linux-mm.kvack.org archive mirror
 help / color / mirror / Atom feed
From: Michal Hocko <mhocko@suse.cz>
To: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Cc: linux-mm@kvack.org, akpm@linux-foundation.org, oleg@redhat.com,
	rientjes@google.com, vdavydov@parallels.com, mst@redhat.com
Subject: Re: [PATCH 1/6] mm,oom_reaper: Reduce find_lock_task_mm() usage.
Date: Mon, 11 Jul 2016 14:02:33 +0200	[thread overview]
Message-ID: <20160711120232.GD1811@dhcp22.suse.cz> (raw)
In-Reply-To: <201607080100.BFB78123.OJFtLVHFFMOSQO@I-love.SAKURA.ne.jp>

On Fri 08-07-16 01:00:13, Tetsuo Handa wrote:
> >From 70de3fe92435095b6ecbb400c61e84a99f639d56 Mon Sep 17 00:00:00 2001
> From: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
> Date: Fri, 8 Jul 2016 00:28:12 +0900
> Subject: [PATCH 1/6] mm,oom_reaper: Reduce find_lock_task_mm() usage.
> 
> Since holding mm_struct with elevated mm_count for a second is harmless,
> we can determine mm_struct and hold it upon entry of oom_reap_task().
> This patch has no functional change. Future patch in this series will
> eliminate find_lock_task_mm() usage from the OOM reaper.

the changelog is quite poor to be honest. It doesn't explain why this
is really needed. What do you think about the following:
"
__oom_reap_task can be simplified a bit if it received a valid mm from
oom_reap_task which might need it as well. We could drop one
find_lock_task_mm call and also make the __oom_reap_task code flow
easier to follow. Moreover this will make later patch in the series
easier to review. Pinning mm's mm_count for longer time is not really
harmfull because this will not pin much memory. 

This patch doesn't introduce any functional change.
"

> 
> Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>

Other than that the patch looks good to me.

Acked-by: Michal Hocko <mhocko@suse.com>

> ---
>  mm/oom_kill.c | 79 ++++++++++++++++++++++++++++-------------------------------
>  1 file changed, 37 insertions(+), 42 deletions(-)
> 
> diff --git a/mm/oom_kill.c b/mm/oom_kill.c
> index 7d0a275..951eb1b 100644
> --- a/mm/oom_kill.c
> +++ b/mm/oom_kill.c
> @@ -452,12 +452,10 @@ static DECLARE_WAIT_QUEUE_HEAD(oom_reaper_wait);
>  static struct task_struct *oom_reaper_list;
>  static DEFINE_SPINLOCK(oom_reaper_lock);
>  
> -static bool __oom_reap_task(struct task_struct *tsk)
> +static bool __oom_reap_task(struct task_struct *tsk, struct mm_struct *mm)
>  {
>  	struct mmu_gather tlb;
>  	struct vm_area_struct *vma;
> -	struct mm_struct *mm = NULL;
> -	struct task_struct *p;
>  	struct zap_details details = {.check_swap_entries = true,
>  				      .ignore_dirty = true};
>  	bool ret = true;
> @@ -478,22 +476,9 @@ static bool __oom_reap_task(struct task_struct *tsk)
>  	 */
>  	mutex_lock(&oom_lock);
>  
> -	/*
> -	 * Make sure we find the associated mm_struct even when the particular
> -	 * thread has already terminated and cleared its mm.
> -	 * We might have race with exit path so consider our work done if there
> -	 * is no mm.
> -	 */
> -	p = find_lock_task_mm(tsk);
> -	if (!p)
> -		goto unlock_oom;
> -	mm = p->mm;
> -	atomic_inc(&mm->mm_count);
> -	task_unlock(p);
> -
>  	if (!down_read_trylock(&mm->mmap_sem)) {
>  		ret = false;
> -		goto mm_drop;
> +		goto unlock_oom;
>  	}
>  
>  	/*
> @@ -503,7 +488,7 @@ static bool __oom_reap_task(struct task_struct *tsk)
>  	 */
>  	if (!mmget_not_zero(mm)) {
>  		up_read(&mm->mmap_sem);
> -		goto mm_drop;
> +		goto unlock_oom;
>  	}
>  
>  	tlb_gather_mmu(&tlb, mm, 0, -1);
> @@ -551,8 +536,6 @@ static bool __oom_reap_task(struct task_struct *tsk)
>  	 * put the oom_reaper out of the way.
>  	 */
>  	mmput_async(mm);
> -mm_drop:
> -	mmdrop(mm);
>  unlock_oom:
>  	mutex_unlock(&oom_lock);
>  	return ret;
> @@ -562,36 +545,45 @@ unlock_oom:
>  static void oom_reap_task(struct task_struct *tsk)
>  {
>  	int attempts = 0;
> +	struct mm_struct *mm = NULL;
> +	struct task_struct *p = find_lock_task_mm(tsk);
> +
> +	/*
> +	 * Make sure we find the associated mm_struct even when the particular
> +	 * thread has already terminated and cleared its mm.
> +	 * We might have race with exit path so consider our work done if there
> +	 * is no mm.
> +	 */
> +	if (!p)
> +		goto done;
> +	mm = p->mm;
> +	atomic_inc(&mm->mm_count);
> +	task_unlock(p);
>  
>  	/* Retry the down_read_trylock(mmap_sem) a few times */
> -	while (attempts++ < MAX_OOM_REAP_RETRIES && !__oom_reap_task(tsk))
> +	while (attempts++ < MAX_OOM_REAP_RETRIES && !__oom_reap_task(tsk, mm))
>  		schedule_timeout_idle(HZ/10);
>  
> -	if (attempts > MAX_OOM_REAP_RETRIES) {
> -		struct task_struct *p;
> +	if (attempts <= MAX_OOM_REAP_RETRIES)
> +		goto done;
>  
> -		pr_info("oom_reaper: unable to reap pid:%d (%s)\n",
> -				task_pid_nr(tsk), tsk->comm);
> +	pr_info("oom_reaper: unable to reap pid:%d (%s)\n",
> +		task_pid_nr(tsk), tsk->comm);
>  
> -		/*
> -		 * If we've already tried to reap this task in the past and
> -		 * failed it probably doesn't make much sense to try yet again
> -		 * so hide the mm from the oom killer so that it can move on
> -		 * to another task with a different mm struct.
> -		 */
> -		p = find_lock_task_mm(tsk);
> -		if (p) {
> -			if (test_and_set_bit(MMF_OOM_NOT_REAPABLE, &p->mm->flags)) {
> -				pr_info("oom_reaper: giving up pid:%d (%s)\n",
> -						task_pid_nr(tsk), tsk->comm);
> -				set_bit(MMF_OOM_REAPED, &p->mm->flags);
> -			}
> -			task_unlock(p);
> -		}
> -
> -		debug_show_all_locks();
> +	/*
> +	 * If we've already tried to reap this task in the past and
> +	 * failed it probably doesn't make much sense to try yet again
> +	 * so hide the mm from the oom killer so that it can move on
> +	 * to another task with a different mm struct.
> +	 */
> +	if (test_and_set_bit(MMF_OOM_NOT_REAPABLE, &mm->flags)) {
> +		pr_info("oom_reaper: giving up pid:%d (%s)\n",
> +			task_pid_nr(tsk), tsk->comm);
> +		set_bit(MMF_OOM_REAPED, &mm->flags);
>  	}
> +	debug_show_all_locks();
>  
> +done:
>  	/*
>  	 * Clear TIF_MEMDIE because the task shouldn't be sitting on a
>  	 * reasonably reclaimable memory anymore or it is not a good candidate
> @@ -603,6 +595,9 @@ static void oom_reap_task(struct task_struct *tsk)
>  
>  	/* Drop a reference taken by wake_oom_reaper */
>  	put_task_struct(tsk);
> +	/* Drop a reference taken above. */
> +	if (mm)
> +		mmdrop(mm);
>  }
>  
>  static int oom_reaper(void *unused)
> -- 
> 1.8.3.1

