[PATCH v2 -mmotm 1/4] cgroups: Fix race between userspace and kernelspace

[PATCH v2 -mmotm 1/4] cgroups: Fix race between userspace and kernelspace

[Kirill A. Shutemov]

eventfd are used to notify about two types of event:
 - control file-specific, like crossing memory threshold;
 - cgroup removing.

To understand what really happen, userspace can check if the cgroup
still exists. To avoid race beetween userspace and kernelspace we have
to notify userspace about cgroup removing only after rmdir of cgroup
directory.

Signed-off-by: Kirill A. Shutemov <kirill@shutemov.name>
Reviewed-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
---
 kernel/cgroup.c |   32 +++++++++++++++++---------------
 1 files changed, 17 insertions(+), 15 deletions(-)

diff --git a/kernel/cgroup.c b/kernel/cgroup.c
index ce9008f..46903cb 100644
--- a/kernel/cgroup.c
+++ b/kernel/cgroup.c
@@ -780,28 +780,15 @@ static struct inode *cgroup_new_inode(mode_t mode, struct super_block *sb)
 static int cgroup_call_pre_destroy(struct cgroup *cgrp)
 {
 	struct cgroup_subsys *ss;
-	struct cgroup_event *event, *tmp;
 	int ret = 0;
 
 	for_each_subsys(cgrp->root, ss)
 		if (ss->pre_destroy) {
 			ret = ss->pre_destroy(ss, cgrp);
 			if (ret)
-				goto out;
+				break;
 		}
 
-	/*
-	 * Unregister events and notify userspace.
-	 */
-	spin_lock(&cgrp->event_list_lock);
-	list_for_each_entry_safe(event, tmp, &cgrp->event_list, list) {
-		list_del(&event->list);
-		eventfd_signal(event->eventfd, 1);
-		schedule_work(&event->remove);
-	}
-	spin_unlock(&cgrp->event_list_lock);
-
-out:
 	return ret;
 }
 
@@ -2991,7 +2978,6 @@ static void cgroup_event_remove(struct work_struct *work)
 	event->cft->unregister_event(cgrp, event->cft, event->eventfd);
 
 	eventfd_ctx_put(event->eventfd);
-	remove_wait_queue(event->wqh, &event->wait);
 	kfree(event);
 }
 
@@ -3009,6 +2995,7 @@ static int cgroup_event_wake(wait_queue_t *wait, unsigned mode,
 	unsigned long flags = (unsigned long)key;
 
 	if (flags & POLLHUP) {
+		remove_wait_queue_locked(event->wqh, &event->wait);
 		spin_lock(&cgrp->event_list_lock);
 		list_del(&event->list);
 		spin_unlock(&cgrp->event_list_lock);
@@ -3457,6 +3444,7 @@ static int cgroup_rmdir(struct inode *unused_dir, struct dentry *dentry)
 	struct dentry *d;
 	struct cgroup *parent;
 	DEFINE_WAIT(wait);
+	struct cgroup_event *event, *tmp;
 	int ret;
 
 	/* the vfs holds both inode->i_mutex already */
@@ -3540,6 +3528,20 @@ again:
 	set_bit(CGRP_RELEASABLE, &parent->flags);
 	check_for_release(parent);
 
+	/*
+	 * Unregister events and notify userspace.
+	 * Notify userspace about cgroup removing only after rmdir of cgroup
+	 * directory to avoid race between userspace and kernelspace
+	 */
+	spin_lock(&cgrp->event_list_lock);
+	list_for_each_entry_safe(event, tmp, &cgrp->event_list, list) {
+		list_del(&event->list);
+		remove_wait_queue(event->wqh, &event->wait);
+		eventfd_signal(event->eventfd, 1);
+		schedule_work(&event->remove);
+	}
+	spin_unlock(&cgrp->event_list_lock);
+
 	mutex_unlock(&cgroup_mutex);
 	return 0;
 }
-- 
1.6.6.2

--
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>

[PATCH v2 -mmotm 2/4] cgroups: remove events before destroying subsystem state objects

[Kirill A. Shutemov]

Events should be removed after rmdir of cgroup directory, but before
destroying subsystem state objects. Let's take reference to cgroup
directory dentry to do that.

