linux-stable/fs/dlm/midcomms.c

1515 lines
39 KiB
C
Raw Normal View History

// SPDX-License-Identifier: GPL-2.0-only
/******************************************************************************
*******************************************************************************
**
** Copyright (C) Sistina Software, Inc. 1997-2003 All rights reserved.
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
** Copyright (C) 2004-2021 Red Hat, Inc. All rights reserved.
**
**
*******************************************************************************
******************************************************************************/
/*
* midcomms.c
*
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
* This is the appallingly named "mid-level" comms layer. It takes care about
* deliver an on application layer "reliable" communication above the used
* lowcomms transport layer.
*
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
* How it works:
*
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
* Each nodes keeps track of all send DLM messages in send_queue with a sequence
* number. The receive will send an DLM_ACK message back for every DLM message
* received at the other side. If a reconnect happens in lowcomms we will send
* all unacknowledged dlm messages again. The receiving side might drop any already
* received message by comparing sequence numbers.
*
* How version detection works:
*
* Due the fact that dlm has pre-configured node addresses on every side
* it is in it's nature that every side connects at starts to transmit
* dlm messages which ends in a race. However DLM_RCOM_NAMES, DLM_RCOM_STATUS
* and their replies are the first messages which are exchanges. Due backwards
* compatibility these messages are not covered by the midcomms re-transmission
* layer. These messages have their own re-transmission handling in the dlm
* application layer. The version field of every node will be set on these RCOM
* messages as soon as they arrived and the node isn't yet part of the nodes
* hash. There exists also logic to detect version mismatched if something weird
* going on or the first messages isn't an expected one.
*
* Termination:
*
* The midcomms layer does a 4 way handshake for termination on DLM protocol
* like TCP supports it with half-closed socket support. SCTP doesn't support
* half-closed socket, so we do it on DLM layer. Also socket shutdown() can be
* interrupted by .e.g. tcp reset itself. Additional there exists the othercon
* paradigm in lowcomms which cannot be easily without breaking backwards
* compatibility. A node cannot send anything to another node when a DLM_FIN
* message was send. There exists additional logic to print a warning if
* DLM wants to do it. There exists a state handling like RFC 793 but reduced
* to termination only. The event "member removal event" describes the cluster
* manager removed the node from internal lists, at this point DLM does not
* send any message to the other node. There exists two cases:
*
* 1. The cluster member was removed and we received a FIN
* OR
* 2. We received a FIN but the member was not removed yet
*
* One of these cases will do the CLOSE_WAIT to LAST_ACK change.
*
*
* +---------+
* | CLOSED |
* +---------+
* | add member/receive RCOM version
* | detection msg
* V
* +---------+
* | ESTAB |
* +---------+
* CLOSE | | rcv FIN
* ------- | | -------
* +---------+ snd FIN / \ snd ACK +---------+
* | FIN |<----------------- ------------------>| CLOSE |
* | WAIT-1 |------------------ | WAIT |
* +---------+ rcv FIN \ +---------+
* | rcv ACK of FIN ------- | CLOSE | member
* | -------------- snd ACK | ------- | removal
* V x V snd FIN V event
* +---------+ +---------+ +---------+
* |FINWAIT-2| | CLOSING | | LAST-ACK|
* +---------+ +---------+ +---------+
* | rcv ACK of FIN | rcv ACK of FIN |
* | rcv FIN -------------- | -------------- |
* | ------- x V x V
* \ snd ACK +---------+ +---------+
* ------------------------>| CLOSED | | CLOSED |
* +---------+ +---------+
*
* NOTE: any state can interrupted by midcomms_close() and state will be
* switched to CLOSED in case of fencing. There exists also some timeout
* handling when we receive the version detection RCOM messages which is
* made by observation.
*
* Future improvements:
*
* There exists some known issues/improvements of the dlm handling. Some
* of them should be done in a next major dlm version bump which makes
* it incompatible with previous versions.
*
* Unaligned memory access:
*
* There exists cases when the dlm message buffer length is not aligned
* to 8 byte. However seems nobody detected any problem with it. This
* can be fixed in the next major version bump of dlm.
*
* Version detection:
*
* The version detection and how it's done is related to backwards
* compatibility. There exists better ways to make a better handling.
* However this should be changed in the next major version bump of dlm.
*
* Tail Size checking:
*
* There exists a message tail payload in e.g. DLM_MSG however we don't
* check it against the message length yet regarding to the receive buffer
* length. That need to be validated.
*
* Fencing bad nodes:
*
* At timeout places or weird sequence number behaviours we should send
* a fencing request to the cluster manager.
*/
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
/* Debug switch to enable a 5 seconds sleep waiting of a termination.
* This can be useful to test fencing while termination is running.
* This requires a setup with only gfs2 as dlm user, so that the
* last umount will terminate the connection.
*
* However it became useful to test, while the 5 seconds block in umount
* just press the reset button. In a lot of dropping the termination
* process can could take several seconds.
*/
#define DLM_DEBUG_FENCE_TERMINATION 0
#include <trace/events/dlm.h>
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
#include <net/tcp.h>
#include "dlm_internal.h"
#include "lowcomms.h"
#include "config.h"
#include "memory.h"
#include "lock.h"
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
#include "util.h"
#include "midcomms.h"
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
/* init value for sequence numbers for testing purpose only e.g. overflows */
#define DLM_SEQ_INIT 0
/* 5 seconds wait to sync ending of dlm */
#define DLM_SHUTDOWN_TIMEOUT msecs_to_jiffies(5000)
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
#define DLM_VERSION_NOT_SET 0
fs: dlm: add send ack threshold and append acks to msgs This patch changes the time when we sending an ack back to tell the other side it can free some message because it is arrived on the receiver node, due random reconnects e.g. TCP resets this is handled as well on application layer to not let DLM run into a deadlock state. The current handling has the following problems: 1. We end in situations that we only send an ack back message of 16 bytes out and no other messages. Whereas DLM has logic to combine so much messages as it can in one send() socket call. This behaviour can be discovered by "trace-cmd start -e dlm_recv" and observing the ret field being 16 bytes. 2. When processing of DLM messages will never end because we receive a lot of messages, we will not send an ack back as it happens when the processing loop ends. This patch introduces a likely and unlikely threshold case. The likely case will send an ack back on a transmit path if the threshold is triggered of amount of processed upper layer protocol. This will solve issue 1 because it will be send when another normal DLM message will be sent. It solves issue 2 because it is not part of the processing loop. There is however a unlikely case, the unlikely case has a bigger threshold and will be triggered when we only receive messages and do not sent any message back. This case avoids that the sending node will keep a lot of message for a long time as we send sometimes ack backs to tell the sender to finally release messages. The atomic cmpxchg() is there to provide a atomically ack send with reset of the upper layer protocol delivery counter. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2023-05-29 21:44:40 +00:00
#define DLM_SEND_ACK_BACK_MSG_THRESHOLD 32
#define DLM_RECV_ACK_BACK_MSG_THRESHOLD (DLM_SEND_ACK_BACK_MSG_THRESHOLD * 8)
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
struct midcomms_node {
int nodeid;
uint32_t version;
atomic_t seq_send;
atomic_t seq_next;
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
/* These queues are unbound because we cannot drop any message in dlm.
* We could send a fence signal for a specific node to the cluster
* manager if queues hits some maximum value, however this handling
* not supported yet.
*/
struct list_head send_queue;
spinlock_t send_queue_lock;
atomic_t send_queue_cnt;
#define DLM_NODE_FLAG_CLOSE 1
#define DLM_NODE_FLAG_STOP_TX 2
#define DLM_NODE_FLAG_STOP_RX 3
fs: dlm: add send ack threshold and append acks to msgs This patch changes the time when we sending an ack back to tell the other side it can free some message because it is arrived on the receiver node, due random reconnects e.g. TCP resets this is handled as well on application layer to not let DLM run into a deadlock state. The current handling has the following problems: 1. We end in situations that we only send an ack back message of 16 bytes out and no other messages. Whereas DLM has logic to combine so much messages as it can in one send() socket call. This behaviour can be discovered by "trace-cmd start -e dlm_recv" and observing the ret field being 16 bytes. 2. When processing of DLM messages will never end because we receive a lot of messages, we will not send an ack back as it happens when the processing loop ends. This patch introduces a likely and unlikely threshold case. The likely case will send an ack back on a transmit path if the threshold is triggered of amount of processed upper layer protocol. This will solve issue 1 because it will be send when another normal DLM message will be sent. It solves issue 2 because it is not part of the processing loop. There is however a unlikely case, the unlikely case has a bigger threshold and will be triggered when we only receive messages and do not sent any message back. This case avoids that the sending node will keep a lot of message for a long time as we send sometimes ack backs to tell the sender to finally release messages. The atomic cmpxchg() is there to provide a atomically ack send with reset of the upper layer protocol delivery counter. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2023-05-29 21:44:40 +00:00
atomic_t ulp_delivered;
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
unsigned long flags;
wait_queue_head_t shutdown_wait;
/* dlm tcp termination state */
#define DLM_CLOSED 1
#define DLM_ESTABLISHED 2
#define DLM_FIN_WAIT1 3
#define DLM_FIN_WAIT2 4
#define DLM_CLOSE_WAIT 5
#define DLM_LAST_ACK 6
#define DLM_CLOSING 7
int state;
spinlock_t state_lock;
/* counts how many lockspaces are using this node
* this refcount is necessary to determine if the
* node wants to disconnect.
*/
int users;
/* not protected by srcu, node_hash lifetime */
void *debugfs;
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
struct hlist_node hlist;
struct rcu_head rcu;
};
struct dlm_mhandle {
const union dlm_packet *inner_p;
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
struct midcomms_node *node;
struct dlm_opts *opts;
struct dlm_msg *msg;
bool committed;
uint32_t seq;
void (*ack_rcv)(struct midcomms_node *node);
/* get_mhandle/commit srcu idx exchange */
int idx;
struct list_head list;
struct rcu_head rcu;
};
static struct hlist_head node_hash[CONN_HASH_SIZE];
static DEFINE_SPINLOCK(nodes_lock);
DEFINE_STATIC_SRCU(nodes_srcu);
/* This mutex prevents that midcomms_close() is running while
* stop() or remove(). As I experienced invalid memory access
* behaviours when DLM_DEBUG_FENCE_TERMINATION is enabled and
* resetting machines. I will end in some double deletion in nodes
* datastructure.
