Skip to contentSkip to conversation

ocfs2: dlmglue: fix false deadlock caused by clearing UPCONVERT_FINISHING too early

PATCH New Eric Ren 3 participants 8 messages Jan 19, 2016 – Jan 22, 2016 OCFS2

lore
Share & tools
lore

Eric Ren

author
#0 lore raw

PATCH 1/1 ocfs2: dlmglue: fix false deadlock caused by clearing UPCONVERT_FINISHING too early

This problem was introduced by commit a19128260107f951d1b4c421cf98b92f8092b069. OCFS2_LOCK_UPCONVERT_FINISHING is set just before clearing OCFS2_LOCK_BUSY. This will prevent dc thread from downconverting immediately, and let mask-waiters in ->l_mask_waiters list whose requesting level is compatible with ->l_level to take the lock. But if we have two waiters in mw list, the first is to get EX lock, and the second is to to get PR lock. The first may fail to get lock and then clear UPCONVERT_FINISHING. It's too early to clear the flag because this second will be also queued again even if ->l_level is PR. As a result, nobody would kick up dc thread, leaving dlmglue a deadlock until another lockres relative thread wake it up.

More specifically, for example:
On node1, there is thread W1 keeping writing; on node2, there are thread R1 and R2 keeping reading; sure this 3 threads make IO on the same shared file. At a time, node2 is receiving ast(0=>3), followed immediately by a bast requesting EX lock on behave of node1. Then this may happen:

node2:                                          node1:
l_level==3; R1(3); R2(3)                        l_level==3
R1(unlock); R1(3=>5, update atime)              W1(3=>5)

BAST
R2(unlock); AST(3=>0)
R2(0=>3)
BAST
AST(0=>3)
set OCFS2_LOCK_UPCONVERT_FINISHING
clear OCFS2_LOCK_BUSY

                                                W1(3=>5)

BAST
dc thread requeue=yes
R1(clear OCFS2_LOCK_UPCONVERT_FINISHING,wait) R2(wait)
...
dlmglue deadlock util dc thread woken up by others

This fix is to clear OCFS2_LOCK_UPCONVERT_FINISHING util OCFS2_LOCK_BUSY has been cleared and every waiters has been looped.

Signed-off-by Eric Ren <[email protected]>
---
 fs/ocfs2/dlmglue.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
fs/ocfs2/dlmglue.c +2 -2
index f92612e..72f8b6c 100644
@@ -824,6 +824,8 @@ static void lockres_clear_flags(struct ocfs2_lock_res *lockres,
 				unsigned long clear)
 {
 	lockres_set_flags(lockres, lockres->l_flags & ~clear);
+	if(clear & OCFS2_LOCK_BUSY)
+		lockres->l_flags &= ~OCFS2_LOCK_UPCONVERT_FINISHING;
 }
 
 static inline void ocfs2_generic_handle_downconvert_action(struct ocfs2_lock_res *lockres)

@@ -1522,8 +1524,6 @@ update_holders:
 
 	ret = 0;
 unlock:
-	lockres_clear_flags(lockres, OCFS2_LOCK_UPCONVERT_FINISHING);
-
 	spin_unlock_irqrestore(&lockres->l_lock, flags);
 out:
 	/*
Signature · 1 line
2.6.2
author
#1 lore raw

PATCH 1/1 ocfs2: dlmglue: fix false deadlock caused by clearing UPCONVERT_FINISHING too early

Hi,

This fix is wrong, becuase it can ensure waking up every waiter, but cannot guarantee every waiter finish trying its "again" patch in __ocfs2_cluster_lock().

Other solutions now on my mind are:
1. Give every waiter an ID. When clearing OCFS2_LOCK_BUSY, we can record those IDs in an array. Process any waiter in mask-waiter list, remove the ID from the arry if its ID is in the array, util array is empty we can then clear OCFS2_LOCK_UPCONVERT_FINISHING.

I think it's a bad idea. It's inefficient to handle the array and the ID control is another problem.

