This patch will use an event based waitqueue to wait for a possible clash
with the ls_remove_name field of dlm_ls instead of doing busy waiting.
Signed-off-by: Alexander Aring <aahringo@redhat.com>
Signed-off-by: David Teigland <teigland@redhat.com>
This patch fixes the following crash by receiving a invalid message:
[ 160.672220] ==================================================================
[ 160.676206] BUG: KASAN: user-memory-access in dlm_user_add_ast+0xc3/0x370
[ 160.679659] Read of size 8 at addr 00000000deadbeef by task kworker/u32:13/319
[ 160.681447]
[ 160.681824] CPU: 10 PID: 319 Comm: kworker/u32:13 Not tainted 5.14.0-rc2+ #399
[ 160.683472] Hardware name: Red Hat KVM/RHEL-AV, BIOS 1.14.0-1.module+el8.6.0+12648+6ede71a5 04/01/2014
[ 160.685574] Workqueue: dlm_recv process_recv_sockets
[ 160.686721] Call Trace:
[ 160.687310] dump_stack_lvl+0x56/0x6f
[ 160.688169] ? dlm_user_add_ast+0xc3/0x370
[ 160.689116] kasan_report.cold.14+0x116/0x11b
[ 160.690138] ? dlm_user_add_ast+0xc3/0x370
[ 160.690832] dlm_user_add_ast+0xc3/0x370
[ 160.691502] _receive_unlock_reply+0x103/0x170
[ 160.692241] _receive_message+0x11df/0x1ec0
[ 160.692926] ? rcu_read_lock_sched_held+0xa1/0xd0
[ 160.693700] ? rcu_read_lock_bh_held+0xb0/0xb0
[ 160.694427] ? lock_acquire+0x175/0x400
[ 160.695058] ? do_purge.isra.51+0x200/0x200
[ 160.695744] ? lock_acquired+0x360/0x5d0
[ 160.696400] ? lock_contended+0x6a0/0x6a0
[ 160.697055] ? lock_release+0x21d/0x5e0
[ 160.697686] ? lock_is_held_type+0xe0/0x110
[ 160.698352] ? lock_is_held_type+0xe0/0x110
[ 160.699026] ? ___might_sleep+0x1cc/0x1e0
[ 160.699698] ? dlm_wait_requestqueue+0x94/0x140
[ 160.700451] ? dlm_process_requestqueue+0x240/0x240
[ 160.701249] ? down_write_killable+0x2b0/0x2b0
[ 160.701988] ? do_raw_spin_unlock+0xa2/0x130
[ 160.702690] dlm_receive_buffer+0x1a5/0x210
[ 160.703385] dlm_process_incoming_buffer+0x726/0x9f0
[ 160.704210] receive_from_sock+0x1c0/0x3b0
[ 160.704886] ? dlm_tcp_shutdown+0x30/0x30
[ 160.705561] ? lock_acquire+0x175/0x400
[ 160.706197] ? rcu_read_lock_sched_held+0xa1/0xd0
[ 160.706941] ? rcu_read_lock_bh_held+0xb0/0xb0
[ 160.707681] process_recv_sockets+0x32/0x40
[ 160.708366] process_one_work+0x55e/0xad0
[ 160.709045] ? pwq_dec_nr_in_flight+0x110/0x110
[ 160.709820] worker_thread+0x65/0x5e0
[ 160.710423] ? process_one_work+0xad0/0xad0
[ 160.711087] kthread+0x1ed/0x220
[ 160.711628] ? set_kthread_struct+0x80/0x80
[ 160.712314] ret_from_fork+0x22/0x30
The issue is that we received a DLM message for a user lock but the
destination lock is a kernel lock. Note that the address which is trying
to derefence is 00000000deadbeef, which is in a kernel lock
lkb->lkb_astparam, this field should never be derefenced by the DLM
kernel stack. In case of a user lock lkb->lkb_astparam is lkb->lkb_ua
(memory is shared by a union field). The struct lkb_ua will be handled
by the DLM kernel stack but on a kernel lock it will contain invalid
data and ends in most likely crashing the kernel.
