| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
Use spin_lock_bh for all spinlocks involved in message processing,
in preparation for softirq message processing. DLM lock requests
from user space involve dlm processing in user context, in addition
to the standard kernel context, necessitating bh variants.
Signed-off-by: Alexander Aring <aahringo@redhat.com>
Signed-off-by: David Teigland <teigland@redhat.com>
|
|
Remove the context parameter for message allocations and
always use GFP_ATOMIC. This prepares for softirq message
processing.
Signed-off-by: Alexander Aring <aahringo@redhat.com>
Signed-off-by: David Teigland <teigland@redhat.com>
|
|
Use the new KMEM_CACHE() macro instead of direct kmem_cache_create
to simplify the creation of SLAB caches.
Signed-off-by: Kunwu Chan <chentao@kylinos.cn>
Signed-off-by: Alexander Aring <aahringo@redhat.com>
Signed-off-by: David Teigland <teigland@redhat.com>
|
|
In case of an final DLM message we can't should not send an ack out
after the final message. This patch moves the ack message before the
messages will be transmitted. If it's the final message and the
receiving node turns into DLM_CLOSED state another ack messages will
being received and turning the receiving node into DLM_ESTABLISHED
again.
Fixes: 1696c75f1864 ("fs: dlm: add send ack threshold and append acks to msgs")
Signed-off-by: Alexander Aring <aahringo@redhat.com>
Signed-off-by: David Teigland <teigland@redhat.com>
|
|
In case we running in a force shutdown in either midcomms or lowcomms
implementation we will make sure we reset all per midcomms node
information.
Fixes: 63e711b08160 ("fs: dlm: create midcomms nodes when configure")
Signed-off-by: Alexander Aring <aahringo@redhat.com>
Signed-off-by: David Teigland <teigland@redhat.com>
|
|
The idea of commit 63e711b08160 ("fs: dlm: create midcomms nodes when
configure") is to set the midcomms node lifetime when a node joins or
leaves the cluster. Currently we can hit the following warning:
[10844.611495] ------------[ cut here ]------------
[10844.615913] WARNING: CPU: 4 PID: 84304 at fs/dlm/midcomms.c:1263
dlm_midcomms_remove_member+0x13f/0x180 [dlm]
or running in a state where we hit a midcomms node usage count in a
negative value:
[ 260.830782] node 2 users dec count -1
The first warning happens when the a specific node does not exists and
it was probably removed but dlm_midcomms_close() which is called when a
node leaves the cluster. The second kernel log message is probably in a
case when dlm_midcomms_addr() is called when a joined the cluster but
due fencing a node leaved the cluster without getting removed from the
lockspace. If the node joins the cluster and it was removed from the
cluster due fencing the first call is to remove the node from lockspaces
triggered by the user space. In both cases if the node wasn't found or
the user count is zero, we should ignore any additional midcomms handling
of dlm_midcomms_remove_member().
Fixes: 63e711b08160 ("fs: dlm: create midcomms nodes when configure")
Signed-off-by: Alexander Aring <aahringo@redhat.com>
Signed-off-by: David Teigland <teigland@redhat.com>
|
|
This patch will lookup existing nodes instead of always creating them
when dlm_midcomms_addr() is called. The idea is here to create midcomms
nodes when user space getting informed that nodes joins the cluster. This
is the case when dlm_midcomms_addr() is called, however it can be called
multiple times by user space to add several address configurations to one
node e.g. when using SCTP. Those multiple times need to be filtered out
and we doing that by looking up if the node exists before. Due configfs
entry it is safe that this function gets only called once at a time.
Fixes: 63e711b08160 ("fs: dlm: create midcomms nodes when configure")
Signed-off-by: Alexander Aring <aahringo@redhat.com>
Signed-off-by: David Teigland <teigland@redhat.com>
|
|
This patch puts the life of a midcomms node the same as a lowcomms
connection. The lowcomms connection lifetime was changed by commit
6f0b0b5d7ae7 ("fs: dlm: remove dlm_node_addrs lookup list"). In the
future the midcomms node instances can be merged with lowcomms
connection structure as the lifetime is the same and states can be
controlled over values or flags.
