October 2013 - aarch64-port-dev

printf format conflict
by Andy Johnson 05 Nov '13

05 Nov '13

I got an error while doing a sim_compile. It complained about a type conflict between the argument and the corresponding format. The following patch solves the problem: # HG changeset patch # User andy johnson andy.johnson(a)linaro.org # Date 1383091105 0 # Node ID 88d8843ae27ffb6f66baf16dfc9b3b34775b91c9 # Parent 2c9491db660f51559026478f7053a0e1054cb277 Fixed conflicting types in printf format argument. diff -r 2c9491db660f -r 88d8843ae27f src/cpu/aarch64/vm/frame_aarch64.cpp --- a/src/cpu/aarch64/vm/frame_aarch64.cpp Tue Oct 29 16:06:56 2013 +0000 +++ b/src/cpu/aarch64/vm/frame_aarch64.cpp Tue Oct 29 23:58:25 2013 +0000 @@ -713,7 +713,7 @@ #define DESCRIBE_FP_OFFSET(name) \ { \ unsigned long *p = (unsigned long *)fp; \ - printf("0x%016lx 0x%016lx %s\n", p + frame::name##_offset, \ + printf("0x%016lx 0x%016lx %s\n", (unsigned long)(p + frame::name##_offset),\ p[frame::name##_offset], #name); \ }