2. Split mask-waiter list into two lists: one for OCFS2_LOCK_BUSY, and another for OCFS2_LOCK_BLOCKED. When OCFS2_LOCK_BUSY being cleared and OCFS2_LOCK_BLOCKED being set, we should process waiters in BUSY list and move waiters who cannot get the lock into BLOCKED list again. And when OCFS2_LOCK_BLOCKED being cleared and OCFS2_LOCK_BUSY being set, we should do things like that.

But is any chance that both OCFS2_LOCK_BUSY and OCFS2_LOCK_BLOCKED are set at the same time? If not, I prefer this one.

What do you think? Any comment would be appreciated.

Thanks,
Eric

Quoted 67 lines · Eric Ren

This problem was introduced by commit a19128260107f951d1b4c421cf98b92f8092b069. OCFS2_LOCK_UPCONVERT_FINISHING is set just before clearing OCFS2_LOCK_BUSY. This will prevent dc thread from downconverting immediately, and let mask-waiters in ->l_mask_waiters list whose requesting level is compatible with ->l_level to take the lock. But if we have two waiters in mw list, the first is to get EX lock, and the second is to to get PR lock. The first may fail to get lock and then clear UPCONVERT_FINISHING. It's too early to clear the flag because this second will be also queued again even if ->l_level is PR. As a result, nobody would kick up dc thread, leaving dlmglue a deadlock until another lockres relative thread wake it up.

More specifically, for example:
On node1, there is thread W1 keeping writing; on node2, there are thread R1 and R2 keeping reading; sure this 3 threads make IO on the same shared file. At a time, node2 is receiving ast(0=>3), followed immediately by a bast requesting EX lock on behave of node1. Then this may happen:

node2:                                          node1:
l_level==3; R1(3); R2(3)                        l_level==3
R1(unlock); R1(3=>5, update atime)              W1(3=>5)

BAST
R2(unlock); AST(3=>0)
R2(0=>3)
BAST
AST(0=>3)
set OCFS2_LOCK_UPCONVERT_FINISHING
clear OCFS2_LOCK_BUSY

                                                W1(3=>5)

BAST
dc thread requeue=yes
R1(clear OCFS2_LOCK_UPCONVERT_FINISHING,wait) R2(wait)
...
dlmglue deadlock util dc thread woken up by others

This fix is to clear OCFS2_LOCK_UPCONVERT_FINISHING util OCFS2_LOCK_BUSY has been cleared and every waiters has been looped.

Signed-off-by Eric Ren <[email protected]>
---
 fs/ocfs2/dlmglue.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
fs/ocfs2/dlmglue.c +2 -2
index f92612e..72f8b6c 100644
@@ -824,6 +824,8 @@ static void lockres_clear_flags(struct ocfs2_lock_res *lockres,
 				unsigned long clear)
 {
 	lockres_set_flags(lockres, lockres->l_flags & ~clear);
+	if(clear & OCFS2_LOCK_BUSY)
+		lockres->l_flags &= ~OCFS2_LOCK_UPCONVERT_FINISHING;
 }
 
 static inline void ocfs2_generic_handle_downconvert_action(struct ocfs2_lock_res *lockres)

@@ -1522,8 +1524,6 @@ update_holders:
 
 	ret = 0;
 unlock:
-	lockres_clear_flags(lockres, OCFS2_LOCK_UPCONVERT_FINISHING);
-
 	spin_unlock_irqrestore(&lockres->l_lock, flags);
 out:
 	/*
Signature · 1 line
2.6.2
author
#2 lore raw

PATCH 1/1 ocfs2: dlmglue: fix false deadlock caused by clearing UPCONVERT_FINISHING too early

In reply to Eric Ren

Hi,

Very sorry, this fix is wrong, becuase it can ensure waking up every waiter, but cannot guarantee every waiter finish trying its "again" patch in __ocfs2_cluster_lock().

Other solutions now on my mind are:
1. Give every waiter an ID. When clearing OCFS2_LOCK_BUSY, we can record those IDs in an array. Process any waiter in mask-waiter list, remove the ID from the arry if its ID is in the array, util array is empty we can then clear OCFS2_LOCK_UPCONVERT_FINISHING.

