powerpc64 head -r345044: WITH_LLVM_LIBUNWIND= based buildworld leads to thrown C++ exceptions segmentation faulting
Mark Millard
marklmi at yahoo.com
Wed Mar 13 03:24:41 UTC 2019
[Explicitly setting r2 to its expected value
in gdb does let the program complete without
failing.]
> On 2019-Mar-12, at 16:34, Mark Millard <marklmi at yahoo.com> wrote:
>
> [The unoptimized code is more revealing so I switch to
> presenting it and what it lead me to: more detail on
> r2 being mishandled.]
>
>> On 2019-Mar-12, at 14:05, Mark Millard <marklmi at yahoo.com> wrote:
>>
>> [I got some libunwind debug output from the a.out run. This
>> and a disassembly of main suggest an instruction pointer
>> address is too large by 0x4 for jumping to the code that
>> would call __cxa_begin_catch .]
>>
>>> On 2019-Mar-12, at 12:20, Mark Millard <marklmi at yahoo.com> wrote:
>>>
>>> [I sometimes experiment with building powerpc64 (and 32-bit) via
>>> more modern toolchains, here a amd64->powerpc64 cross build via
>>> system-clang (so 8.0.0).]
>>>
>>> buildworld with WITH_LLVM_LIBUNWIND= completes for powerpc64
>>> (but not 32-bit powerpc). However, for a system installed
>>> from such for pwoerpc64, the following program (for example)
>>> gets a segmentation fault:
>>>
>>> # more ~/c_tests/exception_test.cpp
>>> #include <exception>
>>>
>>> int main(void)
>>> {
>>> try { throw std::exception(); }
>>> catch (std::exception& e) {}
>>> return 0;
>>> }
>>>
>>> (Note: the same a.out works under a WITHOUT_LLVM_LIBUNWIND=
>>> environment, that was patched for DW_CFA_remember_state and
>>> DW_CFA_restore_state handling, with the system built via
>>> devel/powerpc64-xtoolchain-gcc related materials. So the
>>> failure is on the system library does of things for the
>>> WITH_LLVM_LIBUNWIND= context.)
>>>
>>> Unfortunately:
>>>
>>> A) devel/gdb makes extensive use of thrown C++ exceptions
>>> and so does not work for a powerpc64 system based on
>>> WITH_LLVM_LIBUNWIND= .
>>>
>>> B) The world built is not using dwarf-2 so /usr/libexec/gdb
>>> is not handy/useful.
>>>
>>> C) CFLAGS+=-gdwarf-2 leads to system-clang having an Abort
>>> trap during buildworld's compile of gcrt1.s . (Reference
>>> material later, below.)
>>>
>>> D) lldb crashes in llvm_unreachable in
>>> lldb::RegisterContextSP FreeBSDThread::GetRegisterContext()
>>> on powerpc64. (Reference material later, below.)
>>>
>>> So I've not managed to check the backtrace for the
>>> segmentation fault in the short example.
>>>
>>>
>>>
>>> For reference . . .
