Skip to contentSkip to conversation

optimize freezing tasks by reducing task wakeups

PATCH In review Colin Cross 8 participants 42 messages May 6, 2013 – Jul 23, 2013

Subsystems (13)
lore
Share & tools
lore
#0 lore raw

v2 cover optimize freezing tasks by reducing task wakeups

On slow cpus the large number of task wakeups and context switches triggered by freezing and thawing tasks can take a significant amount of cpu time. This patch series reduces the amount of work done during freezing tasks by avoiding waking up tasks that are already in a freezable state.

The first 4 patches reintroduce 6aa9707099c (lockdep: check that no locks held at freeze time) which was reverted in dbf520a9d7d4, and fix up the known callers with locks held in NFS and CIFS to skip the lockdep check for now. The lockdep check will warn any future incorrect users of the freezable helpers.

The fifth patch reduces the wasted time in try_to_freeze_tasks() by starting with a 1 ms sleep during the first loop and backing off up to an 8 ms sleep if all tasks are not frozen.

The sixth patch modifies the freeze_task() function to skip tasks that have set the PF_FREEZER_SKIP flag by calling freezer_do_not_count(). These tasks will not enter the refrigerator during the suspend/resume cycle unless they woken up by something else, in which case they will enter the refrigerator in freezer_count() before they access any resources that would not be available in suspend or deadlock with another freezing/frozen task.

The rest of the series adds a few more freezable helpers and converts the top call sites that userspace tasks are usually blocked at to freezable helpers. The list of call sites was collected on a Nexus 10 (ARM Exynos 5250 SoC), but all the top call sites other than binder show up at the top of the list on Ubuntu x86-64 as well.

This series cuts the time for freezing tasks from 50 ms to 5 ms when the cpu speed is locked at its lowest setting (200MHz), and reduces the number of context switches and restarted syscalls from 1000 to 25.

v2 moves the skip check to freeze_task(), and expands the commit messages.

v3 adds the patches to reintroduce the lockdep check to this patchset, adds a patch to convert the freezable helpers to static inlines when possible, and splits the patch that adds the new helpers out of the one that converts the existing helpers to use freezer_do_not_count.

#1 lore raw

PATCH v3 1/16 freezer: add unsafe versions of freezable helpers for NFS

NFS calls the freezable helpers with locks held, which is unsafe and will cause lockdep warnings when 6aa9707 "lockdep: check that no locks held at freeze time" is reapplied (it was reverted in dbf520a). NFS shouldn't be doing this, but it has long-running syscalls that must hold a lock but also shouldn't block suspend. Until NFS freeze handling is rewritten to use a signal to exit out of the critical section, add new *_unsafe versions of the helpers that will not run the lockdep test when 6aa9707 is reapplied, and call them from NFS.

In practice the likley result of holding the lock while freezing is that a second task blocked on the lock will never freeze, aborting suspend, but it is possible to manufacture a case using the cgroup freezer, the lock, and the suspend freezer to create a deadlock. Silencing the lockdep warning here will allow problems to be found in other drivers that may have a more serious deadlock risk, and prevent new problems from being added.

---
 fs/nfs/inode.c          |  2 +-
 fs/nfs/nfs3proc.c       |  2 +-
 fs/nfs/nfs4proc.c       |  4 ++--
 include/linux/freezer.h | 42 +++++++++++++++++++++++++++++++++++++++++-
 net/sunrpc/sched.c      |  2 +-
 5 files changed, 46 insertions(+), 6 deletions(-)
fs/nfs/inode.c +1 -1
index 1f94167..53cbee5 100644
@@ -79,7 +79,7 @@ int nfs_wait_bit_killable(void *word)
 {
 	if (fatal_signal_pending(current))
 		return -ERESTARTSYS;
-	freezable_schedule();
+	freezable_schedule_unsafe();
 	return 0;
 }
 EXPORT_SYMBOL_GPL(nfs_wait_bit_killable);
fs/nfs/nfs3proc.c +1 -1
index 43ea96c..ce90eb4 100644
@@ -33,7 +33,7 @@ nfs3_rpc_wrapper(struct rpc_clnt *clnt, struct rpc_message *msg, int flags)
 		res = rpc_call_sync(clnt, msg, flags);
 		if (res != -EJUKEBOX)
 			break;
-		freezable_schedule_timeout_killable(NFS_JUKEBOX_RETRY_TIME);
+		freezable_schedule_timeout_killable_unsafe(NFS_JUKEBOX_RETRY_TIME);
 		res = -ERESTARTSYS;
 	} while (!fatal_signal_pending(current));
 	return res;
fs/nfs/nfs4proc.c +2 -2
index 0ad025e..a236077 100644
@@ -266,7 +266,7 @@ static int nfs4_delay(struct rpc_clnt *clnt, long *timeout)
 		*timeout = NFS4_POLL_RETRY_MIN;
 	if (*timeout > NFS4_POLL_RETRY_MAX)
 		*timeout = NFS4_POLL_RETRY_MAX;
-	freezable_schedule_timeout_killable(*timeout);
+	freezable_schedule_timeout_killable_unsafe(*timeout);
 	if (fatal_signal_pending(current))
 		res = -ERESTARTSYS;
 	*timeout <<= 1;
@@ -4309,7 +4309,7 @@ int nfs4_proc_delegreturn(struct inode *inode, struct rpc_cred *cred, const nfs4
 static unsigned long
 nfs4_set_lock_task_retry(unsigned long timeout)
 {
-	freezable_schedule_timeout_killable(timeout);
+	freezable_schedule_timeout_killable_unsafe(timeout);
 	timeout <<= 1;
 	if (timeout > NFS4_LOCK_MAXTIMEOUT)
 		return NFS4_LOCK_MAXTIMEOUT;
include/linux/freezer.h +41 -1

81 lines not on this page · raw message

net/sunrpc/sched.c +1 -1
index f8529fc..8dcfadc 100644
@@ -254,7 +254,7 @@ static int rpc_wait_bit_killable(void *word)
 {
 	if (fatal_signal_pending(current))
 		return -ERESTARTSYS;
-	freezable_schedule();
+	freezable_schedule_unsafe();
 	return 0;
 }
 
Signature · 1 line
1.8.2.1
#2 lore raw

PATCH v3 2/16 freezer: add unsafe versions of freezable helpers for CIFS

In reply to Colin Cross

CIFS calls wait_event_freezekillable_unsafe with a VFS lock held, which is unsafe and will cause lockdep warnings when 6aa9707 "lockdep: check that no locks held at freeze time" is reapplied (it was reverted in dbf520a). CIFS shouldn't be doing this, but it has long-running syscalls that must hold a lock but also shouldn't block suspend. Until CIFS freeze handling is rewritten to use a signal to exit out of the critical section, add a new wait_event_freezekillable_unsafe helper that will not run the lockdep test when 6aa9707 is reapplied, and call it from CIFS.

In practice the likley result of holding the lock while freezing is that a second task blocked on the lock will never freeze, aborting suspend, but it is possible to manufacture a case using the cgroup freezer, the lock, and the suspend freezer to create a deadlock. Silencing the lockdep warning here will allow problems to be found in other drivers that may have a more serious deadlock risk, and prevent new problems from being added.

---
 include/linux/freezer.h | 13 +++++++++++++
 1 file changed, 13 insertions(+)
include/linux/freezer.h +13 -0
index 5b31e21c..d3c038e 100644
@@ -212,6 +212,16 @@ static inline bool freezer_should_skip(struct task_struct *p)
 	__retval;							\
 })
 
+/* DO NOT ADD ANY NEW CALLERS OF THIS FUNCTION */
+#define wait_event_freezekillable_unsafe(wq, condition)			\
+({									\
+	int __retval;							\
+	freezer_do_not_count();						\
+	__retval = wait_event_killable(wq, (condition));		\
+	freezer_count_unsafe();						\
+	__retval;							\
+})
+
 #define wait_event_freezable(wq, condition)				\
 ({									\
 	int __retval;							\
@@ -277,6 +287,9 @@ static inline void set_freezable(void) {}
 #define wait_event_freezekillable(wq, condition)		\
 		wait_event_killable(wq, condition)
 
+#define wait_event_freezekillable_unsafe(wq, condition)			\
+		wait_event_killable(wq, condition)
+
 #endif /* !CONFIG_FREEZER */
 
 #endif	/* FREEZER_H_INCLUDED */
Signature · 1 line
1.8.2.1
#3 lore raw

Re: [PATCH v3 02/16] freezer: add unsafe versions of freezable helpers for CIFS

Quoted 45 lines · On Mon, 6 May 2013 16:50:07 -0700 Colin Cross <[email protected]> wrote:

CIFS calls wait_event_freezekillable_unsafe with a VFS lock held, which is unsafe and will cause lockdep warnings when 6aa9707 "lockdep: check that no locks held at freeze time" is reapplied (it was reverted in dbf520a). CIFS shouldn't be doing this, but it has long-running syscalls that must hold a lock but also shouldn't block suspend. Until CIFS freeze handling is rewritten to use a signal to exit out of the critical section, add a new wait_event_freezekillable_unsafe helper that will not run the lockdep test when 6aa9707 is reapplied, and call it from CIFS.

In practice the likley result of holding the lock while freezing is that a second task blocked on the lock will never freeze, aborting suspend, but it is possible to manufacture a case using the cgroup freezer, the lock, and the suspend freezer to create a deadlock. Silencing the lockdep warning here will allow problems to be found in other drivers that may have a more serious deadlock risk, and prevent new problems from being added.

---
 include/linux/freezer.h | 13 +++++++++++++
 1 file changed, 13 insertions(+)
include/linux/freezer.h +10 -0
index 5b31e21c..d3c038e 100644
@@ -212,6 +212,16 @@ static inline bool freezer_should_skip(struct task_struct *p)
 	__retval;							\
 })
 
+/* DO NOT ADD ANY NEW CALLERS OF THIS FUNCTION */
+#define wait_event_freezekillable_unsafe(wq, condition)			\
+({									\
+	int __retval;							\
+	freezer_do_not_count();						\
+	__retval = wait_event_killable(wq, (condition));		\
+	freezer_count_unsafe();						\
+	__retval;							\
+})
+
 #define wait_event_freezable(wq, condition)				\
 ({									\
 	int __retval;							\

include/linux/freezer.h static inline void set_freezable(void) {} · line 295

Quoted 9 lines
include/linux/freezer.h +3 -0
@@ -277,6 +287,9 @@ static inline void set_freezable(void) {}
 #define wait_event_freezekillable(wq, condition)		\
 		wait_event_killable(wq, condition)
 
+#define wait_event_freezekillable_unsafe(wq, condition)			\
+		wait_event_killable(wq, condition)
+
 #endif /* !CONFIG_FREEZER */
 
 #endif	/* FREEZER_H_INCLUDED */

I think you also need to convert wait_for_response in the cifs code to use this helper. While it's a pretty straightforward change, you should probably cc [email protected] as well.

Signature · 1 line
Jeff Layton <[email protected]>
#4 lore raw

Re: [PATCH v3 02/16] freezer: add unsafe versions of freezable helpers for CIFS

Quoted 56 lines · Jeff Layton
Quoted 53 lines · On Mon, 6 May 2013 16:50:07 -0700 Colin Cross…

CIFS calls wait_event_freezekillable_unsafe with a VFS lock held, which is unsafe and will cause lockdep warnings when 6aa9707 "lockdep: check that no locks held at freeze time" is reapplied (it was reverted in dbf520a). CIFS shouldn't be doing this, but it has long-running syscalls that must hold a lock but also shouldn't block suspend. Until CIFS freeze handling is rewritten to use a signal to exit out of the critical section, add a new wait_event_freezekillable_unsafe helper that will not run the lockdep test when 6aa9707 is reapplied, and call it from CIFS.

In practice the likley result of holding the lock while freezing is that a second task blocked on the lock will never freeze, aborting suspend, but it is possible to manufacture a case using the cgroup freezer, the lock, and the suspend freezer to create a deadlock. Silencing the lockdep warning here will allow problems to be found in other drivers that may have a more serious deadlock risk, and prevent new problems from being added.

---
 include/linux/freezer.h | 13 +++++++++++++
 1 file changed, 13 insertions(+)
include/linux/freezer.h +13 -0
index 5b31e21c..d3c038e 100644
@@ -212,6 +212,16 @@ static inline bool freezer_should_skip(struct task_struct *p)
      __retval;                                                       \
 })
 
+/* DO NOT ADD ANY NEW CALLERS OF THIS FUNCTION */
+#define wait_event_freezekillable_unsafe(wq, condition)                      \
+({                                                                   \
+     int __retval;                                                   \
+     freezer_do_not_count();                                         \
+     __retval = wait_event_killable(wq, (condition));                \
+     freezer_count_unsafe();                                         \
+     __retval;                                                       \
+})
+
 #define wait_event_freezable(wq, condition)                          \
 ({                                                                   \
      int __retval;                                                   \
@@ -277,6 +287,9 @@ static inline void set_freezable(void) {}
 #define wait_event_freezekillable(wq, condition)             \
              wait_event_killable(wq, condition)
 
+#define wait_event_freezekillable_unsafe(wq, condition)                      \
+             wait_event_killable(wq, condition)
+
 #endif /* !CONFIG_FREEZER */
 
Quoted 8 lines

include/linux/freezer.h static inline void set_freezable(void) {} · line 295

Quoted 1 line
include/linux/freezer.h +0 -0
 #endif       /* FREEZER_H_INCLUDED */

I think you also need to convert wait_for_response in the cifs code to use this helper. While it's a pretty straightforward change, you should probably cc [email protected] as well.

--
Jeff Layton <[email protected]>

Oops, dropped a hunk which is why linux-cifs didn't get cc'd. I will resend it. --
To unsubscribe from this list: send the line "unsubscribe linux-nfs" in the body of a message to [email protected] More majordomo info at http://vger.kernel.org/majordomo-info.html

author Acked-by Pavel Machek
#5 lore raw

Re: [PATCH v4 02/16] freezer: add unsafe versions of freezable helpers for CIFS

CIFS calls wait_event_freezekillable_unsafe with a VFS lock held, which is unsafe and will cause lockdep warnings when 6aa9707 "lockdep: check that no locks held at freeze time" is reapplied (it was reverted in dbf520a). CIFS shouldn't be doing this, but it has long-running syscalls that must hold a lock but also shouldn't block suspend. Until CIFS freeze handling is rewritten to use a signal to exit out of the critical section, add a new wait_event_freezekillable_unsafe helper that will not run the lockdep test when 6aa9707 is reapplied, and call it from CIFS.

In practice the likley result of holding the lock while freezing is that a second task blocked on the lock will never freeze, aborting suspend, but it is possible to manufacture a case using the cgroup freezer, the lock, and the suspend freezer to create a deadlock. Silencing the lockdep warning here will allow problems to be found in other drivers that may have a more serious deadlock risk, and prevent new problems from being added.

---
v4:
Corrected to include CIFS wait_for_response hunk. The rest of this series is still at v3.

 fs/cifs/transport.c     |  2 +-
 include/linux/freezer.h | 13 +++++++++++++
 2 files changed, 14 insertions(+), 1 deletion(-)
fs/cifs/transport.c +1 -1
index 1a52868..e7f22f8 100644
@@ -452,7 +452,7 @@ wait_for_response(struct TCP_Server_Info *server, struct mid_q_entry *midQ)
 {
 	int error;
 
-	error = wait_event_freezekillable(server->response_q,
+	error = wait_event_freezekillable_unsafe(server->response_q,
 				    midQ->mid_state != MID_REQUEST_SUBMITTED);
 	if (error < 0)
 		return -ERESTARTSYS;
include/linux/freezer.h +13 -0
index 5b31e21c..d3c038e 100644
@@ -212,6 +212,16 @@ static inline bool freezer_should_skip(struct task_struct *p)
 	__retval;							\
 })
 
+/* DO NOT ADD ANY NEW CALLERS OF THIS FUNCTION */
+#define wait_event_freezekillable_unsafe(wq, condition)			\
+({									\
+	int __retval;							\
+	freezer_do_not_count();						\
+	__retval = wait_event_killable(wq, (condition));		\
+	freezer_count_unsafe();						\
+	__retval;							\
+})
+
 #define wait_event_freezable(wq, condition)				\
 ({									\
 	int __retval;							\
@@ -277,6 +287,9 @@ static inline void set_freezable(void) {}
 #define wait_event_freezekillable(wq, condition)		\
 		wait_event_killable(wq, condition)
 
+#define wait_event_freezekillable_unsafe(wq, condition)			\
+		wait_event_killable(wq, condition)
+
 #endif /* !CONFIG_FREEZER */
 
 #endif	/* FREEZER_H_INCLUDED */
Signature · 1 line
1.8.2.1
Reviewed-by
#6 lore raw

Re: [PATCH v4 02/16] freezer: add unsafe versions of freezable helpers for CIFS

Quoted 64 lines · On Tue, 7 May 2013 10:52:05 -0700 Colin Cross…

CIFS calls wait_event_freezekillable_unsafe with a VFS lock held, which is unsafe and will cause lockdep warnings when 6aa9707 "lockdep: check that no locks held at freeze time" is reapplied (it was reverted in dbf520a). CIFS shouldn't be doing this, but it has long-running syscalls that must hold a lock but also shouldn't block suspend. Until CIFS freeze handling is rewritten to use a signal to exit out of the critical section, add a new wait_event_freezekillable_unsafe helper that will not run the lockdep test when 6aa9707 is reapplied, and call it from CIFS.

In practice the likley result of holding the lock while freezing is that a second task blocked on the lock will never freeze, aborting suspend, but it is possible to manufacture a case using the cgroup freezer, the lock, and the suspend freezer to create a deadlock. Silencing the lockdep warning here will allow problems to be found in other drivers that may have a more serious deadlock risk, and prevent new problems from being added.

---
v4:
Corrected to include CIFS wait_for_response hunk. The rest of this series is still at v3.

 fs/cifs/transport.c     |  2 +-
 include/linux/freezer.h | 13 +++++++++++++
 2 files changed, 14 insertions(+), 1 deletion(-)
fs/cifs/transport.c +1 -1
index 1a52868..e7f22f8 100644
@@ -452,7 +452,7 @@ wait_for_response(struct TCP_Server_Info *server, struct mid_q_entry *midQ)
 {
 	int error;
 
-	error = wait_event_freezekillable(server->response_q,
+	error = wait_event_freezekillable_unsafe(server->response_q,
 				    midQ->mid_state != MID_REQUEST_SUBMITTED);
 	if (error < 0)
 		return -ERESTARTSYS;
include/linux/freezer.h +10 -0
index 5b31e21c..d3c038e 100644
@@ -212,6 +212,16 @@ static inline bool freezer_should_skip(struct task_struct *p)
 	__retval;							\
 })
 
+/* DO NOT ADD ANY NEW CALLERS OF THIS FUNCTION */
+#define wait_event_freezekillable_unsafe(wq, condition)			\
+({									\
+	int __retval;							\
+	freezer_do_not_count();						\
+	__retval = wait_event_killable(wq, (condition));		\
+	freezer_count_unsafe();						\
+	__retval;							\
+})
+
 #define wait_event_freezable(wq, condition)				\
 ({									\
 	int __retval;							\

include/linux/freezer.h static inline void set_freezable(void) {} · line 295

Quoted 9 lines
include/linux/freezer.h +3 -0
@@ -277,6 +287,9 @@ static inline void set_freezable(void) {}
 #define wait_event_freezekillable(wq, condition)		\
 		wait_event_killable(wq, condition)
 
+#define wait_event_freezekillable_unsafe(wq, condition)			\
+		wait_event_killable(wq, condition)
+
 #endif /* !CONFIG_FREEZER */
 
 #endif	/* FREEZER_H_INCLUDED */

Looks fine...