I think it's a bad idea. It's inefficient to handle the array and the ID control is another problem.

2. Split mask-waiter list into two lists: one for OCFS2_LOCK_BUSY, and another for OCFS2_LOCK_BLOCKED. When OCFS2_LOCK_BUSY being cleared and OCFS2_LOCK_BLOCKED being set, we should process waiters in BUSY list and move waiters who cannot get the lock into BLOCKED list again. And when OCFS2_LOCK_BLOCKED being cleared and OCFS2_LOCK_BUSY being set, we should do things like that.

But is any chance that both OCFS2_LOCK_BUSY and OCFS2_LOCK_BLOCKED are set at the same time? If not, I prefer this one.

What do you think? Any comment would be appreciated.

Thanks,
Eric

Quoted 79 lines
Quoted 1 line
Quoted 1 line

This problem was introduced by commit a19128260107f951d1b4c421cf98b92f8092b069. OCFS2_LOCK_UPCONVERT_FINISHING is set just before clearing OCFS2_LOCK_BUSY. This
will prevent dc thread from downconverting immediately, and let mask-waiters in
->l_mask_waiters list whose requesting level is compatible with ->l_level to take
the lock. But if we have two waiters in mw list, the first is to get EX lock, and
the second is to to get PR lock. The first may fail to get lock and then clear
UPCONVERT_FINISHING. It's too early to clear the flag because this second will be
also queued again even if ->l_level is PR. As a result, nobody would kick up dc
thread, leaving dlmglue a deadlock until another lockres relative thread wake it
up.

More specifically, for example:
On node1, there is thread W1 keeping writing; on node2, there are thread R1 and
R2 keeping reading; sure this 3 threads make IO on the same shared file. At a
time, node2 is receiving ast(0=>3), followed immediately by a bast requesting EX
lock on behave of node1. Then this may happen:

node2:                                          node1: 
l_level==3; R1(3); R2(3)                        l_level==3 
R1(unlock); R1(3=>5, update atime)              W1(3=>5)

BAST
R2(unlock); AST(3=>0)
R2(0=>3)
BAST
AST(0=>3)
set OCFS2_LOCK_UPCONVERT_FINISHING
clear OCFS2_LOCK_BUSY

                                                W1(3=>5)

BAST
dc thread requeue=yes
R1(clear OCFS2_LOCK_UPCONVERT_FINISHING,wait) R2(wait)
...
dlmglue deadlock util dc thread woken up by others

This fix is to clear OCFS2_LOCK_UPCONVERT_FINISHING util OCFS2_LOCK_BUSY has been cleared and every waiters has been looped.

Signed-off-by Eric Ren <[email protected]>

---

 fs/ocfs2/dlmglue.c | 4 ++-- 
 1 file changed, 2 insertions(+), 2 deletions(-)
fs/ocfs2/dlmglue.c +2 -0
index f92612e..72f8b6c 100644 
@@ -824,6 +824,8 @@ static void lockres_clear_flags(struct ocfs2_lock_res  
 *lockres, 
 				unsigned long clear) 
 { 
 	lockres_set_flags(lockres, lockres->l_flags & ~clear); 
+	if(clear & OCFS2_LOCK_BUSY) 
+		lockres->l_flags &= ~OCFS2_LOCK_UPCONVERT_FINISHING; 
 }

static inline void ocfs2_generic_handle_downconvert_action(struct
ocfs2_lock_res *lockres)

fs/ocfs2/dlmglue.c +0 -2
@@ -1522,8 +1524,6 @@ update_holders: 
 
 	ret = 0; 
 unlock: 
-	lockres_clear_flags(lockres, OCFS2_LOCK_UPCONVERT_FINISHING); 
- 
 	spin_unlock_irqrestore(&lockres->l_lock, flags); 
 out: 
 	/*
#3 lore raw

PATCH 1/1 ocfs2: dlmglue: fix false deadlock caused by clearing UPCONVERT_FINISHING too early

In reply to Eric Ren

Hi Eric,

This patch should fix your issue.
"NFS hangs in __ocfs2_cluster_lock due to race with ocfs2_unblock_lock"

Thanks,
Junxiao.