Signed-off-by: Kirill A. Shutemov <kirill-oKw7cIdHH8eLwutG50LtGA@public.gmane.org>
Acked-by: KAMEZAWA Hiroyuki <kamezawa.hioryu-+CUm20s59erQFUHtdCDX3A@public.gmane.org>
---
 include/linux/cgroup.h |    3 ---
 kernel/cgroup.c        |    8 ++++++++
 mm/memcontrol.c        |    9 ---------
 3 files changed, 8 insertions(+), 12 deletions(-)

diff --git a/include/linux/cgroup.h b/include/linux/cgroup.h
index 64cebfe..1719c75 100644
--- a/include/linux/cgroup.h
+++ b/include/linux/cgroup.h
@@ -395,9 +395,6 @@ struct cftype {
 	 * closes the eventfd or on cgroup removing.
 	 * This callback must be implemented, if you want provide
 	 * notification functionality.
-	 *
-	 * Be careful. It can be called after destroy(), so you have
-	 * to keep all nesessary data, until all events are removed.
 	 */
 	int (*unregister_event)(struct cgroup *cgrp, struct cftype *cft,
 			struct eventfd_ctx *eventfd);
diff --git a/kernel/cgroup.c b/kernel/cgroup.c
index 46903cb..d142524 100644
--- a/kernel/cgroup.c
+++ b/kernel/cgroup.c
@@ -2979,6 +2979,7 @@ static void cgroup_event_remove(struct work_struct *work)
 
 	eventfd_ctx_put(event->eventfd);
 	kfree(event);
+	dput(cgrp->dentry);
 }
 
 /*
@@ -3099,6 +3100,13 @@ static int cgroup_write_event_control(struct cgroup *cgrp, struct cftype *cft,
 		goto fail;
 	}
 
+	/*
+	 * Events should be removed after rmdir of cgroup directory, but before
+	 * destroying subsystem state objects. Let's take reference to cgroup
+	 * directory dentry to do that.
+	 */
+	dget(cgrp->dentry);
+
 	spin_lock(&cgrp->event_list_lock);
 	list_add(&event->list, &cgrp->event_list);
 	spin_unlock(&cgrp->event_list_lock);
diff --git a/mm/memcontrol.c b/mm/memcontrol.c
index a443c30..8fe6e7f 100644
--- a/mm/memcontrol.c
+++ b/mm/memcontrol.c
@@ -3358,12 +3358,6 @@ static int mem_cgroup_register_event(struct cgroup *cgrp, struct cftype *cft,
 		}
 	}
 
-	/*
-	 * We need to increment refcnt to be sure that all thresholds
-	 * will be unregistered before calling __mem_cgroup_free()
-	 */
-	mem_cgroup_get(memcg);
-
 	if (type == _MEM)
 		rcu_assign_pointer(memcg->thresholds, thresholds_new);
 	else
@@ -3457,9 +3451,6 @@ assign:
 	/* To be sure that nobody uses thresholds before freeing it */
 	synchronize_rcu();
 
-	for (i = 0; i < thresholds->size - size; i++)
-		mem_cgroup_put(memcg);
-
 	kfree(thresholds);
 unlock:
 	mutex_unlock(&memcg->thresholds_lock);
-- 
1.6.6.2

[PATCH v2 -mmotm 2/4] cgroups: remove events before destroying subsystem state objects

[Kirill A. Shutemov]

Events should be removed after rmdir of cgroup directory, but before
destroying subsystem state objects. Let's take reference to cgroup
directory dentry to do that.

Signed-off-by: Kirill A. Shutemov <kirill@shutemov.name>
Acked-by: KAMEZAWA Hiroyuki <kamezawa.hioryu@jp.fujitsu.com>
---
 include/linux/cgroup.h |    3 ---
 kernel/cgroup.c        |    8 ++++++++
 mm/memcontrol.c        |    9 ---------
 3 files changed, 8 insertions(+), 12 deletions(-)

diff --git a/include/linux/cgroup.h b/include/linux/cgroup.h
index 64cebfe..1719c75 100644
--- a/include/linux/cgroup.h
+++ b/include/linux/cgroup.h
@@ -395,9 +395,6 @@ struct cftype {
 	 * closes the eventfd or on cgroup removing.
 	 * This callback must be implemented, if you want provide
 	 * notification functionality.
-	 *
-	 * Be careful. It can be called after destroy(), so you have
-	 * to keep all nesessary data, until all events are removed.
 	 */
 	int (*unregister_event)(struct cgroup *cgrp, struct cftype *cft,
 			struct eventfd_ctx *eventfd);
diff --git a/kernel/cgroup.c b/kernel/cgroup.c
index 46903cb..d142524 100644
--- a/kernel/cgroup.c
+++ b/kernel/cgroup.c
@@ -2979,6 +2979,7 @@ static void cgroup_event_remove(struct work_struct *work)
 
 	eventfd_ctx_put(event->eventfd);
 	kfree(event);
+	dput(cgrp->dentry);
 }
 
 /*
@@ -3099,6 +3100,13 @@ static int cgroup_write_event_control(struct cgroup *cgrp, struct cftype *cft,
 		goto fail;
 	}
 