4 11

hg: aarch64-port/jdk8/hotspot: 20 new changesets
by aph＠redhat.com 04 Nov '13

04 Nov '13

Changeset: 7bf52c3b41bf Author: aph Date: 2013-10-16 17:06 +0100 URL: http://hg.openjdk.java.net/aarch64-port/jdk8/hotspot/rev/7bf52c3b41bf Add/subtract (shifted regsiter) and Logical (shifted regsiter) ! make/linux/makefiles/adlc.make ! src/cpu/aarch64/vm/aarch64.ad + src/cpu/aarch64/vm/aarch64_ad.m4 Changeset: effb95dde9f8 Author: aph Date: 2013-10-17 13:58 +0100 URL: http://hg.openjdk.java.net/aarch64-port/jdk8/hotspot/rev/effb95dde9f8 Bitfield moves and rotations ! src/cpu/aarch64/vm/aarch64.ad ! src/cpu/aarch64/vm/aarch64_ad.m4 ! src/cpu/aarch64/vm/assembler_aarch64.hpp Changeset: 03f55b66e8cf Author: aph Date: 2013-10-17 16:33 +0100 URL: http://hg.openjdk.java.net/aarch64-port/jdk8/hotspot/rev/03f55b66e8cf Merge - agent/src/share/classes/sun/jvm/hotspot/debugger/cdbg/basic/amd64/AMD64CFrame.java - agent/src/share/classes/sun/jvm/hotspot/debugger/cdbg/basic/x86/X86CFrame.java - make/bsd/makefiles/launcher.make ! make/linux/makefiles/adlc.make - make/linux/makefiles/launcher.make - make/solaris/makefiles/launcher.make - make/windows/makefiles/launcher.make ! src/cpu/aarch64/vm/aarch64.ad ! src/cpu/aarch64/vm/macroAssembler_aarch64.cpp ! src/cpu/aarch64/vm/macroAssembler_aarch64.hpp - src/os/posix/launcher/java_md.c - src/os/posix/launcher/java_md.h - src/os/posix/launcher/launcher.script - src/os/windows/launcher/java_md.c - src/os/windows/launcher/java_md.h - src/os_cpu/bsd_x86/vm/bsd_x86_32.ad - src/os_cpu/bsd_x86/vm/bsd_x86_64.ad - src/os_cpu/linux_sparc/vm/assembler_linux_sparc.cpp - src/os_cpu/linux_x86/vm/linux_x86_32.ad - src/os_cpu/linux_x86/vm/linux_x86_64.ad - src/os_cpu/solaris_sparc/vm/assembler_solaris_sparc.cpp - src/os_cpu/solaris_sparc/vm/solaris_sparc.ad - src/os_cpu/solaris_x86/vm/solaris_x86_32.ad - src/os_cpu/solaris_x86/vm/solaris_x86_64.ad - src/os_cpu/windows_x86/vm/windows_x86_32.ad - src/os_cpu/windows_x86/vm/windows_x86_64.ad - src/share/tools/launcher/java.c - src/share/tools/launcher/java.h - src/share/tools/launcher/jli_util.c - src/share/tools/launcher/jli_util.h - src/share/tools/launcher/wildcard.c - src/share/tools/launcher/wildcard.h - src/share/vm/classfile/genericSignatures.cpp - src/share/vm/classfile/genericSignatures.hpp - src/share/vm/gc_implementation/parallelScavenge/parMarkBitMap.inline.hpp - src/share/vm/runtime/aprofiler.cpp - src/share/vm/runtime/aprofiler.hpp - src/share/vm/trace/traceEventTypes.hpp - test/gc/metaspace/ClassMetaspaceSizeInJmapHeap.java - test/runtime/6878713/Test6878713.sh - test/runtime/6878713/testcase.jar - test/runtime/7020373/Test7020373.sh - test/runtime/7020373/testcase.jar - test/runtime/7051189/Xchecksig.sh - test/runtime/7158804/Test7158804.sh - test/runtime/7196045/Test7196045.java - test/runtime/8000968/Test8000968.sh - test/runtime/8003985/Test8003985.java Changeset: 7ad16844c57e Author: aph Date: 2013-10-17 16:46 +0100 URL: http://hg.openjdk.java.net/aarch64-port/jdk8/hotspot/rev/7ad16844c57e Use XOR encoding for compressed class pointers. ! src/cpu/aarch64/vm/macroAssembler_aarch64.cpp ! src/cpu/aarch64/vm/macroAssembler_aarch64.hpp Changeset: 0fdfde044549 Author: aph Date: 2013-10-17 18:49 +0100 URL: http://hg.openjdk.java.net/aarch64-port/jdk8/hotspot/rev/0fdfde044549 Restore LR for use by C2 runtime ! src/cpu/aarch64/vm/c1_Runtime1_aarch64.cpp Changeset: 78123d7fe1c5 Author: aph Date: 2013-10-22 10:21 +0100 URL: http://hg.openjdk.java.net/aarch64-port/jdk8/hotspot/rev/78123d7fe1c5 Fix offset overflows in profile counters. ! src/cpu/aarch64/vm/c1_LIRAssembler_aarch64.cpp Changeset: c9d218a8476d Author: aph Date: 2013-10-22 10:22 +0100 URL: http://hg.openjdk.java.net/aarch64-port/jdk8/hotspot/rev/c9d218a8476d Fix offset overflow in addptr. ! src/cpu/aarch64/vm/macroAssembler_aarch64.hpp Changeset: 919191537d1c Author: aph Date: 2013-10-22 10:22 +0100 URL: http://hg.openjdk.java.net/aarch64-port/jdk8/hotspot/rev/919191537d1c More shift/add/mask patterns. ! src/cpu/aarch64/vm/aarch64.ad ! src/cpu/aarch64/vm/aarch64_ad.m4 Changeset: c5a148cea134 Author: aph Date: 2013-10-22 18:02 +0100 URL: http://hg.openjdk.java.net/aarch64-port/jdk8/hotspot/rev/c5a148cea134 Move aarch64_enc_save_pc ! src/cpu/aarch64/vm/aarch64.ad Changeset: a669c60b445e Author: aph Date: 2013-10-22 18:37 +0100 URL: http://hg.openjdk.java.net/aarch64-port/jdk8/hotspot/rev/a669c60b445e Whitespace and minor name changes only. ! src/cpu/aarch64/vm/aarch64.ad Changeset: 3a03874840c6 Author: aph Date: 2013-10-22 18:52 +0100 URL: http://hg.openjdk.java.net/aarch64-port/jdk8/hotspot/rev/3a03874840c6 Add indIndexScaledOffsetI2L addessing mode ! src/cpu/aarch64/vm/aarch64.ad Changeset: b962569eb666 Author: aph Date: 2013-10-24 14:05 +0100 URL: http://hg.openjdk.java.net/aarch64-port/jdk8/hotspot/rev/b962569eb666 Temporary commit before reorganaization. ! make/linux/makefiles/adlc.make ! src/cpu/aarch64/vm/aarch64.ad ! src/cpu/aarch64/vm/aarch64_ad.m4 Changeset: 92b2e1520bcb Author: aph Date: 2013-10-25 11:22 +0100 URL: http://hg.openjdk.java.net/aarch64-port/jdk8/hotspot/rev/92b2e1520bcb Record work in progress ! src/cpu/aarch64/vm/aarch64.ad ! src/cpu/aarch64/vm/stubGenerator_aarch64.cpp Changeset: d29d243a63e1 Author: aph Date: 2013-10-25 12:09 +0100 URL: http://hg.openjdk.java.net/aarch64-port/jdk8/hotspot/rev/d29d243a63e1 Record work in progress ! src/cpu/aarch64/vm/aarch64.ad Changeset: 543d6fc7ce4a Author: aph Date: 2013-10-25 14:21 +0100 URL: http://hg.openjdk.java.net/aarch64-port/jdk8/hotspot/rev/543d6fc7ce4a Record work in progress ! src/cpu/aarch64/vm/aarch64.ad + src/cpu/aarch64/vm/ad_encode.m4 Changeset: 0e127d01f89a Author: aph Date: 2013-10-25 16:27 +0100 URL: http://hg.openjdk.java.net/aarch64-port/jdk8/hotspot/rev/0e127d01f89a Merge ! .hgtags ! src/cpu/aarch64/vm/aarch64.ad ! src/cpu/aarch64/vm/c2_globals_aarch64.hpp Changeset: d4c02a1427a6 Author: aph Date: 2013-10-25 16:31 +0100 URL: http://hg.openjdk.java.net/aarch64-port/jdk8/hotspot/rev/d4c02a1427a6 Merge ! src/cpu/aarch64/vm/aarch64.ad ! src/cpu/aarch64/vm/macroAssembler_aarch64.cpp ! src/cpu/aarch64/vm/macroAssembler_aarch64.hpp Changeset: 55c4c3af794e Author: aph Date: 2013-10-25 18:36 +0100 URL: http://hg.openjdk.java.net/aarch64-port/jdk8/hotspot/rev/55c4c3af794e Add a bunch of missing JVMTI methods. ! src/cpu/aarch64/vm/interp_masm_aarch64.cpp ! src/cpu/aarch64/vm/templateTable_aarch64.cpp Changeset: 2d0b9fa2cd76 Author: aph Date: 2013-10-29 14:12 +0000 URL: http://hg.openjdk.java.net/aarch64-port/jdk8/hotspot/rev/2d0b9fa2cd76 Variable rotate instructions ! src/cpu/aarch64/vm/aarch64.ad ! src/cpu/aarch64/vm/aarch64_ad.m4 Changeset: 2c9491db660f Author: aph Date: 2013-10-29 16:06 +0000 URL: http://hg.openjdk.java.net/aarch64-port/jdk8/hotspot/rev/2c9491db660f merge ! src/cpu/aarch64/vm/aarch64.ad