Quoted 65 lines · Eric Ren

This problem was introduced by commit a19128260107f951d1b4c421cf98b92f8092b069. OCFS2_LOCK_UPCONVERT_FINISHING is set just before clearing OCFS2_LOCK_BUSY. This will prevent dc thread from downconverting immediately, and let mask-waiters in ->l_mask_waiters list whose requesting level is compatible with ->l_level to take the lock. But if we have two waiters in mw list, the first is to get EX lock, and the second is to to get PR lock. The first may fail to get lock and then clear UPCONVERT_FINISHING. It's too early to clear the flag because this second will be also queued again even if ->l_level is PR. As a result, nobody would kick up dc thread, leaving dlmglue a deadlock until another lockres relative thread wake it up.

More specifically, for example:
On node1, there is thread W1 keeping writing; on node2, there are thread R1 and R2 keeping reading; sure this 3 threads make IO on the same shared file. At a time, node2 is receiving ast(0=>3), followed immediately by a bast requesting EX lock on behave of node1. Then this may happen:

node2:                                          node1:
l_level==3; R1(3); R2(3)                        l_level==3
R1(unlock); R1(3=>5, update atime)              W1(3=>5)

BAST
R2(unlock); AST(3=>0)
R2(0=>3)
BAST
AST(0=>3)
set OCFS2_LOCK_UPCONVERT_FINISHING
clear OCFS2_LOCK_BUSY

                                                W1(3=>5)

BAST
dc thread requeue=yes
R1(clear OCFS2_LOCK_UPCONVERT_FINISHING,wait) R2(wait)
...
dlmglue deadlock util dc thread woken up by others

This fix is to clear OCFS2_LOCK_UPCONVERT_FINISHING util OCFS2_LOCK_BUSY has been cleared and every waiters has been looped.

Signed-off-by Eric Ren <[email protected]>
---
 fs/ocfs2/dlmglue.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
fs/ocfs2/dlmglue.c +2 -2
index f92612e..72f8b6c 100644
@@ -824,6 +824,8 @@ static void lockres_clear_flags(struct ocfs2_lock_res *lockres,
 				unsigned long clear)
 {
 	lockres_set_flags(lockres, lockres->l_flags & ~clear);
+	if(clear & OCFS2_LOCK_BUSY)
+		lockres->l_flags &= ~OCFS2_LOCK_UPCONVERT_FINISHING;
 }
 
 static inline void ocfs2_generic_handle_downconvert_action(struct ocfs2_lock_res *lockres)

@@ -1522,8 +1524,6 @@ update_holders:
 
 	ret = 0;
 unlock:
-	lockres_clear_flags(lockres, OCFS2_LOCK_UPCONVERT_FINISHING);
-
 	spin_unlock_irqrestore(&lockres->l_lock, flags);
 out:
 	/*
author
#4 lore raw

PATCH 1/1 ocfs2: dlmglue: fix false deadlock caused by clearing UPCONVERT_FINISHING too early

Hi Junxiao,

Quoted 4 lines · Junxiao Bi

Hi Eric,

This patch should fix your issue.
"NFS hangs in __ocfs2_cluster_lock due to race with ocfs2_unblock_lock"

Thanks a lot for bringing up this patch! It hasn't been merged into mainline( at least 4.4), right?

I have found this patch in maillist and it looks good! I'd like to test it right now and give feadback!

Thanks again,
Eric

Quoted 71 lines

Thanks,
Junxiao.

Quoted 65 lines · Eric Ren

This problem was introduced by commit a19128260107f951d1b4c421cf98b92f8092b069. OCFS2_LOCK_UPCONVERT_FINISHING is set just before clearing OCFS2_LOCK_BUSY. This will prevent dc thread from downconverting immediately, and let mask-waiters in ->l_mask_waiters list whose requesting level is compatible with ->l_level to take the lock. But if we have two waiters in mw list, the first is to get EX lock, and the second is to to get PR lock. The first may fail to get lock and then clear UPCONVERT_FINISHING. It's too early to clear the flag because this second will be also queued again even if ->l_level is PR. As a result, nobody would kick up dc thread, leaving dlmglue a deadlock until another lockres relative thread wake it up.

