2020年02月07日 情報科学類 オペレーティングシステム II 筑波大学 システム情報系 新城 靖 <yas@cs.tsukuba.ac.jp>
このページは、次の URL にあります。
http://www.coins.tsukuba.ac.jp/~yas/coins/os2-2019/2020-02-07
あるいは、次のページから手繰っていくこともできます。
http://www.coins.tsukuba.ac.jp/~yas/
http://www.cs.tsukuba.ac.jp/~yas/
試験について
Linux では、割り込みの処理を2つに分ける。
図? 割り込み処理の前半部分と後半部分
割り込みハンドラ(前半部)と後半部の役割分担の目安。
注意1: Tasklet は、task 構造体とはまったく関係ない。名前がよくない。
注意2: Softirq という用語を、割り込み処理の後半部という意味で使う人もい る。
注意3: 伝統的なUnixでは、top half は、システム・コールから派生する上位 層の処理、bottom half は、割り込みから派生する下位層の処理の意味で使わ れることがある。Linux では、top half, bottom half は、割り込み処理の前 半部分と後半部分の意味に使う。
Tasklet で1つの仕事は次のような、struct tasklet_struct で表現される。
linux-5.4.7/include/linux/interrupt.h 592: struct tasklet_struct 593: { 594: struct tasklet_struct *next; 595: unsigned long state; 596: atomic_t count; 597: void (*func)(unsigned long); 598: unsigned long data; 599: };
図? Taskletにおける仕事のキュー
DECLARE_TASKLET(name, func, data) 有効な(count==0) の struct tasklet_struct を宣言する DECLARE_TASKLET_DISABLED(name, func, data) 無効な(count==1) の struct tasklet_struct を宣言する
void tasklet_init(struct tasklet_struct *t, void (*func)(unsigned long), unsigned long data);その他に、生成消滅有効無効に関して次のような操作がある。
void tasklet_handler(unsigned long data) { ... }
void tasklet_schedule(struct tasklet_struct *t) Tasklet t を通常の優先度でスケジュールする void tasklet_hi_schedule(struct tasklet_struct *t) Tasklet t を高優先度でスケジュールするすると、それは「そのうちに」1度だけ実行される。
linux-5.4.7/drivers/net/wireless/ath/ath9k/ath9k.h 982: struct ath_softc { ... 990: struct tasklet_struct intr_tq; 991: struct tasklet_struct bcon_tasklet; ... 1077: }; linux-5.4.7/drivers/net/wireless/ath/ath9k/init.c 651: static int ath9k_init_softc(u16 devid, struct ath_softc *sc, 652: const struct ath_bus_ops *bus_ops) 653: { ... 731: tasklet_init(&sc->intr_tq, ath9k_tasklet, (unsigned long)sc); 732: tasklet_init(&sc->bcon_tasklet, ath9k_beacon_tasklet, 733: (unsigned long)sc); ... 791: }
linux-5.4.7/drivers/net/wireless/ath/ath9k/main.c 481: irqreturn_t ath_isr(int irq, void *dev) 482: { ... 499: struct ath_softc *sc = dev; ... 504: bool sched = false; ... 524: ath9k_hw_getisr(ah, &status, &sync_cause); /* NB: clears ISR too */ ... 543: if (status & SCHED_INTR) 544: sched = true; ... 557: if (status & ATH9K_INT_SWBA) 558: tasklet_schedule(&sc->bcon_tasklet); ... 585: if (sched) { 586: /* turn off every interrupt */ 587: ath9k_hw_kill_interrupts(ah); 588: tasklet_schedule(&sc->intr_tq); 589: } 590: 591: return IRQ_HANDLED; ... 594: } linux-5.4.7/drivers/net/wireless/ath/ath9k/main.c 371: void ath9k_tasklet(unsigned long data) 372: { ... 479: } linux-5.4.7/drivers/net/wireless/ath/ath9k/beacon.c 388: void ath9k_beacon_tasklet(unsigned long data) 389: { ... 508: }
図? Work Queueにおける仕事のキュー
キューにつながれる仕事は、Tasklet の仕事とほとんど同じで、関数へのポイ ンタ func と data からなる。処理の主体が、ワーカ・スレッドと呼ばれるカー ネル・レベルのスレッドである所が違う。
汎用の Work Queue デフォルトのワーカ・スレッドは、kworker/n (nはプロセッ サ番号) とよばれ、プロセッサごとに作られる。1つのスレッドで、様々な要 求元の仕事をこなす。下の例では、1つのプロセッサに5個のスレッドが 作られている。そのうち2つは、nice 値が -20 で高優先度。
$ ps alx|grep worker|wc
22 289 1881
$ ps alx|grep 'worker.*/0'
1 0 4779 2 20 0 0 0 worker S ? 0:00 [kworker/0:2]
1 0 5276 2 20 0 0 0 worker S ? 0:00 [kworker/0:1]
1 0 5479 2 20 0 0 0 worker S ? 0:00 [kworker/0:0]
5 0 12906 2 0 -20 0 0 worker S< ? 0:59 [kworker/0:1H]