Acked-by on v3 2/16
#7 lore raw

Re: [PATCH v3 02/16] freezer: add unsafe versions of freezable helpers for CIFS

In reply to Colin Cross
Quoted 18 lines · Colin Cross

CIFS calls wait_event_freezekillable_unsafe with a VFS lock held, which is unsafe and will cause lockdep warnings when 6aa9707 "lockdep: check that no locks held at freeze time" is reapplied (it was reverted in dbf520a). CIFS shouldn't be doing this, but it has long-running syscalls that must hold a lock but also shouldn't block suspend. Until CIFS freeze handling is rewritten to use a signal to exit out of the critical section, add a new wait_event_freezekillable_unsafe helper that will not run the lockdep test when 6aa9707 is reapplied, and call it from CIFS.

In practice the likley result of holding the lock while freezing is that a second task blocked on the lock will never freeze, aborting suspend, but it is possible to manufacture a case using the cgroup freezer, the lock, and the suspend freezer to create a deadlock. Silencing the lockdep warning here will allow problems to be found in other drivers that may have a more serious deadlock risk, and prevent new problems from being added.

Quoted 1 line
Signature · 6 lines
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
--
To unsubscribe from this list: send the line "unsubscribe linux-nfs" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html
#8 lore raw

PATCH v3 3/16 lockdep: remove task argument from debug_check_no_locks_held

In reply to Colin Cross

The only existing caller to debug_check_no_locks_held calls it with 'current' as the task, and the freezer needs to call debug_check_no_locks_held but doesn't already have a current task pointer, so remove the argument. It is already assuming that the current task is relevant by dumping the current stack trace as part of the warning.

This was originally part of 6aa9707099c (lockdep: check that no locks held at freeze time) which was reverted in dbf520a9d7d4.

Original-author: Mandeep Singh Baines <[email protected]>

---
 include/linux/debug_locks.h |  4 ++--
 kernel/exit.c               |  2 +-
 kernel/lockdep.c            | 17 ++++++++---------
 3 files changed, 11 insertions(+), 12 deletions(-)
include/linux/debug_locks.h +2 -2
index 3bd46f7..a975de1 100644
@@ -51,7 +51,7 @@ struct task_struct;
 extern void debug_show_all_locks(void);
 extern void debug_show_held_locks(struct task_struct *task);
 extern void debug_check_no_locks_freed(const void *from, unsigned long len);
-extern void debug_check_no_locks_held(struct task_struct *task);
+extern void debug_check_no_locks_held(void);
 #else
 static inline void debug_show_all_locks(void)
 {
@@ -67,7 +67,7 @@ debug_check_no_locks_freed(const void *from, unsigned long len)
 }
 
 static inline void
-debug_check_no_locks_held(struct task_struct *task)
+debug_check_no_locks_held(void)
 {
 }
 #endif
kernel/exit.c +1 -1
index 60bc027..51e485c 100644
@@ -835,7 +835,7 @@ void do_exit(long code)
 	/*
 	 * Make sure we are holding no locks:
 	 */
-	debug_check_no_locks_held(tsk);
+	debug_check_no_locks_held();
 	/*
 	 * We can do this unlocked here. The futex code uses this flag
 	 * just to verify whether the pi state cleanup has been done
kernel/lockdep.c +8 -9
index 8a0efac..259db20 100644
@@ -4088,7 +4088,7 @@ void debug_check_no_locks_freed(const void *mem_from, unsigned long mem_len)
 }
 EXPORT_SYMBOL_GPL(debug_check_no_locks_freed);
 
-static void print_held_locks_bug(struct task_struct *curr)
+static void print_held_locks_bug(void)
 {
 	if (!debug_locks_off())
 		return;
@@ -4097,22 +4097,21 @@ static void print_held_locks_bug(struct task_struct *curr)
 
 	printk("\n");
 	printk("=====================================\n");
-	printk("[ BUG: lock held at task exit time! ]\n");
+	printk("[ BUG: %s/%d still has locks held! ]\n",
+	       current->comm, task_pid_nr(current));
 	print_kernel_ident();
 	printk("-------------------------------------\n");
-	printk("%s/%d is exiting with locks still held!\n",
-		curr->comm, task_pid_nr(curr));
-	lockdep_print_held_locks(curr);
-
+	lockdep_print_held_locks(current);
 	printk("\nstack backtrace:\n");
 	dump_stack();
 }
 
-void debug_check_no_locks_held(struct task_struct *task)
+void debug_check_no_locks_held(void)
 {
-	if (unlikely(task->lockdep_depth > 0))
-		print_held_locks_bug(task);
+	if (unlikely(current->lockdep_depth > 0))
+		print_held_locks_bug();
 }
+EXPORT_SYMBOL_GPL(debug_check_no_locks_held);
 
 void debug_show_all_locks(void)
 {
Signature · 1 line
1.8.2.1
Acked-by on v3 3/16
#9 lore raw

Re: [PATCH v3 03/16] lockdep: remove task argument from debug_check_no_locks_held

Quoted 7 lines · Colin Cross

The only existing caller to debug_check_no_locks_held calls it with 'current' as the task, and the freezer needs to call debug_check_no_locks_held but doesn't already have a current task pointer, so remove the argument. It is already assuming that the current task is relevant by dumping the current stack trace as part of the warning.

Quoted 6 lines

This was originally part of 6aa9707099c (lockdep: check that no locks held at freeze time) which was reverted in dbf520a9d7d4.

Original-author: Mandeep Singh Baines <[email protected]>

Signature · 2 lines
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
#10 lore raw

PATCH v3 4/16 lockdep: check that no locks held at freeze time

In reply to Colin Cross

From: Mandeep Singh Baines <[email protected]>

We shouldn't try_to_freeze if locks are held. Holding a lock can cause a deadlock if the lock is later acquired in the suspend or hibernate path (e.g. by dpm). Holding a lock can also cause a deadlock in the case of cgroup_freezer if a lock is held inside a frozen cgroup that is later acquired by a process outside that group.

History:
This patch was originally applied as 6aa9707099c and reverted in dbf520a9d7d4 because NFS was freezing with locks held. It was deemed better to keep the bad freeze point in NFS to allow laptops to suspend consistently. The previous patch in this series converts NFS to call _unsafe versions of the freezable helpers so that lockdep doesn't complain about them until a more correct fix can be applied.

[[email protected]: export debug_check_no_locks_held]
Cc Ben Chan <[email protected]>
Cc Oleg Nesterov <[email protected]>
Cc Tejun Heo <[email protected]>
Cc Rafael J. Wysocki <[email protected]>
Cc Ingo Molnar <[email protected]>
[[email protected]: don't warn if try_to_freeze_unsafe is called]
---
 include/linux/freezer.h | 3 +++
 1 file changed, 3 insertions(+)
include/linux/freezer.h +3 -0
index d3c038e..bcf9e65 100644
@@ -3,6 +3,7 @@
 #ifndef FREEZER_H_INCLUDED
 #define FREEZER_H_INCLUDED
 
+#include <linux/debug_locks.h>
 #include <linux/sched.h>
 #include <linux/wait.h>
 #include <linux/atomic.h>
@@ -60,6 +61,8 @@ static inline bool try_to_freeze_unsafe(void)
 
 static inline bool try_to_freeze(void)
 {
+	if (!(current->flags & PF_NOFREEZE))
+		debug_check_no_locks_held();
 	return try_to_freeze_unsafe();
 }
 
Signature · 1 line
1.8.2.1
Acked-by on v3 4/16
#11 lore raw

Re: [PATCH v3 04/16] lockdep: check that no locks held at freeze time

Quoted 17 lines · Colin Cross

From: Mandeep Singh Baines <[email protected]>

We shouldn't try_to_freeze if locks are held. Holding a lock can cause a deadlock if the lock is later acquired in the suspend or hibernate path (e.g. by dpm). Holding a lock can also cause a deadlock in the case of cgroup_freezer if a lock is held inside a frozen cgroup that is later acquired by a process outside that group.

History:
This patch was originally applied as 6aa9707099c and reverted in dbf520a9d7d4 because NFS was freezing with locks held. It was deemed better to keep the bad freeze point in NFS to allow laptops to suspend consistently. The previous patch in this series converts NFS to call _unsafe versions of the freezable helpers so that lockdep doesn't complain about them until a more correct fix can be applied.

Quoted 11 lines
[akpm-de/[email protected]: export debug_check_no_locks_held]
Cc Ben Chan <[email protected]>
Cc Oleg Nesterov <[email protected]>
Cc Tejun Heo <[email protected]>
Cc Rafael J. Wysocki <[email protected]>
Cc Ingo Molnar <[email protected]>
[[email protected]: don't warn if try_to_freeze_unsafe is called]
Signature · 6 lines
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
--
To unsubscribe from this list: send the line "unsubscribe linux-nfs" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html
#12 lore raw

PATCH v3 5/16 freezer: shorten freezer sleep time using exponential backoff

In reply to Colin Cross

All tasks can easily be frozen in under 10 ms, switch to using an initial 1 ms sleep followed by exponential backoff until 8 ms. Also convert the printed time to ms instead of centiseconds.

---
 kernel/power/process.c | 26 +++++++++++++++-----------
 1 file changed, 15 insertions(+), 11 deletions(-)
kernel/power/process.c +15 -11
index 98088e0..fc0df84 100644
@@ -30,9 +30,10 @@ static int try_to_freeze_tasks(bool user_only)
 	unsigned int todo;
 	bool wq_busy = false;
 	struct timeval start, end;
-	u64 elapsed_csecs64;
-	unsigned int elapsed_csecs;
+	u64 elapsed_msecs64;
+	unsigned int elapsed_msecs;
 	bool wakeup = false;
+	int sleep_usecs = USEC_PER_MSEC;
 
 	do_gettimeofday(&start);
 
@@ -68,22 +69,25 @@ static int try_to_freeze_tasks(bool user_only)
 
 		/*
 		 * We need to retry, but first give the freezing tasks some
-		 * time to enter the refrigerator.
+		 * time to enter the refrigerator.  Start with an initial
+		 * 1 ms sleep followed by exponential backoff until 8 ms.
 		 */
-		msleep(10);
+		usleep_range(sleep_usecs / 2, sleep_usecs);
+		if (sleep_usecs < 8 * USEC_PER_MSEC)
+			sleep_usecs *= 2;
 	}
 
 	do_gettimeofday(&end);
-	elapsed_csecs64 = timeval_to_ns(&end) - timeval_to_ns(&start);
-	do_div(elapsed_csecs64, NSEC_PER_SEC / 100);
-	elapsed_csecs = elapsed_csecs64;
+	elapsed_msecs64 = timeval_to_ns(&end) - timeval_to_ns(&start);
+	do_div(elapsed_msecs64, NSEC_PER_MSEC);
+	elapsed_msecs = elapsed_msecs64;
 
 	if (todo) {
 		printk("\n");
-		printk(KERN_ERR "Freezing of tasks %s after %d.%02d seconds "
+		printk(KERN_ERR "Freezing of tasks %s after %d.%03d seconds "
 		       "(%d tasks refusing to freeze, wq_busy=%d):\n",
 		       wakeup ? "aborted" : "failed",
-		       elapsed_csecs / 100, elapsed_csecs % 100,
+		       elapsed_msecs / 1000, elapsed_msecs % 1000,
 		       todo - wq_busy, wq_busy);
 
 		if (!wakeup) {
@@ -96,8 +100,8 @@ static int try_to_freeze_tasks(bool user_only)
 			read_unlock(&tasklist_lock);
 		}
 	} else {
-		printk("(elapsed %d.%02d seconds) ", elapsed_csecs / 100,
-			elapsed_csecs % 100);
+		printk("(elapsed %d.%03d seconds) ", elapsed_msecs / 1000,
+			elapsed_msecs % 1000);
 	}
 
 	return todo ? -EBUSY : 0;
Signature · 6 lines
1.8.2.1

--
To unsubscribe from this list: send the line "unsubscribe linux-nfs" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html
#13 lore raw

PATCH v3 6/16 freezer: skip waking up tasks with PF_FREEZER_SKIP set

In reply to Colin Cross

Android goes through suspend/resume very often (every few seconds when on a busy wifi network with the screen off), and a significant portion of the energy used to go in and out of suspend is spent in the freezer. If a task has called freezer_do_not_count(), don't bother waking it up. If it happens to wake up later it will call freezer_count() and immediately enter the refrigerator.

Combined with patches to convert freezable helpers to use freezer_do_not_count() and convert common sites where idle userspace tasks are blocked to use the freezable helpers, this reduces the time and energy required to suspend and resume.

---
v2: move check to freeze_task()

 kernel/freezer.c | 12 ++++++++++++
 1 file changed, 12 insertions(+)
kernel/freezer.c +12 -0
index c38893b..8b2afc1 100644
@@ -110,6 +110,18 @@ bool freeze_task(struct task_struct *p)
 {
 	unsigned long flags;
 
+	/*
+	 * This check can race with freezer_do_not_count, but worst case that
+	 * will result in an extra wakeup being sent to the task.  It does not
+	 * race with freezer_count(), the barriers in freezer_count() and
+	 * freezer_should_skip() ensure that either freezer_count() sees
+	 * freezing == true in try_to_freeze() and freezes, or
+	 * freezer_should_skip() sees !PF_FREEZE_SKIP and freezes the task
+	 * normally.
+	 */
+	if (freezer_should_skip(p))
+		return false;
+
 	spin_lock_irqsave(&freezer_lock, flags);
 	if (!freezing(p) || frozen(p)) {
 		spin_unlock_irqrestore(&freezer_lock, flags);
Signature · 6 lines
1.8.2.1

--
To unsubscribe from this list: send the line "unsubscribe linux-nfs" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html
#14 lore raw

PATCH v3 7/16 freezer: convert freezable helpers to freezer_do_not_count()

In reply to Colin Cross

Freezing tasks will wake up almost every userspace task from where it is blocking and force it to run until it hits a call to try_to_sleep(), generally on the exit path from the syscall it is blocking in. On resume each task will run again, usually restarting the syscall and running until it hits the same blocking call as it was originally blocked in.

Convert the existing wait_event_freezable* wrappers to use freezer_do_not_count(). Combined with a previous patch, these tasks will not run during suspend or resume unless they wake up for another reason, in which case they will run until they hit the try_to_freeze() in freezer_count(), and then continue processing the wakeup after tasks are thawed.

This results in a small change in behavior, previously a race between freezing and a normal wakeup would be won by the wakeup, now the task will freeze and then handle the wakeup after thawing.

---
v3:
split this out of the patch that adds new freezable helpers

 include/linux/freezer.h | 22 +++++++---------------
 1 file changed, 7 insertions(+), 15 deletions(-)
include/linux/freezer.h +7 -15
index bcf9e65..c71337af 100644
@@ -228,27 +228,19 @@ static inline bool freezer_should_skip(struct task_struct *p)
 #define wait_event_freezable(wq, condition)				\
 ({									\
 	int __retval;							\
-	for (;;) {							\
-		__retval = wait_event_interruptible(wq, 		\
-				(condition) || freezing(current));	\
-		if (__retval || (condition))				\
-			break;						\
-		try_to_freeze();					\
-	}								\
+	freezer_do_not_count();						\
+	__retval = wait_event_interruptible(wq, (condition));		\
+	freezer_count();						\
 	__retval;							\
 })
 
 #define wait_event_freezable_timeout(wq, condition, timeout)		\
 ({									\
 	long __retval = timeout;					\
-	for (;;) {							\
-		__retval = wait_event_interruptible_timeout(wq,		\
-				(condition) || freezing(current),	\
-				__retval); 				\
-		if (__retval <= 0 || (condition))			\
-			break;						\
-		try_to_freeze();					\
-	}								\
+	freezer_do_not_count();						\
+	__retval = wait_event_interruptible_timeout(wq,	(condition),	\
+				__retval);				\
+	freezer_count();						\
 	__retval;							\
 })
 
Signature · 1 line
1.8.2.1
#15 lore raw

PATCH v3 8/16 freezer: convert freezable helpers to static inline where possible

In reply to Colin Cross

Some of the freezable helpers have to be macros because their condition argument needs to get evaluated every time through the wait loop. Convert the others to static inline to make future changes easier.

---
 include/linux/freezer.h | 58 ++++++++++++++++++++++++-------------------------
 1 file changed, 29 insertions(+), 29 deletions(-)
include/linux/freezer.h +29 -29
index c71337af..8430d4c5 100644
@@ -159,46 +159,46 @@ static inline bool freezer_should_skip(struct task_struct *p)
 }
 
 /*
- * These macros are intended to be used whenever you want allow a sleeping
+ * These functions are intended to be used whenever you want allow a sleeping
  * task to be frozen. Note that neither return any clear indication of
  * whether a freeze event happened while in this function.
  */
 
 /* Like schedule(), but should not block the freezer. */
-#define freezable_schedule()						\
-({									\
-	freezer_do_not_count();						\
-	schedule();							\
-	freezer_count();						\
-})
+static inline void freezable_schedule(void)
+{
+	freezer_do_not_count();
+	schedule();
+	freezer_count();
+}
 
 /* DO NOT ADD ANY NEW CALLERS OF THIS FUNCTION */
-#define freezable_schedule_unsafe()					\
-({									\
-	freezer_do_not_count();						\
-	schedule();							\
-	freezer_count_unsafe();						\
-})
+static inline void freezable_schedule_unsafe(void)
+{
+	freezer_do_not_count();
+	schedule();
+	freezer_count_unsafe();
+}
 
 /* Like schedule_timeout_killable(), but should not block the freezer. */
-#define freezable_schedule_timeout_killable(timeout)			\
-({									\
-	long __retval;							\
-	freezer_do_not_count();						\
-	__retval = schedule_timeout_killable(timeout);			\
-	freezer_count();						\
-	__retval;							\
-})
+static inline long freezable_schedule_timeout_killable(long timeout)
+{
+	long __retval;
+	freezer_do_not_count();
+	__retval = schedule_timeout_killable(timeout);
+	freezer_count();
+	return __retval;
+}
 
 /* DO NOT ADD ANY NEW CALLERS OF THIS FUNCTION */
-#define freezable_schedule_timeout_killable_unsafe(timeout)		\
-({									\
-	long __retval;							\
-	freezer_do_not_count();						\
-	__retval = schedule_timeout_killable(timeout);			\
-	freezer_count_unsafe();						\
-	__retval;							\
-})
+static inline long freezable_schedule_timeout_killable_unsafe(long timeout)
+{
+	long __retval;
+	freezer_do_not_count();
+	__retval = schedule_timeout_killable(timeout);
+	freezer_count_unsafe();
+	return __retval;
+}
 
 /*
  * Freezer-friendly wrappers around wait_event_interruptible(),
Signature · 1 line
1.8.2.1
#16 lore raw

PATCH v3 9/16 freezer: add new freezable helpers using freezer_do_not_count()

In reply to Colin Cross

Freezing tasks will wake up almost every userspace task from where it is blocking and force it to run until it hits a call to try_to_sleep(), generally on the exit path from the syscall it is blocking in. On resume each task will run again, usually restarting the syscall and running until it hits the same blocking call as it was originally blocked in.

