D. 因很多 GPIO 3.3V 因此若要控制 5V or 12V 可能需要 NPN or PNP Tranzistor 或使用 solid state relay (SSR)
特別注意 Ic 電流大小 (看 datasheet) 複習電子學(安排電阻值，自己計算負載所需的電流) 電資學院應該要懂一些基本的

e.g., TIP122 NPN 可達 5A 可用, 就不須 SSR

以前我學弟在美國, 有次請他們研發人員設計一電路, 隔一星期, 對方說沒給電路圖無法用 SPICE!

Image Not Showing Possible Reasons

The image file may be corrupted
The server hosting the image is unavailable
The image path is incorrect
The image format is not supported

Learn More →

這課程很花時間, 可以練吐納養氣腹式呼吸

必學！吐納養氣腹式呼吸，啟動身體自癒力！健康2.0 20161224

李鳳山師父平甩功十分鐘練習【梅門官方】

這單元主要是 Linux Embedded System 課程基礎原理

Boot Loader, Linux OS, 系統程式與網路核心運作原理簡介
Linux System Call, VFS, init 運作原理
Embedded Linux 系統運作原理

PART I

General Embedded Linux Systems

Image Not Showing Possible Reasons

The image file may be corrupted
The server hosting the image is unavailable
The image path is incorrect
The image format is not supported

Learn More →

Start up

power on–>hardware (CPU+ROM code (or Flash), Memory, Flash, ..)
ROM code UP and RUN
ROM code –-Choose Available Boot Loader (e.g., U-Boot)

    Boot Loader UP and RUN-->Check and Choose Linux kernel image (e.g., bzimage) and Root File System (rootfs)
           Load image, maybe,  from flash, serial port, network, ...
    Linux kernel image UP (first part)--decompress Linux (main part) to memory
    RUN Linux main part --->
     START all initial codes
      ......
     mount rootfs   / <<<root
     RUN init (in rootfs)  
            init RUN all apps and INSERT all pre-defined kernel modules 
     START shell, ...
     $>

PC 使用的 GRUB (GRand Unified Bootloader):
Details of GRUB on the PC

Linux kernel 啟動部分:
以 Linux 2.4.0 為例:

https://lxr.linux.no/linux-old+v2.4.0/init/main.c

























































































 520asmlinkage void __init start_kernel(void)
 521{
 522        char * command_line;
 523        unsigned long mempages;
 524        extern char saved_command_line[];
 525/*
 526 * Interrupts are still disabled. Do necessary setups, then
 527 * enable them
 528 */
 529        lock_kernel();
 530        printk(linux_banner);
 531        setup_arch(&command_line);
 532        printk("Kernel command line: %s\n", saved_command_line);
 533        parse_options(command_line);
 534        trap_init();
 535        init_IRQ();
 536        sched_init();
 537        time_init();
 538        softirq_init();
 539
 540        /*
 541         * HACK ALERT! This is early. We're enabling the console before
 542         * we've done PCI setups etc, and console_init() must be aware of
 543         * this. But we do want output early, in case something goes wrong.
 544         */
 545        console_init();
 546#ifdef CONFIG_MODULES
 547        init_modules();
 548#endif
 549        if (prof_shift) {
 550                unsigned int size;
 551                /* only text is profiled */
 552                prof_len = (unsigned long) &_etext - (unsigned long) &_stext;
 553                prof_len >>= prof_shift;
 554                
 555                size = prof_len * sizeof(unsigned int) + PAGE_SIZE-1;
 556                prof_buffer = (unsigned int *) alloc_bootmem(size);
 557        }
 558
 559        kmem_cache_init();
 560        sti();
 561        calibrate_delay();
 562#ifdef CONFIG_BLK_DEV_INITRD
 563        if (initrd_start && !initrd_below_start_ok &&
 564                        initrd_start < min_low_pfn << PAGE_SHIFT) {
 565                printk(KERN_CRIT "initrd overwritten (0x%08lx < 0x%08lx) - "
 566                    "disabling it.\n",initrd_start,min_low_pfn << PAGE_SHIFT);
 567                initrd_start = 0;
 568        }
 569#endif
 570        mem_init();
 571        kmem_cache_sizes_init();
 572#ifdef CONFIG_3215_CONSOLE
 573        con3215_activate();
 574#endif
 575#ifdef CONFIG_PROC_FS
 576        proc_root_init();
 577#endif
 578        mempages = num_physpages;
 579
 580        fork_init(mempages);
 581        proc_caches_init();
 582        vfs_caches_init(mempages);
 583        buffer_init(mempages);
 584        page_cache_init(mempages);
 585        kiobuf_setup();
 586        signals_init();
 587        bdev_init();
 588        inode_init(mempages);
 589#if defined(CONFIG_SYSVIPC)
 590        ipc_init();
 591#endif
 592#if defined(CONFIG_QUOTA)
 593        dquot_init_hash();
 594#endif
 595        check_bugs();
 596        printk("POSIX conformance testing by UNIFIX\n");
 597
 598        /* 
 599         *      We count on the initial thread going ok 
 600         *      Like idlers init is an unlocked kernel thread, which will
 601         *      make syscalls (and thus be locked).
 602         */
 603        smp_init();
 604        kernel_thread(init, NULL, CLONE_FS | CLONE_FILES | CLONE_SIGNAL);
 605        unlock_kernel();
 606        current->need_resched = 1;
 607        cpu_idle();
 608}