-- 
Michal Hocko
SUSE Labs

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

  reply	other threads:[~2016-07-11 12:02 UTC|newest]

Thread overview: 19+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2016-07-07 15:58 [PATCH v2 0/6] Change OOM killer to use list of mm_struct Tetsuo Handa
2016-07-07 16:00 ` [PATCH 1/6] mm,oom_reaper: Reduce find_lock_task_mm() usage Tetsuo Handa
2016-07-11 12:02   ` Michal Hocko [this message]
2016-07-07 16:01 ` [PATCH 2/6] mm,oom_reaper: Do not attempt to reap a task twice Tetsuo Handa
2016-07-11 12:15   ` Michal Hocko
2016-07-07 16:03 ` [PATCH 3/6] mm,oom: Use list of mm_struct used by OOM victims Tetsuo Handa
2016-07-11 12:50   ` Michal Hocko
2016-07-12  6:00     ` Tetsuo Handa
2016-07-12  7:09       ` Michal Hocko
2016-07-07 16:04 ` [PATCH 4/6] mm,oom_reaper: Make OOM reaper use list of mm_struct Tetsuo Handa
2016-07-11 13:16   ` Michal Hocko
2016-07-12 13:38     ` Tetsuo Handa
2016-07-12 13:46       ` Michal Hocko
2016-07-12 13:55         ` Michal Hocko
2016-07-12 14:01           ` Tetsuo Handa
2016-07-07 16:06 ` [PATCH 5/6] mm,oom: Remove OOM_SCAN_ABORT case and signal_struct->oom_victims Tetsuo Handa
2016-07-11 13:19   ` Michal Hocko
2016-07-07 16:07 ` [PATCH 6/6] mm,oom: Stop clearing TIF_MEMDIE on remote thread Tetsuo Handa
2016-07-11 13:22   ` Michal Hocko

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20160711120232.GD1811@dhcp22.suse.cz \
    --to=mhocko@suse.cz \
    --cc=akpm@linux-foundation.org \
    --cc=linux-mm@kvack.org \
    --cc=mst@redhat.com \
    --cc=oleg@redhat.com \
    --cc=penguin-kernel@I-love.SAKURA.ne.jp \
    --cc=rientjes@google.com \
    --cc=vdavydov@parallels.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).