*/
static DEFINE_MUTEX(close_lock);
struct kmem_cache *dlm_midcomms_cache_create(void)
{
return kmem_cache_create("dlm_mhandle", sizeof(struct dlm_mhandle),
0, 0, NULL);
}
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
static inline const char *dlm_state_str(int state)
{
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
switch (state) {
case DLM_CLOSED:
return "CLOSED";
case DLM_ESTABLISHED:
return "ESTABLISHED";
case DLM_FIN_WAIT1:
return "FIN_WAIT1";
case DLM_FIN_WAIT2:
return "FIN_WAIT2";
case DLM_CLOSE_WAIT:
return "CLOSE_WAIT";
case DLM_LAST_ACK:
return "LAST_ACK";
case DLM_CLOSING:
return "CLOSING";
default:
return "UNKNOWN";
}
}
const char *dlm_midcomms_state(struct midcomms_node *node)
{
return dlm_state_str(node->state);
}
unsigned long dlm_midcomms_flags(struct midcomms_node *node)
{
return node->flags;
}
int dlm_midcomms_send_queue_cnt(struct midcomms_node *node)
{
return atomic_read(&node->send_queue_cnt);
}
uint32_t dlm_midcomms_version(struct midcomms_node *node)
{
return node->version;
}
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
static struct midcomms_node *__find_node(int nodeid, int r)
{
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
struct midcomms_node *node;
hlist_for_each_entry_rcu(node, &node_hash[r], hlist) {
if (node->nodeid == nodeid)
return node;
}
return NULL;
}
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
static void dlm_mhandle_release(struct rcu_head *rcu)
{
struct dlm_mhandle *mh = container_of(rcu, struct dlm_mhandle, rcu);
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
dlm_lowcomms_put_msg(mh->msg);
dlm_free_mhandle(mh);
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
}
static void dlm_mhandle_delete(struct midcomms_node *node,
struct dlm_mhandle *mh)
{
list_del_rcu(&mh->list);
atomic_dec(&node->send_queue_cnt);
call_rcu(&mh->rcu, dlm_mhandle_release);
}
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
static void dlm_send_queue_flush(struct midcomms_node *node)
{
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
struct dlm_mhandle *mh;
pr_debug("flush midcomms send queue of node %d\n", node->nodeid);
rcu_read_lock();
fs: dlm: parallelize lowcomms socket handling This patch is rework of lowcomms handling, the main goal was here to handle recvmsg() and sendpage() to run parallel. Parallel in two senses: 1. per connection and 2. that recvmsg()/sendpage() doesn't block each other. Currently recvmsg()/sendpage() cannot run parallel because two workqueues "dlm_recv" and "dlm_send" are ordered workqueues. That means only one work item can be executed. The amount of queue items will be increased about the amount of nodes being inside the cluster. The current two workqueues for sending and receiving can also block each other if the same connection is executed at the same time in dlm_recv and dlm_send workqueue because a per connection mutex for the socket handling. To make it more parallel we introduce one "dlm_io" workqueue which is not an ordered workqueue, the amount of workers are not limited. Due per connection flags SEND/RECV pending we schedule workers ordered per connection and per send and receive task. To get rid of the mutex blocking same workers to do socket handling we switched to a semaphore which handles socket operations as read lock and sock releases as write operations, to prevent sock_release() being called while the socket is being used. There might be more optimization removing the semaphore and replacing it with other synchronization mechanism, however due other circumstances e.g. othercon behaviour it seems complicated to doing this change. I added comments to remove the othercon handling and moving to a different synchronization mechanism as this is done. We need to do that to the next dlm major version upgrade because it is not backwards compatible with the current connect mechanism. The processing of dlm messages need to be still handled by a ordered workqueue. An dlm_process ordered workqueue was introduced which gets filled by the receive worker. This is probably the next bottleneck of DLM but the application can't currently parse dlm messages parallel. A comment was introduced to lift the workqueue context of dlm processing in a non-sleepable softirq to get messages processing done fast. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2022-11-17 22:11:57 +00:00
spin_lock_bh(&node->send_queue_lock);
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
list_for_each_entry_rcu(mh, &node->send_queue, list) {
dlm_mhandle_delete(node, mh);
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
}
fs: dlm: parallelize lowcomms socket handling This patch is rework of lowcomms handling, the main goal was here to handle recvmsg() and sendpage() to run parallel. Parallel in two senses: 1. per connection and 2. that recvmsg()/sendpage() doesn't block each other. Currently recvmsg()/sendpage() cannot run parallel because two workqueues "dlm_recv" and "dlm_send" are ordered workqueues. That means only one work item can be executed. The amount of queue items will be increased about the amount of nodes being inside the cluster. The current two workqueues for sending and receiving can also block each other if the same connection is executed at the same time in dlm_recv and dlm_send workqueue because a per connection mutex for the socket handling. To make it more parallel we introduce one "dlm_io" workqueue which is not an ordered workqueue, the amount of workers are not limited. Due per connection flags SEND/RECV pending we schedule workers ordered per connection and per send and receive task. To get rid of the mutex blocking same workers to do socket handling we switched to a semaphore which handles socket operations as read lock and sock releases as write operations, to prevent sock_release() being called while the socket is being used. There might be more optimization removing the semaphore and replacing it with other synchronization mechanism, however due other circumstances e.g. othercon behaviour it seems complicated to doing this change. I added comments to remove the othercon handling and moving to a different synchronization mechanism as this is done. We need to do that to the next dlm major version upgrade because it is not backwards compatible with the current connect mechanism. The processing of dlm messages need to be still handled by a ordered workqueue. An dlm_process ordered workqueue was introduced which gets filled by the receive worker. This is probably the next bottleneck of DLM but the application can't currently parse dlm messages parallel. A comment was introduced to lift the workqueue context of dlm processing in a non-sleepable softirq to get messages processing done fast. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2022-11-17 22:11:57 +00:00
spin_unlock_bh(&node->send_queue_lock);
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
rcu_read_unlock();
}
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
static void midcomms_node_reset(struct midcomms_node *node)
{
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
pr_debug("reset node %d\n", node->nodeid);
atomic_set(&node->seq_next, DLM_SEQ_INIT);
atomic_set(&node->seq_send, DLM_SEQ_INIT);
fs: dlm: add send ack threshold and append acks to msgs This patch changes the time when we sending an ack back to tell the other side it can free some message because it is arrived on the receiver node, due random reconnects e.g. TCP resets this is handled as well on application layer to not let DLM run into a deadlock state. The current handling has the following problems: 1. We end in situations that we only send an ack back message of 16 bytes out and no other messages. Whereas DLM has logic to combine so much messages as it can in one send() socket call. This behaviour can be discovered by "trace-cmd start -e dlm_recv" and observing the ret field being 16 bytes. 2. When processing of DLM messages will never end because we receive a lot of messages, we will not send an ack back as it happens when the processing loop ends. This patch introduces a likely and unlikely threshold case. The likely case will send an ack back on a transmit path if the threshold is triggered of amount of processed upper layer protocol. This will solve issue 1 because it will be send when another normal DLM message will be sent. It solves issue 2 because it is not part of the processing loop. There is however a unlikely case, the unlikely case has a bigger threshold and will be triggered when we only receive messages and do not sent any message back. This case avoids that the sending node will keep a lot of message for a long time as we send sometimes ack backs to tell the sender to finally release messages. The atomic cmpxchg() is there to provide a atomically ack send with reset of the upper layer protocol delivery counter. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2023-05-29 21:44:40 +00:00
atomic_set(&node->ulp_delivered, 0);
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
node->version = DLM_VERSION_NOT_SET;
node->flags = 0;
dlm_send_queue_flush(node);
node->state = DLM_CLOSED;
wake_up(&node->shutdown_wait);
}
static struct midcomms_node *nodeid2node(int nodeid)
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
{
return __find_node(nodeid, nodeid_hash(nodeid));
}
int dlm_midcomms_addr(int nodeid, struct sockaddr_storage *addr, int len)
{
int ret, idx, r = nodeid_hash(nodeid);
struct midcomms_node *node;
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
ret = dlm_lowcomms_addr(nodeid, addr, len);
if (ret)
return ret;
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
idx = srcu_read_lock(&nodes_srcu);
node = __find_node(nodeid, r);
if (node) {
srcu_read_unlock(&nodes_srcu, idx);
return 0;
}
srcu_read_unlock(&nodes_srcu, idx);
node = kmalloc(sizeof(*node), GFP_NOFS);
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
if (!node)
return -ENOMEM;
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
node->nodeid = nodeid;
spin_lock_init(&node->state_lock);
spin_lock_init(&node->send_queue_lock);
atomic_set(&node->send_queue_cnt, 0);
INIT_LIST_HEAD(&node->send_queue);
init_waitqueue_head(&node->shutdown_wait);
node->users = 0;
midcomms_node_reset(node);
spin_lock(&nodes_lock);
hlist_add_head_rcu(&node->hlist, &node_hash[r]);
spin_unlock(&nodes_lock);
node->debugfs = dlm_create_debug_comms_file(nodeid, node);
return 0;
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
}
static int dlm_send_ack(int nodeid, uint32_t seq)
{
int mb_len = sizeof(struct dlm_header);
struct dlm_header *m_header;
struct dlm_msg *msg;
char *ppc;
msg = dlm_lowcomms_new_msg(nodeid, mb_len, GFP_ATOMIC, &ppc,
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
NULL, NULL);
if (!msg)
return -ENOMEM;
m_header = (struct dlm_header *)ppc;
m_header->h_version = cpu_to_le32(DLM_HEADER_MAJOR | DLM_HEADER_MINOR);
m_header->h_nodeid = cpu_to_le32(dlm_our_nodeid());
m_header->h_length = cpu_to_le16(mb_len);
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
m_header->h_cmd = DLM_ACK;
m_header->u.h_seq = cpu_to_le32(seq);
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
dlm_lowcomms_commit_msg(msg);
dlm_lowcomms_put_msg(msg);
return 0;
}
fs: dlm: add send ack threshold and append acks to msgs This patch changes the time when we sending an ack back to tell the other side it can free some message because it is arrived on the receiver node, due random reconnects e.g. TCP resets this is handled as well on application layer to not let DLM run into a deadlock state. The current handling has the following problems: 1. We end in situations that we only send an ack back message of 16 bytes out and no other messages. Whereas DLM has logic to combine so much messages as it can in one send() socket call. This behaviour can be discovered by "trace-cmd start -e dlm_recv" and observing the ret field being 16 bytes. 2. When processing of DLM messages will never end because we receive a lot of messages, we will not send an ack back as it happens when the processing loop ends. This patch introduces a likely and unlikely threshold case. The likely case will send an ack back on a transmit path if the threshold is triggered of amount of processed upper layer protocol. This will solve issue 1 because it will be send when another normal DLM message will be sent. It solves issue 2 because it is not part of the processing loop. There is however a unlikely case, the unlikely case has a bigger threshold and will be triggered when we only receive messages and do not sent any message back. This case avoids that the sending node will keep a lot of message for a long time as we send sometimes ack backs to tell the sender to finally release messages. The atomic cmpxchg() is there to provide a atomically ack send with reset of the upper layer protocol delivery counter. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2023-05-29 21:44:40 +00:00
static void dlm_send_ack_threshold(struct midcomms_node *node,
uint32_t threshold)
{
uint32_t oval, nval;
bool send_ack;
/* let only send one user trigger threshold to send ack back */
do {
oval = atomic_read(&node->ulp_delivered);
send_ack = (oval > threshold);
/* abort if threshold is not reached */
if (!send_ack)
break;
nval = 0;
/* try to reset ulp_delivered counter */
} while (atomic_cmpxchg(&node->ulp_delivered, oval, nval) != oval);
if (send_ack)
dlm_send_ack(node->nodeid, atomic_read(&node->seq_next));
}
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
static int dlm_send_fin(struct midcomms_node *node,
void (*ack_rcv)(struct midcomms_node *node))
{
int mb_len = sizeof(struct dlm_header);
struct dlm_header *m_header;
struct dlm_mhandle *mh;
char *ppc;
mh = dlm_midcomms_get_mhandle(node->nodeid, mb_len, GFP_ATOMIC, &ppc);
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
if (!mh)
