Bug detailed description:
-------------------------
It fails on arch i386 and works well on x86_64. 
Build igt, run make:
  CC       pm_pc8.o
pm_pc8.c: In function ‘submit_blt_cmd’:
pm_pc8.c:1017:23: error: cast from pointer to integer of different size [-Werror=pointer-to-int-cast]
  objs[1].relocs_ptr = (uint64_t) relocs;
                       ^
pm_pc8.c:1019:24: error: cast from pointer to integer of different size [-Werror=pointer-to-int-cast]
  execbuf.buffers_ptr = (uint64_t) objs;
                        ^
pm_pc8.c: In function ‘gem_execbuf_stress_subtest’:
pm_pc8.c:1162:24: error: cast from pointer to integer of different size [-Werror=pointer-to-int-cast]
  execbuf.buffers_ptr = (uint64_t) objs;
                        ^
cc1: some warnings being treated as errors
make[2]: *** [pm_pc8.o] Error 1
make[2]: Leaving directory `/GFX/Test/Intel_gpu_tools/intel-gpu-tools/tests'
make[1]: *** [all-recursive] Error 1
make[1]: Leaving directory `/GFX/Test/Intel_gpu_tools/intel-gpu-tools'
make: *** [all] Error 2

Bisect shows: c527a29325bb914ea994ba8d9b38668ac8a39798 is the first bad commit
commit c527a29325bb914ea994ba8d9b38668ac8a39798
Author:     Paulo Zanoni <paulo.r.zanoni@intel.com>
AuthorDate: Fri Nov 8 19:19:35 2013 -0200
Commit:     Paulo Zanoni <paulo.r.zanoni@intel.com>
CommitDate: Fri Nov 8 19:38:03 2013 -0200

    tests/pm_pc8: add some GEM tests

    These are more complete tests than the previous test_batch() one. We
    test CPU/GTT mmaps, pread/pwrite and batch buffers.

    Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>