| Summary: | [4.9/5] sys-devel/gcc-4.9.0 - libitm fails to build with fortify enable | ||
|---|---|---|---|
| Product: | Gentoo Linux | Reporter: | Magnus Granberg <zorry> |
| Component: | [OLD] Core system | Assignee: | Gentoo Toolchain Maintainers <toolchain> |
| Status: | RESOLVED FIXED | ||
| Severity: | normal | CC: | hardened, jana |
| Priority: | Normal | ||
| Version: | unspecified | ||
| Hardware: | All | ||
| OS: | Linux | ||
| URL: | https://gcc.gnu.org/PR61164 | ||
| See Also: | https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61164 | ||
| Whiteboard: | |||
| Package list: | Runtime testing required: | --- | |
| Bug Depends on: | |||
| Bug Blocks: | 259417, 508418 | ||
| Attachments: | Disable fortify in libitm | ||
Created attachment 375832 [details, diff]
Disable fortify in libitm
We disable fortify as Debian/Ubuntu do but it need to be fixed.
So, out of curiosity, is this a bootstrap problem ? That is, is this failure caused by something like bug 365015 and once gcc 4.9.0 is built, it can build itself without stumbling into this ? (In reply to Rafał Mużyło from comment #2) > So, out of curiosity, is this a bootstrap problem ? > > That is, is this failure caused by something like bug 365015 and once gcc > 4.9.0 is built, it can build itself without stumbling into this ? it is gcc 4.9.0 (xgcc) that is trying to build the itm lib This might be fallout from http://gcc.gnu.org/git/?p=gcc.git;a=commit;h=f0d26d578d94a3fafc81d191d2168ab3517b3f0e Does reverting that help? Portage 2.2.10 (hardened/linux/amd64, gcc-4.8.2, glibc-2.19, 3.13.9-hardened x86_64) ================================================================= System uname: Linux-3.13.9-hardened-x86_64-Intel-R-_Core-TM-_i7_CPU_Q_720_@_1.60GHz-with-gentoo-2.2 KiB Mem: 8114608 total, 523960 free KiB Swap: 0 total, 0 free Timestamp of tree: Sun, 27 Apr 2014 15:15:01 +0000 ld GNU ld (GNU Binutils) 2.24 app-shells/bash: 4.2_p47 dev-java/java-config: 2.2.0 dev-lang/python: 2.6.9, 2.7.6, 3.2.5-r3, 3.3.5, 3.4.0 dev-util/cmake: 2.8.12.2 dev-util/pkgconfig: 0.28-r1 sys-apps/baselayout: 2.2 sys-apps/openrc: 0.12.4 sys-apps/sandbox: 2.6-r1 sys-devel/autoconf: 2.13, 2.69 sys-devel/automake: 1.8.5-r4, 1.9.6-r3, 1.11.6, 1.12.6, 1.13.4, 1.14.1 sys-devel/binutils: 2.24-r2 sys-devel/gcc: 4.4.4-r2, 4.5.3-r2, 4.6.4, 4.7.3, 4.8.2 sys-devel/gcc-config: 1.8 sys-devel/libtool: 2.4.2 sys-devel/make: 4.0-r1 sys-kernel/linux-headers: 3.14 (virtual/os-headers) sys-libs/glibc: 2.19 ACCEPT_KEYWORDS="~amd64" CFLAGS="-pipe -O2 -march=core2" CXXFLAGS="-pipe -O2 -march=core2" (In reply to Ryan Hill from comment #4) > This might be fallout from > http://gcc.gnu.org/git/?p=gcc.git;a=commit; > h=f0d26d578d94a3fafc81d191d2168ab3517b3f0e > > Does reverting that help? Still the same error There is something weird going on with glibc's "__fortify_function" define. It complains about problems inlining the fortification