Revision as of 03:46, 27 June 2011

Overview

These are notes on the current state of the Arch Linux toolchain (gcc, glibc, binutils, linux-api-headers and their dependencies). This will record all known testsuite failures so that they can be investigated. Future toolchain rebuilds should not suffer from any additional (unexpected) testsuite failures and ideally should fix the current issues...

Current Toolchain

binutils

• Version: 2.21-9 (20110605)

Test suite failures (i686 and x86_64):

FAIL: Build libdnew1a.so with --Bsymbolic-functions --dynamic-list-cpp-new
FAIL: Build libdnew1b.so with --dynamic-list-data --dynamic-list-cpp-new
FAIL: Run with libnew1a.so
FAIL: Run with libnew1b.so
FAIL: plugin claimfile lost symbol

• libnew1?.so
  • Testsuite issue: http://sourceware.org/ml/binutils-cvs/2011-02/msg00119.html
• plugin claimfile lost symbol
  • Testsuite issue: http://sourceware.org/ml/binutils-cvs/2010-11/msg00048.html

Test suite failures (x86_64 only):

FAIL: final_layout.sh
FAIL: exception_separate_shared_12_test
FAIL: exception_separate_shared_21_test
FAIL: ifuncmain1static
FAIL: ifuncmain1picstatic
FAIL: ifuncmain2static
FAIL: ifuncmain2picstatic
FAIL: ifuncmain4static
FAIL: ifuncmain4picstatic
FAIL: ifuncmain5static
FAIL: ifuncmain5picstatic
FAIL: ifuncmain7static
FAIL: ifuncmain7picstatic

(All these also failed on i686 with 20110312)

• final_layout.sh
  • http://sourceware.org/ml/binutils/2011-03/msg00466.html
• exception_separate_shared_{12,21}_test
  • failure due to missing symbols with -Wl,--as-needed: http://sourceware.org/bugzilla/show_bug.cgi?id=12353
• ifuncmain*
  • no idea...

glibc

• Version: 2.13-5 (20110117)

make[2]: *** [/build/src/glibc-build/stdlib/tst-putenv.out] Error 1
make[2]: *** [/build/src/glibc-build/nptl/tst-cancel24] Error 1
make[2]: *** [/build/src/glibc-build/nptl/tst-tls6.out] Error 1
make[2]: *** [/build/src/glibc-build/rt/tst-cpuclock2.out] Error 1
make[2]: *** [/build/src/glibc-build/elf/noload.out] Error 1
make[2]: *** [/build/src/glibc-build/elf/nodelete2.out] Error 1
make[2]: *** [/build/src/glibc-build/elf/next.out] Error 1
make[2]: *** [/build/src/glibc-build/elf/reldep5.out] Error 127
make[2]: *** [/build/src/glibc-build/elf/reldep6.out] Error 1
make[2]: *** [/build/src/glibc-build/elf/reldep8.out] Error 134
make[2]: *** [/build/src/glibc-build/elf/unload4.out] Error 127
make[2]: *** [/build/src/glibc-build/elf/order2.out] Error 1

• rt/tst-cpuclock2.out is a "timing" issue and can be ignored
• the rest are be due to building the test-suite with -Wl,--as-needed.

Note: x86_64 futex bug (worked around in PKGBUILD...)

• http://www.cygwin.com/ml/libc-help/2010-12/msg00017.html
• http://sourceware.org/bugzilla/show_bug.cgi?id=12403

gcc

• Version: 4.6.0-6 (20110603)

Test suite failures (i686 and x86_64):

FAIL: libgomp.graphite/force-parallel-1.c execution test

FAIL: libmudflap.c/pass49-frag.c execution test
FAIL: libmudflap.c/pass49-frag.c output pattern test
FAIL: libmudflap.c/pass49-frag.c execution test
FAIL: libmudflap.c/pass49-frag.c output pattern test
FAIL: libmudflap.c/pass49-frag.c (-static) execution test
FAIL: libmudflap.c/pass49-frag.c (-static) output pattern test
FAIL: libmudflap.c/pass49-frag.c (-static) execution test
FAIL: libmudflap.c/pass49-frag.c (-static) output pattern test
FAIL: libmudflap.c/pass49-frag.c (-O2) execution test
FAIL: libmudflap.c/pass49-frag.c (-O2) output pattern test
FAIL: libmudflap.c/pass49-frag.c (-O2) execution test
FAIL: libmudflap.c/pass49-frag.c (-O2) output pattern test
FAIL: libmudflap.c/pass49-frag.c (-O3) execution test
FAIL: libmudflap.c/pass49-frag.c (-O3) output pattern test
FAIL: libmudflap.c/pass49-frag.c (-O3) execution test
FAIL: libmudflap.c/pass49-frag.c (-O3) output pattern test

FAIL: 20_util/enable_shared_from_this/cons/constexpr.cc (test for excess errors)
FAIL: 20_util/shared_ptr/cons/constexpr.cc (test for excess errors)
FAIL: 20_util/unique_ptr/cons/constexpr.cc (test for excess errors)
FAIL: 20_util/weak_ptr/cons/constexpr.cc (test for excess errors)
FAIL: 30_threads/future/cons/constexpr.cc (test for excess errors)
FAIL: 30_threads/shared_future/cons/constexpr.cc (test for excess errors)

• libgomp.graphite:
  • Seen regularly on i686/x86_64 (in H. J. Lu's test runs) but no bug report...
• libmudflap failures:
  • http://gcc.gnu.org/bugzilla/show_bug.cgi?id=20003
• 20_util/* and 30_threads/*
  • Due to use of -pipe in C{XX}FLAGS: http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48565

Test suite failures (i686 only):

FAIL: gfortran.dg/graphite/vect-pr40979.f90  -O  scan-tree-dump-times vect "vectorized 1 loops" 1

FAIL: libmudflap.c/fail40-frag.c output pattern test
FAIL: libmudflap.c/pass51-frag.c execution test
FAIL: libmudflap.c/pass51-frag.c execution test
FAIL: libmudflap.c/fail40-frag.c (-static) output pattern test
FAIL: libmudflap.c/pass51-frag.c (-static) execution test
FAIL: libmudflap.c/pass51-frag.c (-static) execution test
FAIL: libmudflap.c/fail40-frag.c (-O2) output pattern test
FAIL: libmudflap.c/pass51-frag.c (-O2) execution test
FAIL: libmudflap.c/pass51-frag.c (-O2) execution test
FAIL: libmudflap.c/fail40-frag.c (-O3) output pattern test
FAIL: libmudflap.c/pass51-frag.c (-O3) execution test
FAIL: libmudflap.c/pass51-frag.c (-O3) execution test
FAIL: libmudflap.cth/pass39-frag.c (-O2) (rerun 14) execution test
FAIL: libmudflap.cth/pass39-frag.c (-O2) (rerun 14) output pattern test

• gfortran.dg/graphite/vect-pr40979.f90
  • http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48497

Test suite failures (x86_64 only):

FAIL: ./tmp.go execution,  -O2 -g

linux-api-headers

• Version: 2.6.39.1-1

Dependencies

The following packages were used in generating the test suite results for the current toolchain:

cloog-0.16.2-1
isl-0.06-1
gmp-5.0.2-1
libmpc-0.9-1
mpfr-3.0.1-1
ppl-0.11.2-1
zlib-1.2.5-3

Any updates to these packages must pass their respective test suite completely before entering [testing]. In general it is not necessary to rerun the toolchain packages test suite on updates of these packages unless an soname bump has occurred.