CPUs running for long time periods in the kernel in nohz_full mode
might leave the scheduling-clock interrupt disabled for then full
duration of their in-kernel execution. This can (among other things)
delay grace periods. This commit therefore forces the tick back on
for any nohz_full CPU that is failing to pass through a quiescent state
upon return from interrupt, which the resched_cpu() will induce.
Reported-by: Joel Fernandes <joel@joelfernandes.org>
[ paulmck: Clear ->rcu_forced_tick as reported by Joel Fernandes testing. ]
[ paulmck: Apply Joel Fernandes TICK_DEP_MASK_RCU->TICK_DEP_BIT_RCU fix. ]
Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
This commit causes locking, sleeping, and callback state to be printed
for no-CBs CPUs when the rcutorture writer is delayed sufficiently for
rcutorture to complain.
Signed-off-by: Paul E. McKenney <paulmck@linux.ibm.com>
Use of the rcu_data structure's segmented ->cblist for no-CBs CPUs
takes advantage of unrelated grace periods, thus reducing the memory
footprint in the face of floods of call_rcu() invocations. However,
the ->cblist field is a more-complex rcu_segcblist structure which must
be protected via locking. Even though there are only three entities
which can acquire this lock (the CPU invoking call_rcu(), the no-CBs
grace-period kthread, and the no-CBs callbacks kthread), the contention
on this lock is excessive under heavy stress.
This commit therefore greatly reduces contention by provisioning
an rcu_cblist structure field named ->nocb_bypass within the
rcu_data structure. Each no-CBs CPU is permitted only a limited
number of enqueues onto the ->cblist per jiffy, controlled by a new
nocb_nobypass_lim_per_jiffy kernel boot parameter that defaults to
about 16 enqueues per millisecond (16 * 1000 / HZ). When that limit is
exceeded, the CPU instead enqueues onto the new ->nocb_bypass.
The ->nocb_bypass is flushed into the ->cblist every jiffy or when
the number of callbacks on ->nocb_bypass exceeds qhimark, whichever
happens first. During call_rcu() floods, this flushing is carried out
by the CPU during the course of its call_rcu() invocations. However,
a CPU could simply stop invoking call_rcu() at any time. The no-CBs
grace-period kthread therefore carries out less-aggressive flushing
(every few jiffies or when the number of callbacks on ->nocb_bypass
exceeds (2 * qhimark), whichever comes first). This means that the
no-CBs grace-period kthread cannot be permitted to do unbounded waits
while there are callbacks on ->nocb_bypass. A ->nocb_bypass_timer is
used to provide the needed wakeups.
[ paulmck: Apply Coverity feedback reported by Colin Ian King. ]
Signed-off-by: Paul E. McKenney <paulmck@linux.ibm.com>
The sleep/wakeup of the no-CBs grace-period kthreads is synchronized
using the ->nocb_lock of the first CPU corresponding to that kthread.
This commit provides a separate ->nocb_gp_lock for this purpose, thus
reducing contention on ->nocb_lock.
Signed-off-by: Paul E. McKenney <paulmck@linux.ibm.com>
A given rcu_data structure's ->nocb_lock can be acquired very frequently
by the corresponding CPU and occasionally by the corresponding no-CBs
grace-period and callbacks kthreads. In particular, these two kthreads
will have frequent gaps between ->nocb_lock acquisitions that are roughly
a grace period in duration. This means that any excessive ->nocb_lock
contention will be due to the CPU's acquisitions, and this in turn
enables a very naive contention-avoidance strategy to be quite effective.
This commit therefore modifies rcu_nocb_lock() to first
attempt a raw_spin_trylock(), and to atomically increment a
separate ->nocb_lock_contended across a raw_spin_lock(). This new
->nocb_lock_contended field is checked in __call_rcu_nocb_wake() when
interrupts are enabled, with a spin-wait for contending acquisitions
to complete, thus allowing the kthreads a chance to acquire the lock.