>>>
>>>
>>> For (C) ( -gdwarf-2 use ):
>>>
>>> QUOTES
>>> (gdb) bt
>>> #0 thr_kill () at thr_kill.S:3
>>> #1 0x000000000474afcf in __raise (s=6) at /usr/src/lib/libc/gen/raise.c:52
>>> #2 0x00000000046cd386 in abort () at /usr/src/lib/libc/stdlib/abort.c:79
>>> #3 0x00000000047394ba in __assert (func=<optimized out>, file=<optimized out>, line=<optimized out>, failedexpr=<optimized out>) at /usr/src/lib/libc/gen/assert.c:51
>>> #4 0x000000000429aa9f in resetRootFile () at /usr/src/contrib/llvm/include/llvm/MC/MCDwarf.h:316
>>> #5 parseDirectiveFile () at /usr/src/contrib/llvm/lib/MC/MCParser/AsmParser.cpp:3377
>>> #6 parseStatement () at /usr/src/contrib/llvm/lib/MC/MCParser/AsmParser.cpp:2023
>>> #7 0x000000000428cc12 in Run () at /usr/src/contrib/llvm/lib/MC/MCParser/AsmParser.cpp:884
>>> #8 0x000000000163c649 in ExecuteAssembler () at /usr/src/contrib/llvm/tools/clang/tools/driver/cc1as_main.cpp:503
>>> #9 cc1as_main () at /usr/src/contrib/llvm/tools/clang/tools/driver/cc1as_main.cpp:589
>>> #10 0x0000000001643d10 in ExecuteCC1Tool () at /usr/src/contrib/llvm/tools/clang/tools/driver/driver.cpp:312
>>> #11 main () at /usr/src/contrib/llvm/tools/clang/tools/driver/driver.cpp:382
>>>
>>> void resetRootFile() {
>>> assert(Header.MCDwarfFiles.empty());
>>> Header.RootFile.Name.clear();
>>> Header.resetMD5Usage();
>>> Header.HasSource = false;
>>> }
>>>
>>> --- lib/csu__L ---
>>> cc: error: unable to execute command: Abort trap (core dumped)
>>> cc: error: clang integrated assembler command failed due to signal (use -v to see invocation)
>>> FreeBSD clang version 8.0.0 (branches/release_80 355677) (based on LLVM 8.0.0)
>>> Target: powerpc64-unknown-freebsd13.0
>>> Thread model: posix
>>> InstalledDir: /usr/bin
>>> cc: note: diagnostic msg: PLEASE submit a bug report to https://bugs.freebsd.org/submit/ and include the crash backtrace, preprocessed source, and associated run script.
>>> cc: note: diagnostic msg: Error generating preprocessed source(s) - no preprocessable inputs.
>>> *** [gcrt1.o] Error code 254
>>>
>>> make[5]: stopped in /usr/src/lib/csu/powerpc64
>>> .ERROR_TARGET='gcrt1.o'
>>> .ERROR_META_FILE='/usr/obj/powerpc64vtsc_clang_altbinutils/powerpc.powerpc64/usr/src/powerpc.powerpc64/lib/csu/powerpc64/gcrt1.o.meta'
>>> .MAKE.LEVEL='5'
>>> MAKEFILE=''
>>> .MAKE.MODE='meta missing-filemon=yes missing-meta=yes silent=yes verbose'
>>> _ERROR_CMD='cc -gdwarf-2 -target powerpc64-unknown-freebsd13.0 --sysroot=/usr/obj/powerpc64vtsc_clang_altbinutils/powerpc.powerpc64/usr/src/powerpc.powerpc64/tmp -B/usr/local/powerpc64-unknown-freebsd13.0/bin/ -O2 -pipe -I/usr/src/lib/csu/common -I/usr/src/lib/libc/include -mlongcall -DCRT_IRELOC_SUPPRESS -std=gnu99 -Wsystem-headers -Wall -Wno-format-y2k -W -Wno-unused-parameter -Wstrict-prototypes -Wmissing-prototypes -Wpointer-arith -Wreturn-type -Wcast-qual -Wwrite-strings -Wswitch -Wshadow -Wunused-parameter -Wcast-align -Wchar-subscripts -Winline -Wnested-externs -Wredundant-decls -Wold-style-definition -Wno-pointer-sign -Wthread-safety -Wno-empty-body -Wno-string-plus-int -Wno-unused-const-variable -Qunused-arguments -c -o gcrt1.o gcrt1.s;'