5 0 30659 2 0 -20 0 0 worker S< ? 0:07 [kworker/0:0H]
0 1013 5803 5611 20 0 117076 1016 pipe_w S+ pts/2 0:00 grep --color=auto worker.*/0
$
汎用の Work Queue のワーカ・スレッドの他に、専用のワーカ・スレッドを作
ることもできる。
linux-5.4.7/include/linux/workqueue.h 21: typedef void (*work_func_t)(struct work_struct *work); 102: struct work_struct { 103: atomic_long_t data; 104: struct list_head entry; 105: work_func_t func; ... 109: };
struct work_struct my_work; ... INIT_WORK(&my_work,my_work_handler);
void my_work_handler(struct work_struct *work) { ... }
schedule_work(&work);この結果、INIT_WORK() で設定したハンドラがワーカ・スレッドにより「その うち」に呼び出される。
schedule_work() では、即座に実行される可能性もある。少し後に実行したい (間を取りたい)時には、次の関数を呼ぶ。
schedule_delayed_work(&work,ticks);ticks は、どのくらい間をとるか。単位は、 ticks (jiffiesの単位)。 多くのシステムで10ミリ秒-1ミリ秒で、設定によって異なる。
解決策:
図? 層構造を用いたファイル・システムの実装
解決策
$ ls -l /usr/bin/perl{,5.10.1}
-rwxr-xr-x. 2 root root 13304 Mar 22 2017 /usr/bin/perl
-rwxr-xr-x. 2 root root 13304 Mar 22 2017 /usr/bin/perl5.10.1
$ ls -li /usr/bin/perl{,5.10.1}
1846686 -rwxr-xr-x. 2 root root 13304 Mar 22 2017 /usr/bin/perl
1846686 -rwxr-xr-x. 2 root root 13304 Mar 22 2017 /usr/bin/perl5.10.1
$
$ grep -v '#' /etc/fstab
UUID=9cfbc67e-781c-48d1-8303-1dde8ce87ee9 / ext4 defaults 1 1
UUID=bab1faf1-5f5b-4a2a-b24f-e850a2b0b82d /boot ext4 defaults 1 2
UUID=a1f61ff2-2c99-4c54-8c3e-2178eed3ec10 swap swap defaults 0 0
tmpfs /dev/shm tmpfs defaults 0 0
devpts /dev/pts devpts gid=5,mode=620 0 0
sysfs /sys sysfs defaults 0 0
proc /proc proc defaults 0 0
pentas-fs:/vol0/home /home nfs rw,hard,bg,nfsvers=3,intr 0 0
pentas-fs:/vol0/web /var/www nfs rw,hard,bg,nfsvers=3,intr 0 0
pentas-fs:/vol0/local3 /usr/local3 nfs rw,hard,bg,nfsvers=3,intr 0 0
$ df /
Filesystem 1K-blocks Used Available Use% Mounted on
/dev/sda3 49071944 6721604 39857568 15% /
$ blkid /dev/sda3
/dev/sda3: UUID="9cfbc67e-781c-48d1-8303-1dde8ce87ee9" TYPE="ext4"
$ ls -l /dev/sda3
brw-rw----. 1 root disk 8, 3 Feb 2 10:50 /dev/sda3
$ lsblk
NAME MAJ:MIN RM SIZE RO TYPE MOUNTPOINT
sda 8:0 0 50G 0 disk
|-sda1 8:1 0 512M 0 part /boot
|-sda2 8:2 0 2G 0 part [SWAP]
`-sda3 8:3 0 47.6G 0 part /
sr0 11:0 1 1024M 0 rom
$ ls -l /dev/sda
brw-rw----. 1 root disk 8, 0 Feb 2 10:50 /dev/sda
$
$ grep cd /etc/auto.misc
cd -fstype=iso9660,ro,nosuid,nodev :/dev/cdrom
$
STAT(2) Linux Programmer's Manual STAT(2) ... int stat(const char *path, struct stat *buf); ... struct stat { dev_t st_dev; /* ID of device containing file */ ino_t st_ino; /* inode number */ mode_t st_mode; /* protection */ nlink_t st_nlink; /* number of hard links */ uid_t st_uid; /* user ID of owner */ gid_t st_gid; /* group ID of owner */ dev_t st_rdev; /* device ID (if special file) */ off_t st_size; /* total size, in bytes */ blksize_t st_blksize; /* blocksize for filesystem I/O */ blkcnt_t st_blocks; /* number of blocks allocated */ time_t st_atime; /* time of last access */ time_t st_mtime; /* time of last modification */ time_t st_ctime; /* time of last status change */ };stat コマンドを使うと stat システム・コールで返される値に近いものが表示 される。
$ stat .bashrc
File: `.bashrc'
Size: 240 Blocks: 16 IO Block: 65536 regular file
Device: 14h/20d Inode: 50700660 Links: 1
Access: (0644/-rw-r--r--) Uid: ( 1013/ yas) Gid: ( 510/ prof)