return -ENOMEM;
set_bit(DLM_NODE_FLAG_STOP_TX, &node->flags);
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
mh->ack_rcv = ack_rcv;
m_header = (struct dlm_header *)ppc;
m_header->h_version = cpu_to_le32(DLM_HEADER_MAJOR | DLM_HEADER_MINOR);
m_header->h_nodeid = cpu_to_le32(dlm_our_nodeid());
m_header->h_length = cpu_to_le16(mb_len);
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
m_header->h_cmd = DLM_FIN;
pr_debug("sending fin msg to node %d\n", node->nodeid);
dlm_midcomms_commit_mhandle(mh, NULL, 0);
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
return 0;
}
static void dlm_receive_ack(struct midcomms_node *node, uint32_t seq)
{
struct dlm_mhandle *mh;
rcu_read_lock();
list_for_each_entry_rcu(mh, &node->send_queue, list) {
if (before(mh->seq, seq)) {
if (mh->ack_rcv)
mh->ack_rcv(node);
} else {
/* send queue should be ordered */
break;
}
}
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
fs: dlm: parallelize lowcomms socket handling This patch is rework of lowcomms handling, the main goal was here to handle recvmsg() and sendpage() to run parallel. Parallel in two senses: 1. per connection and 2. that recvmsg()/sendpage() doesn't block each other. Currently recvmsg()/sendpage() cannot run parallel because two workqueues "dlm_recv" and "dlm_send" are ordered workqueues. That means only one work item can be executed. The amount of queue items will be increased about the amount of nodes being inside the cluster. The current two workqueues for sending and receiving can also block each other if the same connection is executed at the same time in dlm_recv and dlm_send workqueue because a per connection mutex for the socket handling. To make it more parallel we introduce one "dlm_io" workqueue which is not an ordered workqueue, the amount of workers are not limited. Due per connection flags SEND/RECV pending we schedule workers ordered per connection and per send and receive task. To get rid of the mutex blocking same workers to do socket handling we switched to a semaphore which handles socket operations as read lock and sock releases as write operations, to prevent sock_release() being called while the socket is being used. There might be more optimization removing the semaphore and replacing it with other synchronization mechanism, however due other circumstances e.g. othercon behaviour it seems complicated to doing this change. I added comments to remove the othercon handling and moving to a different synchronization mechanism as this is done. We need to do that to the next dlm major version upgrade because it is not backwards compatible with the current connect mechanism. The processing of dlm messages need to be still handled by a ordered workqueue. An dlm_process ordered workqueue was introduced which gets filled by the receive worker. This is probably the next bottleneck of DLM but the application can't currently parse dlm messages parallel. A comment was introduced to lift the workqueue context of dlm processing in a non-sleepable softirq to get messages processing done fast. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2022-11-17 22:11:57 +00:00
spin_lock_bh(&node->send_queue_lock);
list_for_each_entry_rcu(mh, &node->send_queue, list) {
if (before(mh->seq, seq)) {
dlm_mhandle_delete(node, mh);
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
} else {
/* send queue should be ordered */
break;
}
}
fs: dlm: parallelize lowcomms socket handling This patch is rework of lowcomms handling, the main goal was here to handle recvmsg() and sendpage() to run parallel. Parallel in two senses: 1. per connection and 2. that recvmsg()/sendpage() doesn't block each other. Currently recvmsg()/sendpage() cannot run parallel because two workqueues "dlm_recv" and "dlm_send" are ordered workqueues. That means only one work item can be executed. The amount of queue items will be increased about the amount of nodes being inside the cluster. The current two workqueues for sending and receiving can also block each other if the same connection is executed at the same time in dlm_recv and dlm_send workqueue because a per connection mutex for the socket handling. To make it more parallel we introduce one "dlm_io" workqueue which is not an ordered workqueue, the amount of workers are not limited. Due per connection flags SEND/RECV pending we schedule workers ordered per connection and per send and receive task. To get rid of the mutex blocking same workers to do socket handling we switched to a semaphore which handles socket operations as read lock and sock releases as write operations, to prevent sock_release() being called while the socket is being used. There might be more optimization removing the semaphore and replacing it with other synchronization mechanism, however due other circumstances e.g. othercon behaviour it seems complicated to doing this change. I added comments to remove the othercon handling and moving to a different synchronization mechanism as this is done. We need to do that to the next dlm major version upgrade because it is not backwards compatible with the current connect mechanism. The processing of dlm messages need to be still handled by a ordered workqueue. An dlm_process ordered workqueue was introduced which gets filled by the receive worker. This is probably the next bottleneck of DLM but the application can't currently parse dlm messages parallel. A comment was introduced to lift the workqueue context of dlm processing in a non-sleepable softirq to get messages processing done fast. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2022-11-17 22:11:57 +00:00
spin_unlock_bh(&node->send_queue_lock);
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
rcu_read_unlock();
}
static void dlm_pas_fin_ack_rcv(struct midcomms_node *node)
{
spin_lock(&node->state_lock);
pr_debug("receive passive fin ack from node %d with state %s\n",
node->nodeid, dlm_state_str(node->state));
switch (node->state) {
case DLM_LAST_ACK:
/* DLM_CLOSED */
midcomms_node_reset(node);
break;
case DLM_CLOSED:
/* not valid but somehow we got what we want */
wake_up(&node->shutdown_wait);
break;
default:
spin_unlock(&node->state_lock);
log_print("%s: unexpected state: %d",
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
__func__, node->state);
WARN_ON_ONCE(1);
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
return;
}
spin_unlock(&node->state_lock);
}
static void dlm_receive_buffer_3_2_trace(uint32_t seq,
const union dlm_packet *p)
{
switch (p->header.h_cmd) {
case DLM_MSG:
trace_dlm_recv_message(dlm_our_nodeid(), seq, &p->message);
break;
case DLM_RCOM:
trace_dlm_recv_rcom(dlm_our_nodeid(), seq, &p->rcom);
break;
default:
break;
}
}
static void dlm_midcomms_receive_buffer(const union dlm_packet *p,
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
struct midcomms_node *node,
uint32_t seq)
{
bool is_expected_seq;
uint32_t oval, nval;
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
do {
oval = atomic_read(&node->seq_next);
is_expected_seq = (oval == seq);
if (!is_expected_seq)
break;
nval = oval + 1;
} while (atomic_cmpxchg(&node->seq_next, oval, nval) != oval);
if (is_expected_seq) {
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
switch (p->header.h_cmd) {
case DLM_FIN:
spin_lock(&node->state_lock);
pr_debug("receive fin msg from node %d with state %s\n",
node->nodeid, dlm_state_str(node->state));
switch (node->state) {
case DLM_ESTABLISHED:
dlm_send_ack(node->nodeid, nval);
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
/* passive shutdown DLM_LAST_ACK case 1
* additional we check if the node is used by
* cluster manager events at all.
*/
if (node->users == 0) {
node->state = DLM_LAST_ACK;
pr_debug("switch node %d to state %s case 1\n",
node->nodeid, dlm_state_str(node->state));
set_bit(DLM_NODE_FLAG_STOP_RX, &node->flags);
dlm_send_fin(node, dlm_pas_fin_ack_rcv);
} else {
node->state = DLM_CLOSE_WAIT;
pr_debug("switch node %d to state %s\n",
node->nodeid, dlm_state_str(node->state));
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
}
break;
case DLM_FIN_WAIT1:
dlm_send_ack(node->nodeid, nval);
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
node->state = DLM_CLOSING;
set_bit(DLM_NODE_FLAG_STOP_RX, &node->flags);
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
pr_debug("switch node %d to state %s\n",
node->nodeid, dlm_state_str(node->state));
break;
case DLM_FIN_WAIT2:
dlm_send_ack(node->nodeid, nval);
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
midcomms_node_reset(node);
pr_debug("switch node %d to state %s\n",
node->nodeid, dlm_state_str(node->state));
break;
case DLM_LAST_ACK:
/* probably remove_member caught it, do nothing */
break;
default:
spin_unlock(&node->state_lock);
log_print("%s: unexpected state: %d",
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
__func__, node->state);
WARN_ON_ONCE(1);
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
return;
}
spin_unlock(&node->state_lock);
break;
default:
WARN_ON_ONCE(test_bit(DLM_NODE_FLAG_STOP_RX, &node->flags));
dlm_receive_buffer_3_2_trace(seq, p);
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
dlm_receive_buffer(p, node->nodeid);
fs: dlm: add send ack threshold and append acks to msgs This patch changes the time when we sending an ack back to tell the other side it can free some message because it is arrived on the receiver node, due random reconnects e.g. TCP resets this is handled as well on application layer to not let DLM run into a deadlock state. The current handling has the following problems: 1. We end in situations that we only send an ack back message of 16 bytes out and no other messages. Whereas DLM has logic to combine so much messages as it can in one send() socket call. This behaviour can be discovered by "trace-cmd start -e dlm_recv" and observing the ret field being 16 bytes. 2. When processing of DLM messages will never end because we receive a lot of messages, we will not send an ack back as it happens when the processing loop ends. This patch introduces a likely and unlikely threshold case. The likely case will send an ack back on a transmit path if the threshold is triggered of amount of processed upper layer protocol. This will solve issue 1 because it will be send when another normal DLM message will be sent. It solves issue 2 because it is not part of the processing loop. There is however a unlikely case, the unlikely case has a bigger threshold and will be triggered when we only receive messages and do not sent any message back. This case avoids that the sending node will keep a lot of message for a long time as we send sometimes ack backs to tell the sender to finally release messages. The atomic cmpxchg() is there to provide a atomically ack send with reset of the upper layer protocol delivery counter. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2023-05-29 21:44:40 +00:00
atomic_inc(&node->ulp_delivered);
/* unlikely case to send ack back when we don't transmit */
dlm_send_ack_threshold(node, DLM_RECV_ACK_BACK_MSG_THRESHOLD);
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
break;
}
} else {
/* retry to ack message which we already have by sending back
* current node->seq_next number as ack.
*/
if (seq < oval)
dlm_send_ack(node->nodeid, oval);
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
log_print_ratelimited("ignore dlm msg because seq mismatch, seq: %u, expected: %u, nodeid: %d",
seq, oval, node->nodeid);
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
}
}
static int dlm_opts_check_msglen(const union dlm_packet *p, uint16_t msglen,
int nodeid)
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
{
int len = msglen;
/* we only trust outer header msglen because
* it's checked against receive buffer length.
*/
if (len < sizeof(struct dlm_opts))
return -1;
len -= sizeof(struct dlm_opts);
if (len < le16_to_cpu(p->opts.o_optlen))
return -1;
len -= le16_to_cpu(p->opts.o_optlen);
switch (p->opts.o_nextcmd) {
case DLM_FIN:
if (len < sizeof(struct dlm_header)) {
log_print("fin too small: %d, will skip this message from node %d",
len, nodeid);
return -1;
}
break;
case DLM_MSG:
if (len < sizeof(struct dlm_message)) {
log_print("msg too small: %d, will skip this message from node %d",
msglen, nodeid);
return -1;
}
break;
case DLM_RCOM:
if (len < sizeof(struct dlm_rcom)) {
log_print("rcom msg too small: %d, will skip this message from node %d",
len, nodeid);
return -1;
}
break;
default:
log_print("unsupported o_nextcmd received: %u, will skip this message from node %d",
p->opts.o_nextcmd, nodeid);
return -1;
}
return 0;
}
static void dlm_midcomms_receive_buffer_3_2(const union dlm_packet *p, int nodeid)
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
{
uint16_t msglen = le16_to_cpu(p->header.h_length);
struct midcomms_node *node;
uint32_t seq;
int ret, idx;
idx = srcu_read_lock(&nodes_srcu);
node = nodeid2node(nodeid);
if (WARN_ON_ONCE(!node))
goto out;
switch (node->version) {
case DLM_VERSION_NOT_SET:
node->version = DLM_VERSION_3_2;
wake_up(&node->shutdown_wait);
log_print("version 0x%08x for node %d detected", DLM_VERSION_3_2,
node->nodeid);
spin_lock(&node->state_lock);
switch (node->state) {
case DLM_CLOSED:
node->state = DLM_ESTABLISHED;
pr_debug("switch node %d to state %s\n",
node->nodeid, dlm_state_str(node->state));
break;
default:
break;
}
spin_unlock(&node->state_lock);
break;
case DLM_VERSION_3_2:
break;
default:
log_print_ratelimited("version mismatch detected, assumed 0x%08x but node %d has 0x%08x",
DLM_VERSION_3_2, node->nodeid, node->version);
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
goto out;
}
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
switch (p->header.h_cmd) {
case DLM_RCOM:
/* these rcom message we use to determine version.
* they have their own retransmission handling and
* are the first messages of dlm.
*
* length already checked.