More specifically, for example:
On node1, there is thread W1 keeping writing; on node2, there are thread R1 and R2 keeping reading; sure this 3 threads make IO on the same shared file. At a time, node2 is receiving ast(0=>3), followed immediately by a bast requesting EX lock on behave of node1. Then this may happen:

node2:                                          node1:
l_level==3; R1(3); R2(3)                        l_level==3
R1(unlock); R1(3=>5, update atime)              W1(3=>5)

BAST
R2(unlock); AST(3=>0)
R2(0=>3)
BAST
AST(0=>3)
set OCFS2_LOCK_UPCONVERT_FINISHING
clear OCFS2_LOCK_BUSY

                                                W1(3=>5)

BAST
dc thread requeue=yes
R1(clear OCFS2_LOCK_UPCONVERT_FINISHING,wait) R2(wait)
...
dlmglue deadlock util dc thread woken up by others

This fix is to clear OCFS2_LOCK_UPCONVERT_FINISHING util OCFS2_LOCK_BUSY has been cleared and every waiters has been looped.

Signed-off-by Eric Ren <[email protected]>
---
 fs/ocfs2/dlmglue.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
fs/ocfs2/dlmglue.c +2 -2
index f92612e..72f8b6c 100644
@@ -824,6 +824,8 @@ static void lockres_clear_flags(struct ocfs2_lock_res *lockres,
 				unsigned long clear)
 {
 	lockres_set_flags(lockres, lockres->l_flags & ~clear);
+	if(clear & OCFS2_LOCK_BUSY)
+		lockres->l_flags &= ~OCFS2_LOCK_UPCONVERT_FINISHING;
 }
 
 static inline void ocfs2_generic_handle_downconvert_action(struct ocfs2_lock_res *lockres)

@@ -1522,8 +1524,6 @@ update_holders:
 
 	ret = 0;
 unlock:
-	lockres_clear_flags(lockres, OCFS2_LOCK_UPCONVERT_FINISHING);
-
 	spin_unlock_irqrestore(&lockres->l_lock, flags);
 out:
 	/*
#5 lore raw

PATCH 1/1 ocfs2: dlmglue: fix false deadlock caused by clearing UPCONVERT_FINISHING too early

Quoted 10 lines · Eric Ren

Hi Junxiao,

Quoted 4 lines · Junxiao Bi

Hi Eric,

This patch should fix your issue.
"NFS hangs in __ocfs2_cluster_lock due to race with ocfs2_unblock_lock"

Thanks a lot for bringing up this patch! It hasn't been merged into mainline( at least 4.4), right?

Right, it is still in linux-next.

Thanks,
Junxiao.

Quoted 78 lines

I have found this patch in maillist and it looks good! I'd like to test it right now and give feadback!

Thanks again,
Eric

Quoted 71 lines

Thanks,
Junxiao.

Quoted 65 lines · Eric Ren

This problem was introduced by commit a19128260107f951d1b4c421cf98b92f8092b069. OCFS2_LOCK_UPCONVERT_FINISHING is set just before clearing OCFS2_LOCK_BUSY. This will prevent dc thread from downconverting immediately, and let mask-waiters in ->l_mask_waiters list whose requesting level is compatible with ->l_level to take the lock. But if we have two waiters in mw list, the first is to get EX lock, and the second is to to get PR lock. The first may fail to get lock and then clear UPCONVERT_FINISHING. It's too early to clear the flag because this second will be also queued again even if ->l_level is PR. As a result, nobody would kick up dc thread, leaving dlmglue a deadlock until another lockres relative thread wake it up.

