Paul Eggert wrote in
<[email protected]>:
|On 2025-09-13 07:56, Steffen Nurpmeso wrote:
|>|What does "strace cp -a xb xc" say? I see this:
|>
|> geteuid() = 1000
|> openat(AT_FDCWD, "c", O_RDONLY|O_PATH|O_DIRECTORY) = -1 ENOENT \
|> (No such file or directory)
|> newfstatat(AT_FDCWD, "b", {st_mode=S_IFLNK|0777, st_size=1, ...}, \
|> AT_SYMLINK_NOFOLLOW) = 0
|> newfstatat(AT_FDCWD, "c", 0x7ffdf869dea0, AT_SYMLINK_NOFOLLOW) \
|> = -1 ENOENT (No such file or directory)
|> readlink("b", "a", 2) = 1
|> symlinkat("a", AT_FDCWD, "c") = 0
|> newfstatat(AT_FDCWD, "c", {st_mode=S_IFLNK|0777, st_size=1, ...}, \
|> AT_SYMLINK_NOFOLLOW) = 0
|> lchown("c", 1000, 1000) = 0
|> fchmodat2(AT_FDCWD, "c", 0777, AT_SYMLINK_NOFOLLOW) = -1 ENOSYS \
|> (Function not implemented)
|> ^^
|
|So far the bug report has been about lchownat. But now you're saying
Well i am just a user, ok? I am saying nothing, i had a build
problem caused by cp(1), the kbd guy thankfully fixed that via
-dPR not -a, i then saw it is from a code line that says
lchownat() in coreutils' cp.c, that was all i know. I stopped
looking into (the horrible and terrible-style) GNU source code
when i switched from Linux to FreeBSD 4.7, which was a fantastic
experience in comparison!
|that cp fails because fchmodat2 fails. So it sounds like we have been
|barking up the wrong tree by worrying about the chown-related syscalls,
|and that we should instead worry about chmod-related syscalls.
Seems like that.
|> I do not know how
|>
|> $ grep -r fchmodat2 /usr/include/
|> /usr/include/bits/syscall.h:#ifdef __NR_fchmodat2
|> /usr/include/bits/syscall.h:# define SYS_fchmodat2 __NR_fchmodat2
|>
|> gets into there, the kernel definitely does not have that as via
|> grep. Generated at glibc build time via
|>
|> # install kernel headers
|> make -C /usr/src/linux mrproper
|> make -C /usr/src/linux INSTALL_HDR_PATH=$PKG/usr headers_install
|
|Recent glibc fchmodat uses fchmodat2 if available. It assumes fchmodat2
|is available if __ASSUME_FCHMODAT2 is nonzero, and __ASSUME_FCHMODAT2 is
|nonzero if you build glibc with headers from Linux kernel 6.6 or later.
Like i said i have the Linux 6.1.* series, and i build glibc with
the Linux kernel that i run (which is a local change, the default
Pkgfile downloads and installs 6.1.86 just for that:
# install kernel headers
#make -C $SRC/linux-$_kernel_version mrproper
#make -C $SRC/linux-$_kernel_version INSTALL_HDR_PATH=$PKG/usr
headers_install
make -C /usr/src/linux mrproper
make -C /usr/src/linux INSTALL_HDR_PATH=$PKG/usr headers_install
...
$SRC/$name-$version/configure \
--prefix=/usr \
--libdir=/usr/lib \
--libexecdir=/usr/lib \
--with-headers=$PKG/usr/include \
--enable-kernel=5.10 \
--enable-stack-protector=strong \
--enable-bind-now \
--enable-multi-arch \
--disable-profile \
--disable-werror \
--without-gd)
|It sounds like your glibc was built with __ASSUME_FCHMODAT2 = 1, but is
|running atop an older kernel lacking fchmodat2. If so, that would
|explain the abovementioned symptoms, and it would be a bug in how your
|glibc was built, not a coreutils bug.
No. It was build exactly like so:
# Description: The C library used in the GNU system
# URL: https://www.gnu.org/software/libc/
# Maintainer: CRUX System Team, core-ports at crux dot nu
name=glibc
version=2.40
release=3
_kernel_version=6.1.86
source=(https://ftp.gnu.org/gnu/glibc/glibc-$version.tar.xz
#https://www.kernel.org/pub/linux/kernel/v6.x/linux-$_kernel_version.tar.xz
glibc-$version.patch
hosts resolv.conf nsswitch.conf host.conf ld.so.conf
locale-gen locale.gen.in)
build() {
# install kernel headers
#make -C $SRC/linux-$_kernel_version mrproper
#make -C $SRC/linux-$_kernel_version INSTALL_HDR_PATH=$PKG/usr
headers_install
make -C /usr/src/linux mrproper
make -C /usr/src/linux INSTALL_HDR_PATH=$PKG/usr headers_install
chown root:root $PKG/usr
patch -p1 -d $SRC/$name-${version:0:4} -i $SRC/$name-$version.patch
mkdir $SRC/build
cd $SRC/build
export libc_cv_slibdir=/lib
$SRC/$name-$version/configure \
--prefix=/usr \
--libdir=/usr/lib \
--libexecdir=/usr/lib \
--with-headers=$PKG/usr/include \
--enable-kernel=5.10 \
--enable-stack-protector=strong \
--enable-bind-now \
--enable-multi-arch \
--disable-profile \
--disable-werror \
--without-gd
make
#make check
make install_root=$PKG install
install -m 0644 $SRC/$name-${version:0:4}/nscd/nscd.conf $PKG/etc
install -d $PKG/var/db/nscd
install -m 0644
$SRC/{hosts,resolv.conf,nsswitch.conf,host.conf,ld.so.conf} $PKG/etc
ln -sf ../usr/share/zoneinfo/UTC $PKG/etc/localtime
install -d $PKG/etc/ld.so.conf.d $PKG/usr/lib/locale
touch $PKG/etc/ld.so.cache
install -Dm755 $SRC/locale-gen $PKG/usr/sbin/locale-gen
install -Dm644 $SRC/locale.gen.in $PKG/etc/locale.gen
sed -e '1,3d' -e 's|/| |g' -e 's|\\| |g' -e 's|^|#|g' \
$SRC/$name-$version/localedata/SUPPORTED >> $PKG/etc/locale.gen
LC_ALL=C ./elf/ld.so --library-path . $PKG/usr/bin/localedef \
--force --quiet \
--inputfile=$SRC/$name-$version/localedata/locales/C \
--charmap=$SRC/$name-$version/localedata/charmaps/UTF-8 \
$PKG/usr/lib/locale/C.UTF-8 || true
rm -r $PKG/usr/share/locale \
$PKG/usr/bin/{tzselect,zdump} \
$PKG/usr/sbin/zic
install -Dm644 $SRC/$name-$version/intl/locale.alias \
$PKG/usr/share/locale/locale.alias
}
The patch was
-rw-rw-r-- 1 ports ports 1186584 Jul 24 17:01 glibc-2.40.patch
and created like
: ${version?'Need $version'}
curl -L -o .glibc-$version.patch \
'https://sourceware.org/git/?p=glibc.git;a=commitdiff_plain;hp=refs/tags/glibc-'$version';h=refs/heads/release/'$version'/master'
(with $version being 2.40).
That is all i know.
--End of <[email protected]>
--steffen
|
|Der Kragenbaer, The moon bear,
|der holt sich munter he cheerfully and one by one
|einen nach dem anderen runter wa.ks himself off
|(By Robert Gernhardt)
