diff options
Diffstat (limited to 'open_issues')
| -rw-r--r-- | open_issues/glibc.mdwn | 136 | ||||
| -rw-r--r-- | open_issues/glibc_madvise_vs_static_linking.mdwn | 5 |
2 files changed, 118 insertions, 23 deletions
diff --git a/open_issues/glibc.mdwn b/open_issues/glibc.mdwn index e8279139..3160c86f 100644 --- a/open_issues/glibc.mdwn +++ b/open_issues/glibc.mdwn @@ -1,5 +1,5 @@ -[[!meta copyright="Copyright © 2007, 2008, 2010, 2011 Free Software Foundation, -Inc."]] +[[!meta copyright="Copyright © 2007, 2008, 2010, 2011, 2012 Free Software +Foundation, Inc."]] [[!meta license="""[[!toggle id="license" text="GFDL 1.2+"]][[!toggleable id="license" text="Permission is granted to copy, distribute and/or modify this @@ -36,8 +36,8 @@ git log --reverse --pretty=fuller --stat=$COLUMNS,$COLUMNS -p -C --cc ..sourcewa --> -Last reviewed up to the [[Git mirror's 9d65ea3a9b83ac3961229ba296a7caf90abce68d -(2011-11-17) sources|source_repositories/glibc]]. +Last reviewed up to the [[Git mirror's d40c5d54cb551acba4ef1617464760c5b3d41a14 +(2012-02-28) sources|source_repositories/glibc]]. * t/dup3 @@ -208,6 +208,16 @@ Last reviewed up to the [[Git mirror's 9d65ea3a9b83ac3961229ba296a7caf90abce68d Then define `_POSIX_MAPPED_FILES`, `_POSIX_SYNCHRONIZED_IO`. + * `sys/epoll.h` + + * `sys/eventfd.h` + + * `sys/inotify.h` + + * `sys/signalfd.h` + + * `sys/timerfd.h` + * Create `t/cleanup_kernel-features.h`. * Add tests from Linux kernel commit messages for `t/dup3` et al. @@ -373,13 +383,15 @@ Last reviewed up to the [[Git mirror's 9d65ea3a9b83ac3961229ba296a7caf90abce68d * [OK] 22a89187139a9083ca73989bfd11597e0f85cb61 (`malloc: Remove all kinds of unused configuration options and dead code.`). `NO_STARTER` changes (should be OK). - * [OK] 02d46fc4b969e25e4ba0c54aa95fa98d7279bd05 (`Simplify malloc - initialization`). + * [high] `pagesize`, 02d46fc4b969e25e4ba0c54aa95fa98d7279bd05 (`Simplify + malloc initialization`); aebae0537dcb408100b88c6b7647a7e858c43237, `BZ + 11929`. Is this all kosher for us? See [[!message-id + "87mxd9hl2n.fsf@kepler.schwinge.homeip.net"]]. * [OK] 83cd14204559abbb52635006832eaf4d2f42514a (`Remove --wth-tls option, TLS support is required`). * a7c8e6a1478de9f990b11e5e853318ccbe4330f2 (`Fix invalid conversion in __cmsg_nxthdr`). Probably just a C++ thing and not relevant for us; - see [[message-id "87r52nk1kx.fsf@kepler.schwinge.homeip.net"]]. + see [[!message-id "87r52nk1kx.fsf@kepler.schwinge.homeip.net"]]. * [high] `__ctype_init`, fd5bdc0924e0cfd1688b632068c1b26f3b0c88da. Probably need to mirror `init-first.c` change. * [high] `__attribute__ ((__leaf__))`, `BZ #13344`, @@ -391,6 +403,23 @@ Last reviewed up to the [[Git mirror's 9d65ea3a9b83ac3961229ba296a7caf90abce68d edc5984d4d18296d7aa3d8f4ed8f7336a743170e + 57769839788e2c62b68d9dfbf4b35052321278ba. <http://gcc.gnu.org/onlinedocs/gcc-4.6.1/gcc/Function-Attributes.html>. + * [low] implement `timespec_get`, + 74033a2507841cf077e31221de2481ff30b43d51. + * [low] `__volatile`, `BZ #13553`, + a784e502472fb3a1afa4d01a47c66b52d23e00f6: + `sysdeps/mach/i386/machine-lock.h:typedef __volatile int + __spin_lock_t;`, `sysdeps/mach/powerpc/machine-lock.h:typedef + __volatile long int __spin_lock_t;` + * [high] 6ee65ed6ddbf04402fad0bec6aa9c73b9d982ae4, hopefully OK. + * [high] `crti`/`crtn`, 3add8e1353d62d77fdd9b4ca363cdfe7006b0efb, + 0e7dfaef514bbb3ec08934c6f7f42953bc149257, should just work. + * 7638c0fda568726f52ee5a88e1eadcddcd9fa290, `EHWPOISON`, does + [[!message-id + "Pine.LNX.4.64.1202191652540.3253@digraph.polyomino.org.uk"]] apply for + us? + * [low] `conformtest`, 3134156779108fe8b46e0f4cd60d837572faaa93 + + 4efeffc1d583597e4f52985b9747269e47b754e2 + + d94a4670800de6e8f088b8630ad5142866127980 -- what does it do for us? # Build @@ -405,10 +434,18 @@ sources|source_repositories/glibc]], run on coulomb.SCHWINGE. $ make install_root=/INVALID 2>&1 | tee log_build_ [...] -This takes up around 400 MiB and needs roughly 120 min on coulomb.SCHWINGE. +This takes up around 500 MiB and needs roughly X min on kepler.SCHWINGE and 100 +min on coulomb.SCHWINGE (GCC 4.4/4.5/4.6). <!-- - $ (make install_root=/INVALID && touch .go-install) 2>&1 | tee log_build_ && test -f .go-install && (make install_root="$PWD".install install && touch .go-check) 2>&1 | tee log_install && test -f .go-check && make -k install_root=/INVALID check TIMEOUTFACTOR=100 2>&1 | tee log_check + $ (make install_root=/INVALID && touch .go-install) 2>&1 | tee log_build_ && test -f .go-install && (make install_root="$PWD".install install && touch .go-check) 2>&1 | tee log_install && test -f .go-check && ln -s /usr/lib/i386-*gnu/libstdc++.so.6 /lib/i386-*gnu/libpthread-stubs.so.0 /lib/i386-*gnu/libgcc_s.so.1 mach/libmachuser.so.1 hurd/libhurduser.so.0.3 ./ && make -k install_root=/INVALID check TIMEOUTFACTOR=100 2>&1 | tee log_check + +Mask out gcc-4.X (with possibly a backslash before the dot), GCC 4.5's column +output for (warning, error) messages, GCC 4.6's `[-Wsomething]` or `[enabled by +default]` identifiers which warning flag triggered. + + $ for f in log_*; do sed -e 's%gcc-4\\\?.[456]%[GCC]%g' -e 's%g++-4\\\?.[456]%[G++]%g' -e 's%\(:[0-9]\+:\)[0-9]\+:%\1%' -e 's% \[\(-W[a-z-]\+\|enabled by default\)\]$%%' < "$f" > "$f".nv; done + $ find ./ -name \*.o -o -name \*.os -o -name \*.oS | while read f; do ~/tmp/gcc/git/contrib/compare-debug --preserve ../Roger_Whittaker.build-gcc-4.4-486.O/"$f" "$f"; done 2>&1 | less $ while read f; do (readelf -a "$f" && objdump -xDrtw "$f") > N && (cd ../Roger_Whittaker.build-gcc-4.4-486.O/ && readelf -a "$f" && objdump -xDrtw "$f") > O && diff -u O N | less; done $ find ./ -name \*.o -o -name \*.os -o -name \*.oS | while read f; do readelf -h "$f" | grep OS/ABI | (read a b && [ x"$b" != x'UNIX - System V' ] && echo "### $f: $b"); done @@ -420,18 +457,48 @@ This takes up around 400 MiB and needs roughly 120 min on coulomb.SCHWINGE. TODO. + * With GCC 4.5, there's a ton of these warnings: + + hurd/hurd.h: In function '__hurd_fail': + hurd/hurd.h:73: warning: case value '0' not in enumerated type 'error_t' + + ... as well as a few individual instances: + + hurdselect.c: In function '_hurd_select': + hurdselect.c:265: warning: case value '0' not in enumerated type 'error_t' + get-host.c: In function '_hurd_get_host_config': + get-host.c:38: warning: case value '0' not in enumerated type 'error_t' + hurdmsg.c: In function '_S_msg_get_init_ints': + hurdmsg.c:186: warning: case value '0' not in enumerated type 'error_t' + hurdmsg.c: In function '_S_msg_set_init_ints': + hurdmsg.c:273: warning: case value '0' not in enumerated type 'error_t' + intr-msg.c: In function '_hurd_intr_rpc_mach_msg': + intr-msg.c:363: warning: case value '0' not in enumerated type 'error_t' + sysdeps/mach/hurd/setitimer.c: In function 'timer_thread': + sysdeps/mach/hurd/setitimer.c:117: warning: case value '0' not in enumerated type 'error_t' + sysdeps/mach/hurd/wait4.c: In function '__wait4': + sysdeps/mach/hurd/wait4.c:40: warning: case value '0' not in enumerated type 'error_t' + sysdeps/mach/hurd/fork.c: In function '__fork': + sysdeps/mach/hurd/fork.c:423: warning: case value '0' not in enumerated type 'error_t' + sysdeps/mach/hurd/spawni.c: In function '__spawni': + sysdeps/mach/hurd/spawni.c:600: warning: case value '0' not in enumerated type 'error_t' + sysdeps/mach/hurd/setpriority.c: In function 'setonepriority': + sysdeps/mach/hurd/setpriority.c:66: warning: case value '0' not in enumerated type 'error_t' + sysdeps/mach/hurd/ioctl.c: In function 'send_rpc': + sysdeps/mach/hurd/ioctl.c:177: warning: case value '0' not in enumerated type 'error_t' + sysdeps/mach/hurd/ioctl.c: In function '__ioctl': + sysdeps/mach/hurd/ioctl.c:306: warning: case value '0' not in enumerated type 'error_t' + # Install TODO. -<!-- $ make install_root="$PWD".install install 2>&1 | tee log_install [...] -This takes up around 50 MiB, and needs roughly 1 min on kepler.SCHWINGE and 3 +This takes up around 100 MiB, and needs roughly X min on kepler.SCHWINGE and 18 min on coulomb.SCHWINGE. ---> ## Analysis @@ -471,25 +538,51 @@ Comparing the results files, [[sum_linux]] to [[sum_hurd]]: There is quite a baseline of failures. + ### Additional Failures Compared to Debian $ bash ~/tmp/glibc/debian/eglibc-2.13/debian/testsuite-checking/convertlog.sh log_check > log_check.filtered $ bash ~/tmp/glibc/debian/eglibc-2.13/debian/testsuite-checking/compare.sh ~/tmp/glibc/debian/eglibc-2.13/debian/testsuite-checking/expected-results-i486-gnu-libc log_check.filtered - * `bug-atexit3.out`, `debug/tst-chk4`, `debug/tst-lfschk4`, - `debug/tst-lfschk5`, `debug/tst-lfschk6`, `debug/tst-chk5`, - `debug/tst-chk6` + * `bug-atexit3.out`, `debug/tst-chk4`, `debug/tst-chk5`, `debug/tst-chk6`, + `debug/tst-lfschk4`, `debug/tst-lfschk5`, `debug/tst-lfschk6` dlopen failed: libstdc++.so.6: cannot open shared object file: No such file or directory + See [[!message-id "20090420002344.11798.qmail@s461.sureserver.com"]]. + Hacked around with `ln -s /usr/lib/i386-*gnu/libstdc++.so.6 + /lib/i386-*gnu/libpthread-stubs.so.0 /lib/i386-*gnu/libgcc_s.so.1 ./`. + This is a bug in the glibc test harness. Should probably use some + `configure` magic akin to the `fixincludes` stuff (`gcc-4.4 + -print-file-name=libstdc++.so.6`, etc.). + + * `debug/tst-chk4`, `debug/tst-chk5`, `debug/tst-chk6`, `debug/tst-lfschk4`, + `debug/tst-lfschk5`, `debug/tst-lfschk6` + + Fail in the same way as the C