To allow tasks to avoid running on every suspend/resume cycle, this patch adds additional freezable wrappers around blocking calls that call freezer_do_not_count(). Combined with the previous patch, these tasks will not run during suspend or resume unless they wake up for another reason, in which case they will run until they hit the try_to_freeze() in freezer_count(), and then continue processing the wakeup after tasks are thawed.

Additional patches will convert the most common locations that userspace blocks in to use freezable helpers.

---
v3:
split out the changes to existing helpers to a separate patch

 include/linux/freezer.h | 61 +++++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 61 insertions(+)
include/linux/freezer.h +61 -0
index 8430d4c5..7fd81b8 100644
@@ -180,6 +180,32 @@ static inline void freezable_schedule_unsafe(void)
 	freezer_count_unsafe();
 }
 
+/*
+ * Like freezable_schedule_timeout(), but should not block the freezer.  Do not
+ * call this with locks held.
+ */
+static inline long freezable_schedule_timeout(long timeout)
+{
+	long __retval;
+	freezer_do_not_count();
+	__retval = schedule_timeout(timeout);
+	freezer_count();
+	return __retval;
+}
+
+/*
+ * Like schedule_timeout_interruptible(), but should not block the freezer.  Do not
+ * call this with locks held.
+ */
+static inline long freezable_schedule_timeout_interruptible(long timeout)
+{
+	long __retval;
+	freezer_do_not_count();
+	__retval = schedule_timeout_interruptible(timeout);
+	freezer_count();
+	return __retval;
+}
+
 /* Like schedule_timeout_killable(), but should not block the freezer. */
 static inline long freezable_schedule_timeout_killable(long timeout)
 {
@@ -201,6 +227,20 @@ static inline long freezable_schedule_timeout_killable_unsafe(long timeout)
 }
 
 /*
+ * Like schedule_hrtimeout_range(), but should not block the freezer.  Do not
+ * call this with locks held.
+ */
+static inline int freezable_schedule_hrtimeout_range(ktime_t *expires,
+		unsigned long delta, const enum hrtimer_mode mode)
+{
+	int __retval;
+	freezer_do_not_count();
+	__retval = schedule_hrtimeout_range(expires, delta, mode);
+	freezer_count();
+	return __retval;
+}
+
+/*
  * Freezer-friendly wrappers around wait_event_interruptible(),
  * wait_event_killable() and wait_event_interruptible_timeout(), originally
  * defined in <linux/wait.h>
@@ -244,6 +284,16 @@ static inline long freezable_schedule_timeout_killable_unsafe(long timeout)
 	__retval;							\
 })
 
+#define wait_event_freezable_exclusive(wq, condition)			\
+({									\
+	int __retval;							\
+	freezer_do_not_count();						\
+	__retval = wait_event_interruptible_exclusive(wq, condition);	\
+	freezer_count();						\
+	__retval;							\
+})
+
+
 #else /* !CONFIG_FREEZER */
 static inline bool frozen(struct task_struct *p) { return false; }
 static inline bool freezing(struct task_struct *p) { return false; }
@@ -267,18 +317,29 @@ static inline void set_freezable(void) {}
 
 #define freezable_schedule_unsafe()  schedule()
 
+#define freezable_schedule_timeout(timeout)  schedule_timeout(timeout)
+
+#define freezable_schedule_timeout_interruptible(timeout)		\
+	schedule_timeout_interruptible(timeout)
+
 #define freezable_schedule_timeout_killable(timeout)			\
 	schedule_timeout_killable(timeout)
 
 #define freezable_schedule_timeout_killable_unsafe(timeout)		\
 	schedule_timeout_killable(timeout)
 
+#define freezable_schedule_hrtimeout_range(expires, delta, mode)	\
+	schedule_hrtimeout_range(expires, delta, mode)
+
 #define wait_event_freezable(wq, condition)				\
 		wait_event_interruptible(wq, condition)
 
 #define wait_event_freezable_timeout(wq, condition, timeout)		\
 		wait_event_interruptible_timeout(wq, condition, timeout)
 
+#define wait_event_freezable_exclusive(wq, condition)			\
+		wait_event_interruptible_exclusive(wq, condition)
+
 #define wait_event_freezekillable(wq, condition)		\
 		wait_event_killable(wq, condition)
 
Signature · 1 line
1.8.2.1
#17 lore raw

PATCH v3 10/16 binder: use freezable blocking calls

In reply to Colin Cross

Avoid waking up every thread sleeping in a binder call during suspend and resume by calling a freezable blocking call. Previous patches modified the freezer to avoid sending wakeups to threads that are blocked in freezable blocking calls.

This call was selected to be converted to a freezable call because it doesn't hold any locks or release any resources when interrupted that might be needed by another freezing task or a kernel driver during suspend, and is a common site where idle userspace tasks are blocked.

---
 drivers/staging/android/binder.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)
drivers/staging/android/binder.c +3 -2
index 24456a0..af8fba4 100644
@@ -20,6 +20,7 @@
 #include <asm/cacheflush.h>
 #include <linux/fdtable.h>
 #include <linux/file.h>
+#include <linux/freezer.h>
 #include <linux/fs.h>
 #include <linux/list.h>
 #include <linux/miscdevice.h>
@@ -2140,13 +2141,13 @@ retry:
 			if (!binder_has_proc_work(proc, thread))
 				ret = -EAGAIN;
 		} else
-			ret = wait_event_interruptible_exclusive(proc->wait, binder_has_proc_work(proc, thread));
+			ret = wait_event_freezable_exclusive(proc->wait, binder_has_proc_work(proc, thread));
 	} else {
 		if (non_block) {
 			if (!binder_has_thread_work(thread))
 				ret = -EAGAIN;
 		} else
-			ret = wait_event_interruptible(thread->wait, binder_has_thread_work(thread));
+			ret = wait_event_freezable(thread->wait, binder_has_thread_work(thread));
 	}
 
 	binder_lock(__func__);
Signature · 1 line
1.8.2.1
#18 lore raw

PATCH v3 11/16 epoll: use freezable blocking call

In reply to Colin Cross

Avoid waking up every thread sleeping in an epoll_wait call during suspend and resume by calling a freezable blocking call. Previous patches modified the freezer to avoid sending wakeups to threads that are blocked in freezable blocking calls.

This call was selected to be converted to a freezable call because it doesn't hold any locks or release any resources when interrupted that might be needed by another freezing task or a kernel driver during suspend, and is a common site where idle userspace tasks are blocked.

---
 fs/eventpoll.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)
fs/eventpoll.c +3 -1
index 9fec183..65245e7 100644
@@ -34,6 +34,7 @@
 #include <linux/mutex.h>
 #include <linux/anon_inodes.h>
 #include <linux/device.h>
+#include <linux/freezer.h>
 #include <asm/uaccess.h>
 #include <asm/io.h>
 #include <asm/mman.h>
@@ -1543,7 +1544,8 @@ fetch_events:
 			}
 
 			spin_unlock_irqrestore(&ep->lock, flags);
-			if (!schedule_hrtimeout_range(to, slack, HRTIMER_MODE_ABS))
+			if (!freezable_schedule_hrtimeout_range(to, slack,
+								HRTIMER_MODE_ABS))
 				timed_out = 1;
 
 			spin_lock_irqsave(&ep->lock, flags);
Signature · 1 line
1.8.2.1
#19 lore raw

PATCH v3 12/16 select: use freezable blocking call

In reply to Colin Cross

Avoid waking up every thread sleeping in a select call during suspend and resume by calling a freezable blocking call. Previous patches modified the freezer to avoid sending wakeups to threads that are blocked in freezable blocking calls.

This call was selected to be converted to a freezable call because it doesn't hold any locks or release any resources when interrupted that might be needed by another freezing task or a kernel driver during suspend, and is a common site where idle userspace tasks are blocked.

---
 fs/select.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)
fs/select.c +3 -1
index 8c1c96c..6b14dc7 100644
@@ -27,6 +27,7 @@
 #include <linux/rcupdate.h>
 #include <linux/hrtimer.h>
 #include <linux/sched/rt.h>
+#include <linux/freezer.h>
 
 #include <asm/uaccess.h>
 
@@ -236,7 +237,8 @@ int poll_schedule_timeout(struct poll_wqueues *pwq, int state,
 
 	set_current_state(state);
 	if (!pwq->triggered)
-		rc = schedule_hrtimeout_range(expires, slack, HRTIMER_MODE_ABS);
+		rc = freezable_schedule_hrtimeout_range(expires, slack,
+							HRTIMER_MODE_ABS);
 	__set_current_state(TASK_RUNNING);
 
 	/*
Signature · 1 line
1.8.2.1
#20 lore raw

PATCH v3 13/16 futex: use freezable blocking call

In reply to Colin Cross

Avoid waking up every thread sleeping in a futex_wait call during suspend and resume by calling a freezable blocking call. Previous patches modified the freezer to avoid sending wakeups to threads that are blocked in freezable blocking calls.

This call was selected to be converted to a freezable call because it doesn't hold any locks or release any resources when interrupted that might be needed by another freezing task or a kernel driver during suspend, and is a common site where idle userspace tasks are blocked.

---
 kernel/futex.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
kernel/futex.c +2 -1
index b26dcfc..d710fae 100644
@@ -61,6 +61,7 @@
 #include <linux/nsproxy.h>
 #include <linux/ptrace.h>
 #include <linux/sched/rt.h>
+#include <linux/freezer.h>
 
 #include <asm/futex.h>
 
@@ -1807,7 +1808,7 @@ static void futex_wait_queue_me(struct futex_hash_bucket *hb, struct futex_q *q,
 		 * is no timeout, or if it has yet to expire.
 		 */
 		if (!timeout || timeout->task)
-			schedule();
+			freezable_schedule();
 	}
 	__set_current_state(TASK_RUNNING);
 }
Signature · 1 line
1.8.2.1
#21 lore raw

3.11-rc regression bisected: s2disk does not work (was Re: [PATCH v3 13/16] futex: use freezable blocking call)

Quoted 1 line · On Mon, 6 May 2013 16:50:18 -0700 Colin Cross <[email protected]> wrote:

Avoid waking up every thread sleeping in a futex_wait call during

[...]

With 3.11-rc s2disk from suspend-utils stopped working: Frozen at displaying 0% of saving image to disk.

echo "1" >/sys/power/state still works.

Bisecting yielded 88c8004fd3a5fdd2378069de86b90b21110d33a4, reverting that from 3.11-rc2 makes s2disk working again.

Signature · 3 lines
MfG,

Michael Leun
#22 lore raw

Re: 3.11-rc regression bisected: s2disk does not work (was Re: [PATCH v3 13/16] futex: use freezable blocking call)

Quoted 6 lines · Michael Leun
Quoted 1 line · On Mon, 6 May 2013 16:50:18 -0700 Colin Cross <[email protected]> wrote:

Avoid waking up every thread sleeping in a futex_wait call during

[...]

Quoted 8 lines

With 3.11-rc s2disk from suspend-utils stopped working: Frozen at displaying 0% of saving image to disk.

echo "1" >/sys/power/state still works.

Bisecting yielded 88c8004fd3a5fdd2378069de86b90b21110d33a4, reverting that from 3.11-rc2 makes s2disk working again.

I think the expanded use of the freezable_* helpers is exposing an existing bug in hibernation. The SNAPSHOT_FREEZE ioctl calls freeze_processes(), which sets the global system_freezing_cnt and pm_freezing. try_to_freeze_tasks then sends every process except current a signal which causes them all to end up in the refrigerator. The current task then returns back to userspace and continues its work to suspend to disk. If that task ever hits a call to try_to_freeze() in the kernel, it will see system_freezing_cnt and pm_freezing=true and freeze, and suspend to disk will hang forever. It could hit try_to_freeze() because of a signal delivered to the task, or from calling any syscall that uses a freezable_* helper like the one I added to sys_futex.

I think the right solution is to add a flag to the freezing task that marks it unfreezable. I think PF_NOFREEZE would work, although it is normally used on kernel threads, can you see if the attached patch helps?

From 0f22f2b357b06208fc7c0b82ce3f0929d00877ca Mon Sep 17 00:00:00 2001
From: Colin Cross <[email protected]>
Date: Mon, 22 Jul 2013 16:53:15 -0700
Subject: [PATCH] power: set PF_NOFREEZE flag on SNAPSHOT_FREEZE task

The task that calls the SNAPSHOT_FREEZE ioctl needs to return back to userspace and continue preparing a suspend-to-disk image. Set the PF_NOFREEZE flag on it so that it doesn't accidentally freeze if it comes across a call to try_to_freeze().

---
 kernel/power/user.c | 2 ++
 1 file changed, 2 insertions(+)
kernel/power/user.c +2 -0
index 4ed81e7..17f9c20 100644
@@ -219,6 +219,7 @@ static long snapshot_ioctl(struct file *filp, unsigned int cmd,
 		sys_sync();
 		printk("done.\n");
 
+		current->flags |= PF_NOFREEZE;
 		error = freeze_processes();
 		if (!error)
 			data->frozen = 1;
@@ -229,6 +230,7 @@ static long snapshot_ioctl(struct file *filp, unsigned int cmd,
 			break;
 		pm_restore_gfp_mask();
 		thaw_processes();
+		current->flags &= ~PF_NOFREEZE;
 		data->frozen = 0;
 		break;
 
Signature · 1 line
1.8.3
#23 lore raw

Re: 3.11-rc regression bisected: s2disk does not work (was Re: [PATCH v3 13/16] futex: use freezable blocking call)

Quoted 5 lines · Colin Cross

I think the right solution is to add a flag to the freezing task that marks it unfreezable. I think PF_NOFREEZE would work, although it is normally used on kernel threads, can you see if the attached patch helps?

Hmm. That does seem to be the right thing to do, but I wonder about the *other* callers of freeze_processes() IOW, kexec and friends.

So maybe we should do this in {freeze|thaw}_processes() itself, and just make the rule be that the caller of freeze_processes() itself is obviously not frozen, and has to be the same one that then thaws things?

Colin? Rafael? Comments?

Linus

#24 lore raw

Re: 3.11-rc regression bisected: s2disk does not work (was Re: [PATCH v3 13/16] futex: use freezable blocking call)

Quoted 10 lines · Linus Torvalds
Quoted 5 lines · Colin Cross

I think the right solution is to add a flag to the freezing task that marks it unfreezable. I think PF_NOFREEZE would work, although it is normally used on kernel threads, can you see if the attached patch helps?

Hmm. That does seem to be the right thing to do, but I wonder about the *other* callers of freeze_processes() IOW, kexec and friends.

Quoted 8 lines

So maybe we should do this in {freeze|thaw}_processes() itself, and just make the rule be that the caller of freeze_processes() itself is obviously not frozen, and has to be the same one that then thaws things?

Colin? Rafael? Comments?

Linus

I was worried about clearing the flag in thaw_processes(). If a kernel thread with PF_NOFREEZE set ever called thaw_processes(), which autosleep might do, it would clear the flag. Or if a different thread called freeze_processes() and thaw_processes(). All the other callers besides the SNAPSHOT_FREEZE ioctl stay in the kernel between freeze_processes() and thaw_processes(), which makes the fanout of places that could call try_to_freeze() much more controllable.

Using a new flag that operates like PF_NOFREEZE but doesn't conflict with it, or a nofreeze_depth counter, would also work.
--
To unsubscribe from this list: send the line "unsubscribe linux-nfs" in the body of a message to [email protected] More majordomo info at http://vger.kernel.org/majordomo-info.html

#25 lore raw

Re: 3.11-rc regression bisected: s2disk does not work (was Re: [PATCH v3 13/16] futex: use freezable blocking call)

Quoted 17 lines · Colin Cross
Quoted 15 lines · Linus Torvalds
Quoted 5 lines · Colin Cross

I think the right solution is to add a flag to the freezing task that marks it unfreezable. I think PF_NOFREEZE would work, although it is normally used on kernel threads, can you see if the attached patch helps?

Hmm. That does seem to be the right thing to do, but I wonder about the *other* callers of freeze_processes() IOW, kexec and friends.

So maybe we should do this in {freeze|thaw}_processes() itself, and just make the rule be that the caller of freeze_processes() itself is obviously not frozen, and has to be the same one that then thaws things?

Quoted 8 lines
Quoted 3 lines

Colin? Rafael? Comments?

Linus

I was worried about clearing the flag in thaw_processes(). If a kernel thread with PF_NOFREEZE set ever called thaw_processes(), which autosleep might do, it would clear the flag. Or if a different thread called freeze_processes() and thaw_processes().

Is that legitimate?

Quoted 6 lines

All the other callers besides the SNAPSHOT_FREEZE ioctl stay in the kernel between freeze_processes() and thaw_processes(), which makes the fanout of places that could call try_to_freeze() much more controllable.

Using a new flag that operates like PF_NOFREEZE but doesn't conflict with it, or a nofreeze_depth counter, would also work.

Well, that would be robust enough. At least if the purpose of that new flag is clearly specified, people hopefully won't be tempted to optimize it away in the future.

Thanks,
Rafael

Signature · 2 lines
I speak only for myself.
Rafael J. Wysocki, Intel Open Source Technology Center.
#26 lore raw

Re: 3.11-rc regression bisected: s2disk does not work (was Re: [PATCH v3 13/16] futex: use freezable blocking call)

Quoted 20 lines · Rafael J. Wysocki
Quoted 19 lines · Colin Cross
Quoted 17 lines · Linus Torvalds
Quoted 5 lines · Colin Cross

I think the right solution is to add a flag to the freezing task that marks it unfreezable. I think PF_NOFREEZE would work, although it is normally used on kernel threads, can you see if the attached patch helps?

Hmm. That does seem to be the right thing to do, but I wonder about the *other* callers of freeze_processes() IOW, kexec and friends.

So maybe we should do this in {freeze|thaw}_processes() itself, and just make the rule be that the caller of freeze_processes() itself is obviously not frozen, and has to be the same one that then thaws things?

Colin? Rafael? Comments?

Quoted 8 lines
Quoted 6 lines
Quoted 1 line

Linus

I was worried about clearing the flag in thaw_processes(). If a kernel thread with PF_NOFREEZE set ever called thaw_processes(), which autosleep might do, it would clear the flag. Or if a different thread called freeze_processes() and thaw_processes().

Is that legitimate?

Nothing precludes it today, but I don't see any need for it. I'll add a comment when I add the flag.

Quoted 7 lines
Quoted 6 lines

All the other callers besides the SNAPSHOT_FREEZE ioctl stay in the kernel between freeze_processes() and thaw_processes(), which makes the fanout of places that could call try_to_freeze() much more controllable.

Using a new flag that operates like PF_NOFREEZE but doesn't conflict with it, or a nofreeze_depth counter, would also work.

Quoted 6 lines

Well, that would be robust enough. At least if the purpose of that new flag is clearly specified, people hopefully won't be tempted to optimize it away in the future.