Access: 2019-01-27 15:29:58.000000000 +0900
Modify: 2018-06-08 10:46:57.004451000 +0900
Change: 2018-06-08 10:46:57.004451000 +0900
$
図? スーパーブロック、inode、dentry、file
int fd1 = open("file1",O_RDONLY); int fd2 = open("file1",O_RDONLY);ファイル名 "file1" で表現されるファイルの inode 構造体は、1 個でも、 file 構造体は、2 個割り当てられる。
ディスク上には対応するデータ構造は存在しない。
linux-5.4.7/include/linux/fs.h 935: struct file { ... 940: struct path f_path; 941: struct inode *f_inode; /* cached value */ 942: const struct file_operations *f_op; ... 950: atomic_long_t f_count; 951: unsigned int f_flags; 952: fmode_t f_mode; ... 954: loff_t f_pos; ... 964: void *private_data; ... 971: struct address_space *f_mapping; ... 973: } __randomize_layout 974: __attribute__((aligned(4))); /* lest something weird decides that 2 is OK */ linux-5.4.7/include/linux/path.h 8: struct path { 9: struct vfsmount *mnt; 10: struct dentry *dentry; 11: } __randomize_layout;
図? C言語によるオブジェクト指向の継承・委譲の実装方法。共通インスタンス変数・関数、固有インスタンス変数関数の置き方
struct fileの操作は、たとえば次のような形で行われる。 第1引数は、struct file *。
struct file *file; file->f_op->read(file, buf, count, pos);f_op には、次のような手続きがある。各ファイルシステム (ext4,nfs,tmpfs,...) ごとに、手続きの実体は異なるが、インタフェースは同じ。
linux-5.4.7/include/linux/fs.h 1821: struct file_operations { 1822: struct module *owner; 1823: loff_t (*llseek) (struct file *, loff_t, int); 1824: ssize_t (*read) (struct file *, char __user *, size_t, loff_t *); 1825: ssize_t (*write) (struct file *, const char __user *, size_t, loff_t *); 1826: ssize_t (*read_iter) (struct kiocb *, struct iov_iter *); 1827: ssize_t (*write_iter) (struct kiocb *, struct iov_iter *); 1828: int (*iopoll)(struct kiocb *kiocb, bool spin); 1829: int (*iterate) (struct file *, struct dir_context *); 1830: int (*iterate_shared) (struct file *, struct dir_context *); 1831: __poll_t (*poll) (struct file *, struct poll_table_struct *); 1832: long (*unlocked_ioctl) (struct file *, unsigned int, unsigned long); 1833: long (*compat_ioctl) (struct file *, unsigned int, unsigned long); 1834: int (*mmap) (struct file *, struct vm_area_struct *); 1835: unsigned long mmap_supported_flags; 1836: int (*open) (struct inode *, struct file *); 1837: int (*flush) (struct file *, fl_owner_t id); 1838: int (*release) (struct inode *, struct file *); 1839: int (*fsync) (struct file *, loff_t, loff_t, int datasync); 1840: int (*fasync) (int, struct file *, int); 1841: int (*lock) (struct file *, int, struct file_lock *); 1842: ssize_t (*sendpage) (struct file *, struct page *, int, size_t, loff_t *, int); 1843: unsigned long (*get_unmapped_area)(struct file *, unsigned long, unsigned long, unsigned long, unsigned long); 1844: int (*check_flags)(int); 1845: int (*flock) (struct file *, int, struct file_lock *); 1846: ssize_t (*splice_write)(struct pipe_inode_info *, struct file *, loff_t *, size_t, unsigned int); 1847: ssize_t (*splice_read)(struct file *, loff_t *, struct pipe_inode_info *, size_t, unsigned int); 1848: int (*setlease)(struct file *, long, struct file_lock **, void **); 1849: long (*fallocate)(struct file *file, int mode, loff_t offset, 1850: loff_t len); 1851: void (*show_fdinfo)(struct seq_file *m, struct file *f); 1852: #ifndef CONFIG_MMU 1853: unsigned (*mmap_capabilities)(struct file *); 1854: #endif 1855: ssize_t (*copy_file_range)(struct file *, loff_t, struct file *, 1856: loff_t, size_t, unsigned int); 1857: loff_t (*remap_file_range)(struct file *file_in, loff_t pos_in, 1858: struct file *file_out, loff_t pos_out, 1859: loff_t len, unsigned int remap_flags); 1860: int (*fadvise)(struct file *, loff_t, loff_t, int); 1861: } __randomize_layout;主な手続きの意味
