On Tue, 2008-07-29 at 19:41 +0100, Mel Gorman wrote:
> Regression tests of 2.0-pre1 on an older distro are reporting FAIL for the
> ELFMAP-related tests. As the version of binutils is too old to support the
> feature, these are not true failures. "Expected FAILs" are confusing as it
> gets difficult to tell what a real and expected failure is over time. This
> patch identifies when binutils is too old and skips the ELFMAP tests.
>
> Comments?
>
> Signed-off-by: Mel Gorman <[EMAIL PROTECTED]>
> ---
> run_tests.sh | 57 +++++++++++++++++++++++++++++++++++++++++----------------
> 1 file changed, 41 insertions(+), 16 deletions(-)
>
> diff --git a/tests/run_tests.sh b/tests/run_tests.sh
> index c98179d..1327de6 100755
> --- a/tests/run_tests.sh
> +++ b/tests/run_tests.sh
> @@ -49,6 +49,21 @@ function check_linkhuge_tests() {
> fi
> }
>
> +function check_elfmap_tests() {
> + # Binutils 2.17 or later are required for the ELFMAP tests to complete
> + # successfully. If the version is too old, we do not bother running the
> + # tests as they would report FAIL when it's not a true failure
> +
> + ELFMAP_SKIP=0
> + BINUTILS_VERSION=`ld -v | sed -e 's/[a-zA-Z() \t]//g'`
> + MAJOR=`echo $BINUTILS_VERSION | cut -d. -f1`
> + MINOR=`echo $BINUTILS_VERSION | cut -d. -f2`
> +
> + if [ $MAJOR -lt 2 ] || [ $MAJOR -eq 2 -a $MINOR -lt 17 ]; then
> + ELFMAP_SKIP=1
> + fi
> +}
Hmm. It isn't actually invalid to run the old linkhuge tests in all
cases with binutils 2.17 or newer. In fact, the only place I know that
the linker scripts are broken is powerpc with binutils 2.17 or newer.
The reason this combination doesn't work is because the system ldscripts
began using the "SPECIAL" keyword. I would suggest the following
function (from an earlier patch posted by me) to determine whether to
skip the linkhuge tests:
+# Up-front checks for the remapping test cases:
+function check_linkhuge_tests() {
+ # In some circumstances, our linker scripts are known to be broken and
+ # they will produce binaries with undefined runtime behavior. In those
+ # cases don't bother running the xNNN.linkhuge tests. This checks if the
+ # system linker scripts use the SPECIAL keyword (for placing the got and
+ # plt). Our linker scripts do not use SPECIAL and are thus broken when the
+ # system scripts use it.
+ ld --verbose | grep -q SPECIAL
+ if [ $? -eq 0 ]; then
+ LINKHUGE_SKIP=1
+ fi
+}
> run_test_bits () {
> BITS=$1
> shift
> @@ -79,6 +94,14 @@ maybe_run_linkhuge_test () {
> fi
> }
>
> +maybe_run_elfmap_test () {
> + if [ "$ELFMAP_SKIP" != "1" ]; then
> + run_test "$@"
> + else
> + skip_test "$@"
> + fi
> +}
> +
> preload_test () {
> run_test LD_PRELOAD=libhugetlbfs.so "$@"
> }
> @@ -104,16 +127,16 @@ elflink_test () {
>
> elflink_rw_test() {
> # Basic tests: None, Read-only, Write-only, Read-Write, exlicit disable
> - run_test linkhuge_rw
> - run_test HUGETLB_ELFMAP=R linkhuge_rw
> - run_test HUGETLB_ELFMAP=W linkhuge_rw
> - run_test HUGETLB_ELFMAP=RW linkhuge_rw
> - run_test HUGETLB_ELFMAP=no linkhuge_rw
> + maybe_run_elfmap_test linkhuge_rw