+	/*
+	 * Events should be removed after rmdir of cgroup directory, but before
+	 * destroying subsystem state objects. Let's take reference to cgroup
+	 * directory dentry to do that.
+	 */
+	dget(cgrp->dentry);
+
 	spin_lock(&cgrp->event_list_lock);
 	list_add(&event->list, &cgrp->event_list);
 	spin_unlock(&cgrp->event_list_lock);
diff --git a/mm/memcontrol.c b/mm/memcontrol.c
index a443c30..8fe6e7f 100644
--- a/mm/memcontrol.c
+++ b/mm/memcontrol.c
@@ -3358,12 +3358,6 @@ static int mem_cgroup_register_event(struct cgroup *cgrp, struct cftype *cft,
 		}
 	}
 
-	/*
-	 * We need to increment refcnt to be sure that all thresholds
-	 * will be unregistered before calling __mem_cgroup_free()
-	 */
-	mem_cgroup_get(memcg);
-
 	if (type == _MEM)
 		rcu_assign_pointer(memcg->thresholds, thresholds_new);
 	else
@@ -3457,9 +3451,6 @@ assign:
 	/* To be sure that nobody uses thresholds before freeing it */
 	synchronize_rcu();
 
-	for (i = 0; i < thresholds->size - size; i++)
-		mem_cgroup_put(memcg);
-
 	kfree(thresholds);
 unlock:
 	mutex_unlock(&memcg->thresholds_lock);
-- 
1.6.6.2

--
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>

[PATCH v2 -mmotm 3/4] cgroups: Add simple listener of cgroup events to documentation

[Kirill A. Shutemov]

An example of cgroup notification API usage.

Signed-off-by: Kirill A. Shutemov <kirill-oKw7cIdHH8eLwutG50LtGA@public.gmane.org>
Reviewed-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu-+CUm20s59erQFUHtdCDX3A@public.gmane.org>
---
 Documentation/cgroups/cgroup_event_listener.c |  103 +++++++++++++++++++++++++
 1 files changed, 103 insertions(+), 0 deletions(-)
 create mode 100644 Documentation/cgroups/cgroup_event_listener.c

diff --git a/Documentation/cgroups/cgroup_event_listener.c b/Documentation/cgroups/cgroup_event_listener.c
new file mode 100644
index 0000000..8c2d7aa
--- /dev/null
+++ b/Documentation/cgroups/cgroup_event_listener.c
@@ -0,0 +1,103 @@
+/*
+ * cgroup_event_listener.c - Simple listener of cgroup events
+ *
+ * Copyright (C) Kirill A. Shutemov <kirill-oKw7cIdHH8eLwutG50LtGA@public.gmane.org>
+ */
+
+#include <assert.h>
+#include <errno.h>
+#include <fcntl.h>
+#include <libgen.h>
+#include <limits.h>
+#include <stdio.h>
+#include <string.h>
+#include <unistd.h>
+
+#include <sys/eventfd.h>
+
+#define USAGE_STR "Usage: cgroup_event_listener <path-to-control-file> <args>\n"
+
+int main(int argc, char **argv)
+{
+	int efd = -1;
+	int cfd = -1;
+	int event_control = -1;
+	char event_control_path[PATH_MAX];
+	int ret;
+
+	if (argc != 3) {
+		fputs(USAGE_STR, stderr);
+		return 1;
+	}
+
+	cfd = open(argv[1], O_RDONLY);
+	if (cfd == -1) {
+		fprintf(stderr, "Cannot open %s: %s\n", argv[1],
+				strerror(errno));
+		goto out;
+	}
+
+	ret = snprintf(event_control_path, PATH_MAX, "%s/cgroup.event_control",
+			dirname(argv[1]));
+	if (ret > PATH_MAX) {
+		fputs("Path to cgroup.event_control is too long\n", stderr);
+		goto out;
+	}
+
+	event_control = open(event_control_path, O_WRONLY);
+	if (event_control == -1) {
+		fprintf(stderr, "Cannot open %s: %s\n", event_control_path,
+				strerror(errno));
+		goto out;
+	}
+
+	efd = eventfd(0, 0);
+	if (efd == -1) {
+		perror("eventfd() failed");
+		goto out;
+	}
+
+	ret = dprintf(event_control, "%d %d %s", efd, cfd, argv[2]);
+	if (ret == -1) {
+		perror("Cannot write to cgroup.event_control");
+		goto out;
+	}
+
+	while (1) {
+		uint64_t result;
+
+		ret = read(efd, &result, sizeof(result));
+		if (ret == -1) {
+			if (errno == EINTR)
+				continue;
+			perror("Cannot read from eventfd");
+			break;
+		}
+		assert(ret == sizeof(result));
+
+		ret = access(event_control_path, W_OK);
+		if ((ret == -1) && (errno == ENOENT)) {
+				puts("The cgroup seems to have removed.");
+				ret = 0;
+				break;
+		}
+
+		if (ret == -1) {
+			perror("cgroup.event_control "
+					"is not accessable any more");
+			break;
+		}
+
+		printf("%s %s: crossed\n", argv[1], argv[2]);
+	}
+
+out:
+	if (efd >= 0)
+		close(efd);
+	if (event_control >= 0)
+		close(event_control);
+	if (cfd >= 0)
+		close(cfd);
+
+	return (ret != 0);
+}
-- 
1.6.6.2

[PATCH v2 -mmotm 3/4] cgroups: Add simple listener of cgroup events to documentation

[Kirill A. Shutemov]

An example of cgroup notification API usage.

