On Wed, Feb 26, 2020 at 11:53:37AM +0000, Cristian Marussi wrote: > Sorry I lost track of this thread....tested now on v5.6rc3 on arm64 on a local > setup (not proper CI) with CONFIG_LKDTM=y and solves the above issue. (missing CONFIG_LKDTM?) Great! > Looking at the general run though I have anyway a bunch of SKIP similar to Naresh > > ./run_kselftest.sh 2>/dev/null | grep "SKIP" > not ok 1 selftests: lkdtm: PANIC.sh # SKIP > not ok 6 selftests: lkdtm: LOOP.sh # SKIP > not ok 7 selftests: lkdtm: EXHAUST_STACK.sh # SKIP > not ok 8 selftests: lkdtm: CORRUPT_STACK.sh # SKIP > not ok 9 selftests: lkdtm: CORRUPT_STACK_STRONG.sh # SKIP > not ok 15 selftests: lkdtm: UNSET_SMEP.sh # SKIP > not ok 16 selftests: lkdtm: DOUBLE_FAULT.sh # SKIP > not ok 18 selftests: lkdtm: OVERWRITE_ALLOCATION.sh # SKIP > not ok 19 selftests: lkdtm: WRITE_AFTER_FREE.sh # SKIP > not ok 21 selftests: lkdtm: WRITE_BUDDY_AFTER_FREE.sh # SKIP > not ok 26 selftests: lkdtm: SOFTLOCKUP.sh # SKIP > not ok 27 selftests: lkdtm: HARDLOCKUP.sh # SKIP > not ok 28 selftests: lkdtm: SPINLOCKUP.sh # SKIP > not ok 29 selftests: lkdtm: HUNG_TASK.sh # SKIP > not ok 59 selftests: lkdtm: REFCOUNT_TIMING.sh # SKIP > not ok 60 selftests: lkdtm: ATOMIC_TIMING.sh # SKIP > > BUT, if I look at one LKDTM test script > (lkdtm/USERCOPY_STACK_FRAME_TO.sh): > > ... > # If the test is commented out, report a skip > if echo "$test" | grep -q '^#' ; then > ... > > and in fact: > > root@sqwt-ubuntu:/opt/ksft# cat lkdtm/tests.txt | grep ^# > #PANIC > #LOOP Hangs the system > #EXHAUST_STACK Corrupts memory on failure > #CORRUPT_STACK Crashes entire system on success > #CORRUPT_STACK_STRONG Crashes entire system on success > #OVERWRITE_ALLOCATION Corrupts memory on failure > #WRITE_AFTER_FREE Corrupts memory on failure > #WRITE_BUDDY_AFTER_FREE Corrupts memory on failure > #SOFTLOCKUP Hangs the system > #HARDLOCKUP Hangs the system > #SPINLOCKUP Hangs the system > #HUNG_TASK Hangs the system > #REFCOUNT_TIMING timing only > #ATOMIC_TIMING timing only > > so only two of the above tests are in fact unexpectedly skipped. Yup, that is operating as expected. :) > not ok 15 selftests: lkdtm: UNSET_SMEP.sh # SKIP > not ok 16 selftests: lkdtm: DOUBLE_FAULT.sh # SKIP These are x86-specific, so that is correct as well. Thanks for following up! -- Kees Cook
