From 3ccfebedd8cf54e291c809c838d8ad5cc00f5688 Mon Sep 17 00:00:00 2001 From: Mathieu Desnoyers Date: Mon, 29 Jan 2018 15:20:11 -0500 Subject: powerpc, membarrier: Skip memory barrier in switch_mm() Allow PowerPC to skip the full memory barrier in switch_mm(), and only issue the barrier when scheduling into a task belonging to a process that has registered to use expedited private. Threads targeting the same VM but which belong to different thread groups is a tricky case. It has a few consequences: It turns out that we cannot rely on get_nr_threads(p) to count the number of threads using a VM. We can use (atomic_read(&mm->mm_users) == 1 && get_nr_threads(p) == 1) instead to skip the synchronize_sched() for cases where the VM only has a single user, and that user only has a single thread. It also turns out that we cannot use for_each_thread() to set thread flags in all threads using a VM, as it only iterates on the thread group. Therefore, test the membarrier state variable directly rather than relying on thread flags. This means membarrier_register_private_expedited() needs to set the MEMBARRIER_STATE_PRIVATE_EXPEDITED flag, issue synchronize_sched(), and only then set MEMBARRIER_STATE_PRIVATE_EXPEDITED_READY which allows private expedited membarrier commands to succeed. membarrier_arch_switch_mm() now tests for the MEMBARRIER_STATE_PRIVATE_EXPEDITED flag. Signed-off-by: Mathieu Desnoyers Acked-by: Thomas Gleixner Acked-by: Peter Zijlstra (Intel) Cc: Alan Stern Cc: Alexander Viro Cc: Andrea Parri Cc: Andrew Hunter Cc: Andy Lutomirski Cc: Avi Kivity Cc: Benjamin Herrenschmidt Cc: Boqun Feng Cc: Dave Watson Cc: David Sehr Cc: Greg Hackmann Cc: Linus Torvalds Cc: Maged Michael Cc: Michael Ellerman Cc: Nicholas Piggin Cc: Paul E. McKenney Cc: Paul Mackerras Cc: Russell King Cc: Will Deacon Cc: linux-api@vger.kernel.org Cc: linux-arch@vger.kernel.org Cc: linuxppc-dev@lists.ozlabs.org Link: http://lkml.kernel.org/r/20180129202020.8515-3-mathieu.desnoyers@efficios.com Signed-off-by: Ingo Molnar --- include/linux/sched/mm.h | 13 ++++++++++++- 1 file changed, 12 insertions(+), 1 deletion(-) (limited to 'include') diff --git a/include/linux/sched/mm.h b/include/linux/sched/mm.h index 3d49b91b674d..26307cdc3969 100644 --- a/include/linux/sched/mm.h +++ b/include/linux/sched/mm.h @@ -215,14 +215,25 @@ static inline void memalloc_noreclaim_restore(unsigned int flags) #ifdef CONFIG_MEMBARRIER enum { MEMBARRIER_STATE_PRIVATE_EXPEDITED_READY = (1U << 0), - MEMBARRIER_STATE_SWITCH_MM = (1U << 1), + MEMBARRIER_STATE_PRIVATE_EXPEDITED = (1U << 1), }; +#ifdef CONFIG_ARCH_HAS_MEMBARRIER_CALLBACKS +#include +#endif + static inline void membarrier_execve(struct task_struct *t) { atomic_set(&t->mm->membarrier_state, 0); } #else +#ifdef CONFIG_ARCH_HAS_MEMBARRIER_CALLBACKS +static inline void membarrier_arch_switch_mm(struct mm_struct *prev, + struct mm_struct *next, + struct task_struct *tsk) +{ +} +#endif static inline void membarrier_execve(struct task_struct *t) { } -- cgit v1.2.3 From 306e060435d7a3aef8f6f033e43b0f581638adce Mon Sep 17 00:00:00 2001 From: Mathieu Desnoyers Date: Mon, 29 Jan 2018 15:20:12 -0500 Subject: membarrier: Document scheduler barrier requirements Document the membarrier requirement on having a full memory barrier in __schedule() after coming from user-space, before storing to rq->curr. It is provided by smp_mb__after_spinlock() in __schedule(). Document that membarrier requires a full barrier on transition from kernel thread to userspace thread. We currently have an implicit barrier from atomic_dec_and_test() in mmdrop() that ensures this. The x86 switch_mm_irqs_off() full barrier is currently provided by many cpumask update operations as well as write_cr3(). Document that write_cr3() provides this barrier. Signed-off-by: Mathieu Desnoyers Acked-by: Thomas Gleixner Acked-by: Peter Zijlstra (Intel) Cc: Andrea Parri Cc: Andrew Hunter Cc: Andy Lutomirski Cc: Avi Kivity Cc: Benjamin Herrenschmidt Cc: Boqun Feng Cc: Dave Watson Cc: David Sehr Cc: Greg Hackmann Cc: H. Peter Anvin Cc: Linus Torvalds Cc: Maged Michael Cc: Michael Ellerman Cc: Paul E. McKenney Cc: Paul Mackerras Cc: Russell King Cc: Will Deacon Cc: linux-api@vger.kernel.org Link: http://lkml.kernel.org/r/20180129202020.8515-4-mathieu.desnoyers@efficios.com Signed-off-by: Ingo Molnar --- arch/x86/mm/tlb.c | 5 +++++ include/linux/sched/mm.h | 5 +++++ kernel/sched/core.c | 37 ++++++++++++++++++++++++++----------- 3 files changed, 36 insertions(+), 11 deletions(-) (limited to 'include') diff --git a/arch/x86/mm/tlb.c b/arch/x86/mm/tlb.c index 5bfe61a5e8e3..9fa7d2e0e15e 100644 --- a/arch/x86/mm/tlb.c +++ b/arch/x86/mm/tlb.c @@ -228,6 +228,11 @@ void switch_mm_irqs_off(struct mm_struct *prev, struct mm_struct *next, #endif this_cpu_write(cpu_tlbstate.is_lazy, false); + /* + * The membarrier system call requires a full memory barrier + * before returning to user-space, after storing to rq->curr. + * Writing