Before midcomms nodes were generated during version detection. This is
not necessary anymore when the nodes are created when the cluster
manager configures DLM via configfs. When a midcomms node is created over
configfs it well set DLM_VERSION_NOT_SET as version. This indicates that
the version of the midcomms node is still unknown and need to be probed
via certain rcom messages.
Signed-off-by: Alexander Aring <aahringo@redhat.com>
Signed-off-by: David Teigland <teigland@redhat.com>
|
|
The dlm receive buffer should be never manipulated as DLM is the last
instance of parsing layer. This patch constify the whole receive buffer
so we are sure it never gets manipulated when it's being parsed.
Signed-off-by: Alexander Aring <aahringo@redhat.com>
Signed-off-by: David Teigland <teigland@redhat.com>
|
|
This patch cleanups the lock order to hold at first the close_lock and
then held the nodes_srcu read lock. Probably it will never be a problem
as nodes_srcu is only a read lock preventing the node pointer getting
freed.
Signed-off-by: Alexander Aring <aahringo@redhat.com>
Signed-off-by: David Teigland <teigland@redhat.com>
|
|
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>
|
|
Currently seq_next is only be read on the receive side which processed
in an ordered way. The seq_send is being protected by locks. To being
able to read the seq_next value on send side as well we convert it to an
atomic_t value. The atomic_cmpxchg() is probably not necessary, however
the atomic_inc() depends on a if coniditional and this should be handled
in an atomic context.
Signed-off-by: Alexander Aring <aahringo@redhat.com>
Signed-off-by: David Teigland <teigland@redhat.com>
|
|
It makes no sense to call midcomms/lowcomms functionality for the local
node as socket functionality is only required for remote nodes. This
patch filters those calls in the upper layer of lockspace membership
handling instead of doing it in midcomms/lowcomms layer as they should
never be aware of local nodeid.
Signed-off-by: Alexander Aring <aahringo@redhat.com>
Signed-off-by: David Teigland <teigland@redhat.com>
|
|
This patch reverts commit 2c3fa6ae4d52 ("dlm: check required context
while close"). The function dlm_midcomms_close(), which will call later
dlm_lowcomms_close(), is called when the cluster manager tells the node
got fenced which means on midcomms/lowcomms layer to disconnect the node
from the cluster communication. The node can rejoin the cluster later.
This patch was ensuring no new message were able to be triggered when we
are in the close() function context. This was done by checking if the
lockspace has been stopped. However there is a missing check that we
only need to check specific lockspaces where the fenced node is member
of. This is currently complicated because there is no way to easily
check if a node is part of a specific lockspace without stopping the
recovery. For now we just revert this commit as it is just a check to
finding possible leaks of stopping lockspaces before close() is called.
Cc: stable@vger.kernel.org
Fixes: 2c3fa6ae4d52 ("dlm: check required context while close")
Signed-off-by: Alexander Aring <aahringo@redhat.com>
Signed-off-by: David Teigland <teigland@redhat.com>
|
|
The wake_up() is already handled inside of midcomms_node_reset() when
switching the state to CLOSED state. So there is not need to call it
after midcomms_node_reset() again.
Signed-off-by: Alexander Aring <aahringo@redhat.com>
Signed-off-by: David Teigland <teigland@redhat.com>
|
|
Currently we can switch at first into DLM_CLOSE_WAIT state and then do
another state change if a condition is true. Instead of doing two state
changes we handle the other state change inside an else branch of this
condition.
Signed-off-by: Alexander Aring <aahringo@redhat.com>
Signed-off-by: David Teigland <teigland@redhat.com>
|
|
There is an API difference between log_print() and other printk()s to
put a newline or not. This one was introduced by mistake because
log_print() adds a newline.
Signed-off-by: Alexander Aring <aahringo@redhat.com>
Signed-off-by: David Teigland <teigland@redhat.com>
|
|
When a shutdown is stuck, time out after 5 seconds instead of
3 minutes. After this timeout we try a forced shutdown.