wrapper functions from C++ mode. When I replace "__fortify_function" by "__fortify_function inline" in /usr/include/bits/stdio2.h, the libitm subdirecty compiles. It will quickly fail in other directories complaining about duplicate "inline" keywords. So apparently, the macros should have figured out to add "inline" to the list of keywords, but for some unknown reason it fails for the "libitm" subdirectory. It's kind of hard to follow how the __fortify_function macro is actually expanded in what situation. It depends on the compiler mode or if C or C++ and whether GNU mode is used and so on... AFAICT the "always_inline" gcc extension is supposed to be used in C mode, and a simple "inline" in c++. Maybe related to the way the compiler is called, a special mode used during gcc building that only enables partial features. (e.g. libitm is written in C++, but doesn't need libstdc++) I guess libitm defines some preprocessor things that throw off glibc because it assumes fortification is not enabled by default or something. Haven't tried to chase down the exact cause. Just checked, a preprocessor run reveils that __fortify_function is turned into a simple "extern" only in libitm, no signs of any inlining keyword... Strange, I looked at a few different files that use vfprintf and the definition of __fortify_function is the same for all. The declaration of vfprintf however is missing an __inline keyword in libitm. should be fixed for gcc-6. we're still carrying Ryan's no-fortify patch until then. nothing left for us here. |
When fortify is enable gcc 4.9 fail to build the itm lib with sys-include -DHAVE_CONFIG_H -I. -I/var/tmp/portage/sys-devel/gcc-4.9.0/work/gcc-4.9.0/libitm -I/var/tmp/portage/sys-devel/gcc-4.9.0/work/gcc-4.9.0/libitm/config/linux/x86 -I/var/tmp/portage/sys-devel/gcc-4.9.0/work/gcc-4.9.0/libitm/config/linux -I/var/tmp/portage/sys-devel/gcc-4.9.0/work/gcc-4.9.0/libitm/config/x86 -I/var/tmp/portage/sys-devel/gcc-4.9.0/work/gcc-4.9.0/libitm/config/posix -I/var/tmp/portage/sys-devel/gcc-4.9.0/work/gcc-4.9.0/libitm/config/generic -I/var/tmp/portage/sys-devel/gcc-4.9.0/work/gcc-4.9.0/libitm -mrtm -pthread -Wall -std=gnu++0x -funwind-tables -fno-exceptions -fno-rtti -fabi-version=4 -pipe -O2 -march=core2 -D_GNU_SOURCE -MT useraction.lo -MD -MP -MF .deps/useraction.Tpo -c /var/tmp/portage/sys-devel/gcc-4.9.0/work/gcc-4.9.0/libitm/useraction.cc -o useraction.o >/dev/null 2>&1 mv -f .deps/sanitizer_libignore.Tpo .deps/sanitizer_libignore.Plo /bin/sh ../libtool --tag=CXX --mode=compile /var/tmp/portage/sys-devel/gcc-4.9.0/work/build/./gcc/xgcc -shared-libgcc -B/var/tmp/portage/sys-devel/gcc-4.9.0/work/build/./gcc -nostdinc++ -L/var/tmp/portage/sys-devel/gcc-4.9.0/work/build/x86_64-pc-linux-gnu/libstdc++-v3/src -L/var/tmp/portage/sys-devel/gcc-4.9.0/work/build/x86_64-pc-linux-gnu/libstdc++-v3/src/.libs -L/var/tmp/portage/sys-devel/gcc-4.9.0/work/build/x86_64-pc-linux-gnu/libstdc++-v3/libsupc++/.libs -B/usr/x86_64-pc-linux-gnu/bin/ -B/usr/x86_64-pc-linux-gnu/lib/ -isystem /usr/x86_64-pc-linux-gnu/include -isystem /usr/x86_64-pc-linux-gnu/sys-include -D_GNU_SOURCE -D_DEBUG -D__STDC_CONSTANT_MACROS -D__STDC_FORMAT_MACROS -D__STDC_LIMIT_MACROS -I. -I/var/tmp/portage/sys-devel/gcc-4.9.0/work/gcc-4.9.0/libsanitizer/sanitizer_common -I.. -I /var/tmp/portage/sys-devel/gcc-4.9.0/work/gcc-4.9.0/libsanitizer/include -isystem /var/tmp/portage/sys-devel/gcc-4.9.0/work/gcc-4.9.0/libsanitizer/include/system -Wall -W -Wno-unused-parameter -Wwrite-strings - pedantic -Wno-long-long -fPIC -fno-builtin -fno-exceptions -fno-rtti -fomit-frame-pointer -funwind-tables -fvisibility=hidden -Wno-variadic-macros -I../../libstdc++-v3/include -I../../libstdc++-v3/include/x86_64-pc-linux-gnu -I/var/tmp/portage/sys-devel/gcc-4.9.0/work/gcc-4.9.0/libsanitizer/../libstdc++-v3/libsupc++ -DSANITIZER_LIBBACKTRACE -DSANITIZER_CP_DEMANGLE -I /var/tmp/portage/sys-devel/gcc-4.9.0/work/gcc-4.9.0/libsanitizer/../libbacktrace -I ../libbacktrace -I /var/tmp/portage/sys-devel/gcc-4.9.0/work/gcc-4.9.0/libsanitizer/../include -include /var/tmp/portage/sys-devel/gcc-4.9.0/work/gcc-4.9.0/libsanitizer/libbacktrace/backtrace-rename.h -pipe -O2 -march=core2 -D_GNU_SOURCE -MT sanitizer_linux.lo -MD -MP -MF .deps/sanitizer_linux.Tpo -c -o sanitizer_linux.lo /var/tmp/portage/sys-devel/gcc-4.9.0/work/gcc-4.9.0/libsanitizer/sanitizer_common/sanitizer_linux.cc libtool: compile: /var/tmp/portage/sys-devel/gcc-4.9.0/work/build/./gcc/xgcc -B/var/tmp/portage/sys-devel/gcc-4.9.0/work/build/./gcc/ -B/usr/x86_64-pc-linux-gnu/bin/ -B/usr/x86_64-pc-linux-gnu/lib/ -isystem /usr/x86_64-pc-linux-gnu/include -isystem /usr/x86_64-pc-linux-gnu/sys-include -DHAVE_CONFIG_H -I. -I/var/tmp/portage/sys-devel/gcc-4.9.0/work/gcc-4.9.0/libquadmath -I /var/tmp/portage/sys-devel/gcc-4.9.0/work/gcc-4.9.0/libquadmath/../include -g -pipe -O2 -march=core2 -m32 -MT math/ctanq.lo -MD -MP -MF math/.deps/ctanq.Tpo -c /var/tmp/portage/sys-devel/gcc-4.9.0/work/gcc-4.9.0/libquadmath/math/ctanq.c -o math/ctanq.o >/dev/null 2>&1 In file included from /usr/include/stdio.h:937:0, from /var/tmp/portage/sys-devel/gcc-4.9.0/work/gcc-4.9.0/libitm/util.cc:27: /usr/include/bits/stdio2.h: In function ‘void GTM::gtm_verror(const char*, __va_list_tag*)’: /usr/include/bits/stdio2.h:124:1: error: inlining failed in call to always_inline ‘int vfprintf(FILE*, const char*, __va_list_tag*)’: function body can be overwritten at link time vfprintf (FILE *__restrict __stream, ^ /var/tmp/portage/sys-devel/gcc-4.9.0/work/gcc-4.9.0/libitm/util.cc:35:31: error: called from here vfprintf (stderr, fmt, list); ^ Makefile:517: recipe for target 'util.lo' failed make[4]: *** [util.lo] Error 1 make[4]: *** Waiting for unfinished jobs.... Debian/Ubuntu have disable fortify in there patchset for gcc 4.9.0 and lib itm