equals("amd64") || arch.equals("x86_64") + || arch.equals("ppc64") || arch.equals("ppc64le") + || arch.equals("aarch64"); unalignedKnown = true; return unaligned; }

3 13

[10] RFR(S): JDK-8184943: AARCH64: Intrinsify hasNegatives
by Dmitrij Pochepko 14 Aug '17

14 Aug '17

Hi everyone, Please review this small webrev [1] that implements an enhancement [2] which adds has_negatives intrinsic to AARCH64 OpenJDK port. This intrinsic performs better than c2-compiled code for every array size tried: ThunderX T88: about 2% for array size = 1 and up to 8.5x for large arrays Cortex A53(R-Pi): shows about the same numbers(really large sizes can't be normally tested there due to small amount of available memory). Intrinsified HasNegatives method checks if provided byte array has any byte with negative value(higher bit set) and intrinsic in general do as following(with various minor optimizations): 1) check array length variable to have lower bits set (0x1, 0x2, 0x4, 0x8) and invoke respective load instruction(ldrb, ldrh, ldrw, ldr) while reducing remaining length variable respectively. So, remaining length is 16*N after this code. Proceed to 2). 2) in case remaining length >= 64, loads data in a loop with 4 ldp instructions(16 bytes each) and invoking prfm (prefetch hint) in case SoftwarePrefetchHintDistance >= 0 once per loop. This new flag (SoftwarePrefetchHintDistance) is introduced to provide configurable software prefetching in dynamically compiled code. This flag can disable software prefetch hint or set prefetch distance. Default distance is set to 3 * dcache_line which shows best performance on armv8 CPUs we have. 64-bytes loop proceed until length < 64, then, proceed to 3). 3) simple 16-byte loading loop until remaining length is 0. Note: It was observed that software prefetching hint improves performance for platforms that do not have hardware prefetching (ThunderX T88), but also for platforms we have in hand which do have hardware prefetching (Cortex A53). Performance testing: JMH-based microbenchmark was developed [3] to test the performance of this enhancement. The performance results on Cortex A53 [4] and ThunderX T88 [5] for this intrinsic are on-par with C2-compiled java code for very small strings and improve the performance with the increase in string length starting from string length of 3 and up to 8x for long strings. Functional testing: Tested by running hotspot jtreg tests on Cortex A53 and ThunderX T88 and comparing the test results diff with vanilla build. No regressions were observed. Specifically, test hotspot/test/compiler/intrinsics/string/TestHasNegatives.java passed on both Cortex A53 and ThunderX T88. [1] webrev: http://cr.openjdk.java.net/~dpochepk/8184943/webrev.01/ [2] CR: https://bugs.openjdk.java.net/browse/JDK-8184943 [3] JMH micro benchmark: http://cr.openjdk.java.net/~dpochepk/8184943/HasNegativesBenchmark/ [4] A53 graph: http://cr.openjdk.java.net/~dpochepk/8184943/Cortex_A53_comparison.png [5] T88 graph: http://cr.openjdk.java.net/~dpochepk/8184943/ThunderX_comparison.png I'll be happy to merge suggestions for improvement of this intrinsic should they come into this review. Thanks, Dmitrij

5 16

Backports from jdk9 to aarch64-port/jdk8u
by Stuart Monteith 11 Aug '17

11 Aug '17

Hi, I've been working on some aarch64 patches that should be backported from jdk9 to aarch64-prot/jdk8u. I've backported the following stack of patches, and I'm currently testing with them. Do these look sensible? Are there other I should consider, that are not specifically AArch64? I've put in 8175887 as it obviously affects JCStress. I've excluded patches that are obviously not applicable to JDK8u - Graal, for example. 8175887: C1 value numbering handling of Unsafe.get*Volatile is incorrect https://bugs.openjdk.java.net/browse/JDK-8175887 http://hg.openjdk.java.net/jdk9/dev/hotspot/rev/2ff05d967fb2 This isn't strictly an aarch64 issue, but is causing JCStress to fail. 8173472: AArch64: C1 comparisons with null only use 32-bit instructions https://bugs.openjdk.java.net/browse/JDK-8173472 http://hg.openjdk.java.net/jdk9/dev/hotspot/rev/b53b0251e250 This is a fix for a serious bug that was causing spurious failures. 8172881: AArch64: assertion failure: the int pressure is incorrect https://bugs.openjdk.java.net/browse/JDK-8172881 http://hg.openjdk.java.net/jdk9/dev/hotspot/rev/bfa8e4b0d4e2 Was causing a SIGSEGV. 8171537: aarch64: compiler/c1/Test6849574.java generates guarantee failure in C1 https://bugs.openjdk.java.net/browse/JDK-8171537 http://hg.openjdk.java.net/jdk9/dev/hotspot/rev/a3bd5804b4be 8171410: aarch64: long multiplyExact shifts by 31 instead of 63 https://bugs.openjdk.java.net/browse/JDK-8171410 http://hg.openjdk.java.net/jdk9/dev/hotspot/rev/da52f1047024 Fixes a bug that was causing a slow path to be executed by mistake. Effectively a performance improvement. 8171129: [aarch64] hs_err logs do not print register mappings https://bugs.openjdk.java.net/browse/JDK-8171129 http://hg.openjdk.java.net/jdk9/dev/hotspot/rev/f3a31c992164 This is a nice-to-have and could be excluded. 8169177: AArch64: SIGSEGV when "-XX:+ZeroTLAB" is specified along with GC options https://bugs.openjdk.java.net/browse/JDK-8169177 http://hg.openjdk.java.net/jdk9/dev/hotspot/rev/9831cd4dd428 8170100: AArch64: Crash in C1-compiled code accessing References https://bugs.openjdk.java.net/browse/JDK-8170100 http://hg.openjdk.java.net/jdk9/dev/hotspot/rev/52d18f20804b 8169901: AArch64: CompareAndExchange intrinsics clobber address register https://bugs.openjdk.java.net/browse/JDK-8169901 http://hg.openjdk.java.net/jdk9/dev/hotspot/rev/f5689e544d44 8169497: Aarch64: Improve internal array handling https://bugs.openjdk.java.net/browse/JDK-8169497 http://hg.openjdk.java.net/jdk9/dev/hotspot/rev/824a8dcba48b I'm currently testing the stack of patches. Shall I post an email with the patch series? It is 800 lines in length. There was a little fixup required for some of the patches. BR, Stuart