Signed-off-by: Alexander Aring <aahringo@redhat.com>
Signed-off-by: David Teigland <teigland@redhat.com>
|
|
The current dlm version detection is very complex due to backwards
compatablilty with earlier dlm protocol versions. It takes some time to
detect if a peer node has a specific DLM version. If it's not detected,
we just cut the socket connection. There could be cases where the local
node has not detected the version yet, but the peer node has. In these
cases, we are trying to shutdown the dlm connection with a FIN/ACK message
exchange to be sure the other peer is ready to shutdown the connection on
dlm application level. However this mechanism is only available on DLM
protocol version 3.2 and we need to be sure the DLM version is detected
before.
To make it more robust we introduce a a "best effort" wait to wait for the
version detection before shutdown the dlm connection. This need to be
done before the kthread recoverd for recovery handling is stopped,
because recovery handling will trigger enough messages to have a version
detection going on.
It is a corner case which was detected by modprobe dlm_locktroture module
and rmmod dlm_locktorture module directly afterwards (in a looping
behaviour). In practice probably nobody would leave a lockspace immediately
after joining it.
Signed-off-by: Alexander Aring <aahringo@redhat.com>
Signed-off-by: David Teigland <teigland@redhat.com>
|
|
This patch ignores unexpected RCOM_NAMES/RCOM_STATUS messages.
To be backwards compatible, those messages are not part of the new
reliable DLM OPTS encapsulation header, and have their own
retransmit handling using sequence number matching When we get
unexpected non dlm opts messages, we should allow them and let
RCOM message handling filter them out using sequence numbers.
Signed-off-by: Alexander Aring <aahringo@redhat.com>
Signed-off-by: David Teigland <teigland@redhat.com>
|
|
This patch mostly reverts commit 4f567acb0b86 ("fs: dlm: remove socket
shutdown handling"). There can be situations where the dlm midcomms nodes
hash and lowcomms connection hash are not equal, but we need to guarantee
that the lowcomms are all closed on a last release of a dlm lockspace,
when a shutdown is invoked. This patch guarantees that we always close
all sockets managed by the lowcomms connection hash, and calls shutdown
for the last message sent. This ensures we don't cut the socket, which
could cause the peer to get a connection reset.
In future we should try to merge the midcomms/lowcomms hashes into one
hash and not handle both in separate hashes.
Signed-off-by: Alexander Aring <aahringo@redhat.com>
Signed-off-by: David Teigland <teigland@redhat.com>
|
|
This patch moves to send a ack back for receiving a FIN message only
when we are in valid states. In other cases and there might be a sender
waiting for a ack we just let it timeout at the senders time and
hopefully all other cleanups will remove the FIN message on their
sending queue. As an example we should never send out an ACK being in
LAST_ACK state or we cannot assume a working socket communication when
we are in CLOSED state.
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>
|
|
This patch moves the send fin handling, which should appear in a specific
state change, into the state change handling while the per node
state_lock is held. I experienced issues with other messages because
we changed the state and a fin message was sent out in a different state.
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>
|
|
Similar to the stop tx flag, the rx flag should warn about a dlm message
being received at DLM_FIN state change, when we are assuming no other
dlm application messages. If we receive a FIN message and we are in the
state DLM_FIN_WAIT2 we call midcomms_node_reset() which puts the
midcomms node into DLM_CLOSED state. Afterwards we should not set the
DLM_NODE_FLAG_STOP_RX flag any more. This patch changes the setting
DLM_NODE_FLAG_STOP_RX in those state changes when we receive a FIN
message and we assume there will be no other dlm application messages
received until we hit DLM_CLOSED state.
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>
|
|
This patch sets the stop tx flag before we commit the dlm message.
This flag will report about unexpected transmissions after we
send the DLM_FIN message out, which should be the last message sent.
When we commit the dlm fin message, it could be that we already
got an ack back and the CLOSED state change already happened.
We should not set this flag when we are in CLOSED state. To avoid this
race we simply set the tx flag before the state change can be in
progress by moving it before dlm_midcomms_commit_mhandle().
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>
|
|
If we release a midcomms node structure, there should be nothing left
inside the dlm midcomms send queue. However, sometimes this is not true
because I believe some DLM_FIN message was not acked... if we run
into a shutdown timeout, then we should be sure there is no pending send
dlm message inside this queue when releasing midcomms node structure.