It can be reproduced with two cluster nodes.
node 2:
dlm_tool join test
echo "862 fooobaar 1 2 1" > /sys/kernel/debug/dlm/test_locks
echo "862 3 1" > /sys/kernel/debug/dlm/test_waiters
node 1:
dlm_tool join test
python:
foo = DLM(h_cmd=3, o_nextcmd=1, h_nodeid=1, h_lockspace=0x77222027, \
m_type=7, m_flags=0x1, m_remid=0x862, m_result=0xFFFEFFFE)
newFile = open("/sys/kernel/debug/dlm/comms/2/rawmsg", "wb")
newFile.write(bytes(foo))
Signed-off-by: Alexander Aring <aahringo@redhat.com>
Signed-off-by: David Teigland <teigland@redhat.com>
This patch adds functionality to put a lkb to the waiters state. It can
be useful to combine this feature with the "rawmsg" debugfs
functionality. It will bring the DLM lkb into a state that a message
will be parsed by the kernel.
Signed-off-by: Alexander Aring <aahringo@redhat.com>
Signed-off-by: David Teigland <teigland@redhat.com>
This patch adds functionality to add an lkb during runtime. This is a
highly debugging feature only, wrong input can crash the kernel. It is a
early state feature as well. The goal is to provide a user interface for
manipulate dlm state and combine it with the rawmsg feature. It is
debugfs functionality, we don't care about UAPI breakage. Even it's
possible to add lkb's/rsb's which could never be exists in such wat by
using normal DLM operation. The user of this interface always need to
think before using this feature, not every crash which happens can really
occur during normal dlm operation.
Future there should be more functionality to add a more realistic lkb
which reflects normal DLM state inside the kernel. For now this is
enough.
Signed-off-by: Alexander Aring <aahringo@redhat.com>
Signed-off-by: David Teigland <teigland@redhat.com>
This patch adds functionality to add a lkb with a specific id range.
Signed-off-by: Alexander Aring <aahringo@redhat.com>
Signed-off-by: David Teigland <teigland@redhat.com>
This patch adds initial support for dlm tracepoints. It will introduce
tracepoints to dlm main functionality dlm_lock()/dlm_unlock() and their
complete ast() callback or blocking bast() callback.
The lock/unlock functionality has a start and end tracepoint, this is
because there exists a race in case if would have a tracepoint at the
end position only the complete/blocking callbacks could occur before. To
work with eBPF tracing and using their lookup hash functionality there
could be problems that an entry was not inserted yet. However use the
start functionality for hash insert and check again in end functionality
if there was an dlm internal error so there is no ast callback. In further
it might also that locks with local masters will occur those callbacks
immediately so we must have such functionality.
I did not make everything accessible yet, although it seems eBPF can be
used to access a lot of internal datastructures if it's aware of the
struct definitions of the running kernel instance. We still can change
it, if you do eBPF experiments e.g. time measurements between lock and
callback functionality you can simple use the local lkb_id field as hash
value in combination with the lockspace id if you have multiple
lockspaces. Otherwise you can simple use trace-cmd for some functionality,
e.g. `trace-cmd record -e dlm` and `trace-cmd report` afterwards.
Signed-off-by: Alexander Aring <aahringo@redhat.com>
Signed-off-by: David Teigland <teigland@redhat.com>
This patch adds union inside the lockspace id to handle it also for
another use case for a different dlm command.
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 uses GFP_ZERO for allocate a page for the internal dlm
sending buffer allocator instead of calling memset zero after every
allocation. An already allocated space will never be reused again.
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>
According to comment in dlm_user_request() ua should be freed
in dlm_free_lkb() after successful attach to lkb.
However ua is attached to lkb not in set_lock_args() but later,
inside request_lock().
Fixes 597d0cae0f ("[DLM] dlm: user locks")
Cc: stable@kernel.org # 2.6.19
Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
Signed-off-by: David Teigland <teigland@redhat.com>
This function was only for debugging. It would be
called in a condition that should not happen, and
should probably have been removed from the final
version of the original commit.
Remove it because it does mutex lock under spin lock.