Signed-off-by: Kirill A. Shutemov <kirill@shutemov.name>
Reviewed-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
---
 Documentation/cgroups/cgroup_event_listener.c |  103 +++++++++++++++++++++++++
 1 files changed, 103 insertions(+), 0 deletions(-)
 create mode 100644 Documentation/cgroups/cgroup_event_listener.c

diff --git a/Documentation/cgroups/cgroup_event_listener.c b/Documentation/cgroups/cgroup_event_listener.c
new file mode 100644
index 0000000..8c2d7aa
--- /dev/null
+++ b/Documentation/cgroups/cgroup_event_listener.c
@@ -0,0 +1,103 @@
+/*
+ * cgroup_event_listener.c - Simple listener of cgroup events
+ *
+ * Copyright (C) Kirill A. Shutemov <kirill@shutemov.name>
+ */
+
+#include <assert.h>
+#include <errno.h>
+#include <fcntl.h>
+#include <libgen.h>
+#include <limits.h>
+#include <stdio.h>
+#include <string.h>
+#include <unistd.h>
+
+#include <sys/eventfd.h>
+
+#define USAGE_STR "Usage: cgroup_event_listener <path-to-control-file> <args>\n"
+
+int main(int argc, char **argv)
+{
+	int efd = -1;
+	int cfd = -1;
+	int event_control = -1;
+	char event_control_path[PATH_MAX];
+	int ret;
+
+	if (argc != 3) {
+		fputs(USAGE_STR, stderr);
+		return 1;
+	}
+
+	cfd = open(argv[1], O_RDONLY);
+	if (cfd == -1) {
+		fprintf(stderr, "Cannot open %s: %s\n", argv[1],
+				strerror(errno));
+		goto out;
+	}
+
+	ret = snprintf(event_control_path, PATH_MAX, "%s/cgroup.event_control",
+			dirname(argv[1]));
+	if (ret > PATH_MAX) {
+		fputs("Path to cgroup.event_control is too long\n", stderr);
+		goto out;
+	}
+
+	event_control = open(event_control_path, O_WRONLY);
+	if (event_control == -1) {
+		fprintf(stderr, "Cannot open %s: %s\n", event_control_path,
+				strerror(errno));
+		goto out;
+	}
+
+	efd = eventfd(0, 0);
+	if (efd == -1) {
+		perror("eventfd() failed");
+		goto out;
+	}
+
+	ret = dprintf(event_control, "%d %d %s", efd, cfd, argv[2]);
+	if (ret == -1) {
+		perror("Cannot write to cgroup.event_control");
+		goto out;
+	}
+
+	while (1) {
+		uint64_t result;
+
+		ret = read(efd, &result, sizeof(result));
+		if (ret == -1) {
+			if (errno == EINTR)
+				continue;
+			perror("Cannot read from eventfd");
+			break;
+		}
+		assert(ret == sizeof(result));
+
+		ret = access(event_control_path, W_OK);
+		if ((ret == -1) && (errno == ENOENT)) {
+				puts("The cgroup seems to have removed.");
+				ret = 0;
+				break;
+		}
+
+		if (ret == -1) {
+			perror("cgroup.event_control "
+					"is not accessable any more");
+			break;
+		}
+
+		printf("%s %s: crossed\n", argv[1], argv[2]);
+	}
+
+out:
+	if (efd >= 0)
+		close(efd);
+	if (event_control >= 0)
+		close(event_control);
+	if (cfd >= 0)
+		close(cfd);
+
+	return (ret != 0);
+}
-- 
1.6.6.2

--
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>

[PATCH v2 -mmotm 4/4] memcg: Update memcg_test.txt to describe memory thresholds

[Kirill A. Shutemov]

Decription of sanity check for memory thresholds.

