The current vnode layout is not smp-friendly by having frequently read data
avoidably sharing cachelines with very frequently modified fields. In
particular v_iflag inspected for VI_DOOMED can be found in the same line with
v_usecount. Instead make it available in the same cacheline as the v_op, v_data
and v_type which all get read all the time.
v_type is avoidably 4 bytes while the necessary data will easily fit in 1.
Shrinking it frees up 3 bytes, 2 of which get used here to introduce a new
flag field with a new value: VIRF_DOOMED.
Reviewed by: kib, jeff
Differential Revision: https://reviews.freebsd.org/D22715
This combined with previous changes significantly depessimizes the behaviour
under contentnion.
In particular the lock1_processes test (locking/unlocking separate files)
from the will-it-scale suite was executed with 128 concurrency on a
4-socket Broadwell with 128 hardware threads.
Operations/second (lock+unlock) go from ~750000 to ~45000000 (6000%)
For reference single-process is ~1680000 (i.e. on stock kernel the resulting
perf is less than *half* of the single-threaded run),
Note this still does not really scale all that well as the locks were just
bolted on top of the current implementation. Significant room for improvement
is still here. In particular the top performance fluctuates depending on the
extent of false sharing in given run (which extends beyond the file).
Added chain+lock pairs were not padded w.r.t. cacheline size.
One big ticket item is the hash used for spreading threads: it used to be the
process pid (which basically serialized all threaded ops). Temporarily the
vnode addr was slapped in instead.
Tested by: pho
1. check if P_ADVLOCK is already set and if so, don't lock to set it
(stolen from DragonFly)
2. when trying for fast path unlock, check that we are doing unlock
first instead of taking the interlock for no reason (e.g. if we want
to *lock*). whilere make it more likely that falling fast path will
not take the interlock either by checking for state
Note the code is severely pessimized both single- and multithreaded.
Mainly focus on files that use BSD 3-Clause license.
The Software Package Data Exchange (SPDX) group provides a specification
to make it easier for automated tools to detect and summarize well known
opensource licenses. We are gradually adopting the specification, noting
that the tags are considered only advisory and do not, in any way,
superceed or replace the license texts.
Special thanks to Wind River for providing access to "The Duke of
Highlander" tool: an older (2014) run over FreeBSD tree was useful as a
starting point.
The checks have quadratic complexity over a number of advisory locks
active for a file and that could be a lot. What's the worse is that the
checks are done while holding ls_lock. That could lead to a long a very
long backlog and performance degradation even if all requested locks are
compatible (e.g. all shared locks).
The checks used to be under INVARIANTS.
Discussed with: kib
MFC after: 2 weeks
Sponsored by: Panzura
interrupt sleeps with the ERESTART on the suspension attempts.
Otherwise, single-threading requests are deferred until the locks are
granted for NFS files, which causes hangs.
When retrying local registration of the remotely-granted adv lock,
allow full suspension and check for suspension, for usual reasons.
Reported by: markj, pho
Reviewed by: jilles
Tested by: pho
Sponsored by: The FreeBSD Foundation
MFC after: 2 weeks
Approved by: re (gjb)
by reinitializing the 'freestate' pointer after freeing the memory.
Obtained from: HardenedBSD (71fab80c5dd3034b71a29a61064625018671bbeb)
PR: 194525
Submitted by: Oliver Pinter <oliver.pinter@hardenedbsd.org>
MFC after: 2 weeks
tools/regression/file/flock/flock.c, which completes the fix in
r192685. When the lock was stolen from us, retry the whole lock
sequence in kernel, instead of returning EINTR to usermode and hoping
that application would handle it correctly by restarting the lock
acquire.
Tested by: pho
Sponsored by: The FreeBSD Foundation
MFC after: 2 weeks
sleep waiting for conditions when the lock may be granted.
To prevent lf_setlock() from accessing possibly freed memory, add reference
counting to the struct lockf_entry. Bump refcount around the sleep.
Make lf_free_lock() return non-zero when structure was freed, and use
this after the sleep to return EINTR to the caller. The error code might
need a clarification, but we cannot return success to usermode, since
the lock is not owned anymore.
Reviewed by: dfr
Tested by: pho
MFC after: 1 month
is acquired. In the lf_purgelocks(), assert that vnode is doomed and set
*statep to NULL before clearing ls_pending list. Otherwise, we allow for
the thread executing lf_advlockasync() to put new pending entry after
state->ls_lock is dropped in lf_purgelocks().
