Hi,
Thanks for your comments.
> From: David Teigland [mailto:teigl...@redhat.com]
> Sent: Friday, September 15, 2017 12:14 AM
>
>> On Thu, Sep 14, 2017 at 09:31:59AM +, tsutomu@toshiba.co.jp wrote:
>> Hi,
>>
>> > Thanks for sorting this out. Could we expand this with a description of
>>
Hi,
> Thanks for sorting this out. Could we expand this with a description of
> the externally visible effects from the change?
This fix has no effect except when using DLM_LKF_NODLCKWT flag.
We are using fsdlm with ocfs2. Currently, DLM_LKF_NODLCKWT flag is used
only by ocfs2. ocfs2 does not ex
In the current implementation, we think that exclusion control
for othercon in tcp_accept_from_sock() and sctp_accept_from_sock()
was not enough. We fix them.
Signed-off-by: Tadashi Miyauchi
Signed-off-by: Tsutomu Owa
---
fs/dlm/lowcomms.c | 6 ++
1 file changed, 6 insertions(+)
diff --git
When an error occurs in kernel_recvmsg or kernel_sendpage and
close_connection is called and receive work is already scheduled,
receive work is canceled. In that case, the receive work will not
be scheduled forever after reconnection, because CF_READ_PENDING
flag is established.
Signed-off-by: Tad
CF_WRITE_PENDING flag has been reanimated to make dlm_send stop properly
when running dlm_lowcomms_stop.
Signed-off-by: Tadashi Miyauchi
Signed-off-by: Tsutomu Owa
---
fs/dlm/lowcomms.c | 9 -
1 file changed, 8 insertions(+), 1 deletion(-)
diff --git a/fs/dlm/lowcomms.c b/fs/dlm/lowcom
If a node sends a DLM_RCOM_STATUS command and an error occurs on the
receiving side, the DLM_RCOM_STATUS_REPLY response may not be returned.
We retransmitted the DLM_RCOM_STATUS command so that we do not wait for
an infinite response.
Signed-off-by: Tadashi Miyauchi
Signed-off-by: Tsutomu Owa
--
---
This version implements Steve Whitehouse's suggestion to put
cond_resched() after the queue_work in function send_to_sock.
This was just a thinko; it does make more sense to do it afterward.
Before this patch the CF_WRITE_PENDING flag was used to indicate
when writes to the socket were pending
When kernel_sendpage(in send_to_sock) and kernel_recvmsg
(in receive_from_sock) return error, close_connection may works at the
same time. At that time, they may wait for each other by cancel_work_sync.
Signed-off-by: Tadashi Miyauchi
Signed-off-by: Tsutomu Owa
---
fs/dlm/lowcomms.c | 8 ++-
When the DLM_LKF_NODLCKWT flag was set, even if conversion deadlock
was detected, the caller of can_be_granted() was unknown.
We change the behavior of can_be_granted() and change it to detect
conversion deadlock regardless of whether the DLM_LKF_NODLCKWT flag
is set or not. And depending on whethe
The sk member of the socket generated by sock_create_kern() is overwritten
by ops->accept(). So the previous sk will not be released.
We use kernel_accept() instead of sock_create_kern() and ops->accept().
Signed-off-by: Tadashi Miyauchi
Signed-off-by: Tsutomu Owa
---
fs/dlm/lowcomms.c | 21 +++
The writequeue and writequeue_lock member of othercon was not initialized.
If lowcomms_state_change() is called from network layer, othercon->swork
may be scheduled. In this case, send_to_sock() will generate a NULL pointer
reference. We avoid this problem by correctly initializing writequeue and
w
dlm_lowcomms_stop() was not functioning properly. Correctly, we have to
wait until all processing is finished with send_workqueue and
recv_workqueue.
This problem causes the following issue. Senario is
1. dlm_send thread:
send_to_sock refers con->writequeue
2. main thread:
dlm_lowcomms_sto
Before this patch, there was a flag in the con structure that was
used to determine whether or not a connect was needed. The bit was
set here and there, and cleared here and there, so it left some
race conditions: the bit was set, work was queued, then the worker
cleared the bit, allowing someone e
In a previous patch I noted that accept() often copies the struct
sock (sk) which overwrites the sock callbacks. However, in testing
we discovered that the dlm connection structures (con) are sometimes
deleted and recreated as connections come and go, and since they're
zeroed out by kmem_cache_zall
When dlm_recoverd_stop() is called between kthread_should_stop() and
set_task_state(TASK_INTERRUPTIBLE), dlm_recoverd will not wake up.
Signed-off-by: Tadashi Miyauchi
Signed-off-by: Tsutomu Owa
---
fs/dlm/recoverd.c | 11 ++-
1 file changed, 10 insertions(+), 1 deletion(-)
diff --git
In the current implementation, we think that exclusion control between
processing to set the callback function to the connection structure and
processing to refer to the connection structure from the callback function
was not enough. We fix them.
Signed-off-by: Tadashi Miyauchi
Signed-off-by: Tsu
save_cb argument is not used. We remove them.
