On Fri 23-10-15 17:02:30, Aristeu Rozanski wrote:
> One of the largest chunks of log messages in a OOM is from dump_stack() and in
> some cases it isn't even necessary to figure out what's going on. In
> systems with multiple tenants/containers with limited resources each
> OOMs can be way more frequent and being able to reduce the amount of log
> output for each situation is useful.
I can see why you want to reduce the amount of information, I guess you
have tried to reduce the loglevel but this hasn't helped because
dump_stack uses default log level which is too low to be usable, right?
Or are there any other reasons?
> This patch adds a sysctl to allow disabling dump_stack() during an OOM while
> keeping the default to behave the same way it behaves today.
I am not sure sysctl is a good way to tell this particular restriction
on the output. What if somebody else doesn't want to see the list of
eligible tasks? Should we add another knob?
Would it make more sense to distinguish different parts of the OOM
report by loglevel properly?
pr_err - killed task report
pr_warning - oom invocation + memory info
pr_notice - task list
pr_info - stack trace
> Cc: Greg Thelen <gthelen@xxxxxxxxxx>
> Cc: Johannes Weiner <hannes@xxxxxxxxxxx>
> Cc: linux-mm@xxxxxxxxx
> Cc: cgroups@xxxxxxxxxxxxxxx
> Signed-off-by: Aristeu Rozanski <arozansk@xxxxxxxxxx>
> ---
> include/linux/oom.h | 1 +
> kernel/sysctl.c | 7 +++++++
> mm/oom_kill.c | 4 +++-
> 3 files changed, 11 insertions(+), 1 deletion(-)
>
> diff --git a/include/linux/oom.h b/include/linux/oom.h
> index 03e6257..bdd03e5 100644
> --- a/include/linux/oom.h
> +++ b/include/linux/oom.h
> @@ -115,6 +115,7 @@ static inline bool task_will_free_mem(struct task_struct *task)
>
> /* sysctls */
> extern int sysctl_oom_dump_tasks;
> +extern int sysctl_oom_dump_stack;
> extern int sysctl_oom_kill_allocating_task;
> extern int sysctl_panic_on_oom;
> #endif /* _INCLUDE_LINUX_OOM_H */
> diff --git a/kernel/sysctl.c b/kernel/sysctl.c
> index e69201d..c812523 100644
> --- a/kernel/sysctl.c
> +++ b/kernel/sysctl.c
> @@ -1176,6 +1176,13 @@ static struct ctl_table vm_table[] = {
> .proc_handler = proc_dointvec,
> },
> {
> + .procname = "oom_dump_stack",
> + .data = &sysctl_oom_dump_stack,
> + .maxlen = sizeof(sysctl_oom_dump_stack),
> + .mode = 0644,
> + .proc_handler = proc_dointvec,
> + },
> + {
> .procname = "overcommit_ratio",
> .data = &sysctl_overcommit_ratio,
> .maxlen = sizeof(sysctl_overcommit_ratio),
> diff --git a/mm/oom_kill.c b/mm/oom_kill.c
> index 1ecc0bc..bdbf83b 100644
> --- a/mm/oom_kill.c
> +++ b/mm/oom_kill.c
> @@ -42,6 +42,7 @@
> int sysctl_panic_on_oom;
> int sysctl_oom_kill_allocating_task;
> int sysctl_oom_dump_tasks = 1;
> +int sysctl_oom_dump_stack = 1;
>
> DEFINE_MUTEX(oom_lock);
>
> @@ -384,7 +385,8 @@ static void dump_header(struct oom_control *oc, struct task_struct *p,
> current->signal->oom_score_adj);
> cpuset_print_task_mems_allowed(current);
> task_unlock(current);
> - dump_stack();
> + if (sysctl_oom_dump_stack)
> + dump_stack();
> if (memcg)
> mem_cgroup_print_oom_info(memcg, p);
> else
> --
> 1.8.3.1
>
> --
> To unsubscribe, send a message with 'unsubscribe linux-mm' in
> the body to majordomo@xxxxxxxxx. For more info on Linux MM,
> see: http://www.linux-mm.org/ .
> Don't email: <a href=mailto:"dont@xxxxxxxxx";> email@xxxxxxxxx </a>
--
Michal Hocko
SUSE Labs
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@xxxxxxxxx. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@xxxxxxxxx";> email@xxxxxxxxx </a>