Signed-off-by: Paul E. McKenney <paulmck@linux.ibm.com>
This commit removes the obsolete nocb_q_count and nocb_q_count_lazy
fields, also removing rcu_get_n_cbs_nocb_cpu(), adjusting
rcu_get_n_cbs_cpu(), and making rcutree_migrate_callbacks() once again
disable the ->cblist fields of offline CPUs.
Signed-off-by: Paul E. McKenney <paulmck@linux.ibm.com>
Currently the RCU callbacks for no-CBs CPUs are queued on a series of
ad-hoc linked lists, which means that these callbacks cannot benefit
from "drive-by" grace periods, thus suffering needless delays prior
to invocation. In addition, the no-CBs grace-period kthreads first
wait for callbacks to appear and later wait for a new grace period,
which means that callbacks appearing during a grace-period wait can
be delayed. These delays increase memory footprint, and could even
result in an out-of-memory condition.
This commit therefore enqueues RCU callbacks from no-CBs CPUs on the
rcu_segcblist structure that is already used by non-no-CBs CPUs. It also
restructures the no-CBs grace-period kthread to be checking for incoming
callbacks while waiting for grace periods. Also, instead of waiting
for a new grace period, it waits for the closest grace period that will
cause some of the callbacks to be safe to invoke. All of these changes
reduce callback latency and thus the number of outstanding callbacks,
in turn reducing the probability of an out-of-memory condition.
Signed-off-by: Paul E. McKenney <paulmck@linux.ibm.com>
As a first step towards making no-CBs CPUs use the ->cblist, this commit
leaves the ->cblist enabled for these CPUs. The main reason to make
no-CBs CPUs use ->cblist is to take advantage of callback numbering,
which will reduce the effects of missed grace periods which in turn will
reduce forward-progress problems for no-CBs CPUs.
Signed-off-by: Paul E. McKenney <paulmck@linux.ibm.com>
Currently, there is one no-CBs rcuo kthread per CPU, and these kthreads
are divided into groups. The first rcuo kthread to come online in a
given group is that group's leader, and the leader both waits for grace
periods and invokes its CPU's callbacks. The non-leader rcuo kthreads
only invoke callbacks.
This works well in the real-time/embedded environments for which it was
intended because such environments tend not to generate all that many
callbacks. However, given huge floods of callbacks, it is possible for
the leader kthread to be stuck invoking callbacks while its followers
wait helplessly while their callbacks pile up. This is a good recipe
for an OOM, and rcutorture's new callback-flood capability does generate
such OOMs.
One strategy would be to wait until such OOMs start happening in
production, but similar OOMs have in fact happened starting in 2018.
It would therefore be wise to take a more proactive approach.
This commit therefore features per-CPU rcuo kthreads that do nothing
but invoke callbacks. Instead of having one of these kthreads act as
leader, each group has a separate rcog kthread that handles grace periods
for its group. Because these rcuog kthreads do not invoke callbacks,
callback floods on one CPU no longer block callbacks from reaching the
rcuc callback-invocation kthreads on other CPUs.
This change does introduce additional kthreads, however:
1. The number of additional kthreads is about the square root of
the number of CPUs, so that a 4096-CPU system would have only
about 64 additional kthreads. Note that recent changes
decreased the number of rcuo kthreads by a factor of two
(CONFIG_PREEMPT=n) or even three (CONFIG_PREEMPT=y), so
this still represents a significant improvement on most systems.
2. The leading "rcuo" of the rcuog kthreads should allow existing
scripting to affinity these additional kthreads as needed, the
same as for the rcuop and rcuos kthreads. (There are no longer
any rcuob kthreads.)
3. A state-machine approach was considered and rejected. Although
this would allow the rcuo kthreads to continue their dual
leader/follower roles, it complicates callback invocation
and makes it more difficult to consolidate rcuo callback
invocation with existing softirq callback invocation.
The introduction of rcuog kthreads should thus be acceptable.
Signed-off-by: Paul E. McKenney <paulmck@linux.ibm.com>
This commit simply rewords comments to prepare for leader nocb kthreads
doing only grace-period work and callback shuffling. This will mean
the addition of replacement kthreads to invoke callbacks. The "leader"
and "follower" thus become less meaningful, so the commit changes no-CB
comments with these strings to "GP" and "CB", respectively. (Give or
take the usual grammatical transformations.)
