https://gcc.gnu.org/bugzilla/show_bug.cgi?id=78535
Bug ID: 78535 Summary: [6.2] invalid code generation with -O1 -fdevirtualize Product: gcc Version: 6.2.0 Status: UNCONFIRMED Severity: normal Priority: P3 Component: ipa Assignee: unassigned at gcc dot gnu.org Reporter: daniel.f.starke at freenet dot de Target Milestone: --- Using GCC: Target: x86_64-w64-mingw32 Configured with: ../../src/gcc-6.2.0/configure --host=x86_64-w64-mingw32 --enable-languages=c,c++ --enable-seh-exceptions --enable-threads=posix --enable-tls --disable-nls --enable-shared=libstdc++ --enable-static --enable-fully-dynamic-string --enable-lto --enable-plugins --enable-libgomp --with-dwarf2 --disable-win32-registry --enable-version-specific-runtime-libs --prefix=/mingw64-64 --with-sysroot=/mingw64-64 --target=x86_64-w64-mingw32 --enable-targets=all --enable-checking=release --with-gmp=/usr/new-gcc/lib/gmp-5.0.5 --with-mpfr=/usr/new-gcc/lib/mpfr-2.4.2 --with-mpc=/usr/new-gcc/lib/mpc-0.9 --with-isl=/usr/new-gcc/lib/isl-0.14.1 --with-cloog=/usr/new-gcc/lib/cloog-0.18.3 --with-host-libstdcxx='-lstdc++ -lsupc++' --disable-cloog-version-check --enable-cloog-backend=isl Thread model: posix I get a segmentation fault when compiling a simple hello world C++ application (libstdc++ was build with lto enabled). #include <iostream> int main() { std::cout << "Hello World!" << std::endl; return 0; } This may be related to PR 71207. The issue occurs when enabling -flto -O1 -fdevirtualize and not with -flto -O0 -fdevirtualize. I tried to isolate the additional -O1 enabled optimization by enabling all options that were enabled by -O1 (by comparing the outputs of -Q -v) manually but the problem did not occur. I noticed that -finline was not enabled this way, however, passing -fno-inline to the problematic command-line (with -O1) did produce the same flags in the -Q -v output as when I enabled all options manually additional to -O0. But it did result in a erroneous executable. So I assume that there are more options set than shown by -Q -v when enabling -O1 that finally lead to the issue or these options were passed differently to the lto stage. The backtrace looks like this: #0 0x0000000000000000 in ?? () No symbol table info available. #1 0x00000000004ded8a in widen (__c=10 '\n', this=0x4e9180 <(anonymous namespace)::ctype_c>) at E:/msys/new-gcc/bin64-64/gcc-6.2.0/x86_64-w64-mingw32/libstdc++-v3/include/bits/locale_facets.h:877 __c = 10 '\n' this = 0x4e9180 <(anonymous namespace)::ctype_c> #2 widen (__c=10 '\n', this=<optimized out>) at E:/msys/new-gcc/bin64-64/gcc-6.2.0/x86_64-w64-mingw32/libstdc++-v3/include/bits/basic_ios.h:450 No locals. #3 std::endl<char, std::char_traits<char> > (__os=...) at E:/msys/new-gcc/bin64-64/gcc-6.2.0/x86_64-w64-mingw32/libstdc++-v3/include/ostream:591 No locals. #4 0x00000000004e6f0e in operator<< () at e:/msys/mingw64-64/lib/gcc/x86_64-w64-mingw32/6.2.0/include/c++/ostream:113 No locals. #5 main () at test.cpp:4