Signed-off-by: Kirill A. Shutemov <kirill-oKw7cIdHH8eLwutG50LtGA@public.gmane.org>
Acked-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu-+CUm20s59erQFUHtdCDX3A@public.gmane.org>
---
 Documentation/cgroups/memcg_test.txt |   21 +++++++++++++++++++++
 1 files changed, 21 insertions(+), 0 deletions(-)

diff --git a/Documentation/cgroups/memcg_test.txt b/Documentation/cgroups/memcg_test.txt
index e011488..4d32e0e 100644
--- a/Documentation/cgroups/memcg_test.txt
+++ b/Documentation/cgroups/memcg_test.txt
@@ -396,3 +396,24 @@ Under below explanation, we assume CONFIG_MEM_RES_CTRL_SWAP=y.
 	memory.stat of both A and B.
 	See 8.2 of Documentation/cgroups/memory.txt to see what value should be
 	written to move_charge_at_immigrate.
+
+ 9.10 Memory thresholds
+	Memory controler implements memory thresholds using cgroups notification
+	API. You can use Documentation/cgroups/cgroup_event_listener.c to test
+	it.
+
+	(Shell-A) Create cgroup and run event listener
+	# mkdir /cgroup/A
+	# ./cgroup_event_listener /cgroup/A/memory.usage_in_bytes 5M
+
+	(Shell-B) Add task to cgroup and try to allocate and free memory
+	# echo $$ >/cgroup/A/tasks
+	# a="$(dd if=/dev/zero bs=1M count=10)"
+	# a=
+
+	You will see message from cgroup_event_listener every time you cross
+	the thresholds.
+
+	Use /cgroup/A/memory.memsw.usage_in_bytes to test memsw thresholds.
+
+	It's good idea to test root cgroup as well.
-- 
1.6.6.2

[PATCH v2 -mmotm 4/4] memcg: Update memcg_test.txt to describe memory thresholds

[Kirill A. Shutemov]

Decription of sanity check for memory thresholds.

Signed-off-by: Kirill A. Shutemov <kirill@shutemov.name>
Acked-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
---
 Documentation/cgroups/memcg_test.txt |   21 +++++++++++++++++++++
 1 files changed, 21 insertions(+), 0 deletions(-)

diff --git a/Documentation/cgroups/memcg_test.txt b/Documentation/cgroups/memcg_test.txt
index e011488..4d32e0e 100644
--- a/Documentation/cgroups/memcg_test.txt
+++ b/Documentation/cgroups/memcg_test.txt
@@ -396,3 +396,24 @@ Under below explanation, we assume CONFIG_MEM_RES_CTRL_SWAP=y.
 	memory.stat of both A and B.
 	See 8.2 of Documentation/cgroups/memory.txt to see what value should be
 	written to move_charge_at_immigrate.
+
+ 9.10 Memory thresholds
+	Memory controler implements memory thresholds using cgroups notification
+	API. You can use Documentation/cgroups/cgroup_event_listener.c to test
+	it.
+
+	(Shell-A) Create cgroup and run event listener
+	# mkdir /cgroup/A
+	# ./cgroup_event_listener /cgroup/A/memory.usage_in_bytes 5M
+
+	(Shell-B) Add task to cgroup and try to allocate and free memory
+	# echo $$ >/cgroup/A/tasks
+	# a="$(dd if=/dev/zero bs=1M count=10)"
+	# a=
+
+	You will see message from cgroup_event_listener every time you cross
+	the thresholds.
+
+	Use /cgroup/A/memory.memsw.usage_in_bytes to test memsw thresholds.
+
+	It's good idea to test root cgroup as well.
-- 
1.6.6.2

--
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>

Re: [PATCH v2 -mmotm 1/4] cgroups: Fix race between userspace and kernelspace

[Li Zefan]

(Late reply for I just came back from a long vacation)

Kirill A. Shutemov wrote:
> eventfd are used to notify about two types of event:
>  - control file-specific, like crossing memory threshold;
>  - cgroup removing.
> 
> To understand what really happen, userspace can check if the cgroup
> still exists. To avoid race beetween userspace and kernelspace we have
> to notify userspace about cgroup removing only after rmdir of cgroup
> directory.
> 
> Signed-off-by: Kirill A. Shutemov <kirill-oKw7cIdHH8eLwutG50LtGA@public.gmane.org>
> Reviewed-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu-+CUm20s59erQFUHtdCDX3A@public.gmane.org>

Acked-by: Li Zefan <lizf-BthXqXjhjHXQFUHtdCDX3A@public.gmane.org>

Re: [PATCH v2 -mmotm 1/4] cgroups: Fix race between userspace and kernelspace

[Li Zefan]

(Late reply for I just came back from a long vacation)

Kirill A. Shutemov wrote:
> eventfd are used to notify about two types of event:
>  - control file-specific, like crossing memory threshold;
>  - cgroup removing.
> 
> To understand what really happen, userspace can check if the cgroup
> still exists. To avoid race beetween userspace and kernelspace we have
> to notify userspace about cgroup removing only after rmdir of cgroup
> directory.
> 
> Signed-off-by: Kirill A. Shutemov <kirill@shutemov.name>
> Reviewed-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>

Acked-by: Li Zefan <lizf@cn.fujitsu.com>

--
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>

Re: [PATCH v2 -mmotm 2/4] cgroups: remove events before destroying subsystem state objects

[Li Zefan]

Kirill A. Shutemov wrote:
> Events should be removed after rmdir of cgroup directory, but before
> destroying subsystem state objects. Let's take reference to cgroup
> directory dentry to do that.
> 
> Signed-off-by: Kirill A. Shutemov <kirill-oKw7cIdHH8eLwutG50LtGA@public.gmane.org>
> Acked-by: KAMEZAWA Hiroyuki <kamezawa.hioryu-+CUm20s59erQFUHtdCDX3A@public.gmane.org>

Looks good.