Signed-off-by: Paul E. McKenney <paulmck@linux.ibm.com>
This commit simply renames rcu_data fields to prepare for leader
nocb kthreads doing only grace-period work and callback shuffling.
This will mean the addition of replacement kthreads to invoke callbacks.
The "leader" and "follower" thus become less meaningful, so the commit
changes no-CB fields with these strings to "gp" and "cb", respectively.
Signed-off-by: Paul E. McKenney <paulmck@linux.ibm.com>
The rcu_data structure's ->deferred_qs field is used to indicate that the
current CPU is blocking an expedited grace period (perhaps a future one).
Given that it is used only for expedited grace periods, its current name
is misleading, so this commit renames it to ->exp_deferred_qs.
Signed-off-by: Paul E. McKenney <paulmck@linux.ibm.com>
When rcu_read_unlock_special() is invoked with interrupts disabled, is
either not in an interrupt handler or is not using RCU_SOFTIRQ, is not
the first RCU read-side critical section in the chain, and either there
is an expedited grace period in flight or this is a NO_HZ_FULL kernel,
the end of the grace period can be unduly delayed. The reason for this
is that it is not safe to do wakeups in this situation.
This commit fixes this problem by using the irq_work subsystem to
force a later interrupt handler in a clean environment. Because
set_tsk_need_resched(current) and set_preempt_need_resched() are
invoked prior to this, the scheduler will force a context switch
upon return from this interrupt (though perhaps at the end of any
interrupted preempt-disable or BH-disable region of code), which will
invoke rcu_note_context_switch() (again in a clean environment), which
will in turn give RCU the chance to report the deferred quiescent state.
Of course, by then this task might be within another RCU read-side
critical section. But that will be detected at that time and reporting
will be further deferred to the outermost rcu_read_unlock(). See
rcu_preempt_need_deferred_qs() and rcu_preempt_deferred_qs() for more
details on the checking.
Suggested-by: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Paul E. McKenney <paulmck@linux.ibm.com>
Some workloads need to change kthread priority for RCU core processing
without affecting other softirq work. This commit therefore introduces
the rcutree.use_softirq kernel boot parameter, which moves the RCU core
work from softirq to a per-CPU SCHED_OTHER kthread named rcuc. Use of
SCHED_OTHER approach avoids the scalability problems that appeared
with the earlier attempt to move RCU core processing to from softirq
to kthreads. That said, kernels built with RCU_BOOST=y will run the
rcuc kthreads at the RCU-boosting priority.
Note that rcutree.use_softirq=0 must be specified to move RCU core
processing to the rcuc kthreads: rcutree.use_softirq=1 is the default.
Reported-by: Thomas Gleixner <tglx@linutronix.de>
Tested-by: Mike Galbraith <efault@gmx.de>
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
[ paulmck: Adjust for invoke_rcu_callbacks() only ever being invoked
from RCU core processing, in contrast to softirq->rcuc transition
in old mainline RCU priority boosting. ]
[ paulmck: Avoid wakeups when scheduler might have invoked rcu_read_unlock()
while holding rq or pi locks, also possibly fixing a pre-existing latent
bug involving raise_softirq()-induced wakeups. ]
Signed-off-by: Paul E. McKenney <paulmck@linux.ibm.com>
This commit further consolidates stall-warning functionality by moving
forward-progress checkers into kernel/rcu/tree_stall.h, updating a
comment or two while in the area. More specifically, this commit moves
show_rcu_gp_kthreads(), rcu_check_gp_start_stall(), rcu_fwd_progress_check(),
sysrq_rcu, sysrq_show_rcu(), sysrq_rcudump_op, and rcu_sysrq_init() from
kernel/rcu/tree.c to kernel/rcu/tree_stall.h.
