Re: [GIT PULL rcu/next] RCU commits for 4.1

[Ingo Molnar <mingo@kernel.org>]

* Paul E. McKenney <paulmck@linux.vnet.ibm.com> wrote:

> Hello, Ingo,
> 
> The changes in this series include the following, but please see the
> note on set 5:
> 
> 1.	Documentation updates.  These were posted to LKML at
> 	https://lkml.org/lkml/2015/3/3/493.
> 
> 2.	Changes permitting use of call_rcu() and friends very early in
> 	boot, for example, before rcu_init() is invoked.  These were
> 	posted to LKML at https://lkml.org/lkml/2015/3/3/510.
> 
> 3.	Miscellaneous fixes.  These were posted to LKML at
> 	https://lkml.org/lkml/2015/3/3/538.
> 
> 4.	In-kernel API to enable and disable expediting of normal RCU
> 	grace periods.  These were posted to LKML at
> 	https://lkml.org/lkml/2015/3/3/557.
> 
> 5.	Improve RCU's handling of outgoing CPUs.  These were posted to
> 	LKML at https://lkml.org/lkml/2015/3/3/578.
> 
> 	Note:  ARM support is lagging a bit.  This might require that
> 	this particular set of commits be deferred.  I am currently
> 	optimistic, but please let me know if you would like to handle
> 	these some other way.
> 
> 6.	NO_HZ_FULL_SYSIDLE fixes.  These were posted to LKML at
> 	https://lkml.org/lkml/2015/3/3/607.
> 
> 7.	Tiny RCU updates to make it more tiny.  These were posted to LKML
> 	at https://lkml.org/lkml/2015/3/3/618.
> 
> These changes (based on v4.0-rc1) are available in the git repository at:
> 
>   git://git.kernel.org/pub/scm/linux/kernel/git/paulmck/linux-rcu.git for-mingo
> 
> for you to fetch changes up to 42528795ac1c8d7ba021797ec004904168956d64:
> 
>   Merge branches 'doc.2015.02.26a', 'earlycb.2015.03.03a', 'fixes.2015.03.03a', 'gpexp.2015.02.26a', 'hotplug.2015.03.20a', 'sysidle.2015.02.26b' and 'tiny.2015.02.26a' into HEAD (2015-03-20 08:31:01 -0700)
> 
> ----------------------------------------------------------------
> 
> Alexander Gordeev (2):
>       rcu: Remove unnecessary condition check in rcu_qsctr_help()
>       rcu: Remove fastpath from __rcu_process_callbacks()
> 
> Lai Jiangshan (1):
>       rcu_tree: Avoid touching rnp->completed when a new GP is started
> 
> Paul E. McKenney (53):
>       rcu: Consolidate rcu_synchronize and wakeme_after_rcu()
>       documentation: Update rcutree.kthread_prio for grace-period kthread use
>       documentation: Update based on on-demand vmstat workers
>       documentation: Update NO_HZ_FULL interaction with POSIX timers
>       documentation: Update per-CPU kthreads documentation
>       documentation: Clarify memory-barrier semantics of atomic operations
>       documentation: Clarify control-dependency pairing
>       rcu: Abstract default callback-list initialization from init_callback_list()
>       rcu: Wire ->rda pointers at compile time
>       rcu: Prevent early-boot RCU callbacks from splatting
>       rcu: Refine diagnostics for lacking kthread for no-CBs callbacks
>       rcu: Avoid clobbering early boot callbacks
>       rcu: Fix a couple of typos in rcu_all_qs() comment header
>       rcu: Drive PROVE_RCU directly off of PROVE_LOCKING
>       rcu: Provide rcu_expedite_gp() and rcu_unexpedite_gp()
>       rcu: Add rcu_expedite_gp() and rcu_unexpedite_gp() to rcutorture
>       rcu: Update from rcu_expedited variable to rcu_gp_is_expedited()
>       rcu: Add Kconfig option to expedite grace periods during boot
>       rcutorture: Make consistent use of variables
>       rcu: Fixes to NO_HZ_FULL sysidle accounting
>       rcu: Tighten up affinity and check for sysidle
>       rcu: Move early-boot callbacks to no-CBs lists for no-CBs CPUs
>       rcu: Move early boot callback tests earlier
>       rcu: Use IS_ENABLED() to CONFIG_RCU_FANOUT_EXACT #ifdef
>       rcu: Improve diagnostics for blocked critical sections in irq
>       rcu: Use IS_ENABLED() to simplify rcu_bootup_announce_oddness()
>       rcu: Add boot-up check for non-default CONFIG_RCU_FANOUT_LEAF values