Acked-by: Li Zefan <lizf-BthXqXjhjHXQFUHtdCDX3A@public.gmane.org>

Re: [PATCH v2 -mmotm 2/4] cgroups: remove events before destroying subsystem state objects

[Li Zefan]

Kirill A. Shutemov wrote:
> Events should be removed after rmdir of cgroup directory, but before
> destroying subsystem state objects. Let's take reference to cgroup
> directory dentry to do that.
> 
> Signed-off-by: Kirill A. Shutemov <kirill@shutemov.name>
> Acked-by: KAMEZAWA Hiroyuki <kamezawa.hioryu@jp.fujitsu.com>

Looks good.

Acked-by: Li Zefan <lizf@cn.fujitsu.com>

--
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>

Re: [PATCH v2 -mmotm 3/4] cgroups: Add simple listener of cgroup events to documentation

[Li Zefan]

Kirill A. Shutemov wrote:
> An example of cgroup notification API usage.
> 
> Signed-off-by: Kirill A. Shutemov <kirill-oKw7cIdHH8eLwutG50LtGA@public.gmane.org>
> Reviewed-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu-+CUm20s59erQFUHtdCDX3A@public.gmane.org>

Acked-by: Li Zefan <lizf-BthXqXjhjHXQFUHtdCDX3A@public.gmane.org>

Re: [PATCH v2 -mmotm 3/4] cgroups: Add simple listener of cgroup events to documentation

[Li Zefan]

Kirill A. Shutemov wrote:
> An example of cgroup notification API usage.
> 
> Signed-off-by: Kirill A. Shutemov <kirill@shutemov.name>
> Reviewed-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>

Acked-by: Li Zefan <lizf@cn.fujitsu.com>

--
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>

Re: [PATCH v2 -mmotm 3/4] cgroups: Add simple listener of cgroup events to documentation

[Li Zefan]

> +	ret = dprintf(event_control, "%d %d %s", efd, cfd, argv[2]);

I found it won't return negative value for invalid input, though
errno is set properly.

try:
# ./cgroup_event_listner /cgroup/cgroup.procs abc

while strace shows write() does return -1:

# strace ./cgroup_event_listner /cgroup/cgroup.procs abc
...
write(6, "7 5 abc"..., 7)               = -1 EINVAL (Invalid argument)

> +	if (ret == -1) {
> +		perror("Cannot write to cgroup.event_control");
> +		goto out;
> +	}

Re: [PATCH v2 -mmotm 3/4] cgroups: Add simple listener of cgroup events to documentation

[Li Zefan]

> +	ret = dprintf(event_control, "%d %d %s", efd, cfd, argv[2]);

I found it won't return negative value for invalid input, though
errno is set properly.

try:
# ./cgroup_event_listner /cgroup/cgroup.procs abc

while strace shows write() does return -1:

# strace ./cgroup_event_listner /cgroup/cgroup.procs abc
...
write(6, "7 5 abc"..., 7)               = -1 EINVAL (Invalid argument)

> +	if (ret == -1) {
> +		perror("Cannot write to cgroup.event_control");
> +		goto out;
> +	}

--
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>

Re: [PATCH v2 -mmotm 1/4] cgroups: Fix race between userspace and kernelspace

[Balbir Singh]

* Kirill A. Shutemov <kirill-oKw7cIdHH8eLwutG50LtGA@public.gmane.org> [2010-02-22 17:43:39]:

> eventfd are used to notify about two types of event:
>  - control file-specific, like crossing memory threshold;
>  - cgroup removing.
> 
> To understand what really happen, userspace can check if the cgroup
> still exists. To avoid race beetween userspace and kernelspace we have
> to notify userspace about cgroup removing only after rmdir of cgroup
> directory.
> 
> Signed-off-by: Kirill A. Shutemov <kirill-oKw7cIdHH8eLwutG50LtGA@public.gmane.org>
> Reviewed-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu-+CUm20s59erQFUHtdCDX3A@public.gmane.org>

That does make sense, looks good to me. You've already got the
necessary acks.