Thanks,
Rafael

OK, I'll add a new flag.
--
To unsubscribe from this list: send the line "unsubscribe linux-nfs" in the body of a message to [email protected] More majordomo info at http://vger.kernel.org/majordomo-info.html

#27 lore raw

Re: 3.11-rc regression bisected: s2disk does not work (was Re: [PATCH v3 13/16] futex: use freezable blocking call)

Quoted 40 lines · Colin Cross
Quoted 20 lines · Rafael J. Wysocki
Quoted 19 lines · Colin Cross
Quoted 17 lines · Linus Torvalds
Quoted 5 lines · Colin Cross

I think the right solution is to add a flag to the freezing task that marks it unfreezable. I think PF_NOFREEZE would work, although it is normally used on kernel threads, can you see if the attached patch helps?

Hmm. That does seem to be the right thing to do, but I wonder about the *other* callers of freeze_processes() IOW, kexec and friends.

So maybe we should do this in {freeze|thaw}_processes() itself, and just make the rule be that the caller of freeze_processes() itself is obviously not frozen, and has to be the same one that then thaws things?

Colin? Rafael? Comments?

Quoted 8 lines
Quoted 6 lines
Quoted 1 line

Linus

I was worried about clearing the flag in thaw_processes(). If a kernel thread with PF_NOFREEZE set ever called thaw_processes(), which autosleep might do, it would clear the flag. Or if a different thread called freeze_processes() and thaw_processes().

Is that legitimate?

Nothing precludes it today, but I don't see any need for it. I'll add a comment when I add the flag.

Quoted 7 lines
Quoted 6 lines

All the other callers besides the SNAPSHOT_FREEZE ioctl stay in the kernel between freeze_processes() and thaw_processes(), which makes the fanout of places that could call try_to_freeze() much more controllable.

Using a new flag that operates like PF_NOFREEZE but doesn't conflict with it, or a nofreeze_depth counter, would also work.

Quoted 8 lines
Quoted 6 lines

Well, that would be robust enough. At least if the purpose of that new flag is clearly specified, people hopefully won't be tempted to optimize it away in the future.

Thanks,
Rafael

OK, I'll add a new flag.

Michael, can you see if this patch works and doesn't throw any warnings during suspend or resume?

If the extra process flag is considered too precious for this (there are only 2 left after this patch) I could get the same functionality by having freeze_processes() reject calls from a PF_KTHREAD|PF_NOFREEZE thread, and use PF_KTHREAD to determine if PF_NOFREEZE should be cleared in thaw_processes().

From 188f58711ad19315a1abaa05000db9d312aaa93f Mon Sep 17 00:00:00 2001
From: Colin Cross <[email protected]>
Date: Mon, 22 Jul 2013 16:53:15 -0700
Subject: [PATCH] power: set PF_SUSPEND_TASK flag on tasks that call
freeze_processes

Calling freeze_processes sets a global flag that will cause any process that calls try_to_freeze to enter the refrigerator. It skips sending a signal to the current task, but if the current task ever hits try_to_freeze all threads will be frozen and the system will deadlock.

Set a new flag, PF_SUSPEND_TASK, on the task that calls freeze_processes. The flag notifies the freezer that the thread is involved in suspend and should not be frozen. Also add a WARN_ON in thaw_processes if the caller does not have the PF_SUSPEND_TASK flag set to catch if a different task calls thaw_processes than the one that called freeze_processes, leaving a task with PF_SUSPEND_TASK permanently set on it.

Threads that spawn off a task with PF_SUSPEND_TASK set (which swsusp does) will also have PF_SUSPEND_TASK set, preventing them from freezing while they are helping with suspend, but they need to be dead by the time suspend is triggered, otherwise they may run when userspace is expected to be frozen. Add a WARN_ON in thaw_processes if more than one thread has the PF_SUSPEND_TASK flag set.

---
 include/linux/sched.h  |  1 +
 kernel/freezer.c       |  2 +-
 kernel/power/process.c | 11 +++++++++++
 3 files changed, 13 insertions(+), 1 deletion(-)
include/linux/sched.h +1 -0
index 50d04b9..d722490 100644
@@ -1628,6 +1628,7 @@ extern void thread_group_cputime_adjusted(struct task_struct *p, cputime_t *ut,
 #define PF_MEMPOLICY	0x10000000	/* Non-default NUMA mempolicy */
 #define PF_MUTEX_TESTER	0x20000000	/* Thread belongs to the rt mutex tester */
 #define PF_FREEZER_SKIP	0x40000000	/* Freezer should not count it as freezable */
+#define PF_SUSPEND_TASK 0x80000000      /* this thread called freeze_processes and should not be frozen */
 
 /*
  * Only the _current_ task can read/write to tsk->flags, but other
kernel/freezer.c +1 -1
index 8b2afc1..b462fa1 100644
@@ -33,7 +33,7 @@ static DEFINE_SPINLOCK(freezer_lock);
  */
 bool freezing_slow_path(struct task_struct *p)
 {
-	if (p->flags & PF_NOFREEZE)
+	if (p->flags & (PF_NOFREEZE | PF_SUSPEND_TASK))
 		return false;
 
 	if (pm_nosig_freezing || cgroup_freezing(p))
kernel/power/process.c +11 -0
index fc0df84..06ec886 100644
@@ -109,6 +109,8 @@ static int try_to_freeze_tasks(bool user_only)
 
 /**
  * freeze_processes - Signal user space processes to enter the refrigerator.
+ * The current thread will not be frozen.  The same process that calls
+ * freeze_processes must later call thaw_processes.
  *
  * On success, returns 0.  On failure, -errno and system is fully thawed.
  */
@@ -120,6 +122,9 @@ int freeze_processes(void)
 	if (error)
 		return error;
 
+	/* Make sure this task doesn't get frozen */
+	current->flags |= PF_SUSPEND_TASK;
+
 	if (!pm_freezing)
 		atomic_inc(&system_freezing_cnt);
 
@@ -168,6 +173,7 @@ int freeze_kernel_threads(void)
 void thaw_processes(void)
 {
 	struct task_struct *g, *p;
+	struct task_struct *curr = current;
 
 	if (pm_freezing)
 		atomic_dec(&system_freezing_cnt);
@@ -182,10 +188,15 @@ void thaw_processes(void)
 
 	read_lock(&tasklist_lock);
 	do_each_thread(g, p) {
+		/* No other threads should have PF_SUSPEND_TASK set */
+		WARN_ON((p != curr) && (p->flags & PF_SUSPEND_TASK));
 		__thaw_task(p);
 	} while_each_thread(g, p);
 	read_unlock(&tasklist_lock);
 
+	WARN_ON(!(curr->flags & PF_SUSPEND_TASK));
+	curr->flags &= ~PF_SUSPEND_TASK;
+
 	usermodehelper_enable();
 
 	schedule();
Signature · 1 line
1.8.3
#28 lore raw

Re: 3.11-rc regression bisected: s2disk does not work (was Re: [PATCH v3 13/16] futex: use freezable blocking call)

Quoted 50 lines · On Tue, 23 Jul 2013 13:31:49 -0700 Colin Cross <[email protected]> wrote:
Quoted 48 lines · Colin Cross
Quoted 22 lines · Rafael J. Wysocki
Quoted 21 lines · Colin Cross
Quoted 19 lines · Linus Torvalds
Quoted 5 lines · Colin Cross

I think the right solution is to add a flag to the freezing task that marks it unfreezable. I think PF_NOFREEZE would work, although it is normally used on kernel threads, can you see if the attached patch helps?

Hmm. That does seem to be the right thing to do, but I wonder about the *other* callers of freeze_processes() IOW, kexec and friends.

So maybe we should do this in {freeze|thaw}_processes() itself, and just make the rule be that the caller of freeze_processes() itself is obviously not frozen, and has to be the same one that then thaws things?

Colin? Rafael? Comments?

Quoted 8 lines
Quoted 6 lines
Quoted 1 line

Linus

I was worried about clearing the flag in thaw_processes(). If a kernel thread with PF_NOFREEZE set ever called thaw_processes(), which autosleep might do, it would clear the flag. Or if a different thread called freeze_processes() and thaw_processes().

Is that legitimate?

Nothing precludes it today, but I don't see any need for it. I'll add a comment when I add the flag.

Quoted 12 lines
Quoted 7 lines

All the other callers besides the SNAPSHOT_FREEZE ioctl stay in the kernel between freeze_processes() and thaw_processes(), which makes the fanout of places that could call try_to_freeze() much more controllable.

Using a new flag that operates like PF_NOFREEZE but doesn't conflict with it, or a nofreeze_depth counter, would also work.

Well, that would be robust enough. At least if the purpose of that new flag is clearly specified, people hopefully won't be tempted to optimize it away in the future.

Quoted 8 lines
Quoted 4 lines
Quoted 2 lines

Thanks,
Rafael

OK, I'll add a new flag.

Michael, can you see if this patch works and doesn't throw any warnings during suspend or resume?

Tried several times with and without threads = y in suspend.conf, tried also to produce high load / much processes / high memory usage.

Worked every time, no WARN seen.

Quoted 5 lines

If the extra process flag is considered too precious for this (there are only 2 left after this patch) I could get the same functionality by having freeze_processes() reject calls from a PF_KTHREAD|PF_NOFREEZE thread, and use PF_KTHREAD to determine if PF_NOFREEZE should be cleared in thaw_processes().

If another solution is considered please do not hesitate to send me the patch for another round of check.

Signature · 3 lines
MfG,

Michael Leun
#29 lore raw

Re: 3.11-rc regression bisected: s2disk does not work (was Re: [PATCH v3 13/16] futex: use freezable blocking call)

In reply to Colin Cross
Quoted 30 lines · On Mon, 22 Jul 2013 16:55:58 -0700 Colin Cross <[email protected]> wrote:
Quoted 6 lines · Michael Leun
Quoted 1 line · On Mon, 6 May 2013 16:50:18 -0700 Colin Cross <[email protected]> wrote:

Avoid waking up every thread sleeping in a futex_wait call during

[...]

Quoted 8 lines

With 3.11-rc s2disk from suspend-utils stopped working: Frozen at displaying 0% of saving image to disk.

echo "1" >/sys/power/state still works.

Bisecting yielded 88c8004fd3a5fdd2378069de86b90b21110d33a4, reverting that from 3.11-rc2 makes s2disk working again.

I think the expanded use of the freezable_* helpers is exposing an existing bug in hibernation. The SNAPSHOT_FREEZE ioctl calls freeze_processes(), which sets the global system_freezing_cnt and pm_freezing. try_to_freeze_tasks then sends every process except current a signal which causes them all to end up in the refrigerator. The current task then returns back to userspace and continues its work to suspend to disk. If that task ever hits a call to try_to_freeze() in the kernel, it will see system_freezing_cnt and pm_freezing=true and freeze, and suspend to disk will hang forever. It could hit try_to_freeze() because of a signal delivered to the task, or from calling any syscall that uses a freezable_* helper like the one I added to sys_futex.

Quoted 4 lines

I think the right solution is to add a flag to the freezing task that marks it unfreezable. I think PF_NOFREEZE would work, although it is normally used on kernel threads, can you see if the attached patch helps?

That patch helps.

BTW, the only machine I can reproduce this bug with is an i7-3630QM notebook. Cannot reproduce on an Core Duo U1400 and cannot reproduce on an i7 M 620.

Are the sysreq backtraces still wanted? If so, any tip, how I could get them saved?

Signature · 3 lines
MfG,

Michael Leun
#30 lore raw

Re: 3.11-rc regression bisected: s2disk does not work (was Re: [PATCH v3 13/16] futex: use freezable blocking call)

Quoted 38 lines · Michael Leun
Quoted 30 lines · On Mon, 22 Jul 2013 16:55:58 -0700 Colin Cross <[email protected]> wrote:
Quoted 6 lines · Michael Leun
Quoted 1 line · On Mon, 6 May 2013 16:50:18 -0700 Colin Cross <[email protected]> wrote:

Avoid waking up every thread sleeping in a futex_wait call during

[...]

Quoted 8 lines

With 3.11-rc s2disk from suspend-utils stopped working: Frozen at displaying 0% of saving image to disk.

echo "1" >/sys/power/state still works.

Bisecting yielded 88c8004fd3a5fdd2378069de86b90b21110d33a4, reverting that from 3.11-rc2 makes s2disk working again.

I think the expanded use of the freezable_* helpers is exposing an existing bug in hibernation. The SNAPSHOT_FREEZE ioctl calls freeze_processes(), which sets the global system_freezing_cnt and pm_freezing. try_to_freeze_tasks then sends every process except current a signal which causes them all to end up in the refrigerator. The current task then returns back to userspace and continues its work to suspend to disk. If that task ever hits a call to try_to_freeze() in the kernel, it will see system_freezing_cnt and pm_freezing=true and freeze, and suspend to disk will hang forever. It could hit try_to_freeze() because of a signal delivered to the task, or from calling any syscall that uses a freezable_* helper like the one I added to sys_futex.

Quoted 4 lines

I think the right solution is to add a flag to the freezing task that marks it unfreezable. I think PF_NOFREEZE would work, although it is normally used on kernel threads, can you see if the attached patch helps?

Quoted 8 lines

That patch helps.

BTW, the only machine I can reproduce this bug with is an i7-3630QM notebook. Cannot reproduce on an Core Duo U1400 and cannot reproduce on an i7 M 620.

Are the sysreq backtraces still wanted? If so, any tip, how I could get them saved?

Typically by setting up a serial console or a netconsole and saving the log from the attached terminal emulator (such as screen or minicom).

Is this what you are asking?

Signature · 3 lines
Darren Hart
Intel Open Source Technology Center
Yocto Project - Linux Kernel
#31 lore raw

Re: 3.11-rc regression bisected: s2disk does not work (was Re: [PATCH v3 13/16] futex: use freezable blocking call)

In reply to Michael Leun
Quoted 44 lines · Michael Leun
Quoted 30 lines · On Mon, 22 Jul 2013 16:55:58 -0700 Colin Cross <[email protected]> wrote:
Quoted 6 lines · Michael Leun
Quoted 1 line · On Mon, 6 May 2013 16:50:18 -0700 Colin Cross <[email protected]> wrote:

Avoid waking up every thread sleeping in a futex_wait call during

[...]

Quoted 8 lines

With 3.11-rc s2disk from suspend-utils stopped working: Frozen at displaying 0% of saving image to disk.

echo "1" >/sys/power/state still works.

Bisecting yielded 88c8004fd3a5fdd2378069de86b90b21110d33a4, reverting that from 3.11-rc2 makes s2disk working again.

I think the expanded use of the freezable_* helpers is exposing an existing bug in hibernation. The SNAPSHOT_FREEZE ioctl calls freeze_processes(), which sets the global system_freezing_cnt and pm_freezing. try_to_freeze_tasks then sends every process except current a signal which causes them all to end up in the refrigerator. The current task then returns back to userspace and continues its work to suspend to disk. If that task ever hits a call to try_to_freeze() in the kernel, it will see system_freezing_cnt and pm_freezing=true and freeze, and suspend to disk will hang forever. It could hit try_to_freeze() because of a signal delivered to the task, or from calling any syscall that uses a freezable_* helper like the one I added to sys_futex.

Quoted 4 lines

I think the right solution is to add a flag to the freezing task that marks it unfreezable. I think PF_NOFREEZE would work, although it is normally used on kernel threads, can you see if the attached patch helps?

That patch helps.

BTW, the only machine I can reproduce this bug with is an i7-3630QM notebook. Cannot reproduce on an Core Duo U1400 and cannot reproduce on an i7 M 620.

Quoted 9 lines

Are the sysreq backtraces still wanted? If so, any tip, how I could get them saved?

--
MfG,

Michael Leun

Any chance that the failing machine has threads=y in the suspend.conf file?

Rafael, it appears that swsusp's suspend.c spawns new threads after calling the SNAPSHOT_FREEZE ioctl. The PF_NOFREEZE (or the new flag) will get copied to those new threads, but nothing will clear the flag.
Should I just assume that the userspace suspend code will kill those
threads before continuing with suspend? Or maybe add a WARN_ON in the kernel if any threads besides current have the new flag set when the suspend ops that assume all of userspace is frozen are called?

#32 lore raw

Re: 3.11-rc regression bisected: s2disk does not work (was Re: [PATCH v3 13/16] futex: use freezable blocking call)

Quoted 49 lines · On Tue, 23 Jul 2013 11:29:57 -0700 Colin Cross <[email protected]> wrote:
Quoted 47 lines · Michael Leun
Quoted 31 lines · On Mon, 22 Jul 2013 16:55:58 -0700 Colin Cross <[email protected]> wrote:
Quoted 7 lines · Michael Leun
Quoted 2 lines · On Mon, 6 May 2013 16:50:18 -0700 Colin Cross <[email protected]> wrote:

Avoid waking up every thread sleeping in a futex_wait call during

[...]

Quoted 8 lines

With 3.11-rc s2disk from suspend-utils stopped working: Frozen at displaying 0% of saving image to disk.

echo "1" >/sys/power/state still works.

Bisecting yielded 88c8004fd3a5fdd2378069de86b90b21110d33a4, reverting that from 3.11-rc2 makes s2disk working again.

I think the expanded use of the freezable_* helpers is exposing an existing bug in hibernation. The SNAPSHOT_FREEZE ioctl calls freeze_processes(), which sets the global system_freezing_cnt and pm_freezing. try_to_freeze_tasks then sends every process except current a signal which causes them all to end up in the refrigerator. The current task then returns back to userspace and continues its work to suspend to disk. If that task ever hits a call to try_to_freeze() in the kernel, it will see system_freezing_cnt and pm_freezing=true and freeze, and suspend to disk will hang forever. It could hit try_to_freeze() because of a signal delivered to the task, or from calling any syscall that uses a freezable_* helper like the one I added to sys_futex.

Quoted 4 lines

I think the right solution is to add a flag to the freezing task that marks it unfreezable. I think PF_NOFREEZE would work, although it is normally used on kernel threads, can you see if the attached patch helps?

That patch helps.

BTW, the only machine I can reproduce this bug with is an i7-3630QM notebook. Cannot reproduce on an Core Duo U1400 and cannot reproduce on an i7 M 620.

Are the sysreq backtraces still wanted? If so, any tip, how I could get them saved?

Quoted 1 line · Darren Hart

Typically by setting up a serial console or a netconsole and saving

[...]

Quoted 1 line

Is this what you are asking?

Yes, and it indeed works - I halfway expected the net / netconsole stuff being already frozen in that situation...

Thanks, Darren - see below for the backtraces.

Quoted 3 lines

Any chance that the failing machine has threads=y in the suspend.conf file?

Yes, that indeed is the trigger / difference, enabling that on the U4100 (its not a U1400) machine makes that fail also and disabling makes it work on the i7-3630QM.