ones, `tst-chk1..3`. + * `io/ftwtest`, `posix/globtest`, `iconvdata/iconv-test`, `intl/tst-gettext`, `malloc/tst-mtrace`, `elf/tst-pathopt`, `iconvdata/tst-tables`, `grp/tst_fgetgrent`, `dlfcn/tststatic`, `dlfcn/tststatic2`, - `posix/wordexp-tst`, `localedata/bug-setlocale1.out` + `posix/wordexp-tst`, `localedata/bug-setlocale1.out`, `posix/tst-getconf` /home/thomas/tmp/glibc/tschwinge/Roger_Whittaker.build-gcc-4.4-486.O/io/ftwtest: error while loading shared libraries: libmachuser.so.1: cannot open shared object file: No such file or directory - Should be using the build-directory one anyway! + Looking into `localedata/bug-setlocale1.c`, it is clear what it going on: + only the root of the build directory is added for `--library-path`, but + none of the other directories that are additionally used. This is a bug in + the glibc test harness. Hacked around by `ln -s mach/libmachuser.so.1 + hurd/libhurduser.so.0.3 ./`. Hopefully the other instances are similar. + + * `posix/tst-getconf` + + Ends with: + + getconf POSIX_ALLOC_SIZE_MIN /: /home/thomas/tmp/glibc/tschwinge/Roger_Whittaker.build-gcc-4.4-486/posix/getconf: pathconf: /: Invalid argument + + * `dlfcn/tststatic`, `dlfcn/tststatic2` + + No output, SEGFAULT. * `math/test-idouble`, `math/test-ifloat`, `math/test-ildoubl`, `math/test-ldouble` @@ -524,11 +617,6 @@ There is quite a baseline of failures. tst-ether_line.c:19: error: 'ETH_ALEN' undeclared (first use in this function) - * `posix/tst-getconf` - - /bin/sh -e tst-getconf.sh /home/thomas/tmp/glibc/tschwinge/Roger_Whittaker.build-gcc-4.4-486.O/ /home/thomas/tmp/glibc/tschwinge/Roger_Whittaker.build-gcc-4.4-486.O/elf/ ld.so.1 - make[2]: *** [/home/thomas/tmp/glibc/tschwinge/Roger_Whittaker.build-gcc-4.4-486.O/posix/tst-getconf.out] Error 127 - * `time/tst-mktime2` tst-mktime2.c:132: error: 'INT_MAX' undeclared (first use in this function) @@ -556,6 +644,10 @@ There is quite a baseline of failures. /media/erich/home/thomas/tmp/glibc/tschwinge/Roger_Whittaker/string/test-strnlen.c:87: undefined reference to `MIN' + * `assert/test-assert.out` + + Fails sometimes... + * `stdlib/bug-getcontext.out` getcontext failed, errno: 1073741902. diff --git a/open_issues/glibc_madvise_vs_static_linking.mdwn b/open_issues/glibc_madvise_vs_static_linking.mdwn index bfda0f74..43b8314f 100644 --- a/open_issues/glibc_madvise_vs_static_linking.mdwn +++ b/open_issues/glibc_madvise_vs_static_linking.mdwn @@ -11,6 +11,8 @@ License|/fdl]]."]]"""]] [[!tag open_issue_glibc]] +[[!sourceware_bug 4822]]. + $ echo 'int main() {}' | gcc -o /dev/null -static -x c - /usr/lib/gcc/i486-gnu/4.4.5/../../../libcrt.a(malloc.o): In function `_int_free': (.text+0xdc3): warning: warning: madvise is not implemented and will always fail @@ -24,7 +26,8 @@ case of MADV_DONTNEED), but may influence its performance. The kernel is free to ignore the advice.* (`man madvise`), so we may simply want to turn it into a no-op in glibc, avoiding the link-time warning. -2011-07: This is what Samuel has done for Debian glibc. +2011-07: This is what Samuel has [done for Debian +glibc](http://anonscm.debian.org/viewvc/pkg-glibc/glibc-package/trunk/debian/patches/hurd-i386/local-madvise_warn.diff). GCC c5db973fdab3db3e13db575e5650c0bcfd3630f4 (2011-10-17) makes use of this. As we now export the symbol (and `MADV_DONTNEED`, too), GCC will no longer |