-- 
	Three Cheers,
	Balbir

Re: [PATCH v2 -mmotm 1/4] cgroups: Fix race between userspace and kernelspace

[Balbir Singh]

* Kirill A. Shutemov <kirill@shutemov.name> [2010-02-22 17:43:39]:

> eventfd are used to notify about two types of event:
>  - control file-specific, like crossing memory threshold;
>  - cgroup removing.
> 
> To understand what really happen, userspace can check if the cgroup
> still exists. To avoid race beetween userspace and kernelspace we have
> to notify userspace about cgroup removing only after rmdir of cgroup
> directory.
> 
> Signed-off-by: Kirill A. Shutemov <kirill@shutemov.name>
> Reviewed-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>

That does make sense, looks good to me. You've already got the
necessary acks.

-- 
	Three Cheers,
	Balbir

--
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>

Re: [PATCH v2 -mmotm 2/4] cgroups: remove events before destroying subsystem state objects

[Balbir Singh]

* Kirill A. Shutemov <kirill-oKw7cIdHH8eLwutG50LtGA@public.gmane.org> [2010-02-22 17:43:40]:

> Events should be removed after rmdir of cgroup directory, but before
> destroying subsystem state objects. Let's take reference to cgroup
> directory dentry to do that.
> 
> Signed-off-by: Kirill A. Shutemov <kirill-oKw7cIdHH8eLwutG50LtGA@public.gmane.org>
> Acked-by: KAMEZAWA Hiroyuki <kamezawa.hioryu-+CUm20s59erQFUHtdCDX3A@public.gmane.org>

Looks good, but remember the mem_cgroup data structure will can
disappear after the rmdir

-- 
	Three Cheers,
	Balbir

Re: [PATCH v2 -mmotm 2/4] cgroups: remove events before destroying subsystem state objects

[Balbir Singh]

* Kirill A. Shutemov <kirill@shutemov.name> [2010-02-22 17:43:40]:

> Events should be removed after rmdir of cgroup directory, but before
> destroying subsystem state objects. Let's take reference to cgroup
> directory dentry to do that.
> 
> Signed-off-by: Kirill A. Shutemov <kirill@shutemov.name>
> Acked-by: KAMEZAWA Hiroyuki <kamezawa.hioryu@jp.fujitsu.com>

Looks good, but remember the mem_cgroup data structure will can
disappear after the rmdir

-- 
	Three Cheers,
	Balbir

--
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>

Re: [PATCH v2 -mmotm 3/4] cgroups: Add simple listener of cgroup events to documentation

[Kirill A. Shutemov]

On Wed, Feb 24, 2010 at 5:30 AM, Li Zefan <lizf@cn.fujitsu.com> wrote:
>> +     ret = dprintf(event_control, "%d %d %s", efd, cfd, argv[2]);
>
> I found it won't return negative value for invalid input, though
> errno is set properly.

It looks like a glibc bug. I've file bug to glibc bugzilla:

http://sourceware.org/bugzilla/show_bug.cgi?id=11319

I'll fix cgroup_event_listener.c. Thanks!

> try:
> # ./cgroup_event_listner /cgroup/cgroup.procs abc
>
> while strace shows write() does return -1:
>
> # strace ./cgroup_event_listner /cgroup/cgroup.procs abc
> ...
> write(6, "7 5 abc"..., 7)               = -1 EINVAL (Invalid argument)
>
>> +     if (ret == -1) {
>> +             perror("Cannot write to cgroup.event_control");
>> +             goto out;
>> +     }
>
> --
> 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>
>
_______________________________________________
Containers mailing list
Containers@lists.linux-foundation.org
https://lists.linux-foundation.org/mailman/listinfo/containers

Re: [PATCH v2 -mmotm 3/4] cgroups: Add simple listener of cgroup events to documentation

[Kirill A. Shutemov]

On Wed, Feb 24, 2010 at 5:30 AM, Li Zefan <lizf@cn.fujitsu.com> wrote:
>> +     ret = dprintf(event_control, "%d %d %s", efd, cfd, argv[2]);
>
> I found it won't return negative value for invalid input, though
> errno is set properly.

It looks like a glibc bug. I've file bug to glibc bugzilla:

http://sourceware.org/bugzilla/show_bug.cgi?id=11319

I'll fix cgroup_event_listener.c. Thanks!