*/
switch (p->rcom.rc_type) {
case cpu_to_le32(DLM_RCOM_NAMES):
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
fallthrough;
case cpu_to_le32(DLM_RCOM_NAMES_REPLY):
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
fallthrough;
case cpu_to_le32(DLM_RCOM_STATUS):
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
fallthrough;
case cpu_to_le32(DLM_RCOM_STATUS_REPLY):
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
break;
default:
log_print("unsupported rcom type received: %u, will skip this message from node %d",
le32_to_cpu(p->rcom.rc_type), nodeid);
goto out;
}
WARN_ON_ONCE(test_bit(DLM_NODE_FLAG_STOP_RX, &node->flags));
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
dlm_receive_buffer(p, nodeid);
break;
case DLM_OPTS:
seq = le32_to_cpu(p->header.u.h_seq);
ret = dlm_opts_check_msglen(p, msglen, nodeid);
if (ret < 0) {
log_print("opts msg too small: %u, will skip this message from node %d",
msglen, nodeid);
goto out;
}
p = (union dlm_packet *)((unsigned char *)p->opts.o_opts +
le16_to_cpu(p->opts.o_optlen));
/* recheck inner msglen just if it's not garbage */
msglen = le16_to_cpu(p->header.h_length);
switch (p->header.h_cmd) {
case DLM_RCOM:
if (msglen < sizeof(struct dlm_rcom)) {
log_print("inner rcom msg too small: %u, will skip this message from node %d",
msglen, nodeid);
goto out;
}
break;
case DLM_MSG:
if (msglen < sizeof(struct dlm_message)) {
log_print("inner msg too small: %u, will skip this message from node %d",
msglen, nodeid);
goto out;
}
break;
case DLM_FIN:
if (msglen < sizeof(struct dlm_header)) {
log_print("inner fin too small: %u, will skip this message from node %d",
msglen, nodeid);
goto out;
}
break;
default:
log_print("unsupported inner h_cmd received: %u, will skip this message from node %d",
msglen, nodeid);
goto out;
}
dlm_midcomms_receive_buffer(p, node, seq);
break;
case DLM_ACK:
seq = le32_to_cpu(p->header.u.h_seq);
dlm_receive_ack(node, seq);
break;
default:
log_print("unsupported h_cmd received: %u, will skip this message from node %d",
p->header.h_cmd, nodeid);
break;
}
out:
srcu_read_unlock(&nodes_srcu, idx);
}
static void dlm_midcomms_receive_buffer_3_1(const union dlm_packet *p, int nodeid)
{
uint16_t msglen = le16_to_cpu(p->header.h_length);
struct midcomms_node *node;
int idx;
idx = srcu_read_lock(&nodes_srcu);
node = nodeid2node(nodeid);
if (WARN_ON_ONCE(!node)) {
srcu_read_unlock(&nodes_srcu, idx);
return;
}
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
switch (node->version) {
case DLM_VERSION_NOT_SET:
node->version = DLM_VERSION_3_1;
wake_up(&node->shutdown_wait);
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
log_print("version 0x%08x for node %d detected", DLM_VERSION_3_1,
node->nodeid);
break;
case DLM_VERSION_3_1:
break;
default:
log_print_ratelimited("version mismatch detected, assumed 0x%08x but node %d has 0x%08x",
DLM_VERSION_3_1, node->nodeid, node->version);
srcu_read_unlock(&nodes_srcu, idx);
return;
}
srcu_read_unlock(&nodes_srcu, idx);
switch (p->header.h_cmd) {
case DLM_RCOM:
/* length already checked */
break;
case DLM_MSG:
if (msglen < sizeof(struct dlm_message)) {
log_print("msg too small: %u, will skip this message from node %d",
msglen, nodeid);
return;
}
break;
default:
log_print("unsupported h_cmd received: %u, will skip this message from node %d",
p->header.h_cmd, nodeid);
return;
}
dlm_receive_buffer(p, nodeid);
}
fs: dlm: parallelize lowcomms socket handling This patch is rework of lowcomms handling, the main goal was here to handle recvmsg() and sendpage() to run parallel. Parallel in two senses: 1. per connection and 2. that recvmsg()/sendpage() doesn't block each other. Currently recvmsg()/sendpage() cannot run parallel because two workqueues "dlm_recv" and "dlm_send" are ordered workqueues. That means only one work item can be executed. The amount of queue items will be increased about the amount of nodes being inside the cluster. The current two workqueues for sending and receiving can also block each other if the same connection is executed at the same time in dlm_recv and dlm_send workqueue because a per connection mutex for the socket handling. To make it more parallel we introduce one "dlm_io" workqueue which is not an ordered workqueue, the amount of workers are not limited. Due per connection flags SEND/RECV pending we schedule workers ordered per connection and per send and receive task. To get rid of the mutex blocking same workers to do socket handling we switched to a semaphore which handles socket operations as read lock and sock releases as write operations, to prevent sock_release() being called while the socket is being used. There might be more optimization removing the semaphore and replacing it with other synchronization mechanism, however due other circumstances e.g. othercon behaviour it seems complicated to doing this change. I added comments to remove the othercon handling and moving to a different synchronization mechanism as this is done. We need to do that to the next dlm major version upgrade because it is not backwards compatible with the current connect mechanism. The processing of dlm messages need to be still handled by a ordered workqueue. An dlm_process ordered workqueue was introduced which gets filled by the receive worker. This is probably the next bottleneck of DLM but the application can't currently parse dlm messages parallel. A comment was introduced to lift the workqueue context of dlm processing in a non-sleepable softirq to get messages processing done fast. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2022-11-17 22:11:57 +00:00
int dlm_validate_incoming_buffer(int nodeid, unsigned char *buf, int len)
{
fs: dlm: rework receive handling This patch reworks the current receive handling of dlm. As I tried to change the send handling to fix reorder issues I took a look into the receive handling and simplified it, it works as the following: Each connection has a preallocated receive buffer with a minimum length of 4096. On receive, the upper layer protocol will process all dlm message until there is not enough data anymore. If there exists "leftover" data at the end of the receive buffer because the dlm message wasn't fully received it will be copied to the begin of the preallocated receive buffer. Next receive more data will be appended to the previous "leftover" data and processing will begin again. This will remove a lot of code of the current mechanism. Inside the processing functionality we will ensure with a memmove() that the dlm message should be memory aligned. To have a dlm message always started at the beginning of the buffer will reduce some amount of memmove() calls because src and dest pointers are the same. The cluster attribute "buffer_size" becomes a new meaning, it's now the size of application layer receive buffer size. If this is changed during runtime the receive buffer will be reallocated. It's important that the receive buffer size has at minimum the size of the maximum possible dlm message size otherwise the received message cannot be placed inside the receive buffer size. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2020-09-24 14:31:26 +00:00
const unsigned char *ptr = buf;
const struct dlm_header *hd;
uint16_t msglen;
fs: dlm: rework receive handling This patch reworks the current receive handling of dlm. As I tried to change the send handling to fix reorder issues I took a look into the receive handling and simplified it, it works as the following: Each connection has a preallocated receive buffer with a minimum length of 4096. On receive, the upper layer protocol will process all dlm message until there is not enough data anymore. If there exists "leftover" data at the end of the receive buffer because the dlm message wasn't fully received it will be copied to the begin of the preallocated receive buffer. Next receive more data will be appended to the previous "leftover" data and processing will begin again. This will remove a lot of code of the current mechanism. Inside the processing functionality we will ensure with a memmove() that the dlm message should be memory aligned. To have a dlm message always started at the beginning of the buffer will reduce some amount of memmove() calls because src and dest pointers are the same. The cluster attribute "buffer_size" becomes a new meaning, it's now the size of application layer receive buffer size. If this is changed during runtime the receive buffer will be reallocated. It's important that the receive buffer size has at minimum the size of the maximum possible dlm message size otherwise the received message cannot be placed inside the receive buffer size. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2020-09-24 14:31:26 +00:00
int ret = 0;
fs: dlm: rework receive handling This patch reworks the current receive handling of dlm. As I tried to change the send handling to fix reorder issues I took a look into the receive handling and simplified it, it works as the following: Each connection has a preallocated receive buffer with a minimum length of 4096. On receive, the upper layer protocol will process all dlm message until there is not enough data anymore. If there exists "leftover" data at the end of the receive buffer because the dlm message wasn't fully received it will be copied to the begin of the preallocated receive buffer. Next receive more data will be appended to the previous "leftover" data and processing will begin again. This will remove a lot of code of the current mechanism. Inside the processing functionality we will ensure with a memmove() that the dlm message should be memory aligned. To have a dlm message always started at the beginning of the buffer will reduce some amount of memmove() calls because src and dest pointers are the same. The cluster attribute "buffer_size" becomes a new meaning, it's now the size of application layer receive buffer size. If this is changed during runtime the receive buffer will be reallocated. It's important that the receive buffer size has at minimum the size of the maximum possible dlm message size otherwise the received message cannot be placed inside the receive buffer size. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2020-09-24 14:31:26 +00:00
while (len >= sizeof(struct dlm_header)) {
hd = (struct dlm_header *)ptr;
/* no message should be more than DLM_MAX_SOCKET_BUFSIZE or
* less than dlm_header size.
*
* Some messages does not have a 8 byte length boundary yet
* which can occur in a unaligned memory access of some dlm
* messages. However this problem need to be fixed at the
* sending side, for now it seems nobody run into architecture
* related issues yet but it slows down some processing.
* Fixing this issue should be scheduled in future by doing
* the next major version bump.
fs: dlm: rework receive handling This patch reworks the current receive handling of dlm. As I tried to change the send handling to fix reorder issues I took a look into the receive handling and simplified it, it works as the following: Each connection has a preallocated receive buffer with a minimum length of 4096. On receive, the upper layer protocol will process all dlm message until there is not enough data anymore. If there exists "leftover" data at the end of the receive buffer because the dlm message wasn't fully received it will be copied to the begin of the preallocated receive buffer. Next receive more data will be appended to the previous "leftover" data and processing will begin again. This will remove a lot of code of the current mechanism. Inside the processing functionality we will ensure with a memmove() that the dlm message should be memory aligned. To have a dlm message always started at the beginning of the buffer will reduce some amount of memmove() calls because src and dest pointers are the same. The cluster attribute "buffer_size" becomes a new meaning, it's now the size of application layer receive buffer size. If this is changed during runtime the receive buffer will be reallocated. It's important that the receive buffer size has at minimum the size of the maximum possible dlm message size otherwise the received message cannot be placed inside the receive buffer size. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2020-09-24 14:31:26 +00:00
*/
msglen = le16_to_cpu(hd->h_length);
if (msglen > DLM_MAX_SOCKET_BUFSIZE ||
msglen < sizeof(struct dlm_header)) {
log_print("received invalid length header: %u from node %d, will abort message parsing",
msglen, nodeid);
fs: dlm: rework receive handling This patch reworks the current receive handling of dlm. As I tried to change the send handling to fix reorder issues I took a look into the receive handling and simplified it, it works as the following: Each connection has a preallocated receive buffer with a minimum length of 4096. On receive, the upper layer protocol will process all dlm message until there is not enough data anymore. If there exists "leftover" data at the end of the receive buffer because the dlm message wasn't fully received it will be copied to the begin of the preallocated receive buffer. Next receive more data will be appended to the previous "leftover" data and processing will begin again. This will remove a lot of code of the current mechanism. Inside the processing functionality we will ensure with a memmove() that the dlm message should be memory aligned. To have a dlm message always started at the beginning of the buffer will reduce some amount of memmove() calls because src and dest pointers are the same. The cluster attribute "buffer_size" becomes a new meaning, it's now the size of application layer receive buffer size. If this is changed during runtime the receive buffer will be reallocated. It's important that the receive buffer size has at minimum the size of the maximum possible dlm message size otherwise the received message cannot be placed inside the receive buffer size. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2020-09-24 14:31:26 +00:00
return -EBADMSG;
}
fs: dlm: rework receive handling This patch reworks the current receive handling of dlm. As I tried to change the send handling to fix reorder issues I took a look into the receive handling and simplified it, it works as the following: Each connection has a preallocated receive buffer with a minimum length of 4096. On receive, the upper layer protocol will process all dlm message until there is not enough data anymore. If there exists "leftover" data at the end of the receive buffer because the dlm message wasn't fully received it will be copied to the begin of the preallocated receive buffer. Next receive more data will be appended to the previous "leftover" data and processing will begin again. This will remove a lot of code of the current mechanism. Inside the processing functionality we will ensure with a memmove() that the dlm message should be memory aligned. To have a dlm message always started at the beginning of the buffer will reduce some amount of memmove() calls because src and dest pointers are the same. The cluster attribute "buffer_size" becomes a new meaning, it's now the size of application layer receive buffer size. If this is changed during runtime the receive buffer will be reallocated. It's important that the receive buffer size has at minimum the size of the maximum possible dlm message size otherwise the received message cannot be placed inside the receive buffer size. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2020-09-24 14:31:26 +00:00
/* caller will take care that leftover
* will be parsed next call with more data
*/
if (msglen > len)
break;
fs: dlm: parallelize lowcomms socket handling This patch is rework of lowcomms handling, the main goal was here to handle recvmsg() and sendpage() to run parallel. Parallel in two senses: 1. per connection and 2. that recvmsg()/sendpage() doesn't block each other. Currently recvmsg()/sendpage() cannot run parallel because two workqueues "dlm_recv" and "dlm_send" are ordered workqueues. That means only one work item can be executed. The amount of queue items will be increased about the amount of nodes being inside the cluster. The current two workqueues for sending and receiving can also block each other if the same connection is executed at the same time in dlm_recv and dlm_send workqueue because a per connection mutex for the socket handling. To make it more parallel we introduce one "dlm_io" workqueue which is not an ordered workqueue, the amount of workers are not limited. Due per connection flags SEND/RECV pending we schedule workers ordered per connection and per send and receive task. To get rid of the mutex blocking same workers to do socket handling we switched to a semaphore which handles socket operations as read lock and sock releases as write operations, to prevent sock_release() being called while the socket is being used. There might be more optimization removing the semaphore and replacing it with other synchronization mechanism, however due other circumstances e.g. othercon behaviour it seems complicated to doing this change. I added comments to remove the othercon handling and moving to a different synchronization mechanism as this is done. We need to do that to the next dlm major version upgrade because it is not backwards compatible with the current connect mechanism. The processing of dlm messages need to be still handled by a ordered workqueue. An dlm_process ordered workqueue was introduced which gets filled by the receive worker. This is probably the next bottleneck of DLM but the application can't currently parse dlm messages parallel. A comment was introduced to lift the workqueue context of dlm processing in a non-sleepable softirq to get messages processing done fast. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2022-11-17 22:11:57 +00:00
ret += msglen;
len -= msglen;
ptr += msglen;
}
return ret;
}
/*
* Called from the low-level comms layer to process a buffer of
* commands.