>>> .CURDIR='/usr/src/lib/csu/powerpc64'
>>> .MAKE='make'
>>> .OBJDIR='/usr/obj/powerpc64vtsc_clang_altbinutils/powerpc.powerpc64/usr/src/powerpc.powerpc64/lib/csu/powerpc64'
>>> .TARGETS='all'
>>> DESTDIR='/usr/obj/powerpc64vtsc_clang_altbinutils/powerpc.powerpc64/usr/src/powerpc.powerpc64/tmp'
>>> LD_LIBRARY_PATH=''
>>> MACHINE='powerpc'
>>> MACHINE_ARCH='powerpc64'
>>> MAKEOBJDIRPREFIX=''
>>> MAKESYSPATH='/usr/src/share/mk'
>>> MAKE_VERSION='20181221'
>>> PATH='/usr/obj/powerpc64vtsc_clang_altbinutils/powerpc.powerpc64/usr/src/powerpc.powerpc64/tmp/usr/sbin:/usr/obj/powerpc64vtsc_clang_altbinutils/powerpc.powerpc64/usr/src/powerpc.powerpc64/tmp/usr/bin:/usr/obj/powerpc64vtsc_clang_altbinutils/powerpc.powerpc64/usr/src/powerpc.powerpc64/tmp/legacy/usr/sbin:/usr/obj/powerpc64vtsc_clang_altbinutils/powerpc.powerpc64/usr/src/powerpc.powerpc64/tmp/legacy/usr/bin:/usr/obj/powerpc64vtsc_clang_altbinutils/powerpc.powerpc64/usr/src/powerpc.powerpc64/tmp/legacy/bin::/sbin:/bin:/usr/sbin:/usr/bin'
>>> SRCTOP='/usr/src'
>>> OBJTOP='/usr/obj/powerpc64vtsc_clang_altbinutils/powerpc.powerpc64/usr/src/powerpc.powerpc64'
>>> .MAKE.MAKEFILES='/usr/src/share/mk/sys.mk /usr/src/share/mk/local.sys.env.mk /usr/src/share/mk/src.sys.env.mk /root/src.configs/src.conf.powerpc64-clang_altbinutils-bootstrap.amd64-host /usr/src/share/mk/bsd.mkopt.mk /usr/src/share/mk/src.sys.obj.mk /usr/src/share/mk/auto.obj.mk /usr/src/share/mk/bsd.suffixes.mk /root/src.configs/make.conf /usr/src/share/mk/local.sys.mk /usr/src/share/mk/src.sys.mk /dev/null /usr/src/lib/csu/powerpc64/Makefile /usr/src/share/mk/bsd.lib.mk /usr/src/share/mk/bsd.init.mk /usr/src/share/mk/bsd.opts.mk /usr/src/share/mk/bsd.cpu.mk /usr/src/share/mk/local.init.mk /usr/src/share/mk/src.init.mk /usr/src/lib/csu/powerpc64/../Makefile.inc /usr/src/share/mk/src.opts.mk /usr/src/share/mk/bsd.own.mk /usr/src/share/mk/bsd.compiler.mk /usr/src/share/mk/bsd.linker.mk /usr/src/lib/csu/powerpc64/../../Makefile.inc /usr/src/share/mk/bsd.libnames.mk /usr/src/share/mk/src.libnames.mk /usr/src/share/mk/bsd.symver.mk /usr/src/share/mk/bsd.nls.mk /usr/src/share/mk/bsd.confs.mk /usr/src/share/mk/bsd.files.mk /usr/src/share/mk/bsd.dirs.mk /usr/src/share/mk/bsd.incs.mk /usr/src/share/mk/bsd.links.mk /usr/src/share/mk/bsd.dep.mk /usr/src/share/mk/bsd.clang-analyze.mk /usr/src/share/mk/bsd.obj.mk /usr/src/share/mk/bsd.subdir.mk /usr/src/share/mk/bsd.sys.mk'
>>> .PATH='. /usr/src/lib/csu/powerpc64 /usr/src/lib/csu/common'
>>> 1 error
>>> END QUOTES
>>>
>>>
>>> For (D) (lldb):
>>>
>>> QUOTES
>>> CPU not supported
>>> UNREACHABLE executed at /usr/src/contrib/llvm/tools/lldb/source/Plugins/Process/FreeBSD/FreeBSDThread.cpp:192!
