When __call_rcu_core() notices excessive numbers of callbacks pending
on the current CPU, we know that at least one of them is not yet
classified, namely the one that was just now queued. Therefore, it
is not necessary to invoke rcu_start_gp() and thus not necessary to
acquire the root rcu_node structure's ->lock. This commit therefore
replaces the rcu_start_gp() with rcu_accelerate_cbs(), thus replacing
an acquisition of the root rcu_node structure's ->lock with that of
this CPU's leaf rcu_node structure.
This decreases contention on the root rcu_node structure's ->lock.
Reported-by: Nicholas Piggin <npiggin@gmail.com>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Tested-by: Nicholas Piggin <npiggin@gmail.com>
The rcu_migrate_callbacks() function invokes rcu_advance_cbs()
twice, ignoring the return value. This is OK at pressent because of
failsafe code that does the wakeup when needed. However, this failsafe
code acquires the root rcu_node structure's lock frequently, while
rcu_migrate_callbacks() does so only once per CPU-offline operation.
This commit therefore makes rcu_migrate_callbacks()
wake up the RCU GP kthread when either call to rcu_advance_cbs()
returns true, thus removing need for the failsafe code.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Tested-by: Nicholas Piggin <npiggin@gmail.com>
There is no longer any need for ->need_future_gp[] to count the number of
requests for future grace periods, so this commit converts the additions
to assignments to "true" and reduces the size of each element to one byte.
While we are in the area, fix an obsolete comment.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Tested-by: Nicholas Piggin <npiggin@gmail.com>
Currently, the rcu_future_needs_gp() function checks only the current
element of the ->need_future_gps[] array, which might miss elements that
were offset from the expected element, for example, due to races with
the start or the end of a grace period. This commit therefore makes
rcu_future_needs_gp() use the need_any_future_gp() macro to check all
of the elements of this array.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Tested-by: Nicholas Piggin <npiggin@gmail.com>
The rcu_cbs_completed() function provides the value of ->completed
at which new callbacks can safely be invoked. This is recorded in
two-element ->need_future_gp[] arrays in the rcu_node structure, and
the elements of these arrays corresponding to the just-completed grace
period are zeroed at the end of that grace period. However, the
rcu_cbs_completed() function can return the current ->completed value
plus either one or two, so it is possible for the corresponding
->need_future_gp[] entry to be cleared just after it was set, thus
losing a request for a future grace period.
This commit avoids this race by expanding ->need_future_gp[] to four
elements.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Tested-by: Nicholas Piggin <npiggin@gmail.com>
Currently, rcu_gp_cleanup() scans the rcu_node tree in order to reset
state to reflect the end of the grace period. It also checks to see
whether a new grace period is needed, but in a number of cases, rather
than directly cause the new grace period to be immediately started, it
instead leaves the grace-period-needed state where various fail-safes
can find it. This works fine, but results in higher contention on the
root rcu_node structure's ->lock, which is undesirable, and contention
on that lock has recently become noticeable.
This commit therefore makes rcu_gp_cleanup() immediately start a new
grace period if there is any need for one.
It is quite possible that it will later be necessary to throttle the
grace-period rate, but that can be dealt with when and if.
Reported-by: Nicholas Piggin <npiggin@gmail.com>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Tested-by: Nicholas Piggin <npiggin@gmail.com>
The rcu_gp_kthread() function immediately sleeps waiting to be notified
of the need for a new grace period, which currently works because there
are a number of code sequences that will provide the needed wakeup later.
However, some of these code sequences need to acquire the root rcu_node
structure's ->lock, and contention on that lock has started manifesting.
This commit therefore makes rcu_gp_kthread() check for early-boot activity
when it starts up, omitting the initial sleep in that case.
