Skip to contentSkip to conversation

mm: do not call mem_cgroup_free() from within mem_cgroup_alloc()

PATCH Superseded Tahsin Erdogan 2 participants 5 messages Mar 4, 2017 – Mar 7, 2017

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 mm: do not call mem_cgroup_free() from within mem_cgroup_alloc()

mem_cgroup_free() indirectly calls wb_domain_exit() which is not prepared to deal with a struct wb_domain object that hasn't executed wb_domain_init(). For instance, the following warning message is printed by lockdep if alloc_percpu() fails in mem_cgroup_alloc():

INFO: trying to register non-static key.
the code is fine but needs lockdep annotation.
turning off the locking correctness validator.

  CPU: 1 PID: 1950 Comm: mkdir Not tainted 4.10.0+ #151
  Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS Bochs 01/01/2011
  Call Trace:
   dump_stack+0x67/0x99
   register_lock_class+0x36d/0x540
   __lock_acquire+0x7f/0x1a30
   ? irq_work_queue+0x73/0x90
   ? wake_up_klogd+0x36/0x40
   ? console_unlock+0x45d/0x540
   ? vprintk_emit+0x211/0x2e0
   lock_acquire+0xcc/0x200
   ? try_to_del_timer_sync+0x60/0x60
   del_timer_sync+0x3c/0xc0
   ? try_to_del_timer_sync+0x60/0x60
   wb_domain_exit+0x14/0x20
   mem_cgroup_free+0x14/0x40
   mem_cgroup_css_alloc+0x3f9/0x620
   cgroup_apply_control_enable+0x190/0x390
   cgroup_mkdir+0x290/0x3d0
   kernfs_iop_mkdir+0x58/0x80
   vfs_mkdir+0x10e/0x1a0
   SyS_mkdirat+0xa8/0xd0
   SyS_mkdir+0x14/0x20
   entry_SYSCALL_64_fastpath+0x18/0xad

Fix mem_cgroup_alloc() by doing more granular clean up in case of failures.

Fixes 0b8f73e104285 ("mm: memcontrol: clean up alloc, online, offline, free functions")
---
 mm/memcontrol.c | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)
mm/memcontrol.c +4 -1
index c52ec893e241..9a9d5630df91 100644
@@ -4194,9 +4194,12 @@ static struct mem_cgroup *mem_cgroup_alloc(void)
 	idr_replace(&mem_cgroup_idr, memcg, memcg->id.id);
 	return memcg;
 fail:
+	for_each_node(node)
+		free_mem_cgroup_per_node_info(memcg, node);
+	free_percpu(memcg->stat);
 	if (memcg->id.id > 0)
 		idr_remove(&mem_cgroup_idr, memcg->id.id);
-	mem_cgroup_free(memcg);
+	kfree(memcg);
 	return NULL;
 }
 
Signature · 7 lines
2.12.0.rc1.440.g5b76565f74-goog

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to [email protected].  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"[email protected]"> [email protected] </a>
maintainer
#1 lore raw
Quoted 33 lines · Tahsin Erdogan

mem_cgroup_free() indirectly calls wb_domain_exit() which is not prepared to deal with a struct wb_domain object that hasn't executed wb_domain_init(). For instance, the following warning message is printed by lockdep if alloc_percpu() fails in mem_cgroup_alloc():

INFO: trying to register non-static key.
the code is fine but needs lockdep annotation.
turning off the locking correctness validator.

  CPU: 1 PID: 1950 Comm: mkdir Not tainted 4.10.0+ #151
  Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS Bochs 01/01/2011
  Call Trace:
   dump_stack+0x67/0x99
   register_lock_class+0x36d/0x540
   __lock_acquire+0x7f/0x1a30
   ? irq_work_queue+0x73/0x90
   ? wake_up_klogd+0x36/0x40
   ? console_unlock+0x45d/0x540
   ? vprintk_emit+0x211/0x2e0
   lock_acquire+0xcc/0x200
   ? try_to_del_timer_sync+0x60/0x60
   del_timer_sync+0x3c/0xc0
   ? try_to_del_timer_sync+0x60/0x60
   wb_domain_exit+0x14/0x20
   mem_cgroup_free+0x14/0x40
   mem_cgroup_css_alloc+0x3f9/0x620
   cgroup_apply_control_enable+0x190/0x390
   cgroup_mkdir+0x290/0x3d0
   kernfs_iop_mkdir+0x58/0x80
   vfs_mkdir+0x10e/0x1a0
   SyS_mkdirat+0xa8/0xd0
   SyS_mkdir+0x14/0x20
   entry_SYSCALL_64_fastpath+0x18/0xad
Quoted 5 lines

Fix mem_cgroup_alloc() by doing more granular clean up in case of failures.

