Description of problem: RHEL-7.3-20160629.n.0 ltptest hits EWD at madvise06 Version-Release number of selected component (if applicable): distro: RHEL-7.3-20160629.n.0 kernel: 4.5.0-0.44.el7 task: /kernel/distribution/ltp/lite 20160510-7 How reproducible: unknown 1. Install ARM system with distro: RHEL-7.3-20160629.n.0 2. Install kernel 4.5.0-0.44.el7 3. Run task /kernel/distribution/ltp/lite 20160510-7 Actual results: https://beaker.engineering.redhat.com/jobs/1388712 https://beaker.engineering.redhat.com/recipes/2836998#task42573224 http://beaker-archive.app.eng.bos.redhat.com/beaker-logs/2016/06/13887/1388712/2836998/console.log ---<-snip->--- [ 2002.927199] ltptest madvise06 start [ 2008.033105] restraintd[4229]: *** Current Time: Thu Jun 30 18:56:47 2016 Localwatchdog at: Thu Jun 30 21:46:46 2016 ---<-snip->--- Expected results: successful completion on ltp-lite task Additional info:
This has likely same cause as: Bug 1351249 - [RHELSA-7.3][ltp-lite] 4.5.0-0.41.el7 ltptest hits EWD at mtest01w madvise06 wants to allocate and use entire RAM: https://github.com/linux-test-project/ltp/blob/master/testcases/kernel/syscalls/madvise/madvise06.c
If Comment #3 is correct and bz1351249 has the same cause, bz1351249 is in VERIFIED due to a fix in kernel-aarch64-4.5.0-0.45.el7. It looks like the bz in question (1352669) was found in kernel-4.5.0-0.44.el7 (oh, so close) so this needs a retest.
All, Looking at current KT1 testing, madvise06 is not currently being run: https://beaker.engineering.redhat.com/recipes/3098701#task45981010 http://beaker-archive.app.eng.bos.redhat.com/beaker-logs/2016/09/15191/1519140/3098701/45981010/226997148/resultoutputfile.log Looks like it was intentionally removed based on this BZ (Bug 1352669):) See here: http://pkgs.devel.redhat.com/cgit/tests/kernel/commit/?id=7b657eb745fc4 pbunyan will rerun recent ltp-lite task, reserve host, and manually run mdavise06 subtask. I will follow up when testing completes. pbunyan high5s jbastian :) best, -pbunyan
All, Manual re-testing /kernel/distribution/ltp/lite mdavise06 subtask looks good. LiWang (liwang) - please remove mdavise06 from /distribution/ltp/include/knownissue.sh for aarch64: Here is suggested patch: ---<-snip->--- --- ./knownissue.sh.orig 2016-09-27 16:38:40.953579388 -0400 +++ ./knownissue.sh 2016-09-27 16:39:21.745065364 -0400 @@ -123,12 +123,6 @@ function knownissue_unfix() # disable futex_wake04 until we fix Bug 1087896 echo "futex_wake04" >> ${kn_unfix} - - kvercmp "$cver" '4.5.0-0.41.el7.aarch64' - if [ $kver_ret -ge 0 ]; then - # Bug 1352669 - [RHELSA-7.3] ltptest hits EWD at madvise06 - echo "madvise06" >> ${kn_unfix} - fi fi ---<-snip->--- ================== testing reference: ================== ------------------- manual test results ------------------- mnt/testarea/ltp/testcases/bin/mfind /mnt -name mtest0* /mnt/testarea/ltp/testcases/bin/mtest01 /mnt/tests/kernel/distribution/ltp/include/patches/20150903/mtest01-sigchld.patch /mnt/tests/kernel/distribution/ltp/lite/patches/20150903/mtest01.c /mnt/tests/kernel/distribution/ltp/lite/ltp-full-20160920/testcases/kernel/mem/mtest05 /mnt/tests/kernel/distribution/ltp/lite/ltp-full-20160920/testcases/kernel/mem/mtest06 /mnt/tests/kernel/distribution/ltp/lite/ltp-full-20160920/testcases/kernel/mem/mtest01 /mnt/tests/kernel/distribution/ltp/lite/ltp-full-20160920/testcases/kernel/mem/mtest01/mtest01.c /mnt/tests/kernel/distribution/ltp/lite/ltp-full-20160920/testcases/kernel/mem/mtest01/mtest01 /mnt/tests/kernel/distribution/ltp/lite/ltp-full-20160920/testcases/kernel/mem/mtest07 [root@apm-mustang-ev3-11 lite]# find /mnt -name madvise06 /mnt/testarea/ltp/testcases/bin/madvise06 /mnt/tests/kernel/distribution/ltp/lite/ltp-full-20160920/testcases/kernel/syscalls/madvise/madvise06 [root@apm-mustang-ev3-11 lite]# /mnt/testarea/ltp/testcases/bin/mtest01 mtest01 0 TINFO : Total memory already used on system = 1777600 kbytes mtest01 1 TPASS : 0 kbytes allocated only. [root@apm-mustang-ev3-11 lite]# /mnt/testarea/ltp/testcases/bin/mtest01 -p80 mtest01 0 TINFO : Total memory already used on system = 1777984 kbytes mtest01 0 TINFO : Total memory used needed to reach maximum = 8307712 kbytes mtest01 0 TINFO : Filling up 80% of ram which is 6529728 kbytes mtest01 1 TPASS : 6529728 kbytes allocated only. [root@apm-mustang-ev3-11 lite]# /mnt/testarea/ltp/testcases/bin/mtest01 -p80 w mtest01 0 TINFO : Total memory already used on system = 1778240 kbytes mtest01 0 TINFO : Total memory used needed to reach maximum = 8307712 kbytes mtest01 0 TINFO : Filling up 80% of ram which is 6529472 kbytes mtest01 1 TPASS : 6529472 kbytes allocated only. [root@apm-mustang-ev3-11 lite]# /mnt/testarea/ltp/testcases/bin/madvise06 tst_test.c:756: INFO: Timeout per run is 0h 05m 00s madvise06.c:65: INFO: dropping caches madvise06.c:139: INFO: SwapCached (before madvise): 686592 madvise06.c:153: INFO: SwapCached (after madvise): 995520 madvise06.c:155: PASS: Regression test pass Summary: passed 1 failed 0 skipped 0 warnings 0 [root@apm-mustang-ev3-11 lite]# /mnt/testarea/ltp/testcases/bin/madvise06 tst_test.c:756: INFO: Timeout per run is 0h 05m 00s madvise06.c:65: INFO: dropping caches madvise06.c:139: INFO: SwapCached (before madvise): 810560 madvise06.c:153: INFO: SwapCached (after madvise): 1118976 madvise06.c:155: PASS: Regression test pass Summary: passed 1 failed 0 skipped 0 warnings 0 [root@apm-mustang-ev3-11 lite]# /mnt/testarea/ltp/testcases/bin/madvise06 tst_test.c:756: INFO: Timeout per run is 0h 05m 00s madvise06.c:65: INFO: dropping caches madvise06.c:139: INFO: SwapCached (before madvise): 962560 madvise06.c:153: INFO: SwapCached (after madvise): 1270976 madvise06.c:155: PASS: Regression test pass Summary: passed 1 failed 0 skipped 0 warnings 0 [root@apm-mustang-ev3-11 lite]# uname -a Linux apm-mustang-ev3-11.lab.eng.rdu.redhat.com 4.5.0-13.el7.aarch64 #1 SMP Thu Sep 22 09:30:49 EDT 2016 aarch64 aarch64 aarch64 GNU/Linux [root@apm-mustang-ev3-11 lite]# Best, -pbunyan
Created attachment 1205314 [details] comment6-suggested-patch
Paul, rather than removing madvise06 completely from the known issues list, I moved it to the fixed section, so it will still be disabled on kernels older than 4.5.0-0.45.el7. See: http://pkgs.devel.redhat.com/cgit/tests/kernel/commit/?id=a35e6a35259d http://pkgs.devel.redhat.com/cgit/tests/kernel/commit/?id=5af5b61bb937 (typo fix) I started a new run of ltp/lite to verify madvise06 runs: https://beaker.engineering.redhat.com/jobs/1536509
(In reply to Jeff Bastian from comment #8) > Paul, rather than removing madvise06 completely from the known issues list, > I moved it to the fixed section, so it will still be disabled on kernels > older than 4.5.0-0.45.el7. > > See: > http://pkgs.devel.redhat.com/cgit/tests/kernel/commit/?id=a35e6a35259d > http://pkgs.devel.redhat.com/cgit/tests/kernel/commit/?id=5af5b61bb937 (typo > fix) > > I started a new run of ltp/lite to verify madvise06 runs: > https://beaker.engineering.redhat.com/jobs/1536509 Ok Jeff. I see that madvise06 ran and passed :) https://beaker.engineering.redhat.com/jobs/1536509 http://beaker-archive.app.eng.bos.redhat.com/beaker-logs/2016/10/15365/1536509/3135826/46469895/229290409/test_log-RHELKT1LITE.FILTERED.log ---<-snip->--- madvise06 PASS 0 ---<-snip->--- kernel-distribution-ltp-include-20160920-3 looks good. Thank you. best, -pbunyan
I will close this since the original problem is fixed in the current kernel and madvise06 is running again in the Beaker LTP tasks.