Reported-by: Nicholas Piggin <npiggin@gmail.com>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Tested-by: Nicholas Piggin <npiggin@gmail.com>
Accessors for the ->need_future_gp[] array are currently open-coded,
which makes them difficult to change. To improve maintainability, this
commit adds need_future_gp_mask() to compute the indexing mask from the
array size, need_future_gp_element() to access the element corresponding
to the specified grace-period number, and need_any_future_gp() to
determine if any future grace period is needed. This commit also applies
need_future_gp_element() to existing open-coded single-element accesses.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Tested-by: Nicholas Piggin <npiggin@gmail.com>
The rcu_start_future_gp() function uses a sloppy check for a grace
period being in progress, which works today because there are a number
of code sequences that resolve the resulting races. However, some of
these race-resolution code sequences must acquire the root rcu_node
structure's ->lock, and contention on that lock has started manifesting.
This commit therefore makes rcu_start_future_gp() check more precise,
eliminating the sloppy lockless check of the rcu_state structure's ->gpnum
and ->completed fields. The effect is that rcu_start_future_gp() will
sometimes unnecessarily attempt to start a new grace period, but this
overhead will be reduced later using funnel locking.
Reported-by: Nicholas Piggin <npiggin@gmail.com>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Tested-by: Nicholas Piggin <npiggin@gmail.com>
When rcu_cbs_completed() is invoked on a non-root rcu_node structure,
it unconditionally assumes that two grace periods must complete before
the callbacks at hand can be invoked. This is overly conservative because
if that non-root rcu_node structure believes that no grace period is in
progress, and if the corresponding rcu_state structure's ->gpnum field
has not yet been incremented, then these callbacks may safely be invoked
after only one grace period has completed.
This change is required to permit grace-period start requests to use
funnel locking, which is in turn permitted to reduce root rcu_node ->lock
contention, which has been observed by Nick Piggin. Furthermore, such
contention will likely be increased by the merging of RCU-bh, RCU-preempt,
and RCU-sched, so it makes sense to take steps to decrease it.
This commit therefore improves the accuracy of rcu_cbs_completed() when
invoked on a non-root rcu_node structure as described above.
Reported-by: Nicholas Piggin <npiggin@gmail.com>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Tested-by: Nicholas Piggin <npiggin@gmail.com>
This commit adds rcu_first_leaf_node() that returns a pointer to
the first leaf rcu_node structure in the specified RCU flavor and an
rcu_is_leaf_node() that returns true iff the specified rcu_node structure
is a leaf. This commit also uses these macros where appropriate.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Tested-by: Nicholas Piggin <npiggin@gmail.com>
The current cleanup_srcu_struct() flushes work, which prevents it
from being invoked from some workqueue contexts, as well as from
atomic (non-blocking) contexts. This patch therefore introduced a
cleanup_srcu_struct_quiesced(), which can be invoked only after all
activity on the specified srcu_struct has completed. This restriction
allows cleanup_srcu_struct_quiesced() to be invoked from workqueue
contexts as well as from atomic contexts.
Suggested-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Tested-by: Nitzan Carmi <nitzanc@mellanox.com>
Tested-by: Nicholas Piggin <npiggin@gmail.com>
Because rcu_eqs_special_set() is declared only in internal header
kernel/rcu/tree.h and stubbed in include/linux/rcutiny.h, it is
inaccessible outside of the RCU implementation. This patch therefore
moves the rcu_eqs_special_set() declaration to include/linux/rcutree.h,
which allows it to be used in non-rcu kernel code.
Signed-off-by: Yury Norov <ynorov@caviumnetworks.com>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Tested-by: Nicholas Piggin <npiggin@gmail.com>
The header comment for rcu_bind_gp_kthread() refers to sysidle, which
is no longer with us. However, it is still important to bind RCU's
grace-period kthreads to the housekeeping CPU(s), so rather than remove
rcu_bind_gp_kthread(), this commit updates the comment.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Tested-by: Nicholas Piggin <npiggin@gmail.com>
The __rcu_read_lock() and __rcu_read_unlock() functions were moved
to kernel/rcu/update.c in order to implement tiny preemptible RCU.
However, tiny preemptible RCU was removed from the kernel a long time
ago, so this commit belatedly moves them back into the only remaining
preemptible-RCU code.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Tested-by: Nicholas Piggin <npiggin@gmail.com>
Commit e31d28b6ab ("trace: Eliminate cond_resched_rcu_qs() in favor
of cond_resched()") substituted cond_resched() for the earlier call
to cond_resched_rcu_qs(). However, the new-age cond_resched() does
not do anything to help RCU-tasks grace periods because (1) RCU-tasks
is only enabled when CONFIG_PREEMPT=y and (2) cond_resched() is a
complete no-op when preemption is enabled. This situation results
in hangs when running the trace benchmarks.