604 kernel_thread(init, NULL, CLONE_FS | CLONE_FILES | CLONE_SIGNAL);


































static int init(void * unused)
 762{
 763        lock_kernel();
 764        do_basic_setup();
 765
 766        /*
 767         * Ok, we have completed the initial bootup, and
 768         * we're essentially up and running. Get rid of the
 769         * initmem segments and start the user-mode stuff..
 770         */
 771        free_initmem();
 772        unlock_kernel();
 773
 774        if (open("/dev/console", O_RDWR, 0) < 0)
 775                printk("Warning: unable to open an initial console.\n");
 776
 777        (void) dup(0);
 778        (void) dup(0);
 779        
 780        /*
 781         * We try each of these until one succeeds.
 782         *
 783         * The Bourne shell can be used instead of init if we are 
 784         * trying to recover a really broken machine.
 785         */
 786
 787        if (execute_command)
 788                execve(execute_command,argv_init,envp_init);
 789        execve("/sbin/init",argv_init,envp_init);
 790        execve("/etc/init",argv_init,envp_init);
 791        execve("/bin/init",argv_init,envp_init);
 792        execve("/bin/sh",argv_init,envp_init);
 793        panic("No init found.  Try passing init= option to kernel.");
 794}

dup, dup2, dup3 - duplicate a file descriptor
execve - execute program
How to Set Environment Variables in Linux

Futher Reading:
How to read environment variables of a process

In general, "init" is the leading process with pid (process id)=0

You can look the result of pstree:
laikc@laikc-VirtualBox:~/course$ pstree
init─┬─NetworkManager─┬─dhclient
│ ├─dnsmasq
│ └─2*[{NetworkManager}]
├─accounts-daemon───{accounts-daemon}
├─atd
├─bluetoothd
├─console-kit-dae───64*[{console-kit-dae}]
├─cron
├─cupsd
├─2*[dbus-daemon]
├─6*[getty]
├─gvfs-fuse-daemo───3*[{gvfs-fuse-daemo}]
├─gvfsd
├─lightdm─┬─Xorg
│ ├─lightdm─┬─lightdm-greeter───lightdm-gtk-gre───{lightdm-gtk-gre}
│ │ └─{lightdm}
│ ├─lightdm
│ └─2*[{lightdm}]
├─modem-manager
├─named───3*[{named}]
├─ntpd
├─polkitd───{polkitd}
├─rsyslogd───3*[{rsyslogd}]
├─samba───4*[samba]
├─smbd───smbd
├─sshd─┬─sshd───sshd───bash───bash───pstree
│ └─sshd───sshd───bash
├─udevd───2*[udevd]
├─upowerd───2*[{upowerd}]
├─upstart-socket-
└─upstart-udev-br

Linux VFS & System Call

Linux: VFS & System Call

Anatomy of the Linux file system (dentry, inode)

Abstract List (亦可用 Doubly Linked List)

Register Object (register_filesystem) and Use Object

devfs

procfs

Linux v2.4.31 net/socket.c














 114static struct file_operations socket_file_ops = {
 115        llseek:         no_llseek,
 116        read:           sock_read,
 117        write:          sock_write,
 118        poll:           sock_poll,
 119        ioctl:          sock_ioctl,
 120        mmap:           sock_mmap,
 121        open:           sock_no_open,   /* special open code to disallow open via /proc */
 122        release:        sock_close,
 123        fasync:         sock_fasync,
 124        readv:          sock_readv,
 125        writev:         sock_writev,
 126        sendpage:       sock_sendpage
 127};

