Re: [PATCH v3] mm/oom_kill: show oom eligibility when displaying the current memory state of all tasks
From: David Rientjes <rientjes@google.com>
Date: 2021-08-02 03:49:10
Also in:
lkml
On Fri, 30 Jul 2021, Aaron Tomlin wrote:
quoted hunk ↗ jump to hunk
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 verydiff --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".
quoted hunk ↗ jump to hunk
+ 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