buffer.c

一份精简的linux内核源代码

/*
 *  linux/fs/buffer.c
 *
 *  (C) 1991  Linus Torvalds
 */

/*
 *  'buffer.c' implements the buffer-cache functions. Race-conditions have
 * been avoided by NEVER letting a interrupt change a buffer (except for the
 * data, of course), but instead letting the caller do it. NOTE! As interrupts
 * can wake up a caller, some cli-sti sequences are needed to check for
 * sleep-on-calls. These should be extremely quick, though (I hope).
 */

/*
 * NOTE! There is one discordant note here: checking floppies for
 * disk change. This is where it fits best, I think, as it should
 * invalidate changed floppy-disk-caches.
 */

#include <stdarg.h>
 
#include <linux/config.h>
#include <linux/sched.h>
#include <linux/kernel.h>
#include <asm/system.h>
#include <asm/io.h>

extern int end; /*内核代码末端地址，由链接程序生成*/
struct buffer_head * start_buffer = (struct buffer_head *) &end;/*缓冲区开始位置*/
struct buffer_head * hash_table[NR_HASH];/*NR_HASH＝307*/
static struct buffer_head * free_list;/*空闲缓冲链表头指针*/
static struct task_struct * buffer_wait = NULL;/*等待空闲缓冲块而睡眠的任务队列*/
int NR_BUFFERS = 0;/*系统含有缓冲块个数，371*/

static inline void wait_on_buffer(struct buffer_head * bh)/*若指定的缓冲块上锁，则当前进程挂起（在b_wait队列中等待）*/
{
	cli();
	while (bh->b_lock)
		sleep_on(&bh->b_wait);
	sti();
}

int sys_sync(void)/*同步块设备和内存缓冲区中的数据*/
{
	int i;
	struct buffer_head * bh;

	sync_inodes();		/* 将所有修改过的i节点结构写入缓冲区,59 */
	bh = start_buffer;
	for (i=0 ; i<NR_BUFFERS ; i++,bh++) {/*将缓冲区中的所有被修改过的数据写入设备中*/
		wait_on_buffer(bh);/*？？某个进程修改已写盘的缓冲块怎么办*/
		if (bh->b_dirt)
			ll_rw_block(WRITE,bh);
	}
	return 0;
}

int sync_dev(int dev)/*对指定设备进行高速缓冲区与设备上数据的同步*/
{
	int i;
	struct buffer_head * bh;

	bh = start_buffer;
	for (i=0 ; i<NR_BUFFERS ; i++,bh++) {/*对缓冲区中指定设备的缓冲块（修改过的）进行写盘*/
		if (bh->b_dev != dev)
			continue;
		wait_on_buffer(bh);
		if (bh->b_dev == dev && bh->b_dirt)
			ll_rw_block(WRITE,bh);
	}
	sync_inodes(); /* 将所有修改过的i节点结构写入缓冲区,59*/
	bh = start_buffer;
	for (i=0 ; i<NR_BUFFERS ; i++,bh++) {{/*对缓冲区中指定设备的缓冲块（修改过的）进行写盘*/
		if (bh->b_dev != dev)
			continue;
		wait_on_buffer(bh);
		if (bh->b_dev == dev && bh->b_dirt)
			ll_rw_block(WRITE,bh);
	}
	return 0;
}

void inline invalidate_buffers(int dev){/*使指定设备在高速缓冲区中的数据无效*/
{
	int i;
	struct buffer_head * bh;

	bh = start_buffer;
	for (i=0 ; i<NR_BUFFERS ; i++,bh++) {
		if (bh->b_dev != dev)
			continue;
		wait_on_buffer(bh);
		if (bh->b_dev == dev)
			bh->b_uptodate = bh->b_dirt = 0;{/*缓冲块中的数据无效*/
	}
}

/*
 * This routine checks whether a floppy has been changed, and
 * invalidates all buffer-cache-entries in that case. This
 * is a relatively slow routine, so we have to try to minimize using
 * it. Thus it is called only upon a 'mount' or 'open'. This
 * is the best way of combining speed and utility, I think.
 * People changing diskettes in the middle of an operation deserve
 * to loose :-)
 *
 * NOTE! Although currently this is only for floppies, the idea is
 * that any additional removable block-device will use this routine,
 * and that mount/open needn't know that floppies/whatever are
 * special.
 */
void check_disk_change(int dev){/*flopy disk 不看*/
{
	int i;

	if (MAJOR(dev) != 2)
		return;
	if (!floppy_change(dev & 0x03))
		return;
	for (i=0 ; i<NR_SUPER ; i++)
		if (super_block[i].s_dev == dev)
			put_super(super_block[i].s_dev);
	invalidate_inodes(dev);
	invalidate_buffers(dev);
}