A number of potential fixes were discussed on LKML
(https://lkml.kernel.org/r/20180224151240.0d63a059@vmware.local.home),
including making cond_resched() not be a no-op; making cond_resched()
not be a no-op, but only when running tracing benchmarks; reverting
the aforementioned commit (which works because cond_resched_rcu_qs()
does provide an RCU-tasks quiescent state; and adding a call to the
scheduler/RCU rcu_note_voluntary_context_switch() function. All were
deemed unsatisfactory, either due to added cond_resched() overhead or
due to magic functions inviting cargo culting.
This commit renames cond_resched_rcu_qs() to cond_resched_tasks_rcu_qs(),
which provides a clear hint as to what this function is doing and
why and where it should be used, and then replaces the call to
cond_resched() with cond_resched_tasks_rcu_qs() in the trace benchmark's
benchmark_event_kthread() function.
Reported-by: Steven Rostedt <rostedt@goodmis.org>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Tested-by: Nicholas Piggin <npiggin@gmail.com>
Commit ae91aa0adb ("rcu: Remove debugfs tracing") removed the
RCU debugfs tracing code, but did not remove the no-longer used
->exp_workdone{0,1,2,3} fields in the srcu_data structure. This commit
therefore removes these fields along with the code that uselessly
updates them.
Signed-off-by: Byungchul Park <byungchul.park@lge.com>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Tested-by: Nicholas Piggin <npiggin@gmail.com>
If an excessive number of callbacks have been queued, but the NOCB
leader kthread's wakeup must be deferred, then we should wake up the
leader unconditionally once it is safe to do so.
This was handled correctly in commit fbce7497ee ("rcu: Parallelize and
economize NOCB kthread wakeups"), but then commit 8be6e1b15c ("rcu:
Use timer as backstop for NOCB deferred wakeups") passed RCU_NOCB_WAKE
instead of the correct RCU_NOCB_WAKE_FORCE to wake_nocb_leader_defer().
As an interesting aside, RCU_NOCB_WAKE_FORCE is never passed to anything,
which should have been taken as a hint. ;-)
This commit therefore passes RCU_NOCB_WAKE_FORCE instead of RCU_NOCB_WAKE
to wake_nocb_leader_defer() when a callback is queued onto a NOCB CPU
that already has an excessive number of callbacks pending.
Signed-off-by: Byungchul Park <byungchul.park@lge.com>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Tested-by: Nicholas Piggin <npiggin@gmail.com>
Commit 44c65ff2e3 ("rcu: Eliminate NOCBs CPU-state Kconfig options")
made allocation of rcu_nocb_mask depend only on the rcu_nocbs=,
nohz_full=, or isolcpus= kernel boot parameters. However, it failed
to change the initial value of rcu_init_nohz()'s local variable
need_rcu_nocb_mask to false, which can result in useless allocation
of an all-zero rcu_nocb_mask. This commit therefore fixes this bug by
changing the initial value of need_rcu_nocb_mask to false.
While we are in the area, also correct the error message that is printed
when someone specifies that can-never-exist CPUs should be NOCBs CPUs.
Reported-by: Byungchul Park <byungchul.park@lge.com>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Acked-by: Byungchul Park <byungchul.park@lge.com>
Tested-by: Nicholas Piggin <npiggin@gmail.com>
The rcu_preempt_do_callbacks() function was introduced in commit
09223371dea(rcu: Use softirq to address performance regression), where it
was necessary to handle kernel builds both containing and not containing
RCU-preempt. Since then, various changes (most notably f8b7fc6b51
("rcu: use softirq instead of kthreads except when RCU_BOOST=y")) have
resulted in this function being invoked only from rcu_kthread_do_work(),
which is present only in kernels containing RCU-preempt, which in turn
means that the rcu_preempt_do_callbacks() function is no longer needed.