Signed-off-by: Tadashi Miyauchi
Signed-off-by: Tsutomu Owa
---
fs/dlm/lowcomms.c | 16
1 file changed, 8 insertions(+), 8 deletions(-)
diff --git a/fs/dlm/lowcomms.c b/fs/dlm/lowcomms.c
index f5eea54..585a327 100644
--- a/fs/dlm/lo
If an error occurs in the sending / receiving process, if othercon
exists, sending / receiving processing using othercon may also result
in an error. We fix to pre-close othercon as well.
Signed-off-by: Tadashi Miyauchi
Signed-off-by: Tsutomu Owa
---
fs/dlm/lowcomms.c | 4 ++--
1 file changed,
dlm_cb_seq is 64 bits. If dlm_cb_seq overflows and returns to 0,
dlm_rem_lkb_callback() will not work properly.
Signed-off-by: Tadashi Miyauchi
Signed-off-by: Tsutomu Owa
---
fs/dlm/ast.c | 2 ++
1 file changed, 2 insertions(+)
diff --git a/fs/dlm/ast.c b/fs/dlm/ast.c
index 07fed83..562fa8c 10
If reconnection fails while executing dlm_lowcomms_stop,
dlm_send will not stop.
Signed-off-by: Tadashi Miyauchi
Signed-off-by: Tsutomu Owa
---
fs/dlm/lowcomms.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/fs/dlm/lowcomms.c b/fs/dlm/lowcomms.c
index 4b33614..e2067a6 100644
--- a/fs/dlm/
Hi,
This series of patches (2nd version after previous review on August) is to
fix various bugs. This patch set is against the mainline kernel.
We'd like reviewed to make sure those changes are fine.
Patch number 01/02/03 were posted on this mailing list.
However, I've modifed 03 to correct a bu
Hi,
Thanks a lot for your patch. Yes, we're thinking about sending the next
revised patch set
after fixing the [13/17] which has currenly been tested.
We'll re-send the patch set in a few days.
thanks,
-- owa
-Original Message-
From: Bob Peterson [mailto:rpete...@redhat.com]
Sent: Sa
Hi,
> This patch could do with a description, but the addition of locking for
> the callbacks is definitely something that needs doing,
The reason for this fix is not because something went wrong.
In the current implementation, we think that exclusion control between
processing to set the callb
Hi,
>> After looking more closely, this is a subtle form of conversion deadlock,
>> and this exact case is described in the comment here:
thanks, we withdraw this patch for now since we need to look into more.
-- owa
-Original Message-
From: David Teigland [mailto:teigl...@redhat.com]
Hi,
> This works, but have you looked elsewhere in the kernel for kthread
> examples we could copy that do a race-free check without a timeout?
If you refer to other implementations in kernel, the following
modifications may be better.
The important thing is to call kthread_should_stop() after
se
Hi, thank you for your review.
> I'm curious to know more about this case. Could you describe a little
> more about what happens?
Errors (sk_err=104) occurred in fs/dlm/lowcomms.c::lowcomms_error_report().
In this case, DLM_RCOM_STATUS packet was lost at receiving side, as a result
no DLM_RCOM_S
Hi, thank you for your review.
>> +cond_resched();
>> +queue_work(send_workqueue, &con->swork);
> I think it would make more sense to call cond_resched() after the
> queue_work() since we want the queued work to run soon after it has been
> queued I think,
Well, we're fine with that order
Hi,
Thanks a lot for your quick response and comment.
We hope to make corrections better and incorporate them into mainline by
proceeding
with the review.
The primary test which we've been using for those reported cases is to mount
and umount from multiple nodes at the same time.
i.e.;
mkfs.oc
Signed-off-by: Tadashi Miyauchi
Signed-off-by: Tsutomu Owa
---
fs/dlm/lowcomms.c | 29 +
1 file changed, 17 insertions(+), 12 deletions(-)
diff --git a/fs/dlm/lowcomms.c b/fs/dlm/lowcomms.c
index 3e7c096..f4ba4a8 100644
--- a/fs/dlm/lowcomms.c
+++ b/fs/dlm/lowcomms.c
dlm_cb_seq is 64 bits. If dlm_cb_seq overflows and returns to 0,
dlm_rem_lkb_callback() will not work properly.
Signed-off-by: Tadashi Miyauchi
Signed-off-by: Tsutomu Owa
---
fs/dlm/ast.c | 2 ++
1 file changed, 2 insertions(+)
diff --git a/fs/dlm/ast.c b/fs/dlm/ast.c
index 07fed83..562fa8c 10
When an error occurs in kernel_recvmsg or kernel_sendpage and
close_connection is called and receive work is already scheduled,
receive work is canceled. In that case, the receive work will not
be scheduled forever after reconnection, because CF_READ_PENDING
flag is established.
Signed-off-by: Tad
The sk member of the socket generated by sock_create_kern() is overwritten
by ops->accept(). So the previous sk will not be released.
