LTO and libelf (and FreeBSD)

[Richard Guenther]

On Sat, May 22, 2010 at 10:29 PM, Steve Kargl
<sgk@troutmask.apl.washington.edu> wrote:
> Guys,
>
> I only read the gcc@ archive, so sorry about breaking the thread.
> Testing with gfortran finds
>
> FreeBSD's libelf with no patches.
>
>                === gfortran Summary ===
>
> # of expected passes            34177
> # of unexpected failures        40
> # of expected failures          33
> # of unresolved testcases       37
> # of unsupported tests          266
> gfortran  version 4.6.0 20100522 (experimental) (GCC)
>
>
> FreeBSD's libelf with Kia's patch
>
>                === gfortran Summary ===
>
> # of expected passes            34204
> # of unexpected failures        19
> # of expected failures          33
> # of unresolved testcases       16
> # of unsupported tests          266
> gfortran  version 4.6.0 20100522 (experimental) (GCC)
>
>
> No patches.  libelf from http://www.mr511.de/software/libelf-0.8.12.tar.gz
>
>                === gfortran Summary ===
>
> # of expected passes            34226
> # of unexpected failures        3
> # of expected failures          33
> # of unsupported tests          266
> gfortran  version 4.6.0 20100522 (experimental) (GCC)
>
>
> Richi's lto-elf.c patch with libelf from URL://libelf-0.8.12.tar.gz
>
>                === gfortran Summary ===
>
> # of expected passes            34226
> # of unexpected failures        3
> # of expected failures          33
> # of unsupported tests          266
> gfortran  version 4.6.0 20100522 (experimental) (GCC)
>
>
> FreeBSD's libelf with Kia and Richi's patches.
>
>                === gfortran Summary ===
>
> # of expected passes            34204
> # of unexpected failures        19
> # of expected failures          33
> # of unresolved testcases       16
> # of unsupported tests          266
> /usr/home/kargl/gcc/obj4x/gcc/testsuite/gfortran/../../gfortran  version 4.6.0 20100522 (experimental) (GCC)
>

Hm, so you didn't test FreeBSD's libelf without Kias patch but my GCC patch
applied.  (at least my patch doesn't make the situation worse for any case
it seems)

I would apply my patch as a reasonable workaround for (parts of) the
FreeBSD libelf problems if it alone makes a difference.

Richard.

> Three of the failures are common to all tests.  There are
>
> FAIL: gfortran.dg/array_constructor_11.f90  -O3 -g  (internal compiler error)
> FAIL: gfortran.dg/array_constructor_11.f90  -O3 -g  (test for excess errors)
> FAIL: gfortran.dg/ltrans-7.f90  -O  scan-tree-dump-times ltrans "transformed
>
>
> The 16 remaining failures with FreeBSD's libelf  and Kia patch are of
> the following form.
>
>
> Executing on host: /usr/home/kargl/gcc/obj4x/gcc/testsuite/gfortran/../../gfortran -B/usr/home/kargl/gcc/obj4x/gcc/te
> stsuite/gfortran/../../ f_lto_20091015-1_0.o f_lto_20091015-1_1.o f_lto_20091015-1_2.o  -O0 -fwhopr -w  -B/usr/home/kargl/gcc/obj4x/i386-unknown-freebsd9.0/./libgfortran/.libs -L/usr/home/kargl/gcc/obj4x/i386-unknown-freebsd9.0/./libgfortran/.libs -L/usr/home/kargl/gcc/obj4x/i386-unknown-freebsd9.0/./libgfortran/.libs -L/usr/home/kargl/gcc/obj4x/i386-unknown-freebsd9.0/./libiberty      -o gfortran-dg-lto-20091015-1-01    (timeout = 300)
> lto1: fatal error: elf_update() failed: Layout constraint violation
> compilation terminated.
> lto-wrapper: /usr/home/kargl/gcc/obj4x/gcc/testsuite/gfortran/../../gfortran returned 1 exit status
> collect2: lto-wrapper returned 1 exit status
> compiler exited with status 1
> output is:
> lto1: fatal error: elf_update() failed: Layout constraint violation
> compilation terminated.
> lto-wrapper: /usr/home/kargl/gcc/obj4x/gcc/testsuite/gfortran/../../gfortran returned 1 exit status
> collect2: lto-wrapper returned 1 exit status
>
> FAIL: gfortran.dg/lto/20091015-1 f_lto_20091015-1_0.o-f_lto_20091015-1_2.o link
>
> --
> Steve
>