smp: Cleanup smp_call_function*()

Get rid of the __call_single_node union and cleanup the API a little to avoid external code relying on the structure layout as much. Signed-off-by: Peter Zijlstra (Intel) <[email protected]> Reviewed-by: Frederic Weisbecker <[email protected]>
vieko · Nov 24, 2020 · 545b8c8 · 545b8c8
1 parent 7a9f50a
commit 545b8c8
Show file tree

Hide file tree

Showing 13 changed files with 60 additions and 94 deletions.
diff --git a/arch/mips/kernel/process.c b/arch/mips/kernel/process.c
@@ -702,7 +702,6 @@ unsigned long arch_align_stack(unsigned long sp)
 	return sp & ALMASK;
 }
 
-static DEFINE_PER_CPU(call_single_data_t, backtrace_csd);
 static struct cpumask backtrace_csd_busy;
 
 static void handle_backtrace(void *info)
@@ -711,6 +710,9 @@ static void handle_backtrace(void *info)
 	cpumask_clear_cpu(smp_processor_id(), &backtrace_csd_busy);
 }
 
+static DEFINE_PER_CPU(call_single_data_t, backtrace_csd) =
+	CSD_INIT(handle_backtrace, NULL);
+
 static void raise_backtrace(cpumask_t *mask)
 {
 	call_single_data_t *csd;
@@ -730,7 +732,6 @@ static void raise_backtrace(cpumask_t *mask)
 		}
 
 		csd = &per_cpu(backtrace_csd, cpu);
-		csd->func = handle_backtrace;
 		smp_call_function_single_async(cpu, csd);
 	}
 }

diff --git a/arch/mips/kernel/smp.c b/arch/mips/kernel/smp.c
@@ -687,36 +687,23 @@ EXPORT_SYMBOL(flush_tlb_one);
 
 #ifdef CONFIG_GENERIC_CLOCKEVENTS_BROADCAST
 
-static DEFINE_PER_CPU(call_single_data_t, tick_broadcast_csd);
-
-void tick_broadcast(const struct cpumask *mask)
-{
-	call_single_data_t *csd;
-	int cpu;
-
-	for_each_cpu(cpu, mask) {
-		csd = &per_cpu(tick_broadcast_csd, cpu);
-		smp_call_function_single_async(cpu, csd);
-	}
-}
-
 static void tick_broadcast_callee(void *info)
 {
 	tick_receive_broadcast();
 }
 
-static int __init tick_broadcast_init(void)
+static DEFINE_PER_CPU(call_single_data_t, tick_broadcast_csd) =
+	CSD_INIT(tick_broadcast_callee, NULL);
+
+void tick_broadcast(const struct cpumask *mask)
 {
 	call_single_data_t *csd;
 	int cpu;
 
-	for (cpu = 0; cpu < NR_CPUS; cpu++) {
+	for_each_cpu(cpu, mask) {
 		csd = &per_cpu(tick_broadcast_csd, cpu);
-		csd->func = tick_broadcast_callee;
+		smp_call_function_single_async(cpu, csd);
 	}
-
-	return 0;
 }
-early_initcall(tick_broadcast_init);
 
 #endif /* CONFIG_GENERIC_CLOCKEVENTS_BROADCAST */
diff --git a/arch/s390/pci/pci_irq.c b/arch/s390/pci/pci_irq.c
@@ -178,9 +178,7 @@ static void zpci_handle_fallback_irq(void)
 		if (atomic_inc_return(&cpu_data->scheduled) > 1)
 			continue;
 
-		cpu_data->csd.func = zpci_handle_remote_irq;
-		cpu_data->csd.info = &cpu_data->scheduled;
-		cpu_data->csd.flags = 0;
+		INIT_CSD(&cpu_data->csd, zpci_handle_remote_irq, &cpu_data->scheduled);
 		smp_call_function_single_async(cpu, &cpu_data->csd);
 	}
 }

diff --git a/arch/x86/kernel/cpuid.c b/arch/x86/kernel/cpuid.c
@@ -74,10 +74,9 @@ static ssize_t cpuid_read(struct file *file, char __user *buf,
 
 	init_completion(&cmd.done);
 	for (; count; count -= 16) {
-		call_single_data_t csd = {
-			.func = cpuid_smp_cpuid,
-			.info = &cmd,
-		};
+		call_single_data_t csd;
+
+		INIT_CSD(&csd, cpuid_smp_cpuid, &cmd);
 
 		cmd.regs.eax = pos;
 		cmd.regs.ecx = pos >> 32;

diff --git a/arch/x86/lib/msr-smp.c b/arch/x86/lib/msr-smp.c
@@ -169,12 +169,11 @@ static void __wrmsr_safe_on_cpu(void *info)
 int rdmsr_safe_on_cpu(unsigned int cpu, u32 msr_no, u32 *l, u32 *h)
 {
 	struct msr_info_completion rv;
-	call_single_data_t csd = {
-		.func	= __rdmsr_safe_on_cpu,
-		.info	= &rv,
-	};
+	call_single_data_t csd;
 	int err;
 