#define _hashfn(dev,block) (((unsigned)(dev^block))%NR_HASH){/*dev异或block再除以NR_HASH*/
#define hash(dev,block) hash_table[_hashfn(dev,block)]

static inline void remove_from_queues(struct buffer_head * bh){/*从hash队列和空闲缓冲队列中移走缓冲块*/
{
/* remove from hash-queue */
	if (bh->b_next)
		bh->b_next->b_prev = bh->b_prev;
	if (bh->b_prev)
		bh->b_prev->b_next = bh->b_next;
	if (hash(bh->b_dev,bh->b_blocknr) == bh)
		hash(bh->b_dev,bh->b_blocknr) = bh->b_next;
/* remove from free list */
	if (!(bh->b_prev_free) || !(bh->b_next_free))
		panic("Free block list corrupted");
	bh->b_prev_free->b_next_free = bh->b_next_free;
	bh->b_next_free->b_prev_free = bh->b_prev_free;
	if (free_list == bh)
		free_list = bh->b_next_free;
}

static inline void insert_into_queues(struct buffer_head * bh){/*将缓冲块移入hash队列中和空闲缓冲队列尾*/
{
/* put at end of free list */
	bh->b_next_free = free_list;
	bh->b_prev_free = free_list->b_prev_free;
	free_list->b_prev_free->b_next_free = bh;
	free_list->b_prev_free = bh;
/* put the buffer in new hash-queue if it has a device */
	bh->b_prev = NULL;
	bh->b_next = NULL;
	if (!bh->b_dev)
		return;
	bh->b_next = hash(bh->b_dev,bh->b_blocknr);
	hash(bh->b_dev,bh->b_blocknr) = bh;
	bh->b_next->b_prev = bh;
}

static struct buffer_head * find_buffer(int dev, int block)/*利用hash表在高速缓冲中寻找给定设备和指定块号的缓冲块*/
{		
	struct buffer_head * tmp;

	for (tmp = hash(dev,block) ; tmp != NULL ; tmp = tmp->b_next)
		if (tmp->b_dev==dev && tmp->b_blocknr==block)
			return tmp;
	return NULL;
}

/*
 * Why like this, I hear you say... The reason is race-conditions.
 * As we don't lock buffers (unless we are readint them, that is),
 * something might happen to it while we sleep (ie a read-error
 * will force it bad). This shouldn't really happen currently, but
 * the code is ready.
 */
struct buffer_head * get_hash_table(int dev, int block)/*利用hash表在高速缓冲中寻找给定设备和指定块号的缓冲块，并将其上锁*/
{
	struct buffer_head * bh;

	for (;;) {
		if (!(bh=find_buffer(dev,block)))
			return NULL;
		bh->b_count++;
		wait_on_buffer(bh);
		if (bh->b_dev == dev && bh->b_blocknr == block)/*防止在进程睡眠时，该缓冲块的dev或block发生变化*/
			return bh;
		bh->b_count--;
	}
}

/*
 * Ok, this is getblk, and it isn't very clear, again to hinder
 * race-conditions. Most of the code is seldom used, (ie repeating),
 * so it should be much more efficient than it looks.
 *
 * The algoritm is changed: hopefully better, and an elusive bug removed.
 */
#define BADNESS(bh) (((bh)->b_dirt<<1)+(bh)->b_lock)/*设某一缓冲块的权重*/
struct buffer_head * getblk(int dev,int block)/*取高速缓冲区中指定的缓冲块，若直接找到则返回，否则在缓冲区中设置一个新项*/
{
	struct buffer_head * tmp, * bh;

repeat:
	if (bh = get_hash_table(dev,block))
		return bh;
	tmp = free_list;
	do {
		if (tmp->b_count)/*先除去已被别的进程引用的缓冲块*//*！一个未被引用的缓冲块不一定就是干净的或没有上锁的*/
			continue;
		if (!bh || BADNESS(tmp)<BADNESS(bh)) {/*在缓冲队列中找出权值最小的缓冲块*/
			bh = tmp;
			if (!BADNESS(tmp))
				break;
		}
/* and repeat until we find something good */
	} while ((tmp = tmp->b_next_free) != free_list);
	if (!bh) {/*无空闲缓冲块（b_count非0）则挂起当前进程*/
		sleep_on(&buffer_wait);
		goto repeat;
	}
	wait_on_buffer(bh);/*有空闲缓冲块，但该块未解锁*/
	if (bh->b_count)
		goto repeat;
	while (bh->b_dirt) {/*写盘（可能刚写入某一块，但该块立刻又被更改，但这不影响）*/
		sync_dev(bh->b_dev);
		wait_on_buffer(bh);
		if (bh->b_count)
			goto repeat;
	}
/* NOTE!! While we slept waiting for this block, somebody else might */
/* already have added "this" block to the cache. check it */
	if (find_buffer(dev,block))
		goto repeat;
/* OK, FINALLY we know that this buffer is the only one of it's kind, */
/* and that it's unused (b_count=0), unlocked (b_lock=0), and clean */
	bh->b_count=1;/*设缓冲头结构*/
	bh->b_dirt=0;
	bh->b_uptodate=0;
	remove_from_queues(bh);
	bh->b_dev=dev;
	bh->b_blocknr=block;
	insert_into_queues(bh);
	return bh;
}

void brelse(struct buffer_head * buf)/*释放缓冲块（b_count-1，唤醒等待该缓冲块的进程）*/
{
	if (!buf)
		return;
	wait_on_buffer(buf);
	if (!(buf->b_count--))
		panic("Trying to free free buffer");
	wake_up(&buffer_wait);
}