Fixes 0b8f73e104285 ("mm: memcontrol: clean up alloc, online, offline, free functions")

Please do not duplicate mem_cgroup_free code and rather add __mem_cgroup_free which does everything except for wb_domain_exit. An alternative would be to teach memcg_wb_domain_exit to not call wb_domain_exit if it hasn't been initialized yet. The first option seems easier.

Thanks!

Quoted 30 lines
---
 mm/memcontrol.c | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)
mm/memcontrol.c +4 -1
index c52ec893e241..9a9d5630df91 100644
@@ -4194,9 +4194,12 @@ static struct mem_cgroup *mem_cgroup_alloc(void)
 	idr_replace(&mem_cgroup_idr, memcg, memcg->id.id);
 	return memcg;
 fail:
+	for_each_node(node)
+		free_mem_cgroup_per_node_info(memcg, node);
+	free_percpu(memcg->stat);
 	if (memcg->id.id > 0)
 		idr_remove(&mem_cgroup_idr, memcg->id.id);
-	mem_cgroup_free(memcg);
+	kfree(memcg);
 	return NULL;
 }
 
Signature · 7 lines
2.12.0.rc1.440.g5b76565f74-goog

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to [email protected].  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"[email protected]"> [email protected] </a>
Signature · 8 lines
Michal Hocko
SUSE Labs

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to [email protected].  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"[email protected]"> [email protected] </a>
#2 lore raw

PATCH v2 1/1 mm: do not call mem_cgroup_free() from within mem_cgroup_alloc()

mem_cgroup_free() indirectly calls wb_domain_exit() which is not prepared to deal with a struct wb_domain object that hasn't executed wb_domain_init(). For instance, the following warning message is printed by lockdep if alloc_percpu() fails in mem_cgroup_alloc():

INFO: trying to register non-static key.
the code is fine but needs lockdep annotation.
turning off the locking correctness validator.

  CPU: 1 PID: 1950 Comm: mkdir Not tainted 4.10.0+ #151
  Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS Bochs 01/01/2011
  Call Trace:
   dump_stack+0x67/0x99
   register_lock_class+0x36d/0x540
   __lock_acquire+0x7f/0x1a30
   ? irq_work_queue+0x73/0x90
   ? wake_up_klogd+0x36/0x40
   ? console_unlock+0x45d/0x540
   ? vprintk_emit+0x211/0x2e0
   lock_acquire+0xcc/0x200
   ? try_to_del_timer_sync+0x60/0x60
   del_timer_sync+0x3c/0xc0
   ? try_to_del_timer_sync+0x60/0x60
   wb_domain_exit+0x14/0x20
   mem_cgroup_free+0x14/0x40
   mem_cgroup_css_alloc+0x3f9/0x620
   cgroup_apply_control_enable+0x190/0x390
   cgroup_mkdir+0x290/0x3d0
   kernfs_iop_mkdir+0x58/0x80
   vfs_mkdir+0x10e/0x1a0
   SyS_mkdirat+0xa8/0xd0
   SyS_mkdir+0x14/0x20
   entry_SYSCALL_64_fastpath+0x18/0xad

Add __mem_cgroup_free() which skips wb_domain_exit(). This is used by both mem_cgroup_free() and mem_cgroup_alloc() clean up.