This commit therefore inlines rcu_preempt_do_callbacks() into its
sole remaining caller and also removes the rcu_state_p and rcu_data_p
indirection for added clarity.
Signed-off-by: Byungchul Park <byungchul.park@lge.com>
Reviewed-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
[ paulmck: Remove the rcu_state_p and rcu_data_p indirection. ]
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Tested-by: Nicholas Piggin <npiggin@gmail.com>
Currently some callsites of sync_rcu_preempt_exp_done() are not called
with the corresponding rcu_node's ->lock held, which could introduces
bugs as per Paul:
o CPU 0 in sync_rcu_preempt_exp_done() reads ->exp_tasks and
sees that it is NULL.
o CPU 1 blocks within an RCU read-side critical section, so
it enqueues the task and points ->exp_tasks at it and
clears CPU 1's bit in ->expmask.
o All other CPUs clear their bits in ->expmask.
o CPU 0 reads ->expmask, sees that it is zero, so incorrectly
concludes that all quiescent states have completed, despite
the fact that ->exp_tasks is non-NULL.
To fix this, sync_rcu_preempt_exp_unlocked() is introduced to replace
lockless callsites of sync_rcu_preempt_exp_done().
Further, a lockdep annotation is added into sync_rcu_preempt_exp_done()
to prevent mis-use in the future.
Signed-off-by: Boqun Feng <boqun.feng@gmail.com>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Tested-by: Nicholas Piggin <npiggin@gmail.com>
Since commit d9a3da0699 ("rcu: Add expedited grace-period support
for preemptible RCU"), there are comments for some funtions in
rcu_report_exp_rnp()'s call-chain saying that exp_mutex or its
predecessors needs to be held.
However, exp_mutex and its predecessors were used only to synchronize
between GPs, and it is clear that all variables visited by those functions
are under the protection of rcu_node's ->lock. Moreover, those functions
are currently called without held exp_mutex, and seems that doesn't
introduce any trouble.
So this patch fixes this problem by updating the comments to match the
current code.
Signed-off-by: Boqun Feng <boqun.feng@gmail.com>
Fixes: d9a3da0699 ("rcu: Add expedited grace-period support for preemptible RCU")
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Tested-by: Nicholas Piggin <npiggin@gmail.com>
The latency of RCU expedited grace periods grows with increasing numbers
of CPUs, eventually failing to be all that expedited. Much of the growth
in latency is in the initialization phase, so this commit uses workqueues
to carry out this initialization concurrently on a rcu_node-by-rcu_node
basis.
This change makes use of a new rcu_par_gp_wq because flushing a work
item from another work item running from the same workqueue can result
in deadlock.
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Tested-by: Nicholas Piggin <npiggin@gmail.com>
This patch only refactors the existing sockmap code. This will allow
much of the psock initialization code path and bpf helper codes to
work for both sockmap bpf map types that are backed by an array, the
currently supported type, and the new hash backed bpf map type
sockhash.
Most the fallout comes from three changes,
- Pushing bpf programs into an independent structure so we
can use it from the htab struct in the next patch.
- Generalizing helpers to use void *key instead of the hardcoded
u32.
- Instead of passing map/key through the metadata we now do
the lookup inline. This avoids storing the key in the metadata
which will be useful when keys can be longer than 4 bytes. We
rename the sk pointers to sk_redir at this point as well to
avoid any confusion between the current sk pointer and the
redirect pointer sk_redir.
Signed-off-by: John Fastabend <john.fastabend@gmail.com>
Acked-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
The schedutil driver sets sg_policy->next_freq to UINT_MAX on certain
occasions to discard the cached value of next freq:
- In sugov_start(), when the schedutil governor is started for a group
of CPUs.
- And whenever we need to force a freq update before rate-limit
duration, which happens when:
- there is an update in cpufreq policy limits.
- Or when the utilization of DL scheduling class increases.
In return, get_next_freq() doesn't return a cached next_freq value but
recalculates the next frequency instead.
But having special meaning for a particular value of frequency makes the
code less readable and error prone. We recently fixed a bug where the
UINT_MAX value was considered as valid frequency in
sugov_update_single().