to CR3 provides that full memory barrier. + */ if (real_prev == next) { VM_WARN_ON(this_cpu_read(cpu_tlbstate.ctxs[prev_asid].ctx_id) != next->context.ctx_id); diff --git a/include/linux/sched/mm.h b/include/linux/sched/mm.h index 26307cdc3969..b84e0fde1d72 100644 --- a/include/linux/sched/mm.h +++ b/include/linux/sched/mm.h @@ -39,6 +39,11 @@ static inline void mmgrab(struct mm_struct *mm) extern void __mmdrop(struct mm_struct *); static inline void mmdrop(struct mm_struct *mm) { + /* + * The implicit full barrier implied by atomic_dec_and_test() is + * required by the membarrier system call before returning to + * user-space, after storing to rq->curr. + */ if (unlikely(atomic_dec_and_test(&mm->mm_count))) __mmdrop(mm); } diff --git a/kernel/sched/core.c b/kernel/sched/core.c index ead0c2135d47..11bf4d48d2d3 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -2703,6 +2703,12 @@ static struct rq *finish_task_switch(struct task_struct *prev) finish_arch_post_lock_switch(); fire_sched_in_preempt_notifiers(current); + /* + * When transitioning from a kernel thread to a userspace + * thread, mmdrop()'s implicit full barrier is required by the + * membarrier system call, because the current ->active_mm can + * become the current mm without going through switch_mm(). + */ if (mm) mmdrop(mm); if (unlikely(prev_state == TASK_DEAD)) { @@ -2808,6 +2814,13 @@ context_switch(struct rq *rq, struct task_struct *prev, */ arch_start_context_switch(prev); + /* + * If mm is non-NULL, we pass through switch_mm(). If mm is + * NULL, we will pass through mmdrop() in finish_task_switch(). + * Both of these contain the full memory barrier required by + * membarrier after storing to rq->curr, before returning to + * user-space. + */ if (!mm) { next->active_mm = oldmm; mmgrab(oldmm); @@ -3344,6 +3357,9 @@ static void __sched notrace __schedule(bool preempt) * Make sure that signal_pending_state()->signal_pending() below * can't be reordered with __set_current_state(TASK_INTERRUPTIBLE) * done by the caller to avoid the race with signal_wake_up(). + * + * The membarrier system call requires a full memory barrier + * after coming from user-space, before storing to rq->curr. */ rq_lock(rq, &rf); smp_mb__after_spinlock(); @@ -3391,17 +3407,16 @@ static void __sched notrace __schedule(bool preempt) /* * The membarrier system call requires each architecture * to have a full memory barrier after updating - * rq->curr, before returning to user-space. For TSO - * (e.g. x86), the architecture must provide its own - * barrier in switch_mm(). For weakly ordered machines - * for which spin_unlock() acts as a full memory - * barrier, finish_lock_switch() in common code takes - * care of this barrier. For weakly ordered machines for - * which spin_unlock() acts as a RELEASE barrier (only - * arm64 and PowerPC), arm64 has a full barrier in - * switch_to(), and PowerPC has - * smp_mb__after_unlock_lock() before - * finish_lock_switch(). + * rq->curr, before returning to user-space. + * + * Here are the schemes providing that barrier on the + * various architectures: + * - mm ? switch_mm() : mmdrop() for x86, s390, sparc, PowerPC. + * switch_mm() rely on membarrier_arch_switch_mm() on PowerPC. + * - finish_lock_switch() for weakly-ordered + * architectures where spin_unlock is a full barrier, + * - switch_to() for arm64 (weakly-ordered, spin_unlock + * is a RELEASE barrier), */ ++*switch_count; -- cgit v1.2.3 From c5f58bd58f432be5d92df33c5458e0bcbee3aadf Mon Sep 17 00:00:00 2001 From: Mathieu Desnoyers Date: Mon, 29 Jan 2018 15:20:13 -0500 Subject: membarrier: Provide GLOBAL_EXPEDITED command Allow expedited membarrier to be used for data shared between processes through shared memory. Processes wishing to receive the membarriers register with MEMBARRIER_CMD_REGISTER_GLOBAL_EXPEDITED. Those which want to issue membarrier invoke MEMBARRIER_CMD_GLOBAL_EXPEDITED. This allows extremely simple kernel-level implementation: we have almost everything we need with the PRIVATE_EXPEDITED barrier code. All we need to do is to add a flag in the mm_struct that will be used to check whether we need to send the IPI to the current thread of each CPU. There is a slight downside to this approach compared to targeting specific shared memory users: when performing a membarrier operation, all registered "global" receivers will get the barrier, even if they don't share a memory mapping with the sender issuing MEMBARRIER_CMD_GLOBAL_EXPEDITED. This registration approach seems to fit the requirement of not disturbing processes that really deeply care about real-time: they simply should not register with MEMBARRIER_CMD_REGISTER_GLOBAL_EXPEDITED. In order to align the membarrier command names, the "MEMBARRIER_CMD_SHARED" command is renamed to "MEMBARRIER_CMD_GLOBAL", keeping an alias of MEMBARRIER_CMD_SHARED to MEMBARRIER_CMD_GLOBAL for UAPI header backward compatibility. Signed-off-by: Mathieu Desnoyers Acked-by: Thomas