3 6

Re: [aarch64-port-dev ] Fireside chat - Thursday 13th July
by Stuart Monteith 01 Aug '17

01 Aug '17

Ok everyone, I've scheduled the chat for 16:00 to 16:45 today (28th July). The details are unchanged: To join in, please join the chat in Bluejeans at the following URL: https://bluejeans.com/791239268 In order to join in, please join the chat at Bluejeans here: https://bluejeans.com/791239268 Alternatively you may dial in using one of the following numbers: http://bluejeans.com/numbers and enter the Meeting ID: 791239268 Please do not use any of the 'freefone' numbers, because although they may be free for you they cost us $$$$. BR, Stuart On 27 July 2017 at 14:38, Jones, Joel <Joel.Jones(a)cavium.com> wrote: > Friday I won't be able to attend, but that's OK, because my train is running > late this morning and I'd miss part or all of the meeting this morning > anyway. > > Get Outlook for iOS > ________________________________ > From: aarch64-port-dev <aarch64-port-dev-bounces(a)openjdk.java.net> on behalf > of Andrew Haley <aph(a)redhat.com> > Sent: Thursday, July 27, 2017 1:10:46 AM > To: White, Derek; Stuart Monteith; aarch64-port-dev(a)openjdk.java.net > Subject: Re: [aarch64-port-dev ] Fireside chat - Thursday 13th July > > On 27/07/17 05:30, White, Derek wrote: >> Assuming you meant 27th July -> 28th July, yes, fine with me. > > Sure. > > -- > Andrew Haley > Java Platform Lead Engineer > Red Hat UK Ltd. <https://www.redhat.com> > EAC8 43EB D3EF DB98 CC77 2FAD A5CD 6035 332F A671

3 3

hg: aarch64-port/jdk8u/hotspot: 8161190: AArch64: Fix overflow in immediate cmp instruction
by felix.yang＠linaro.org 28 Jul '17

28 Jul '17

Changeset: 8aae9841c393 Author: aph Date: 2016-07-22 17:05 +0100 URL: http://hg.openjdk.java.net/aarch64-port/jdk8u/hotspot/rev/8aae9841c393 8161190: AArch64: Fix overflow in immediate cmp instruction Summary: Use subs instead of cmp to compare BlockZeroingLowLimit. Subs can check and handle immediate out of range correctly. Reviewed-by: aph Contributed-by: yang.zhang(a)linaro.org ! src/cpu/aarch64/vm/macroAssembler_aarch64.cpp ! src/cpu/aarch64/vm/stubGenerator_aarch64.cpp

1 0

hg: aarch64-port/jdk8u-shenandoah/hotspot: 2 new changesets
by roman＠kennke.org 27 Jul '17

27 Jul '17

Changeset: 26bb4cd21d34 Author: rkennke Date: 2017-07-27 15:51 +0000 URL: http://hg.openjdk.java.net/aarch64-port/jdk8u-shenandoah/hotspot/rev/26bb4c… Add missing cmpoops() declaration to AArch64 macro assembler. Back out matrix related code from AArch64 interpreter. ! src/cpu/aarch64/vm/macroAssembler_aarch64.cpp ! src/cpu/aarch64/vm/macroAssembler_aarch64.hpp ! src/cpu/aarch64/vm/templateTable_aarch64.cpp Changeset: 7898e9f14077 Author: rkennke Date: 2017-07-27 17:56 +0200 URL: http://hg.openjdk.java.net/aarch64-port/jdk8u-shenandoah/hotspot/rev/7898e9… Added tag aarch64-shenandoah-jdk8u141-b16-shenandoah-merge-2017-07-27-02 for changeset 26bb4cd21d34 ! .hgtags

1 0

hg: aarch64-port/jdk8u-shenandoah/langtools: Added tag aarch64-shenandoah-jdk8u141-b16-shenandoah-merge-2017-07-27-02 for changeset 870e665267d6
by roman＠kennke.org 27 Jul '17

27 Jul '17

Changeset: 9083934e7562 Author: rkennke Date: 2017-07-27 17:56 +0200 URL: http://hg.openjdk.java.net/aarch64-port/jdk8u-shenandoah/langtools/rev/9083… Added tag aarch64-shenandoah-jdk8u141-b16-shenandoah-merge-2017-07-27-02 for changeset 870e665267d6 ! .hgtags

1 0

hg: aarch64-port/jdk8u-shenandoah/jaxp: Added tag aarch64-shenandoah-jdk8u141-b16-shenandoah-merge-2017-07-27-02 for changeset 2a7a979a243d
by roman＠kennke.org 27 Jul '17

27 Jul '17

Changeset: 36ff6751b2fe Author: rkennke Date: 2017-07-27 17:56 +0200 URL: http://hg.openjdk.java.net/aarch64-port/jdk8u-shenandoah/jaxp/rev/36ff6751b… Added tag aarch64-shenandoah-jdk8u141-b16-shenandoah-merge-2017-07-27-02 for changeset 2a7a979a243d ! .hgtags

1 0