[ 1405.527138] SysRq : Changing Loglevel
[ 1405.527220] Loglevel set to 9
[ 1407.845730] SysRq : Show backtrace of all active CPUs
[ 1407.845818] sending NMI to all CPUs:
[ 1407.845835] NMI backtrace for cpu 4
[ 1407.845870] CPU: 4 PID: 0 Comm: swapper/4 Not tainted 3.11.0-rc2 #1
[ 1407.845911] Hardware name: CLEVO                             P15xEMx/P15xEMx, BIOS 4.6.5 01/24/2013
[ 1407.845967] task: ffff880803540000 ti: ffff88080353a000 task.ti: ffff88080353a000
[ 1407.846002] RIP: 0010:[<ffffffff812ac403>]  [<ffffffff812ac403>] intel_idle+0xa3/0xf0
[ 1407.846046] RSP: 0000:ffff88080353bde8  EFLAGS: 00000046
[ 1407.846072] RAX: 0000000000000030 RBX: 0000000000000010 RCX: 0000000000000001
[ 1407.846104] RDX: 0000000000000000 RSI: ffff88080353bfd8 RDI: 0000000000000004
[ 1407.846137] RBP: ffff88080353be18 R08: 0000000000000057 R09: 000000000fde67ee
[ 1407.846169] R10: 0000000000000000 R11: 00000000003567bb R12: 0000000000000005
[ 1407.846201] R13: 0000000000000030 R14: 0000000000000004 R15: ffffffff81a56dd0
[ 1407.846246] FS:  0000000000000000(0000) GS:ffff88082f300000(0000) knlGS:0000000000000000
[ 1407.846283] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[ 1407.846310] CR2: 0000000000000000 CR3: 0000000001a0b000 CR4: 00000000001407e0
[ 1407.846342] Stack:
[ 1407.846355]  ffff88080353be18 0000000481098d4d ffff88082f319e00 ffffffff81a56c00
[ 1407.846401]  000001473bc66eed 0000000000000005 ffff88080353be78 ffffffff81371eea
[ 1407.846452]  000000000000010a 000000000df017f3 000000000000010a 000000000df017f3
[ 1407.846498] Call Trace:
[ 1407.846520]  [<ffffffff81371eea>] cpuidle_enter_state+0x4a/0xd0
[ 1407.846550]  [<ffffffff81372026>] cpuidle_idle_call+0xb6/0x260
[ 1407.846580]  [<ffffffff8100c699>] arch_cpu_idle+0x9/0x20
[ 1407.846607]  [<ffffffff81097800>] cpu_startup_entry+0x80/0x280
[ 1407.846637]  [<ffffffff8109f301>] ? clockevents_config_and_register+0x21/0x30
[ 1407.846672]  [<ffffffff8102cc2c>] start_secondary+0x1cc/0x270
[ 1407.846709] Code: 28 e0 ff ff 83 e2 08 75 22 31 d2 48 83 c0 10 48 89 d1 0f 01 c8 0f ae f0 48 8b 86 38 e0 ff ff a8 08 75 08 b1 01 4c 89 e8 0f 01 c9 <85> 1d 8f ab 7a 00 75 0e 48 8d 75 dc bf 05 00 00 00 e8 37 2a df 
[ 1407.847137] NMI backtrace for cpu 0
[ 1407.847140] INFO: NMI handler (arch_trigger_all_cpu_backtrace_handler) took too long to run: 1.301 msecs
[ 1407.847198] CPU: 0 PID: 0 Comm: swapper/0 Not tainted 3.11.0-rc2 #1
[ 1407.847220] Hardware name: CLEVO                             P15xEMx/P15xEMx, BIOS 4.6.5 01/24/2013
[ 1407.847254] task: ffffffff81a10440 ti: ffffffff81a00000 task.ti: ffffffff81a00000
[ 1407.847283] RIP: 0010:[<ffffffff8126667f>] [ 1407.847387] RBP: ffff88082f203b98 R08: 0000000000000001 R09: 000000000000066c
[ 1407.847414] R10: ffffffff81a1ec40 R11: 0000000000000000 R12: 000000000000006c
[ 1407.847441] R13: 0000000000000086 R14: 0000000000000001 R15: 0000000000000009
[ 1407.847467] FS:  0000000000000000(0000) GS:ffff88082f200000(0000) knlGS:0000000000000000
[ 1407.847495] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[ 1407.847516] CR2: 00007fffb526ebdc CR3: 0000000001a0b000 CR4: 00000000001407f0
[ 1407.847540] Stack:
[ 1407.847550]  ffff88082f203bb8 ffffffff81030500 0000000000000000 ffffffff81a5bd40
[ 1407.847589]  ffff88082f203bc8 ffffffff812ec099 ffff88082f203c08 ffffffff812ec547
[ 1407.847624]  ffff88082f203c88 ffff88080e80b400 0000000000000026 0000000000000001
[ 1407.848858] Call Trace:
[ 1407.850070]  <IRQ> 
[ 1407.850079]  [<ffffffff81030500>] arch_trigger_all_cpu_backtrace+0x80/0xa0
[ 1407.852518]  [<ffffffff812ec099>] sysrq_handle_showallcpus+0x9/0x10
[ 1407.853731]  [<ffffffff812ec547>] __handle_sysrq+0x127/0x190
[ 1407.854922]  [<ffffffff812ec92e>] sysrq_filter+0x33e/0x380
[ 1407.856114]  [<ffffffff81335522>] input_to_handler+0x52/0xf0
[ 1407.857308]  [<ffffffff813375a9>] input_pass_values.part.9+0x169/0x170
[ 1407.858508]  [<ffffffff813388c7>] input_handle_event+0x117/0x530
[ 1407.859706]  [<ffffffff81338de2>] input_event+0x52/0x70
[ 1407.860910]  [<ffffffff81340457>] atkbd_interrupt+0x5e7/0x6b0
[ 1407.862117]  [<ffffffff813329ed>] serio_interrupt+0x4d/0xa0
[ 1407.863317]  [<ffffffff81333e4a>] i8042_interrupt+0x1ba/0x3a0
[ 1407.864513]  [<ffffffff810708a1>] ? raw_notifier_call_chain+0x11/0x20
[ 1407.865716]  [<ffffffff810984f8>] ? timekeeping_update.constprop.8+0x38/0x80
[ 1407.866926]  [<ffffffff812a1bb0>] ? fbcon_add_cursor_timer+0x100/0x100
[ 1407.868139]  [<ffffffff810cc78d>] handle_irq_event_percpu+0x6d/0x240
[ 1407.869357]  [<ffffffff810cc9a3>] handle_irq_event+0x43/0x70
[ 1407.870572]  [<ffffffff810cf07f>] handle_edge_irq+0x6f/0x110
[ 1407.871788]  [<ffffffff81004aed>] handle_irq+0x1d/0x30
[ 1407.872996]  [<ffffffff810045c5>] do_IRQ+0x55/0xd0
[ 1407.874202]  [<ffffffff814989ea>] common_interrupt+0x6a/0x6a
[ 1407.875410]  <EOI> 
[ 1407.875419]  [<ffffffff810a0b7f>] ? tick_program_event+0x1f/0x30
[ 1407.877811]  [<ffffffff81371ef6>] ? cpuidle_enter_state+0x56/0xd0
[ 1407.879001]  [<ffffffff81371ef2>] ? cpuidle_enter_state+0x52/0xd0
[ 1407.880171]  [<ffffffff81372026>] cpuidle_idle_call+0xb6/0x260
[ 1407.881330]  [<ffffffff8100c699>] arch_cpu_idle+0x9/0x20
[ 1407.882484]  [<ffffffff81097800>] cpu_startup_entry+0x80/0x280
[ 1407.883642]  [<ffffffff81485380>] rest_init+0x80/0x90
[ 1407.884795]  [<ffffffff81ab2e49>] start_kernel+0x3aa/0x3b7
[ 1407.885944]  [<ffffffff81ab289e>] ? repair_env_string+0x5e/0x5e
[ 1407.887093]  [<ffffffff81ab25a3>] x86_64_start_reservations+0x2a/0x2c
[ 1407.888236]  [<ffffffff81ab269d>] x86_64_start_kernel+0xf8/0xfc
[ 1407.889366] Code: 4c 89 4d f8 c7 45 b8 10 00 00 00 48 89 45 c8 e8 38 ff ff ff c9 c3 66 0f 1f 44 00 00 8d 4e 3f 85 f6 55 0f 49 ce 48 89 e5 c1 f9 06 <85> c9 7e 61 48 83 3f 00 75 57 48 8d 57 08 31 c0 eb 12 0f 1f 80 
[ 1407.890771] NMI backtrace for cpu 1
[ 1407.890773] INFO: NMI handler (arch_trigger_all_cpu_backtrace_handler) took too long to run: 44.935 msecs
[ 1407.894228] CPU: 1 PID: 0 Comm: swapper/1 Not tainted 3.11.0-rc2 #1
[ 1407.895746] Hardware name: CLEVO                             P15xEMx/P15xEMx, BIOS 4.6.5 01/24/2013
[ 1407.897286] task: ffff88080350aea0 ti: ffff880803534000 task.ti: ffff880803534000
[ 1407.899298] RIP: 0010:[<ffffffff812ac403>]  [<ffffffff812ac403>] intel_idle+0xa3/0xf0
[ 1407.901326] RSP: 0018:ffff880803535de8  EFLAGS: 00000046
[ 1407.902884] RAX: 0000000000000030 RBX: 0000000000000010 RCX: 0000000000000001
[ 1407.904456] RDX: 0000000000000000 RSI: ffff880803535fd8 RDI: 0000000000000001
[ 1407.906026] RBP: ffff880803535e18 R08: 0000000000000057 R09: 000000000ff56e74
[ 1407.907597] R10: 0000000000000000 R11: 00000000003567c1 R12: 0000000000000005
[ 1407.909164] R13: 0000000000000030 R14: 0000000000000004 R15: ffffffff81a56dd0
[ 1407.910669] FS:  0000000000000000(0000) GS:ffff88082f240000(0000) knlGS:0000000000000000
[ 1407.912575] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[ 1407.914455] CR2: 00007ff17fd28000 CR3: 0000000001a0b000 CR4: 00000000001407e0
[ 1407.915879] Stack:
[ 1407.917716]  ffff880803535e18 0000000181098d4d ffff88082f259e00 ffffffff81a56c00
[ 1407.919152]  00000146e25e4673 0000000000000005 ffff880803535e78 ffffffff81371eea
[ 1407.920573]  000000000000010b 000000002c46c98c 000000000000010b 000000002c46c98c
[ 1407.921986] Call Trace:
[ 1407.923370]  [<ffffffff81371eea>] cpuidle_enter_state+0x4a/0xd0
[ 1407.924769]  [<ffffffff81372026>] cpuidle_idle_call+0xb6/0x260
[ 1407.926166]  [<ffffffff8100c699>] arch_cpu_idle+0x9/0x20
[ 1407.927561]  [<ffffffff81097800>] cpu_startup_entry+0x80/0x280
[ 1407.928963]  [<ffffffff8109f301>] ? clockevents_config_and_register+0x21/0x30
[ 1407.930368]  [<ffffffff8102cc2c>] start_secondary+0x1cc/0x270
[ 1407.932200] Code: 28 e0 ff ff 83 e2 08 75 22 31 d2 48 83 c0 10 48 89 d1 0f 01 c8 0f ae f0 48 8b 86 38 e0 ff ff a8 08 75 08 b1 01 4c 89 e8 0f 01 c9 <85> 1d 8f ab 7a 00 75 0e 48 8d 75 dc bf 05 00 00 00 e8 37 2a df 
[ 1407.933922] NMI backtrace for cpu 5
[ 1407.933924] INFO: NMI handler (arch_trigger_all_cpu_backtrace_handler) took too long to run: 88.055 msecs
[ 1407.936433] CPU: 5 PID: 0 Comm: swapper/5 Not tainted 3.11.0-rc2 #1
[ 1407.937700] Hardware name: CLEVO                             P15xEMx/P15xEMx, BIOS 4.6.5 01/24/2013
[ 1407.938987] task: ffff880803541750 ti: ffff88080353c000 task.ti: ffff88080353c000
[ 1407.940270] RIP: 0010:[<ffffffff812ac403>]  [<ffffffff812ac403>] intel_idle+0xa3/0xf0
[ 1407.941576] RSP: 0000:ffff88080353dde8  EFLAGS: 00000046
[ 1407.942873] RAX: 0000000000000030 RBX: 0000000000000010 RCX: 0000000000000001
[ 1407.944181] RDX: 0000000000000000 RSI: ffff88080353dfd8 RDI: 0000000000000005
[ 1407.945488] RBP: ffff88080353de18 R08: 0000000000000057 R09: 000000000fd6bb17
[ 1407.946798] R10: 0000000000000000 R11: 00000000003567d7 R12: 0000000000000005
[ 1407.948105] R13: 0000000000000030 R14: 0000000000000004 R15: ffffffff81a56dd0
[ 1407.949415] FS:  0000000000000000(0000) GS:ffff88082f340000(0000) knlGS:0000000000000000
[ 1407.950735] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[ 1407.952055] CR2: 0000000000000000 CR3: 0000000001a0b000 CR4: 00000000001407e0
[ 1407.953389] Stack:
[ 1407.954719]  ffff88080353de18 0000000581098d4d ffff88082f359e00 ffffffff81a56c00
[ 1407.956084]  000001475993cee2 0000000000000005 ffff88080353de78 ffffffff81371eea
[ 1407.957448]  0000000000000109 000000002b8fbaa2 0000000000000109 000000002b8fbaa2
[ 1407.958812] Call Trace:
[ 1407.960160]  [<ffffffff81371eea>] cpuidle_enter_state+0x4a/0xd0
[ 1407.961519]  [<ffffffff81372026>] cpuidle_idle_call+0xb6/0x260
[ 1407.962879]  [<ffffffff8100c699>] arch_cpu_idle+0x9/0x20
[ 1407.964240]  [<ffffffff81097800>] cpu_startup_entry+0x80/0x280
[ 1407.965602]  [<ffffffff8109f301>] ? clockevents_config_and_register+0x21/0x30
[ 1407.966978]  [<ffffffff8102cc2c>] start_secondary+0x1cc/0x270
[ 1407.968352] Code: 28 e0 ff ff 83 e2 08 75 22 31 d2 48 83 c0 10 48 89 d1 0f 01 c8 0f ae f0 48 8b 86 38 e0 ff ff a8 08 75 08 b1 01 4c 89 e8 0f 01 c9 <85> 1d 8f ab 7a 00 75 0e 48 8d 75 dc bf 05 00 00 00 e8 37 2a df 
[ 1407.970046] NMI backtrace for cpu 6
[ 1407.970048] INFO: NMI handler (arch_trigger_all_cpu_backtrace_handler) took too long to run: 124.179 msecs
[ 1407.974761] CPU: 6 PID: 0 Comm: swapper/6 Not tainted 3.11.0-rc2 #1
[ 1407.976613] Hardware name: CLEVO                             P15xEMx/P15xEMx, BIOS 4.6.5 01/24/2013
[ 1407.978491] task: ffff880803542ea0 ti: ffff88080353e000 task.ti: ffff88080353e000
[ 1407.980381] RIP: 0010:[<ffffffff812ac403>]  [<ffffffff812ac403>] intel_idle+0xa3/0xf0
[ 1407.982855] RSP: 0018:ffff88080353fde8  EFLAGS: 00000046
[ 1407.984756] RAX: 0000000000000030 RBX: 0000000000000010 RCX: 0000000000000001
[ 1407.986674] RDX: 0000000000000000 RSI: ffff88080353ffd8 RDI: 0000000000000006
[ 1407.988590] RBP: ffff88080353fe18 R08: 0000000000000057 R09: 000000000fcf0e3f
[ 1407.991088] R10: 0000000000000000 R11: 00000000003567da R12: 0000000000000005
[ 1407.993018] R13: 0000000000000030 R14: 0000000000000004 R15: ffffffff81a56dd0
[ 1407.994875] FS:  0000000000000000(0000) GS:ffff88082f380000(0000) knlGS:0000000000000000
[ 1407.996698] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[ 1407.998485] CR2: 0000000000000000 CR3: 0000000001a0b000 CR4: 00000000001407e0
[ 1408.000272] Stack:
[ 1408.002572]  ffff88080353fe18 0000000681098d4d ffff88082f399e00 ffffffff81a56c00
[ 1408.004361]  0000014777613532 0000000000000005 ffff88080353fe78 ffffffff81371eea
[ 1408.006674]  0000000000000109 000000000d948c6f 0000000000000109 000000000d948c6f
[ 1408.008443] Call Trace:
[ 1408.010189]  [<ffffffff81371eea>] cpuidle_enter_state+0x4a/0xd0
[ 1408.011951]  [<ffffffff81372026>] cpuidle_idle_call+0xb6/0x260
[ 1408.014250]  [<ffffffff8100c699>] arch_cpu_idle+0x9/0x20
[ 1408.016546]  [<ffffffff81097800>] cpu_startup_entry+0x80/0x280
[ 1408.018843]  [<ffffffff8109f301>] ? clockevents_config_and_register+0x21/0x30
[ 1408.021144]  [<ffffffff8102cc2c>] start_secondary+0x1cc/0x270
[ 1408.022885] Code: 28 e0 ff ff 83 e2 08 75 22 31 d2 48 83 c0 10 48 89 d1 0f 01 c8 0f ae f0 48 8b 86 38 e0 ff ff a8 08 75 08 b1 01 4c 89 e8 0f 01 c9 <85> 1d 8f ab 7a 00 75 0e 48 8d 75 dc bf 05 00 00 00 e8 37 2a df 
[ 1408.024957] INFO: NMI handler (arch_trigger_all_cpu_backtrace_handler) took too long to run: 179.088 msecs
[ 1408.024958] NMI backtrace for cpu 2
[ 1408.024960] CPU: 2 PID: 0 Comm: swapper/2 Not tainted 3.11.0-rc2 #1
[ 1408.024961] Hardware name: CLEVO                             P15xEMx/P15xEMx, BIOS 4.6.5 01/24/2013
[ 1408.024962] task: ffff88080350c5f0 ti: ffff880803536000 task.ti: ffff880803536000
[ 1408.024964] RIP: 0010:[<ffffffff812ac403>]  [<ffffffff812ac403>] intel_idle+0xa3/0xf0
[ 1408.024965] RSP: 0000:ffff880803537de8  EFLAGS: 00000046
[ 1408.024966] RAX: 0000000000000030 RBX: 0000000000000010 RCX: 0000000000000001
[ 1408.024967] RDX: 0000000000000000 RSI: ffff880803537fd8 RDI: 0000000000000002
[ 1408.024967] RBP: ffff880803537e18 R08: 0000000000000057 R09: 000000000fedc1a1
[ 1408.024968] R10: 0000000000000000 R11: 00000000003567d7 R12: 0000000000000005
[ 1408.024969] R13: 0000000000000030 R14: 0000000000000004 R15: ffffffff81a56dd0
[ 1408.024970] FS:  0000000000000000(0000) GS:ffff88082f280000(0000) knlGS:0000000000000000
[ 1408.024971] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[ 1408.024971] CR2: 0000000000000000 CR3: 0000000001a0b000 CR4: 00000000001407e0
[ 1408.024972] Stack:
[ 1408.024974]  ffff880803537e18 0000000281098d4d ffff88082f299e00 ffffffff81a56c00
[ 1408.024975]  00000147002b9b55 0000000000000005 ffff880803537e78 ffffffff81371eea
[ 1408.024976]  000000000000010b 000000000e4baf99 000000000000010b 000000000e4baf99
[ 1408.024977] Call Trace:
[ 1408.024979]  [<ffffffff81371eea>] cpuidle_enter_state+0x4a/0xd0
[ 1408.024982]  [<ffffffff81372026>] cpuidle_idle_call+0xb6/0x260
[ 1408.024984]  [<ffffffff8100c699>] arch_cpu_idle+0x9/0x20
[ 1408.024985]  [<ffffffff81097800>] cpu_startup_entry+0x80/0x280
[ 1408.024987]  [<ffffffff8109f301>] ? clockevents_config_and_register+0x21/0x30
[ 1408.024989]  [<ffffffff8102cc2c>] start_secondary+0x1cc/0x270
[ 1408.025006] Code: 28 e0 ff ff 83 e2 08 75 22 31 d2 48 83 c0 10 48 89 d1 0f 01 c8 0f ae f0 48 8b 86 38 e0 ff ff a8 08 75 08 b1 01 4c 89 e8 0f 01 c9 <85> 1d 8f ab 7a 00 75 0e 48 8d 75 dc bf 05 00 00 00 e8 37 2a df 
[ 1408.025008] INFO: NMI handler (arch_trigger_all_cpu_backtrace_handler) took too long to run: 179.139 msecs
[ 1408.025009] NMI backtrace for cpu 7
[ 1408.025011] CPU: 7 PID: 0 Comm: swapper/7 Not tainted 3.11.0-rc2 #1
[ 1408.025011] Hardware name: CLEVO                             P15xEMx/P15xEMx, BIOS 4.6.5 01/24/2013
[ 1408.025013] task: ffff8808035445f0 ti: ffff880803550000 task.ti: ffff880803550000
[ 1408.025016] RIP: 0010:[<ffffffff812ac403>]  [<ffffffff812ac403>] intel_idle+0xa3/0xf0
[ 1408.025016] RSP: 0000:ffff880803551de8  EFLAGS: 00000046
[ 1408.025017] RAX: 0000000000000030 RBX: 0000000000000010 RCX: 0000000000000001
[ 1408.025018] RDX: 0000000000000000 RSI: ffff880803551fd8 RDI: 0000000000000007
[ 1408.025018] RBP: ffff880803551e18 R08: 0000000000000057 R09: 000000000007a079
[ 1408.025019] R10: 0000000000000000 R11: 000000000020e8dc R12: 0000000000000005
[ 1408.025020] R13: 0000000000000030 R14: 0000000000000004 R15: ffffffff81a56dd0
[ 1408.025021] FS:  0000000000000000(0000) GS:ffff88082f3c0000(0000) knlGS:0000000000000000
[ 1408.025022] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[ 1408.025022] CR2: 0000000000000000 CR3: 0000000001a0b000 CR4: 00000000001407e0
[ 1408.025023] Stack:
[ 1408.025025]  ffff880803551e18 0000000781098d4d ffff88082f3d9e00 ffffffff81a56c00
[ 1408.025026]  0000014777614583 0000000000000005 ffff880803551e78 ffffffff81371eea
[ 1408.025027]  0000000000000000 000000001dcad99a 0000000000000000 000000001dcad99a
[ 1408.025028] Call Trace:
[ 1408.025030]  [<ffffffff81371eea>] cpuidle_enter_state+0x4a/0xd0
[ 1408.025032]  [<ffffffff81372026>] cpuidle_idle_call+0xb6/0x260
[ 1408.025035]  [<ffffffff8100c699>] arch_cpu_idle+0x9/0x20
[ 1408.025036]  [<ffffffff81097800>] cpu_startup_entry+0x80/0x280
[ 1408.025038]  [<ffffffff8109f301>] ? clockevents_config_and_register+0x21/0x30
[ 1408.025040]  [<ffffffff8102cc2c>] start_secondary+0x1cc/0x270
[ 1408.025057] Code: 28 e0 ff ff 83 e2 08 75 22 31 d2 48 83 c0 10 48 89 d1 0f 01 c8 0f ae f0 48 8b 86 38 e0 ff ff a8 08 75 08 b1 01 4c 89 e8 0f 01 c9 <85> 1d 8f ab 7a 00 75 0e 48 8d 75 dc bf 05 00 00 00 e8 37 2a df 
[ 1408.025057] NMI backtrace for cpu 3
[ 1408.025059] INFO: NMI handler (arch_trigger_all_cpu_backtrace_handler) took too long to run: 179.189 msecs
[ 1408.025061] CPU: 3 PID: 0 Comm: swapper/3 Not tainted 3.11.0-rc2 #1
[ 1408.025061] Hardware name: CLEVO                             P15xEMx/P15xEMx, BIOS 4.6.5 01/24/2013
[ 1408.025062] task: ffff88080350dd40 ti: ffff880803538000 task.ti: ffff880803538000
[ 1408.025064] RIP: 0010:[<ffffffff812ac403>]  [<ffffffff812ac403>] intel_idle+0xa3/0xf0
[ 1408.025065] RSP: 0000:ffff880803539de8  EFLAGS: 00000046
[ 1408.025065] RAX: 0000000000000030 RBX: 0000000000000010 RCX: 0000000000000001
[ 1408.025066] RDX: 0000000000000000 RSI: ffff880803539fd8 RDI: 0000000000000003
[ 1408.025066] RBP: ffff880803539e18 R08: 0000000000000057 R09: 00000000001e939c
[ 1408.025067] R10: 0000000000000000 R11: 00000000001c6983 R12: 0000000000000005
[ 1408.025067] R13: 0000000000000030 R14: 0000000000000004 R15: ffffffff81a56dd0
[ 1408.025068] FS:  0000000000000000(0000) GS:ffff88082f2c0000(0000) knlGS:0000000000000000
[ 1408.025068] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[ 1408.025069] CR2: 0000000000000000 CR3: 0000000001a0b000 CR4: 00000000001407e0
[ 1408.025069] Stack:
[ 1408.025070]  ffff880803539e18 0000000381098d4d ffff88082f2d9e00 ffffffff81a56c00
[ 1408.025071]  0000014780148c27 0000000000000005 ffff880803539e78 ffffffff81371eea
[ 1408.025071]  0000000000000002 00000000003b0659 0000000000000002 00000000003b0659
[ 1408.025072] Call Trace:
[ 1408.025073]  [<ffffffff81371eea>] cpuidle_enter_state+0x4a/0xd0
[ 1408.025075]  [<ffffffff81372026>] cpuidle_idle_call+0xb6/0x260
[ 1408.025076]  [<ffffffff8100c699>] arch_cpu_idle+0x9/0x20
[ 1408.025078]  [<ffffffff81097800>] cpu_startup_entry+0x80/0x280
[ 1408.025079]  [<ffffffff8109f301>] ? clockevents_config_and_register+0x21/0x30
[ 1408.025080]  [<ffffffff8102cc2c>] start_secondary+0x1cc/0x270
[ 1408.025090] Code: 28 e0 ff ff 83 e2 08 75 22 31 d2 48 83 c0 10 48 89 d1 0f 01 c8 0f ae f0 48 8b 86 38 e0 ff ff a8 08 75 08 b1 01 4c 89 e8 0f 01 c9 <85> 1d 8f ab 7a 00 75 0e 48 8d 75 dc bf 05 00 00 00 e8 37 2a df 
[ 1408.025091] INFO: NMI handler (arch_trigger_all_cpu_backtrace_handler) took too long to run: 179.222 msecs
Signature · 3 lines
MfG,