Gleixner Acked-by: Peter Zijlstra (Intel) Cc: Andrea Parri Cc: Andrew Hunter Cc: Andy Lutomirski Cc: Avi Kivity Cc: Benjamin Herrenschmidt Cc: Boqun Feng Cc: Dave Watson Cc: David Sehr Cc: Greg Hackmann Cc: H. Peter Anvin Cc: Linus Torvalds Cc: Maged Michael Cc: Michael Ellerman Cc: Paul E. McKenney Cc: Paul Mackerras Cc: Russell King Cc: Will Deacon Cc: linux-api@vger.kernel.org Link: http://lkml.kernel.org/r/20180129202020.8515-5-mathieu.desnoyers@efficios.com Signed-off-by: Ingo Molnar --- arch/powerpc/include/asm/membarrier.h | 3 +- include/linux/sched/mm.h | 6 +- include/uapi/linux/membarrier.h | 42 ++++++++++-- kernel/sched/membarrier.c | 120 +++++++++++++++++++++++++++++++--- 4 files changed, 153 insertions(+), 18 deletions(-) (limited to 'include') diff --git a/arch/powerpc/include/asm/membarrier.h b/arch/powerpc/include/asm/membarrier.h index 98ff4f1fcf2b..6e20bb5c74ea 100644 --- a/arch/powerpc/include/asm/membarrier.h +++ b/arch/powerpc/include/asm/membarrier.h @@ -13,7 +13,8 @@ static inline void membarrier_arch_switch_mm(struct mm_struct *prev, * store to rq->curr. */ if (likely(!(atomic_read(&next->membarrier_state) & - MEMBARRIER_STATE_PRIVATE_EXPEDITED) || !prev)) + (MEMBARRIER_STATE_PRIVATE_EXPEDITED | + MEMBARRIER_STATE_GLOBAL_EXPEDITED)) || !prev)) return; /* diff --git a/include/linux/sched/mm.h b/include/linux/sched/mm.h index b84e0fde1d72..1c4e40c5efaf 100644 --- a/include/linux/sched/mm.h +++ b/include/linux/sched/mm.h @@ -219,8 +219,10 @@ static inline void memalloc_noreclaim_restore(unsigned int flags) #ifdef CONFIG_MEMBARRIER enum { - MEMBARRIER_STATE_PRIVATE_EXPEDITED_READY = (1U << 0), - MEMBARRIER_STATE_PRIVATE_EXPEDITED = (1U << 1), + MEMBARRIER_STATE_PRIVATE_EXPEDITED_READY = (1U << 0), + MEMBARRIER_STATE_PRIVATE_EXPEDITED = (1U << 1), + MEMBARRIER_STATE_GLOBAL_EXPEDITED_READY = (1U << 2), + MEMBARRIER_STATE_GLOBAL_EXPEDITED = (1U << 3), }; #ifdef CONFIG_ARCH_HAS_MEMBARRIER_CALLBACKS diff --git a/include/uapi/linux/membarrier.h b/include/uapi/linux/membarrier.h index 4e01ad7ffe98..d252506e1b5e 100644 --- a/include/uapi/linux/membarrier.h +++ b/include/uapi/linux/membarrier.h @@ -31,7 +31,7 @@ * enum membarrier_cmd - membarrier system call command * @MEMBARRIER_CMD_QUERY: Query the set of supported commands. It returns * a bitmask of valid commands. - * @MEMBARRIER_CMD_SHARED: Execute a memory barrier on all running threads. + * @MEMBARRIER_CMD_GLOBAL: Execute a memory barrier on all running threads. * Upon return from system call, the caller thread * is ensured that all running threads have passed * through a state where all memory accesses to @@ -40,6 +40,28 @@ * (non-running threads are de facto in such a * state). This covers threads from all processes * running on the system. This command returns 0. + * @MEMBARRIER_CMD_GLOBAL_EXPEDITED: + * Execute a memory barrier on all running threads + * of all processes which previously registered + * with MEMBARRIER_CMD_REGISTER_GLOBAL_EXPEDITED. + * Upon return from system call, the caller thread + * is ensured that all running threads have passed + * through a state where all memory accesses to + * user-space addresses match program order between + * entry to and return from the system call + * (non-running threads are de facto in such a + * state). This only covers threads from processes + * which registered with + * MEMBARRIER_CMD_REGISTER_GLOBAL_EXPEDITED. + * This command returns 0. Given that + * registration is about the intent to receive + * the barriers, it is valid to invoke + * MEMBARRIER_CMD_GLOBAL_EXPEDITED from a + * non-registered process. + * @MEMBARRIER_CMD_REGISTER_GLOBAL_EXPEDITED: + * Register the process intent to receive + * MEMBARRIER_CMD_GLOBAL_EXPEDITED memory + * barriers. Always returns 0. * @MEMBARRIER_CMD_PRIVATE_EXPEDITED: * Execute a memory barrier on each running * thread belonging to the same process as the current @@ -64,18 +86,24 @@ * Register the process intent to use * MEMBARRIER_CMD_PRIVATE_EXPEDITED. Always * returns 0. + * @MEMBARRIER_CMD_SHARED: + * Alias to MEMBARRIER_CMD_GLOBAL. Provided for + * header backward compatibility. * * Command to be passed to the membarrier system call. The commands need to * be a single bit each, except for MEMBARRIER_CMD_QUERY which is assigned to * the value 0. */ enum membarrier_cmd { - MEMBARRIER_CMD_QUERY = 0, - MEMBARRIER_CMD_SHARED = (1 << 0), - /* reserved for MEMBARRIER_CMD_SHARED_EXPEDITED (1 << 1) */ - /* reserved for MEMBARRIER_CMD_PRIVATE (1 << 2) */ - MEMBARRIER_CMD_PRIVATE_EXPEDITED = (1 << 3), - MEMBARRIER_CMD_REGISTER_PRIVATE_EXPEDITED = (1 << 4), + MEMBARRIER_CMD_QUERY = 0, + MEMBARRIER_CMD_GLOBAL = (1 << 0), + MEMBARRIER_CMD_GLOBAL_EXPEDITED = (1 << 1), + MEMBARRIER_CMD_REGISTER_GLOBAL_EXPEDITED = (1 << 2), + MEMBARRIER_CMD_PRIVATE_EXPEDITED = (1 << 3), + MEMBARRIER_CMD_REGISTER_PRIVATE_EXPEDITED = (1 << 4), + + /* Alias for header backward compatibility. */ + MEMBARRIER_CMD_SHARED = MEMBARRIER_CMD_GLOBAL, }; #endif /* _UAPI_LINUX_MEMBARRIER_H */ diff --git a/kernel/sched/membarrier.c b/kernel/sched/membarrier.c index 678577267a9a..d2087d5f9837 100644 --- a/kernel/sched/membarrier.c +++ b/kernel/sched/membarrier.c @@ -27,7 +27,9 @@ * except MEMBARRIER_CMD_QUERY. */ #define MEMBARRIER_CMD_BITMASK \ - (MEMBARRIER_CMD_SHARED | MEMBARRIER_CMD_PRIVATE_EXPEDITED \ + (MEMBARRIER_CMD_GLOBAL | MEMBARRIER_CMD_GLOBAL_EXPEDITED \ + | MEMBARRIER_CMD_REGISTER_GLOBAL_EXPEDITED \ + | MEMBARRIER_CMD_PRIVATE_EXPEDITED \ | MEMBARRIER_CMD_REGISTER_PRIVATE_EXPEDITED) static void ipi_mb(void *info) @@ -35,6 +37,73 @@ static void ipi_mb(void *info) smp_mb(); /* IPIs should be serializing but paranoid. */ } +static int membarrier_global_expedited(void) +{ + int cpu; + bool fallback = false; + cpumask_var_t tmpmask; + + if (num_online_cpus() == 1) + return 0; + + /* + * Matches memory barriers around rq->curr modification in + * scheduler. + */ + smp_mb(); /* system call entry is not a mb. */ + + /* + * Expedited membarrier commands guarantee that they won't + * block, hence the GFP_NOWAIT allocation flag and fallback + * implementation. + */ + if (!zalloc_cpumask_var(&tmpmask, GFP_NOWAIT)) { + /* Fallback for OOM. */ + fallback = true; + } + + cpus_read_lock(); + for_each_online_cpu(cpu) { + struct task_struct *p; + + /* + * Skipping the current CPU is OK even through we can be + * migrated at any point. The current CPU, at the point + * where we read raw_smp_processor_id(), is ensured to + * be in program order with respect to the caller + * thread. Therefore, we can skip this CPU from the + * iteration. + */ + if (cpu == raw_smp_processor_id()) + continue; + rcu_read_lock(); + p = task_rcu_dereference(&cpu_rq(cpu)->curr); + if (p && p->mm && (atomic_read(&p->mm->membarrier_state) & + MEMBARRIER_STATE_GLOBAL_EXPEDITED)) { + if (!fallback) + __cpumask_set_cpu(cpu, tmpmask); + else + smp_call_function_single(cpu, ipi_mb, NULL, 1); + } + rcu_read_unlock(); + } + if (!fallback) { + preempt_disable(); + smp_call_function_many(tmpmask, ipi_mb, NULL, 1); + preempt_enable(); + free_cpumask_var(tmpmask); + } + cpus_read_unlock(); + + /* + * Memory barrier on the caller thread _after_ we finished + * waiting for the last IPI. Matches memory barriers around + * rq->curr modification in scheduler. + */ + smp_mb(); /* exit from system call is not a mb */ + return 0; +} + static int membarrier_private_expedited(void) { int cpu; @@ -105,7 +174,38 @@ static int membarrier_private_expedited(void) return 0; } -static void membarrier_register_private_expedited(void) +static int membarrier_register_global_expedited(void) +{ + struct task_struct *p = current; + struct mm_struct *mm = p->mm; + + if (atomic_read(&mm->membarrier_state) & + MEMBARRIER_STATE_GLOBAL_EXPEDITED_READY) + return 0; + atomic_or(MEMBARRIER_STATE_GLOBAL_EXPEDITED, &mm->membarrier_state); + if (atomic_read(&mm->mm_users) == 1 && get_nr_threads(p) == 1) { + /* + * For single mm user, single threaded process, we can + * simply issue a memory barrier after setting + * MEMBARRIER_STATE_GLOBAL_EXPEDITED to guarantee that + * no memory access following registration is reordered + * before registration. + */ + smp_mb(); + } else { + /* + * For multi-mm user threads, we need to ensure all + * future scheduler executions will observe the new + * thread flag state for this mm. + */ + synchronize_sched(); + } + atomic_or(MEMBARRIER_STATE_GLOBAL_EXPEDITED_READY, + &mm->membarrier_state); + return 0; +} + +static int membarrier_register_private_expedited(void) { struct task_struct *p = current; struct mm_struct *mm = p->mm; @@ -117,7 +217,7 @@ static void membarrier_register_private_expedited(void) */ if (atomic_read(&mm->membarrier_state) & MEMBARRIER_STATE_PRIVATE_EXPEDITED_READY) - return; + return 0; atomic_or(MEMBARRIER_STATE_PRIVATE_EXPEDITED, &mm->membarrier_state); if (!(atomic_read(&mm->mm_users) == 1 && get_nr_threads(p) == 1)) { /* @@ -128,6 +228,7 @@ static void membarrier_register_private_expedited(void) } atomic_or(MEMBARRIER_STATE_PRIVATE_EXPEDITED_READY, &mm->membarrier_state); + return 0; } /** @@ -167,21 +268,24 @@ SYSCALL_DEFINE2(membarrier, int, cmd, int, flags) int cmd_mask = MEMBARRIER_CMD_BITMASK; if (tick_nohz_full_enabled()) - cmd_mask &= ~MEMBARRIER_CMD_SHARED; + cmd_mask &= ~MEMBARRIER_CMD_GLOBAL; return cmd_mask; } - case MEMBARRIER_CMD_SHARED: - /* MEMBARRIER_CMD_SHARED is not compatible with nohz_full. */ + case MEMBARRIER_CMD_GLOBAL: + /* MEMBARRIER_CMD_GLOBAL is not compatible with nohz_full. */ if (tick_nohz_full_enabled()) return -EINVAL; if (num_online_cpus() > 1) synchronize_sched(); return 0; + case MEMBARRIER_CMD_GLOBAL_EXPEDITED: + return membarrier_global_expedited(); + case MEMBARRIER_CMD_REGISTER_GLOBAL_EXPEDITED: + return membarrier_register_global_expedited(); case MEMBARRIER_CMD_PRIVATE_EXPEDITED: return membarrier_private_expedited(); case MEMBARRIER_CMD_REGISTER_PRIVATE_EXPEDITED: - membarrier_register_private_expedited(); - return 0; + return membarrier_register_private_expedited(); default: return -EINVAL; } -- cgit v1.2.3 From e61938a921a46347d7725badc40ec436ebfff977 Mon Sep 17 00:00:00 2001 From: Mathieu Desnoyers Date: Mon, 29 Jan 2018 15:20:15 -0500 Subject: locking: Introduce sync_core_before_usermode() Introduce an architecture function that ensures the current CPU issues a core serializing instruction before returning to usermode. This is needed for the membarrier "sync_core" command. Architectures defining the sync_core_before_usermode() static inline need to select ARCH_HAS_SYNC_CORE_BEFORE_USERMODE. Signed-off-by: Mathieu Desnoyers Acked-by: Thomas Gleixner Acked-by: Peter Zijlstra (Intel) Cc: Andrea Parri Cc: Andrew Hunter Cc: Andy Lutomirski Cc: Arnd Bergmann Cc: Avi Kivity Cc: Benjamin Herrenschmidt Cc: Boqun Feng Cc: Dave Watson Cc: David Sehr Cc: Greg Hackmann Cc: H. Peter Anvin Cc: Linus Torvalds Cc: Maged Michael Cc: Michael Ellerman Cc: Paul E. McKenney Cc: Paul Mackerras Cc: Russell King Cc: Will Deacon Cc: linux-api@vger.kernel.org Cc: linux-arch@vger.kernel.org Link: http://lkml.kernel.org/r/20180129202020.8515-7-mathieu.desnoyers@efficios.com Signed-off-by: Ingo Molnar --- include/linux/sync_core.h | 21 +++++++++++++++++++++ init/Kconfig | 3 +++ 2 files changed, 24 insertions(+) create mode 100644 include/linux/sync_core.h (limited to 'include') diff --git a/include/linux/sync_core.h b/include/linux/sync_core.h new file mode 100644 index 000000000000..013da4b8b327 --- /dev/null +++ b/include/linux/sync_core.h @@ -0,0 +1,21 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +#ifndef _LINUX_SYNC_CORE_H +#define _LINUX_SYNC_CORE_H + +#ifdef CONFIG_ARCH_HAS_SYNC_CORE_BEFORE_USERMODE +#include +#else +/* + * This is a dummy sync_core_before_usermode() implementation that can be used + * on all architectures which return to user-space through core serializing + * instructions. + * If your architecture returns to user-space through non-core-serializing + * instructions, you need to write your own functions. + */ +static inline void sync_core_before_usermode(void) +{ +} +#endif + +#endif /* _LINUX_SYNC_CORE_H */ + diff --git a/init/Kconfig b/init/Kconfig index 837adcf075d9..535421facea8 100644 --- a/init/Kconfig +++ b/init/Kconfig @@ -1918,3 +1918,6 @@ config ASN1 functions to call on what tags. source "kernel/Kconfig.locks" + +config ARCH_HAS_SYNC_CORE_BEFORE_USERMODE + bool -- cgit v1.2.3 From 70216e18e519a54a2f13569e8caff99a092a92d6 Mon Sep 17 00:00:00 2001 From: Mathieu Desnoyers Date: Mon, 29 Jan 2018 15:20:17 -0500 Subject: membarrier: Provide core serializing command, *_SYNC_CORE Provide core serializing membarrier command to support memory reclaim by JIT. Each architecture needs to explicitly opt into that support by documenting in their architecture code how they provide the core serializing instructions required when returning from the membarrier IPI, and after the scheduler has updated the curr->mm pointer (before going back to user-space). They should then select ARCH_HAS_MEMBARRIER_SYNC_CORE to enable support for that command on their architecture. Architectures selecting this feature need to either document that they issue core serializing instructions when returning to user-space, or implement their architecture-specific sync_core_before_usermode(). Signed-off-by: Mathieu Desnoyers Acked-by: Thomas Gleixner Acked-by: Peter Zijlstra (Intel) Cc: Andrea Parri Cc: Andrew Hunter Cc: Andy Lutomirski Cc: Avi Kivity Cc: Benjamin Herrenschmidt Cc: Boqun Feng Cc: Dave Watson Cc: David Sehr Cc: Greg Hackmann Cc: H. Peter Anvin Cc: Linus Torvalds Cc: Maged Michael Cc: Michael Ellerman Cc: Paul E. McKenney Cc: Paul Mackerras Cc: Russell King Cc: Will Deacon Cc: linux-api@vger.kernel.org Cc: linux-arch@vger.kernel.org Link: http://lkml.kernel.org/r/20180129202020.8515-9-mathieu.desnoyers@efficios.com Signed-off-by: Ingo Molnar --- include/linux/sched/mm.h | 18 ++++++++++++++ include/uapi/linux/membarrier.h | 32 ++++++++++++++++++++++++- init/Kconfig | 3 +++ kernel/sched/core.c | 18 ++++++++++---- kernel/sched/membarrier.c | 53 +++++++++++++++++++++++++++++++---------- 5 files changed, 106 insertions(+), 18 deletions(-) (limited to 'include') diff --git a/include/linux/sched/mm.h b/include/linux/sched/mm.h index 1c4e40c5efaf..03a169087a18 100644 --- a/include/linux/sched/mm.h +++ b/include/linux/sched/mm.h @@ -7,6 +7,7 @@ #include #include #include +#include /* * Routines for handling mm_structs @@ -223,12 +224,26 @@ enum { MEMBARRIER_STATE_PRIVATE_EXPEDITED = (1U << 1), MEMBARRIER_STATE_GLOBAL_EXPEDITED_READY = (1U << 2), MEMBARRIER_STATE_GLOBAL_EXPEDITED = (1U << 3), + MEMBARRIER_STATE_PRIVATE_EXPEDITED_SYNC_CORE_READY = (1U << 4), + MEMBARRIER_STATE_PRIVATE_EXPEDITED_SYNC_CORE = (1U << 5), +}; + +enum { + MEMBARRIER_FLAG_SYNC_CORE = (1U << 0), }; #ifdef CONFIG_ARCH_HAS_MEMBARRIER_CALLBACKS #include #endif +static inline void membarrier_mm_sync_core_before_usermode(struct mm_struct *mm) +{ + if (likely(!