Signed-off-by: David Teigland <teigland@redhat.com>
When the DLM_LKF_NODLCKWT flag was set, even if conversion deadlock
was detected, the caller of can_be_granted() was unknown.
We change the behavior of can_be_granted() and change it to detect
conversion deadlock regardless of whether the DLM_LKF_NODLCKWT flag
is set or not. And depending on whether the DLM_LKF_NODLCKWT flag
is set or not, we change the behavior at the caller of can_be_granted().
This fix has no effect except when using DLM_LKF_NODLCKWT flag.
Currently, ocfs2 uses the DLM_LKF_NODLCKWT flag and does not expect a
cancel operation from conversion deadlock when calling dlm_lock().
ocfs2 is implemented to perform a cancel operation by requesting
BASTs (callback).
Signed-off-by: Tadashi Miyauchi <miyauchi@toshiba-tops.co.jp>
Signed-off-by: Tsutomu Owa <tsutomu.owa@toshiba.co.jp>
Signed-off-by: David Teigland <teigland@redhat.com>
Replace the specification of a data structure by a pointer dereference
as the parameter for the operator "sizeof" to make the corresponding size
determination a bit safer according to the Linux coding style convention.
Signed-off-by: Markus Elfring <elfring@users.sourceforge.net>
Signed-off-by: David Teigland <teigland@redhat.com>
A multiplication for the size determination of a memory allocation
indicated that an array data structure should be processed.
Thus use the corresponding function "kcalloc".
This issue was detected by using the Coccinelle software.
Signed-off-by: Markus Elfring <elfring@users.sourceforge.net>
Signed-off-by: David Teigland <teigland@redhat.com>
No point in going through loops and hoops instead of just comparing the
values.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: Peter Zijlstra <peterz@infradead.org>
ktime_set(S,N) was required for the timespec storage type and is still
useful for situations where a Seconds and Nanoseconds part of a time value
needs to be converted. For anything where the Seconds argument is 0, this
is pointless and can be replaced with a simple assignment.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: Peter Zijlstra <peterz@infradead.org>
A process may exit, leaving an orphan lock in the lockspace.
This adds the capability for another process to acquire the
orphan lock. Acquiring the orphan just moves the lock from
the orphan list onto the acquiring process's list of locks.
An adopting process must specify the resource name and mode
of the lock it wants to adopt. If a matching lock is found,
the lock is moved to the caller's 's list of locks, and the
lkid of the lock is returned like the lkid of a new lock.
If an orphan with a different mode is found, then -EAGAIN is
returned. If no orphan lock is found on the resource, then
-ENOENT is returned. No async completion is used because
the result is immediately available.
Also, when orphans are purged, allow a zero nodeid to refer
to the local nodeid so the caller does not need to look up
the local nodeid.
Signed-off-by: David Teigland <teigland@redhat.com>
The log messages relating to the progress of recovery
are minimal and very often useful. Change these to
the KERN_INFO level so they are always available.
Signed-off-by: David Teigland <teigland@redhat.com>
We pass the freed "r" pointer back to the caller. It's harmless but it
upsets the static checkers.
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: David Teigland <teigland@redhat.com>
For lockspaces with an LVB length above 64 bytes, avoid truncating
the LVB while exchanging it with another node in the cluster.
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
Signed-off-by: David Teigland <teigland@redhat.com>
Convert to the much saner new idr interface. Error return values from
recover_idr_add() mix -1 and -errno. The conversion doesn't change
that but it looks iffy.
Signed-off-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Keep track of whether a toss list contains any
shrinkable rsbs. If not, dlm_scand can avoid
scanning the list for rsbs to shrink. Unnecessary
scanning can otherwise waste a lot of time because
the toss lists can contain a large number of rsbs
that are non-shrinkable (directory records).
Signed-off-by: David Teigland <teigland@redhat.com>
When a node is removed that held a PW/EX lock, the
existing master node should invalidate the lvb on the
resource due to the purged lock.