>       rcu: Reverse rcu_dereference_check() conditions
>       torture: Avoid script syntax error when insufficient CPUs
>       rcu: Get rcu_sched_force_quiescent_state() where it belongs
>       smpboot: Add common code for notification from dying CPU
>       x86: Use common outgoing-CPU-notification code
>       blackfin: Use common outgoing-CPU-notification code
>       metag: Use common outgoing-CPU-notification code
>       rcu: Consolidate offline-CPU callback initialization
>       rcu: Put all orphan-callback-related code under same comment
>       rcu: Simplify sync_rcu_preempt_exp_init()
>       rcu: Eliminate empty HOTPLUG_CPU ifdef
>       rcu: Detect stalls caused by failure to propagate up rcu_node tree
>       rcu: Provide diagnostic option to slow down grace-period initialization
>       rcutorture: Enable slow grace-period initializations
>       rcu: Remove event tracing from rcu_cpu_notify(), used by offline CPUs
>       rcu: Rework preemptible expedited bitmask handling
>       rcu: Move rcu_report_unblock_qs_rnp() to common code
>       rcu: Process offlining and onlining only at grace-period start
>       rcu: Eliminate ->onoff_mutex from rcu_node structure
>       cpu: Make CPU-offline idle-loop transition point more precise
>       rcu: Handle outgoing CPUs on exit from idle loop
>       rcutorture: Default to grace-period-initialization delays
>       rcu: Add diagnostics to grace-period cleanup
>       rcu: Yet another fix for preemption and CPU hotplug
>       rcu: Associate quiescent-state reports with grace period
>       Merge branches 'doc.2015.02.26a', 'earlycb.2015.03.03a', 'fixes.2015.03.03a', 'gpexp.2015.02.26a', 'hotplug.2015.03.20a', 'sysidle.2015.02.26b' and 'tiny.2015.02.26a' into HEAD
> 
> Yao Dongdong (1):
>       rcu: Remove redundant check of cpu_online()
> 
>  Documentation/atomic_ops.txt                       |  45 +--
>  Documentation/kernel-parameters.txt                |  20 +-
>  Documentation/kernel-per-CPU-kthreads.txt          |  34 +-
>  Documentation/memory-barriers.txt                  |  42 +-
>  Documentation/timers/NO_HZ.txt                     |  10 +-
>  arch/blackfin/mach-common/smp.c                    |   6 +-
>  arch/metag/kernel/smp.c                            |   5 +-
>  arch/x86/include/asm/cpu.h                         |   2 -
>  arch/x86/include/asm/smp.h                         |   2 +-
>  arch/x86/kernel/smpboot.c                          |  39 +-
>  arch/x86/xen/smp.c                                 |  46 ++-
>  include/linux/cpu.h                                |  14 +
>  include/linux/lockdep.h                            |   7 +-
>  include/linux/rcupdate.h                           |  40 +-
>  include/linux/srcu.h                               |   2 +-
>  init/Kconfig                                       |  13 +
>  kernel/cpu.c                                       |   4 +-
>  kernel/rcu/rcutorture.c                            |  27 +-
>  kernel/rcu/srcu.c                                  |  19 +-
>  kernel/rcu/tiny.c                                  |  14 +-
>  kernel/rcu/tree.c                                  | 437 +++++++++++++++------
>  kernel/rcu/tree.h                                  |  11 +-
>  kernel/rcu/tree_plugin.h                           | 267 +++++++------
>  kernel/rcu/tree_trace.c                            |   4 +-
>  kernel/rcu/update.c                                |  72 +++-
>  kernel/sched/idle.c                                |   9 +-
>  kernel/smpboot.c                                   | 156 ++++++++
>  lib/Kconfig.debug                                  |  35 +-
>  tools/testing/selftests/rcutorture/bin/kvm.sh      |   2 +-
>  .../selftests/rcutorture/configs/rcu/CFcommon      |   1 +
>  30 files changed, 948 insertions(+), 437 deletions(-)

Pulled into tip:core/rcu, thanks a lot Paul!

	Ingo