Michael Leun
#33 lore raw

Re: 3.11-rc regression bisected: s2disk does not work (was Re: [PATCH v3 13/16] futex: use freezable blocking call)

Quoted 65 lines · Michael Leun
Quoted 49 lines · On Tue, 23 Jul 2013 11:29:57 -0700 Colin Cross…
Quoted 47 lines · Michael Leun
Quoted 31 lines · On Mon, 22 Jul 2013 16:55:58 -0700 Colin Cross…
Quoted 7 lines · Michael Leun
Quoted 2 lines · On Mon, 6 May 2013 16:50:18 -0700 Colin Cross…

Avoid waking up every thread sleeping in a futex_wait call during

[...]

Quoted 8 lines

With 3.11-rc s2disk from suspend-utils stopped working: Frozen at displaying 0% of saving image to disk.

echo "1" >/sys/power/state still works.

Bisecting yielded 88c8004fd3a5fdd2378069de86b90b21110d33a4, reverting that from 3.11-rc2 makes s2disk working again.

I think the expanded use of the freezable_* helpers is exposing an existing bug in hibernation. The SNAPSHOT_FREEZE ioctl calls freeze_processes(), which sets the global system_freezing_cnt and pm_freezing. try_to_freeze_tasks then sends every process except current a signal which causes them all to end up in the refrigerator. The current task then returns back to userspace and continues its work to suspend to disk. If that task ever hits a call to try_to_freeze() in the kernel, it will see system_freezing_cnt and pm_freezing=true and freeze, and suspend to disk will hang forever. It could hit try_to_freeze() because of a signal delivered to the task, or from calling any syscall that uses a freezable_* helper like the one I added to sys_futex.

Quoted 4 lines

I think the right solution is to add a flag to the freezing task that marks it unfreezable. I think PF_NOFREEZE would work, although it is normally used on kernel threads, can you see if the attached patch helps?

That patch helps.

BTW, the only machine I can reproduce this bug with is an i7-3630QM notebook. Cannot reproduce on an Core Duo U1400 and cannot reproduce on an i7 M 620.

Are the sysreq backtraces still wanted? If so, any tip, how I could get them saved?

Quoted 1 line · Darren Hart

Typically by setting up a serial console or a netconsole and saving

[...]

Quoted 1 line

Is this what you are asking?

Yes, and it indeed works - I halfway expected the net / netconsole stuff being already frozen in that situation...

Thanks, Darren - see below for the backtraces.

Quoted 1 line
Quoted 6 lines
Quoted 2 lines

Any chance that the failing machine has threads=y in the suspend.conf file?

Yes, that indeed is the trigger / difference, enabling that on the U4100 (its not a U1400) machine makes that fail also and disabling makes it work on the i7-3630QM.

Thanks, if you get a chance sysrq w might be interesting but I think we have enough info to solve the problem.
--
To unsubscribe from this list: send the line "unsubscribe linux-nfs" in the body of a message to [email protected] More majordomo info at http://vger.kernel.org/majordomo-info.html

#34 lore raw

Re: 3.11-rc regression bisected: s2disk does not work (was Re: [PATCH v3 13/16] futex: use freezable blocking call)

Quoted 71 lines · On Tue, 23 Jul 2013 12:29:57 -0700 Colin Cross…
Quoted 69 lines · Michael Leun
Quoted 50 lines · On Tue, 23 Jul 2013 11:29:57 -0700 Colin Cross…
Quoted 48 lines · Michael Leun
Quoted 32 lines · On Mon, 22 Jul 2013 16:55:58 -0700 Colin Cross…
Quoted 7 lines · Michael Leun
Quoted 2 lines · On Mon, 6 May 2013 16:50:18 -0700 Colin Cross…

Avoid waking up every thread sleeping in a futex_wait call during

[...]

Quoted 8 lines

With 3.11-rc s2disk from suspend-utils stopped working:
Frozen at displaying 0% of saving image to disk.

echo "1" >/sys/power/state still works.

Bisecting yielded 88c8004fd3a5fdd2378069de86b90b21110d33a4, reverting that from 3.11-rc2 makes s2disk working again.

I think the expanded use of the freezable_* helpers is exposing an existing bug in hibernation. The SNAPSHOT_FREEZE ioctl calls freeze_processes(), which sets the global system_freezing_cnt and pm_freezing. try_to_freeze_tasks then sends every process except current a signal which causes them all to end up in the refrigerator. The current task then returns back to userspace and continues its work to suspend to disk. If that task ever hits a call to try_to_freeze() in the kernel, it will see system_freezing_cnt and pm_freezing=true and freeze, and suspend to disk will hang forever. It could hit try_to_freeze() because of a signal delivered to the task, or from calling any syscall that uses a freezable_* helper like the one I added to sys_futex.

Quoted 4 lines

I think the right solution is to add a flag to the freezing task that marks it unfreezable. I think PF_NOFREEZE would work, although it is normally used on kernel threads, can you see if the attached patch helps?

That patch helps.

BTW, the only machine I can reproduce this bug with is an i7-3630QM notebook. Cannot reproduce on an Core Duo U1400 and cannot reproduce on an i7 M 620.

Are the sysreq backtraces still wanted? If so, any tip, how I could get them saved?

Quoted 1 line · Darren Hart

Typically by setting up a serial console or a netconsole and saving

[...]

Quoted 1 line

Is this what you are asking?

Yes, and it indeed works - I halfway expected the net / netconsole stuff being already frozen in that situation...

Thanks, Darren - see below for the backtraces.

Quoted 3 lines

Any chance that the failing machine has threads=y in the suspend.conf file?

Quoted 7 lines
Quoted 3 lines

Yes, that indeed is the trigger / difference, enabling that on the U4100 (its not a U1400) machine makes that fail also and disabling makes it work on the i7-3630QM.

Thanks, if you get a chance sysrq w might be interesting but I think we have enough info to solve the problem.

Now that I've set up everything this is no big effort...

