25.08.2015 21:18, Mark H Weaver пишет: > Andrei Borzenkov writes: > >> 24.08.2015 11:12, Andrei Borzenkov пишет: >>> Could you test attached patch for both gcc and clang? Use >>> >>> configure TARGET_CC=clang >> >> Sorry, was broken, fixed version attached. > > Your patch works for me with gcc, although config.log shows this warning > when building the test program in your new code with the correct flags > for gcc (-mabi=32): > > ld: warning: cannot find entry symbol __start; defaulting to 00000000004000d0 > Well, I just reused existing checks in this case; we worry about it later. > Here's the relevant excerpt: > > --8<---------------cut here---------------start------------->8--- > configure:24700: gcc -o conftest -Os -Wall -W -Wshadow -Wpointer-arith -Wundef -Wchar-subscripts -Wcomment -Wdeprecated-declarations -Wdisabled-optimization -Wdiv-by-zero -Wfloat-equal -Wformat-extra-args -Wformat-security -Wformat-y2k -Wimplicit -Wimplicit-function-declaration -Wimplicit-int -Wmain -Wmissing-braces -Wmissing-format-attribute -Wmultichar -Wparentheses -Wreturn-type -Wsequence-point -Wshadow -Wsign-compare -Wswitch -Wtrigraphs -Wunknown-pragmas -Wunused -Wunused-function -Wunused-label -Wunused-parameter -Wunused-value -Wunused-variable -Wwrite-strings -Wnested-externs -Wstrict-prototypes -g -Wredundant-decls -Wmissing-prototypes -Wmissing-declarations -Wcast-align -Wextra -Wattributes -Wendif-labels -Winit-self -Wint-to-pointer-cast -Winvalid-pch -Wmissing-field-initializers -Wnonnull -Woverflow -Wvla -Wpointer-to-int-cast -Wstrict-aliasing -Wvariadic-macros -Wvolatile-register-var -Wpointer-sign -Wmissing-include-dirs -Wmissing-prototypes -Wmissing-declarations -Wformat=2 -mabi=32 -Werror -Wall -W -DGRUB_CPU_MIPSEL=1 -DGRUB_MACHINE_MIPS_LOONGSON=1 -DGRUB_MACHINE=MIPS_LOONGSON -nostdlib -static conftest.c >&5 > ld: warning: cannot find entry symbol __start; defaulting to 00000000004000d0 > configure:24700: $? = 0 > configure:24708: result: -mabi=32 > --8<---------------cut here---------------end--------------->8--- > > I also worry that linking an O32 executable might be fragile on a pure > non-multilib N32 system such as GNU Guix, where we lack even the gcc > runtime library (libgcc_s) for O32. For now it seems to work, but I > wonder if it might be more robust to do the test compilation with -c and > -freestanding? > It is using -nostdlib; is it not enough? > clang: with TARGET_CC=clang, it fails for me. Here are the > relevant excerpts from config.log: > > --8<---------------cut here---------------start------------->8--- > configure:24700: /gnu/store/yjmchvkz0kmz4s1sxc0p89imfik68qzg-clang-3.6.0/bin/clang -o conftest -Os -Wall -W -Wshadow -Wpointer-arith -Wundef -Wchar-subscripts -Wcomment -Wdeprecated-declarations -Wdisabled-optimization -Wdiv-by-zero -Wfloat-equal -Wformat-extra-args -Wformat-security -Wformat-y2k -Wimplicit -Wimplicit-function-declaration -Wimplicit-int -Wmain -Wmissing-braces -Wmissing-format-attribute -Wmultichar -Wparentheses -Wreturn-type -Wsequence-point -Wshadow -Wsign-compare -Wswitch -Wtrigraphs -Wunknown-pragmas -Wunused -Wunused-function -Wunused-label -Wunused-parameter -Wunused-value -Wunused-variable -Wwrite-strings -Wnested-externs -Wstrict-prototypes -g -Wredundant-decls -Wmissing-prototypes -Wmissing-declarations -Wcast-align -Wextra -Wattributes -Wendif-labels -Winit-self -Wint-to-pointer-cast -Winvalid-pch -Wmissing-field-initializers -Wnonnull -Woverflow -Wvla -Wpointer-to-int-cast -Wstrict-aliasing -Wvariadic-macros -Wvolatile-register-var -Wpointer-sign -Wmis sing-include-dirs -Wmissing-prototypes -Wmissing-declarations -Wformat=2 -mabi=32 -Werror -Wall -W -DGRUB_CPU_MIPSEL=1 -DGRUB_MACHINE_MIPS_LOONGSON=1 -DGRUB_MACHINE=MIPS_LOONGSON -nostdlib -static conftest.c >&5 > error: unknown target CPU 'mips32r2' > configure:24700: $? = 1 > configure: failed program was: > [...] > configure:24700: /gnu/store/yjmchvkz0kmz4s1sxc0p89imfik68qzg-clang-3.6.0/bin/clang -o conftest -Os -Wall -W -Wshadow -Wpointer-arith -Wundef -Wchar-subscripts -Wcomment -Wdeprecated-declarations -Wdisabled-optimization -Wdiv-by-zero -Wfloat-equal -Wformat-extra-args -Wformat-security -Wformat-y2k -Wimplicit -Wimplicit-function-declaration -Wimplicit-int -Wmain -Wmissing-braces -Wmissing-format-attribute -Wmultichar -Wparentheses -Wreturn-type -Wsequence-point -Wshadow -Wsign-compare -Wswitch -Wtrigraphs -Wunknown-pragmas -Wunused -Wunused-function -Wunused-label -Wunused-parameter -Wunused-value -Wunused-variable -Wwrite-strings -Wnested-externs -Wstrict-prototypes -g -Wredundant-decls -Wmissing-prototypes -Wmissing-declarations -Wcast-align -Wextra -Wattributes -Wendif-labels -Winit-self -Wint-to-pointer-cast -Winvalid-pch -Wmissing-field-initializers -Wnonnull -Woverflow -Wvla -Wpointer-to-int-cast -Wstrict-aliasing -Wvariadic-macros -Wvolatile-register-var -Wpointer-sign -Wmis sing-include-dirs -Wmissing-prototypes -Wmissing-declarations -Wformat=2 -target mips -mabi=32 -Werror -Wall -W -DGRUB_CPU_MIPSEL=1 -DGRUB_MACHINE_MIPS_LOONGSON=1 -DGRUB_MACHINE=MIPS_LOONGSON -nostdlib -static conftest.c >&5 > ld: /tmp/conftest-efc530.o: compiled for a big endian system and target is little endian > ld: /tmp/conftest-efc530.o: endianness incompatible with that of the selected emulation This is the same problem we had on powerpc. Your binutils are set to use little-endian output by default; we need to explicitly enable either big or little endian in this case as well. Could you try attached patch to verify? Although I guess we need to unify endianness checks for all supported platforms. > ld: failed to merge target specific data of file /tmp/conftest-efc530.o > ld: warning: cannot find entry symbol __start; defaulting to 00000000004000f0 > collect2: error: ld returned 1 exit status > clang-3.6: error: linker (via gcc) command failed with exit code 1 (use -v to see invocation) > configure:24700: $? = 1 > configure: failed program was: > [...] > --8<---------------cut here---------------end--------------->8--- > > I guess that "-target mips" is interpreted as requesting big endian. > > I suspect that simply passing -mabi=32 is the right thing to do for > clang as well, but that the clang we have built in GNU Guix doesn't > include support for targetting the O32 ABI. As far as I can tell, clang can enable full platform (i.e. MIPS) but not platform variant. So if clang can compile for MIPS at all it should be able to produce both versions. The problem is, endianness in clang is property of target; and what makes it even worse, options that are passed to backend toolchains are also target dependent. > It's possible that the only > way to target O32 from a non-multilib N32 system using clang is to do > cross-compilation. Or perhaps we need to explicitly request support for > 'mips32r2' when building clang. > > Mark >