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>
|
|
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>
|
|
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>
|
|
Since commit 489d8e559c65 ("fs: dlm: add reliable connection if
reconnect") we have functionality like TCP offers for half-closed
sockets on dlm application protocol layer. This feature is required
because the cluster manager events about leaving resource memberships
can be locally already occurred but other cluster nodes having a pending
leaving membership over the cluster manager protocol happening. In this
time the local dlm node already shutdown it's connection and don't
transmit anymore any new dlm messages, but however it still needs to be
able to accept dlm messages because the pending leave membership request
of the cluster manager protocol which the dlm kernel implementation has
no control about it.
We have this functionality on the application for two reasons, the main
reason is that SCTP does not support such functionality on socket
layer. But we can do it inside application layer.
Another small issue is that this feature is broken in the TCP world
because some NAT devices does not implement such functionality
correctly. This is the same reason why the reliable connection session
layer in DLM exists. We give up on middle devices in the networking
which sends e.g. TCP resets out. In DLM we cannot have any message
dropping and we ensure it over a session layer that it can't happen.
Back to the half-closed grace shutdown handling. It's not necessary
anymore to do it on socket layer (which is only support for TCP sockets)
because we do it on application layer. This patch removes this handling,
if there are still issues then we have a problem on the application
layer for such handling.
Signed-off-by: Alexander Aring <aahringo@redhat.com>
Signed-off-by: David Teigland <teigland@redhat.com>
|
|
This patch introduces leftovers of init, start, stop and exit
functionality. The dlm application layer should always call the midcomms
layer which getting aware of such event and redirect it to the lowcomms
layer. Some functionality which is currently handled inside the start
functionality of midcomms and lowcomms should be handled in the init
functionality as it only need to be initialized once when dlm is loaded.
Signed-off-by: Alexander Aring <aahringo@redhat.com>
Signed-off-by: David Teigland <teigland@redhat.com>
|
|
In DLM when we send a dlm message it is easy to add the lock resource
name, but additional lookup is required when to trace the receive
message side. The idea here is to move the lookup work to the user by
using a lookup to find the right send message with recv message. As note
DLM can't drop any message which is guaranteed by a special session
layer.
For doing the lookup a 3 tupel is required as an unique identification
which is dst nodeid, src nodeid and sequence number. This patch adds the
destination nodeid to the dlm message trace points. The source nodeid is
given by the h_nodeid field inside the header.
Signed-off-by: Alexander Aring <aahringo@redhat.com>
Signed-off-by: David Teigland <teigland@redhat.com>
|
|
To not get the console spammed about WARN_ON() of invalid states in the
dlm midcomms hot path handling we switch to WARN_ON_ONCE() to get it
only once that there might be an issue with the midcomms state handling.
Signed-off-by: Alexander Aring <aahringo@redhat.com>
Signed-off-by: David Teigland <teigland@redhat.com>
|
|
This patch allows to give the use control about the allocation context
based on a per message basis. Currently all messages forced to be
created under GFP_NOFS context.
Signed-off-by: Alexander Aring <aahringo@redhat.com>
Signed-off-by: David Teigland <teigland@redhat.com>
|
|
This patch adds tracepoints for send and recv cases of dlm messages and
dlm rcom messages. In case of send and dlm message we add the dlm rsb
resource name this dlm messages belongs to. This has the advantage to
follow dlm messages on a per lock basis. In case of recv message the
resource name can be extracted by follow the send message sequence
number.
The dlm message DLM_MSG_PURGE doesn't belong to a lock request and will
not set the resource name in a dlm_message trace. The same for all rcom
messages.
There is additional handling required for this debugging functionality
which is tried to be small as possible. Also the midcomms layer gets
aware of lock resource names, for now this is required to make a
connection between sequence number and lock resource names. It is for
debugging purpose only.
Signed-off-by: Alexander Aring <aahringo@redhat.com>
Signed-off-by: David Teigland <teigland@redhat.com>
|
|
To allow more than just dereferencing the inner header we directly point
to the inner dlm packet which allows us to dereference the header, rcom
or message structure.