>>> Abort trap (core dumped)
>>>
>>> (gdb) bt
>>> #0 0x0000000813715208 in .__sys_thr_kill () at thr_kill.S:3
>>> #1 0x00000008137147cc in __raise (s=<optimized out>) at /usr/src/lib/libc/gen/raise.c:52
>>> #2 0x000000081366b5d8 in abort () at /usr/src/lib/libc/stdlib/abort.c:79
>>> #3 0x0000000011df6fb8 in llvm::llvm_unreachable_internal () at /usr/src/contrib/llvm/lib/Support/ErrorHandling.cpp:222
>>> #4 0x00000000103aaaf8 in FreeBSDThread::GetRegisterContext () at /usr/src/contrib/llvm/tools/lldb/source/Plugins/Process/FreeBSD/FreeBSDThread.cpp:192
>>> #5 0x00000000105807d4 in lldb_private::Thread::SetupForResume () at /usr/src/contrib/llvm/tools/lldb/source/Target/Thread.cpp:613
>>> #6 0x0000000010571bc8 in lldb_private::ThreadList::WillResume () at /usr/src/contrib/llvm/tools/lldb/source/Target/ThreadList.cpp:541
>>> #7 0x00000000105da23c in lldb_private::Process::PrivateResume () at /usr/src/contrib/llvm/tools/lldb/source/Target/Process.cpp:3281
>>> #8 0x00000000105a00c8 in lldb_private::Target::Launch () at /usr/src/contrib/llvm/tools/lldb/source/Target/Target.cpp:2922
>>> #9 0x000000001073f550 in CommandObjectProcessLaunch::DoExecute () at /usr/src/contrib/llvm/tools/lldb/source/Commands/CommandObjectProcess.cpp:221
>>> #10 0x00000000106c36c4 in lldb_private::CommandObjectParsed::Execute () at /usr/src/contrib/llvm/tools/lldb/source/Interpreter/CommandObject.cpp:975
>>> #11 0x00000000106d8b44 in lldb_private::CommandInterpreter::HandleCommand () at /usr/src/contrib/llvm/tools/lldb/source/Interpreter/CommandInterpreter.cpp:1761
>>> #12 0x00000000106da0a0 in lldb_private::CommandInterpreter::IOHandlerInputComplete () at /usr/src/contrib/llvm/tools/lldb/source/Interpreter/CommandInterpreter.cpp:2801
>>> #13 0x00000000107c0a08 in lldb_private::IOHandlerEditline::Run () at /usr/src/contrib/llvm/tools/lldb/source/Core/IOHandler.cpp:558
>>> #14 0x0000000010346e5c in lldb_private::Debugger::ExecuteIOHandlers () at /usr/src/contrib/llvm/tools/lldb/source/Core/Debugger.cpp:988
>>> #15 0x00000000106c8ddc in lldb_private::CommandInterpreter::RunCommandInterpreter () at /usr/src/contrib/llvm/tools/lldb/source/Interpreter/CommandInterpreter.cpp:3003
>>> #16 0x000000001034feb4 in lldb::SBDebugger::RunCommandInterpreter () at /usr/src/contrib/llvm/tools/lldb/source/API/SBDebugger.cpp:935
>>> #17 0x00000000101de878 in Driver::MainLoop () at /usr/src/contrib/llvm/tools/lldb/tools/driver/Driver.cpp:756
>>> #18 0x00000000101a0088 in main () at /usr/src/contrib/llvm/tools/lldb/tools/driver/Driver.cpp:936
>>>
>>> lldb::RegisterContextSP FreeBSDThread::GetRegisterContext() {
>>> if (!m_reg_context_sp) {
>>> m_posix_thread = nullptr;
>>>
>>> RegisterInfoInterface *reg_interface = nullptr;
>>> const ArchSpec &target_arch = GetProcess()->GetTarget().GetArchitecture();
>>>
>>> switch (target_arch.GetMachine()) {
>>> case llvm::Triple::aarch64:
>>> reg_interface = new RegisterInfoPOSIX_arm64(target_arch);
>>> break;
>>> case llvm::Triple::arm:
>>> reg_interface = new RegisterInfoPOSIX_arm(target_arch);
>>> break;
>>> case llvm::Triple::ppc:
>>> #ifndef __powerpc64__
>>> reg_interface = new RegisterContextFreeBSD_powerpc32(target_arch);
>>> break;
>>> #endif
>>> case llvm::Triple::ppc64:
>>> reg_interface = new RegisterContextFreeBSD_powerpc64(target_arch);
>>> break;
>>> case llvm::Triple::mips64:
>>> reg_interface = new RegisterContextFreeBSD_mips64(target_arch);
>>> break;
>>> case llvm::Triple::x86:
>>> reg_interface = new RegisterContextFreeBSD_i386(target_arch);
>>> break;
>>> case llvm::Triple::x86_64:
>>> reg_interface = new RegisterContextFreeBSD_x86_64(target_arch);
>>> break;
>>> default:
>>> llvm_unreachable("CPU not supported");
>>> }
>>> END QUOTES.