All we need is a flag which can be used to discard the value of
sg_policy->next_freq and we already have need_freq_update for that. Lets
reuse it instead of setting next_freq to UINT_MAX.
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Reviewed-by: Joel Fernandes (Google) <joel@joelfernandes.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
This reverts commit e2cabe48c2.
Lifting the restriction that the sugov kthread is bound to the
policy->related_cpus for a system with a slow switching cpufreq driver,
which is able to perform DVFS from any cpu (e.g. cpufreq-dt), is not
only not beneficial it also harms Enery-Aware Scheduling (EAS) on
systems with asymmetric cpu capacities (e.g. Arm big.LITTLE).
The sugov kthread which does the update for the little cpus could
potentially run on a big cpu. It could prevent that the big cluster goes
into deeper idle states although all the tasks are running on the little
cluster.
Example: hikey960 w/ 4.16.0-rc6-+
Arm big.LITTLE with per-cluster DVFS
root@h960:~# cat /proc/cpuinfo | grep "^CPU part"
CPU part : 0xd03 (Cortex-A53, little cpu)
CPU part : 0xd03
CPU part : 0xd03
CPU part : 0xd03
CPU part : 0xd09 (Cortex-A73, big cpu)
CPU part : 0xd09
CPU part : 0xd09
CPU part : 0xd09
root@h960:/sys/devices/system/cpu/cpufreq# ls
policy0 policy4 schedutil
root@h960:/sys/devices/system/cpu/cpufreq# cat policy*/related_cpus
0 1 2 3
4 5 6 7
(1) w/o the revert:
root@h960:~# ps -eo pid,class,rtprio,pri,psr,comm | awk 'NR == 1 ||
/sugov/'
PID CLS RTPRIO PRI PSR COMMAND
1489 #6 0 140 1 sugov:0
1490 #6 0 140 0 sugov:4
The sugov kthread sugov:4 responsible for policy4 runs on cpu0. (In this
case both sugov kthreads run on little cpus).
cross policy (cluster) remote callback example:
...
migration/1-14 [001] enqueue_task_fair: this_cpu=1 cpu_of(rq)=5
migration/1-14 [001] sugov_update_shared: this_cpu=1 sg_cpu->cpu=5
sg_cpu->sg_policy->policy->related_cpus=4-7
sugov:4-1490 [000] sugov_work: this_cpu=0
sg_cpu->sg_policy->policy->related_cpus=4-7
...
The remote callback (this_cpu=1, target_cpu=5) is executed on cpu=0.
(2) w/ the revert:
root@h960:~# ps -eo pid,class,rtprio,pri,psr,comm | awk 'NR == 1 ||
/sugov/'
PID CLS RTPRIO PRI PSR COMMAND
1491 #6 0 140 2 sugov:0
1492 #6 0 140 4 sugov:4
The sugov kthread sugov:4 responsible for policy4 runs on cpu4.
cross policy (cluster) remote callback example:
...
migration/1-14 [001] enqueue_task_fair: this_cpu=1 cpu_of(rq)=7
migration/1-14 [001] sugov_update_shared: this_cpu=1 sg_cpu->cpu=7
sg_cpu->sg_policy->policy->related_cpus=4-7
sugov:4-1492 [004] sugov_work: this_cpu=4
sg_cpu->sg_policy->policy->related_cpus=4-7
...
The remote callback (this_cpu=1, target_cpu=7) is executed on cpu=4.
Now the sugov kthread executes again on the policy (cluster) for which
the Operating Performance Point (OPP) should be changed.
It avoids the problem that an otherwise idle policy (cluster) is running
schedutil (the sugov kthread) for another one.
Signed-off-by: Dietmar Eggemann <dietmar.eggemann@arm.com>
Acked-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Recognizing that the audit context is an internal audit value, use an
access function to set the audit context pointer for the task
rather than reaching directly into the task struct to set it.