Signed-off-by: Alexander Aring <aahringo@redhat.com>
Signed-off-by: David Teigland <teigland@redhat.com>
|
|
This patch adds a WARN_ON() check to validate the right context while
dlm_midcomms_close() is called. Even before commit 489d8e559c65
("fs: dlm: add reliable connection if reconnect") in this context
dlm_lowcomms_close() flushes all ongoing transmission triggered by dlm
application stack. If we do that, it's required that no new message will
be triggered by the dlm application stack. The function
dlm_midcomms_close() is not called often so we can check if all
lockspaces are in such context.
Signed-off-by: Alexander Aring <aahringo@redhat.com>
Signed-off-by: David Teigland <teigland@redhat.com>
|
|
This patch is a cleanup to move the byte order conversion to compile
time. In a simple comparison like this it's possible to move it to
static values so the compiler will always convert those values at
compile time.
Signed-off-by: Alexander Aring <aahringo@redhat.com>
Signed-off-by: David Teigland <teigland@redhat.com>
|
|
This patch changes to use __le types directly in the dlm header
structure which is casted at the right dlm message buffer positions.
The main goal what is reached here is to remove sparse warnings
regarding to host to little byte order conversion or vice versa. Leaving
those sparse issues ignored and always do it in out/in functionality
tends to leave it unknown in which byte order the variable is being
handled.
Signed-off-by: Alexander Aring <aahringo@redhat.com>
Signed-off-by: David Teigland <teigland@redhat.com>
|
|
This patch will adds #ifndef __CHECKER__ for false positives warnings
about an imbalance lock/unlock srcu handling. Which are shown by running
sparse checks:
fs/dlm/midcomms.c:1065:20: warning: context imbalance in 'dlm_midcomms_get_mhandle' - wrong count at exit
Using __CHECKER__ will tell sparse to ignore these sections.
Those imbalances are false positive because from upper layer it is
always required to call a function in sequence, e.g. if
dlm_midcomms_get_mhandle() is successful there must be a
dlm_midcomms_commit_mhandle() call afterwards.
Signed-off-by: Alexander Aring <aahringo@redhat.com>
Signed-off-by: David Teigland <teigland@redhat.com>
|
|
This patch will introduce a kmem cache for allocating message handles
which are needed for midcomms layer to take track of lowcomms messages.
Signed-off-by: Alexander Aring <aahringo@redhat.com>
Signed-off-by: David Teigland <teigland@redhat.com>
|
|
Eliminate the following coccinelle check warning:
fs/dlm/midcomms.c:972:2-3
Reported-by: Zeal Robot <zealci@zte.com.cn>
Signed-off-by: Zhang Mingyu <zhang.mingyu@zte.com.cn>
Signed-off-by: David Teigland <teigland@redhat.com>
|
|
This patch adds a dlm functionality to send a raw dlm message to a
specific cluster node. This raw message can be build by user space and
send out by writing the message to "rawmsg" dlm debugfs file.
There is a in progress scapy dlm module which provides a easy build of
DLM messages in user space. For example:
DLM(h_cmd=3, o_nextcmd=1, h_nodeid=1, h_lockspace=0xe4f48a18, ...)
The goal is to provide an easy reproducable state to crash DLM or to
fuzz the DLM kernel stack if there are possible ways to crash it.
Note: that if the sequence number is zero and dlm version is not set to
3.1 the kernel will automatic will set a right sequence number, otherwise
DLM stack testing is not possible.
Signed-off-by: Alexander Aring <aahringo@redhat.com>
Signed-off-by: David Teigland <teigland@redhat.com>
|
|
This patch changes the dlm_lowcomms_new_msg() function pointer private data
from "struct mhandle *" to "void *" to provide different structures than
just "struct mhandle".
Signed-off-by: Alexander Aring <aahringo@redhat.com>
Signed-off-by: David Teigland <teigland@redhat.com>
|
|
This patch moves version conversion to little endian from a runtime
variable to compile time constant.
Signed-off-by: Alexander Aring <aahringo@redhat.com>
Signed-off-by: David Teigland <teigland@redhat.com>
|
|
This patch improves the debug output for midcomms layer by also printing
out the nodeid where users counter belongs to.
Signed-off-by: Alexander Aring <aahringo@redhat.com>
Signed-off-by: David Teigland <teigland@redhat.com>
|
|
This patch changes that we don't ack each message. Lowcomms will take
care about to send an ack back after a bulk of messages was processed.
