Skip to content

Latest commit

 

History

History
185 lines (147 loc) · 6.96 KB

linux-kernel-dev-reading-note.org

File metadata and controls

185 lines (147 loc) · 6.96 KB
Raw

Linux kernel development重读笔记

ch02 overview

编译kernel:

make menuconfig
make defconfig
# 每次编译之前必敲：
make oldconfig
make时只想关注错误信息
make > /dev/null

ch03进程管理

OS提供两个抽象:

A virtualized processor and virtual memory. The virtual processor gives the process the illusion that is alone monopolizes the system, despite possibly sharing the processor among hundreds of other processes. Virtual memory let the process allocate and manage memory as if it alone owned all the momory in the system.

需要注意的是Linux的threads共享share memory的抽象，但是每个线程都各自拥有虚拟化的processor。

进程始于fork()终于exit()。

kernel stack的大小：一般是page的2倍。即stack在32位上为8KB，64位上为16KB

Unix 特点：

  • simple， system call非常少
  • everything is a file
  • written in C
  • fast process creation time
  • robust IPC

build kernel

make defconfig # defaults for your architecture

kernel characteristics

  • The kernel has access to neither the C library nor the standard C headers.
  • The kernel is coded in GNU C.
  • The kernel lacks the memory protection afforded to user-space.
  • The kernel cannot easily execute floating-point operations.
  • The kernel has a small per-process fixed-size stack.
  • Because the kernel has asynchronous interrupts, is preemptive, and supports SMP,
  • synchronization and concurrency are major concerns within the kernel.
  • Portability is important.

process

Each thread includes a unique program counter, process stack, and set of processor registers.

调度对象： The kernel schedules individual threads, not processes.

current宏得到当前进程的kernel空间，这个结构体中保存了指向task_struct的指针。

static __always_inline struct task_struct *get_current(void)
{
  return this_cpu_read_stable(current_task);
}

linkage什么意思？

首先作为linux操作系统,它不一定就只运行在X86平台下面,还有其他平台例如ARM,PPC， 达芬奇等等，所以在不同的处理器结构上不能保证都是通过 局部栈传递参数的，可能此时就有朋 友就会问：不放在栈中能放在哪里呢？熟悉ARM的朋友一定知道ARM对函数调用过程中的传参定义 了一套规则，叫 ATPCS（内地叫AAPCS），规则中明确指出ARM中R0-R4都是作为通用寄存器使 用，在函数调用时处理器从R0-R4中获取参数，在函数返回时再 将需要返回的参数一次存到R0- R4中，也就是说可以将函数参数直接存放在寄存器中，所以为了严格区别函数参数的存放位置， 引入了两个标记，即 asmlinkage和FASTCALL，前者表示将函数参数存放在局部栈中,后者则 是通知编译器将函数参数用寄存器保存起来

调度

调度的策略决定总体上OS的感觉：

The scheduler policy in Unix systems tends to explicitly favor I/O-bound processes, thus providing good process response time. Linux, aiming to provide good interactive response and desktop performance, optimizes for process response (low latency), thus favoring I/O-bound processes over processor-bound processors.As we will see, this is done in a creative manner that does not neglect processor-bound processes.

进程有两个不交叉的属性，nice值和实时优先级，nice值越高优先级越低，而实时优先级相反， Linux implements real-time priorities in accordance with the relevant Unix standards, specifically POSIX.1b.

sched chain:

#define SCHED_NORMAL   0
#define SCHED_FIFO    1 #rt
#define SCHED_RR    2 #rt
#define SCHED_BATCH   3
/* SCHED_ISO: reserved but not implemented yet */
#define SCHED_IDLE    5
#define SCHED_DEADLINE    6

四种类型的调度器：

  1. stop_sched_class: The stop_sched_class is to stop cpu, using on SMP system, for

load balancing and cpu hotplug. This class have the highest scheduling priority.

If your system does not define CONFIG_SMP, you can try to remove this class, there are several files need to be changed for successful compilation.

  1. rt_sched_class
  2. fair_sched_class
  3. idle_sched_class

    chain together

    #define sched_class_highest (&stop_sched_class)
#define for_each_class(class) \
     for (class = sched_class_highest; class; class = class->next)

extern const struct sched_class stop_sched_class;
extern const struct sched_class dl_sched_class;
extern const struct sched_class rt_sched_class;
extern const struct sched_class fair_sched_class;
extern const struct sched_class idle_sched_class;

    优先级

    #define MAX_NICE  19
#define MIN_NICE  -20
#define NICE_WIDTH  (MAX_NICE - MIN_NICE + 1)

/*
* Priority of a process goes from 0..MAX_PRIO-1, valid RT
* priority is 0..MAX_RT_PRIO-1, and SCHED_NORMAL/SCHED_BATCH
* tasks are in the range MAX_RT_PRIO..MAX_PRIO-1. Priority
* values are inverted: lower p->prio value means higher priority.
*
* The MAX_USER_RT_PRIO value allows the actual maximum
* RT priority to be separate from the value exported to
* user-space.  This allows kernel threads to set their
* priority to a value higher than any user task. Note:
* MAX_RT_PRIO must not be smaller than MAX_USER_RT_PRIO.
*/

#define MAX_USER_RT_PRIO  100
#define MAX_RT_PRIO   MAX_USER_RT_PRIO

#define MAX_PRIO    (MAX_RT_PRIO + NICE_WIDTH)
#define DEFAULT_PRIO    (MAX_RT_PRIO + NICE_WIDTH / 2)

ch05: system calls

The existence of these interfaces, and the fact that applications are not free to directly do whatever they want, is key to providing a stable system.

Syscall provide:

  1. First, it provides an abstracted hardware interface for userspace.
  2. System calls ensure system security and stability.
  3. A single common layer between user-space and the rest of the

system allows for the virtualized system provided to processe.

Provide mechanism, not policy.

The process does not refer to the syscall by name.

current syscall count

#define NR_syscalls 327 /* sizeof(syscalls_64) # */

ch06: interrupt

Interrupt context is time-critical because the interrupt handler interrupts other code. Code should be quick and simple.

Every process on the system previously needed two pages of contiguous, nonswappable kernel memory.

Perhaps not surprising, the implementation of the interrupt handling system in Linux is architecture-dependent.

ch08: sync

Ensuring that unsafe concurrency is prevented and that race conditions do not occur is called synchronization.

The kernel provides a set of interfaces that implement these atomic instructions

  • locking: Threads hold locks locks protect data.