[  343.801889] Loglevel set to 9
[  347.336205]  [<ffffffff81496df4>] schedule+0x24/0x70
[  347.337184]  [<ffffffff81097aad>] __refrigerator+0x4d/0x140
[  347.338170]  [<ffffffff8107601e>] ? __wake_up+0x4e/0x70
[  347.339163]  [<ffffffffa0198866>] kjournald2+0x236/0x240 [jbd2]
[  347.340156]  [<ffffffff8106ba70>] ? finish_wait+0x80/0x80
[  347.341162]  [<ffffffffa0198630>] ? journal_init_common+0x160/0x160 [jbd2]
[  347.342162]  [<ffffffff8106b27b>] kthread+0xbb/0xc0
[  347.343143]  [<ffffffff8106b1c0>] ? kthread_create_on_node+0x130/0x130
[  347.344116]  [<ffffffff814990ac>] ret_from_fork+0x7c/0xb0
[  347.345059]  [<ffffffff8106b1c0>] ? kthread_create_on_node+0x130/0x130
[  347.345983] systemd-journal D ffff88082f252d40     0   526      1 0x00000000
[  347.346919]  ffff8807f63e7dd8 0000000000000082 ffff8807f72e1750 ffff8807f63e7fd8
[  347.347871]  ffff8807f63e7fd8 ffff8807f63e7fd8 ffff88080350aea0 ffff8807f72e1750
[  347.348827]  ffff8807f63e7dc8 ffff8807f72e1750 ffff8807f72e1750 ffff8807f72e1750
[  347.349784] Call Trace:
[  347.350733]  [<ffffffff81496df4>] schedule+0x24/0x70
[  347.351688]  [<ffffffff81097aad>] __refrigerator+0x4d/0x140
[  347.352641]  [<ffffffff810bab88>] ? cgroup_freezing+0x28/0x40
[  347.353592]  [<ffffffff811b3a30>] ep_poll+0x320/0x340
[  347.354533]  [<ffffffff81392361>] ? sock_ioctl+0x71/0x2a0
[  347.355473]  [<ffffffff8107b120>] ? try_to_wake_up+0x2b0/0x2b0
[  347.356414]  [<ffffffff811b4aa5>] SyS_epoll_wait+0xd5/0x100
[  347.357350]  [<ffffffff81499152>] system_call_fastpath+0x16/0x1b
[  347.358286] kauditd         D ffff88082f312d40     0   527      2 0x00000000
[  347.359236]  ffff8807f618bde8 0000000000000046 ffff8807f7361750 ffff8807f618bfd8
[  347.360198]  ffff8807f618bfd8 ffff8807f618bfd8 ffff880803540000 ffff8807f7361750
[  347.361165]  ffff8807f7259380 ffff8807f7361750 ffff8807f7361750 ffff8807f7361750
[  347.362132] Call Trace:
[  347.363084]  [<ffffffff81496df4>] schedule+0x24/0x70
[  347.364043]  [<ffffffff81097aad>] __refrigerator+0x4d/0x140
[  347.365001]  [<ffffffff8107601e>] ? __wake_up+0x4e/0x70
[  347.365959]  [<ffffffff810c0c5a>] kauditd_thread+0x1aa/0x1b0
[  347.366913]  [<ffffffff8107b120>] ? try_to_wake_up+0x2b0/0x2b0
[  347.367860]  [<ffffffff810c0ab0>] ? audit_printk_skb+0x70/0x70
[  347.368805]  [<ffffffff8106b27b>] kthread+0xbb/0xc0
[  347.369746]  [<ffffffff8106b1c0>] ? kthread_create_on_node+0x130/0x130
[  347.370685]  [<ffffffff814990ac>] ret_from_fork+0x7c/0xb0
[  347.371621]  [<ffffffff8106b1c0>] ? kthread_create_on_node+0x130/0x130
[  347.372565] systemd-udevd   D ffff88082f252d40     0   553      1 0x00000000
[  347.373524]  ffff8807f5fdbdd8 0000000000000086 ffff8807f72b5d40 ffff8807f5fdbfd8
[  347.374496]  ffff8807f5fdbfd8 ffff8807f5fdbfd8 ffff88080350aea0 ffff8807f72b5d40
[  347.375471]  0000000000000000 ffff8807f72b5d40 ffff8807f72b5d40 ffff8807f72b5d40
[  347.376444] Call Trace:
[  347.377402]  [<ffffffff81496df4>] schedule+0x24/0x70
[  347.378369]  [<ffffffff81097aad>] __refrigerator+0x4d/0x140
[  347.379333]  [<ffffffff811b3a30>] ep_poll+0x320/0x340
[  347.380296]  [<ffffffff8107b120>] ? try_to_wake_up+0x2b0/0x2b0
[  347.381254]  [<ffffffff811b4aa5>] SyS_epoll_wait+0xd5/0x100
[  347.382206]  [<ffffffff81499152>] system_call_fastpath+0x16/0x1b
[  347.383158] haveged         D ffff88082f352d40     0   858      1 0x00000000
[  347.384129]  ffff8807f641b908 0000000000000082 ffff8807f45add40 ffff8807f641bfd8[  347.385106]  ffff8807f641bfd8 ffff8807f641bfd8 ffff880803541750 ffff8807f45add40
[  347.386083]  ffff8807f641b8f8 ffff8807f45add40 ffff8807f45add40 ffff8807f45add40[  347.389000]  [<ffffffff81097aad>] __refrigerator+0x4d/0x140
[  347.389976]  [<ffffffff810bab88>] ? cgroup_freezing+0x28/0x40
[  347.390949]  [<ffffffff81180699>] poll_schedule_timeout+0xa9/0xb0
[  347.391915]  [<ffffffff81180fd5>] do_select+0x6f5/0x840
[  347.393845]  [<ffffffff81180580>] ? __pollwait+0xf0/0xf0
[  347.396686]  [<ffffffff81105834>] ? filemap_fault+0x84/0x460
[  347.397623]  [<ffffffff81103512>] ? unlock_page+0x22/0x30
[  347.401351]  [<ffffffff81181303>] core_sys_select+0x1e3/0x310
[  347.402278]  [<ffffffff81306c7b>] ? credit_entropy_bits.part.7+0x18b/0x1f0
[  347.403209]  [<ffffffff8130710a>] ? random_ioctl+0x16a/0x190
[  347.404140]  [<ffffffff8130710a>] ? random_ioctl+0x16a/0x190
[  347.540893]  ffff8807f7378d00 ffff8808029f8000 ffff8808029f8000[  347.543682]  [<ffffffff81097aad>] __refrigerator+0x4d/0x140
[  347.544609]  [<ffffffff8107601e>] ? __wake_up+0x4e/0x70
[  347.546466]  [<ffffffff8106ba70>] ? finish_wait+0x80/0x80
[  347.549269]  [<ffffffff8106b1c0>] ? kthread_create_on_node+0x130/0x130
[  347.553048]  ffff8807f4de7d18 0000000000000086 ffff8807f6e02ea0 ffff8807f4de7fd8[  347.554022]  ffff8807f4de7fd8
 ffff8807f6e02ea0 ffff8807f6e02ea0
[  347.556887]  [<ffffffff81496df4>] schedule+0x24/0x70
[  347.557837]  [<ffffffff81097aad>] __refrigerator+0x4d/0x140
[  347.559725]  [<ffffffff8105b14c>] get_signal_to_deliver+0x5fc/0x650
 ffff88082f212d40 
[  347.577740]  [<ffffffff81180fd5>] do_select+0x6f5/0x840
[  347.584244]  [<ffffffff81128d6e>] ? __do_fault+0x1ee/0x520
[  347.590613]  [<ffffffff81499152>] system_call_fastpath+0x16/0x1b
[  347.592425]  ffff8807f5ce9908 ffff8807f4a82ea0 ffff8807f5ce9fd8[  347.594265]  0000000000007530 ffff8807f4a82ea0
[  347.597922]  [<ffffffff8106ed7f>] ? hrtimer_start_range_ns+0xf/0x20
 ffff88082f212d40 [  347.813797]  ffff8807fc6d5a68 0000000000000086 ffff8807fc6d5fd8
[  347.814556]  ffff8807fc6d5fd8 ffffffff81a10440[  347.956796]  [<ffffffff81097aad>] __refrigerator+0x4d/0x140
[  347.957668]  [<ffffffff810bab88>] ? cgroup_freezing+0x28/0x40
[  347.958543]  [<ffffffff8105b14c>] get_signal_to_deliver+0x5fc/0x650
[  347.959418]  [<ffffffff81002392>] do_signal+0x52/0x5a0
[  347.960290]  [<ffffffff8107b120>] ? try_to_wake_up+0x2b0/0x2b0
[  347.961160]  [<ffffffff812e7857>] ? tty_ldisc_deref+0x37/0xa0
[  347.962029]  [<ffffffff812dfd61>] ? tty_read+0xa1/0x100
[  347.962894]  [<ffffffff81002945>] do_notify_resume+0x65/0x80
[  347.963762]  [<ffffffff8116e21d>] ? SyS_read+0x4d/0xa0
[  347.964627]  [<ffffffff814993da>] int_signal+0x12/0x17
[  347.965489] bash            D ffff88082f212d40     0  2780   2707 0x00000004
[  347.966369]  ffff8807f99b7d18 0000000000000086 ffff8807f9cfc5f0 ffff8807f99b7fd8
[  347.967258]  ffff8807f99b7fd8 ffff8807f99b7fd8 ffffffff81a10440 ffff8807f9cfc5f0
[  347.968140]  ffff8807f99b7d08 ffff8807f9cfc5f0 ffff8807f9cfc5f0 ffff8807f9cfc5f0
[  347.969014] Call Trace:
[  347.969873]  [<ffffffff81496df4>] schedule+0x24/0x70
[  347.970743]  [<ffffffff81097aad>] __refrigerator+0x4d/0x140
[  347.971606]  [<ffffffff810bab88>] ? cgroup_freezing+0x28/0x40
[  347.972469]  [<ffffffff8105b14c>] get_signal_to_deliver+0x5fc/0x650
[  347.973333]  [<ffffffff81002392>] do_signal+0x52/0x5a0
[  347.974193]  [<ffffffff8107b120>] ? try_to_wake_up+0x2b0/0x2b0
[  347.975056]  [<ffffffff812e7857>] ? tty_ldisc_deref+0x37/0xa0
[  347.975917]  [<ffffffff812dfd61>] ? tty_read+0xa1/0x100
[  347.976774]  [<ffffffff81002945>] do_notify_resume+0x65/0x80
[  347.977634]  [<ffffffff8116e21d>] ? SyS_read+0x4d/0xa0
[  347.978492]  [<ffffffff814993da>] int_signal+0x12/0x17
[  347.979348] bash            D ffff88082f252d40     0  2786   2700 0x00000004
[  347.980221]  ffff8807f98d3d18 0000000000000086 ffff8807f9f92ea0 ffff8807f98d3fd8
[  347.981113]  ffff8807f98d3fd8 ffff8807f98d3fd8 ffff8807f9de0000 ffff8807f9f92ea0
[  347.981996]  ffff8807f98d3d08 ffff8807f9f92ea0 ffff8807f9f92ea0 ffff8807f9f92ea0
[  347.982868] Call Trace:
[  347.983722]  [<ffffffff81496df4>] schedule+0x24/0x70
[  347.984587]  [<ffffffff81097aad>] __refrigerator+0x4d/0x140
[  347.985448]  [<ffffffff810bab88>] ? cgroup_freezing+0x28/0x40
[  347.986306]  [<ffffffff8105b14c>] get_signal_to_deliver+0x5fc/0x650
[  347.987168]  [<ffffffff81002392>] do_signal+0x52/0x5a0
[  347.988027]  [<ffffffff8107b120>] ? try_to_wake_up+0x2b0/0x2b0
[  347.988890]  [<ffffffff812e7857>] ? tty_ldisc_deref+0x37/0xa0
[  347.989748]  [<ffffffff812dfd61>] ? tty_read+0xa1/0x100
[  347.990603]  [<ffffffff81002945>] do_notify_resume+0x65/0x80
[  347.991460]  [<ffffffff8116e21d>] ? SyS_read+0x4d/0xa0
[  347.992316]  [<ffffffff814993da>] int_signal+0x12/0x17
[  347.993170] bash            D ffff88082f352d40     0  2792   2707 0x00000004
[  347.994042]  ffff8807f9f25d18 0000000000000086 ffff8807f9cf8000 ffff8807f9f25fd8
[  347.994930]  ffff8807f9f25fd8 ffff8807f9f25fd8 ffff8807fa300000 ffff8807f9cf8000
[  347.995810]  ffff8807f9f25d08 ffff8807f9cf8000 ffff8807f9cf8000 ffff8807f9cf8000
[  347.996680] Call Trace:
[  347.997533]  [<ffffffff81496df4>] schedule+0x24/0x70
[  347.998397]  [<ffffffff81097aad>] __refrigerator+0x4d/0x140
[  347.999254]  [<ffffffff810bab88>] ? cgroup_freezing+0x28/0x40
[  348.000111]  [<ffffffff8105b14c>] get_signal_to_deliver+0x5fc/0x650
[  348.000971]  [<ffffffff81002392>] do_signal+0x52/0x5a0
[  348.001831]  [<ffffffff8107b120>] ? try_to_wake_up+0x2b0/0x2b0
[  348.002687]  [<ffffffff812e7857>] ? tty_ldisc_deref+0x37/0xa0
[  348.003540]  [<ffffffff812dfd61>] ? tty_read+0xa1/0x100
[  348.004395]  [<ffffffff81002945>] do_notify_resume+0x65/0x80
[  348.005248]  [<ffffffff8116e21d>] ? SyS_read+0x4d/0xa0
[  348.006102]  [<ffffffff814993da>] int_signal+0x12/0x17
[  348.006955] bash            D ffff88082f252d40     0  2798   2700 0x00000004
[  348.007823]  ffff8807f9e41d18 0000000000000082 ffff8807f9de0000 ffff8807f9e41fd8
[  348.008709]  ffff8807f9e41fd8 ffff8807f9e41fd8 ffff88080350aea0 ffff8807f9de0000
[  348.009588]  ffff8807f9e41d08 ffff8807f9de0000 ffff8807f9de0000 ffff8807f9de0000
[  348.010460] Call Trace:
[  348.011313]  [<ffffffff81496df4>] schedule+0x24/0x70
[  348.012179]  [<ffffffff81097aad>] __refrigerator+0x4d/0x140
[  348.013036]  [<ffffffff810bab88>] ? cgroup_freezing+0x28/0x40
[  348.013893]  [<ffffffff8105b14c>] get_signal_to_deliver+0x5fc/0x650
[  348.014755]  [<ffffffff81002392>] do_signal+0x52/0x5a0
[  348.015611]  [<ffffffff8107b120>] ? try_to_wake_up+0x2b0/0x2b0
[  348.016469]  [<ffffffff812e7857>] ? tty_ldisc_deref+0x37/0xa0
[  348.017325]  [<ffffffff812dfd61>] ? tty_read+0xa1/0x100
[  348.018179]  [<ffffffff81002945>] do_notify_resume+0x65/0x80
[  348.158814]  ffff8807fc441fd8 ffff8807fc441fd8 ffff880803540000 ffff8807f9c4dd40
[  348.159718]  ffff8807fc441a58 ffff8807f9c4dd40 ffff8807f9c4dd40 ffff8807f9c4dd40
[  348.160622] Call Trace:
[  348.161503]  [<ffffffff81496df4>] schedule+0x24/0x70
[  348.162386]  [<ffffffff81097aad>] __refrigerator+0x4d/0x140
[  348.163269]  [<ffffffff8106ed7f>] ? hrtimer_start_range_ns+0xf/0x20
[  348.164158]  [<ffffffff810bab88>] ? cgroup_freezing+0x28/0x40
[  348.165039]  [<ffffffff81180699>] poll_schedule_timeout+0xa9/0xb0
[  348.165921]  [<ffffffff81181b3d>] do_sys_poll+0x3ed/0x5b0
[  348.166798]  [<ffffffff81180580>] ? __pollwait+0xf0/0xf0
[  348.167676]  [<ffffffff81180580>] ? __pollwait+0xf0/0xf0
[  348.168544]  [<ffffffff81180580>] ? __pollwait+0xf0/0xf0
[  348.169405]  [<ffffffff81180580>] ? __pollwait+0xf0/0xf0
[  348.170255]  [<ffffffff81180580>] ? __pollwait+0xf0/0xf0
[  348.171096]  [<ffffffff81394c28>] ? SYSC_recvfrom+0x118/0x140
[  348.171938]  [<ffffffff81098e17>] ? ktime_get_ts+0x47/0xe0
[  348.172781]  [<ffffffff811808c2>] ? poll_select_set_timeout+0x72/0x90
[  348.173624]  [<ffffffff81181dcd>] SyS_poll+0x6d/0x100
[  348.174468]  [<ffffffff81499152>] system_call_fastpath+0x16/0x1b
[  348.175314] systemd-sleep   D ffff88082f292d40     0  3262      1 0x00000004
[  348.176167]  ffff8807fc7e1d18 0000000000000086 ffff8807f6fa8000 ffff8807fc7e1fd8
[  348.177024]  ffff8807fc7e1fd8 ffff8807fc7e1fd8 ffff88080350c5f0 ffff8807f6fa8000
[  348.177881]  ffff8807fc7e1d08 ffff8807f6fa8000 ffff8807f6fa8000 ffff8807f6fa8000
[  348.178740] Call Trace:
[  348.179575]  [<ffffffff81496df4>] schedule+0x24/0x70
[  348.180418]  [<ffffffff81097aad>] __refrigerator+0x4d/0x140
[  348.181264]  [<ffffffff810bab88>] ? cgroup_freezing+0x28/0x40
[  348.182109]  [<ffffffff8105b14c>] get_signal_to_deliver+0x5fc/0x650
[  348.182956]  [<ffffffff81075ba4>] ? finish_task_switch+0x44/0xd0
[  348.183803]  [<ffffffff81002392>] do_signal+0x52/0x5a0
[  348.184647]  [<ffffffff8106bda1>] ? remove_wait_queue+0x51/0x60
[  348.185494]  [<ffffffff8104acd3>] ? do_wait+0x123/0x280
[  348.186335]  [<ffffffff8107b27b>] ? wake_up_new_task+0xfb/0x1a0
[  348.187181]  [<ffffffff81002945>] do_notify_resume+0x65/0x80
[  348.188028]  [<ffffffff81049a70>] ? task_stopped_code+0x50/0x50
[  348.188880]  [<ffffffff814993da>] int_signal+0x12/0x17
[  348.189725] pm-hibernate    D ffff88082f3d2d40     0  3264   3262 0x00000004
[  348.190576]  ffff8807fc16bd18 0000000000000086 ffff8807fa2eaea0 ffff8807fc16bfd8
[  348.191439]  ffff8807fc16bfd8 ffff8807fc16bfd8 ffff8808035445f0 ffff8807fa2eaea0
[  348.192307]  ffff8807fc16bd08 ffff8807fa2eaea0 ffff8807fa2eaea0 ffff8807fa2eaea0
[  348.193163] Call Trace:
[  348.194004]  [<ffffffff81496df4>] schedule+0x24/0x70
[  348.194852]  [<ffffffff81097aad>] __refrigerator+0x4d/0x140
[  348.195704]  [<ffffffff810bab88>] ? cgroup_freezing+0x28/0x40
[  348.196554]  [<ffffffff8105b14c>] get_signal_to_deliver+0x5fc/0x650
[  348.197406]  [<ffffffff810801e3>] ? pick_next_task_fair+0x63/0x180
[  348.198258]  [<ffffffff81075ba4>] ? finish_task_switch+0x44/0xd0
[  348.199114]  [<ffffffff81002392>] do_signal+0x52/0x5a0
[  348.199968]  [<ffffffff8106bda1>] ? remove_wait_queue+0x51/0x60
[  348.200817]  [<ffffffff8104acd3>] ? do_wait+0x123/0x280
[  348.201657]  [<ffffffff8107b27b>] ? wake_up_new_task+0xfb/0x1a0
[  348.202494]  [<ffffffff81002945>] do_notify_resume+0x65/0x80
[  348.203331]  [<ffffffff81049a70>] ? task_stopped_code+0x50/0x50
[  348.204166]  [<ffffffff814993da>] int_signal+0x12/0x17
[  348.204999] s2disk          D ffff88082f392d40     0  3483   3264 0x00000000
[  348.205841]  ffff8807fc10fbf8 0000000000000082 ffff8807fa5d0000 ffff8807fc10ffd8
[  348.206699]  ffff8807fc10ffd8 ffff8807fc10ffd8 ffff880803542ea0 ffff8807fa5d0000
[  348.207563]  ffff8807f725de40 ffff8807fa5d0000 ffff8807fa5d0000 ffff8807fa5d0000
[  348.208423] Call Trace:
[  348.209270]  [<ffffffff81496df4>] schedule+0x24/0x70
[  348.210124]  [<ffffffff81097aad>] __refrigerator+0x4d/0x140
[  348.210979]  [<ffffffff810a2ba5>] futex_wait_queue_me+0x125/0x140
[  348.211838]  [<ffffffff810a3381>] futex_wait+0x181/0x290
[  348.212691]  [<ffffffff810a4e3c>] do_futex+0x11c/0xb30
[  348.213545]  [<ffffffff810a58e3>] SyS_futex+0x93/0x1a0
[  348.214389]  [<ffffffff81075ba4>] ? finish_task_switch+0x44/0xd0
[  348.215229]  [<ffffffff81079572>] ? schedule_tail+0x22/0xa0
[  348.216063]  [<ffffffff81499152>] system_call_fastpath+0x16/0x1b
[  348.216907] Sched Debug Version: v0.10, 3.11.0-rc2 #1
[  348.217736] ktime                                   : 347737.351160
[  348.218572] sched_clk                               : 348216.906329
[  348.219406] cpu_clk                                 : 348216.906356
[  348.220232] jiffies                                 : 4295015036
[  348.221053] sched_clock_stable                      : 1
[  348.221871] 
[  348.222674] sysctl_sched
[  348.223478]   .sysctl_sched_latency                    : 24.000000
[  348.345542]   .nr_running                    : 0
[  348.346103]   .load                          : 0
[  348.346660]   .runnable_load_avg             : 0
[  348.347219]   .blocked_load_avg              : 0
[  348.347774]   .tg_load_contrib               : 0
[  348.348327]   .tg_runnable_contrib           : 0
[  348.348878]   .tg_load_avg                   : 0
[  348.349429]   .tg->runnable_avg              : 5
[  348.349985]   .avg->runnable_avg_sum         : 36
[  348.350539]   .avg->runnable_avg_period      : 48194
[  348.351092] 
[  348.351092] rt_rq[1]:/system/systemd-hibernate.service
[  348.352186]   .rt_nr_running                 : 0
[  348.352747]   .rt_throttled                  : 0
[  348.353307]   .rt_time                       : 0.000000
[  348.353867]   .rt_runtime                    : 0.000000
[  348.354421] 
[  348.354421] rt_rq[1]:/system/bluetooth.service
[  348.355507]   .rt_nr_running                 : 0
[  348.356066]   .rt_throttled                  : 0
[  348.356625]   .rt_time                       : 0.000000
[  348.357183]   .rt_runtime                    : 0.000000
[  348.357740] 
[  348.357740] rt_rq[1]:/system/udisks2.service
[  348.358841]   .rt_nr_running                 : 0
[  348.359406]   .rt_throttled                  : 0
[  348.359972]   .rt_time                       : 0.000000
[  348.360540]   .rt_runtime                    : 0.000000
[  348.361108] 
[  348.361108] rt_rq[1]:/system/polkit.service
[  348.362228]   .rt_nr_running                 : 0
[  348.362801]   .rt_throttled                  : 0
[  348.363375]   .rt_time                       : 0.000000
[  348.363948]   .rt_runtime                    : 0.000000
[  348.364521] 
[  348.364521] rt_rq[1]:/system/upower.service
[  348.365605]   .rt_nr_running                 : 0
[  348.366130]   .rt_throttled                  : 0
[  348.366657]   .rt_time                       : 0.000000
[  348.367179]   .rt_runtime                    : 0.000000
[  348.367702] 
[  348.367702] rt_rq[1]:/system/postfix.service/control
[  348.368740]   .rt_nr_running                 : 0
[  348.369269]   .rt_throttled                  : 0
[  348.369799]   .rt_time                       : 0.000000
[  348.370323]   .rt_runtime                    : 0.000000
[  348.370844] 
[  348.370844] rt_rq[1]:/system/sshd.service
[  348.371876]   .rt_nr_running                 : 0
[  348.372408]   .rt_throttled                  : 0
[  348.372946]   .rt_time                       : 0.000000
[  348.373489]   .rt_runtime                    : 0.000000
[  348.374028] 
[  348.374028] rt_rq[1]:/system/postfix.service
[  348.375095]   .rt_nr_running                 : 0
[  348.375641]   .rt_throttled                  : 0
[  348.376184]   .rt_time                       : 0.000000
[  348.376732]   .rt_runtime                    : 0.000000
[  348.377276] 
[  348.377276] rt_rq[1]:/system/cron.service
[  348.378354]   .rt_nr_running                 : 0
[  348.378900]   .rt_throttled                  : 0
[  348.379447]   .rt_time                       : 0.000000
[  348.379994]   .rt_runtime                    : 0.000000
[  348.380541] 
[  348.380541] rt_rq[1]:/system/xdm.service
[  348.381614]   .rt_nr_running                 : 0
[  348.382163]   .rt_throttled                  : 0
[  348.382714]   .rt_time                       : 0.000000
[  348.383264]   .rt_runtime                    : 0.000000
[  348.383811] 
[  348.383811] rt_rq[1]:/system/dbus.service
[  348.384893]   .rt_nr_running                 : 0
[  348.385448]   .rt_throttled                  : 0
[  348.386000]   .rt_time                       : 0.000000
[  348.386554]   .rt_runtime                    : 0.000000
[  348.387105] 
[  348.387105] rt_rq[1]:/system/[email protected]/tty1
[  348.388188]   .rt_nr_running                 : 0
[  348.388735]   .rt_throttled                  : 0
[  348.389284]   .rt_time                       : 0.000000
[  348.389834]   .rt_runtime                    : 0.000000
[  348.390385] 
[  348.390385] rt_rq[1]:/system/[email protected]
[  348.391467]   .rt_nr_running                 : 0
[  348.392015]   .rt_throttled                  : 0
[  348.392562]   .rt_time                       : 0.000000
[  348.393112]   .rt_runtime                    : 0.000000
[  348.393661] 
[  348.393661] rt_rq[1]:/system/systemd-logind.service
[  348.394744]   .rt_nr_running                 : 0
[  348.395291]   .rt_throttled                  : 0
[  348.395836]   .rt_time                       : 0.000000
[  348.396385]   .rt_runtime                    : 0.000000
[  348.396934] 
[  348.396934] rt_rq[1]:/system/rsyslog.service
[  348.398015]   .rt_nr_running                 : 0
[  348.398564]   .rt_throttled                  : 0
[  348.399110]   .rt_time                       : 0.000000
[  348.399659]   .rt_runtime                    : 0.000000
[  348.400211] 
[  348.400211] rt_rq[1]:/system/haveged.service
[  348.401292]   .rt_nr_running                 : 0
[  348.401842]   .rt_throttled                  : 0
[  348.402390]   .rt_time                       : 0.000000
[  348.402940]   .rt_runtime                    : 0.000000
[  348.403487] 
[  348.403487] rt_rq[1]:/system/[email protected]
[  348.404567]   .rt_nr_running                 : 0
[  348.405111]   .rt_throttled                  : 0
[  348.405656]   .rt_time                       : 0.000000
[  348.406202]   .rt_runtime                    : 0.000000
[  348.406750] 
[  348.406750] rt_rq[1]:/system/systemd-udevd.service
[  348.407831]   .rt_nr_running                 : 0
[  348.408379]   .rt_throttled                  : 0
[  348.408926]   .rt_time                       : 0.000000
[  348.409475]   .rt_runtime                    : 0.000000
[  348.410024] 
[  348.410024] rt_rq[1]:/system/systemd-journald.service
[  348.411110]   .rt_nr_running                 : 0
[  348.411659]   .rt_throttled                  : 0
[  348.412207]   .rt_time                       : 0.000000
[  348.412756]   .rt_runtime                    : 0.000000
[  348.413308] 
[  348.413308] rt_rq[1]:/system
[  348.414385]   .rt_nr_running                 : 0
[  348.414931]   .rt_throttled                  : 0
[  348.415473]   .rt_time                       : 0.000000
[  348.416019]   .rt_runtime                    : 0.000000
[  348.416564] 
[  348.416564] rt_rq[1]:/
[  348.417627]   .rt_nr_running                 : 0
[  348.418163]   .rt_throttled                  : 0
[  348.418700]   .rt_time                       : 0.000000
[  348.419240]   .rt_runtime                    : 950.000000
[  348.419782] 
[  348.419782] runnable tasks:
[  348.419782]             task   PID         tree-key  switches  prio     exec-runtime         sum-exec        sum-sleep
[  348.419782] ----------------------------------------------------------------------------------------------------------
[  348.421983] 
[  348.422555] cpu#2, 2394.479 MHz
[  348.423140]   .nr_running                    : 0
[  348.423729]   .load                          : 0
[  348.424313]   .nr_switches                   : 46347
[  348.424893]   .nr_load_updates               : 12960
[  348.425473]   .nr_uninterruptible            : -42
[  348.426053]   .next_balance                  : 4295.014754
[  348.426641]   .curr->pid                     : 0
[  348.427228]   .clock                         : 347934.367461
[  348.427814]   .cpu_load[0]                   : 0
[  348.428394]   .cpu_load[1]                   : 0
[  348.552446]   .rt_nr_running                 : 0
[  348.552973]   .rt_throttled                  : 0
[  348.636728]   .rt_time                       : 0.000000
[  348.554024]   .rt_runtime                    : 0.000000
[  348.560864] 
[  348.560864] rt_rq[3]:/system/postfix.service
[  348.561921]   .rt_nr_running                 : 0
[  348.562464]   .rt_throttled                  : 0
[  348.780964]   .rt_nr_running                 : 0
[  348.563556]   .rt_runtime                    : 0.000000
[  348.564097] 
[  348.564097] rt_rq[3]:/system/cron.service
[  348.566793]   .rt_runtime                    : 0.000000
[  348.567332] 
[  348.567332] rt_rq[3]:/system/xdm.service
[  348.568396]   .rt_nr_running                 : 0
[  348.568941]   .rt_throttled                  : 0
[  348.570035]   .rt_runtime                    : 0.000000
[  348.570581] 
[  348.570581] rt_rq[3]:/system/dbus.service
[  348.571659]   .rt_nr_running                 : 0
[  348.803324]   .cpu_load[4]                   : 0
[  348.573864] 
[  348.573864] rt_rq[3]:/system/[email protected]/tty1
[  348.576024]   .rt_time                       : 0.000000
[  348.576568]   .rt_runtime                    : 0.000000
[  348.578181]   .rt_nr_running                 : 0
[  348.578722]   .rt_throttled                  : 0
[  348.579263]   .rt_time                       : 0.000000
[  348.580349] 
[  348.580349] rt_rq[3]:/system/systemd-logind.service
[  348.583049]   .rt_runtime                    : 0.000000
[  348.583592] 
[  348.583592] rt_rq[3]:/system/rsyslog.service
[  348.586294]   .rt_runtime                    : 0.000000
[  348.586841] 
[  348.586841] rt_rq[3]:/system/haveged.service
[  348.588455]   .rt_throttled                  : 0
[  348.588998]   .rt_time                       : 0.000000
[  348.590087] 
[  348.590087] rt_rq[3]:/system/[email protected]
[  348.591164]   .rt_nr_running                 : 0
[  348.591706]   .rt_throttled                  : 0
[  348.592247]   .rt_time                       : 0.000000
[  348.592791]   .rt_runtime                    : 0.000000
[  348.593335] 
[  348.593335] rt_rq[3]:/system/systemd-udevd.service
[  348.594409]   .rt_nr_running                 : 0
[  348.594956]   .rt_throttled                  : 0
[  348.596046]   .rt_runtime                    : 0.000000
[  348.596592] 
[  348.596592] rt_rq[3]:/system/systemd-journald.service
[  348.597675]   .rt_nr_running                 : 0
[  348.598224]   .rt_throttled                  : 0
[  348.598769]   .rt_time                       : 0.000000
[  348.599317]   .rt_runtime                    : 0.000000
[  348.599868] 
[  348.599868] rt_rq[3]:/system
[  348.600942]   .rt_nr_running                 : 0
[  348.601485]   .rt_throttled                  : 0
[  348.602027]   .rt_time                       : 0.000000
[  348.602571]   .rt_runtime                    : 0.000000
[  348.603115] 
[  348.603115] rt_rq[3]:/
[  348.604181]   .rt_nr_running                 : 0
[  348.604715]   .rt_throttled                  : 0
[  348.605252]   .rt_time                       : 0.000000
[  348.605791]   .rt_runtime                    : 950.000000
[  348.606331] 
[  348.606331] runnable tasks:
[  348.606331]             task   PID         tree-key  switches  prio     exec-runtime         sum-exec        sum-sleep
[  348.606331] ----------------------------------------------------------------------------------------------------------
[  348.609111] cpu#4, 2394.479 MHz
[  348.609693]   .nr_running                    : 0
[  348.610281]   .load                          : 0
[  348.610866]   .nr_switches                   : 26395
[  348.611446]   .nr_load_updates               : 9172
[  348.612608]   .next_balance                  : 4295.011756
[  348.613786]   .clock                         : 348434.750881
[  348.614956]   .cpu_load[1]                   : 0
[  348.615536]   .cpu_load[2]                   : 0
[  348.616690]   .cpu_load[4]                   : 0
[  348.617263]   .yld_count                     : 2324
[  348.618398]   .sched_goidle                  : 11227
[  348.619530]   .ttwu_count                    : 16564
[  348.620668] 
[  348.620668] cfs_rq[4]:/
[  348.621770]   .exec_clock                    : 2510.563054
[  348.624032]   .spread                        : 0.000000
[  348.624597]   .spread0                       : 673.526556
[  348.626834]   .runnable_load_avg             : 0
[  348.627389]   .blocked_load_avg              : 0
[  348.629587]   .tg->runnable_avg              : 4
[  348.630138]   .avg->runnable_avg_sum         : 0
[  348.632321]   .rt_nr_running                 : 0
[  348.633430]   .rt_time                       : 0.000000
[  348.763005]   .rt_runtime                    : 0.000000
[  348.769521]   .rt_runtime                    : 0.000000
[  348.776606] 
[  348.776606] rt_rq[5]:/system/[email protected]
[  348.777683]   .rt_nr_running                 : 0
[  348.782614]   .rt_runtime                    : 0.000000
[  348.785355]   .rt_time                       : 0.000000
[  348.787539]   .rt_nr_running                 : 0
[  348.788081]   .rt_throttled                  : 0
[  348.792389]   .rt_runtime                    : 950.000000
[  348.796880]   .load                          : 0
[  348.797467]   .nr_switches                   : 20163
[  348.963314]   .rt_runtime                    : 0.000000
[  348.966042]   .rt_time                       : 0.000000
[  348.968773]   .rt_throttled                  : 0
[  348.971503]   .rt_nr_running                 : 0
[  348.976949] 
[  348.976949] rt_rq[7]:/
[  348.979618]   .rt_runtime                    : 950.000000
[  348.980156] 
[  348.980156] runnable tasks:
[  348.980156]             task   PID         tree-key  switches  prio     exec-runtime         sum-exec        sum-sleep
[  348.980156] ----------------------------------------------------------------------------------------------------------
[  348.982352] 
Signature · 8 lines
MfG,