Currently it's only when the whole receive buffer was processed, there
might better positions to send an ack back but only the lowcomms
implementation know when there are more data to receive. This patch has
also disadvantages that we might retransmit more on errors, however this
is a very rare case.
Tested with make_panic on gfs2 with three nodes by running:
trace-cmd record -p function -l 'dlm_send_ack' sleep 100
and
trace-cmd report | wc -l
Before patch:
- 20548
- 21376
- 21398
After patch:
- 18338
- 20679
- 19949
Signed-off-by: Alexander Aring <aahringo@redhat.com>
Signed-off-by: David Teigland <teigland@redhat.com>
|
|
This patch will evaluate the message length if a dlm opts header can fit
in before accessing it if a node lookup fails. The invalid sequence
error means that the version detection failed and an unexpected message
arrived. For debugging such situation the type of arrived message is
important to know.
Signed-off-by: Alexander Aring <aahringo@redhat.com>
Signed-off-by: David Teigland <teigland@redhat.com>
|
|
This patch fixes a race between mhandle deletion in case of receiving an
acknowledge and flush of all pending mhandle in cases of an timeout or
resetting node states.
Fixes: 489d8e559c65 ("fs: dlm: add reliable connection if reconnect")
Reported-by: Guillaume Nault <gnault@redhat.com>
Signed-off-by: Alexander Aring <aahringo@redhat.com>
Signed-off-by: David Teigland <teigland@redhat.com>
|
|
This patch renames DEFAULT_BUFFER_SIZE to DLM_MAX_SOCKET_BUFSIZE and
LOWCOMMS_MAX_TX_BUFFER_LEN to DLM_MAX_APP_BUFSIZE as they are proper
names to define what's behind those values. The DLM_MAX_SOCKET_BUFSIZE
defines the maximum size of buffer which can be handled on socket layer,
the DLM_MAX_APP_BUFSIZE defines the maximum size of buffer which can be
handled by the DLM application layer.
Signed-off-by: Alexander Aring <aahringo@redhat.com>
Signed-off-by: David Teigland <teigland@redhat.com>
|
|
There are spelling mistake in log messages. Fix these.
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: David Teigland <teigland@redhat.com>
|
|
This patch adds functionality to debug midcomms per connection state
inside a comms directory which is similar like dlm configfs. Currently
there exists the possibility to read out two attributes which is the
send queue counter and the version of each midcomms node state.
Signed-off-by: Alexander Aring <aahringo@redhat.com>
Signed-off-by: David Teigland <teigland@redhat.com>
|
|
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>
|
|
This patch makes the void pointer handle for lowcomms functionality per
message and not per page allocation entry. A refcount handling for the
handle was added to keep the message alive until the user doesn't need
it anymore.
There exists now a per message callback which will be called when
allocating a new buffer. This callback will be guaranteed to be called
according the order of the sending buffer, which can be used that the
caller increments a sequence number for the dlm message handle.
For transition process we cast the dlm_mhandle to dlm_msg and vice versa
until the midcomms layer will implement a specific dlm_mhandle structure.
Signed-off-by: Alexander Aring <aahringo@redhat.com>
Signed-off-by: David Teigland <teigland@redhat.com>
|
|
This patch prepares hooks to redirect to the midcomms layer which will
be used by the midcomms re-transmit handling.
There exists the new concept of stateless buffers allocation and
commits. This can be used to bypass the midcomms re-transmit handling. It
is used by RCOM_STATUS and RCOM_NAMES messages, because they have their
own ping-like re-transmit handling. As well these two messages will be
used to determine the DLM version per node, because these two messages
are per observation the first messages which are exchanged.
Cluster manager events for node membership are added to add support for
half-closed connections in cases that the peer connection get to
an end of file but DLM still holds membership of the node. In
this time DLM can still trigger new message which we should allow. After
the cluster manager node removal event occurs it safe to close the
connection.