Fixes 0b8f73e104285 ("mm: memcontrol: clean up alloc, online, offline, free functions")

---
v2:
Added __mem_cgroup_free()

 mm/memcontrol.c | 11 ++++++++---
 1 file changed, 8 insertions(+), 3 deletions(-)
mm/memcontrol.c +8 -3
index c52ec893e241..e7d900c5f2d0 100644
@@ -4135,17 +4135,22 @@ static void free_mem_cgroup_per_node_info(struct mem_cgroup *memcg, int node)
 	kfree(memcg->nodeinfo[node]);
 }
 
-static void mem_cgroup_free(struct mem_cgroup *memcg)
+static void __mem_cgroup_free(struct mem_cgroup *memcg)
 {
 	int node;
 
-	memcg_wb_domain_exit(memcg);
 	for_each_node(node)
 		free_mem_cgroup_per_node_info(memcg, node);
 	free_percpu(memcg->stat);
 	kfree(memcg);
 }
 
+static void mem_cgroup_free(struct mem_cgroup *memcg)
+{
+	memcg_wb_domain_exit(memcg);
+	__mem_cgroup_free(memcg);
+}
+
 static struct mem_cgroup *mem_cgroup_alloc(void)
 {
 	struct mem_cgroup *memcg;
@@ -4196,7 +4201,7 @@ static struct mem_cgroup *mem_cgroup_alloc(void)
 fail:
 	if (memcg->id.id > 0)
 		idr_remove(&mem_cgroup_idr, memcg->id.id);
-	mem_cgroup_free(memcg);
+	__mem_cgroup_free(memcg);
 	return NULL;
 }
 
Signature · 7 lines
2.12.0.rc1.440.g5b76565f74-goog

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to [email protected].  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"[email protected]"> [email protected] </a>
maintainer Acked-by on v2 1/1
#3 lore raw
Quoted 33 lines · Tahsin Erdogan

mem_cgroup_free() indirectly calls wb_domain_exit() which is not prepared to deal with a struct wb_domain object that hasn't executed wb_domain_init(). For instance, the following warning message is printed by lockdep if alloc_percpu() fails in mem_cgroup_alloc():

INFO: trying to register non-static key.
the code is fine but needs lockdep annotation.
turning off the locking correctness validator.

  CPU: 1 PID: 1950 Comm: mkdir Not tainted 4.10.0+ #151
  Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS Bochs 01/01/2011
  Call Trace:
   dump_stack+0x67/0x99
   register_lock_class+0x36d/0x540
   __lock_acquire+0x7f/0x1a30
   ? irq_work_queue+0x73/0x90
   ? wake_up_klogd+0x36/0x40
   ? console_unlock+0x45d/0x540
   ? vprintk_emit+0x211/0x2e0
   lock_acquire+0xcc/0x200
   ? try_to_del_timer_sync+0x60/0x60
   del_timer_sync+0x3c/0xc0
   ? try_to_del_timer_sync+0x60/0x60
   wb_domain_exit+0x14/0x20
   mem_cgroup_free+0x14/0x40
   mem_cgroup_css_alloc+0x3f9/0x620
   cgroup_apply_control_enable+0x190/0x390
   cgroup_mkdir+0x290/0x3d0
   kernfs_iop_mkdir+0x58/0x80
   vfs_mkdir+0x10e/0x1a0
   SyS_mkdirat+0xa8/0xd0
   SyS_mkdir+0x14/0x20
   entry_SYSCALL_64_fastpath+0x18/0xad
Quoted 5 lines

Add __mem_cgroup_free() which skips wb_domain_exit(). This is used by both mem_cgroup_free() and mem_cgroup_alloc() clean up.

Fixes 0b8f73e104285 ("mm: memcontrol: clean up alloc, online, offline, free functions")

Thanks!

Quoted 47 lines

---
v2:
Added __mem_cgroup_free()

 mm/memcontrol.c | 11 ++++++++---
 1 file changed, 8 insertions(+), 3 deletions(-)
mm/memcontrol.c +8 -3
index c52ec893e241..e7d900c5f2d0 100644
@@ -4135,17 +4135,22 @@ static void free_mem_cgroup_per_node_info(struct mem_cgroup *memcg, int node)
 	kfree(memcg->nodeinfo[node]);
 }
 
-static void mem_cgroup_free(struct mem_cgroup *memcg)
+static void __mem_cgroup_free(struct mem_cgroup *memcg)
 {
 	int node;
 
-	memcg_wb_domain_exit(memcg);
 	for_each_node(node)
 		free_mem_cgroup_per_node_info(memcg, node);
 	free_percpu(memcg->stat);
 	kfree(memcg);
 }
 
