Hide Forgot
Created attachment 477723 [details] objects and script to reproduce the crash on rhel6.0 x86_64 Description of problem: Testing a build on rhel6, I noticed the following: gcc -o gen-LINUX_V6/prosody_thread ../libutil/gen-LINUX_V6/generic_io.o gen-LINUX_V6/prosody_thread.o -lpthread 2>&1 | t0 errs/gen-LINUX_V6/prosody_thread collect2: ld terminated with signal 11 [Segmentation fault], core dumped /usr/bin/ld: i386 architecture of input file `../libutil/gen-LINUX_V6/generic_io.o' is incompatible with i386:x86-64 output /usr/bin/ld: i386 architecture of input file `gen-LINUX_V6/prosody_thread.o' is incompatible with i386:x86-64 output Version-Release number of selected component (if applicable): binutils-2.20.51.0.2-5.11.el6.x86_64 How reproducible: everytime, with attached test case Steps to Reproduce: 1. untar attached source, run crash.sh on x86_64 arch 2. 3. Actual results: [root@localhost ld_crash]# ./crash.sh collect2: ld terminated with signal 11 [Segmentation fault], core dumped /usr/bin/ld: i386 architecture of input file `generic_io.o' is incompatible with i386:x86-64 output /usr/bin/ld: i386 architecture of input file `prosody_thread.o' is incompatible with i386:x86-64 output Expected results: Not sure, probably a successful link :) or another error but no crash. Additional info:
Just use -m32.
Anthony Symons, Does using the gcc command line option -m32, fix your problem?
Since RHEL 6.1 External Beta has begun, and this bug remains unresolved, it has been rejected as it is not proposed as exception or blocker. Red Hat invites you to ask your support representative to propose this request, if appropriate and relevant, in the next release of Red Hat Enterprise Linux.
Even though this is a user-error mixing objects of different types, the linker still shouldn't segfault. Instead it should gracefully exit after issuing the error.
Moving to 6.3
Created attachment 550962 [details] Detect attempts to link mixed binaries and prevent the linker from seg-faulting
I have uploaded a proposed patch to fix this bug. It is based on a patch applied to the mainline FSF binutils sources. (Thanks to Jeff Law for extracting this part of the patch for me). The patch also includes a new linker testcase that reproduces the bug without requiring pre-built binaries. With the full patch applied and the linker testsuite run for an x86_64-* targeted set of the binutils the following new entries will appear in the ld.sum file: PASS: build 32-bit object PASS: build 64-bit object PASS: link mixed objects Without the bugfix part of the patch applied (binutils-rh676194.patch) but with the new testcase applied, the following new entries should appear in the ld.sum file: PASS: build 32-bit object PASS: build 64-bit object FAIL: link mixed objects
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory, and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. http://rhn.redhat.com/errata/RHBA-2012-0872.html