Description of problem: Running KT1 on target system listed in comment#1, /kernel/distribution/ltp/lite had the following failure: move_pages04 1 TFAIL : move_pages04.c:131: status[1] is -14 Version-Release number of selected component (if applicable): distro: RHEL-7.3-20160707.2 Server aarch64 kernel: 4.5.0-3.el7 task: /kernel/distribution/ltp/lite 20160510-7 How reproducible: unknown Steps to Reproduce: 1. Install target system listed in comment#1 with RHEL-7.3-20160707.2 Server aarch64 2. Install kernel-4.5.0-3.el7 3. run /kernel/distribution/ltp/lite 20160510-7 Actual results: https://beaker.engineering.redhat.com/jobs/1435778 https://beaker.engineering.redhat.com/recipes/2939557#task44065244 http://beaker-archive.app.eng.bos.redhat.com/beaker-logs/2016/08/14357/1435778/2939557/44065244/217420387/resultoutputfile.log ---<-snip->--- move_pages04 FAIL 1 ---<-snip->--- http://beaker-archive.app.eng.bos.redhat.com/beaker-logs/2016/08/14357/1435778/2939557/44065244/RHELKT1LITE.FILTERED.run.log ---<-snip->--- <<<test_start>>> tag=move_pages04 stime=1470443314 cmdline="move_pages.sh 04" contacts="" analysis=exit <<<test_output>>> move_pages04 1 TFAIL : move_pages04.c:131: status[1] is -14 <<<execution_status>>> initiation_status="ok" duration=0 termination_type=exited termination_id=1 corefile=no cutime=1 cstime=0 <<<test_end>>> ---<-snip->--- Expected results: move_pages04 subtask completes successfully Additional info:
All, Issue persists and is isolated to cavium-thunderx2 hosts. distro: RHEL-7.3-20160817.1 Server aarch64 kernel: 4.5.0-5.el7 host: cavium-thunderx2 https://beaker.engineering.redhat.com/recipes/2993875#task44806350 http://beaker-archive.app.eng.bos.redhat.com/beaker-logs/2016/08/14624/1462439/2993875/44806350/RHELKT1LITE.FILTERED.run.log ---<-snip->--- <<<test_start>>> tag=move_pages04 stime=1471986926 cmdline="move_pages.sh 04" contacts="" analysis=exit <<<test_output>>> move_pages04 1 TFAIL : move_pages04.c:131: status[1] is -14 <<<execution_status>>> initiation_status="ok" duration=0 termination_type=exited termination_id=1 corefile=no cutime=1 cstime=0 <<<test_end>>> ---<-snip->--- Best, -pbunyan
Yeah - it'll be specific to NUMA systems, such as ThunderX. Following up.
Looks like test needs an update after following kernel commit: commit d899844e9c98c9c74b4d9926fd3bd66a225f6978 Author: Kirill A. Shutemov <kirill.shutemov.com> Date: Fri Sep 4 15:47:53 2015 -0700 mm: fix status code which move_pages() returns for zero page The manpage for move_pages(2) specifies that status code for zero page is supposed to be -EFAULT. Currently kernel return -ENOENT in this case. follow_page() can do it for us, if we would ask for FOLL_DUMP. The use of FOLL_DUMP also means that the upper layer page tables pages are no longer allocated. Signed-off-by: Kirill A. Shutemov <kirill.shutemov.com> Reviewed-by: Christoph Lameter <cl> Cc: Hugh Dickins <hughd> Signed-off-by: Andrew Morton <akpm> Signed-off-by: Linus Torvalds <torvalds>
(In reply to Jan Stancek from comment #6) > Looks like test needs an update after following kernel commit: Posted: http://lists.linux.it/pipermail/ltp/2016-October/002728.html # uname -r 4.8.0-1.el7.test.x86_64 # numactl -H available: 2 nodes (0-1) node 0 cpus: 0 1 2 3 node 0 size: 5933 MB node 0 free: 3952 MB node 1 cpus: 4 5 6 7 node 1 size: 6047 MB node 1 free: 3759 MB node distances: node 0 1 0: 10 20 1: 20 10 # ./move_pages04 move_pages04 1 TPASS : status[1] has expected value
(In reply to Jan Stancek from comment #7) > (In reply to Jan Stancek from comment #6) > > Looks like test needs an update after following kernel commit: > > Posted: http://lists.linux.it/pipermail/ltp/2016-October/002728.html https://github.com/linux-test-project/ltp/commit/d539a004dde3b760f610ef7cae90a96de8489ec8
So is this a test problem that only is exhibited on NUMA enabled systems? If so, the bz should be moved somewhere so it can be addressed properly. Thanks.
(In reply to John Feeney from comment #9) > So is this a test problem that only is exhibited on NUMA enabled systems? Yes. > If so, the bz should be moved somewhere so it can be addressed properly. It's been already addressed, both upstream (comment 8) and internally (tests/kernel commit 970e332d14a). Closing as NOTABUG.