/*
 * bread() reads a specified block and returns the buffer that contains
 * it. It returns NULL if the block was unreadable.
 */
struct buffer_head * bread(int dev,int block)/*从设备上读取设备块*/
{
	struct buffer_head * bh;

	if (!(bh=getblk(dev,block)))/*此处返回的块一定是未上锁的*/
		panic("bread: getblk returned NULL\n");
	if (bh->b_uptodate)/*若缓冲区中已有则直接返回*/
		return bh;
	ll_rw_block(READ,bh);/*底层读设备操作*/
	wait_on_buffer(bh);
	if (bh->b_uptodate)
		return bh;
	brelse(bh);
	return NULL;
}

#define COPYBLK(from,to) \
__asm__("cld\n\t" \
	"rep\n\t" \
	"movsl\n\t" \
	::"c" (BLOCK_SIZE/4),"S" (from),"D" (to) \
	:"cx","di","si")

/*
 * bread_page reads four buffers into memory at the desired address. It's
 * a function of its own, as there is some speed to be got by reading them
 * all at the same time, not waiting for one to be read, and then another
 * etc.
 */
void bread_page(unsigned long address,int dev,int b[4])/*读入4个缓冲块并将其存入指定的地址处（该函数仅用于memory.c中的do_no_page中）*/
{
	struct buffer_head * bh[4];
	int i;

	for (i=0 ; i<4 ; i++)
		if (b[i]) {
			if (bh[i] = getblk(dev,b[i]))
				if (!bh[i]->b_uptodate)
					ll_rw_block(READ,bh[i]);
		} else
			bh[i] = NULL;
	for (i=0 ; i<4 ; i++,address += BLOCK_SIZE)
		if (bh[i]) {
			wait_on_buffer(bh[i]);
			if (bh[i]->b_uptodate)
				COPYBLK((unsigned long) bh[i]->b_data,address);
			brelse(bh[i]);
		}
}

/*
 * Ok, breada can be used as bread, but additionally to mark other
 * blocks for reading as well. End the argument list with a negative
 * number.
 */
struct buffer_head * breada(int dev,int first, ...)/*预读入一些块*/
{
	va_list args;/*函数可变参数*/
	struct buffer_head * bh, *tmp;

	va_start(args,first);/*读第一个块*/
	if (!(bh=getblk(dev,first)))
		panic("bread: getblk returned NULL\n");
	if (!bh->b_uptodate)
		ll_rw_block(READ,bh);
	while ((first=va_arg(args,int))>=0) {/*预读几个块*/
		tmp=getblk(dev,first);
		if (tmp) {
			if (!tmp->b_uptodate)
				ll_rw_block(READA,bh);
			tmp->b_count--;/*因为是预读，所以仅仅是把该块读入缓冲区，但引用计数仍为0*/
		}
	}
	va_end(args);
	wait_on_buffer(bh);
	if (bh->b_uptodate)
		return bh;
	brelse(bh);
	return (NULL);
}

void buffer_init(long buffer_end)/*缓冲区初始化*/
{
	struct buffer_head * h = start_buffer;
	void * b;
	int i;

	if (buffer_end == 1<<20)        /*去除显存*/
		b = (void *) (640*1024);
	else
		b = (void *) buffer_end;
	while ( (b -= BLOCK_SIZE) >= ((void *) (h+1)) ) { /*若剩余内存仍够分配一个缓冲块（1kb），则分配*/
		h->b_dev = 0; /*使用该缓冲块的设备号*/
		h->b_dirt = 0; /*缓冲块修改标志*/
		h->b_count = 0; /*缓冲块应用计数*/
		h->b_lock = 0; /*缓冲块锁定标志*/
		h->b_uptodate = 0; /*缓冲块更新标志*/
		h->b_wait = NULL;/*指向等待该缓冲块解锁的进程*/
		h->b_next = NULL;/*指向具有相同hash值的下一个缓冲头*/
		h->b_prev = NULL;/*指向具有相同hash值的上一个缓冲头*/
		h->b_data = (char *) b;/*指向对应缓冲块数据块*/
		h->b_prev_free = h-1;/*指向链表中前一项*/
		h->b_next_free = h+1;/*指向链表中下一项*/
		h++;                  /*指向下一个新缓冲头位置*/
		NR_BUFFERS++;         /*缓冲区块数累加*/
		if (b == (void *) 0x100000)/*跳过显存*/
			b = (void *) 0xA0000;
	}
	h--;
	free_list = start_buffer;/*形成一个环*/
	free_list->b_prev_free = h;
	h->b_next_free = free_list;
	for (i=0;i<NR_HASH;i++)
		hash_table[i]=NULL;
}