>>
>>
>>
>> I ran into libunwind having LIBUNWIND_PRINT_UNWINDING and
>> LIBUNWIND_PRINT_APIS so I can report for the segmentation
>> fault:
>>
>> # export LIBUNWIND_PRINT_UNWINDING=""
>> # export LIBUNWIND_PRINT_APIS=""
>> # ./a.out
>> . . .
>
> Using unoptimized code instead:
>
> # c++ -g exception_test.cpp
> libunwind: __register_frame_info(0x137d6610, 0x13ae3150)
> libunwind: __register_frame_info(0x137d6610, 0x13ae3150)
> libunwind: __deregister_frame_info(0x137d6610)
> libunwind: __deregister_frame_info(0x137d6610)
> # ./a.out
> libunwind: _Unwind_RaiseException(ex_obj=0x810043060)
> libunwind: unw_init_local(cursor=0x3fffffffffffcb48, context=0x3fffffffffffd0e0)
> libunwind: unw_step(cursor=0x3fffffffffffcb48)
> libunwind: unw_get_proc_info(cursor=0x3fffffffffffcb48, &info=0x3fffffffffffd830)
> libunwind: unw_get_proc_name(cursor=0x3fffffffffffcb48, &buf=0x3fffffffffffd628, bufLen=512)
> libunwind: unw_get_reg(cursor=0x3fffffffffffcb48, regNum=-1, &value=0x3fffffffffffd618)
> libunwind: unwind_phase1(ex_ojb=0x810043060): pc=0x81019d954, start_ip=0x81019d860, func=.anonymous., lsda=0x0, personality=0x0
> libunwind: unw_step(cursor=0x3fffffffffffcb48)
> libunwind: unw_get_proc_info(cursor=0x3fffffffffffcb48, &info=0x3fffffffffffd830)
> libunwind: unw_get_proc_name(cursor=0x3fffffffffffcb48, &buf=0x3fffffffffffd628, bufLen=512)
> libunwind: unw_get_reg(cursor=0x3fffffffffffcb48, regNum=-1, &value=0x3fffffffffffd618)
> libunwind: unwind_phase1(ex_ojb=0x810043060): pc=0x10000dac, start_ip=0x10000d64, func=.anonymous., lsda=0x10000fe0, personality=0x8101b5360
> libunwind: unwind_phase1(ex_ojb=0x810043060): calling personality function 0x8101b5360
> libunwind: unw_get_proc_info(cursor=0x3fffffffffffcb48, &info=0x3fffffffffffc950)
> libunwind: _Unwind_GetLanguageSpecificData(context=0x3fffffffffffcb48) => 0x10000fe0
> libunwind: unw_get_proc_info(cursor=0x3fffffffffffcb48, &info=0x3fffffffffffc890)
> libunwind: _Unwind_GetRegionStart(context=0x3fffffffffffcb48) => 0x10000d64
> libunwind: unw_get_reg(cursor=0x3fffffffffffcb48, regNum=-1, &value=0x3fffffffffffc8c8)
> libunwind: _Unwind_GetIP(context=0x3fffffffffffcb48) => 0x10000dac
> libunwind: unw_get_proc_info(cursor=0x3fffffffffffcb48, &info=0x3fffffffffffc880)
> libunwind: _Unwind_GetRegionStart(context=0x3fffffffffffcb48) => 0x10000d64
> libunwind: unw_get_reg(cursor=0x3fffffffffffcb48, regNum=-2, &value=0x3fffffffffffd828)
> libunwind: unwind_phase1(ex_ojb=0x810043060): _URC_HANDLER_FOUND
