[patch V5 00/72] locking, sched: The PREEMPT-RT locking infrastructure

From: Thomas Gleixner
Date: Sun Aug 15 2021 - 17:28:51 EST


Folks,

the following series is an update to V4 which can be found here:

https://lore.kernel.org/r/20210811120348.855823694@xxxxxxxxxxxxx

It contains the bulk of the PREEMPT-RT locking infrastructure. In
PREEMPT-RT enabled kernels the following locking primitives are substituted
by RT-Mutex based variants:

mutex, ww_mutex, rw_semaphore, spinlock, rwlock

semaphores are not substituted because they do not provide strict owner
semantics.

Of course raw_spinlocks are not touched either as they protect low level
operations in the scheduler, timers and hardware access.

The most interesting parts of the series which need a lot of eyeballs
are:

- the scheduler bits which provide the infrastructure for spinlock and
rwlock substitution to ensure that the task state is preserved when
blocking on such a lock and a regular wakeup is handled correctly and
not lost

- the rtmutex core implementation to handle lock contention on spinlocks
and rwlocks correctly vs. the task state

- the rw_semaphore/rwlock substitutions which utilize the same
implementation vs. the reader/writer handling

- The new rtmutex based ww_mutex implementation.

- the PI futex related bits to handle the interaction between blocking
on the underlying rtmutex and contention on the hash bucket lock which
is converted to a 'sleeping spinlock'.

The rest surely needs a thorough review as well, but those parts are pretty
straight forward: quite some code restructuring and the actual wrapper
functions to replace the existing !RT implementations.

The series survived internal testing in RT kernels and is part of the upcoming
v5.14-rc6-rt9 release.

For !RT kernels there is no functional change.

The series is also available from git:

git://git.kernel.org/pub/scm/linux/kernel/git/tglx/devel.git rtmutex

Changes vs. V4

- A few lockdep wait_type fixes in static and runtime initializers.
Sebastian noticed while working on getting the lockdep selftests
reenabled on RT.

- Missing might_sleep() invocations in RT spin/rwlocks (Sebastian)

- Add explicit owner init for local locks when lockdep is enabled instead
of relying on zero initialized memory.

- Add the RT variants for local locks, which is the last lock type
getting special treatment on RT.

The lockdep selftest changes are not yet ready and will be posted in a
separate submission.


Thanks,

tglx
---
b/drivers/staging/media/atomisp/pci/atomisp_ioctl.c | 4
b/include/linux/debug_locks.h | 3
b/include/linux/local_lock_internal.h | 90 +
b/include/linux/mutex.h | 92 +
b/include/linux/preempt.h | 4
b/include/linux/rbtree.h | 30
b/include/linux/rbtree_types.h | 34
b/include/linux/rtmutex.h | 63 -
b/include/linux/rwbase_rt.h | 38
b/include/linux/rwlock_rt.h | 140 ++
b/include/linux/rwlock_types.h | 53
b/include/linux/rwsem.h | 78 +
b/include/linux/sched.h | 119 +-
b/include/linux/sched/wake_q.h | 8
b/include/linux/spinlock.h | 15
b/include/linux/spinlock_api_smp.h | 3
b/include/linux/spinlock_rt.h | 159 ++
b/include/linux/spinlock_types.h | 89 -
b/include/linux/spinlock_types_raw.h | 73 +
b/include/linux/ww_mutex.h | 50
b/kernel/Kconfig.locks | 2
b/kernel/futex.c | 556 ++++++---
b/kernel/locking/Makefile | 3
b/kernel/locking/mutex-debug.c | 5
b/kernel/locking/mutex.c | 431 -------
b/kernel/locking/mutex.h | 48
b/kernel/locking/rtmutex.c | 1134 +++++++++-----------
b/kernel/locking/rtmutex_api.c | 590 ++++++++++
b/kernel/locking/rtmutex_common.h | 122 +-
b/kernel/locking/rwbase_rt.c | 263 ++++
b/kernel/locking/rwsem.c | 109 +
b/kernel/locking/spinlock.c | 7
b/kernel/locking/spinlock_debug.c | 5
b/kernel/locking/spinlock_rt.c | 263 ++++
b/kernel/locking/ww_mutex.h | 569 ++++++++++
b/kernel/locking/ww_rt_mutex.c | 76 +
b/kernel/rcu/tree_plugin.h | 6
b/kernel/sched/core.c | 109 +
b/lib/Kconfig.debug | 11
b/lib/test_lockup.c | 8
kernel/locking/mutex-debug.h | 29
41 files changed, 3940 insertions(+), 1551 deletions(-)