與 Linux VFS 溝通

socket(AF_INET, SOCK_STREAM, 0)

net/ipv4/af_inet.c






















 945struct proto_ops inet_stream_ops = {
 946        family:         PF_INET,
 947
 948        release:        inet_release,
 949        bind:           inet_bind,
 950        connect:        inet_stream_connect,
 951        socketpair:     sock_no_socketpair,
 952        accept:         inet_accept,
 953        getname:        inet_getname, 
 954        poll:           tcp_poll,
 955        ioctl:          inet_ioctl,
 956        listen:         inet_listen,
 957        shutdown:       inet_shutdown,
 958        setsockopt:     inet_setsockopt,
 959        getsockopt:     inet_getsockopt,
 960        sendmsg:        inet_sendmsg,
 961        recvmsg:        inet_recvmsg,
 962        mmap:           sock_no_mmap,
 963        sendpage:       tcp_sendpage
 964};
 965

GPIO
因同學有來問, 因此增加說明
這對Embedded Linux 設計很重要的基礎
[Day 7]-【STM32系列】淺入淺出之 General Purpose Input/Output 介紹 (上)

GPIO: You can control GPIO from user space
e.g. control /sys/class/gpio/gpio49/value

/sys/class/gpio/gpio49#> ls
active_low direction edge power subsystem uevent value
#> echo out > direction
#> echo 1 > value
#> echo 0 > value
注意優缺點! 有些應用不適宜! 此外 debounce 也須處理!!! Appendix A.

SPI
/dev/spidevxx.yy 其中 xx 可能 0, yy 可能 0,1,.. yy 代表 chip select (哪一個 SPI)
user space
fd=open("/dev/spidevxx.yy ", …
ioctl (fd, …chip mode…)
…
read/write /ioctl
close

對一般的應用可更進一步參考:
http://www.raspberry-projects.com/pi/programming-in-c/spi/using-the-spi-interface

其中 int SpiWriteAndRead 是重要的做法,還有 memset(&spi[i], 0, sizeof (spi[i])); 也類似我於課程中提的問題! 小心! 很多人忘記這個動作!

用 ioctl 來取代 read/write 有內部設計的原因 , I2C, USB 都有類似的用法

想像

   user app<<<==================>>> SPI device

SPI interface (CS, CLK, MISO, MOSI) bit frame <–-> device

   client   <<<  Network  programming   >>> server

兩者觀念一樣!! 也是

                        螺絲<-----螺紋---->螺帽

的關係

特別提醒
如果使用 GPIO, SPI 等裝置要注意 TIMING 的問題, 若使用 USER SPACE 可能有精準度的問題!

Benchmarking Raspberry Pi GPIO Speed

有關SPI, I²C, NAND Flash, NOR Flash, …可參考

Essential Linux Device Drivers, by Sreekrishnan Venkateswaran

I2C Introduction – Part 1 (Basics)
I2C Protocol Introduction – Part 2 (Advanced Topics)

We will focus on socket layer in this course (network programming).

Remark.

For example, "printf" is not a system call. Try run "strace your_hello_program" , you find it's "write" not printf!

Moreover, fopen, fprintf, strcmp, strcpy, … are not system calls. They are functions in dynamic lib libc…so.

Try "ldd your_hello_program" ! You can print the shared objects (shared libraries) ! This is an important method to build your new root file system with .shared libraries. See example "man ldd".

// user name is laikc
vdso (virtual dynamic shared object, see man 7 vdso)
Why does the vDSO exist at all? There are some system calls the
kernel provides that user-space code ends up using frequently, to
the point that such calls can dominate overall performance. This
is due both to the frequency of the call as well as the context-
switch overhead that results from exiting user space and entering the kernel.

In x86 32-bit systems, you can
trigger a software interrupt (int $0x80) to tell the kernel you
wish to make a system call.

e.g., gettimeofday

laikc@laikc-VirtualBox:/tmp/demo/laikc/course$ ldd ./ex1_1
linux-vdso.so.1 => (0x00007fff31d0d000)
libc.so.6 => /lib/x86_64-linux-gnu/libc.so.6 (0x00007ff323de0000)
/lib64/ld-linux-x86-64.so.2 (0x00007ff3241b5000)
laikc@laikc-VirtualBox:/tmp/demo/laikc/course$

Performance issue: "copy_to_ user" & "copy_from_ user"

#include <unistd.h>
ssize_t read(int fd, void *buf, size_t count);

e.g.,

User space <–> Kernel space
read(fd, buf, count)/recv <––––> read(fd, buf,…)/recv
<<<<–-copy_to_ user

write(fd, buf,count)/send <––––-> write(fd, buf, …)/send
copy from user–->>>>>

Virtual Device:

Then we can use the remote devices. For example, mouse, keyboard, sound, memory, disk, …display…

Network programming is an important skill!"