*/
int dlm_process_incoming_buffer(int nodeid, unsigned char *buf, int len)
{
const unsigned char *ptr = buf;
const struct dlm_header *hd;
uint16_t msglen;
int ret = 0;
while (len >= sizeof(struct dlm_header)) {
hd = (struct dlm_header *)ptr;
msglen = le16_to_cpu(hd->h_length);
if (msglen > len)
break;
switch (hd->h_version) {
case cpu_to_le32(DLM_VERSION_3_1):
dlm_midcomms_receive_buffer_3_1((const union dlm_packet *)ptr, nodeid);
fs: dlm: rework receive handling This patch reworks the current receive handling of dlm. As I tried to change the send handling to fix reorder issues I took a look into the receive handling and simplified it, it works as the following: Each connection has a preallocated receive buffer with a minimum length of 4096. On receive, the upper layer protocol will process all dlm message until there is not enough data anymore. If there exists "leftover" data at the end of the receive buffer because the dlm message wasn't fully received it will be copied to the begin of the preallocated receive buffer. Next receive more data will be appended to the previous "leftover" data and processing will begin again. This will remove a lot of code of the current mechanism. Inside the processing functionality we will ensure with a memmove() that the dlm message should be memory aligned. To have a dlm message always started at the beginning of the buffer will reduce some amount of memmove() calls because src and dest pointers are the same. The cluster attribute "buffer_size" becomes a new meaning, it's now the size of application layer receive buffer size. If this is changed during runtime the receive buffer will be reallocated. It's important that the receive buffer size has at minimum the size of the maximum possible dlm message size otherwise the received message cannot be placed inside the receive buffer size. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2020-09-24 14:31:26 +00:00
break;
case cpu_to_le32(DLM_VERSION_3_2):
dlm_midcomms_receive_buffer_3_2((const union dlm_packet *)ptr, nodeid);
fs: dlm: rework receive handling This patch reworks the current receive handling of dlm. As I tried to change the send handling to fix reorder issues I took a look into the receive handling and simplified it, it works as the following: Each connection has a preallocated receive buffer with a minimum length of 4096. On receive, the upper layer protocol will process all dlm message until there is not enough data anymore. If there exists "leftover" data at the end of the receive buffer because the dlm message wasn't fully received it will be copied to the begin of the preallocated receive buffer. Next receive more data will be appended to the previous "leftover" data and processing will begin again. This will remove a lot of code of the current mechanism. Inside the processing functionality we will ensure with a memmove() that the dlm message should be memory aligned. To have a dlm message always started at the beginning of the buffer will reduce some amount of memmove() calls because src and dest pointers are the same. The cluster attribute "buffer_size" becomes a new meaning, it's now the size of application layer receive buffer size. If this is changed during runtime the receive buffer will be reallocated. It's important that the receive buffer size has at minimum the size of the maximum possible dlm message size otherwise the received message cannot be placed inside the receive buffer size. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2020-09-24 14:31:26 +00:00
break;
default:
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
log_print("received invalid version header: %u from node %d, will skip this message",
le32_to_cpu(hd->h_version), nodeid);
break;
fs: dlm: rework receive handling This patch reworks the current receive handling of dlm. As I tried to change the send handling to fix reorder issues I took a look into the receive handling and simplified it, it works as the following: Each connection has a preallocated receive buffer with a minimum length of 4096. On receive, the upper layer protocol will process all dlm message until there is not enough data anymore. If there exists "leftover" data at the end of the receive buffer because the dlm message wasn't fully received it will be copied to the begin of the preallocated receive buffer. Next receive more data will be appended to the previous "leftover" data and processing will begin again. This will remove a lot of code of the current mechanism. Inside the processing functionality we will ensure with a memmove() that the dlm message should be memory aligned. To have a dlm message always started at the beginning of the buffer will reduce some amount of memmove() calls because src and dest pointers are the same. The cluster attribute "buffer_size" becomes a new meaning, it's now the size of application layer receive buffer size. If this is changed during runtime the receive buffer will be reallocated. It's important that the receive buffer size has at minimum the size of the maximum possible dlm message size otherwise the received message cannot be placed inside the receive buffer size. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2020-09-24 14:31:26 +00:00
}
ret += msglen;
len -= msglen;
fs: dlm: rework receive handling This patch reworks the current receive handling of dlm. As I tried to change the send handling to fix reorder issues I took a look into the receive handling and simplified it, it works as the following: Each connection has a preallocated receive buffer with a minimum length of 4096. On receive, the upper layer protocol will process all dlm message until there is not enough data anymore. If there exists "leftover" data at the end of the receive buffer because the dlm message wasn't fully received it will be copied to the begin of the preallocated receive buffer. Next receive more data will be appended to the previous "leftover" data and processing will begin again. This will remove a lot of code of the current mechanism. Inside the processing functionality we will ensure with a memmove() that the dlm message should be memory aligned. To have a dlm message always started at the beginning of the buffer will reduce some amount of memmove() calls because src and dest pointers are the same. The cluster attribute "buffer_size" becomes a new meaning, it's now the size of application layer receive buffer size. If this is changed during runtime the receive buffer will be reallocated. It's important that the receive buffer size has at minimum the size of the maximum possible dlm message size otherwise the received message cannot be placed inside the receive buffer size. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2020-09-24 14:31:26 +00:00
ptr += msglen;
}
fs: dlm: rework receive handling This patch reworks the current receive handling of dlm. As I tried to change the send handling to fix reorder issues I took a look into the receive handling and simplified it, it works as the following: Each connection has a preallocated receive buffer with a minimum length of 4096. On receive, the upper layer protocol will process all dlm message until there is not enough data anymore. If there exists "leftover" data at the end of the receive buffer because the dlm message wasn't fully received it will be copied to the begin of the preallocated receive buffer. Next receive more data will be appended to the previous "leftover" data and processing will begin again. This will remove a lot of code of the current mechanism. Inside the processing functionality we will ensure with a memmove() that the dlm message should be memory aligned. To have a dlm message always started at the beginning of the buffer will reduce some amount of memmove() calls because src and dest pointers are the same. The cluster attribute "buffer_size" becomes a new meaning, it's now the size of application layer receive buffer size. If this is changed during runtime the receive buffer will be reallocated. It's important that the receive buffer size has at minimum the size of the maximum possible dlm message size otherwise the received message cannot be placed inside the receive buffer size. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2020-09-24 14:31:26 +00:00
return ret;
}
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
void dlm_midcomms_unack_msg_resend(int nodeid)
{
struct midcomms_node *node;
struct dlm_mhandle *mh;
int idx, ret;
idx = srcu_read_lock(&nodes_srcu);
node = nodeid2node(nodeid);
if (WARN_ON_ONCE(!node)) {
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
srcu_read_unlock(&nodes_srcu, idx);
return;
}
/* old protocol, we don't support to retransmit on failure */
switch (node->version) {
case DLM_VERSION_3_2:
break;
default:
srcu_read_unlock(&nodes_srcu, idx);
return;
}
rcu_read_lock();
list_for_each_entry_rcu(mh, &node->send_queue, list) {
if (!mh->committed)
continue;
ret = dlm_lowcomms_resend_msg(mh->msg);
if (!ret)
log_print_ratelimited("retransmit dlm msg, seq %u, nodeid %d",
mh->seq, node->nodeid);
}
rcu_read_unlock();
srcu_read_unlock(&nodes_srcu, idx);
}
static void dlm_fill_opts_header(struct dlm_opts *opts, uint16_t inner_len,
uint32_t seq)
{
opts->o_header.h_cmd = DLM_OPTS;
opts->o_header.h_version = cpu_to_le32(DLM_HEADER_MAJOR | DLM_HEADER_MINOR);
opts->o_header.h_nodeid = cpu_to_le32(dlm_our_nodeid());
opts->o_header.h_length = cpu_to_le16(DLM_MIDCOMMS_OPT_LEN + inner_len);
opts->o_header.u.h_seq = cpu_to_le32(seq);
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
}
static void midcomms_new_msg_cb(void *data)
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
{
struct dlm_mhandle *mh = data;
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
atomic_inc(&mh->node->send_queue_cnt);
fs: dlm: parallelize lowcomms socket handling This patch is rework of lowcomms handling, the main goal was here to handle recvmsg() and sendpage() to run parallel. Parallel in two senses: 1. per connection and 2. that recvmsg()/sendpage() doesn't block each other. Currently recvmsg()/sendpage() cannot run parallel because two workqueues "dlm_recv" and "dlm_send" are ordered workqueues. That means only one work item can be executed. The amount of queue items will be increased about the amount of nodes being inside the cluster. The current two workqueues for sending and receiving can also block each other if the same connection is executed at the same time in dlm_recv and dlm_send workqueue because a per connection mutex for the socket handling. To make it more parallel we introduce one "dlm_io" workqueue which is not an ordered workqueue, the amount of workers are not limited. Due per connection flags SEND/RECV pending we schedule workers ordered per connection and per send and receive task. To get rid of the mutex blocking same workers to do socket handling we switched to a semaphore which handles socket operations as read lock and sock releases as write operations, to prevent sock_release() being called while the socket is being used. There might be more optimization removing the semaphore and replacing it with other synchronization mechanism, however due other circumstances e.g. othercon behaviour it seems complicated to doing this change. I added comments to remove the othercon handling and moving to a different synchronization mechanism as this is done. We need to do that to the next dlm major version upgrade because it is not backwards compatible with the current connect mechanism. The processing of dlm messages need to be still handled by a ordered workqueue. An dlm_process ordered workqueue was introduced which gets filled by the receive worker. This is probably the next bottleneck of DLM but the application can't currently parse dlm messages parallel. A comment was introduced to lift the workqueue context of dlm processing in a non-sleepable softirq to get messages processing done fast. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2022-11-17 22:11:57 +00:00
spin_lock_bh(&mh->node->send_queue_lock);
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
list_add_tail_rcu(&mh->list, &mh->node->send_queue);
fs: dlm: parallelize lowcomms socket handling This patch is rework of lowcomms handling, the main goal was here to handle recvmsg() and sendpage() to run parallel. Parallel in two senses: 1. per connection and 2. that recvmsg()/sendpage() doesn't block each other. Currently recvmsg()/sendpage() cannot run parallel because two workqueues "dlm_recv" and "dlm_send" are ordered workqueues. That means only one work item can be executed. The amount of queue items will be increased about the amount of nodes being inside the cluster. The current two workqueues for sending and receiving can also block each other if the same connection is executed at the same time in dlm_recv and dlm_send workqueue because a per connection mutex for the socket handling. To make it more parallel we introduce one "dlm_io" workqueue which is not an ordered workqueue, the amount of workers are not limited. Due per connection flags SEND/RECV pending we schedule workers ordered per connection and per send and receive task. To get rid of the mutex blocking same workers to do socket handling we switched to a semaphore which handles socket operations as read lock and sock releases as write operations, to prevent sock_release() being called while the socket is being used. There might be more optimization removing the semaphore and replacing it with other synchronization mechanism, however due other circumstances e.g. othercon behaviour it seems complicated to doing this change. I added comments to remove the othercon handling and moving to a different synchronization mechanism as this is done. We need to do that to the next dlm major version upgrade because it is not backwards compatible with the current connect mechanism. The processing of dlm messages need to be still handled by a ordered workqueue. An dlm_process ordered workqueue was introduced which gets filled by the receive worker. This is probably the next bottleneck of DLM but the application can't currently parse dlm messages parallel. A comment was introduced to lift the workqueue context of dlm processing in a non-sleepable softirq to get messages processing done fast. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2022-11-17 22:11:57 +00:00
spin_unlock_bh(&mh->node->send_queue_lock);
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
mh->seq = atomic_fetch_inc(&mh->node->seq_send);
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
}
static struct dlm_msg *dlm_midcomms_get_msg_3_2(struct dlm_mhandle *mh, int nodeid,
int len, gfp_t allocation, char **ppc)
{
struct dlm_opts *opts;
struct dlm_msg *msg;
msg = dlm_lowcomms_new_msg(nodeid, len + DLM_MIDCOMMS_OPT_LEN,
allocation, ppc, midcomms_new_msg_cb, mh);
if (!msg)
return NULL;
opts = (struct dlm_opts *)*ppc;
mh->opts = opts;
/* add possible options here */
dlm_fill_opts_header(opts, len, mh->seq);
*ppc += sizeof(*opts);
mh->inner_p = (const union dlm_packet *)*ppc;
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
return msg;
}
/* avoid false positive for nodes_srcu, unlock happens in
* dlm_midcomms_commit_mhandle which is a must call if success
*/
#ifndef __CHECKER__
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
struct dlm_mhandle *dlm_midcomms_get_mhandle(int nodeid, int len,
gfp_t allocation, char **ppc)
{
struct midcomms_node *node;
struct dlm_mhandle *mh;
struct dlm_msg *msg;
int idx;
idx = srcu_read_lock(&nodes_srcu);
node = nodeid2node(nodeid);
if (WARN_ON_ONCE(!node))
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
goto err;
/* this is a bug, however we going on and hope it will be resolved */
WARN_ON_ONCE(test_bit(DLM_NODE_FLAG_STOP_TX, &node->flags));
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
mh = dlm_allocate_mhandle(allocation);
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
if (!mh)
goto err;
mh->committed = false;
mh->ack_rcv = NULL;
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
mh->idx = idx;
mh->node = node;
switch (node->version) {
case DLM_VERSION_3_1:
msg = dlm_lowcomms_new_msg(nodeid, len, allocation, ppc,
NULL, NULL);
if (!msg) {
dlm_free_mhandle(mh);
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
goto err;
}
break;
case DLM_VERSION_3_2:
/* send ack back if necessary */
dlm_send_ack_threshold(node, DLM_SEND_ACK_BACK_MSG_THRESHOLD);
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
msg = dlm_midcomms_get_msg_3_2(mh, nodeid, len, allocation,
ppc);
if (!msg) {
dlm_free_mhandle(mh);
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
goto err;
}
break;
default:
dlm_free_mhandle(mh);
WARN_ON_ONCE(1);
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
goto err;
}
mh->msg = msg;
/* keep in mind that is a must to call
* dlm_midcomms_commit_msg() which releases
* nodes_srcu using mh->idx which is assumed
* here that the application will call it.
