From 465725bcde0f98efd5cd6bf305060e1cac0a6741 Mon Sep 17 00:00:00 2001 From: Thomas Schwinge Date: Tue, 6 Dec 2016 15:45:26 +0100 Subject: binutils-gdb: GCC 6 --- open_issues/binutils.mdwn | 29 ++++++++++++++++++++--------- open_issues/gcc.mdwn | 6 ++++++ 2 files changed, 26 insertions(+), 9 deletions(-) (limited to 'open_issues') diff --git a/open_issues/binutils.mdwn b/open_issues/binutils.mdwn index ed4db498..afc0405f 100644 --- a/open_issues/binutils.mdwn +++ b/open_issues/binutils.mdwn @@ -702,7 +702,7 @@ hard-coded in `gdb/remote.c`, and `global_thread_id_to_ptid` used instead of and laplace.SCHWINGE. $ export LC_ALL=C - $ ../W._C._Handy/configure --prefix="$PWD".install --enable-gold --enable-plugins --with-sysroot=/ SHELL=/bin/dash CC=gcc-4.9 CXX=g++-4.9 --disable-werror --disable-build-with-cxx 2>&1 | tee log_build + $ ../W._C._Handy/configure --prefix="$PWD".install --enable-gold --enable-plugins --with-sysroot=/ SHELL=/bin/dash CC=gcc-6 CXX=g++-6 --disable-werror --disable-build-with-cxx 2>&1 | tee log_build [...] $ make 2>&1 | tee log_build_ [...] @@ -717,13 +717,13 @@ MIG-generated stub files; thus no `-Werror` until that is resolved. below. This needs to be addressed soon, as GDB will soon *require a C++ compiler*, [[!message-id "ab08f2c8-3b52-3a7a-2236-89833bd09129@redhat.com"]]. -This takes up around 1.5 GiB, and runs for [[16 min|performance#measure]] on -kepler.SCHWINGE and [[21 min|performance#measure]] on laplace.SCHWINGE. +This takes up around 1.7 GiB, and runs for [[19 min|performance#measure]] on +kepler.SCHWINGE and [[24 min|performance#measure]] on laplace.SCHWINGE. @@ -830,8 +830,8 @@ formats, and more emulation vectors. $ make install 2>&1 | tee log_install [...] -This takes up around 220 MiB, and runs for [[0 min|performance#measure]] on -kepler.SCHWINGE and [[2 min|performance#measure]] on laplace.SCHWINGE. +This takes up around 250 MiB, and runs for [[0 min|performance#measure]] on +kepler.SCHWINGE and [[3 min|performance#measure]] on laplace.SCHWINGE. ## Analysis @@ -843,10 +843,21 @@ kepler.SCHWINGE and [[2 min|performance#measure]] on laplace.SCHWINGE. # Testsuite - $ make -k check 2>&1 | tee log_test +At least for GNU/Linux, Debian's GCC 6 is configured with +`--enable-default-pie` -- which the *gdb*, *gold*, and *ld* testsuites don't +like, so we have to cancel that for testing. +`CC_FOR_TARGET` and `CXX_FOR_TARGET` in `RUNTESTFLAGS` is considered by the *gdb* testsuite, +`CC_FOR_TARGET` and `CXX_FOR_TARGET` on the `make` command line by the *ld* testsuite, +`CC` and `CXX` on the `make` command line by the *gold* testsuite +but only when invoked (from the top level) as `make check-gold` and not when run as part of `make check`... +I couldn't figure out a command-line variant for *gold* testsuite, +so instead we apply an ugly hack to `gold/testsuite/Makefile`. + + $ sed -e 's%^\([A-Z]\+ = g..-6\)%override \1 -fno-pie -fno-PIE -no-pie%g' -i gold/testsuite/Makefile + $ make -k check RUNTESTFLAGS=CC_FOR_TARGET='gcc-6\ -fno-pie\ -fno-PIE\ -no-pie'\ CXX_FOR_TARGET='g++-6\ -fno-pie\ -fno-PIE\ -no-pie' CC_FOR_TARGET='gcc-6 -fno-pie -fno-PIE -no-pie' CXX_FOR_TARGET='g++-6 -fno-pie -fno-PIE -no-pie' 2>&1 | tee log_test [...] -This runs for [[23 min|performance#measure]] on kepler.SCHWINGE and [[73 +This runs for [[24 min|performance#measure]] on kepler.SCHWINGE and [[86 min|performance#measure]] on laplace.SCHWINGE. When running `make -k check 2>&1 | tee log_test`, at the end of the testsuite diff --git a/open_issues/gcc.mdwn b/open_issues/gcc.mdwn index c742f06e..7850d024 100644 --- a/open_issues/gcc.mdwn +++ b/open_issues/gcc.mdwn @@ -1061,6 +1061,12 @@ a050099a416f013bda35832b878d9a57b0cbb231 (gcc-6-branch branch point; ee1bfdb0d611ed2e1057071f39589e6ebcd87983 `Remove __seg_tls before first release`. + * `--enable-default-pie` + + On 2016-12-05 noticed that per `gcc -v` output, `gcc version 6.2.1 20161124 + (Debian 6.2.1-5)` for GNU/Linux is configured with + `--enable-default-pie`, but GNU/Hurd is not. + # Build -- cgit v1.2.3