linux-5.4.7/include/linux/dcache.h 89: struct dentry { ... 94: struct dentry *d_parent; /* parent directory */ 95: struct qstr d_name; 96: struct inode *d_inode; /* Where the name belongs to - NULL is 97: * negative */ 98: unsigned char d_iname[DNAME_INLINE_LEN]; /* small names */ 99: 100: /* Ref lookup also touches following */ 101: struct lockref d_lockref; /* per-dentry lock and refcount */ 102: const struct dentry_operations *d_op; 103: struct super_block *d_sb; /* The root of the dentry tree */ ... 105: void *d_fsdata; /* fs-specific data */ ... 111: struct list_head d_child; /* child of parent list */ 112: struct list_head d_subdirs; /* our children */ ... 121: } __randomize_layout; 284: static inline unsigned d_count(const struct dentry *dentry) 285: { 286: return dentry->d_lockref.count; 287: } 33: #define HASH_LEN_DECLARE u32 hash; u32 len 47: struct qstr { 48: union { 49: struct { 50: HASH_LEN_DECLARE; 51: }; 52: u64 hash_len; 53: }; 54: const unsigned char *name; 55: }; 83: # define DNAME_INLINE_LEN 40 /* 128 bytes */
linux-5.4.7/include/linux/fs.h 628: struct inode { 629: umode_t i_mode; 630: unsigned short i_opflags; 631: kuid_t i_uid; 632: kgid_t i_gid; ... 640: const struct inode_operations *i_op; 641: struct super_block *i_sb; ... 649: unsigned long i_ino; ... 658: const unsigned int i_nlink; ... 661: dev_t i_rdev; 662: loff_t i_size; 663: struct timespec64 i_atime; 664: struct timespec64 i_mtime; 665: struct timespec64 i_ctime; 666: spinlock_t i_lock; /* i_blocks, i_bytes, maybe i_size */ 667: unsigned short i_bytes; 668: u8 i_blkbits; 669: u8 i_write_hint; 670: blkcnt_t i_blocks; ... 683: struct hlist_node i_hash; ... 697: struct hlist_head i_dentry; ... 701: atomic_t i_count; ... 737: void *i_private; /* fs or device private pointer */ 738: } __randomize_layout;
struct inode *inode; ... inode->i_op->create(inode, name, mode, true);i_op には、次のような手続きがある。各ファイルシステム (ext4,nfs,tmpfs,...) ごとに、手続きの実体は異なるが、インタフェースは同じ。
linux-5.4.7/include/linux/fs.h 1863: struct inode_operations { 1864: struct dentry * (*lookup) (struct inode *,struct dentry *, unsigned int); 1865: const char * (*get_link) (struct dentry *, struct inode *, struct delayed_call *); 1866: int (*permission) (struct inode *, int); 1867: struct posix_acl * (*get_acl)(struct inode *, int); 1868: 1869: int (*readlink) (struct dentry *, char __user *,int); 1870: 1871: int (*create) (struct inode *,struct dentry *, umode_t, bool); 1872: int (*link) (struct dentry *,struct inode *,struct dentry *); 1873: int (*unlink) (struct inode *,struct dentry *); 1874: int (*symlink) (struct inode *,struct dentry *,const char *); 1875: int (*mkdir) (struct inode *,struct dentry *,umode_t); 1876: int (*rmdir) (struct inode *,struct dentry *); 1877: int (*mknod) (struct inode *,struct dentry *,umode_t,dev_t); 