Signed-off-by: Paul E. McKenney <paulmck@linux.ibm.com>
The rcu_iw_handler() function's sole purpose in life is to indicate
whether a stalled CPU had interrupts disabled, so it belongs in
kernel/rcu/tree_stall.h. This commit therefore makes that move,
clarifying its header comment while in the area.
Signed-off-by: Paul E. McKenney <paulmck@linux.ibm.com>
This commit does only code movement, removal of now-unneeded forward
declarations, and addition of comments. It organizes the functions
that implement RCU CPU stall warnings for normal grace periods into
three categories:
1. Control of RCU CPU stall warnings, including computing timeouts.
2. Interaction of stall warnings with grace periods.
3. Actual printing of the RCU CPU stall-warning messages.
Signed-off-by: Paul E. McKenney <paulmck@linux.ibm.com>
This commit further consolidates the stall-warning code by moving
print_cpu_stall_info() and its helper functions along with
zero_cpu_stall_ticks() to kernel/rcu/tree_stall.h.
Signed-off-by: Paul E. McKenney <paulmck@linux.ibm.com>
The print_cpu_stall_info_begin() and print_cpu_stall_info_end() print a
single character each onto the console, and are a holdover from a time
when RCU CPU stall warning messages could be abbreviated using a long-gone
Kconfig option. This commit therefore adds these single characters to
already-printed strings in the calling functions, and then eliminates
both print_cpu_stall_info_begin() and print_cpu_stall_info_end().
Signed-off-by: Paul E. McKenney <paulmck@linux.ibm.com>
The rcu_print_detail_task_stall(), rcu_print_task_stall_begin(), and
rcu_print_task_stall_end() functions were defined to allow long-gone
Kconfig options to provide an abbreviated RCU CPU stall warning printout.
This commit saves a few lines of code by inlining them into their sole
callers.
While in the area, a useless call of rcu_print_detail_task_stall_rnp()
on the root rcu_node structure was eliminated. If there is only one
rcu_node structure, its tasks get printed twice, but if there are more,
the root rcu_node structure is guaranteed to have an empty list of blocked
tasks, hence the uselessness. (Long ago, root rcu_node structures with
non-empty ->blkd_tasks lists could happen, but no longer.)
Signed-off-by: Paul E. McKenney <paulmck@linux.ibm.com>
This commit completes the process of consolidating the code for RCU CPU
stall warnings for normal grace periods by moving the remaining such
code from kernel/rcu/tree.c to kernel/rcu/tree_stall.h.
Signed-off-by: Paul E. McKenney <paulmck@linux.ibm.com>
Replace the license boiler plate with a SPDX license identifier.
While in the area, update an email address.
Signed-off-by: Paul E. McKenney <paulmck@linux.ibm.com>
[ paulmck: Update .h file SPDX comment format per Joe Perches. ]
Reviewed-by: Thomas Gleixner <tglx@linutronix.de>
srcu_queue_delayed_work_on() disables preemption (and therefore CPU
hotplug in RCU's case) and then checks based on its own accounting if a
CPU is online. If the CPU is online it uses queue_delayed_work_on()
otherwise it fallbacks to queue_delayed_work().
The problem here is that queue_work() on -RT does not work with disabled
preemption.
queue_work_on() works also on an offlined CPU. queue_delayed_work_on()
has the problem that it is possible to program a timer on an offlined
CPU. This timer will fire once the CPU is online again. But until then,
the timer remains programmed and nothing will happen.
Add a local timer which will fire (as requested per delay) on the local
CPU and then enqueue the work on the specific CPU.
RCUtorture testing with SRCU-P for 24h showed no problems.
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Paul E. McKenney <paulmck@linux.ibm.com>
The name rcu_check_callbacks() arguably made sense back in the early
2000s when RCU was quite a bit simpler than it is today, but it has
become quite misleading, especially with the advent of dyntick-idle
and NO_HZ_FULL. The rcu_check_callbacks() function is RCU's hook into
the scheduling-clock interrupt, and is now but one of many ways that
callbacks get promoted to invocable state.
