From: Pavel Emelianov <xemul@openvz.org>
To: Andrew Morton <akpm@osdl.org>,
Balbir Singh <balbir@linux.vnet.ibm.com>,
Vaidyanathan Srinivasan <svaidy@linux.vnet.ibm.com>
Cc: Linux Containers <containers@lists.osdl.org>,
Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
devel@openvz.org
Subject: [PATCH 6/8] Per container OOM killer
Date: Mon, 04 Jun 2007 17:40:17 +0400 [thread overview]
Message-ID: <46641641.8070702@openvz.org> (raw)
In-Reply-To: <466412C5.1060104@openvz.org>
When container is completely out of memory some tasks should die.
This is unfair to kill the current task, so a task with the largest
RSS is chosen and killed. The code re-uses current OOM killer
select_bad_process() for task selection.
Signed-off-by: Pavel Emelianov <xemul@openvz.org>
---
--- ./include/linux/rss_container.h.rssoom 2007-06-04 12:06:27.000000000 +0400
+++ ./include/linux/rss_container.h 2007-06-04 12:21:11.000000000 +0400
@@ -73,6 +73,8 @@ void container_rss_add(struct page_conta
void container_rss_del(struct page_container *);
void container_rss_release(struct page_container *);
+void container_out_of_memory(struct rss_container *);
+
void mm_init_container(struct mm_struct *mm, struct task_struct *tsk);
void mm_free_container(struct mm_struct *mm);
#else
--- ./mm/oom_kill.c.rssoom 2007-06-04 12:05:26.000000000 +0400
+++ ./mm/oom_kill.c 2007-06-04 12:21:11.000000000 +0400
@@ -24,6 +24,7 @@
#include <linux/cpuset.h>
#include <linux/module.h>
#include <linux/notifier.h>
+#include <linux/rss_container.h>
int sysctl_panic_on_oom;
/* #define DEBUG */
@@ -47,7 +48,8 @@ int sysctl_panic_on_oom;
* of least surprise ... (be careful when you change it)
*/
-unsigned long badness(struct task_struct *p, unsigned long uptime)
+unsigned long badness(struct task_struct *p, unsigned long uptime,
+ struct rss_container *rss)
{
unsigned long points, cpu_time, run_time, s;
struct mm_struct *mm;
@@ -60,6 +62,13 @@ unsigned long badness(struct task_struct
return 0;
}
+#ifdef CONFIG_RSS_CONTAINER
+ if (rss != NULL && mm->rss_container != rss) {
+ task_unlock(p);
+ return 0;
+ }
+#endif
+
/*
* The memory size of the process is the basis for the badness.
*/
@@ -204,7 +213,8 @@ static inline int constrained_alloc(stru
*
* (not docbooked, we don't want this one cluttering up the manual)
*/
-static struct task_struct *select_bad_process(unsigned long *ppoints)
+static struct task_struct *select_bad_process(unsigned long *ppoints,
+ struct rss_container *rss)
{
struct task_struct *g, *p;
struct task_struct *chosen = NULL;
@@ -258,7 +268,7 @@ static struct task_struct *select_bad_pr
if (p->oomkilladj == OOM_DISABLE)
continue;
- points = badness(p, uptime.tv_sec);
+ points = badness(p, uptime.tv_sec, rss);
if (points > *ppoints || !chosen) {
chosen = p;
*ppoints = points;
@@ -444,7 +454,7 @@ retry:
* Rambo mode: Shoot down a process and hope it solves whatever
* issues we may have.
*/
- p = select_bad_process(&points);
+ p = select_bad_process(&points, NULL);
if (PTR_ERR(p) == -1UL)
goto out;
@@ -473,3 +483,27 @@ out:
if (!test_thread_flag(TIF_MEMDIE))
schedule_timeout_uninterruptible(1);
}
+
+#ifdef CONFIG_RSS_CONTAINER
+void container_out_of_memory(struct rss_container *rss)
+{
+ unsigned long points = 0;
+ struct task_struct *p;
+
+ container_lock();
+ read_lock(&tasklist_lock);
+retry:
+ p = select_bad_process(&points, rss);
+ if (PTR_ERR(p) == -1UL)
+ goto out;
+
+ if (!p)
+ p = current;
+
+ if (oom_kill_process(p, points, "Container out of memory"))
+ goto retry;
+out:
+ read_unlock(&tasklist_lock);
+ container_unlock();
+}
+#endif
next prev parent reply other threads:[~2007-06-04 13:37 UTC|newest]
Thread overview: 16+ messages / expand[flat|nested] mbox.gz Atom feed top
2007-06-04 13:25 [PATCH 0/8] RSS controller based on process containers (v3.1) Pavel Emelianov
2007-06-04 13:28 ` [PATCH 1/8] Resource counters Pavel Emelianov
2007-06-04 13:29 ` [PATCH 2/8] Add container pointer on struct page Pavel Emelianov
2007-06-04 13:31 ` [PATCH 3/8] Add container pointer on mm_struct Pavel Emelianov
2007-06-04 13:33 ` [PATCH 4/8] Scanner changes needed to implement per-container scanner Pavel Emelianov
2007-06-04 13:38 ` [PATCH 5/8] RSS container core Pavel Emelianov
2007-06-04 13:40 ` Pavel Emelianov [this message]
2007-06-04 13:41 ` [PATCH 7/8] Per-container pages reclamation Pavel Emelianov
2007-06-04 13:46 ` [PATCH 8/8] RSS accounting hooks over the code Pavel Emelianov
2007-06-08 12:03 ` [PATCH 0/8] RSS controller based on process containers (v3.1) Herbert Poetzl
2007-06-08 12:39 ` Pavel Emelianov
2007-06-08 15:37 ` Herbert Poetzl
2007-06-08 17:07 ` Balbir Singh
2007-06-08 17:44 ` Vaidyanathan Srinivasan
-- strict thread matches above, loose matches on Subject: below --
2007-05-30 15:24 [PATCH 0/8] RSS controller based on process containers (v3) Pavel Emelianov
2007-05-30 15:35 ` [PATCH 6/8] Per container OOM killer Pavel Emelianov
2007-04-09 12:22 [PATCH 0/8] RSS controller based on process containers (v2) Pavel Emelianov
2007-04-09 12:56 ` [PATCH 6/8] Per container OOM killer Pavel Emelianov
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=46641641.8070702@openvz.org \
--to=xemul@openvz.org \
--cc=akpm@osdl.org \
--cc=balbir@linux.vnet.ibm.com \
--cc=containers@lists.osdl.org \
--cc=devel@openvz.org \
--cc=linux-kernel@vger.kernel.org \
--cc=svaidy@linux.vnet.ibm.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).