+++ This bug was initially created as a clone of Bug #1952505 +++ Description of problem: The version of DejaGnu in Fedora is old and has a bug that causes problems for GCC testing. See details in https://lists.gnu.org/archive/html/bug-dejagnu/2018-07/msg00000.html https://gcc.gnu.org/bugzilla/show_bug.cgi?id=100203 Version-Release number of selected component (if applicable): dejagnu-1.6.1-7.fc32 dejagnu-1.6.1-8.fc33 dejagnu-1.6.1-9.fc34 How reproducible: Always. Actual results: Tests that deadlock and timeout do not get killed. Expected results: Tests are killed if they time out. Additional info: The patch at https://lists.gnu.org/archive/html/bug-dejagnu/2018-07/msg00002.html is in the 1.6.2 release from November 2018, but has never made it into Fedora. Other distros either ship 1.6.2 or include that patch. --- Additional comment from Jonathan Wakely on 2021-04-22 12:19:01 UTC --- P.S. I'd be happy to do the update myself if that's OK. --- Additional comment from Jakub Jelinek on 2021-04-27 17:24:10 UTC --- gcc needs to remove some tests otherwise these hangs kill the builds almost always. Otherwise it is only sometimes.
If we do it, verification might be done by running a build of the packages that depend on dejagnu build-time: # dnf repoquery --disablerepo=* --enablerepo=*-source --whatrequires dejagnu --quiet automake-0:1.16.2-6.el9.src binutils-0:2.35.2-13.el9.src compat-libgfortran-48-0:4.8.5-36.4.el9.1.src dwz-0:0.14-3.el9.src environment-modules-0:5.0.1-1.el9.src findutils-1:4.8.0-5.el9.src gcc-0:11.2.1-6.1.el9.src libffi-0:3.4.2-7.el9.src ltrace-0:0.7.91-43.el9.src mingw-binutils-0:2.37-1.el9.src A tmt test is now proposed in Fedora, the same should work fine for RHEL: https://src.fedoraproject.org/rpms/dejagnu/pull-request/3#request_diff
A PR for rebasing to 1.6.3 is ready at https://gitlab.com/redhat/centos-stream/rpms/dejagnu/-/merge_requests/6 However, I've realized that staring RHEL-9, the ownership of dejagnu was transferred to the sst_pt_gcc_glibc and this bug was not re-assigned properly. Doing it now. Marek, assigning to you as you're the default assignee for dejagnu in rhel-9 -- can you confirm you can take care of it further? Some sanity test already exists and PR is prepared, so should not need much, but I don't feel our SST should work on the component instead of yours.
OK, I'll finish up the rebase.
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 (new packages: dejagnu), and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://access.redhat.com/errata/RHBA-2022:4305