introducing since release of software to be recognised. This should hopefully
allow the builds to progress a littles further on systems such as the POWER8
which features a little endian 64-bit PowerPC CPU identified as ppc64le.
This change introduced ABI incompatible change with older versions shipped
on NetBSD. This back out code that is currently not working correctly due
to TLS-based std::call_once implementation in GNU libstdc++.
Error when starting gnuchash:
/usr/pkg/lib/libwebkitgtk-1.0.so.0: Undefined symbol "_ZSt15__once_callable" (symnum = 1705)
PR 51139
Reported by <wiz>
Currently std::call_once with libstdc++ works only with static linking.
Disable code path using __thread types and introduce FIXME_PR_51139.
Problem discussed in PR 51139
Functional std::call_once is required in LLVM and LLDB codebase.
Example code to test std::call_once:
#include <iostream>
#include <thread>
#include <mutex>
#include <cstdlib>
std::once_flag flag;
int main(int argc, char **argv)
{
std::call_once(flag, [](){ std::cout << "Simple example: called once\n"; });
return EXIT_SUCCESS;
}
Sponsored by <The NetBSD Foundation>
backport fix from upstream:
URL: https://gcc.gnu.org/viewcvs?rev=238789&root=gcc&view=rev
Log:
rs6000: Fix logic for when to emit .machine (PR71216)
The current logic determining whether to use .machine in the generated
asm code puts it there if the compiler is not configured with a default
target cpu, _or_ no -mcpu= was given on the command line. It should
be "and" instead.
PR target/71216
* config/rs6000/rs6000.c (rs6000_file_start): Fix condition for
when to emit a ".machine" pseudo-op.
Modified:
branches/gcc-5-branch/gcc/ChangeLog
branches/gcc-5-branch/gcc/config/rs6000/rs6000.c
GNU ld for alpha is so clever that the redundant GP load in function
entrypoint is skipped. we must therefore skip initial GP loads; otherwise
breakpoints in function entrypoints can also be skipped.
Reported to upstream (Bug 20969):
https://sourceware.org/bugzilla/show_bug.cgi?id=20969
ok martin
sevan