> libunwind: unw_init_local(cursor=0x3fffffffffffcb48, context=0x3fffffffffffd0e0)
> libunwind: unwind_phase2(ex_ojb=0x810043060)
> libunwind: unw_step(cursor=0x3fffffffffffcb48)
> libunwind: unw_get_reg(cursor=0x3fffffffffffcb48, regNum=-2, &value=0x3fffffffffffca58)
> libunwind: unw_get_proc_info(cursor=0x3fffffffffffcb48, &info=0x3fffffffffffca10)
> libunwind: unw_get_proc_name(cursor=0x3fffffffffffcb48, &buf=0x3fffffffffffc810, bufLen=512)
> libunwind: unwind_phase2(ex_ojb=0x810043060): start_ip=0x81019d860, func=.anonymous., sp=0x3fffffffffffd8e0, lsda=0x0, personality=0x0
> libunwind: unw_step(cursor=0x3fffffffffffcb48)
> libunwind: unw_get_reg(cursor=0x3fffffffffffcb48, regNum=-2, &value=0x3fffffffffffca58)
> libunwind: unw_get_proc_info(cursor=0x3fffffffffffcb48, &info=0x3fffffffffffca10)
> libunwind: unw_get_proc_name(cursor=0x3fffffffffffcb48, &buf=0x3fffffffffffc810, bufLen=512)
> libunwind: unwind_phase2(ex_ojb=0x810043060): start_ip=0x10000d64, func=.anonymous., sp=0x3fffffffffffd980, lsda=0x10000fe0, personality=0x8101b5360
> libunwind: unw_get_proc_info(cursor=0x3fffffffffffcb48, &info=0x3fffffffffffc610)
> libunwind: _Unwind_GetLanguageSpecificData(context=0x3fffffffffffcb48) => 0x10000fe0
> libunwind: _Unwind_SetIP(context=0x3fffffffffffcb48, value=0x10000db4)
> libunwind: unw_set_reg(cursor=0x3fffffffffffcb48, regNum=-1, value=0x10000db4)
> libunwind: _Unwind_SetGR(context=0x3fffffffffffcb48, reg=3, value=0x810043060)
> libunwind: unw_set_reg(cursor=0x3fffffffffffcb48, regNum=3, value=0x810043060)
> libunwind: _Unwind_SetGR(context=0x3fffffffffffcb48, reg=4, value=0x1)
> libunwind: unw_set_reg(cursor=0x3fffffffffffcb48, regNum=4, value=0x1)
> libunwind: unwind_phase2(ex_ojb=0x810043060): _URC_INSTALL_CONTEXT
> libunwind: unw_get_reg(cursor=0x3fffffffffffcb48, regNum=-1, &value=0x3fffffffffffc810)
> libunwind: unw_get_reg(cursor=0x3fffffffffffcb48, regNum=-2, &value=0x3fffffffffffca58)
> libunwind: unwind_phase2(ex_ojb=0x810043060): re-entering user code with ip=0x10000db4, sp=0x3fffffffffffd980
> libunwind: unw_resume(cursor=0x3fffffffffffcb48)
> Segmentation fault (core dumped)
>
> The below shows that the 0x10000db4 in this case seems right --and
> also that the code sequence does not set r2. Apparently the
> throw handling was supposed to set it so the optimized code
> skips setting it. (Later below it seems r2 should have been
> set by something but was not correctly set.)
