All of lore.kernel.org
 help / color / mirror / Atom feed
From: David Rientjes <rientjes@google.com>
To: Andrew Morton <akpm@linux-foundation.org>
Cc: Michal Hocko <mhocko@suse.com>,
	Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>,
	linux-mm@kvack.org, linux-kernel@vger.kernel.org
Subject: [patch] mm, oom: prevent additional oom kills before memory is freed
Date: Wed, 14 Jun 2017 16:43:03 -0700 (PDT)	[thread overview]
Message-ID: <alpine.DEB.2.10.1706141632100.93071@chino.kir.corp.google.com> (raw)

If mm->mm_users is not incremented because it is already zero by the oom
reaper, meaning the final refcount has been dropped, do not set
MMF_OOM_SKIP prematurely.

__mmput() may not have had a chance to do exit_mmap() yet, so memory from
a previous oom victim is still mapped.  __mput() naturally requires no
references on mm->mm_users to do exit_mmap().

Without this, several processes can be oom killed unnecessarily and the
oom log can show an abundance of memory available if exit_mmap() is in
progress at the time the process is skipped.

Signed-off-by: David Rientjes <rientjes@google.com>
---
 mm/oom_kill.c | 13 ++++++-------
 1 file changed, 6 insertions(+), 7 deletions(-)

diff --git a/mm/oom_kill.c b/mm/oom_kill.c
--- a/mm/oom_kill.c
+++ b/mm/oom_kill.c
@@ -531,6 +531,7 @@ static bool __oom_reap_task_mm(struct task_struct *tsk, struct mm_struct *mm)
 					 NULL);
 	}
 	tlb_finish_mmu(&tlb, 0, -1);
+	set_bit(MMF_OOM_SKIP, &mm->flags);
 	pr_info("oom_reaper: reaped process %d (%s), now anon-rss:%lukB, file-rss:%lukB, shmem-rss:%lukB\n",
 			task_pid_nr(tsk), tsk->comm,
 			K(get_mm_counter(mm, MM_ANONPAGES)),
@@ -562,7 +563,11 @@ static void oom_reap_task(struct task_struct *tsk)
 	if (attempts <= MAX_OOM_REAP_RETRIES)
 		goto done;
 
-
+	/*
+	 * Hide this mm from OOM killer because it cannot be reaped since
+	 * mm->mmap_sem cannot be acquired.
+	 */
+	set_bit(MMF_OOM_SKIP, &mm->flags);
 	pr_info("oom_reaper: unable to reap pid:%d (%s)\n",
 		task_pid_nr(tsk), tsk->comm);
 	debug_show_all_locks();
@@ -570,12 +575,6 @@ static void oom_reap_task(struct task_struct *tsk)
 done:
 	tsk->oom_reaper_list = NULL;
 
-	/*
-	 * Hide this mm from OOM killer because it has been either reaped or
-	 * somebody can't call up_write(mmap_sem).
-	 */
-	set_bit(MMF_OOM_SKIP, &mm->flags);
-
 	/* Drop a reference taken by wake_oom_reaper */
 	put_task_struct(tsk);
 }

WARNING: multiple messages have this Message-ID (diff)
From: David Rientjes <rientjes@google.com>
To: Andrew Morton <akpm@linux-foundation.org>
Cc: Michal Hocko <mhocko@suse.com>,
	Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>,
	linux-mm@kvack.org, linux-kernel@vger.kernel.org
Subject: [patch] mm, oom: prevent additional oom kills before memory is freed
Date: Wed, 14 Jun 2017 16:43:03 -0700 (PDT)	[thread overview]
Message-ID: <alpine.DEB.2.10.1706141632100.93071@chino.kir.corp.google.com> (raw)

If mm->mm_users is not incremented because it is already zero by the oom
reaper, meaning the final refcount has been dropped, do not set
MMF_OOM_SKIP prematurely.

__mmput() may not have had a chance to do exit_mmap() yet, so memory from
a previous oom victim is still mapped.  __mput() naturally requires no
references on mm->mm_users to do exit_mmap().

Without this, several processes can be oom killed unnecessarily and the
oom log can show an abundance of memory available if exit_mmap() is in
progress at the time the process is skipped.

Signed-off-by: David Rientjes <rientjes@google.com>
---
 mm/oom_kill.c | 13 ++++++-------
 1 file changed, 6 insertions(+), 7 deletions(-)

diff --git a/mm/oom_kill.c b/mm/oom_kill.c
--- a/mm/oom_kill.c
+++ b/mm/oom_kill.c
@@ -531,6 +531,7 @@ static bool __oom_reap_task_mm(struct task_struct *tsk, struct mm_struct *mm)
 					 NULL);
 	}
 	tlb_finish_mmu(&tlb, 0, -1);
+	set_bit(MMF_OOM_SKIP, &mm->flags);
 	pr_info("oom_reaper: reaped process %d (%s), now anon-rss:%lukB, file-rss:%lukB, shmem-rss:%lukB\n",
 			task_pid_nr(tsk), tsk->comm,
 			K(get_mm_counter(mm, MM_ANONPAGES)),
@@ -562,7 +563,11 @@ static void oom_reap_task(struct task_struct *tsk)
 	if (attempts <= MAX_OOM_REAP_RETRIES)
 		goto done;
 