*/
return mh;
err:
srcu_read_unlock(&nodes_srcu, idx);
return NULL;
}
#endif
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
static void dlm_midcomms_commit_msg_3_2_trace(const struct dlm_mhandle *mh,
const void *name, int namelen)
{
switch (mh->inner_p->header.h_cmd) {
case DLM_MSG:
trace_dlm_send_message(mh->node->nodeid, mh->seq,
&mh->inner_p->message,
name, namelen);
break;
case DLM_RCOM:
trace_dlm_send_rcom(mh->node->nodeid, mh->seq,
&mh->inner_p->rcom);
break;
default:
/* nothing to trace */
break;
}
}
static void dlm_midcomms_commit_msg_3_2(struct dlm_mhandle *mh,
const void *name, int namelen)
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
{
/* nexthdr chain for fast lookup */
mh->opts->o_nextcmd = mh->inner_p->header.h_cmd;
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
mh->committed = true;
dlm_midcomms_commit_msg_3_2_trace(mh, name, namelen);
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
dlm_lowcomms_commit_msg(mh->msg);
}
/* avoid false positive for nodes_srcu, lock was happen in
* dlm_midcomms_get_mhandle
*/
#ifndef __CHECKER__
void dlm_midcomms_commit_mhandle(struct dlm_mhandle *mh,
const void *name, int namelen)
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
{
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
switch (mh->node->version) {
case DLM_VERSION_3_1:
srcu_read_unlock(&nodes_srcu, mh->idx);
dlm_lowcomms_commit_msg(mh->msg);
dlm_lowcomms_put_msg(mh->msg);
/* mh is not part of rcu list in this case */
dlm_free_mhandle(mh);
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
break;
case DLM_VERSION_3_2:
fs: dlm: fix use after free in midcomms commit While working on processing dlm message in softirq context I experienced the following KASAN use-after-free warning: [ 151.760477] ================================================================== [ 151.761803] BUG: KASAN: use-after-free in dlm_midcomms_commit_mhandle+0x19d/0x4b0 [ 151.763414] Read of size 4 at addr ffff88811a980c60 by task lock_torture/1347 [ 151.765284] CPU: 7 PID: 1347 Comm: lock_torture Not tainted 6.1.0-rc4+ #2828 [ 151.766778] Hardware name: Red Hat KVM/RHEL-AV, BIOS 1.16.0-3.module+el8.7.0+16134+e5908aa2 04/01/2014 [ 151.768726] Call Trace: [ 151.769277] <TASK> [ 151.769748] dump_stack_lvl+0x5b/0x86 [ 151.770556] print_report+0x180/0x4c8 [ 151.771378] ? kasan_complete_mode_report_info+0x7c/0x1e0 [ 151.772241] ? dlm_midcomms_commit_mhandle+0x19d/0x4b0 [ 151.773069] kasan_report+0x93/0x1a0 [ 151.773668] ? dlm_midcomms_commit_mhandle+0x19d/0x4b0 [ 151.774514] __asan_load4+0x7e/0xa0 [ 151.775089] dlm_midcomms_commit_mhandle+0x19d/0x4b0 [ 151.775890] ? create_message.isra.29.constprop.64+0x57/0xc0 [ 151.776770] send_common+0x19f/0x1b0 [ 151.777342] ? remove_from_waiters+0x60/0x60 [ 151.778017] ? lock_downgrade+0x410/0x410 [ 151.778648] ? __this_cpu_preempt_check+0x13/0x20 [ 151.779421] ? rcu_lockdep_current_cpu_online+0x88/0xc0 [ 151.780292] _convert_lock+0x46/0x150 [ 151.780893] convert_lock+0x7b/0xc0 [ 151.781459] dlm_lock+0x3ac/0x580 [ 151.781993] ? 0xffffffffc0540000 [ 151.782522] ? torture_stop+0x120/0x120 [dlm_locktorture] [ 151.783379] ? dlm_scan_rsbs+0xa70/0xa70 [ 151.784003] ? preempt_count_sub+0xd6/0x130 [ 151.784661] ? is_module_address+0x47/0x70 [ 151.785309] ? torture_stop+0x120/0x120 [dlm_locktorture] [ 151.786166] ? 0xffffffffc0540000 [ 151.786693] ? lockdep_init_map_type+0xc3/0x360 [ 151.787414] ? 0xffffffffc0540000 [ 151.787947] torture_dlm_lock_sync.isra.3+0xe9/0x150 [dlm_locktorture] [ 151.789004] ? torture_stop+0x120/0x120 [dlm_locktorture] [ 151.789858] ? 0xffffffffc0540000 [ 151.790392] ? lock_torture_cleanup+0x20/0x20 [dlm_locktorture] [ 151.791347] ? delay_tsc+0x94/0xc0 [ 151.791898] torture_ex_iter+0xc3/0xea [dlm_locktorture] [ 151.792735] ? torture_start+0x30/0x30 [dlm_locktorture] [ 151.793606] lock_torture+0x177/0x270 [dlm_locktorture] [ 151.794448] ? torture_dlm_lock_sync.isra.3+0x150/0x150 [dlm_locktorture] [ 151.795539] ? lock_torture_stats+0x80/0x80 [dlm_locktorture] [ 151.796476] ? do_raw_spin_lock+0x11e/0x1e0 [ 151.797152] ? mark_held_locks+0x34/0xb0 [ 151.797784] ? _raw_spin_unlock_irqrestore+0x30/0x70 [ 151.798581] ? __kthread_parkme+0x79/0x110 [ 151.799246] ? trace_preempt_on+0x2a/0xf0 [ 151.799902] ? __kthread_parkme+0x79/0x110 [ 151.800579] ? preempt_count_sub+0xd6/0x130 [ 151.801271] ? __kasan_check_read+0x11/0x20 [ 151.801963] ? __kthread_parkme+0xec/0x110 [ 151.802630] ? lock_torture_stats+0x80/0x80 [dlm_locktorture] [ 151.803569] kthread+0x192/0x1d0 [ 151.804104] ? kthread_complete_and_exit+0x30/0x30 [ 151.804881] ret_from_fork+0x1f/0x30 [ 151.805480] </TASK> [ 151.806111] Allocated by task 1347: [ 151.806681] kasan_save_stack+0x26/0x50 [ 151.807308] kasan_set_track+0x25/0x30 [ 151.807920] kasan_save_alloc_info+0x1e/0x30 [ 151.808609] __kasan_slab_alloc+0x63/0x80 [ 151.809263] kmem_cache_alloc+0x1ad/0x830 [ 151.809916] dlm_allocate_mhandle+0x17/0x20 [ 151.810590] dlm_midcomms_get_mhandle+0x96/0x260 [ 151.811344] _create_message+0x95/0x180 [ 151.811994] create_message.isra.29.constprop.64+0x57/0xc0 [ 151.812880] send_common+0x129/0x1b0 [ 151.813467] _convert_lock+0x46/0x150 [ 151.814074] convert_lock+0x7b/0xc0 [ 151.814648] dlm_lock+0x3ac/0x580 [ 151.815199] torture_dlm_lock_sync.isra.3+0xe9/0x150 [dlm_locktorture] [ 151.816258] torture_ex_iter+0xc3/0xea [dlm_locktorture] [ 151.817129] lock_torture+0x177/0x270 [dlm_locktorture] [ 151.817986] kthread+0x192/0x1d0 [ 151.818518] ret_from_fork+0x1f/0x30 [ 151.819369] Freed by task 1336: [ 151.819890] kasan_save_stack+0x26/0x50 [ 151.820514] kasan_set_track+0x25/0x30 [ 151.821128] kasan_save_free_info+0x2e/0x50 [ 151.821812] __kasan_slab_free+0x107/0x1a0 [ 151.822483] kmem_cache_free+0x204/0x5e0 [ 151.823152] dlm_free_mhandle+0x18/0x20 [ 151.823781] dlm_mhandle_release+0x2e/0x40 [ 151.824454] rcu_core+0x583/0x1330 [ 151.825047] rcu_core_si+0xe/0x20 [ 151.825594] __do_softirq+0xf4/0x5c2 [ 151.826450] Last potentially related work creation: [ 151.827238] kasan_save_stack+0x26/0x50 [ 151.827870] __kasan_record_aux_stack+0xa2/0xc0 [ 151.828609] kasan_record_aux_stack_noalloc+0xb/0x20 [ 151.829415] call_rcu+0x4c/0x760 [ 151.829954] dlm_mhandle_delete+0x97/0xb0 [ 151.830718] dlm_process_incoming_buffer+0x2fc/0xb30 [ 151.831524] process_dlm_messages+0x16e/0x470 [ 151.832245] process_one_work+0x505/0xa10 [ 151.832905] worker_thread+0x67/0x650 [ 151.833507] kthread+0x192/0x1d0 [ 151.834046] ret_from_fork+0x1f/0x30 [ 151.834900] The buggy address belongs to the object at ffff88811a980c30 which belongs to the cache dlm_mhandle of size 88 [ 151.836894] The buggy address is located 48 bytes inside of 88-byte region [ffff88811a980c30, ffff88811a980c88) [ 151.839007] The buggy address belongs to the physical page: [ 151.839904] page:0000000076cf5d62 refcount:1 mapcount:0 mapping:0000000000000000 index:0x0 pfn:0x11a980 [ 151.841378] flags: 0x8000000000000200(slab|zone=2) [ 151.842141] raw: 8000000000000200 0000000000000000 dead000000000122 ffff8881089b43c0 [ 151.843401] raw: 0000000000000000 0000000000220022 00000001ffffffff 0000000000000000 [ 151.844640] page dumped because: kasan: bad access detected [ 151.845822] Memory state around the buggy address: [ 151.846602] ffff88811a980b00: fb fb fb fb fc fc fc fc fa fb fb fb fb fb fb fb [ 151.847761] ffff88811a980b80: fb fb fb fc fc fc fc fa fb fb fb fb fb fb fb fb [ 151.848921] >ffff88811a980c00: fb fb fc fc fc fc fa fb fb fb fb fb fb fb fb fb [ 151.850076] ^ [ 151.851085] ffff88811a980c80: fb fc fc fc fc fa fb fb fb fb fb fb fb fb fb fb [ 151.852269] ffff88811a980d00: fc fc fc fc fa fb fb fb fb fb fb fb fb fb fb fc [ 151.853428] ================================================================== [ 151.855618] Disabling lock debugging due to kernel taint It is accessing a mhandle in dlm_midcomms_commit_mhandle() and the mhandle was freed by a call_rcu() call in dlm_process_incoming_buffer(), dlm_mhandle_delete(). It looks like it was freed because an ack of this message was received. There is a short race between committing the dlm message to be transmitted and getting an ack back. If the ack is faster than returning from dlm_midcomms_commit_msg_3_2(), then we run into a use-after free because we still need to reference the mhandle when calling srcu_read_unlock(). To avoid that, we don't allow that mhandle to be freed between dlm_midcomms_commit_msg_3_2() and srcu_read_unlock() by using rcu read lock. We can do that because mhandle is protected by rcu handling. Cc: stable@vger.kernel.org Fixes: 489d8e559c65 ("fs: dlm: add reliable connection if reconnect") Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2023-01-12 22:10:32 +00:00
/* held rcu read lock here, because we sending the
* dlm message out, when we do that we could receive
* an ack back which releases the mhandle and we
* get a use after free.