More specifically, for example:
On node1, there is thread W1 keeping writing; on node2, there are thread R1 and R2 keeping reading; sure this 3 threads make IO on the same shared file. At a time, node2 is receiving ast(0=>3), followed immediately by a bast requesting EX lock on behave of node1. Then this may happen:

node2:                                          node1:
l_level==3; R1(3); R2(3)                        l_level==3
R1(unlock); R1(3=>5, update atime)              W1(3=>5)

BAST
R2(unlock); AST(3=>0)
R2(0=>3)
BAST
AST(0=>3)
set OCFS2_LOCK_UPCONVERT_FINISHING
clear OCFS2_LOCK_BUSY

                                                W1(3=>5)

BAST
dc thread requeue=yes
R1(clear OCFS2_LOCK_UPCONVERT_FINISHING,wait) R2(wait)
...
dlmglue deadlock util dc thread woken up by others

This fix is to clear OCFS2_LOCK_UPCONVERT_FINISHING util OCFS2_LOCK_BUSY has been cleared and every waiters has been looped.

Signed-off-by Eric Ren <[email protected]>
---
 fs/ocfs2/dlmglue.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
fs/ocfs2/dlmglue.c +2 -2
index f92612e..72f8b6c 100644
@@ -824,6 +824,8 @@ static void lockres_clear_flags(struct ocfs2_lock_res *lockres,
 				unsigned long clear)
 {
 	lockres_set_flags(lockres, lockres->l_flags & ~clear);
+	if(clear & OCFS2_LOCK_BUSY)
+		lockres->l_flags &= ~OCFS2_LOCK_UPCONVERT_FINISHING;
 }
 
 static inline void ocfs2_generic_handle_downconvert_action(struct ocfs2_lock_res *lockres)

@@ -1522,8 +1524,6 @@ update_holders:
 
 	ret = 0;
 unlock:
-	lockres_clear_flags(lockres, OCFS2_LOCK_UPCONVERT_FINISHING);
-
 	spin_unlock_irqrestore(&lockres->l_lock, flags);
 out:
 	/*
#6 lore raw

PATCH 1/1 ocfs2: dlmglue: fix false deadlock caused by clearing UPCONVERT_FINISHING too early

Quoted 4 lines · On Thu, 21 Jan 2016 16:18:38 +0800 Junxiao Bi <[email protected]> wrote:
Quoted 3 lines · Eric Ren

Hi Junxiao,

On Thu, Jan 21, 2016 at 03:10:20PM +0800, Junxiao Bi wrote:

Quoted 8 lines
Quoted 7 lines
Quoted 4 lines

Hi Eric,

This patch should fix your issue.
"NFS hangs in __ocfs2_cluster_lock due to race with ocfs2_unblock_lock"

Thanks a lot for bringing up this patch! It hasn't been merged into mainline( at least 4.4), right?

Right, it is still in linux-next.

I'll be sending it to Linus today.

author
#7 lore raw

PATCH 1/1 ocfs2: dlmglue: fix false deadlock caused by clearing UPCONVERT_FINISHING too early

Hi all,

Quoted 8 lines · Andrew Morton
Quoted 6 lines · On Thu, 21 Jan 2016 16:18:38 +0800 Junxiao Bi <[email protected]> wrote:
Quoted 5 lines · Eric Ren

Hi Junxiao,

Quoted 2 lines · Junxiao Bi

Hi Eric,

Quoted 8 lines
Quoted 6 lines
Quoted 5 lines
Quoted 2 lines

This patch should fix your issue.
"NFS hangs in __ocfs2_cluster_lock due to race with ocfs2_unblock_lock"

Thanks a lot for bringing up this patch! It hasn't been merged into mainline( at least 4.4), right?

Right, it is still in linux-next.

I'll be sending it to Linus today.

Thanks! This patch can also avoid the deadlock in my case. It make sense into mainline now. But there still remains another problem, that is, it may not be fair enough for the node which have more than one threads resting on mask-waiter list during OCFS2_LOCK_BUSY was set. Now, it only give one of those waiters the chance of retry, then clear UPCONVERT_FINISHING, and then start the downconvert thread.

I think it should be better to let every waiter retry beforing down conversion. However, it's more complex. We could discuss and fix it in a new thread later;-)

THX,
Eric

Quoted 7 lines

_______________________________________________
Ocfs2-devel mailing list
Ocfs2-devel at oss.oracle.com
https://oss.oracle.com/mailman/listinfo/ocfs2-devel