> + maybe_run_elfmap_test HUGETLB_ELFMAP=R linkhuge_rw
> + maybe_run_elfmap_test HUGETLB_ELFMAP=W linkhuge_rw
> + maybe_run_elfmap_test HUGETLB_ELFMAP=RW linkhuge_rw
> + maybe_run_elfmap_test HUGETLB_ELFMAP=no linkhuge_rw
>
> # Test we don't blow up if HUGETLB_MINIMAL_COPY is disabled
> - run_test HUGETLB_MINIMAL_COPY=no HUGETLB_ELFMAP=R linkhuge_rw
> - run_test HUGETLB_MINIMAL_COPY=no HUGETLB_ELFMAP=W linkhuge_rw
> - run_test HUGETLB_MINIMAL_COPY=no HUGETLB_ELFMAP=RW linkhuge_rw
> + maybe_run_elfmap_test HUGETLB_MINIMAL_COPY=no HUGETLB_ELFMAP=R
> linkhuge_rw
> + maybe_run_elfmap_test HUGETLB_MINIMAL_COPY=no HUGETLB_ELFMAP=W
> linkhuge_rw
> + maybe_run_elfmap_test HUGETLB_MINIMAL_COPY=no HUGETLB_ELFMAP=RW
> linkhuge_rw
> }
>
> elfshare_test () {
> @@ -154,14 +177,14 @@ elflink_and_share_test () {
>
> elflink_rw_and_share_test () {
> clear_hpages
> - run_test HUGETLB_ELFMAP=R HUGETLB_SHARE=1 linkhuge_rw
> - run_test HUGETLB_ELFMAP=R HUGETLB_SHARE=1 linkhuge_rw
> + maybe_run_elfmap_test HUGETLB_ELFMAP=R HUGETLB_SHARE=1 linkhuge_rw
> + maybe_run_elfmap_test HUGETLB_ELFMAP=R HUGETLB_SHARE=1 linkhuge_rw
> clear_hpages
> - run_test HUGETLB_ELFMAP=W HUGETLB_SHARE=1 linkhuge_rw
> - run_test HUGETLB_ELFMAP=W HUGETLB_SHARE=1 linkhuge_rw
> + maybe_run_elfmap_test HUGETLB_ELFMAP=W HUGETLB_SHARE=1 linkhuge_rw
> + maybe_run_elfmap_test HUGETLB_ELFMAP=W HUGETLB_SHARE=1 linkhuge_rw
> clear_hpages
> - run_test HUGETLB_ELFMAP=RW HUGETLB_SHARE=1 linkhuge_rw
> - run_test HUGETLB_ELFMAP=RW HUGETLB_SHARE=1 linkhuge_rw
> + maybe_run_elfmap_test HUGETLB_ELFMAP=RW HUGETLB_SHARE=1 linkhuge_rw
> + maybe_run_elfmap_test HUGETLB_ELFMAP=RW HUGETLB_SHARE=1 linkhuge_rw
> clear_hpages
> }
>
> @@ -251,6 +274,8 @@ functional_tests () {
>
> # Run the remapping tests' up-front checks
> check_linkhuge_tests
> +check_elfmap_tests
> +
> # Original elflink tests
> elflink_test HUGETLB_VERBOSE=0 linkhuge_nofd # Lib error msgs expected
> elflink_test linkhuge
>
> -------------------------------------------------------------------------
> This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
> Build the coolest Linux based applications with Moblin SDK & win great prizes
> Grand prize is a trip for two to an Open Source event anywhere in the world
> http://moblin-contest.org/redirect.php?banner_id=100&url=/
> _______________________________________________
> Libhugetlbfs-devel mailing list
> [email protected]
> https://lists.sourceforge.net/lists/listinfo/libhugetlbfs-devel
>
--
Adam Litke - (agl at us.ibm.com)
IBM Linux Technology Center
-------------------------------------------------------------------------
This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
Build the coolest Linux based applications with Moblin SDK & win great prizes
Grand prize is a trip for two to an Open Source event anywhere in the world
http://moblin-contest.org/redirect.php?banner_id=100&url=/
_______________________________________________
Libhugetlbfs-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/libhugetlbfs-devel