+	INIT_CSD(&csd, __rdmsr_safe_on_cpu, &rv);
+
 	memset(&rv, 0, sizeof(rv));
 	init_completion(&rv.done);
 	rv.msr.msr_no = msr_no;

diff --git a/block/blk-mq.c b/block/blk-mq.c
@@ -671,9 +671,7 @@ bool blk_mq_complete_request_remote(struct request *rq)
 		return false;
 
 	if (blk_mq_complete_need_ipi(rq)) {
-		rq->csd.func = __blk_mq_complete_request_remote;
-		rq->csd.info = rq;
-		rq->csd.flags = 0;
+		INIT_CSD(&rq->csd, __blk_mq_complete_request_remote, rq);
 		smp_call_function_single_async(rq->mq_ctx->cpu, &rq->csd);
 	} else {
 		if (rq->q->nr_hw_queues > 1)

diff --git a/drivers/cpuidle/coupled.c b/drivers/cpuidle/coupled.c
@@ -674,8 +674,7 @@ int cpuidle_coupled_register_device(struct cpuidle_device *dev)
 	coupled->refcnt++;
 
 	csd = &per_cpu(cpuidle_coupled_poke_cb, dev->cpu);
-	csd->func = cpuidle_coupled_handle_poke;
-	csd->info = (void *)(unsigned long)dev->cpu;
+	INIT_CSD(csd, cpuidle_coupled_handle_poke, (void *)(unsigned long)dev->cpu);
 
 	return 0;
 }

diff --git a/drivers/net/ethernet/cavium/liquidio/lio_core.c b/drivers/net/ethernet/cavium/liquidio/lio_core.c
@@ -729,13 +729,8 @@ static void liquidio_napi_drv_callback(void *arg)
 	    droq->cpu_id == this_cpu) {
 		napi_schedule_irqoff(&droq->napi);
 	} else {
-		call_single_data_t *csd = &droq->csd;
-
-		csd->func = napi_schedule_wrapper;
-		csd->info = &droq->napi;
-		csd->flags = 0;
-
-		smp_call_function_single_async(droq->cpu_id, csd);
+		INIT_CSD(&droq->csd, napi_schedule_wrapper, &droq->napi);
+		smp_call_function_single_async(droq->cpu_id, &droq->csd);
 	}
 }
 

diff --git a/include/linux/smp.h b/include/linux/smp.h
@@ -21,24 +21,23 @@ typedef bool (*smp_cond_func_t)(int cpu, void *info);
  * structure shares (partial) layout with struct irq_work
  */
 struct __call_single_data {
-	union {
-		struct __call_single_node node;
-		struct {
-			struct llist_node llist;
-			unsigned int flags;
-#ifdef CONFIG_64BIT
-			u16 src, dst;
-#endif
-		};
-	};
+	struct __call_single_node node;
 	smp_call_func_t func;
 	void *info;
 };
 
+#define CSD_INIT(_func, _info) \
+	(struct __call_single_data){ .func = (_func), .info = (_info), }
+
 /* Use __aligned() to avoid to use 2 cache lines for 1 csd */
 typedef struct __call_single_data call_single_data_t
 	__aligned(sizeof(struct __call_single_data));
 
+#define INIT_CSD(_csd, _func, _info)		\
+do {						\
+	*(_csd) = CSD_INIT((_func), (_info));	\
+} while (0)
+
 /*
  * Enqueue a llist_node on the call_single_queue; be very careful, read
  * flush_smp_call_function_queue() in detail.

diff --git a/kernel/debug/debug_core.c b/kernel/debug/debug_core.c
@@ -225,8 +225,6 @@ NOKPROBE_SYMBOL(kgdb_skipexception);
  * Default (weak) implementation for kgdb_roundup_cpus
  */
 
-static DEFINE_PER_CPU(call_single_data_t, kgdb_roundup_csd);
-
 void __weak kgdb_call_nmi_hook(void *ignored)
 {
 	/*
@@ -241,6 +239,9 @@ void __weak kgdb_call_nmi_hook(void *ignored)
 }
 NOKPROBE_SYMBOL(kgdb_call_nmi_hook);
 
+static DEFINE_PER_CPU(call_single_data_t, kgdb_roundup_csd) =
+	CSD_INIT(kgdb_call_nmi_hook, NULL);
+
 void __weak kgdb_roundup_cpus(void)
 {
 	call_single_data_t *csd;
@@ -267,7 +268,6 @@ void __weak kgdb_roundup_cpus(void)
 			continue;
 		kgdb_info[cpu].rounding_up = true;
 
-		csd->func = kgdb_call_nmi_hook;
 		ret = smp_call_function_single_async(cpu, csd);
 		if (ret)
 			kgdb_info[cpu].rounding_up = false;

diff --git a/kernel/sched/core.c b/kernel/sched/core.c
@@ -320,14 +320,6 @@ void update_rq_clock(struct rq *rq)
 	update_rq_clock_task(rq, delta);
 }
 
-static inline void
-rq_csd_init(struct rq *rq, call_single_data_t *csd, smp_call_func_t func)
-{
-	csd->flags = 0;
-	csd->func = func;
-	csd->info = rq;
-}
-
 #ifdef CONFIG_SCHED_HRTICK
 /*
  * Use HR-timers to deliver accurate preemption points.
@@ -428,7 +420,7 @@ void hrtick_start(struct rq *rq, u64 delay)
 static void hrtick_rq_init(struct rq *rq)
 {
 #ifdef CONFIG_SMP
-	rq_csd_init(rq, &rq->hrtick_csd, __hrtick_start);
+	INIT_CSD(&rq->hrtick_csd, __hrtick_start, rq);
 #endif
 	hrtimer_init(&rq->hrtick_timer, CLOCK_MONOTONIC, HRTIMER_MODE_REL_HARD);
 	rq->hrtick_timer.function = hrtick;
@@ -7774,7 +7766,7 @@ void __init sched_init(void)
 		rq->last_blocked_load_update_tick = jiffies;
 		atomic_set(&rq->nohz_flags, 0);
 
-		rq_csd_init(rq, &rq->nohz_csd, nohz_csd_func);
+		INIT_CSD(&rq->nohz_csd, nohz_csd_func, rq);
 #endif
 #ifdef CONFIG_HOTPLUG_CPU
 		rcuwait_init(&rq->hotplug_wait);