This commit therefore changes the name to rcu_sched_clock_irq(),
which is the same number of characters and clearly indicates this
function's relation to the rest of the Linux kernel. In addition, for
the sake of consistency, rcu_flavor_check_callbacks() is also renamed
to rcu_flavor_sched_clock_irq().
While in the area, the header comments for both functions are reworked.
Signed-off-by: Paul E. McKenney <paulmck@linux.ibm.com>
If a grace period fails to start (for example, because you commented
out the last two lines of rcu_accelerate_cbs_unlocked()), rcu_core()
will invoke rcu_check_gp_start_stall(), which will notice and complain.
However, this complaint is lacking crucial debugging information such
as when the last wakeup executed and what the value of ->gp_seq was at
that time. This commit therefore removes the current pr_alert() from
rcu_check_gp_start_stall(), instead invoking show_rcu_gp_kthreads(),
which has been updated to print the needed information, which is collected
by rcu_gp_kthread_wake().
Signed-off-by: Paul E. McKenney <paulmck@linux.ibm.com>
The rcu_cpu_kthread_cpu used to provide debugfs information, but is no
longer used. This commit therefore removes it.
Signed-off-by: Paul E. McKenney <paulmck@linux.ibm.com>
Given that RCU has a perfectly good per-CPU rcu_data structure, most
per-CPU quantities should be stored there.
This commit therefore moves the rcu_cpu_has_work per-CPU variable to
the rcu_data structure. This also makes this variable unconditionally
present, which should be acceptable given the memory reduction due to the
RCU flavor consolidation and also due to simplifications this will enable.
Signed-off-by: Paul E. McKenney <paulmck@linux.ibm.com>
The rcu_cpu_kthread_loops variable used to provide debugfs information,
but is no longer used. This commit therefore removes it.
Signed-off-by: Paul E. McKenney <paulmck@linux.ibm.com>
Given that RCU has a perfectly good per-CPU rcu_data structure, most
per-CPU quantities should be stored there.
This commit therefore moves the rcu_cpu_kthread_status per-CPU variable
to the rcu_data structure. This also makes this variable unconditionally
present, which should be acceptable given the memory reduction due to the
RCU flavor consolidation and also due to simplifications this will enable.
Signed-off-by: Paul E. McKenney <paulmck@linux.ibm.com>
Given that RCU has a perfectly good per-CPU rcu_data structure, most
per-CPU quantities should be stored there.
This commit therefore moves the rcu_cpu_kthread_task per-CPU variable to
the rcu_data structure. This also makes this variable unconditionally
present, which should be acceptable given the memory reduction due to the
RCU flavor consolidation and also due to simplifications this will enable.
Signed-off-by: Paul E. McKenney <paulmck@linux.ibm.com>
Back when there were multiple flavors of RCU, it was necessary to
separately count lazy and non-lazy callbacks for each CPU. These counts
were used in CONFIG_RCU_FAST_NO_HZ kernels to determine how long a newly
idle CPU should be allowed to sleep before handling its RCU callbacks.
But now that there is only one flavor, the callback counts for a given
CPU's sole rcu_data structure are the counts for that CPU.
This commit therefore removes the rcu_data structure's ->nonlazy_posted
and ->nonlazy_posted_snap fields, the rcu_idle_count_callbacks_posted()
and rcu_cpu_has_callbacks() functions, repurposes the rcu_data structure's
->all_lazy field to record the laziness state at the beginning of the
latest idle sojourn, and modifies CONFIG_RCU_FAST_NO_HZ RCU CPU stall
warnings accordingly.
Signed-off-by: Paul E. McKenney <paulmck@linux.ibm.com>
Back when there could be multiple RCU flavors running in the same kernel
at the same time, it was necessary to specify the expedited grace-period
IPI handler at runtime. Now that there is only one RCU flavor, the
IPI handler can be determined at build time. There is therefore no
longer any reason for the RCU-preempt and RCU-sched IPI handlers to
have different names, nor is there any reason to pass these handlers in
function arguments and in the data structures enclosing workqueues.
This commit therefore makes all these changes, pushing the specification
of the expedited grace-period IPI handler down to the point of use.