>
> (gdb) disass main
> Dump of assembler code for function main():
> 0x0000000010000d64 <+0>: mflr r0
> 0x0000000010000d68 <+4>: std r31,-8(r1)
> 0x0000000010000d6c <+8>: std r0,16(r1)
> 0x0000000010000d70 <+12>: stdu r1,-160(r1)
> 0x0000000010000d74 <+16>: mr r31,r1
> 0x0000000010000d78 <+20>: li r3,0
> 0x0000000010000d7c <+24>: stw r3,148(r31)
> 0x0000000010000d80 <+28>: li r3,8
> 0x0000000010000d84 <+32>: bl 0x100007a0 <00000018.plt_call.__cxa_allocate_exception@@CXXABI_1.3>
> 0x0000000010000d88 <+36>: ld r2,40(r1)
> 0x0000000010000d8c <+40>: std r3,112(r31)
> 0x0000000010000d90 <+44>: bl 0x10000e00 <std::exception::exception()>
> 0x0000000010000d94 <+48>: nop
> 0x0000000010000d98 <+52>: ld r4,-32728(r2)
> 0x0000000010000d9c <+56>: nop
> 0x0000000010000da0 <+60>: ld r5,-32720(r2)
> 0x0000000010000da4 <+64>: ld r3,112(r31)
> 0x0000000010000da8 <+68>: bl 0x10000800 <00000018.plt_call.__cxa_throw@@CXXABI_1.3>
> 0x0000000010000dac <+72>: ld r2,40(r1)
> 0x0000000010000db0 <+76>: b 0x10000df4 <main()+144>
> 0x0000000010000db4 <+80>: mr r5,r4
> 0x0000000010000db8 <+84>: std r3,136(r31)
> 0x0000000010000dbc <+88>: stw r5,132(r31)
> 0x0000000010000dc0 <+92>: b 0x10000dc4 <main()+96>
> 0x0000000010000dc4 <+96>: ld r3,136(r31)
> 0x0000000010000dc8 <+100>: bl 0x100007c0 <00000018.plt_call.__cxa_begin_catch@@CXXABI_1.3>
> 0x0000000010000dcc <+104>: ld r2,40(r1)
> 0x0000000010000dd0 <+108>: std r3,120(r31)
> 0x0000000010000dd4 <+112>: bl 0x100007e0 <00000018.plt_call.__cxa_end_catch@@CXXABI_1.3>
> 0x0000000010000dd8 <+116>: ld r2,40(r1)
> 0x0000000010000ddc <+120>: li r3,0
> 0x0000000010000de0 <+124>: addi r1,r1,160
> 0x0000000010000de4 <+128>: ld r0,16(r1)
> 0x0000000010000de8 <+132>: ld r31,-8(r1)
> 0x0000000010000dec <+136>: mtlr r0
> 0x0000000010000df0 <+140>: blr
> 0x0000000010000df4 <+144>: .long 0x0
> 0x0000000010000df8 <+148>: .long 0x0
> 0x0000000010000dfc <+152>: .long 0x0
> End of assembler dump.
>
> Notably lr seems to be: 0x10000dcc which
> is just after:
>
> bl 0x100007c0 <00000018.plt_call.__cxa_begin_catch@@CXXABI_1.3>
>
> Note also the pc, r12, and ctr all having the failure
> address: 0x81042b900 .
>
> #0 0x000000081042b900 in ?? () from /lib/libc.so.7
> (gdb) x/32i 0x000000081042b8F0
> 0x81042b8f0 <getfsstat at got.plt>: .long 0x8
> 0x81042b8f4 <getfsstat at got.plt+4>: vmrglh v1,v30,v7
> 0x81042b8f8: .long 0x8
> 0x81042b8fc: vsubeuqm v1,v30,v7,v17
> => 0x81042b900: .long 0x8
> 0x81042b904: vpmsumw v1,v30,v7
> 0x81042b908 <_citrus_bcs_skip_ws_len at got.plt>: .long 0x8
> 0x81042b90c <_citrus_bcs_skip_ws_len at got.plt+4>: .long 0x103e3c92
> 0x81042b910: .long 0x8
> 0x81042b914: .long 0x103e3c9b
> 0x81042b918: .long 0x8
>
> (gdb) info reg
> r0 0x810563d10 34633825552
> r1 0x3fffffffffffd980 4611686018427378048
> r2 0x0 0
> r3 0x810043060 34628448352
> r4 0x1 1
> r5 0x1 1
> r6 0x8103d9490 34632209552
> r7 0x0 0
> r8 0x29 41
> r9 0x4e 78
> r10 0x3fffffffffffc8f8 4611686018427373816
> r11 0x81056c28c 34633859724
> r12 0x81042b900 34632546560
> r13 0x81005f020 34628562976
> r14 0x0 0
> r15 0x0 0
> r16 0x0 0
> r17 0x0 0
> r18 0x0 0
> r19 0x0 0
> r20 0x0 0
> r21 0x0 0
> r22 0x0 0
> r23 0x0 0
> r24 0x0 0
> r25 0x0 0
> r26 0x0 0
> r27 0x0 0
> r28 0x1 1
> r29 0x3fffffffffffdb78 4611686018427378552
> r30 0x3fffffffffffdb88 4611686018427378568
