2011/5/20 KOSAKI Motohiro <kosaki.motohiro@xxxxxxxxxxxxxx>:
> CAI Qian reported his kernel did hang-up if he ran fork intensive
> workload and then invoke oom-killer.
>
> The problem is, current oom calculation uses 0-1000 normalized value
> (The unit is a permillage of system-ram). Its low precision make
> a lot of same oom score. IOW, in his case, all processes have smaller
> oom score than 1 and internal calculation round it to 1.
>
> Thus oom-killer kill ineligible process. This regression is caused by
> commit a63d83f427 (oom: badness heuristic rewrite).
>
> The solution is, the internal calculation just use number of pages
> instead of permillage of system-ram. And convert it to permillage
> value at displaying time.
>
> This patch doesn't change any ABI (included Â/proc/<pid>/oom_score_adj)
> even though current logic has a lot of my dislike thing.
>
> Reported-by: CAI Qian <caiqian@xxxxxxxxxx>
> Signed-off-by: KOSAKI Motohiro <kosaki.motohiro@xxxxxxxxxxxxxx>
> ---
> Âfs/proc/base.c   Â|  13 ++++++----
> Âinclude/linux/oom.h | Â Â7 +----
> Âmm/oom_kill.c    |  60 +++++++++++++++++++++++++++++++++-----------------
> Â3 files changed, 49 insertions(+), 31 deletions(-)
>
> diff --git a/fs/proc/base.c b/fs/proc/base.c
> index dfa5327..d6b0424 100644
> --- a/fs/proc/base.c
> +++ b/fs/proc/base.c
> @@ -476,14 +476,17 @@ static const struct file_operations proc_lstats_operations = {
>
> Âstatic int proc_oom_score(struct task_struct *task, char *buffer)
> Â{
> - Â Â Â unsigned long points = 0;
> + Â Â Â unsigned long points;
> + Â Â Â unsigned long ratio = 0;
> + Â Â Â unsigned long totalpages = totalram_pages + total_swap_pages + 1;
Does we need +1?
oom_badness does have the check.
>
> Â Â Â Âread_lock(&tasklist_lock);
> - Â Â Â if (pid_alive(task))
> - Â Â Â Â Â Â Â points = oom_badness(task, NULL, NULL,
> - Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â totalram_pages + total_swap_pages);
> + Â Â Â if (pid_alive(task)) {
> + Â Â Â Â Â Â Â points = oom_badness(task, NULL, NULL, totalpages);
> + Â Â Â Â Â Â Â ratio = points * 1000 / totalpages;
> + Â Â Â }
> Â Â Â Âread_unlock(&tasklist_lock);
> - Â Â Â return sprintf(buffer, "%lu\n", points);
> + Â Â Â return sprintf(buffer, "%lu\n", ratio);
> Â}
>
> Âstruct limit_names {
> diff --git a/include/linux/oom.h b/include/linux/oom.h
> index 5e3aa83..0f5b588 100644
> --- a/include/linux/oom.h
> +++ b/include/linux/oom.h
> @@ -40,7 +40,8 @@ enum oom_constraint {
> Â Â Â ÂCONSTRAINT_MEMCG,
> Â};
>
> -extern unsigned int oom_badness(struct task_struct *p, struct mem_cgroup *mem,
> +/* The badness from the OOM killer */
> +extern unsigned long oom_badness(struct task_struct *p, struct mem_cgroup *mem,
> Â Â Â Â Â Â Â Â Â Â Â Âconst nodemask_t *nodemask, unsigned long totalpages);
> Âextern int try_set_zonelist_oom(struct zonelist *zonelist, gfp_t gfp_flags);
> Âextern void clear_zonelist_oom(struct zonelist *zonelist, gfp_t gfp_flags);
> @@ -62,10 +63,6 @@ static inline void oom_killer_enable(void)
> Â Â Â Âoom_killer_disabled = false;
> Â}
>
> -/* The badness from the OOM killer */
> -extern unsigned long badness(struct task_struct *p, struct mem_cgroup *mem,
> - Â Â Â Â Â Â Â Â Â Â const nodemask_t *nodemask, unsigned long uptime);
> -
> Âextern struct task_struct *find_lock_task_mm(struct task_struct *p);
>
> Â/* sysctls */
> diff --git a/mm/oom_kill.c b/mm/oom_kill.c
> index e6a6c6f..8bbc3df 100644
> --- a/mm/oom_kill.c
> +++ b/mm/oom_kill.c
> @@ -132,10 +132,12 @@ static bool oom_unkillable_task(struct task_struct *p,
> Â* predictable as possible. ÂThe goal is to return the highest value for the
> Â* task consuming the most memory to avoid subsequent oom failures.
> Â*/
> -unsigned int oom_badness(struct task_struct *p, struct mem_cgroup *mem,
> +unsigned long oom_badness(struct task_struct *p, struct mem_cgroup *mem,
> Â Â Â Â Â Â Â Â Â Â Âconst nodemask_t *nodemask, unsigned long totalpages)
> Â{
> - Â Â Â int points;
> + Â Â Â unsigned long points;
> + Â Â Â unsigned long score_adj = 0;
> +
>
> Â Â Â Âif (oom_unkillable_task(p, mem, nodemask))
> Â Â Â Â Â Â Â Âreturn 0;
> @@ -160,7 +162,7 @@ unsigned int oom_badness(struct task_struct *p, struct mem_cgroup *mem,
> Â Â Â Â */
> Â Â Â Âif (p->flags & PF_OOM_ORIGIN) {
> Â Â Â Â Â Â Â Âtask_unlock(p);
> - Â Â Â Â Â Â Â return 1000;
> + Â Â Â Â Â Â Â return ULONG_MAX;
> Â Â Â Â}
>
> Â Â Â Â/*
> @@ -176,33 +178,49 @@ unsigned int oom_badness(struct task_struct *p, struct mem_cgroup *mem,
> Â Â Â Â */
> Â Â Â Âpoints = get_mm_rss(p->mm) + p->mm->nr_ptes;
> Â Â Â Âpoints += get_mm_counter(p->mm, MM_SWAPENTS);
> -
> - Â Â Â points *= 1000;
> - Â Â Â points /= totalpages;
> Â Â Â Âtask_unlock(p);
>
> Â Â Â Â/*
> Â Â Â Â * Root processes get 3% bonus, just like the __vm_enough_memory()
> Â Â Â Â * implementation used by LSMs.
> + Â Â Â Â*
> + Â Â Â Â* XXX: Too large bonus, example, if the system have tera-bytes memory..
> Â Â Â Â */
> - Â Â Â if (has_capability_noaudit(p, CAP_SYS_ADMIN))
> - Â Â Â Â Â Â Â points -= 30;
> + Â Â Â if (has_capability_noaudit(p, CAP_SYS_ADMIN)) {
> + Â Â Â Â Â Â Â if (points >= totalpages / 32)
> + Â Â Â Â Â Â Â Â Â Â Â points -= totalpages / 32;
> + Â Â Â Â Â Â Â else
> + Â Â Â Â Â Â Â Â Â Â Â points = 0;
Odd. Why do we initialize points with 0?
I think the idea is good.
--
Kind regards,
Minchan Kim
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@xxxxxxxxxx For more info on Linux MM,
see: http://www.linux-mm.org/ .
Fight unfair telecom internet charges in Canada: sign http://stopthemeter.ca/
Don't email: <a href