c86d3c2853
http://mail-index.NetBSD.org/source-changes/2014/06/29/msg055885.html --- Tweak LIB1ASMFUNCS order to avoid linker warnings on libgcc_s build with -O2. Without this change, ld complains as the following: >> libgcc_s_pic.a(_float.pico):(.text+0x8): relocation truncated to fit: >> R_68K_PC16 against symbol `$_exception_handler' defined in .text section in >> libgcc_s_pic.a(_floatex.pico) _float.S and _double.S refer `$_exception_handler' declared in _floatex.S and linking the _floatex.S first seems to work around these warnings (probably caused by pic relative jump addresses). See port-m68k@ posts for more details: http://mail-index.NetBSD.org/port-m68k/2014/06/22/msg000488.html --- Note m68k/defs.mk is manually edited to avoid extra diffs. |
||
---|---|---|
.. | ||
dist | ||
lib | ||
usr.bin | ||
gcc2gcc.old | ||
gcc2netbsd | ||
Makefile | ||
Makefile.gcc_path | ||
Makefile.hooks | ||
README.gcc48 |
$Id: README.gcc48,v 1.4 2014/03/10 19:55:51 mrg Exp $ usr.bin: note: -static-libstdc++ -static-libgcc -rdynamic -lz lib: G_NOEXCEPTION_FLAGS -- need to use this new stuff: cc1objcplus collect2 -- need to re-introduce? gcc-ar, -nm and -ranlib, for lto plugin stuff - what is libitm? - asan/libsanitizer x86_64 port: - copied manually i386-builtin-types.inc instead of generating it. arch/feature list. anything not here has been switched already: architecture tools kernels libgcc native-gcc make release runs atf ------------ ----- ------- ------ ---------- ------------ ---- --- coldfire yes N/A[9] yes yes yes[7] ? ? i386 yes yes yes yes yes yes ia64 yes ? yes[6] yes no[6] m68000 yes ? yes no[3] m68k yes yes yes yes no[7] mips64eb yes yes[5] yes yes mips64el yes yes yes yes yes[8] no mipseb yes yes[5] yes yes yes mipsel yes yes yes yes yes yes no[2] powerpc yes yes[1] yes yes yes yes yes powerpc64 yes[4] sh3eb yes ? yes yes no[7] sh3el yes ? yes yes no[7] vax yes yes yes yes no[7] x86_64 yes yes yes yes yes ? [1] ofppc/pegasos has issues with symtab loading. sandpoint works fine in multi-user, found an annoyance by com(4) change. [2] weird failures. [3] relocations fail in libgcc [4] cc1 gets segv compiling libgcc. [5] mips*eb kernels seem to work as well as gcc 4.5 ones. [6] ia64 port is missing libc, libkvm and more support. libgcc has some uncommited hacks. [7] m68k/vax/sh3 gcc seem to randomly get -fstack-protector/-Wstack-protector wrong, for example m68k flags alarm() as unprotected in libc/gen/alarm.c:53, vax ifinit() in sbin/routed/if.c:680. --> hacked around [8] LOONGSON kernel has general problems [9] no kernels in evbcf port yet