We use kernel_accept() instead of sock_create_kern() and ops->accept().
Signed-off-by: Tadashi Miyauchi
Signed-off-by: Tsutomu Owa
---
fs/dlm/lowcomms.c | 21 +++
If there is a lock resource conflict on multiple nodes, the lock on
convert queue may not be granted forever.
EX.)
grant queue:
node0 grmode NL / rqmode IV
node1 grmode NL / rqmode IV
convert queue:
node2 grmode NL / rqmode EX
node3 grmode PR / rqmode EX
wait queue:
node4 grm
If reconnection fails while executing dlm_lowcomms_stop,
dlm_send will not stop.
Signed-off-by: Tadashi Miyauchi
Signed-off-by: Tsutomu Owa
---
fs/dlm/lowcomms.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/fs/dlm/lowcomms.c b/fs/dlm/lowcomms.c
index 438fa92..74ed41e 100644
--- a/fs/dlm/
Signed-off-by: Tadashi Miyauchi
Signed-off-by: Tsutomu Owa
---
fs/dlm/lowcomms.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/fs/dlm/lowcomms.c b/fs/dlm/lowcomms.c
index 4b53c32..76c6cdd 100644
--- a/fs/dlm/lowcomms.c
+++ b/fs/dlm/lowcomms.c
@@ -703,7 +703,7 @@ static
When dlm_recoverd_stop() is called between kthread_should_stop() and
set_task_state(), dlm_recoverd will not wake up.
Signed-off-by: Tadashi Miyauchi
Signed-off-by: Tsutomu Owa
---
fs/dlm/recoverd.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/fs/dlm/recoverd.c b/fs/dlm
CF_WRITE_PENDING flag has been reanimated to make dlm_send stop properly
when running dlm_lowcomms_stop.
Signed-off-by: Tadashi Miyauchi
Signed-off-by: Tsutomu Owa
---
fs/dlm/lowcomms.c | 9 -
1 file changed, 8 insertions(+), 1 deletion(-)
diff --git a/fs/dlm/lowcomms.c b/fs/dlm/lowcom
If a node sends a DLM_RCOM_STATUS command and an error occurs on the
receiving side, the DLM_RCOM_STATUS_REPLY response may not be returned.
We retransmitted the DLM_RCOM_STATUS command so that we do not wait for
an infinite response.
Signed-off-by: Tadashi Miyauchi
Signed-off-by: Tsutomu Owa
--
dlm_lowcomms_stop() was not functioning properly. Correctly, we have to
wait until all processing is finished with send_workqueue and
recv_workqueue.
This problem causes the following issue. Scenario is
1. dlm_send thread:
send_to_sock refers con->writequeue
2. main thread:
dlm_lowcomms_st
Signed-off-by: Tadashi Miyauchi
Signed-off-by: Tsutomu Owa
---
fs/dlm/lowcomms.c | 6 ++
1 file changed, 6 insertions(+)
diff --git a/fs/dlm/lowcomms.c b/fs/dlm/lowcomms.c
index aec3c59..4b53c32 100644
--- a/fs/dlm/lowcomms.c
+++ b/fs/dlm/lowcomms.c
@@ -801,16 +801,19 @@ static int tcp_acce
When kernel_sendpage (in send_to_sock) and kernel_recvmsg
(in receive_from_sock) return error, close_connection may works at the
same time. At that time, they may wait for each other by cancel_work_sync.
Signed-off-by: Tadashi Miyauchi
Signed-off-by: Tsutomu Owa
---
fs/dlm/lowcomms.c | 8 ++
From: Bob Peterson
Before this patch the CF_WRITE_PENDING flag was used to indicate
when writes to the socket were pending. This caused race conditions
whereby one process set the bit and another cleared it. Instead,
we just check to see if there's anything there to be sent. This
makes the code m
From: Bob Peterson
In a previous patch I noted that accept() often copies the struct
sock (sk) which overwrites the sock callbacks. However, in testing
we discovered that the dlm connection structures (con) are sometimes
deleted and recreated as connections come and go, and since they're
zeroed o
To remove an unused argument
Signed-off-by: Tadashi Miyauchi
Signed-off-by: Tsutomu Owa
---
fs/dlm/lowcomms.c | 16
1 file changed, 8 insertions(+), 8 deletions(-)
diff --git a/fs/dlm/lowcomms.c b/fs/dlm/lowcomms.c
index ede4522..be62cf1 100644
--- a/fs/dlm/lowcomms.c
+++ b/fs
From: Bob Peterson
Before this patch, there was a flag in the con structure that was
used to determine whether or not a connect was needed. The bit was
set here and there, and cleared here and there, so it left some
race conditions: the bit was set, work was queued, then the worker
cleared the bi
Hi,
This series of patches is to fix various bugs. This patch set is against the
mainline kernel.
I'd like reviewed to make sure those changes are fine.
Patch number 01/02/03 were posted on this mailing list by Bob Peterson
.
Patch 03 is modified to correct a bug acquiring rwlock multiple time
46 matches
Mail list logo