Signed-off-by: Alexander Aring <aahringo@redhat.com>
Signed-off-by: David Teigland <teigland@redhat.com>
|
|
This patch removes unaligned memory access handling for receiving
midcomms messages. This handling will not fix the unaligned memory
access in general. All messages should be length aligned to 8 bytes,
there exists cases where this isn't the case. It's part of the sending
handling to not send such messages. As the sending handling itself, with
the internal allocator of page buffers, can occur in unaligned memory
access of dlm message fields we just ignore that problem for now as it
seems this code is used by architecture which can handle it.
This patch adds a comment to take care about that problem in a major
bump of dlm protocol.
Signed-off-by: Alexander Aring <aahringo@redhat.com>
Signed-off-by: David Teigland <teigland@redhat.com>
|
|
This patch adds an additional check for minimum dlm header size which is
an invalid dlm message and signals a broken stream. A msglen field cannot
be less than the dlm header size because the field is inclusive header
lengths.
Signed-off-by: Alexander Aring <aahringo@redhat.com>
Signed-off-by: David Teigland <teigland@redhat.com>
|
|
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>
|
|
Based on 1 normalized pattern(s):
this copyrighted material is made available to anyone wishing to use
modify copy or redistribute it subject to the terms and conditions
of the gnu general public license v 2
extracted by the scancode license scanner the SPDX license identifier
GPL-2.0-only
has been chosen to replace the boilerplate/reference in 45 file(s).
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Richard Fontana <rfontana@redhat.com>
Reviewed-by: Allison Randal <allison@lohutok.net>
Reviewed-by: Steve Winslow <swinslow@gmail.com>
Reviewed-by: Alexios Zavras <alexios.zavras@intel.com>
Cc: linux-spdx@vger.kernel.org
Link: https://lkml.kernel.org/r/20190528170027.342746075@linutronix.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
|
Use ls_allocation for memory allocations, which a cluster fs sets to
GFP_NOFS. Use GFP_NOFS for allocations when no lockspace struct is
available. Taking dlm locks needs to avoid calling back into the
cluster fs because write-out can require taking dlm locks.
Cc: Christine Caulfield <ccaulfie@redhat.com>
Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
Signed-off-by: David Teigland <teigland@redhat.com>
|
|
* check that length is large enough to cover the non-variable part of message or
rcom resp. (after checking that it's large enough to cover the header, of
course).
* kill more pointless casts
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: David Teigland <teigland@redhat.com>
|
|
gcc does not guarantee that an auto buffer is 64bit aligned.
This change allows sparc64 to work.
Signed-off-by: Fabio M. Di Nitto <fabbione@ubuntu.com>
Signed-off-by: David Teigland <teigland@redhat.com>
|
|
Introduce a per-lockspace rwsem that's held in read mode by dlm_recv
threads while working in the dlm. This allows dlm_recv activity to be
suspended when the lockspace transitions to, from and between recovery
cycles.
The specific bug prompting this change is one where an in-progress
recovery cycle is aborted by a new recovery cycle. While dlm_recv was
processing a recovery message, the recovery cycle was aborted and
dlm_recoverd began cleaning up. dlm_recv decremented recover_locks_count
on an rsb after dlm_recoverd had reset it to zero. This is fixed by
suspending dlm_recv (taking write lock on the rwsem) before aborting the
current recovery.
The transitions to/from normal and recovery modes are simplified by using
this new ability to block dlm_recv. The switch from normal to recovery
mode means dlm_recv goes from processing locking messages, to saving them
for later, and vice versa. Races are avoided by blocking dlm_recv when
setting the flag that switches between modes.
Signed-off-by: David Teigland <teigland@redhat.com>
Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
|
|
Add a "ci_" prefix to the fields in the dlm_config_info struct so that we
can use macros to add configfs functions to access them (in a later
patch). No functional changes in this patch, just naming changes.
Signed-off-by: David Teigland <teigland@redhat.com>
Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
|
|
Signed-off-by: David Teigland <teigland@redhat.com>
Signed-off-by: Steve Whitehouse <swhiteho@redhat.com>
|
|
This is the core of the distributed lock manager which is required
to use GFS2 as a cluster filesystem. It is also used by CLVM and
can be used as a standalone lock manager independantly of either
of these two projects.
It implements VAX-style locking modes.
Signed-off-by: David Teigland <teigland@redhat.com>
Signed-off-by: Steve Whitehouse <swhiteho@redhat.com>
|