*/
rcu_read_lock();
dlm_midcomms_commit_msg_3_2(mh, name, namelen);
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
srcu_read_unlock(&nodes_srcu, mh->idx);
fs: dlm: fix use after free in midcomms commit While working on processing dlm message in softirq context I experienced the following KASAN use-after-free warning: [ 151.760477] ================================================================== [ 151.761803] BUG: KASAN: use-after-free in dlm_midcomms_commit_mhandle+0x19d/0x4b0 [ 151.763414] Read of size 4 at addr ffff88811a980c60 by task lock_torture/1347 [ 151.765284] CPU: 7 PID: 1347 Comm: lock_torture Not tainted 6.1.0-rc4+ #2828 [ 151.766778] Hardware name: Red Hat KVM/RHEL-AV, BIOS 1.16.0-3.module+el8.7.0+16134+e5908aa2 04/01/2014 [ 151.768726] Call Trace: [ 151.769277] <TASK> [ 151.769748] dump_stack_lvl+0x5b/0x86 [ 151.770556] print_report+0x180/0x4c8 [ 151.771378] ? kasan_complete_mode_report_info+0x7c/0x1e0 [ 151.772241] ? dlm_midcomms_commit_mhandle+0x19d/0x4b0 [ 151.773069] kasan_report+0x93/0x1a0 [ 151.773668] ? dlm_midcomms_commit_mhandle+0x19d/0x4b0 [ 151.774514] __asan_load4+0x7e/0xa0 [ 151.775089] dlm_midcomms_commit_mhandle+0x19d/0x4b0 [ 151.775890] ? create_message.isra.29.constprop.64+0x57/0xc0 [ 151.776770] send_common+0x19f/0x1b0 [ 151.777342] ? remove_from_waiters+0x60/0x60 [ 151.778017] ? lock_downgrade+0x410/0x410 [ 151.778648] ? __this_cpu_preempt_check+0x13/0x20 [ 151.779421] ? rcu_lockdep_current_cpu_online+0x88/0xc0 [ 151.780292] _convert_lock+0x46/0x150 [ 151.780893] convert_lock+0x7b/0xc0 [ 151.781459] dlm_lock+0x3ac/0x580 [ 151.781993] ? 0xffffffffc0540000 [ 151.782522] ? torture_stop+0x120/0x120 [dlm_locktorture] [ 151.783379] ? dlm_scan_rsbs+0xa70/0xa70 [ 151.784003] ? preempt_count_sub+0xd6/0x130 [ 151.784661] ? is_module_address+0x47/0x70 [ 151.785309] ? torture_stop+0x120/0x120 [dlm_locktorture] [ 151.786166] ? 0xffffffffc0540000 [ 151.786693] ? lockdep_init_map_type+0xc3/0x360 [ 151.787414] ? 0xffffffffc0540000 [ 151.787947] torture_dlm_lock_sync.isra.3+0xe9/0x150 [dlm_locktorture] [ 151.789004] ? torture_stop+0x120/0x120 [dlm_locktorture] [ 151.789858] ? 0xffffffffc0540000 [ 151.790392] ? lock_torture_cleanup+0x20/0x20 [dlm_locktorture] [ 151.791347] ? delay_tsc+0x94/0xc0 [ 151.791898] torture_ex_iter+0xc3/0xea [dlm_locktorture] [ 151.792735] ? torture_start+0x30/0x30 [dlm_locktorture] [ 151.793606] lock_torture+0x177/0x270 [dlm_locktorture] [ 151.794448] ? torture_dlm_lock_sync.isra.3+0x150/0x150 [dlm_locktorture] [ 151.795539] ? lock_torture_stats+0x80/0x80 [dlm_locktorture] [ 151.796476] ? do_raw_spin_lock+0x11e/0x1e0 [ 151.797152] ? mark_held_locks+0x34/0xb0 [ 151.797784] ? _raw_spin_unlock_irqrestore+0x30/0x70 [ 151.798581] ? __kthread_parkme+0x79/0x110 [ 151.799246] ? trace_preempt_on+0x2a/0xf0 [ 151.799902] ? __kthread_parkme+0x79/0x110 [ 151.800579] ? preempt_count_sub+0xd6/0x130 [ 151.801271] ? __kasan_check_read+0x11/0x20 [ 151.801963] ? __kthread_parkme+0xec/0x110 [ 151.802630] ? lock_torture_stats+0x80/0x80 [dlm_locktorture] [ 151.803569] kthread+0x192/0x1d0 [ 151.804104] ? kthread_complete_and_exit+0x30/0x30 [ 151.804881] ret_from_fork+0x1f/0x30 [ 151.805480] </TASK> [ 151.806111] Allocated by task 1347: [ 151.806681] kasan_save_stack+0x26/0x50 [ 151.807308] kasan_set_track+0x25/0x30 [ 151.807920] kasan_save_alloc_info+0x1e/0x30 [ 151.808609] __kasan_slab_alloc+0x63/0x80 [ 151.809263] kmem_cache_alloc+0x1ad/0x830 [ 151.809916] dlm_allocate_mhandle+0x17/0x20 [ 151.810590] dlm_midcomms_get_mhandle+0x96/0x260 [ 151.811344] _create_message+0x95/0x180 [ 151.811994] create_message.isra.29.constprop.64+0x57/0xc0 [ 151.812880] send_common+0x129/0x1b0 [ 151.813467] _convert_lock+0x46/0x150 [ 151.814074] convert_lock+0x7b/0xc0 [ 151.814648] dlm_lock+0x3ac/0x580 [ 151.815199] torture_dlm_lock_sync.isra.3+0xe9/0x150 [dlm_locktorture] [ 151.816258] torture_ex_iter+0xc3/0xea [dlm_locktorture] [ 151.817129] lock_torture+0x177/0x270 [dlm_locktorture] [ 151.817986] kthread+0x192/0x1d0 [ 151.818518] ret_from_fork+0x1f/0x30 [ 151.819369] Freed by task 1336: [ 151.819890] kasan_save_stack+0x26/0x50 [ 151.820514] kasan_set_track+0x25/0x30 [ 151.821128] kasan_save_free_info+0x2e/0x50 [ 151.821812] __kasan_slab_free+0x107/0x1a0 [ 151.822483] kmem_cache_free+0x204/0x5e0 [ 151.823152] dlm_free_mhandle+0x18/0x20 [ 151.823781] dlm_mhandle_release+0x2e/0x40 [ 151.824454] rcu_core+0x583/0x1330 [ 151.825047] rcu_core_si+0xe/0x20 [ 151.825594] __do_softirq+0xf4/0x5c2 [ 151.826450] Last potentially related work creation: [ 151.827238] kasan_save_stack+0x26/0x50 [ 151.827870] __kasan_record_aux_stack+0xa2/0xc0 [ 151.828609] kasan_record_aux_stack_noalloc+0xb/0x20 [ 151.829415] call_rcu+0x4c/0x760 [ 151.829954] dlm_mhandle_delete+0x97/0xb0 [ 151.830718] dlm_process_incoming_buffer+0x2fc/0xb30 [ 151.831524] process_dlm_messages+0x16e/0x470 [ 151.832245] process_one_work+0x505/0xa10 [ 151.832905] worker_thread+0x67/0x650 [ 151.833507] kthread+0x192/0x1d0 [ 151.834046] ret_from_fork+0x1f/0x30 [ 151.834900] The buggy address belongs to the object at ffff88811a980c30 which belongs to the cache dlm_mhandle of size 88 [ 151.836894] The buggy address is located 48 bytes inside of 88-byte region [ffff88811a980c30, ffff88811a980c88) [ 151.839007] The buggy address belongs to the physical page: [ 151.839904] page:0000000076cf5d62 refcount:1 mapcount:0 mapping:0000000000000000 index:0x0 pfn:0x11a980 [ 151.841378] flags: 0x8000000000000200(slab|zone=2) [ 151.842141] raw: 8000000000000200 0000000000000000 dead000000000122 ffff8881089b43c0 [ 151.843401] raw: 0000000000000000 0000000000220022 00000001ffffffff 0000000000000000 [ 151.844640] page dumped because: kasan: bad access detected [ 151.845822] Memory state around the buggy address: [ 151.846602] ffff88811a980b00: fb fb fb fb fc fc fc fc fa fb fb fb fb fb fb fb [ 151.847761] ffff88811a980b80: fb fb fb fc fc fc fc fa fb fb fb fb fb fb fb fb [ 151.848921] >ffff88811a980c00: fb fb fc fc fc fc fa fb fb fb fb fb fb fb fb fb [ 151.850076] ^ [ 151.851085] ffff88811a980c80: fb fc fc fc fc fa fb fb fb fb fb fb fb fb fb fb [ 151.852269] ffff88811a980d00: fc fc fc fc fa fb fb fb fb fb fb fb fb fb fb fc [ 151.853428] ================================================================== [ 151.855618] Disabling lock debugging due to kernel taint It is accessing a mhandle in dlm_midcomms_commit_mhandle() and the mhandle was freed by a call_rcu() call in dlm_process_incoming_buffer(), dlm_mhandle_delete(). It looks like it was freed because an ack of this message was received. There is a short race between committing the dlm message to be transmitted and getting an ack back. If the ack is faster than returning from dlm_midcomms_commit_msg_3_2(), then we run into a use-after free because we still need to reference the mhandle when calling srcu_read_unlock(). To avoid that, we don't allow that mhandle to be freed between dlm_midcomms_commit_msg_3_2() and srcu_read_unlock() by using rcu read lock. We can do that because mhandle is protected by rcu handling. Cc: stable@vger.kernel.org Fixes: 489d8e559c65 ("fs: dlm: add reliable connection if reconnect") Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2023-01-12 22:10:32 +00:00
rcu_read_unlock();
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
break;
default:
srcu_read_unlock(&nodes_srcu, mh->idx);
WARN_ON_ONCE(1);
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
break;
}
}
#endif
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
int dlm_midcomms_start(void)
{
return dlm_lowcomms_start();
}
void dlm_midcomms_stop(void)
{
dlm_lowcomms_stop();
}
void dlm_midcomms_init(void)
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
{
int i;
for (i = 0; i < CONN_HASH_SIZE; i++)
INIT_HLIST_HEAD(&node_hash[i]);
dlm_lowcomms_init();
}
static void midcomms_node_release(struct rcu_head *rcu)
{
struct midcomms_node *node = container_of(rcu, struct midcomms_node, rcu);
WARN_ON_ONCE(atomic_read(&node->send_queue_cnt));
dlm_send_queue_flush(node);
kfree(node);
}
void dlm_midcomms_exit(void)
{
struct midcomms_node *node;
int i, idx;
idx = srcu_read_lock(&nodes_srcu);
for (i = 0; i < CONN_HASH_SIZE; i++) {
hlist_for_each_entry_rcu(node, &node_hash[i], hlist) {
dlm_delete_debug_comms_file(node->debugfs);
spin_lock(&nodes_lock);
hlist_del_rcu(&node->hlist);
spin_unlock(&nodes_lock);
call_srcu(&nodes_srcu, &node->rcu, midcomms_node_release);
}
}
srcu_read_unlock(&nodes_srcu, idx);
dlm_lowcomms_exit();
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
}
static void dlm_act_fin_ack_rcv(struct midcomms_node *node)
{
spin_lock(&node->state_lock);
pr_debug("receive active fin ack from node %d with state %s\n",
node->nodeid, dlm_state_str(node->state));
switch (node->state) {
case DLM_FIN_WAIT1:
node->state = DLM_FIN_WAIT2;
pr_debug("switch node %d to state %s\n",
node->nodeid, dlm_state_str(node->state));
break;
case DLM_CLOSING:
midcomms_node_reset(node);
pr_debug("switch node %d to state %s\n",
node->nodeid, dlm_state_str(node->state));
break;
case DLM_CLOSED:
/* not valid but somehow we got what we want */
wake_up(&node->shutdown_wait);
break;
default:
spin_unlock(&node->state_lock);
log_print("%s: unexpected state: %d",
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
__func__, node->state);
WARN_ON_ONCE(1);
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
return;
}
spin_unlock(&node->state_lock);
}
void dlm_midcomms_add_member(int nodeid)
{
struct midcomms_node *node;
int idx;
idx = srcu_read_lock(&nodes_srcu);
node = nodeid2node(nodeid);
if (WARN_ON_ONCE(!node)) {
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
srcu_read_unlock(&nodes_srcu, idx);
return;
}
spin_lock(&node->state_lock);
if (!node->users) {
pr_debug("receive add member from node %d with state %s\n",
node->nodeid, dlm_state_str(node->state));
switch (node->state) {
case DLM_ESTABLISHED:
break;
case DLM_CLOSED:
node->state = DLM_ESTABLISHED;
pr_debug("switch node %d to state %s\n",
node->nodeid, dlm_state_str(node->state));
break;
default:
/* some invalid state passive shutdown
* was failed, we try to reset and
* hope it will go on.
