From Fedora Project Wiki

Revision as of 15:26, 30 November 2009 by Valce (talk | contribs) (Example: SystemTap)

Understanding the (Proposed) Change to DSO Linking


The default behaviour for ld is to not link objects that are listed as dependencies of another linked object. This is dangerous if the other object is ever changed to occlude the object on which your program depended, causing your program to break without any change to your code.

For example : has:

 NEEDED            Shared library: []
 NEEDED            Shared library: []

Under the old system, a program that links with libxml2 and uses dlopen may not link with libdl, and a program that links with libxml2 and uses gzopen may not link with libz. While these programs will work, they will break if libxml2 is ever changed to omit the dependency on libdl/libz.

What's the difference?

For example (courtesy Roland McGrath):

 ==> foo1.c <==
 #include <stdio.h>
 extern int foo ();
 main ()
   printf ("%d\n", foo ());
 ==> foo2.c <==
 extern int foo ();
 int bar () { return foo (); }
 ==> foo3.c <==
 int foo () { return 0; }

Prepare position-independent code:

gcc -g -fPIC -c foo1.c foo2.c foo3.c


gcc -shared -o foo3.o

Generate, linking

gcc -shared -o foo2.o

The proposed change will affect the next step: Creating foo1.


A call to gcc will succeed quietly, even though the link to is only implicit.

 gcc -o foo1 foo1.o -Wl,--rpath-link=.


The call to gcc will fail, prompting the user to explicitly link the required shared object.

 gcc -o foo1 foo1.o -Wl,--rpath-link=.
/usr/bin/ld: foo1.o: undefined reference to symbol 'foo'
/usr/bin/ld: note: 'foo' is defined in DSO ./ so try adding it to the linker command line

So, the difference is whether you can refer to a symbol that's in a DSO that you didn't list explicitly in your link line, but that is a DT_NEEDED dependency of one of those (or recursively of those, I think).

The big difference is that with the proposed change in place, ld will no longer skip linking needed libraries by default. The current default behaviour will lead ld to skip linking with a library if it is listed as a needed by another library that the program uses. In abstract terms, if libA is needed by libB and your program requires both libA and libB, your program may only link to libB. Then if another version of libB comes out that does not list libA as a needed library, then a recompilation will mysteriously break.

What do I do?

If you encounter this error, the error message will prompt you to explicitly link to the DSO that you need. From the foo example, adding will get rid of the error:

gcc -o foo1 foo1.o -Wl,--rpath-link=.

Example: SystemTap

This example uses the SystemTap package and shows how to find and fix any errors that may be created by the proposed change to DSO Linkage.

Checked out SystemTap using CVS. You can use fedora-cvs systemtap or point your cvs to:

In the devel folder, run 'make srpm' to produce a source rpm. Open one of the .cfg files in /etc/mock and add the following between [fedora] and [local]:


You may wish to rename the .cfg file to fedora-test and set config_opts['root']='fedora-test' at the top of the file. This will create a mock config file that uses the new version of ld.

Now run mock -r fedora-test systemtap-1.0.src.rpm. The following error will appear in the log file :

RPM build errors:
/usr/bin/ld.bfd: stap.13165.test: undefined reference to symbol 'pthread_cancel@@GLIBC_2.0'
/usr/bin/ld.bfd: note: 'pthread_cancel@@GLIBC_2.0' is defined in DSO /lib/ so try adding it to the linker command line
/lib/ could not read symbols: Invalid operation
*** /usr/bin/ld: ld behavior mismatch! ***
*** /usr/bin/ld.bfd succeeeded ***
*** /usr/bin/ld.bfd --no-add-needed exits 1 ***
*** arguments: --eh-frame-hdr --build-id -m elf_i386 --hash-style=gnu -dynamic-linker /lib/ -pie -o stap 
/usr/lib/gcc/i686-redhat-linux/4.4.2/../../../Scrt1.o /usr/lib/gcc/i686- redhat-linux/4.4.2/../../../crti.o /usr/lib/gcc/i686-redhat-linux/4.4.2/crtbeginS.o
-L/usr/lib/gcc/i686-redhat-linux/4.4.2 -L/usr/lib/gcc/i686-redhat-linux/4.4.2 -L/usr/lib/gcc/i686-redhat- linux/4.4.2/../../.. -z relro -z now stap-main.o 
stap-parse.o stap-staptree.o stap-elaborate.o stap-translate.o stap-tapsets.o stap-buildrun.o stap-loc2c.o stap-hash.o stap-mdfour.o  stap-cache.o stap-util.o
stap-coveragedb.o stap-dwarf_wrappers.o stap-tapset-been.o stap-tapset-procfs.o stap-tapset-timers.o stap-tapset-perfmon.o stap-tapset-mark.o stap-tapset-itrace.o
stap-tapset-utrace.o stap-task_finder.o stap-dwflpp.o stap-rpm_finder.o --start-group -ldw -lebl --end-group -lelf -lsqlite3 -lrpm -ldl -lstdc++ -lm -lgcc_s -lgcc 
-lc -lgcc_s -lgcc /usr/lib /gcc/i686-redhat-linux/4.4.2/crtendS.o /usr/lib/gcc/i686-redhat-linux/4.4.2/../../../crtn.o 
collect2: ld returned 1 exit status
make[2]: *** [stap] Error 1
make[1]: *** [all-recursive] Error 1
make: *** [all] Error 2

This message means that Systemtap currently uses /lib/ without explicitly linking to it. To fix this, go into the file and add -lpthread where it is needed:

 LIBS="$LIBS -ldl -lpthread"
 rpm_LIBS="-lrpm -lpthread"

For testing purposes, it may help to untar the tar.gz created by make srpm and modifying the configuration files in there before re-tarring it and running make srpm again.

Once -lpthread has been added to the correct location, mock -r fedora-test systemtap-1.0.src.rpm should complete with no further errors.