Signed-off-by: Paul E. McKenney <paulmck@linux.ibm.com>
Given RCU flavor consolidation, the name rcu_spawn_all_nocb_kthreads()
is quite misleading. It no longer ever creates more than one kthread,
and it does so only for the specified CPU. This commit therefore changes
this name to the more descriptive rcu_spawn_cpu_nocb_kthread(), and also
fixes up a similar issue in its header comment while in the area.
Signed-off-by: Paul E. McKenney <paulmck@linux.ibm.com>
If rcutorture's forward-progress tests fail while a grace period is not
in progress, it is useful to print the time since the last grace period
ended as a way to detect failure to launch a new grace period. This
commit therefore makes this change.
Signed-off-by: Paul E. McKenney <paulmck@linux.ibm.com>
The RCU CPU stall warnings print an estimate of the total number of
RCU callbacks queued in the system, but this estimate leaves out
the callbacks queued for nocbs CPUs. This commit therefore introduces
rcu_get_n_cbs_cpu(), which gives an accurate callback estimate for
both nocbs and normal CPUs, and uses this new function as needed.
This commit also introduces a rcu_get_n_cbs_nocb_cpu() helper function
that returns the number of callbacks for nocbs CPUs or zero otherwise,
and also uses this function in place of direct access to ->nocb_q_count
while in the area (fewer characters, you see).
Signed-off-by: Paul E. McKenney <paulmck@linux.ibm.com>
Recent changes have removed the old ->gp_seq_needed field from the
rcu_state structure, which in turn obsoleted a couple of comments in
the rcu_node and rcu_data structures. This commit therefore updates
these comments accordingly.
Signed-off-by: Joel Fernandes (Google) <joel@joelfernandes.org>
Cc: <kernel-team@android.com>
Signed-off-by: Paul E. McKenney <paulmck@linux.ibm.com>
The rcu_bh_state and rcu_sched_state variables were removed during the
RCU flavor consolidations, but external declarations remain in tree.h.
This commit therefore removes these obsolete declarations.
Signed-off-by: Joel Fernandes (Google) <joel@joelfernandes.org>
Cc: <kernel-team@android.com>
Signed-off-by: Paul E. McKenney <paulmck@linux.ibm.com>
1e64b15a4b ("rcu: Fix grace-period hangs due to race with CPU offline")
added spinlock_t ofl_lock to the rcu_state structure, then takes it with
preemption disabled during CPU offline, which gives the -rt patchset's
sleeping spinlock heartburn.
This commit therefore converts ->ofl_lock to raw_spinlock_t.
Signed-off-by: Mike Galbraith <efault@gmx.de>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
The rcu_data structure's ->dynticks_fqs is incremented but never
accesses. Its ->cond_resched_completed field isn't used at all.
This commit therefore removes both fields.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
This commit move ->dynticks from the rcu_dynticks structure to the
rcu_data structure, replacing the field of the same name. It also updates
the code to access ->dynticks from the rcu_data structure and to use the
rcu_data structure rather than following to now-gone ->dynticks field
to the now-gone rcu_dynticks structure. While in the area, this commit
also fixes up comments.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
This commit removes ->dynticks_nesting and ->dynticks_nmi_nesting from
the rcu_dynticks structure and updates the code to access them from the
rcu_data structure.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
This commit removes ->rcu_need_heavy_qs and ->rcu_urgent_qs from the
rcu_dynticks structure and updates the code to access them from the
rcu_data structure.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
This commit removes ->all_lazy, ->nonlazy_posted and ->nonlazy_posted_snap
from the rcu_dynticks structure and updates the code to access them from
the rcu_data structure.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
This commit removes ->last_accelerate and ->last_advance_all from the
rcu_dynticks structure and updates the code to access them from the
rcu_data structure.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
This commit removes ->tick_nohz_enabled_snap from the rcu_dynticks
structure and updates the code to access it from the rcu_data
structure.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Paul E. McKenney
Paul E. McKenney
Sebastian Andrzej Siewior
Joel Fernandes (Google)
Mike Galbraith
Paul E. McKenney