> try:
> # ./cgroup_event_listner /cgroup/cgroup.procs abc
>
> while strace shows write() does return -1:
>
> # strace ./cgroup_event_listner /cgroup/cgroup.procs abc
> ...
> write(6, "7 5 abc"..., 7)               = -1 EINVAL (Invalid argument)
>
>> +     if (ret == -1) {
>> +             perror("Cannot write to cgroup.event_control");
>> +             goto out;
>> +     }
>
> --
> 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>
>

--
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>

Re: [PATCH v2 -mmotm 2/4] cgroups: remove events before destroying subsystem state objects

[Kirill A. Shutemov]

On Wed, Feb 24, 2010 at 10:40 AM, Balbir Singh
<balbir-23VcF4HTsmIX0ybBhKVfKdBPR1lH4CV8@public.gmane.org> wrote:
> * Kirill A. Shutemov <kirill-oKw7cIdHH8eLwutG50LtGA@public.gmane.org> [2010-02-22 17:43:40]:
>
>> Events should be removed after rmdir of cgroup directory, but before
>> destroying subsystem state objects. Let's take reference to cgroup
>> directory dentry to do that.
>>
>> Signed-off-by: Kirill A. Shutemov <kirill-oKw7cIdHH8eLwutG50LtGA@public.gmane.org>
>> Acked-by: KAMEZAWA Hiroyuki <kamezawa.hioryu-+CUm20s59erQFUHtdCDX3A@public.gmane.org>
>
> Looks good, but remember the mem_cgroup data structure will can
> disappear after the rmdir

IIUC, struct mem_cgroup can be freed only after ->destroy(), which can
be called only if there is no references to cgroup directory dentry.

Have I missed something?

Re: [PATCH v2 -mmotm 2/4] cgroups: remove events before destroying subsystem state objects

[Kirill A. Shutemov]

On Wed, Feb 24, 2010 at 10:40 AM, Balbir Singh
<balbir@linux.vnet.ibm.com> wrote:
> * Kirill A. Shutemov <kirill@shutemov.name> [2010-02-22 17:43:40]:
>
>> Events should be removed after rmdir of cgroup directory, but before
>> destroying subsystem state objects. Let's take reference to cgroup
>> directory dentry to do that.
>>
>> Signed-off-by: Kirill A. Shutemov <kirill@shutemov.name>
>> Acked-by: KAMEZAWA Hiroyuki <kamezawa.hioryu@jp.fujitsu.com>
>
> Looks good, but remember the mem_cgroup data structure will can
> disappear after the rmdir

IIUC, struct mem_cgroup can be freed only after ->destroy(), which can
be called only if there is no references to cgroup directory dentry.

Have I missed something?

--
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>

Re: [PATCH v2 -mmotm 2/4] cgroups: remove events before destroying subsystem state objects

[Balbir Singh]

* Kirill A. Shutemov <kirill-oKw7cIdHH8eLwutG50LtGA@public.gmane.org> [2010-02-24 13:42:15]:

> On Wed, Feb 24, 2010 at 10:40 AM, Balbir Singh
> <balbir-23VcF4HTsmIX0ybBhKVfKdBPR1lH4CV8@public.gmane.org> wrote:
> > * Kirill A. Shutemov <kirill-oKw7cIdHH8eLwutG50LtGA@public.gmane.org> [2010-02-22 17:43:40]:
> >
> >> Events should be removed after rmdir of cgroup directory, but before
> >> destroying subsystem state objects. Let's take reference to cgroup
> >> directory dentry to do that.
> >>
> >> Signed-off-by: Kirill A. Shutemov <kirill-oKw7cIdHH8eLwutG50LtGA@public.gmane.org>
> >> Acked-by: KAMEZAWA Hiroyuki <kamezawa.hioryu-+CUm20s59erQFUHtdCDX3A@public.gmane.org>
> >
> > Looks good, but remember the mem_cgroup data structure will can
> > disappear after the rmdir
> 
> IIUC, struct mem_cgroup can be freed only after ->destroy(), which can
> be called only if there is no references to cgroup directory dentry.
>

No.. You've got it right, it disappears after the last dput(). 

-- 
	Three Cheers,
	Balbir

Re: [PATCH v2 -mmotm 2/4] cgroups: remove events before destroying subsystem state objects

[Balbir Singh]

* Kirill A. Shutemov <kirill@shutemov.name> [2010-02-24 13:42:15]:

> On Wed, Feb 24, 2010 at 10:40 AM, Balbir Singh
> <balbir@linux.vnet.ibm.com> wrote:
> > * Kirill A. Shutemov <kirill@shutemov.name> [2010-02-22 17:43:40]:
> >
> >> Events should be removed after rmdir of cgroup directory, but before
> >> destroying subsystem state objects. Let's take reference to cgroup
> >> directory dentry to do that.
> >>
> >> Signed-off-by: Kirill A. Shutemov <kirill@shutemov.name>
> >> Acked-by: KAMEZAWA Hiroyuki <kamezawa.hioryu@jp.fujitsu.com>
> >
> > Looks good, but remember the mem_cgroup data structure will can
> > disappear after the rmdir
> 
> IIUC, struct mem_cgroup can be freed only after ->destroy(), which can
> be called only if there is no references to cgroup directory dentry.
>

No.. You've got it right, it disappears after the last dput(). 

-- 
	Three Cheers,
	Balbir

--
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>