Signed-off-by: Richard Guy Briggs <rgb@redhat.com>
[PM: merge fuzz in audit.h]
Signed-off-by: Paul Moore <paul@paul-moore.com>
Currently, we cannot parse build_id in nmi context because of
up_read(¤t->mm->mmap_sem), this makes stackmap with build_id
less useful. This patch enables parsing build_id in nmi by putting
the up_read() call in irq_work. To avoid memory allocation in nmi
context, we use per cpu variable for the irq_work. As a result, only
one irq_work per cpu is allowed. If the irq_work is in-use, we
fallback to only report ips.
Cc: Alexei Starovoitov <ast@kernel.org>
Cc: Daniel Borkmann <daniel@iogearbox.net>
Cc: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Song Liu <songliubraving@fb.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Recognizing that the audit context is an internal audit value, use an
access function to retrieve the audit context pointer for the task
rather than reaching directly into the task struct to get it.
Signed-off-by: Richard Guy Briggs <rgb@redhat.com>
[PM: merge fuzz in auditsc.c and selinuxfs.c, checkpatch.pl fixes]
Signed-off-by: Paul Moore <paul@paul-moore.com>
Use a macro, "AUDIT_SID_UNSET", to replace each instance of
initialization and comparison to an audit session ID.
Signed-off-by: Richard Guy Briggs <rgb@redhat.com>
Signed-off-by: Paul Moore <paul@paul-moore.com>
Same numerical value (for now at least), but a much better documentation
of intent.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Threads share an address space and each can change the protections of the
same address space to trap NUMA faults. This is redundant and potentially
counter-productive as any thread doing the update will suffice. Potentially
only one thread is required but that thread may be idle or it may not have
any locality concerns and pick an unsuitable scan rate.
This patch uses independent scan period but they are staggered based on
the number of address space users when the thread is created. The intent
is that threads will avoid scanning at the same time and have a chance
to adapt their scan rate later if necessary. This reduces the total scan
activity early in the lifetime of the threads.
The different in headline performance across a range of machines and
workloads is marginal but the system CPU usage is reduced as well as overall
scan activity. The following is the time reported by NAS Parallel Benchmark
using unbound openmp threads and a D size class:
4.17.0-rc1 4.17.0-rc1
vanilla stagger-v1r1
Time bt.D 442.77 ( 0.00%) 419.70 ( 5.21%)
Time cg.D 171.90 ( 0.00%) 180.85 ( -5.21%)
Time ep.D 33.10 ( 0.00%) 32.90 ( 0.60%)
Time is.D 9.59 ( 0.00%) 9.42 ( 1.77%)
Time lu.D 306.75 ( 0.00%) 304.65 ( 0.68%)
Time mg.D 54.56 ( 0.00%) 52.38 ( 4.00%)
Time sp.D 1020.03 ( 0.00%) 903.77 ( 11.40%)
Time ua.D 400.58 ( 0.00%) 386.49 ( 3.52%)
Note it's not a universal win but we have no prior knowledge of which
thread matters but the number of threads created often exceeds the size
of the node when the threads are not bound. However, there is a reducation
of overall system CPU usage:
4.17.0-rc1 4.17.0-rc1
vanilla stagger-v1r1
sys-time-bt.D 48.78 ( 0.00%) 48.22 ( 1.15%)
sys-time-cg.D 25.31 ( 0.00%) 26.63 ( -5.22%)
sys-time-ep.D 1.65 ( 0.00%) 0.62 ( 62.42%)
sys-time-is.D 40.05 ( 0.00%) 24.45 ( 38.95%)
sys-time-lu.D 37.55 ( 0.00%) 29.02 ( 22.72%)
sys-time-mg.D 47.52 ( 0.00%) 34.92 ( 26.52%)
sys-time-sp.D 119.01 ( 0.00%) 109.05 ( 8.37%)
sys-time-ua.D 51.52 ( 0.00%) 45.13 ( 12.40%)
NUMA scan activity is also reduced:
NUMA alloc local 1042828 1342670
NUMA base PTE updates 140481138 93577468
NUMA huge PMD updates 272171 180766
NUMA page range updates 279832690 186129660
NUMA hint faults 1395972 1193897
NUMA hint local faults 877925 855053
NUMA hint local percent 62 71
NUMA pages migrated 12057909 9158023
Similar observations are made for other thread-intensive workloads. System
CPU usage is lower even though the headline gains in performance tend to be
small. For example, specjbb 2005 shows almost no difference in performance
but scan activity is reduced by a third on a 4-socket box. I didn't find
a workload (thread intensive or otherwise) that suffered badly.
Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Matt Fleming <matt@codeblueprint.co.uk>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: linux-kernel@vger.kernel.org
Link: http://lkml.kernel.org/r/20180504154109.mvrha2qo5wdl65vr@techsingularity.net
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Pull x86/pti updates from Thomas Gleixner:
"A mixed bag of fixes and updates for the ghosts which are hunting us.
The scheduler fixes have been pulled into that branch to avoid
conflicts.
- A set of fixes to address a khread_parkme() race which caused lost
wakeups and loss of state.
- A deadlock fix for stop_machine() solved by moving the wakeups
outside of the stopper_lock held region.
- A set of Spectre V1 array access restrictions. The possible
problematic spots were discuvered by Dan Carpenters new checks in
smatch.
- Removal of an unused file which was forgotten when the rest of that
functionality was removed"
* 'x86-pti-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
x86/vdso: Remove unused file
perf/x86/cstate: Fix possible Spectre-v1 indexing for pkg_msr
perf/x86/msr: Fix possible Spectre-v1 indexing in the MSR driver
perf/x86: Fix possible Spectre-v1 indexing for x86_pmu::event_map()
perf/x86: Fix possible Spectre-v1 indexing for hw_perf_event cache_*
perf/core: Fix possible Spectre-v1 indexing for ->aux_pages[]
sched/autogroup: Fix possible Spectre-v1 indexing for sched_prio_to_weight[]
sched/core: Fix possible Spectre-v1 indexing for sched_prio_to_weight[]
sched/core: Introduce set_special_state()
kthread, sched/wait: Fix kthread_parkme() completion issue
kthread, sched/wait: Fix kthread_parkme() wait-loop
sched/fair: Fix the update of blocked load when newly idle
stop_machine, sched: Fix migrate_swap() vs. active_balance() deadlock
Pull scheduler fix from Thomas Gleixner:
"Revert the new NUMA aware placement approach which turned out to
create more problems than it solved"
* 'sched-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
Revert "sched/numa: Delay retrying placement for automatic NUMA balance after wake_affine()"
So far, MSIs have been used to signal edge-triggered interrupts, as
a write is a good model for an edge (you can't "unwrite" something).
On the other hand, routing zillions of wires in an SoC because you
need level interrupts is a bit extreme.
People have come up with a variety of schemes to support this, which
involves sending two messages: one to signal the interrupt, and one
to clear it. Since the kernel cannot represent this, we've ended up
with side-band mechanisms that are pretty awful.
Instead, let's acknoledge the requirement, and ensure that, under the
right circumstances, the irq_compose_msg and irq_write_msg can take
as a parameter an array of two messages instead of a pointer to a
single one. We also add some checking that the compose method only
clobbers the second message if the MSI domain has been created with
the MSI_FLAG_LEVEL_CAPABLE flags.
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: Rob Herring <robh@kernel.org>
Cc: Jason Cooper <jason@lakedaemon.net>
Cc: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Cc: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>
Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Cc: Miquel Raynal <miquel.raynal@bootlin.com>
Link: https://lkml.kernel.org/r/20180508121438.11301-2-marc.zyngier@arm.com
Checking the equality of cpumask for both new and old tick device doesn't
ensure that it's CPU local device. This will cause issue if a low rating
clockevent tick device is registered first followed by the registration
of higher rating clockevent tick device.
In such case, clockevents_released list will never get emptied as both
the devices get selected as preferred one and we will loop forever in
clockevents_notify_released.
Signed-off-by: Sudeep Holla <sudeep.holla@arm.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Link: https://lkml.kernel.org/r/1525881728-4858-1-git-send-email-sudeep.holla@arm.com
This reverts commit 7347fc87df.
Srikar Dronamra pointed out that while the commit in question did show
a performance improvement on ppc64, it did so at the cost of disabling
active CPU migration by automatic NUMA balancing which was not the intent.