2 1

Building with BUILTIN_SIM generates SIGSEGV when running java -version
by Andrew McDermott 31 Oct '13

31 Oct '13

Hi, I'm building the tip of aarch64-port via the sim_configure/sim_compile scripts but running `java -version' leads to a crash[1]. The error stems from an assert() in: jdk8/hotspot/src/share/vm/memory/allocation.cpp:692 #ifndef ALLOW_OPERATOR_NEW_USAGE void* operator new(size_t size) throw() { assert(false, "Should not call global operator new"); return 0; } Digging a little deeper I see that in: jdk8/hotspot/make/linux/makefiles/aarch64.make we have: ifeq ($(BUILTIN_SIM), true) CFLAGS += -DBUILTIN_SIM #CFLAGS += -DBUILTIN_SIM -DALLOW_OPERATOR_NEW_USAGE endif Is there any reason why the "-DALLOW_OPERATOR_NEW_USAGE" is not enabled by default. If I add this then the SEGV goes away (obviously!). Or, if this is deliberately disabled, is there any value if I create a patch which replaces uses of 'new char[]' with, say, class stringStream in parts of the builtin simulator code? [1] java -version # To suppress the following error report, specify this argument # after -XX: or in .hotspotrc: SuppressErrorAt=/allocation.cpp:697 # # A fatal error has been detected by the Java Runtime Environment: # # Internal Error (/home/aim/jdk8/hotspot/src/share/vm/memory/allocation.cpp:697), pid=12534, tid=140431156107008 # assert(false) failed: Should not call global operator new[] # # JRE version: (8.0) (build ) # Java VM: OpenJDK 64-Bit Client VM (25.0-b52-debug mixed mode linux-aarch64 ) # Failed to write core dump. Core dumps have been disabled. To enable core dumping, try "ulimit -c unlimited" before starting Java again # # An error report file with more information is saved as: # /home/aim/jdk8/hs_err_pid12534.log # # If you would like to submit a bug report, please visit: # http://bugreport.sun.com/bugreport/crash.jsp # Current thread is 140431156107008 Dumping core ... LOOPING... -- andy

1 1

hg: aarch64-port/jdk8/hotspot: Added missing AbsD and AbsF instructions
by adinn＠redhat.com 31 Oct '13

31 Oct '13

Changeset: 4de6d9b68bc3 Author: adinn Date: 2013-10-31 10:52 +0000 URL: http://hg.openjdk.java.net/aarch64-port/jdk8/hotspot/rev/4de6d9b68bc3 Added missing AbsD and AbsF instructions ! src/cpu/aarch64/vm/aarch64.ad

1 0

Clone shifts on addressing modes
by Andrew Haley 31 Oct '13

31 Oct '13

This patch improves code for loads and stores, turning this: sxtw x10, w2 lsl x10, x10, #1 add xmethod, x4, x10 ldrsh w13, [xmethod,#24] ;*saload cmp w2, w11 b.cs 0x00007fffed199480 ;*caload add x10, x5, x10 ldrh w10, [x10,#24] into this: add xscratch1, x4, #0x18 ldrsh w13, [xscratch1,w2,sxtw #1] ;*saload cmp w2, w11 b.cs 0x00007fffed199478 ;*caload add xscratch1, x5, #0x18 ldrh w10, [xscratch1,w2,sxtw #1] Andrew. # HG changeset patch # User aph # Date 1383155220 0 # Node ID b92e284531b2d7aa6bc8b94405a5488a28b33f4f # Parent ed0fe5cfe272133e22484d7c99375e1099e4cabb Clone shifts on addressing modes diff -r ed0fe5cfe272 -r b92e284531b2 src/cpu/aarch64/vm/aarch64.ad --- a/src/cpu/aarch64/vm/aarch64.ad Wed Oct 30 08:47:21 2013 +0000 +++ b/src/cpu/aarch64/vm/aarch64.ad Wed Oct 30 17:47:00 2013 +0000 @@ -1591,7 +1591,7 @@ // Should the Matcher clone shifts on addressing modes, expecting them // to be subsumed into complex addressing expressions or compute them // into registers? True for Intel but false for most RISCs -const bool Matcher::clone_shift_expressions = false; +const bool Matcher::clone_shift_expressions = true; // Do we need to mask the count passed to shift instructions or does // the cpu only look at the lower 5/6 bits anyway?