Previously, the existing master node was invalidating
the lvb if it found only NL/CR locks on the resource
during recovery for the removed node. This could lead
to cases where it invalidated the lvb and shouldn't
have, or cases where it should have invalidated and
didn't.
When recovery selects a *new* master node for a
resource, and that new master finds only NL/CR locks
on the resource after lock recovery, it should
invalidate the lvb. This case was handled correctly
(but was incorrectly applied to the existing master
case also.)
When a process exits while holding a PW/EX lock,
the lvb on the resource should be invalidated.
This was not happening.
The lvb contents and VALNOTVALID flag should be
recovered before granting locks in recovery so that
the recovered lvb state is provided in the callback.
The lvb was being recovered after the lock was granted.
Signed-off-by: David Teigland <teigland@redhat.com>
I don't know exactly how, but in some cases, a dir
record is not removed, or a new one is created when
it shouldn't be. The result is that the dir node
lookup returns a master node where the rsb does not
exist. In this case, The master node will repeatedly
return -EBADR for requests, and the lock requests will
be stuck.
Until all possible ways for this to happen can be
eliminated, a simple and effective way to recover from
this situation is for the supposed master node to send
a standard remove message to the dir node when it
receives a request for a resource it has no rsb for.
Signed-off-by: David Teigland <teigland@redhat.com>
The process of rebuilding locks on a new master during
recovery could re-order the locks on the convert queue,
creating an "in place" conversion deadlock that would
not be resolved. Fix this by not considering queue
order when granting conversions after recovery.
Signed-off-by: David Teigland <teigland@redhat.com>
It was possible for a remove message on an old
rsb to be sent after a lookup message on a new
rsb, where the rsbs were for the same resource
name. This could lead to a missing directory
entry for the new rsb.
It is fixed by keeping a copy of the resource
name being removed until after the remove has
been sent. A lookup checks if this in-progress
remove matches the name it is looking up.
Signed-off-by: David Teigland <teigland@redhat.com>
Remove the dir hash table (dirtbl), and use
the rsb hash table (rsbtbl) as the resource
directory. It has always been an unnecessary
duplication of information.
This improves efficiency by using a single rsbtbl
lookup in many cases where both rsbtbl and dirtbl
lookups were needed previously.
This eliminates the need to handle cases of rsbtbl
and dirtbl being out of sync.
In many cases there will be memory savings because
the dir hash table no longer exists.
Signed-off-by: David Teigland <teigland@redhat.com>
The "nodir" mode (statically assign master nodes instead
of using the resource directory) has always been highly
experimental, and never seriously used. This commit
fixes a number of problems, making nodir much more usable.
- Major change to recovery: recover all locks and restart
all in-progress operations after recovery. In some
cases it's not possible to know which in-progess locks
to recover, so recover all. (Most require recovery
in nodir mode anyway since rehashing changes most
master nodes.)
- Change the way nodir mode is enabled, from a command
line mount arg passed through gfs2, into a sysfs
file managed by dlm_controld, consistent with the
other config settings.
- Allow recovering MSTCPY locks on an rsb that has not
yet been turned into a master copy.
- Ignore RCOM_LOCK and RCOM_LOCK_REPLY recovery messages
from a previous, aborted recovery cycle. Base this
on the local recovery status not being in the state
where any nodes should be sending LOCK messages for the
current recovery cycle.
- Hold rsb lock around dlm_purge_mstcpy_locks() because it
may run concurrently with dlm_recover_master_copy().
- Maintain highbast on process-copy lkb's (in addition to
the master as is usual), because the lkb can switch
back and forth between being a master and being a
process copy as the master node changes in recovery.
- When recovering MSTCPY locks, flag rsb's that have
non-empty convert or waiting queues for granting
at the end of recovery. (Rename flag from LOCKS_PURGED
to RECOVER_GRANT and similar for the recovery function,
because it's not only resources with purged locks
that need grant a grant attempt.)
- Replace a couple of unnecessary assertion panics with
error messages.
Signed-off-by: David Teigland <teigland@redhat.com>
Change some existing error/debug messages to
collect more useful information, and add
some new error/debug messages to address
recently found problems.