1878: int (*rename) (struct inode *, struct dentry *, 1879: struct inode *, struct dentry *, unsigned int); 1880: int (*setattr) (struct dentry *, struct iattr *); 1881: int (*getattr) (const struct path *, struct kstat *, u32, unsigned int); 1882: ssize_t (*listxattr) (struct dentry *, char *, size_t); 1883: int (*fiemap)(struct inode *, struct fiemap_extent_info *, u64 start, 1884: u64 len); 1885: int (*update_time)(struct inode *, struct timespec64 *, int); 1886: int (*atomic_open)(struct inode *, struct dentry *, 1887: struct file *, unsigned open_flag, 1888: umode_t create_mode); 1889: int (*tmpfile) (struct inode *, struct dentry *, umode_t); 1890: int (*set_acl)(struct inode *, struct posix_acl *, int); 1891: } ____cacheline_aligned;
linux-5.4.7/include/linux/fs.h 1424: struct super_block { ... 1429: loff_t s_maxbytes; /* Max file size */ 1430: struct file_system_type *s_type; 1431: const struct super_operations *s_op; ... 1438: struct dentry *s_root; ... 1469: void *s_fs_info; /* Filesystem private info */ ... 1533: struct list_lru s_dentry_lru; 1534: struct list_lru s_inode_lru; ... 1547: struct list_head s_inodes; /* all inodes */ ... 1551: } __randomize_layout;
p->files->fdt->fd[fd]
の struct file を表
す。開いているファイルの数が小さい時は、
p->files->fd_array[fd]
と同じ。
多くのファイルを開くプロセスでは、
p->files->fd_array[NR_OPEN_DEFAULT]
で足りなくなった時は、
expand_files(), expand_fdtable() で拡張する。
これらの関数では、kmalloc() 等でメモリを割り当てる。
linux-5.4.7/include/linux/sched.h 624: struct task_struct { ... 908: struct files_struct *files; ... 1286: }; linux-5.4.7/include/linux/fdtable.h 24: #define NR_OPEN_DEFAULT BITS_PER_LONG ... 48: struct files_struct { ... 56: struct fdtable __rcu *fdt; 57: struct fdtable fdtab; ... 66: struct file __rcu * fd_array[NR_OPEN_DEFAULT]; 67: }; linux-5.4.7/include/asm-generic/bitsperlong.h 8: #ifdef CONFIG_64BIT 9: #define BITS_PER_LONG 64 10: #else 11: #define BITS_PER_LONG 32 12: #endif /* CONFIG_64BIT */
図? task_struct、ファイル記述子、file構造体、その他
linux-5.4.7/fs/read_write.c 595: SYSCALL_DEFINE3(read, unsigned int, fd, char __user *, buf, size_t, count) 596: { 597: return ksys_read(fd, buf, count); 598: } 576: ssize_t ksys_read(unsigned int fd, char __user *buf, size_t count) 577: { 578: struct fd f = fdget_pos(fd); 579: ssize_t ret = -EBADF; 580: 581: if (f.file) { 582: loff_t pos, *ppos = file_ppos(f.file); 583: if (ppos) { 584: pos = *ppos; 585: ppos = &pos; 586: } 587: ret = vfs_read(f.file, buf, count, ppos); 588: if (ret >= 0 && ppos) 589: f.file->f_pos = pos; 590: fdput_pos(f); 591: } 592: return ret; 593: } linux-5.4.7/include/linux/file.h 34: struct fd { 35: struct file *file; 36: unsigned int flags; 37: };
linux-5.4.7/fs/read_write.c 446: ssize_t vfs_read(struct file *file, char __user *buf, size_t count, loff_t *pos) 447: { 448: ssize_t ret; 449: 450: if (!(file->f_mode & FMODE_READ)) 451: return -EBADF; 452: if (!