-
+	/*
+	 * Hide this mm from OOM killer because it cannot be reaped since
+	 * mm->mmap_sem cannot be acquired.
+	 */
+	set_bit(MMF_OOM_SKIP, &mm->flags);
 	pr_info("oom_reaper: unable to reap pid:%d (%s)\n",
 		task_pid_nr(tsk), tsk->comm);
 	debug_show_all_locks();
@@ -570,12 +575,6 @@ static void oom_reap_task(struct task_struct *tsk)
 done:
 	tsk->oom_reaper_list = NULL;
 
-	/*
-	 * Hide this mm from OOM killer because it has been either reaped or
-	 * somebody can't call up_write(mmap_sem).
-	 */
-	set_bit(MMF_OOM_SKIP, &mm->flags);
-
 	/* Drop a reference taken by wake_oom_reaper */
 	put_task_struct(tsk);
 }

--
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:[~2017-06-14 23:43 UTC|newest]

Thread overview: 70+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2017-06-14 23:43 David Rientjes [this message]
2017-06-14 23:43 ` [patch] mm, oom: prevent additional oom kills before memory is freed David Rientjes
2017-06-15 10:39 ` Michal Hocko
2017-06-15 10:39   ` Michal Hocko
2017-06-15 10:53   ` Tetsuo Handa
2017-06-15 10:53     ` Tetsuo Handa
2017-06-15 11:01     ` Michal Hocko
2017-06-15 11:01       ` Michal Hocko
2017-06-15 11:32       ` Tetsuo Handa
2017-06-15 11:32         ` Tetsuo Handa
2017-06-15 12:03         ` Michal Hocko
2017-06-15 12:03           ` Michal Hocko
2017-06-15 12:13           ` Michal Hocko
2017-06-15 12:13             ` Michal Hocko
2017-06-15 13:01             ` Tetsuo Handa
2017-06-15 13:01               ` Tetsuo Handa
2017-06-15 13:22               ` Michal Hocko
2017-06-15 13:22                 ` Michal Hocko
2017-06-15 21:43                 ` Tetsuo Handa
2017-06-15 21:43                   ` Tetsuo Handa
2017-06-15 21:37               ` David Rientjes
2017-06-15 21:37                 ` David Rientjes
2017-06-15 12:20       ` Michal Hocko
2017-06-15 12:20         ` Michal Hocko
2017-06-15 21:26   ` David Rientjes
2017-06-15 21:26     ` David Rientjes
2017-06-15 21:41     ` Michal Hocko
2017-06-15 21:41       ` Michal Hocko
2017-06-15 22:03       ` David Rientjes
2017-06-15 22:03         ` David Rientjes
2017-06-15 22:12         ` Michal Hocko
2017-06-15 22:12           ` Michal Hocko
2017-06-15 22:42           ` David Rientjes
2017-06-15 22:42             ` David Rientjes
2017-06-16  8:06             ` Michal Hocko
2017-06-16  8:06               ` Michal Hocko
2017-06-16  0:54           ` Tetsuo Handa
2017-06-16  0:54             ` Tetsuo Handa
2017-06-16  4:00             ` Tetsuo Handa
2017-06-16  4:00               ` Tetsuo Handa
2017-06-16  8:39             ` Michal Hocko
2017-06-16  8:39               ` Michal Hocko
2017-06-16 10:27               ` Tetsuo Handa
2017-06-16 10:27                 ` Tetsuo Handa
2017-06-16 11:02                 ` Michal Hocko
2017-06-16 11:02                   ` Michal Hocko
2017-06-16 14:26                   ` Re: [patch] mm, oom: prevent additional oom kills before memoryis freed Tetsuo Handa
2017-06-16 14:26                     ` Tetsuo Handa
2017-06-16 14:42                     ` Michal Hocko
2017-06-16 14:42                       ` Michal Hocko
2017-06-17 13:30                       ` Re: [patch] mm, oom: prevent additional oom kills before memory is freed Tetsuo Handa
2017-06-17 13:30                         ` Tetsuo Handa
2017-06-23 12:38                         ` Michal Hocko
2017-06-23 12:38                           ` Michal Hocko
2017-06-16 12:22       ` Tetsuo Handa
2017-06-16 12:22         ` Tetsuo Handa
2017-06-16 14:12         ` Michal Hocko
2017-06-16 14:12           ` Michal Hocko
2017-06-17  5:17           ` [PATCH] mm,oom_kill: Close race window of needlessly selecting new victims Tetsuo Handa
2017-06-17  5:17             ` Tetsuo Handa
2017-06-20 22:12             ` David Rientjes
2017-06-20 22:12               ` David Rientjes
2017-06-21  2:17               ` Tetsuo Handa
2017-06-21 20:31                 ` David Rientjes
2017-06-21 20:31                   ` David Rientjes
2017-06-22  0:53                   ` Tetsuo Handa
2017-06-23 12:45                     ` Michal Hocko
2017-06-23 12:45                       ` Michal Hocko
2017-06-21 13:18               ` Michal Hocko
2017-06-21 13:18                 ` 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=alpine.DEB.2.10.1706141632100.93071@chino.kir.corp.google.com \
    --to=rientjes@google.com \
    --cc=akpm@linux-foundation.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-mm@kvack.org \
    --cc=mhocko@suse.com \
    --cc=penguin-kernel@I-love.SAKURA.ne.jp \
    /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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.