Signed-off-by: David Teigland <teigland@redhat.com>
If the rsb is found in the "keep" tree, but is
not the right type (i.e. not MASTER), we can
return immediately with the result. There's
no point in going on to search the "toss" list
as if we hadn't found it.
Signed-off-by: David Teigland <teigland@redhat.com>
An outstanding remote operation (an lkb on the "waiter"
list) could sometimes miss being resent during recovery.
The decision was based on the lkb_nodeid field, which
could have changed during an earlier aborted recovery,
so it no longer represents the actual remote destination.
The lkb_wait_nodeid is always the actual remote node,
so it is the best value to use.
Signed-off-by: David Teigland <teigland@redhat.com>
The QUECVT flag should not prevent conversions from
being granted immediately when the convert queue is
empty.
Signed-off-by: David Teigland <teigland@redhat.com>
The function used to find an rsb during directory
recovery was searching the single linear list of
rsb's. This wasted a lot of time compared to
using the standard hash table to find the rsb.
Signed-off-by: David Teigland <teigland@redhat.com>
Change the linked lists to rb_tree's in the rsb
hash table to speed up searches. Slow rsb searches
were having a large impact on gfs2 performance due
to the large number of dlm locks gfs2 uses.
Signed-off-by: Bob Peterson <rpeterso@redhat.com>
Signed-off-by: David Teigland <teigland@redhat.com>
Instead of creating our own kthread (dlm_astd) to deliver
callbacks for all lockspaces, use a per-lockspace workqueue
to deliver the callbacks. This eliminates complications and
slowdowns from many lockspaces sharing the same thread.
Signed-off-by: David Teigland <teigland@redhat.com>
By pre-allocating rsb structs before searching the hash
table, they can be inserted immediately. This avoids
always having to repeat the search when adding the struct
to hash list.
This also adds space to the rsb struct for a max resource
name, so an rsb allocation can be used by any request.
The constant size also allows us to finally use a slab
for the rsb structs.
Signed-off-by: David Teigland <teigland@redhat.com>
This is simpler and quicker than the hash table, and
avoids needing to search the hash list for every new
lkid to check if it's used.
Signed-off-by: David Teigland <teigland@redhat.com>
In fs/dlm/lock.c in the dlm_scan_waiters() function there are 3 small
issues:
1) There's no need to test the return value of the allocation and do a
memset if is succeedes. Just use kzalloc() to obtain zeroed memory.
2) Since kfree() handles NULL pointers gracefully, the test of
'warned' against NULL before the kfree() after the loop is completely
pointless. Remove it.
3) The arguments to kmalloc() (now kzalloc()) were swapped. Thanks to
Dr. David Alan Gilbert for pointing this out.
Signed-off-by: Jesper Juhl <jj@chaosbits.net>
Signed-off-by: David Teigland <teigland@redhat.com>
kmalloc a stub message struct during recovery instead of sharing the
struct in the lockspace. This leaves the lockspace stub_ms only for
faking downconvert replies, where it is never modified and sharing
is not a problem.
Also improve the debug messages in the same recovery function.
Signed-off-by: David Teigland <teigland@redhat.com>
Add an option (disabled by default) to print a warning message
when a lock has been waiting a configurable amount of time for
a reply message from another node. This is mainly for debugging.
Signed-off-by: David Teigland <teigland@redhat.com>
Change how callbacks are recorded for locks. Previously, information
about multiple callbacks was combined into a couple of variables that
indicated what the end result should be. In some situations, we
could not tell from this combined state what the exact sequence of
callbacks were, and would end up either delivering the callbacks in
the wrong order, or suppress redundant callbacks incorrectly. This
new approach records all the data for each callback, leaving no
uncertainty about what needs to be delivered.
Signed-off-by: David Teigland <teigland@redhat.com>
Alexander Aring
Gustavo A. R. Silva
Thomas Gleixner
Vasily Averin
David Teigland
tsutomu.owa@toshiba.co.jp
Markus Elfring
Dan Carpenter
Bart Van Assche
Tejun Heo
Bob Peterson
Jesper Juhl
Linus Torvalds
Lucas De Marchi