From a35d16905efc6ad5523d864a5c6efcb1e657e386 Mon Sep 17 00:00:00 2001 From: Byungchul Park Date: Mon, 5 Aug 2019 18:22:27 -0400 Subject: rcu: Add basic support for kfree_rcu() batching Recently a discussion about stability and performance of a system involving a high rate of kfree_rcu() calls surfaced on the list [1] which led to another discussion how to prepare for this situation. This patch adds basic batching support for kfree_rcu(). It is "basic" because we do none of the slab management, dynamic allocation, code moving or any of the other things, some of which previous attempts did [2]. These fancier improvements can be follow-up patches and there are different ideas being discussed in those regards. This is an effort to start simple, and build up from there. In the future, an extension to use kfree_bulk and possibly per-slab batching could be done to further improve performance due to cache-locality and slab-specific bulk free optimizations. By using an array of pointers, the worker thread processing the work would need to read lesser data since it does not need to deal with large rcu_head(s) any longer. Torture tests follow in the next patch and show improvements of around 5x reduction in number of grace periods on a 16 CPU system. More details and test data are in that patch. There is an implication with rcu_barrier() with this patch. Since the kfree_rcu() calls can be batched, and may not be handed yet to the RCU machinery in fact, the monitor may not have even run yet to do the queue_rcu_work(), there seems no easy way of implementing rcu_barrier() to wait for those kfree_rcu()s that are already made. So this means a kfree_rcu() followed by an rcu_barrier() does not imply that memory will be freed once rcu_barrier() returns. Another implication is higher active memory usage (although not run-away..) until the kfree_rcu() flooding ends, in comparison to without batching. More details about this are in the second patch which adds an rcuperf test. Finally, in the near future we will get rid of kfree_rcu() special casing within RCU such as in rcu_do_batch and switch everything to just batching. Currently we don't do that since timer subsystem is not yet up and we cannot schedule the kfree_rcu() monitor as the timer subsystem's lock are not initialized. That would also mean getting rid of kfree_call_rcu_nobatch() entirely. [1] http://lore.kernel.org/lkml/20190723035725-mutt-send-email-mst@kernel.org [2] https://lkml.org/lkml/2017/12/19/824 Cc: kernel-team@android.com Cc: kernel-team@lge.com Co-developed-by: Byungchul Park Signed-off-by: Byungchul Park Signed-off-by: Joel Fernandes (Google) [ paulmck: Applied 0day and Paul Walmsley feedback on ->monitor_todo. ] [ paulmck: Make it work during early boot. ] [ paulmck: Add a crude early boot self-test. ] [ paulmck: Style adjustments and experimental docbook structure header. ] Link: https://lore.kernel.org/lkml/alpine.DEB.2.21.9999.1908161931110.32497@viisi.sifive.com/T/#me9956f66cb611b95d26ae92700e1d901f46e8c59 Signed-off-by: Paul E. McKenney --- include/linux/rcutiny.h | 6 ++++++ 1 file changed, 6 insertions(+) (limited to 'include/linux/rcutiny.h') diff --git a/include/linux/rcutiny.h b/include/linux/rcutiny.h index 37b6f0c2b79d..1bd166aab6f3 100644 --- a/include/linux/rcutiny.h +++ b/include/linux/rcutiny.h @@ -39,6 +39,11 @@ static inline void kfree_call_rcu(struct rcu_head *head, rcu_callback_t func) call_rcu(head, func); } +static inline void kfree_call_rcu_nobatch(struct rcu_head *head, rcu_callback_t func) +{ + call_rcu(head, func); +} + void rcu_qs(void); static inline void rcu_softirq_qs(void) @@ -85,6 +90,7 @@ static inline void rcu_scheduler_starting(void) { } static inline void rcu_end_inkernel_boot(void) { } static inline bool rcu_is_watching(void) { return true; } static inline void rcu_momentary_dyntick_idle(void) { } +static inline void kfree_rcu_scheduler_running(void) { } /* Avoid RCU read-side critical sections leaking across. */ static inline void rcu_all_qs(void) { barrier(); } -- cgit v1.2.3 From 189a6883dcf7fa70e17403ae4225c60ffc9e404b Mon Sep 17 00:00:00 2001 From: "Joel Fernandes (Google)" Date: Fri, 30 Aug 2019 12:36:33 -0400 Subject: rcu: Remove kfree_call_rcu_nobatch() Now that the kfree_rcu() special-casing has been removed from tree RCU, this commit removes kfree_call_rcu_nobatch() since it is no longer