This is the operation principle of network thin client!
You can, also, boot Windows OS via PXE.

願意再學請看以下:

作者 Greg Kroah-Hartman 提供

Linux Kernel in a Nutshell
This is the web site for the book, Linux Kernel in a Nutshell, by Greg Kroah-Hartman, published by O'Reilly.

License
This book is available under the terms of the Creative Commons Attribution-ShareAlike 2.5 license. That means that you are free to download and redistribute it. The development of the book was made possible, however, by those who purchase a copy from O'Reilly or elsewhere.

http://www.kroah.com/lkn/

請需要的讀者參考

第一步:

對於一般學生可以進一步學習 x86 Linux

利用載入 Linux kernel source code
重新編譯並可開重新多重開機, 網路有很多善心人士提供許多資料, 我就不寫了.
小心 kernel : 重新編譯重要事項:
(1) make menuconfig 與你CPU 有關, 因此 Google 時請加 VM VirtualBox 或你 VMware 的名稱
最好用 VM 來測試你的 kernel, 才不會麻煩
(2) 不要用太新的版本, 學習用舊的即可, 尤其x86 很雜, 安全性很麻煩, 因練習用就儘可不用, 日後正式用途需要考慮!

如果用量不是很大, Embedded Linux Board也夠用! 省電安全!
樹莓派或 TI BeagleBoard 夠一般 web server, 一般控制, NAS, … 等用了!
(但要 DIY, 請把本課程學好, 才做的好)

:::

後面給進階的使用者

Appendix: System Call

Kernel Part: Depend on kernel version and arch
User Part: Depend on gcc version
Basic file: Check unistd.h, syscall.h, …
For example,

#define __NR_exit 1
#define __NR_fork 2
#define __NR_read 3
#define __NR_write 4
#define __NR_open 5
#define __NR_close 6
#define __NR_waitpid 7

…

Important command strace:

strace: trace system calls and signals

In general, user space: read(…) –>[sys call](3, …) –>kernel [sys call] (__NR_read …) –…

"strace" command help u to see your running system calls and using files, e.g.,"#> strace ./ex1"
If your network program has "gethostname (or gethostname2)", try to see using files.

Be careful, "gethostname" is not thread-safe. If you need not this function, don't use it.

You can add your new system call to kernel source (depends on kernel version…)

Remarks.
I am sorry, I can not give u a unique answer.
If you need this future, please see your kernel source.
In practice, you need not this future. This is a good excise for understanding system calls.

Advanced Users

I2C Introduction – Part 1 (Basics)

Appendix: Build a new Linux kernel and Rootfs
x86:
please google!

other ARCH: (Linux architecture)
step 1. toolchain
step 2. linux kernel
step 3. root file system (you can use busybox)
step 4. create your init

usual command (mount, umount, dd, losetup, mke2fs, ..)

mount:

如何mount 特殊 file system 請 google 查詢

dd:

有用的command, 必需小心的危險command, 用錯會哭哭!!
請看 https://en.wikipedia.org/wiki/Dd_(Unix)
例如
dd if=/dev/zero bs=1024 count=1000000 of=file_1GB

複製至 of=/tmp/zero_copy

dd if=file_1GB of=/tmp/zero_copy

dd if=/dev/zero of=/tmp/rootimg bs=1k count=2000 通常用來產生 2MB rootimg 但原因是…很多人不知道!

事實上很多人 rootfs 愈來愈大就是沒真正了解學過知識的重點! 其實就是~~全為0 的檔案, 再複製你真正的檔案, 這要 gzip 才壓縮更多!! 否則一堆檔案垃圾當壓縮檔愈來愈大

Now making it look like a block device instead of just a regular file (See man 8 losetup)

$ losetup /dev/loop0 rootimg

Creating an ext2 file system with the loop device

$ mke2fs -c /dev/loop0

You can mount in /mnt/rootfs (mkdir /mnt/rootfs)

$ mount -t ext2 /dev/loop0 /mnt/rootfs

You can use /mnt/rootfs as your new root file system

prepare your files…dirty work…
…
…

Finally, unmount and detach

$ umount /dev/loop0
$ losetup -d /dev/loop0

ARM Arch!