> r31 0x3fffffffffffd980 4611686018427378048
> pc 0x81042b900 0x81042b900
> msr <unavailable>
> cr 0x28000802 671090690
> lr 0x10000dcc 0x10000dcc <main()+104>
> ctr 0x81042b900 34632546560
> xer 0x0 0
> fpscr 0x0 0
> vscr <unavailable>
> vrsave <unavailable>
>
> And the following shows how r12 and ctr were filled in by
> code that expected r2 to be correct:
>
> (gdb) x/32i 0x100007c0
> 0x100007c0 <00000018.plt_call.__cxa_begin_catch@@CXXABI_1.3>: std r2,40(r1)
> 0x100007c4 <00000018.plt_call.__cxa_begin_catch@@CXXABI_1.3+4>: ld r12,-32608(r2)
> 0x100007c8 <00000018.plt_call.__cxa_begin_catch@@CXXABI_1.3+8>: mtctr r12
> 0x100007cc <00000018.plt_call.__cxa_begin_catch@@CXXABI_1.3+12>: ld r11,-32592(r2)
> 0x100007d0 <00000018.plt_call.__cxa_begin_catch@@CXXABI_1.3+16>: ld r2,-32600(r2)
> 0x100007d4 <00000018.plt_call.__cxa_begin_catch@@CXXABI_1.3+20>: bctr
> 0x100007d8 <00000018.plt_call.__cxa_begin_catch@@CXXABI_1.3+24>: .long 0x0
> 0x100007dc <00000018.plt_call.__cxa_begin_catch@@CXXABI_1.3+28>: .long 0x0
> . . .
>
> Overall: r2 seem to be mishandled in the exception handling.
Just before:
0x0000000010000dc8 <+100>: bl 0x100007c0 <00000018.plt_call.__cxa_begin_catch@@CXXABI_1.3>
(gdb) set $r2=0x10019300
(gdb) c
Continuing.
Program exited normally.
r2 has the wrong value and needs to have been set by:
ld r2,40(r1)
(expressed as an additional instruction in teh prelude to
bl 0x100007c0 <00000018.plt_call.__cxa_begin_catch@@CXXABI_1.3>)
That shows how I got the 0x10019300 value as well: 40(r1)
lookup.
This makes the optimized code interesting because it had
the "ld r2,40(r1)" but it was skipped by that code being
started at 0x0000000010000dac .
0x0000000010000da8 <+68>: ld r2,40(r1)
0x0000000010000dac <+72>: bl 0x100007c0 <00000018.plt_call.__cxa_begin_catch@@CXXABI_1.3>
(An earlier message of this sequence has more detail
for the optimized code.)
> Note on the gdb use:
>
> The above devel/gdb activity was executed from a
> devel/powerpc64-xtoolchain-gcc built world that was based on
> WITHOUT_LLVM_LIBUNWIND= and and my patched libgcc_s material.
> This means it suffered from mismatches with the clang/libunwind
> world (that I chroot to):
>
> QUOTE
> warning: .dynamic section for "/usr/lib/libc++.so.1" is not at the expected address (wrong library or version mismatch?)
>
> warning: .dynamic section for "/lib/libcxxrt.so.1" is not at the expected address (wrong library or version mismatch?)
>
> warning: .dynamic section for "/lib/libm.so.5" is not at the expected address (wrong library or version mismatch?)
>
> warning: .dynamic section for "/lib/libc.so.7" is not at the expected address (wrong library or version mismatch?)
>
> warning: .dynamic section for "/lib/libgcc_s.so.1" is not at the expected address (wrong library or version mismatch?)
>
> warning: .dynamic section for "/libexec/ld-elf.so.1" is not at the expected address (wrong library or version mismatch?)
> END QUOTE
>
> But any thrown exceptions the gdb may have used worked in my
> patched WITHOUT_LLVM_LIBUNWIND= environment. This gdb is not
> limited to dwarf-2.
>
> The material from the a.out is not misinterpreted even if
> library code details might be.
===
Mark Millard
marklmi at yahoo.com
( dsl-only.net went
away in early 2018-Mar)
More information about the freebsd-ppc
mailing list