1 1

hg: aarch64-port/jdk8/hotspot: Fixed conflicting types in printf format argument.
by aph＠redhat.com 30 Oct '13

30 Oct '13

Changeset: 3857ada3f2e9 Author: andy johnson andy.johnson(a)linaro.org Date: 2013-10-30 18:06 +0000 URL: http://hg.openjdk.java.net/aarch64-port/jdk8/hotspot/rev/3857ada3f2e9 Fixed conflicting types in printf format argument. ! src/cpu/aarch64/vm/frame_aarch64.cpp

1 0

hg: aarch64-port/jdk8/hotspot: Clone shifts on addressing modes
by aph＠redhat.com 30 Oct '13

30 Oct '13

Changeset: b92e284531b2 Author: aph Date: 2013-10-30 17:47 +0000 URL: http://hg.openjdk.java.net/aarch64-port/jdk8/hotspot/rev/b92e284531b2 Clone shifts on addressing modes ! src/cpu/aarch64/vm/aarch64.ad

1 0

hg: aarch64-port/jdk8/hotspot: Fix Add/subtract (extended) pattern which did not match anything
by aph＠redhat.com 30 Oct '13

30 Oct '13

Changeset: ed0fe5cfe272 Author: aph Date: 2013-10-30 08:47 +0000 URL: http://hg.openjdk.java.net/aarch64-port/jdk8/hotspot/rev/ed0fe5cfe272 Fix Add/subtract (extended) pattern which did not match anything ! src/cpu/aarch64/vm/aarch64.ad ! src/cpu/aarch64/vm/aarch64_ad.m4

1 0

Fixed register pressure problem at last!
by Andrew Dinn 29 Oct '13

29 Oct '13

I finally worked out what was causing the register pressure problem which was causing eclipse to break. The root problem was to do with a few of the instructions definitions which were still declaring their output register dst to be of type iRegX instead of iRegXNoSp (where X is one of P, L or I). The specific instruction which was causing the problem was ConvI2B but there were about half a dozen others which also needed to be corrected. This manifests as a register pressure error during construction of the physical interference graph in the situation where the instruction produces a live range which crosses a callout to the VM. The callout is followed by a fat projection (special type of MachProj node) which is meant to kill all the save-on-call (SOC) registers (i.e. cause them to be spilled). Since the live range built by the instruction includes R0-R31, which are no-save (NS), the kill merely reduces it to a live range containing R27-R31 rather than an empty range. The register pressure count for the block is reduced by 1 after the fat projection is processed but the register pressure count associated with the live ranges remains as it was. Luckily an assert detects this disparity. I am continuing to run Eclipse on beyond this point. Perhaps we could consider including Andrew's optimizations now? That would at least mean that I hit Eclipse problems quicker (since it substantially reduces the instruction count for the compiled code). regards, Andrew Dinn -----------

1 0

hg: aarch64-port/jdk8/hotspot: corrected register match rules, ensured rule outputs are all iRegXNoSp
by adinn＠redhat.com 29 Oct '13

29 Oct '13

Changeset: 74bfec34e208 Author: adinn Date: 2013-10-29 15:49 +0000 URL: http://hg.openjdk.java.net/aarch64-port/jdk8/hotspot/rev/74bfec34e208 corrected register match rules, ensured rule outputs are all iRegXNoSp the iRegPXXX operand declarations were matching registers the wrong way round and as a consequence were allowing matches which were unnecessary. they are now correct and this avoids unnecessary work in the generated DFA code. more seriously, some of the rules were declaring their outputs with type iRegX instead of iRegXNoSp i.e. allowing the allocator to allocate the destination output from the special register set R27 to R31. this results in an asser in the register pressure computation in build_ifg_physical. specifically, this happens when a which rule generates an output declared with type iRegX produces a live range which crosses a VM callout. the fat projection following the callout kills the SOC registers but does not remove the special registers from the range, leaving the range still live even when it ought to have been yanked. this is caught by the assert because the register pressure for the block gets decremented by the amount associated with the yanked projection and it does not tally with the count associated with the live ranges. ! src/cpu/aarch64/vm/aarch64.ad ! src/cpu/aarch64/vm/c2_globals_aarch64.hpp

1 0