一步步教你：如何用Qemu來模擬ARM系統

Using VM:
Lubuntu 12.04: Compile QEMU 2.0.0
Create USB linux (using syslinux):
還有 4M Linux

Use QEMU:
QEMU (ARM Arch) (含 toolchain, kernel, rootfs 產生) (若初級使用請 google, 有很多!)

Special File System: device files system, sys file system, proc file system

device files system

$:> ls -al /dev
major minor
total 724
…
lrwxrwxrwx 1 root root 9 Oct 14 22:51 cdrom -> /dev/scd1
lrwxrwxrwx 1 root root 9 Oct 14 22:52 cdrom1 -> /dev/scd0
crw–––- 1 root tty 5, 1 Jan 19 20:47 console

crw–w–– 1 root video 29, 1 Mar 15 2002 fb0autodetect
…
brw-rw–– 1 root disk 3, 64 Mar 15 2002 hdb
brw-rw–– 1 root disk 22, 0 Mar 15 2002 hdc
brw-rw–– 1 root disk 22, 64 Mar 15 2002 hdd
…

crw-r––- 1 root kmem 1, 1 Sep 28 18:06 mem

crw-rw-rw- 1 root root 1, 3 Sep 28 18:06 null
…
brw-rw–– 1 root disk 8, 0 Mar 15 2002 sda //major number=8, minor number=0
brw-rw–– 1 root disk 8, 1 Mar 15 2002 sda1
brw-rw–– 1 root disk 8, 2 Mar 15 2002 sda2
brw-rw–– 1 root disk 8, 3 Mar 15 2002 sda3
brw-rw–– 1 root disk 8, 4 Mar 15 2002 sda4
brw-rw–– 1 root disk 8, 16 Mar 15 2002 sdb
brw-rw–– 1 root disk 8, 17 Mar 15 2002 sdb1
brw-rw–– 1 root disk 8, 18 Mar 15 2002 sdb2
brw-rw–– 1 root disk 8, 19 Mar 15 2002 sdb3
brw-rw–– 1 root disk 8, 20 Mar 15 2002 sdb4
…
lrwxrwxrwx 1 root root 4 Sep 28 18:05 stderr -> fd/2
lrwxrwxrwx 1 root root 4 Sep 28 18:05 stdin -> fd/0
lrwxrwxrwx 1 root root 4 Sep 28 18:05 stdout -> fd/1
crw-rw-rw- 1 root tty 5, 0 Sep 28 18:06 tty
crw–––- 1 root root 4, 0 Sep 28 18:06 tty0
crw–––- 1 root root 4, 1 Jan 19 14:59 tty1
…
crw-rw–– 1 root dialout 4, 64 Mar 15 2002 ttyS0
crw-rw–– 1 root dialout 4, 65 Mar 15 2002 ttyS1
crw-rw–– 1 root dialout 4, 66 Mar 15 2002 ttyS2
crw-rw–– 1 root dialout 4, 67 Mar 15 2002 ttyS3
crw-rw–– 1 root dialout 188, 0 Mar 15 2002 ttyUSB0 //不一定在這裡
crw-rw–– 1 root dialout 188, 1 Mar 15 2002 ttyUSB1
cr–r–r– 1 root root 1, 9 Jan 19 20:46 urandom
drwxr-xr-x 2 root root 4096 Sep 28 18:05 usb
…
crw-rw-rw- 1 root root 1, 5 Sep 28 18:06 zero

//下面這個是自己寫的 device (假設的)
crw-rw-rw- 1 root root 100, 0 Sep 28 18:06 mydev

crw-rw–– c 表示 char device
b???… b 表示 block device

//問題來了
(1) /dev/mydev 如何產生 ? major number=100, minor number=0 如何產生?
(2) user space 如何使用?

(1)

引用寫的很好的 Writing a Linux Kernel Module — Part 2: A Character Device
的file_operations 資料:















































struct file_operations {

   struct module *owner;                             // Pointer to the LKM that owns the structure

   loff_t (*llseek) (struct file *, loff_t, int);    // Change current read/write position in a file

   ssize_t (*read) (struct file *, char __user *, size_t, loff_t *);    // Used to retrieve data from the device

   ssize_t (*write) (struct file *, const char __user *, size_t, loff_t *);   // Used to send data to the device

   ssize_t (*aio_read) (struct kiocb *, const struct iovec *, unsigned long, loff_t);  // Asynchronous read

   ssize_t (*aio_write) (struct kiocb *, const struct iovec *, unsigned long, loff_t); // Asynchronous write

   ssize_t (*read_iter) (struct kiocb *, struct iov_iter *);            // possibly asynchronous read

   ssize_t (*write_iter) (struct kiocb *, struct iov_iter *);           // possibly asynchronous write

   int (*iterate) (struct file *, struct dir_context *);                // called when VFS needs to read the directory contents

   unsigned int (*poll) (struct file *, struct poll_table_struct *);    // Does a read or write block?