Michael Leun

--
To unsubscribe from this list: send the line "unsubscribe linux-nfs" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html
#35 lore raw

Re: 3.11-rc regression bisected: s2disk does not work (was Re: [PATCH v3 13/16] futex: use freezable blocking call)

In reply to Colin Cross
Quoted 58 lines · Colin Cross
Quoted 44 lines · Michael Leun
Quoted 30 lines · On Mon, 22 Jul 2013 16:55:58 -0700 Colin Cross <[email protected]> wrote:
Quoted 6 lines · Michael Leun
Quoted 1 line · On Mon, 6 May 2013 16:50:18 -0700 Colin Cross <[email protected]> wrote:

Avoid waking up every thread sleeping in a futex_wait call during

[...]

Quoted 8 lines

With 3.11-rc s2disk from suspend-utils stopped working: Frozen at displaying 0% of saving image to disk.

echo "1" >/sys/power/state still works.

Bisecting yielded 88c8004fd3a5fdd2378069de86b90b21110d33a4, reverting that from 3.11-rc2 makes s2disk working again.

I think the expanded use of the freezable_* helpers is exposing an existing bug in hibernation. The SNAPSHOT_FREEZE ioctl calls freeze_processes(), which sets the global system_freezing_cnt and pm_freezing. try_to_freeze_tasks then sends every process except current a signal which causes them all to end up in the refrigerator. The current task then returns back to userspace and continues its work to suspend to disk. If that task ever hits a call to try_to_freeze() in the kernel, it will see system_freezing_cnt and pm_freezing=true and freeze, and suspend to disk will hang forever. It could hit try_to_freeze() because of a signal delivered to the task, or from calling any syscall that uses a freezable_* helper like the one I added to sys_futex.

Quoted 4 lines

I think the right solution is to add a flag to the freezing task that marks it unfreezable. I think PF_NOFREEZE would work, although it is normally used on kernel threads, can you see if the attached patch helps?

That patch helps.

BTW, the only machine I can reproduce this bug with is an i7-3630QM notebook. Cannot reproduce on an Core Duo U1400 and cannot reproduce on an i7 M 620.

Quoted 9 lines

Are the sysreq backtraces still wanted? If so, any tip, how I could get them saved?

--
MfG,

Michael Leun

Any chance that the failing machine has threads=y in the suspend.conf file?

Quoted 7 lines

Rafael, it appears that swsusp's suspend.c spawns new threads after calling the SNAPSHOT_FREEZE ioctl. The PF_NOFREEZE (or the new flag) will get copied to those new threads, but nothing will clear the flag.
Should I just assume that the userspace suspend code will kill those
threads before continuing with suspend? Or maybe add a WARN_ON in the kernel if any threads besides current have the new flag set when the suspend ops that assume all of userspace is frozen are called?

Those threads should be killed by user space. They are only spawned for image saving/compression/encryption and should be waited for after that.

Thanks,
Rafael

Signature · 2 lines
I speak only for myself.
Rafael J. Wysocki, Intel Open Source Technology Center.
#36 lore raw

Re: 3.11-rc regression bisected: s2disk does not work (was Re: [PATCH v3 13/16] futex: use freezable blocking call)

In reply to Michael Leun
Quoted 6 lines · Michael Leun
Quoted 1 line · On Mon, 6 May 2013 16:50:18 -0700 Colin Cross…

Avoid waking up every thread sleeping in a futex_wait call during

[...]

Quoted 7 lines

With 3.11-rc s2disk from suspend-utils stopped working: Frozen at displaying 0% of saving image to disk.

echo "1" >/sys/power/state still works.

Bisecting yielded 88c8004fd3a5fdd2378069de86b90b21110d33a4, reverting that from 3.11-rc2 makes s2disk working again.

Would id be possible to get all the backtraces using magic sysrq?

...actually...

I see what could happen. Before, system hibernated in state where all the futexes were unlocked. Now, it can happen that we attempt s2disk with futex held. s2disk should not depend on other parts of userspace, and should not take futexes, but maybe it does...?
Pavel

Signature · 6 lines
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
--
To unsubscribe from this list: send the line "unsubscribe linux-nfs" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html
#37 lore raw

PATCH v3 14/16 nanosleep: use freezable blocking call

In reply to Colin Cross

Avoid waking up every thread sleeping in a nanosleep call during suspend and resume by calling a freezable blocking call. Previous patches modified the freezer to avoid sending wakeups to threads that are blocked in freezable blocking calls.

This call was selected to be converted to a freezable call because it doesn't hold any locks or release any resources when interrupted that might be needed by another freezing task or a kernel driver during suspend, and is a common site where idle userspace tasks are blocked.

---
 kernel/hrtimer.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
kernel/hrtimer.c +2 -1
index 14be27f..e036276 100644
@@ -47,6 +47,7 @@
 #include <linux/sched/sysctl.h>
 #include <linux/sched/rt.h>
 #include <linux/timer.h>
+#include <linux/freezer.h>
 
 #include <asm/uaccess.h>
 
@@ -1525,7 +1526,7 @@ static int __sched do_nanosleep(struct hrtimer_sleeper *t, enum hrtimer_mode mod
 			t->task = NULL;
 
 		if (likely(t->task))
-			schedule();
+			freezable_schedule();
 
 		hrtimer_cancel(&t->timer);
 		mode = HRTIMER_MODE_ABS;
Signature · 1 line
1.8.2.1
#38 lore raw

PATCH v3 15/16 sigtimedwait: use freezable blocking call

In reply to Colin Cross

Avoid waking up every thread sleeping in a sigtimedwait call during suspend and resume by calling a freezable blocking call. Previous patches modified the freezer to avoid sending wakeups to threads that are blocked in freezable blocking calls.

This call was selected to be converted to a freezable call because it doesn't hold any locks or release any resources when interrupted that might be needed by another freezing task or a kernel driver during suspend, and is a common site where idle userspace tasks are blocked.

---
 kernel/signal.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
kernel/signal.c +1 -1
index 598dc06..10a70a0 100644
@@ -2845,7 +2845,7 @@ int do_sigtimedwait(const sigset_t *which, siginfo_t *info,
 		recalc_sigpending();
 		spin_unlock_irq(&tsk->sighand->siglock);
 
-		timeout = schedule_timeout_interruptible(timeout);
+		timeout = freezable_schedule_timeout_interruptible(timeout);
 
 		spin_lock_irq(&tsk->sighand->siglock);
 		__set_task_blocked(tsk, &tsk->real_blocked);
Signature · 1 line
1.8.2.1
#39 lore raw

PATCH v3 16/16 af_unix: use freezable blocking calls in read

In reply to Colin Cross

Avoid waking up every thread sleeping in read call on an AF_UNIX socket during suspend and resume by calling a freezable blocking call. Previous patches modified the freezer to avoid sending wakeups to threads that are blocked in freezable blocking calls.

This call was selected to be converted to a freezable call because it doesn't hold any locks or release any resources when interrupted that might be needed by another freezing task or a kernel driver during suspend, and is a common site where idle userspace tasks are blocked.

---
 net/unix/af_unix.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
net/unix/af_unix.c +2 -1
index 2db702d..2bcac57 100644
@@ -114,6 +114,7 @@
 #include <linux/mount.h>
 #include <net/checksum.h>
 #include <linux/security.h>
+#include <linux/freezer.h>
 
 struct hlist_head unix_socket_table[2 * UNIX_HASH_SIZE];
 EXPORT_SYMBOL_GPL(unix_socket_table);
@@ -1880,7 +1881,7 @@ static long unix_stream_data_wait(struct sock *sk, long timeo)
 
 		set_bit(SOCK_ASYNC_WAITDATA, &sk->sk_socket->flags);
 		unix_state_unlock(sk);
-		timeo = schedule_timeout(timeo);
+		timeo = freezable_schedule_timeout(timeo);
 		unix_state_lock(sk);
 		clear_bit(SOCK_ASYNC_WAITDATA, &sk->sk_socket->flags);
 	}
Signature · 1 line
1.8.2.1
Acked-by on the v2 series
#40 lore raw
In reply to Colin Cross

Hello,

Quoted 5 lines · Colin Cross

On slow cpus the large number of task wakeups and context switches triggered by freezing and thawing tasks can take a significant amount of cpu time. This patch series reduces the amount of work done during freezing tasks by avoiding waking up tasks that are already in a freezable state.

For the whole series,

Thanks a lot!

Signature · 1 line
tejun
#41 lore raw
Quoted 9 lines · Tejun Heo

Hello,

Quoted 5 lines · Colin Cross

On slow cpus the large number of task wakeups and context switches triggered by freezing and thawing tasks can take a significant amount of cpu time. This patch series reduces the amount of work done during freezing tasks by avoiding waking up tasks that are already in a freezable state.

Quoted 5 lines

For the whole series,

Thanks a lot!

All 16 patches queued up as v3.11 material.

Many thanks to everyone involved,
Rafael

Signature · 2 lines
I speak only for myself.
Rafael J. Wysocki, Intel Open Source Technology Center.