| Message ID | 1619684694-116827-1-git-send-email-wangxin410@huawei.com |
|---|---|
| State | Accepted |
| Headers | show |
| Series | [v2] controllers/cpuset: Restore the value of cpuset.sched_load_balance | expand |
Hi, > When we run the cpuhotplug03 test case on the arm64 machine, > we conclude that no error occurs when the cpuhotplug03 case is executed alone, > but once the cpuset_sched_domains case is executed first, > After cpuhotplug03 is executed, the error "cpuhotplug03 1 TFAIL: No cpuhotplug_do_spin_loop processes found on" occurs. > The cpuset_sched_domains test case changes the value of cpuset.sched_load_balance in the cpuset cgroup subsystem, > but does not restore the value at the end of the test. > Modify the cpuset_funcs.sh file. The test result shows that the problem is solved successfully. Reviewed-by: Petr Vorel <pvorel@suse.cz> Kind regards, Petr
Hi Petr, On Thu, Apr 29, 2021 at 4:08 PM Wang Xin <wangxin410@huawei.com> wrote: > > When we run the cpuhotplug03 test case on the arm64 machine, > we conclude that no error occurs when the cpuhotplug03 case is executed alone, > but once the cpuset_sched_domains case is executed first, > After cpuhotplug03 is executed, the error "cpuhotplug03 1 TFAIL: No cpuhotplug_do_spin_loop processes found on" occurs. > The cpuset_sched_domains test case changes the value of cpuset.sched_load_balance in the cpuset cgroup subsystem, > but does not restore the value at the end of the test. > Modify the cpuset_funcs.sh file. The test result shows that the problem is solved successfully. > > Signed-off-by: Wang Xin <wangxin410@huawei.com> Reviewed-by: Li Wang <liwang@redhat.com> This just does saves/restore action in setup/cleanup phase, looks no harmful to merge in this release. -- Regards, Li Wang
Hi Li, > Reviewed-by: Li Wang <liwang@redhat.com> > This just does saves/restore action in setup/cleanup phase, looks no > harmful to merge in this release. You right, merged! Thank you both. Kind regards, Petr
diff --git a/testcases/kernel/controllers/cpuset/cpuset_funcs.sh b/testcases/kernel/controllers/cpuset/cpuset_funcs.sh index 00aad0e..f3ba1d5 100755 --- a/testcases/kernel/controllers/cpuset/cpuset_funcs.sh +++ b/testcases/kernel/controllers/cpuset/cpuset_funcs.sh @@ -60,6 +60,8 @@ CPUSET_TMP="/tmp/cpuset_tmp" CLONE_CHILDREN="/dev/cpuset/cgroup.clone_children" CHILDREN_VALUE="0" HOTPLUG_CPU="1" +SCHED_LB="/dev/cpuset/cpuset.sched_load_balance" +SCHED_LB_VALUE="0" cpuset_log() { @@ -169,6 +171,7 @@ setup() fi CHILDREN_VALUE="`cat $CLONE_CHILDREN`" + SCHED_LB_VALUE="`cat $SCHED_LB`" } # Write the cleanup function @@ -180,6 +183,7 @@ cleanup() } echo $CHILDREN_VALUE > $CLONE_CHILDREN + echo $SCHED_LB_VALUE > $SCHED_LB find "$CPUSET" -type d | sort | sed -n '2,$p' | tac | while read subdir do
When we run the cpuhotplug03 test case on the arm64 machine, we conclude that no error occurs when the cpuhotplug03 case is executed alone, but once the cpuset_sched_domains case is executed first, After cpuhotplug03 is executed, the error "cpuhotplug03 1 TFAIL: No cpuhotplug_do_spin_loop processes found on" occurs. The cpuset_sched_domains test case changes the value of cpuset.sched_load_balance in the cpuset cgroup subsystem, but does not restore the value at the end of the test. Modify the cpuset_funcs.sh file. The test result shows that the problem is solved successfully. Signed-off-by: Wang Xin <wangxin410@huawei.com> --- v1 -> v2: modified the format testcases/kernel/controllers/cpuset/cpuset_funcs.sh | 4 ++++ 1 file changed, 4 insertions(+) -- 2.6.2