The practice of creating symbolic links is somewhat fragile. Always

make copies instead.

There's too many times that we can't run the new binaries with old
libraries. Making the links when things are known to be 'safe' is a
nice optimization, but a copy of all the binaries is only 30MB, so
saving the copies at the cost of increased support when new symbols
are added and used as part of the bootstrap seems to be unwise.

There may be additional optimizations possible here, especially for
!FreeBSD hosts. However, that's beyond the scope of the problem I'm
trying to fix with make failing mid-way through an installworld across
change r363679. This optimization there caused us to run a new binary
with an old library once a new make was installed due to the symbolic
link. One could just copy make, but then other binaries fail as well,
so rather than play whack-a-mole, I opted to take us back to the old
way.  Before r340157 or so we did copies (thogh of a lot fewer
artifacts), and we didn't have issues like this.

Reviewed by: arichards@
Differential Revision: https://reviews.freebsd.org/D25967
This commit is contained in:
Warner Losh 2020-08-07 16:26:56 +00:00
parent 1f325602e4
commit 33176cdc87
3 changed files with 25 additions and 21 deletions

View File

@ -2293,13 +2293,12 @@ ${_bt}-links: .PHONY
.for _tool in ${_bootstrap_tools_links}
${_bt}-link-${_tool}: .PHONY .MAKE
@if [ ! -e "${WORLDTMP}/legacy/bin/${_tool}" ]; then \
@rm -f "${WORLDTMP}/legacy/bin/${_tool}"; \
source_path=`which ${_tool}`; \
if [ ! -e "$${source_path}" ] ; then \
echo "Cannot find host tool '${_tool}'"; false; \
fi; \
ln -sfnv "$${source_path}" "${WORLDTMP}/legacy/bin/${_tool}"; \
fi
cp -f "$${source_path}" "${WORLDTMP}/legacy/bin/${_tool}"
${_bt}-links: ${_bt}-link-${_tool}
.endfor

View File

@ -26,6 +26,12 @@ NOTE TO PEOPLE WHO THINK THAT FreeBSD 13.x IS SLOW:
disable the most expensive debugging functionality run
"ln -s 'abort:false,junk:false' /etc/malloc.conf".)
20200807:
Makefile.inc has been updated to work around the issue documented in
20200729. It was a case where the optimization of using symbolic links
to point to binaries created a situation where we'd run new binaries
with old libraries starting midway through the installworld process.
20200729:
r363679 has redefined some undefined behavior in regcomp(3); notably,
extraneous escapes of most ordinary characters will no longer be

View File

@ -119,26 +119,25 @@ _host_abs_tools_to_symlink= ${_make_abs}:make ${_make_abs}:bmake
host-symlinks:
@echo "Linking host tools into ${DESTDIR}/bin"
.for _tool in ${_host_tools_to_symlink}
@if [ ! -e "${DESTDIR}/bin/${_tool}" ]; then \
source_path=`which ${_tool}`; \
@source_path=`which ${_tool}`; \
if [ ! -e "$${source_path}" ] ; then \
echo "Cannot find host tool '${_tool}'"; false; \
fi; \
ln -sfnv "$${source_path}" "${DESTDIR}/bin/${_tool}"; \
fi
rm -f "${DESTDIR}/bin/${_tool}"; \
cp -f "$${source_path}" "${DESTDIR}/bin/${_tool}"
.endfor
.for _tool in ${_host_abs_tools_to_symlink}
@source_path="${_tool:S/:/ /:[1]}"; \
target_path="${DESTDIR}/bin/${_tool:S/:/ /:[2]}"; \
if [ ! -e "$${target_path}" ] ; then \
if [ ! -e "$${source_path}" ] ; then \
echo "Host tool '${src_path}' is missing"; false; \
fi; \
ln -sfnv "$${source_path}" "$${target_path}"; \
fi
rm -f "$${target_path}"; \
cp -f "$${source_path}" "$${target_path}"
.endfor
.if exists(/usr/libexec/flua)
ln -sf /usr/libexec/flua ${DESTDIR}/usr/libexec/flua
rm -f ${DESTDIR}/usr/libexec/flua
cp -f /usr/libexec/flua ${DESTDIR}/usr/libexec/flua
.endif
# Create all the directories that are needed during the legacy, bootstrap-tools