------- Comment #27 from dominiq at lps dot ens dot fr 2008-02-04 13:32 ------- I have followed the advice of comment #19 and I can reproduced the behavior it describes. In particular I have implemented the patch for testsuite/libjava.jvmti/interp/natgetstacktrace.cc:
- usleep (10); + usleep (40); With that the test suite pass in 32 bit mode: === libjava Summary for unix === # of expected passes 2550 However I have several failures with -m64 and some of them are crashes, for example: Process: pr18278 [76644] Path: ./pr18278 Identifier: pr18278 Version: ??? (???) Code Type: PPC-64 (Native) Parent Process: expect [24348] Date/Time: 2008-02-04 14:21:08.366 +0100 OS Version: Mac OS X 10.5.1 (9B18) Report Version: 6 Exception Type: EXC_BAD_ACCESS (SIGSEGV) Exception Codes: KERN_INVALID_ADDRESS at 0x0000000000c13be4 Crashed Thread: 0 Thread 0 Crashed: 0 ??? 0x0000000000c13be4 0 + 12663780 1 libgcj.9.dylib 0x0000000100c13be4 GC_push_roots + 292 2 libgcj.9.dylib 0x0000000100c12f84 GC_mark_some + 516 3 libgcj.9.dylib 0x0000000100c07a14 GC_stopped_mark + 148 4 libgcj.9.dylib 0x0000000100c07e38 GC_try_to_collect_inner + 264 5 libgcj.9.dylib 0x0000000100c153d8 GC_init_inner + 808 6 libgcj.9.dylib 0x0000000100c15640 GC_init + 32 7 libgcj.9.dylib 0x0000000100c0e124 GC_init_gcj_malloc + 52 8 libgcj.9.dylib 0x0000000100069ee4 _Jv_InitGC() + 132 (boehm.cc:543) 9 libgcj.9.dylib 0x00000001000077a4 _Jv_CreateJavaVM(_Jv_VMInitArgs*) + 484 (prims.cc:1631) 10 libgcj.9.dylib 0x00000001000080c0 _Jv_RunMain(_Jv_VMInitArgs*, java::lang::Class*, char const*, int, char const**, bool) + 112 (prims.cc:1719) 11 pr18278 0x0000000100000a80 main + 92 (ccPzLWva.i:12) 12 pr18278 0x00000001000009d0 start + 68 13 ??? 0x00007fff5fbfdbc8 0 + 140734799797192 Thread 0 crashed with PPC Thread State 64: srr0: 0x0000000000c13be4 srr1: 0x900000000000f030 dar: 0x00000000e3df6000 dsisr: 0x00000000 r0: 0x0000000000c13be4 r1: 0x00007fff5fbfd0e0 r2: 0x00000001019e7668 r3: 0x0000000100c13ad8 r4: 0x0000000000000000 r5: 0x000000000000aae0 r6: 0x0000000000000000 r7: 0x0000000000000000 r8: 0x0000000000000000 r9: 0x0000000000000000 r10: 0x0000000000000000 r11: 0x0000000101a00588 r12: 0x0000000100c11720 r13: 0x00007fff704f7510 r14: 0x0000000000000000 r15: 0x0000000000000000 r16: 0x0000000000000000 r17: 0x0000000000000000 r18: 0x0000000000000000 r19: 0x0000000000000000 r20: 0x0000000000000000 r21: 0x0000000000000000 r22: 0x0000000000000000 r23: 0x0000000000000000 r24: 0x0000000000000000 r25: 0x0000000100001110 r26: 0x00007fff5fbfd290 r27: 0x0000000000000001 r28: 0x00000001019e7a10 r29: 0x00000001019e7890 r30: 0x0000000000000004 r31: 0x0000000100c13ad8 cr: 0x24000028 xer: 0x0000000020000000 lr: 0x0000000000c13be4 ctr: 0x0000000100c11720 vrsave: 0x00000000 Binary Images: 0x100000000 - 0x100000fff +pr18278 ??? (???) <5b712d1c60062c01100e8888dee76cc5> /opt/gcc/darwin_buildw/powerpc-apple-darwin9/libjava/testsuite/pr18278 0x100003000 - 0x101299fff +libgcj.9.dylib ??? (???) <e0c4b449408785b7148ca320a1d7122a> /opt/gcc/darwin_buildw/powerpc-apple-darwin9/ppc64/libjava/.libs/libgcj.9.dylib 0x103cf5000 - 0x103decffc libiconv.2.dylib ??? (???) <ebe5ffdf970cc9899b9fab929ff6a74b> /usr/lib/libiconv.2.dylib 0x7fff5fc00000 - 0x7fff5fc32423 dyld 95.3 (???) <b7d5ff5cb2eb5909a2ce29a4af0da410> /usr/lib/dyld 0x7fff80222000 - 0x7fff8022fffa libgcc_s_ppc64.1.dylib ??? (???) <25e2718ff9370df1a631bb0e4b78be45> /usr/lib/libgcc_s_ppc64.1.dylib 0x7fff826e3000 - 0x7fff82876ffb libSystem.B.dylib ??? (???) <1f468045268776257bae1c0a55cce0a4> /usr/lib/libSystem.B.dylib 0x7fff84579000 - 0x7fff8457effa libmathCommon.A.dylib ??? (???) /usr/lib/system/libmathCommon.A.dylib 0x7fff845ae000 - 0x7fff845c0fff libz.1.dylib ??? (???) <63b0d59c3ff2ce4561723d3ea723ea9c> /usr/lib/libz.1.dylib 0xffffffffffff8000 - 0xffffffffffff9703 libSystem.B.dylib ??? (???) /usr/lib/libSystem.B.dylib It seems that ReportCrash is taking a lot of time for each report, so the test is becoming very sloooow. -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=35035