(file->f_mode & FMODE_CAN_READ)) 453: return -EINVAL; 454: if (unlikely(!access_ok(buf, count))) 455: return -EFAULT; 456: 457: ret = rw_verify_area(READ, file, pos, count); 458: if (!ret) { 459: if (count > MAX_RW_COUNT) 460: count = MAX_RW_COUNT; 461: ret = __vfs_read(file, buf, count, pos); 462: if (ret > 0) { 463: fsnotify_access(file); 464: add_rchar(current, ret); 465: } 466: inc_syscr(current); 467: } 468: 469: return ret; 470: } 421: ssize_t __vfs_read(struct file *file, char __user *buf, size_t count, 422: loff_t *pos) 423: { 424: if (file->f_op->read) 425: return file->f_op->read(file, buf, count, pos); 426: else if (file->f_op->read_iter) 427: return new_sync_read(file, buf, count, pos); 428: else 429: return -EINVAL; 430: } 403: static ssize_t new_sync_read(struct file *filp, char __user *buf, size_t len, loff_t *ppos) 404: { 405: struct iovec iov = { .iov_base = buf, .iov_len = len }; ... 411: kiocb.ki_pos = (ppos ? *ppos : 0); 412: iov_iter_init(&iter, READ, &iov, 1, len); 413: 414: ret = call_read_iter(filp, &kiocb, &iter); ... 416: if (ppos) 417: *ppos = kiocb.ki_pos; 418: return ret; 419: } linux-5.4.7/include/linux/fs.h 1893: static inline ssize_t call_read_iter(struct file *file, struct kiocb *kio, 1894: struct iov_iter *iter) 1895: { 1896: return file->f_op->read_iter(kio, iter); 1897: }
linux-5.4.7/fs/ext4/file.c 512: const struct file_operations ext4_file_operations = { 513: .llseek = ext4_llseek, 514: .read_iter = ext4_file_read_iter, 515: .write_iter = ext4_file_write_iter, 516: .unlocked_ioctl = ext4_ioctl, 517: #ifdef CONFIG_COMPAT 518: .compat_ioctl = ext4_compat_ioctl, 519: #endif 520: .mmap = ext4_file_mmap, 521: .mmap_supported_flags = MAP_SYNC, 522: .open = ext4_file_open, 523: .release = ext4_release_file, 524: .fsync = ext4_sync_file, 525: .get_unmapped_area = thp_get_unmapped_area, 526: .splice_read = generic_file_splice_read, 527: .splice_write = iter_file_splice_write, 528: .fallocate = ext4_fallocate, 529: }; 531: const struct inode_operations ext4_file_inode_operations = { ... 538: }; linux-5.4.7/fs/ext4/super.c 1410: static const struct super_operations ext4_sops = { ... 1438: };
linux-5.4.7/fs/ext4/ext4.h 952: struct ext4_inode_info { ... 1019: struct inode vfs_inode; ... 1095: }; 1560: static inline struct ext4_inode_info *EXT4_I(struct inode *inode) 1561: { 1562: return container_of(inode, struct ext4_inode_info, vfs_inode); 1563: } linux-5.4.7/include/linux/kernel.h 985: /** 986: * container_of - cast a member of a structure out to the containing structure 987: * @ptr: the pointer to the member. 988: * @type: the type of the container struct this is embedded in. 989: * @member: the name of the member within the struct. 990: * 991: */ 992: #define container_of(ptr, type, member) ({ \ 993: void *__mptr = (void *)(ptr); \ 994: BUILD_BUG_ON_MSG(!__same_type(*(ptr), ((type *)0)->member) && \ 995: !__same_type(*(ptr), void), \ 996: "pointer type mismatch in container_of()"); \ 997: ((type *)(__mptr - offsetof(type, member))); }) 998:
図? Ext4 ファイルシステムで使う構造体 ext4_inode_info での struct inode の保持
linux-5.4.7/fs/ext4/file.c 65: static ssize_t ext4_file_read_iter(struct kiocb *iocb, struct iov_iter *to) 66: { ... 77: return generic_file_read_iter(iocb, to); 78: }
linux-5.4.7/mm/filemap.c 2274: generic_file_read_iter(struct kiocb *iocb, struct iov_iter *iter) 2275: { ... 2324: retval = generic_file_buffered_read(iocb, iter, retval); 2325: out: 2326: return retval; 2327: }