   long (*unlocked_ioctl) (struct file *, unsigned int, unsigned long); // Called by the ioctl system call

   long (*compat_ioctl) (struct file *, unsigned int, unsigned long);   // Called by the ioctl system call

   int (*mmap) (struct file *, struct vm_area_struct *);                // Called by mmap system call

   int (*mremap)(struct file *, struct vm_area_struct *);               // Called by memory remap system call 

   int (*open) (struct inode *, struct file *);             // first operation performed on a device file

   int (*flush) (struct file *, fl_owner_t id);             // called when a process closes its copy of the descriptor

   int (*release) (struct inode *, struct file *);          // called when a file structure is being released

   int (*fsync) (struct file *, loff_t, loff_t, int datasync);  // notify device of change in its FASYNC flag

   int (*aio_fsync) (struct kiocb *, int datasync);         // synchronous notify device of change in its FASYNC flag

   int (*fasync) (int, struct file *, int);                 // asynchronous notify device of change in its FASYNC flag

   int (*lock) (struct file *, int, struct file_lock *);    // used to implement file locking

   …

};

基本想法: Object Model (簡化說明)

(a)
Write a kernel module mydev.c ->gcc (You Need Makefile, Kernel Source and strip)->mydev.ko

kernel module mydev.c (完整寫法…可參考http://derekmolloy.ie/writing-a-linux-kernel-module-part-2-a-character-device/ , 他的好例子複雜一點, register_chrdev(100,"mydev",&myfp); 可以取代register_chrdev(0, DEVICE_NAME, &fops);但必須確保 major 100 可用!)

一般而言, 作者 register_chrdev(0, DEVICE_NAME, &fops) 是比較好的方式~ 這裡要讓讀者練習與了解 mknod

(a.1) write myfp ===>file_operations  // 模擬檔案操作

(a.2) you must register a char device

    register_chrdev(100,"mydev",&myfp);

#!> insmod mydev
/* become a virtual char device /
#!> mknod /dev/mydev c 100 0
/ make a device node for user space app*/

Then you can write your app hellodev.c













user space (open, read, write, close, ioctl, ...) <===> kernel space

                                                                     

fd= open ("/dev/mydev",...)    <===> file_operations open

ioctl(fd,        <===>file_operations ioctl (unlocked_ioctl or compat_ioctl)
      
read(fd,...)/write       <===> file_operations read /write

close(fd)                <===> file_operations read /release

select(fd, ...           <===> file_operations  poll

Appendix A.

虛擬程式: square wave

    loop:
        echo 1 > /sys/class/gpio/gpio49/value

           sleep_msec 1
           echo  0 > /sys/class/gpio/gpio49/value
           sleep_msec 1
      goto loop

會如何? 請思考可能的問題! 我上課會說! 事實上這虛擬程式有限制

mmap:

https://jasonblog.github.io/note/linux_driver/mmap_driver_implementation.html

Linux的二號功臣 Alan Cox

How to Write Linux Mouse Drivers, by Alan Cox

Kernel Korner - Sleeping in the Kernel

Embedded Linux System, Boot, Linux VFS, System Call, SPI, GPIO, I²C, … 課程基礎原理

tags: embedded linux overview VFS

這單元主要是 Linux Embedded System 課程基礎原理

PART I

Multiplexing and Resource Sharing

Linux: VFS & System Call

Abstract List (亦可用 Doubly Linked List)

Register Object (register_filesystem) and Use Object

usual command (mount, umount, dd, losetup, mke2fs, ..)

prepare your files…dirty work… … …

Finally, unmount and detach

ARM Arch!

Special File System: device files system, sys file system, proc file system

Read more

Metric space, vector space, inner product space

\mathcal{Probability \ Measure \ and \ Conditional \ Expectation}

使用 pthread 的 echo server (有bug! why?)

QEMU linux kernel and rootfs (busybox-1.22.0 and linux-4.4.50)

tags: `embedded` `linux` `overview` `VFS`

prepare your files…dirty work…
…
…