IPI 通信（SMP）【轉】

轉自：https://winddoing.github.io/post/60164.html

IPI (Interrupt-Procecesorr Interrupt): 處理中間的中斷

主要應用是一個處理器讓另一個處理器做特定的事情（function 和 sched）

 

              +---------------------------+-+
system boot | request_percpu_irq（） +
 | mailbox irq handle +
 +--+-----------------------+--+
 | |
 | |
 +--v--+ +---v-+
 | CPU0| > CPU1|
 +--+--+ +----++
 | |
 +----+----+ +-----+-----+
 |mailbox0 | +-->mailbox1 |
 +---------+ | +-----------+
 |
 +---------------+ |
system run A send IPI CPU1 |
 write mailbox1 |
 | |
 | +-----v-----------------+---+
 +------+----+ | 1. 讀取mailbox中的action +
 | Task A | | 2. 通過action判斷IPI類型 +
 | | | 3. 進行function和sched處理+
 +-----------+ | +
 +---------------------------+

在多核處理器中，每一個 CPU 核有一個 mailbox(相當於郵箱)，如果需要進行 IPI 通信時，其主要通過 IPI 的中斷實現。假設 CPU0 需要給 CPU1 發送一個 action(actionI 的類型：SMP_CALL_FUNCTION,SMP_RESCHEDULE_YOURSELF 等) 時，只需要 CPU0 向 CPU1 的 mailbox 中寫於 action 的 id（相當於信），此時 CPU1 將產生一個 IPI 中斷（表明收到信），mailbox 的中斷處理程序將讀取 mailbox（相當於看信）中的 action，判斷 action 的類型進行相應的處理。

MIPS 架構下的 IPI 通信

1. 關閉中斷后還會發送 IPI

MIPS 接口

 

struct plat_smp_ops {
 void (*send_ipi_single)(int cpu, unsigned int action);
 void (*send_ipi_mask)(const struct cpumask *mask, unsigned int action);
 ...
}

IPI 通信就是多個處理器之間的交流。
send_ipi_single： 一對一聊天
send_ipi_mask : 群發，mask 表示群發的成員（CPU）

action 類型

 

/* Octeon - Tell another core to flush its icache */

/* Used by kexec crashdump to save all cpu's state */

file: arch/mips/include/asm/smp.h

1. 不同的 action (活動) 何時將產生？
2. 各自都有什么作用？

SMP_RESCHEDULE_YOURSELF

SMP_RESCHEDULE_YOURSELF 將直接調用 scheduler_ipi. 將任務插入目標 CPU 的運行隊列。

 

/*
 * this function sends a 'reschedule' IPI to another CPU.
 * it goes straight through and wastes no time serializing
 * anything. Worst case is that we lose a reschedule ...
 */
static inline void smp_send_reschedule(int cpu)
{
 extern struct plat_smp_ops *mp_ops; /* private */

 mp_ops->send_ipi_single(cpu, SMP_RESCHEDULE_YOURSELF);
}

file: arch/mips/include/asm/smp.h

SMP_CALL_FUNCTION

SMP_CALL_FUNCTION: 將特定的函數在目標 CPU 上運行

• 內核回調接口：
   

static inline void arch_send_call_function_single_ipi(int cpu)
{
 extern struct plat_smp_ops *mp_ops; /* private */

 mp_ops->send_ipi_mask(&cpumask_of_cpu(cpu), SMP_CALL_FUNCTION);
}

static inline void arch_send_call_function_ipi_mask(const struct cpumask *mask)
{
 extern struct plat_smp_ops *mp_ops; /* private */

 mp_ops->send_ipi_mask(mask, SMP_CALL_FUNCTION);
}

file: arch/mips/include/asm/smp.h

 

/*
 * smp_call_function_single - Run a function on a specific CPU
 * @func: The function to run. This must be fast and non-blocking.
 * @info: An arbitrary pointer to pass to the function.
 * @wait: If true, wait until function has completed on other CPUs.
 *
 * Returns 0 on success, else a negative status code.
 */

smp_call_function_single
 \->generic_exec_single
 \->arch_send_call_function_single_ipi

/**
 * smp_call_function_many(): Run a function on a set of other CPUs.
 * @mask: The set of cpus to run on (only runs on online subset).
 * @func: The function to run. This must be fast and non-blocking.
 * @info: An arbitrary pointer to pass to the function.
 * @wait: If true, wait (atomically) until function has completed
 * on other CPUs.
 *
 * If @wait is true, then returns once @func has returned.
 *
 * You must not call this function with disabled interrupts or from a
 * hardware interrupt handler or from a bottom half handler. Preemption
 * must be disabled when calling this function.
 */

smp_call_function_many
 \->arch_send_call_function_ipi_mask

file: kernel/smp.c

刷新 TLB

多核進行 TLB 的同步？