needed. Signed-off-by: Joel Fernandes (Google) Signed-off-by: Paul E. McKenney --- Documentation/admin-guide/kernel-parameters.txt | 4 ---- include/linux/rcutiny.h | 5 ----- include/linux/rcutree.h | 1 - kernel/rcu/rcuperf.c | 10 +--------- kernel/rcu/tree.c | 18 ++++-------------- 5 files changed, 5 insertions(+), 33 deletions(-) (limited to 'include/linux/rcutiny.h') diff --git a/Documentation/admin-guide/kernel-parameters.txt b/Documentation/admin-guide/kernel-parameters.txt index 3ce270b56f3a..ed83d6d90cc3 100644 --- a/Documentation/admin-guide/kernel-parameters.txt +++ b/Documentation/admin-guide/kernel-parameters.txt @@ -3991,10 +3991,6 @@ Number of loops doing rcuperf.kfree_alloc_num number of allocations and frees. - rcuperf.kfree_no_batch= [KNL] - Use the non-batching (less efficient) version of kfree_rcu(). - This is useful for comparing with the batched version. - rcuperf.nreaders= [KNL] Set number of RCU readers. The value -1 selects N, where N is the number of CPUs. A value diff --git a/include/linux/rcutiny.h b/include/linux/rcutiny.h index 1bd166aab6f3..b2b2dc990da9 100644 --- a/include/linux/rcutiny.h +++ b/include/linux/rcutiny.h @@ -39,11 +39,6 @@ static inline void kfree_call_rcu(struct rcu_head *head, rcu_callback_t func) call_rcu(head, func); } -static inline void kfree_call_rcu_nobatch(struct rcu_head *head, rcu_callback_t func) -{ - call_rcu(head, func); -} - void rcu_qs(void); static inline void rcu_softirq_qs(void) diff --git a/include/linux/rcutree.h b/include/linux/rcutree.h index 6a65d3a16dbd..2f787b9029d1 100644 --- a/include/linux/rcutree.h +++ b/include/linux/rcutree.h @@ -34,7 +34,6 @@ static inline void rcu_virt_note_context_switch(int cpu) void synchronize_rcu_expedited(void); void kfree_call_rcu(struct rcu_head *head, rcu_callback_t func); -void kfree_call_rcu_nobatch(struct rcu_head *head, rcu_callback_t func); void rcu_barrier(void); bool rcu_eqs_special_set(int cpu); diff --git a/kernel/rcu/rcuperf.c b/kernel/rcu/rcuperf.c index c1e25fd10f2a..da94b89cd531 100644 --- a/kernel/rcu/rcuperf.c +++ b/kernel/rcu/rcuperf.c @@ -593,7 +593,6 @@ rcu_perf_shutdown(void *arg) torture_param(int, kfree_nthreads, -1, "Number of threads running loops of kfree_rcu()."); torture_param(int, kfree_alloc_num, 8000, "Number of allocations and frees done in an iteration."); torture_param(int, kfree_loops, 10, "Number of loops doing kfree_alloc_num allocations and frees."); -torture_param(int, kfree_no_batch, 0, "Use the non-batching (slower) version of kfree_rcu()."); static struct task_struct **kfree_reader_tasks; static int kfree_nrealthreads; @@ -632,14 +631,7 @@ kfree_perf_thread(void *arg) if (!alloc_ptr) return -ENOMEM; - if (!kfree_no_batch) { - kfree_rcu(alloc_ptr, rh); - } else { - rcu_callback_t cb; - - cb = (rcu_callback_t)(unsigned long)offsetof(struct kfree_obj, rh); - kfree_call_rcu_nobatch(&(alloc_ptr->rh), cb); - } + kfree_rcu(alloc_ptr, rh); } cond_resched(); diff --git a/kernel/rcu/tree.c b/kernel/rcu/tree.c index a8dd612098bf..31d2d9255d95 100644 --- a/kernel/rcu/tree.c +++ b/kernel/rcu/tree.c @@ -2763,8 +2763,10 @@ static void kfree_rcu_work(struct work_struct *work) rcu_lock_acquire(&rcu_callback_map); trace_rcu_invoke_kfree_callback(rcu_state.name, head, offset); - /* Could be possible to optimize with kfree_bulk in future */ - kfree((void *)head - offset); + if (!WARN_ON_ONCE(!__is_kfree_rcu_offset(offset))) { + /* Could be optimized with kfree_bulk() in future. */ + kfree((void *)head - offset); + } rcu_lock_release(&rcu_callback_map); cond_resched_tasks_rcu_qs(); @@ -2835,16 +2837,6 @@ static void kfree_rcu_monitor(struct work_struct *work) spin_unlock_irqrestore(&krcp->lock, flags); } -/* - * This version of kfree_call_rcu does not do batching of kfree_rcu() requests. - * Used only by rcuperf torture test for comparison with kfree_rcu_batch(). - */ -void kfree_call_rcu_nobatch(struct rcu_head *head, rcu_callback_t func) -{ - __call_rcu(head, func); -} -EXPORT_SYMBOL_GPL(kfree_call_rcu_nobatch); - /* * Queue a request for lazy invocation of kfree() after a grace period. * @@ -2864,8 +2856,6 @@ void kfree_call_rcu(struct rcu_head *head, rcu_callback_t func) unsigned long flags; struct kfree_rcu_cpu *krcp; - head->func = func; - local_irq_save(flags); // For safely calling this_cpu_ptr(). krcp = this_cpu_ptr(&krc); if (krcp->initialized) -- cgit v1.2.3