*/
log_print("reset node %d because shutdown stuck",
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
node->nodeid);
midcomms_node_reset(node);
node->state = DLM_ESTABLISHED;
break;
}
}
node->users++;
pr_debug("node %d users inc count %d\n", nodeid, node->users);
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
spin_unlock(&node->state_lock);
srcu_read_unlock(&nodes_srcu, idx);
}
void dlm_midcomms_remove_member(int nodeid)
{
struct midcomms_node *node;
int idx;
idx = srcu_read_lock(&nodes_srcu);
node = nodeid2node(nodeid);
/* in case of dlm_midcomms_close() removes node */
if (!node) {
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
srcu_read_unlock(&nodes_srcu, idx);
return;
}
spin_lock(&node->state_lock);
/* case of dlm_midcomms_addr() created node but
* was not added before because dlm_midcomms_close()
* removed the node
*/
if (!node->users) {
spin_unlock(&node->state_lock);
srcu_read_unlock(&nodes_srcu, idx);
return;
}
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
node->users--;
pr_debug("node %d users dec count %d\n", nodeid, node->users);
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
/* hitting users count to zero means the
* other side is running dlm_midcomms_stop()
* we meet us to have a clean disconnect.
*/
if (node->users == 0) {
pr_debug("receive remove member from node %d with state %s\n",
node->nodeid, dlm_state_str(node->state));
switch (node->state) {
case DLM_ESTABLISHED:
break;
case DLM_CLOSE_WAIT:
/* passive shutdown DLM_LAST_ACK case 2 */
node->state = DLM_LAST_ACK;
pr_debug("switch node %d to state %s case 2\n",
node->nodeid, dlm_state_str(node->state));
set_bit(DLM_NODE_FLAG_STOP_RX, &node->flags);
dlm_send_fin(node, dlm_pas_fin_ack_rcv);
break;
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
case DLM_LAST_ACK:
/* probably receive fin caught it, do nothing */
break;
case DLM_CLOSED:
/* already gone, do nothing */
break;
default:
log_print("%s: unexpected state: %d",
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
__func__, node->state);
break;
}
}
spin_unlock(&node->state_lock);
srcu_read_unlock(&nodes_srcu, idx);
}
void dlm_midcomms_version_wait(void)
{
struct midcomms_node *node;
int i, idx, ret;
idx = srcu_read_lock(&nodes_srcu);
for (i = 0; i < CONN_HASH_SIZE; i++) {
hlist_for_each_entry_rcu(node, &node_hash[i], hlist) {
ret = wait_event_timeout(node->shutdown_wait,
node->version != DLM_VERSION_NOT_SET ||
node->state == DLM_CLOSED ||
test_bit(DLM_NODE_FLAG_CLOSE, &node->flags),
DLM_SHUTDOWN_TIMEOUT);
if (!ret || test_bit(DLM_NODE_FLAG_CLOSE, &node->flags))
pr_debug("version wait timed out for node %d with state %s\n",
node->nodeid, dlm_state_str(node->state));
}
}
srcu_read_unlock(&nodes_srcu, idx);
}
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
static void midcomms_shutdown(struct midcomms_node *node)
{
int ret;
/* old protocol, we don't wait for pending operations */
switch (node->version) {
case DLM_VERSION_3_2:
break;
default:
return;
}
spin_lock(&node->state_lock);
pr_debug("receive active shutdown for node %d with state %s\n",
node->nodeid, dlm_state_str(node->state));
switch (node->state) {
case DLM_ESTABLISHED:
node->state = DLM_FIN_WAIT1;
pr_debug("switch node %d to state %s case 2\n",
node->nodeid, dlm_state_str(node->state));
dlm_send_fin(node, dlm_act_fin_ack_rcv);
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
break;
case DLM_CLOSED:
/* we have what we want */
break;
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
default:
/* busy to enter DLM_FIN_WAIT1, wait until passive
* done in shutdown_wait to enter DLM_CLOSED.
*/
break;
}
spin_unlock(&node->state_lock);
if (DLM_DEBUG_FENCE_TERMINATION)
msleep(5000);
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
/* wait for other side dlm + fin */
ret = wait_event_timeout(node->shutdown_wait,
node->state == DLM_CLOSED ||
test_bit(DLM_NODE_FLAG_CLOSE, &node->flags),
DLM_SHUTDOWN_TIMEOUT);
if (!ret)
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
pr_debug("active shutdown timed out for node %d with state %s\n",
node->nodeid, dlm_state_str(node->state));
else
pr_debug("active shutdown done for node %d with state %s\n",
node->nodeid, dlm_state_str(node->state));
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
}
void dlm_midcomms_shutdown(void)
{
struct midcomms_node *node;
int i, idx;
mutex_lock(&close_lock);
idx = srcu_read_lock(&nodes_srcu);
for (i = 0; i < CONN_HASH_SIZE; i++) {
hlist_for_each_entry_rcu(node, &node_hash[i], hlist) {
midcomms_shutdown(node);
}
}
dlm_lowcomms_shutdown();
for (i = 0; i < CONN_HASH_SIZE; i++) {
hlist_for_each_entry_rcu(node, &node_hash[i], hlist) {
midcomms_node_reset(node);
}
}
srcu_read_unlock(&nodes_srcu, idx);
mutex_unlock(&close_lock);
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
}
int dlm_midcomms_close(int nodeid)
{
struct midcomms_node *node;
int idx, ret;
idx = srcu_read_lock(&nodes_srcu);
/* Abort pending close/remove operation */
node = nodeid2node(nodeid);
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
if (node) {
/* let shutdown waiters leave */
set_bit(DLM_NODE_FLAG_CLOSE, &node->flags);
wake_up(&node->shutdown_wait);
}
srcu_read_unlock(&nodes_srcu, idx);
synchronize_srcu(&nodes_srcu);
mutex_lock(&close_lock);
idx = srcu_read_lock(&nodes_srcu);
node = nodeid2node(nodeid);
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
if (!node) {
srcu_read_unlock(&nodes_srcu, idx);
mutex_unlock(&close_lock);
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
return dlm_lowcomms_close(nodeid);
}
ret = dlm_lowcomms_close(nodeid);
dlm_delete_debug_comms_file(node->debugfs);
spin_lock(&nodes_lock);
hlist_del_rcu(&node->hlist);
spin_unlock(&nodes_lock);
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
srcu_read_unlock(&nodes_srcu, idx);
/* wait that all readers left until flush send queue */
synchronize_srcu(&nodes_srcu);
/* drop all pending dlm messages, this is fine as
* this function get called when the node is fenced
*/
dlm_send_queue_flush(node);
call_srcu(&nodes_srcu, &node->rcu, midcomms_node_release);
fs: dlm: add reliable connection if reconnect This patch introduce to make a tcp lowcomms connection reliable even if reconnects occurs. This is done by an application layer re-transmission handling and sequence numbers in dlm protocols. There are three new dlm commands: DLM_OPTS: This will encapsulate an existing dlm message (and rcom message if they don't have an own application side re-transmission handling). As optional handling additional tlv's (type length fields) can be appended. This can be for example a sequence number field. However because in DLM_OPTS the lockspace field is unused and a sequence number is a mandatory field it isn't made as a tlv and we put the sequence number inside the lockspace id. The possibility to add optional options are still there for future purposes. DLM_ACK: Just a dlm header to acknowledge the receive of a DLM_OPTS message to it's sender. DLM_FIN: This provides a 4 way handshake for connection termination inclusive support for half-closed connections. It's provided on application layer because SCTP doesn't support half-closed sockets, the shutdown() call can interrupted by e.g. TCP resets itself and a hard logic to implement it because the othercon paradigm in lowcomms. The 4-way termination handshake also solve problems to synchronize peer EOF arrival and that the cluster manager removes the peer in the node membership handling of DLM. In some cases messages can be still transmitted in this time and we need to wait for the node membership event. To provide a reliable connection the node will retransmit all unacknowledges message to it's peer on reconnect. The receiver will then filtering out the next received message and drop all messages which are duplicates. As RCOM_STATUS and RCOM_NAMES messages are the first messages which are exchanged and they have they own re-transmission handling, there exists logic that these messages must be first. If these messages arrives we store the dlm version field. This handling is on DLM 3.1 and after this patch 3.2 the same. A backwards compatibility handling has been added which seems to work on tests without tcpkill, however it's not recommended to use DLM 3.1 and 3.2 at the same time, because DLM 3.2 tries to fix long term bugs in the DLM protocol. Signed-off-by: Alexander Aring <aahringo@redhat.com> Signed-off-by: David Teigland <teigland@redhat.com>
2021-05-21 19:08:46 +00:00
mutex_unlock(&close_lock);
return ret;
}
/* debug functionality to send raw dlm msg from user space */
struct dlm_rawmsg_data {
struct midcomms_node *node;
void *buf;
};
static void midcomms_new_rawmsg_cb(void *data)
{
struct dlm_rawmsg_data *rd = data;
struct dlm_header *h = rd->buf;
switch (h->h_version) {
case cpu_to_le32(DLM_VERSION_3_1):
break;
default:
switch (h->h_cmd) {
case DLM_OPTS:
if (!h->u.h_seq)
h->u.h_seq = cpu_to_le32(atomic_fetch_inc(&rd->node->seq_send));
break;
default:
break;
}
break;
}
}
int dlm_midcomms_rawmsg_send(struct midcomms_node *node, void *buf,
int buflen)
{
struct dlm_rawmsg_data rd;
struct dlm_msg *msg;
char *msgbuf;
rd.node = node;
rd.buf = buf;
msg = dlm_lowcomms_new_msg(node->nodeid, buflen, GFP_NOFS,
&msgbuf, midcomms_new_rawmsg_cb, &rd);
if (!msg)
return -ENOMEM;
memcpy(msgbuf, buf, buflen);
dlm_lowcomms_commit_msg(msg);
return 0;
}