+static void mem_cgroup_free(struct mem_cgroup *memcg)
+{
+	memcg_wb_domain_exit(memcg);
+	__mem_cgroup_free(memcg);
+}
+
 static struct mem_cgroup *mem_cgroup_alloc(void)
 {
 	struct mem_cgroup *memcg;
@@ -4196,7 +4201,7 @@ static struct mem_cgroup *mem_cgroup_alloc(void)
 fail:
 	if (memcg->id.id > 0)
 		idr_remove(&mem_cgroup_idr, memcg->id.id);
-	mem_cgroup_free(memcg);
+	__mem_cgroup_free(memcg);
 	return NULL;
 }
 
Signature · 1 line
2.12.0.rc1.440.g5b76565f74-goog
Signature · 8 lines
Michal Hocko
SUSE Labs

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to [email protected].  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"[email protected]"> [email protected] </a>
maintainer
#4 lore raw
In reply to Tahsin Erdogan

[CC Andrew]

Quoted 85 lines · Tahsin Erdogan

mem_cgroup_free() indirectly calls wb_domain_exit() which is not prepared to deal with a struct wb_domain object that hasn't executed wb_domain_init(). For instance, the following warning message is printed by lockdep if alloc_percpu() fails in mem_cgroup_alloc():

INFO: trying to register non-static key.
the code is fine but needs lockdep annotation.
turning off the locking correctness validator.

  CPU: 1 PID: 1950 Comm: mkdir Not tainted 4.10.0+ #151
  Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS Bochs 01/01/2011
  Call Trace:
   dump_stack+0x67/0x99
   register_lock_class+0x36d/0x540
   __lock_acquire+0x7f/0x1a30
   ? irq_work_queue+0x73/0x90
   ? wake_up_klogd+0x36/0x40
   ? console_unlock+0x45d/0x540
   ? vprintk_emit+0x211/0x2e0
   lock_acquire+0xcc/0x200
   ? try_to_del_timer_sync+0x60/0x60
   del_timer_sync+0x3c/0xc0
   ? try_to_del_timer_sync+0x60/0x60
   wb_domain_exit+0x14/0x20
   mem_cgroup_free+0x14/0x40
   mem_cgroup_css_alloc+0x3f9/0x620
   cgroup_apply_control_enable+0x190/0x390
   cgroup_mkdir+0x290/0x3d0
   kernfs_iop_mkdir+0x58/0x80
   vfs_mkdir+0x10e/0x1a0
   SyS_mkdirat+0xa8/0xd0
   SyS_mkdir+0x14/0x20
   entry_SYSCALL_64_fastpath+0x18/0xad

Add __mem_cgroup_free() which skips wb_domain_exit(). This is used by both mem_cgroup_free() and mem_cgroup_alloc() clean up.

Fixes 0b8f73e104285 ("mm: memcontrol: clean up alloc, online, offline, free functions")

---
v2:
Added __mem_cgroup_free()

 mm/memcontrol.c | 11 ++++++++---
 1 file changed, 8 insertions(+), 3 deletions(-)
mm/memcontrol.c +8 -3
index c52ec893e241..e7d900c5f2d0 100644
@@ -4135,17 +4135,22 @@ static void free_mem_cgroup_per_node_info(struct mem_cgroup *memcg, int node)
 	kfree(memcg->nodeinfo[node]);
 }
 
-static void mem_cgroup_free(struct mem_cgroup *memcg)
+static void __mem_cgroup_free(struct mem_cgroup *memcg)
 {
 	int node;
 
-	memcg_wb_domain_exit(memcg);
 	for_each_node(node)
 		free_mem_cgroup_per_node_info(memcg, node);
 	free_percpu(memcg->stat);
 	kfree(memcg);
 }
 
+static void mem_cgroup_free(struct mem_cgroup *memcg)
+{
+	memcg_wb_domain_exit(memcg);
+	__mem_cgroup_free(memcg);
+}
+
 static struct mem_cgroup *mem_cgroup_alloc(void)
 {
 	struct mem_cgroup *memcg;
@@ -4196,7 +4201,7 @@ static struct mem_cgroup *mem_cgroup_alloc(void)
 fail:
 	if (memcg->id.id > 0)
 		idr_remove(&mem_cgroup_idr, memcg->id.id);
-	mem_cgroup_free(memcg);
+	__mem_cgroup_free(memcg);
 	return NULL;
 }
 
Signature · 1 line
2.12.0.rc1.440.g5b76565f74-goog
Signature · 8 lines
Michal Hocko
SUSE Labs

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