linux-5.4.7/mm/filemap.c 2009: static ssize_t generic_file_buffered_read(struct kiocb *iocb, 2010: struct iov_iter *iter, ssize_t written) 2011: { 2012: struct file *filp = iocb->ki_filp; 2013: struct address_space *mapping = filp->f_mapping; 2014: struct inode *inode = mapping->host; 2015: struct file_ra_state *ra = &filp->f_ra; 2016: loff_t *ppos = &iocb->ki_pos; 2017: pgoff_t index; ... 2020: unsigned long offset; /* offset into pagecache page */ ... 2022: int error = 0; ... 2028: index = *ppos >> PAGE_SHIFT; ... 2032: offset = *ppos & ~PAGE_MASK; 2033: 2034: for (;;) { 2035: struct page *page; ... 2038: unsigned long nr, ret; ... 2041: find_page: ... 2047: page = find_get_page(mapping, index); 2048: if (!page) { 2049: if (iocb->ki_flags & IOCB_NOWAIT) 2050: goto would_block; 2051: page_cache_sync_readahead(mapping, 2052: ra, filp, 2053: index, last_index - index); 2054: page = find_get_page(mapping, index); 2055: if (unlikely(page == NULL)) 2056: goto no_cached_page; 2057: } ... 2096: page_ok: ... 2114: nr = PAGE_SIZE; ... 2122: nr = nr - offset; ... 2144: ret = copy_page_to_iter(page, offset, nr, iter); 2145: offset += ret; 2146: index += offset >> PAGE_SHIFT; 2147: offset &= ~PAGE_MASK; ... 2150: put_page(page); 2151: written += ret; 2152: if (!iov_iter_count(iter)) 2153: goto out; ... 2160: page_not_up_to_date: ... 2166: page_not_up_to_date_locked: ... 2180: readpage: ... 2227: no_cached_page: ... 2248: } ... 2252: out: ... 2257: *ppos = ((loff_t)index << PAGE_SHIFT) + offset; 2258: file_accessed(filp); 2259: return written ? written : error; 2260: }
linux-5.4.7/fs/namei.c 3844: SYSCALL_DEFINE2(mkdir, const char __user *, pathname, umode_t, mode) 3845: { 3846: return do_mkdirat(AT_FDCWD, pathname, mode); 3847: } 3814: long do_mkdirat(int dfd, const char __user *pathname, umode_t mode) 3815: { 3816: struct dentry *dentry; 3817: struct path path; 3818: int error; 3819: unsigned int lookup_flags = LOOKUP_DIRECTORY; 3820: 3821: retry: 3822: dentry = user_path_create(dfd, pathname, &path, lookup_flags); ... 3830: error = vfs_mkdir(path.dentry->d_inode, dentry, mode); ... 3836: return error; 3837: }
linux-5.4.7/fs/namei.c 3788: int vfs_mkdir(struct inode *dir, struct dentry *dentry, umode_t mode) 3789: { ... 3796: if (!dir->i_op->mkdir) 3797: return -EPERM; 3798: 3799: mode &= (S_IRWXUGO|S_ISVTX); ... 3807: error = dir->i_op->mkdir(dir, dentry, mode); ... 3810: return error; 3811: }
void f(int arg1, int arg2) { 省略; }これを実現するために、どのような Tasklet のハンドラと初期化コードを書け ばよいか。以下の空欄を埋めなさい。
static struct tasklet_struct tl1; void tasklet_handler(unsigned long data) { /* Tasklet ハンドラ */ int arg1, arg2; arg1 = 省略; arg2 = 省略; /*空欄(a)*/ その他の仕事; } 初期化 { /*空欄(b)*/(&tl1, /*空欄(c)*/, 0 ); }
irqreturn_t irq_handler(int irq, void *dev) { /*空欄(d)*/ return IRQ_HANDLED; }
SYSCALL_DEFINE2(flock, unsigned int, fd, unsigned int, cmd) { struct fd f = fdget(fd); ... if (f.file->f_op->/*空欄(e)*/) error = f.file->f_op->/*空欄(f)*/(/*空欄(g)*/, (can_sleep) ? F_SETLKW : F_SETLK, lock); else error = 省略; ... fdput(f); ... return error; }
int vfs_symlink(struct inode *dir, struct dentry *dentry, const char *oldname) { ... if (!dir->i_op->/*空欄(h)*/) return -EPERM; ... error = dir->i_op->/*空欄(h)*/(/*空欄(j)*/, dentry, oldname); ... return error; }