On Fri, 30 Jul 2021, Aaron Tomlin wrote:
> Documentation/admin-guide/sysctl/vm.rst | 5 ++--
> mm/oom_kill.c | 31 +++++++++++++++++++++----
> 2 files changed, 30 insertions(+), 6 deletions(-)
>
> diff --git a/Documentation/admin-guide/sysctl/vm.rst b/Documentation/admin-guide/sysctl/vm.rst
> index 003d5cc3751b..4c79fa00ddb3 100644
> --- a/Documentation/admin-guide/sysctl/vm.rst
> +++ b/Documentation/admin-guide/sysctl/vm.rst
> @@ -650,8 +650,9 @@ oom_dump_tasks
> Enables a system-wide task dump (excluding kernel threads) to be produced
> when the kernel performs an OOM-killing and includes such information as
> pid, uid, tgid, vm size, rss, pgtables_bytes, swapents, oom_score_adj
> -score, and name. This is helpful to determine why the OOM killer was
> -invoked, to identify the rogue task that caused it, and to determine why
> +score, oom eligibility status and name. This is helpful to determine why
> +the OOM killer was invoked, to identify the rogue task that caused it, and
> +to determine why
> the OOM killer chose the task it did to kill.
>
> If this is set to zero, this information is suppressed. On very
> diff --git a/mm/oom_kill.c b/mm/oom_kill.c
> index c729a4c4a1ac..36daa6917b62 100644
> --- a/mm/oom_kill.c
> +++ b/mm/oom_kill.c
> @@ -160,6 +160,27 @@ static inline bool is_sysrq_oom(struct oom_control *oc)
> return oc->order == -1;
> }
>
> +/**
> + * is_task_eligible_oom - determine if and why a task cannot be OOM killed
> + * @tsk: task to check
> + *
> + * Needs to be called with task_lock().
> + */
> +static const char * const is_task_oom_eligible(struct task_struct *p)
> +{
> + long adj;
> +
> + adj = (long)p->signal->oom_score_adj;
> + if (adj == OOM_SCORE_ADJ_MIN)
> + return "M";
oom_score_adj is shown already in the tasklist dump, I'm not sure what
value this adds.
> + else if (test_bit(MMF_OOM_SKIP, &p->mm->flags)
> + return "R";
> + else if (in_vfork(p))
> + return "V";
This is going to be racy, we can't show that a task that is emitted as
part of the tasklist dump was did not have in_vfork() == true at the time
oom_badness() was called.
Wouldn't it be better to simply print the output of oom_badness() to the
tasklist dump instead so we get complete information?
We could simply special case a LONG_MIN return value as -1000 or "min".
> + else
> + return "";
> +}
> +
> /* return true if the task is not adequate as candidate victim task. */
> static bool oom_unkillable_task(struct task_struct *p)
> {
> @@ -401,12 +422,13 @@ static int dump_task(struct task_struct *p, void *arg)
> return 0;
> }
>
> - pr_info("[%7d] %5d %5d %8lu %8lu %8ld %8lu %5hd %s\n",
> + pr_info("[%7d] %5d %5d %8lu %8lu %8ld %8lu %5hd %1s %s\n",
> task->pid, from_kuid(&init_user_ns, task_uid(task)),
> task->tgid, task->mm->total_vm, get_mm_rss(task->mm),
> mm_pgtables_bytes(task->mm),
> get_mm_counter(task->mm, MM_SWAPENTS),
> - task->signal->oom_score_adj, task->comm);
> + task->signal->oom_score_adj, is_task_oom_eligible(task),
> + task->comm);
> task_unlock(task);
>
> return 0;
> @@ -420,12 +442,13 @@ static int dump_task(struct task_struct *p, void *arg)
> * memcg, not in the same cpuset, or bound to a disjoint set of mempolicy nodes
> * are not shown.
> * State information includes task's pid, uid, tgid, vm size, rss,
> - * pgtables_bytes, swapents, oom_score_adj value, and name.
> + * pgtables_bytes, swapents, oom_score_adj value, oom eligibility status
> + * and name.
> */
> static void dump_tasks(struct oom_control *oc)
> {
> pr_info("Tasks state (memory values in pages):\n");
> - pr_info("[ pid ] uid tgid total_vm rss pgtables_bytes swapents oom_score_adj name\n");
> + pr_info("[ pid ] uid tgid total_vm rss pgtables_bytes swapents oom_score_adj oom_skipped name\n");
>
> if (is_memcg_oom(oc))
> mem_cgroup_scan_tasks(oc->memcg, dump_task, oc);
> --
> 2.31.1
>
>