(atomic_read(&mm->membarrier_state) & + MEMBARRIER_STATE_PRIVATE_EXPEDITED_SYNC_CORE))) + return; + sync_core_before_usermode(); +} + static inline void membarrier_execve(struct task_struct *t) { atomic_set(&t->mm->membarrier_state, 0); @@ -244,6 +259,9 @@ static inline void membarrier_arch_switch_mm(struct mm_struct *prev, static inline void membarrier_execve(struct task_struct *t) { } +static inline void membarrier_mm_sync_core_before_usermode(struct mm_struct *mm) +{ +} #endif #endif /* _LINUX_SCHED_MM_H */ diff --git a/include/uapi/linux/membarrier.h b/include/uapi/linux/membarrier.h index d252506e1b5e..5891d7614c8c 100644 --- a/include/uapi/linux/membarrier.h +++ b/include/uapi/linux/membarrier.h @@ -73,7 +73,7 @@ * to and return from the system call * (non-running threads are de facto in such a * state). This only covers threads from the - * same processes as the caller thread. This + * same process as the caller thread. This * command returns 0 on success. The * "expedited" commands complete faster than * the non-expedited ones, they never block, @@ -86,6 +86,34 @@ * Register the process intent to use * MEMBARRIER_CMD_PRIVATE_EXPEDITED. Always * returns 0. + * @MEMBARRIER_CMD_PRIVATE_EXPEDITED_SYNC_CORE: + * In addition to provide memory ordering + * guarantees described in + * MEMBARRIER_CMD_PRIVATE_EXPEDITED, ensure + * the caller thread, upon return from system + * call, that all its running threads siblings + * have executed a core serializing + * instruction. (architectures are required to + * guarantee that non-running threads issue + * core serializing instructions before they + * resume user-space execution). This only + * covers threads from the same process as the + * caller thread. This command returns 0 on + * success. The "expedited" commands complete + * faster than the non-expedited ones, they + * never block, but have the downside of + * causing extra overhead. If this command is + * not implemented by an architecture, -EINVAL + * is returned. A process needs to register its + * intent to use the private expedited sync + * core command prior to using it, otherwise + * this command returns -EPERM. + * @MEMBARRIER_CMD_REGISTER_PRIVATE_EXPEDITED_SYNC_CORE: + * Register the process intent to use + * MEMBARRIER_CMD_PRIVATE_EXPEDITED_SYNC_CORE. + * If this command is not implemented by an + * architecture, -EINVAL is returned. + * Returns 0 on success. * @MEMBARRIER_CMD_SHARED: * Alias to MEMBARRIER_CMD_GLOBAL. Provided for * header backward compatibility. @@ -101,6 +129,8 @@ enum membarrier_cmd { MEMBARRIER_CMD_REGISTER_GLOBAL_EXPEDITED = (1 << 2), MEMBARRIER_CMD_PRIVATE_EXPEDITED = (1 << 3), MEMBARRIER_CMD_REGISTER_PRIVATE_EXPEDITED = (1 << 4), + MEMBARRIER_CMD_PRIVATE_EXPEDITED_SYNC_CORE = (1 << 5), + MEMBARRIER_CMD_REGISTER_PRIVATE_EXPEDITED_SYNC_CORE = (1 << 6), /* Alias for header backward compatibility. */ MEMBARRIER_CMD_SHARED = MEMBARRIER_CMD_GLOBAL, diff --git a/init/Kconfig b/init/Kconfig index 535421facea8..e37f4b2a6445 100644 --- a/init/Kconfig +++ b/init/Kconfig @@ -1415,6 +1415,9 @@ config USERFAULTFD config ARCH_HAS_MEMBARRIER_CALLBACKS bool +config ARCH_HAS_MEMBARRIER_SYNC_CORE + bool + config EMBEDDED bool "Embedded system" option allnoconfig_y diff --git a/kernel/sched/core.c b/kernel/sched/core.c index 11bf4d48d2d3..ee420d78e674 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -2704,13 +2704,21 @@ static struct rq *finish_task_switch(struct task_struct *prev) fire_sched_in_preempt_notifiers(current); /* - * When transitioning from a kernel thread to a userspace - * thread, mmdrop()'s implicit full barrier is required by the - * membarrier system call, because the current ->active_mm can - * become the current mm without going through switch_mm(). + * When switching through a kernel thread, the loop in + * membarrier_{private,global}_expedited() may have observed that + * kernel thread and not issued an IPI. It is therefore possible to + * schedule between user->kernel->user threads without passing though + * switch_mm(). Membarrier requires a barrier after storing to + * rq->curr, before returning to userspace, so provide them here: + * + * - a full memory barrier for {PRIVATE,GLOBAL}_EXPEDITED, implicitly + * provided by mmdrop(), + * - a sync_core for SYNC_CORE. */ - if (mm) + if (mm) { + membarrier_mm_sync_core_before_usermode(mm); mmdrop(mm); + } if (unlikely(prev_state == TASK_DEAD)) { if (prev->sched_class->task_dead) prev->sched_class->task_dead(prev); diff --git a/kernel/sched/membarrier.c b/kernel/sched/membarrier.c index d2087d5f9837..5d0762633639 100644 --- a/kernel/sched/membarrier.c +++ b/kernel/sched/membarrier.c @@ -26,11 +26,20 @@ * Bitmask made from a "or" of all commands within enum membarrier_cmd, * except MEMBARRIER_CMD_QUERY. */ +#ifdef CONFIG_ARCH_HAS_MEMBARRIER_SYNC_CORE +#define MEMBARRIER_PRIVATE_EXPEDITED_SYNC_CORE_BITMASK \ + (MEMBARRIER_CMD_PRIVATE_EXPEDITED_SYNC_CORE \ + | MEMBARRIER_CMD_REGISTER_PRIVATE_EXPEDITED_SYNC_CORE) +#else +#define MEMBARRIER_PRIVATE_EXPEDITED_SYNC_CORE_BITMASK 0 +#endif + #define MEMBARRIER_CMD_BITMASK \ (MEMBARRIER_CMD_GLOBAL | MEMBARRIER_CMD_GLOBAL_EXPEDITED \ | MEMBARRIER_CMD_REGISTER_GLOBAL_EXPEDITED \ | MEMBARRIER_CMD_PRIVATE_EXPEDITED \ - | MEMBARRIER_CMD_REGISTER_PRIVATE_EXPEDITED) + | MEMBARRIER_CMD_REGISTER_PRIVATE_EXPEDITED \ + | MEMBARRIER_PRIVATE_EXPEDITED_SYNC_CORE_BITMASK) static void ipi_mb(void *info) { @@ -104,15 +113,23 @@ static int membarrier_global_expedited(void) return 0; } -static int membarrier_private_expedited(void) +static int membarrier_private_expedited(int flags) { int cpu; bool fallback = false; cpumask_var_t tmpmask; - if (!(atomic_read(¤t->mm->membarrier_state) - & MEMBARRIER_STATE_PRIVATE_EXPEDITED_READY)) - return -EPERM; + if (flags & MEMBARRIER_FLAG_SYNC_CORE) { + if (!IS_ENABLED(CONFIG_ARCH_HAS_MEMBARRIER_SYNC_CORE)) + return -EINVAL; + if (!(atomic_read(¤t->mm->membarrier_state) & + MEMBARRIER_STATE_PRIVATE_EXPEDITED_SYNC_CORE_READY)) + return -EPERM; + } else { + if (!(atomic_read(¤t->mm->membarrier_state) & + MEMBARRIER_STATE_PRIVATE_EXPEDITED_READY)) + return -EPERM; + } if (num_online_cpus() == 1) return 0; @@ -205,20 +222,29 @@ static int membarrier_register_global_expedited(void) return 0; } -static int membarrier_register_private_expedited(void) +static int membarrier_register_private_expedited(int flags) { struct task_struct *p = current; struct mm_struct *mm = p->mm; + int state = MEMBARRIER_STATE_PRIVATE_EXPEDITED_READY; + + if (flags & MEMBARRIER_FLAG_SYNC_CORE) { + if (!IS_ENABLED(CONFIG_ARCH_HAS_MEMBARRIER_SYNC_CORE)) + return -EINVAL; + state = MEMBARRIER_STATE_PRIVATE_EXPEDITED_SYNC_CORE_READY; + } /* * We need to consider threads belonging to different thread * groups, which use the same mm. (CLONE_VM but not * CLONE_THREAD). */ - if (atomic_read(&mm->membarrier_state) - & MEMBARRIER_STATE_PRIVATE_EXPEDITED_READY) + if (atomic_read(&mm->membarrier_state) & state) return 0; atomic_or(MEMBARRIER_STATE_PRIVATE_EXPEDITED, &mm->membarrier_state); + if (flags & MEMBARRIER_FLAG_SYNC_CORE) + atomic_or(MEMBARRIER_STATE_PRIVATE_EXPEDITED_SYNC_CORE, + &mm->membarrier_state); if (!(atomic_read(&mm->mm_users) == 1 && get_nr_threads(p) == 1)) { /* * Ensure all future scheduler executions will observe the @@ -226,8 +252,7 @@ static int membarrier_register_private_expedited(void) */ synchronize_sched(); } - atomic_or(MEMBARRIER_STATE_PRIVATE_EXPEDITED_READY, - &mm->membarrier_state); + atomic_or(state, &mm->membarrier_state); return 0; } @@ -283,9 +308,13 @@ SYSCALL_DEFINE2(membarrier, int, cmd, int, flags) case MEMBARRIER_CMD_REGISTER_GLOBAL_EXPEDITED: return membarrier_register_global_expedited(); case MEMBARRIER_CMD_PRIVATE_EXPEDITED: - return membarrier_private_expedited(); + return membarrier_private_expedited(0); case MEMBARRIER_CMD_REGISTER_PRIVATE_EXPEDITED: - return membarrier_register_private_expedited(); + return membarrier_register_private_expedited(0); + case MEMBARRIER_CMD_PRIVATE_EXPEDITED_SYNC_CORE: + return membarrier_private_expedited(MEMBARRIER_FLAG_SYNC_CORE); + case MEMBARRIER_CMD_REGISTER_PRIVATE_EXPEDITED_SYNC_CORE: + return membarrier_register_private_expedited(MEMBARRIER_FLAG_SYNC_CORE); default: return -EINVAL; } -- cgit v1.2.3 From 32e839dda3ba576943365f0f5817ce5c843137dc Mon Sep 17 00:00:00 2001 From: Mel Gorman Date: Tue, 30 Jan 2018 10:45:55 +0000 Subject: sched/fair: Use a recently used CPU as an idle candidate and the basis for SIS The select_idle_sibling() (SIS) rewrite in commit: 10e2f1acd010 ("sched/core: Rewrite and improve select_idle_siblings()") ... replaced a domain iteration with a search that broadly speaking does a wrapped walk of the scheduler domain sharing a last-level-cache. While this had a number of improvements, one consequence is that two tasks that share a waker/wakee relationship push each other around a socket. Even though two tasks may be active, all cores are evenly used. This is great from a search perspective and spreads a load across individual cores, but it has adverse consequences for cpufreq. As each CPU has relatively low utilisation, cpufreq may decide the utilisation is too low to used a higher P-state and overall computation throughput suffers. While individual cpufreq and cpuidle drivers may compensate by artifically boosting P-state (at c0) or avoiding lower C-states (during idle), it does not help if hardware-based cpufreq (e.g. HWP) is used. This patch tracks a recently used CPU based on what CPU a task was running on when it last was a waker a CPU it was recently using when a task is a wakee. During SIS, the recently used CPU is used as a target if it's still allowed by the task and is idle. The benefit may be non-obvious so consider an example of two tasks communicating back and forth. Task A may be an application doing