Readjusted linker issues etc.

robert baker robertmbaker at
Sun May 17 18:09:10 PDT 2009

As I mentioned the other day I did go ahead and build the temp system
again. This time I used linux- along with the corresponding
grsec patch. I have had no issues I can trace back to the kernel
includes. So at least early on that looks good.

I was experiencing problems with the re-adjusted linker before I
rebuilt the temporary system, but didn't look at it much because I was
trying to document the reboot procedure. These same issues did crop
back up on my latest rebuild. Taking a look  at the ld-new that we
copy into /tools/bin during binutils-pass2 I found it was considerably
smaller than expected. Reading the file with a text editor shed some
light on my problems.

# ld-new - temporary wrapper script for .libs/ld-new
# Generated by (GNU libtool 1.2435 2007/03/18 18:44:42) 2.1a
# The ld-new program cannot be directly executed until all the libtool
# libraries that it depends on are installed.
# This wrapper script should never be moved out of the build directory.
# If it is, it will not operate correctly.

Going back to my binutils obj/ directory and executing the below
commad installed a working  re-adjusted linker.
make -C ld install

I haven't explored a solution that doesn't require keeping the build
directory from the previous chapter just yet. Instead I have been
reviewing the test suite failures that I am experiencing later in the
build. Running the following for the binutils test suite results in a
small handfull of errors.
make CFLAGS="-fPIC -fno-stack-protector -U_FORTIFY_SOURCE -nonow
-nopie" -k check

binutils tests
FAIL: run objcopy of executable
FAIL: run stripped executable
FAIL: run stripped executable with saving a symbol
FAIL: objdump -W
FAIL: readelf -wa (compressed)
ld tests
FAIL: ld-elf/seg
FAIL: vers4
FAIL: Check --gc-section/-r/-u
FAIL: ld-i386/hidden2
FAIL: S-records
FAIL: S-records with constructors

The good news is with similar CFLAGS, and
LDFLAGS="-Wl,-no-fatal-warning" all the glibc tests pass.

Robert Baker

More information about the hlfs-dev mailing list