The issue was that a serious flaw in the logic failed to ever active balance
if SD_WAKE_AFFINE was disabled on scheduler domains. Even when it's enabled,
the logic is still bizarre and against the original intent.
Investigation showed that fixing the patch in either the way he suggested,
using the correct comparison for jiffies values or introducing a new
numa_migrate_deferred variable in task_struct all perform similarly to a
revert with a mix of gains and losses depending on the workload, machine
and socket count.
The original intent of the commit was to handle a problem whereby
wake_affine, idle balancing and automatic NUMA balancing disagree on the
appropriate placement for a task. This was particularly true for cases where
a single task was a massive waker of tasks but where wake_wide logic did
not apply. This was particularly noticeable when a futex (a barrier) woke
all worker threads and tried pulling the wakees to the waker nodes. In that
specific case, it could be handled by tuning MPI or openMP appropriately,
but the behavior is not illogical and was worth attempting to fix. However,
the approach was wrong. Given that we're at rc4 and a fix is not obvious,
it's better to play safe, revert this commit and retry later.
Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: efault@gmx.de
Cc: ggherdovich@suse.cz
Cc: hpa@zytor.com
Cc: matt@codeblueprint.co.uk
Cc: mpe@ellerman.id.au
Link: http://lkml.kernel.org/r/20180509163115.6fnnyeg4vdm2ct4v@techsingularity.net
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Merge misc fixes from Andrew Morton:
"13 fixes"
* emailed patches from Andrew Morton <akpm@linux-foundation.org>:
rbtree: include rcu.h
scripts/faddr2line: fix error when addr2line output contains discriminator
ocfs2: take inode cluster lock before moving reflinked inode from orphan dir
mm, oom: fix concurrent munlock and oom reaper unmap, v3
mm: migrate: fix double call of radix_tree_replace_slot()
proc/kcore: don't bounds check against address 0
mm: don't show nr_indirectly_reclaimable in /proc/vmstat
mm: sections are not offlined during memory hotremove
z3fold: fix reclaim lock-ups
init: fix false positives in W+X checking
lib/find_bit_benchmark.c: avoid soft lockup in test_find_first_bit()
KASAN: prohibit KASAN+STRUCTLEAK combination
MAINTAINERS: update Shuah's email address
The bpf syscall and selftests conflicts were trivial
overlapping changes.
The r8169 change involved moving the added mdelay from 'net' into a
different function.
A TLS close bug fix overlapped with the splitting of the TLS state
into separate TX and RX parts. I just expanded the tests in the bug
fix from "ctx->conf == X" into "ctx->tx_conf == X && ctx->rx_conf
== X".
Signed-off-by: David S. Miller <davem@davemloft.net>
load_module() creates W+X mappings via __vmalloc_node_range() (from
layout_and_allocate()->move_module()->module_alloc()) by using
PAGE_KERNEL_EXEC. These mappings are later cleaned up via
"call_rcu_sched(&freeinit->rcu, do_free_init)" from do_init_module().
This is a problem because call_rcu_sched() queues work, which can be run
after debug_checkwx() is run, resulting in a race condition. If hit,
the race results in a nasty splat about insecure W+X mappings, which
results in a poor user experience as these are not the mappings that
debug_checkwx() is intended to catch.
This issue is observed on multiple arm64 platforms, and has been
artificially triggered on an x86 platform.
Address the race by flushing the queued work before running the
arch-defined mark_rodata_ro() which then calls debug_checkwx().
Link: http://lkml.kernel.org/r/1525103946-29526-1-git-send-email-jhugo@codeaurora.org
Fixes: e1a58320a3 ("x86/mm: Warn on W^X mappings")
Signed-off-by: Jeffrey Hugo <jhugo@codeaurora.org>
Reported-by: Timur Tabi <timur@codeaurora.org>
Reported-by: Jan Glauber <jan.glauber@caviumnetworks.com>
Acked-by: Kees Cook <keescook@chromium.org>
Acked-by: Ingo Molnar <mingo@kernel.org>
Acked-by: Will Deacon <will.deacon@arm.com>
Acked-by: Laura Abbott <labbott@redhat.com>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Stephen Smalley <sds@tycho.nsa.gov>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