IO where task B is a kworker or kthread like journald. Task A may issue IO, wake B and B wakes up A on completion. With the existing scheme this may look like the following (potentially different IDs if SMT is in use but similar principal applies). A (cpu 0) wake B (wakes on cpu 1) B (cpu 1) wake A (wakes on cpu 2) A (cpu 2) wake B (wakes on cpu 3) etc. A careful reader may wonder why CPU 0 was not idle when B wakes A the first time and it's simply due to the fact that A can be rescheduled to another CPU and the pattern is that prev == target when B tries to wakeup A and the information about CPU 0 has been lost. With this patch, the pattern is more likely to be: A (cpu 0) wake B (wakes on cpu 1) B (cpu 1) wake A (wakes on cpu 0) A (cpu 0) wake B (wakes on cpu 1) etc i.e. two communicating casts are more likely to use just two cores instead of all available cores sharing a LLC. The most dramatic speedup was noticed on dbench using the XFS filesystem on UMA as clients interact heavily with workqueues in that configuration. Note that a similar speedup is not observed on ext4 as the wakeup pattern is different: 4.15.0-rc9 4.15.0-rc9 waprev-v1 biasancestor-v1 Hmean 1 287.54 ( 0.00%) 817.01 ( 184.14%) Hmean 2 1268.12 ( 0.00%) 1781.24 ( 40.46%) Hmean 4 1739.68 ( 0.00%) 1594.47 ( -8.35%) Hmean 8 2464.12 ( 0.00%) 2479.56 ( 0.63%) Hmean 64 1455.57 ( 0.00%) 1434.68 ( -1.44%) The results can be less dramatic on NUMA where automatic balancing interferes with the test. It's also known that network benchmarks running on localhost also benefit quite a bit from this patch (roughly 10% on netperf RR for UDP and TCP depending on the machine). Hackbench also seens small improvements (6-11% depending on machine and thread count). The facebook schbench was also tested but in most cases showed little or no different to wakeup latencies. Signed-off-by: Mel Gorman Signed-off-by: Peter Zijlstra (Intel) Cc: Linus Torvalds Cc: Matt Fleming Cc: Mike Galbraith Cc: Peter Zijlstra Cc: Thomas Gleixner Link: http://lkml.kernel.org/r/20180130104555.4125-5-mgorman@techsingularity.net Signed-off-by: Ingo Molnar --- include/linux/sched.h | 8 ++++++++ kernel/sched/core.c | 1 + kernel/sched/fair.c | 22 ++++++++++++++++++++-- 3 files changed, 29 insertions(+), 2 deletions(-) (limited to 'include') diff --git a/include/linux/sched.h b/include/linux/sched.h index 166144c04ef6..92744e3f1556 100644 --- a/include/linux/sched.h +++ b/include/linux/sched.h @@ -555,6 +555,14 @@ struct task_struct { unsigned long wakee_flip_decay_ts; struct task_struct *last_wakee; + /* + * recent_used_cpu is initially set as the last CPU used by a task + * that wakes affine another task. Waker/wakee relationships can + * push tasks around a CPU where each wakeup moves to the next one. + * Tracking a recently used CPU allows a quick search for a recently + * used CPU that may be idle. + */ + int recent_used_cpu; int wake_cpu; #endif int on_rq; diff --git a/kernel/sched/core.c b/kernel/sched/core.c index b40540e68104..36f113ac6353 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -2461,6 +2461,7 @@ void wake_up_new_task(struct task_struct *p) * Use __set_task_cpu() to avoid calling sched_class::migrate_task_rq, * as we're not fully set-up yet. */ + p->recent_used_cpu = task_cpu(p); __set_task_cpu(p, select_task_rq(p, task_cpu(p), SD_BALANCE_FORK, 0)); #endif rq = __task_rq_lock(p, &rf); diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index db45b3554682..5eb3ffc9be84 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -6197,7 +6197,7 @@ static int select_idle_cpu(struct task_struct *p, struct sched_domain *sd, int t static int select_idle_sibling(struct task_struct *p, int prev, int target) { struct sched_domain *sd; - int i; + int i, recent_used_cpu; if (idle_cpu(target)) return target; @@ -6208,6 +6208,21 @@ static int select_idle_sibling(struct task_struct *p, int prev, int target) if (prev != target && cpus_share_cache(prev, target) && idle_cpu(prev)) return prev; + /* Check a recently used CPU as a potential idle candidate */ + recent_used_cpu = p->recent_used_cpu; + if (recent_used_cpu != prev && + recent_used_cpu != target && + cpus_share_cache(recent_used_cpu, target) && + idle_cpu(recent_used_cpu) && + cpumask_test_cpu(p->recent_used_cpu, &p->cpus_allowed)) { + /* + * Replace recent_used_cpu with prev as it is a potential + * candidate for the next wake. + */ + p->recent_used_cpu = prev; + return recent_used_cpu; + } + sd = rcu_dereference(per_cpu(sd_llc, target)); if (!sd) return target; @@ -6375,9 +6390,12 @@ select_task_rq_fair(struct task_struct *p, int prev_cpu, int sd_flag, int wake_f if (!sd) { pick_cpu: - if (sd_flag & SD_BALANCE_WAKE) /* XXX always ? */ + if (sd_flag & SD_BALANCE_WAKE) { /* XXX always ? */ new_cpu = select_idle_sibling(p, prev_cpu, new_cpu); + if (want_affine) + current->recent_used_cpu = cpu; + } } else { new_cpu = find_idlest_cpu(sd, p, cpu, prev_cpu, sd_flag); } -- cgit v1.2.3