Reviewed by: dfr
Tested by: pho
MFC after: 1 month
provides the correct semantics for flock(2) style locks which are used by the
lockf(1) command line tool and the pidfile(3) library. It also implements
recovery from server restarts and ensures that dirty cache blocks are written
to the server before obtaining locks (allowing multiple clients to use file
locking to safely share data).
Sponsored by: Isilon Systems
PR: 94256
MFC after: 2 weeks
filesystem-specific vnode data to the struct vnode. Provide the
default implementation for the vop_advlock and vop_advlockasync.
Purge the locks on the vnode reclaim by using the lf_purgelocks().
The default implementation is augmented for the nfs and smbfs.
In the nfs_advlock, push the Giant inside the nfs_dolock.
Before the change, the vop_advlock and vop_advlockasync have taken the
unlocked vnode and dereferenced the fs-private inode data, racing with
with the vnode reclamation due to forced unmount. Now, the vop_getattr
under the shared vnode lock is used to obtain the inode size, and
later, in the lf_advlockasync, after locking the vnode interlock, the
VI_DOOMED flag is checked to prevent an operation on the doomed vnode.
The implementation of the lf_purgelocks() is submitted by dfr.
Reported by: kris
Tested by: kris, pho
Discussed with: jeff, dfr
MFC after: 2 weeks
user-mode lock manager, build a kernel with the NFSLOCKD option and
add '-k' to 'rpc_lockd_flags' in rc.conf.
Highlights include:
* Thread-safe kernel RPC client - many threads can use the same RPC
client handle safely with replies being de-multiplexed at the socket
upcall (typically driven directly by the NIC interrupt) and handed
off to whichever thread matches the reply. For UDP sockets, many RPC
clients can share the same socket. This allows the use of a single
privileged UDP port number to talk to an arbitrary number of remote
hosts.
* Single-threaded kernel RPC server. Adding support for multi-threaded
server would be relatively straightforward and would follow
approximately the Solaris KPI. A single thread should be sufficient
for the NLM since it should rarely block in normal operation.
* Kernel mode NLM server supporting cancel requests and granted
callbacks. I've tested the NLM server reasonably extensively - it
passes both my own tests and the NFS Connectathon locking tests
running on Solaris, Mac OS X and Ubuntu Linux.
* Userland NLM client supported. While the NLM server doesn't have
support for the local NFS client's locking needs, it does have to
field async replies and granted callbacks from remote NLMs that the
local client has contacted. We relay these replies to the userland
rpc.lockd over a local domain RPC socket.
* Robust deadlock detection for the local lock manager. In particular
it will detect deadlocks caused by a lock request that covers more
than one blocking request. As required by the NLM protocol, all
deadlock detection happens synchronously - a user is guaranteed that
if a lock request isn't rejected immediately, the lock will
eventually be granted. The old system allowed for a 'deferred
deadlock' condition where a blocked lock request could wake up and
find that some other deadlock-causing lock owner had beaten them to
the lock.
* Since both local and remote locks are managed by the same kernel
locking code, local and remote processes can safely use file locks
for mutual exclusion. Local processes have no fairness advantage
compared to remote processes when contending to lock a region that
has just been unlocked - the local lock manager enforces a strict
first-come first-served model for both local and remote lockers.
Sponsored by: Isilon Systems
PR: 95247 107555 115524 116679
MFC after: 2 weeks
1.38 in 2001. Break out of the FOREACH_THREAD_IN_PROC loop when we've
discovered a new proc in the chain.
- Increment i and check for maxlockdepth once per matching process not
once per thread. This didn't properly terminate the loop before.
- Fix a bug which has existed potentially since rev 1.1. waitblock->lf_next
can be NULL when a thread has been woken-up but not yet scheduled. Check
for this condition rather than blindly dereferencing.
Found by: libMicro
requiring the per-process spinlock to only requiring the process lock.
- Reflect these changes in the proc.h documentation and consumers throughout
the kernel. This is a substantial reduction in locking cost for these
fields and was made possible by recent changes to threading support.
to protect this datastructure instead.
- Preallocate an extra lockf structure in case we want to split a lock
on insert or delete.
- msleep() on the vnode interlock when blocking on a lock.
Reviewed by: rwatson
Approved by: re
- Use thread_lock() rather than sched_lock for per-thread scheduling
sychronization.
- Use the per-process spinlock rather than the sched_lock for per-process
scheduling synchronization.
Tested by: kris, current@
Tested on: i386, amd64, ULE, 4BSD, libthr, libkse, PREEMPTION, etc.
Discussed with: kris, attilio, kmacy, jhb, julian, bde (small parts each)
