/*
 *  linux/fs/buffer.c
 *
 *  Copyright (C) 1991, 1992  Linus Torvalds
 */

/*
 *  'buffer.c' implements the buffer-cache functions. Race-conditions have
 * been avoided by NEVER letting an interrupt change a buffer (except for the
 * data, of course), but instead letting the caller do it.
 */

/*
 * 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/errno.h>
#include <linux/sched.h>
#include <linux/kernel.h>
#include <linux/major.h>
#include <linux/string.h>
#include <linux/locks.h>
#include <linux/errno.h>

#include <asm/system.h>
#include <asm/io.h>

#ifdef CONFIG_SCSI
#ifdef CONFIG_BLK_DEV_SR
extern int check_cdrom_media_change(int, int);
#endif
#ifdef CONFIG_BLK_DEV_SD
extern int check_scsidisk_media_change(int, int);
extern int revalidate_scsidisk(int, int);
#endif
#endif
#ifdef CONFIG_CDU31A
extern int check_cdu31a_media_change(int, int);
#endif
#ifdef CONFIG_MCD
extern int check_mcd_media_change(int, int);
#endif

static int grow_buffers(int pri, int size);

static struct buffer_head * hash_table[NR_HASH];
static struct buffer_head * free_list = NULL;
static struct buffer_head * unused_list = NULL;
static struct wait_queue * buffer_wait = NULL;

int nr_buffers = 0;
int buffermem = 0;
int nr_buffer_heads = 0;
static int min_free_pages = 20;	/* nr free pages needed before buffer grows */
extern int *blksize_size[];

/*
 * Rewrote the wait-routines to use the "new" wait-queue functionality,
 * and getting rid of the cli-sti pairs. The wait-queue routines still
 * need cli-sti, but now it's just a couple of 386 instructions or so.
 *
 * Note that the real wait_on_buffer() is an inline function that checks
 * if 'b_wait' is set before calling this, so that the queues aren't set
 * up unnecessarily.
 */
void __wait_on_buffer(struct buffer_head * bh)
{
	struct wait_queue wait = { current, NULL };

	bh->b_count++;
	add_wait_queue(&bh->b_wait, &wait);
repeat:
	current->state = TASK_UNINTERRUPTIBLE;
	if (bh->b_lock) {
		schedule();
		goto repeat;
	}
	remove_wait_queue(&bh->b_wait, &wait);
	bh->b_count--;
	current->state = TASK_RUNNING;
}

/* Call sync_buffers with wait!=0 to ensure that the call does not
   return until all buffer writes have completed.  Sync() may return
   before the writes have finished; fsync() may not. */

static int sync_buffers(dev_t dev, int wait)
{
	int i, retry, pass = 0, err = 0;
	struct buffer_head * bh;

	/* One pass for no-wait, three for wait:
	   0) write out all dirty, unlocked buffers;
	   1) write out all dirty buffers, waiting if locked;
	   2) wait for completion by waiting for all buffers to unlock.
	 */
repeat:
	retry = 0;
	bh = free_list;
	for (i = nr_buffers*2 ; i-- > 0 ; bh = bh->b_next_free) {
		if (dev && bh->b_dev != dev)
			continue;
#ifdef 0 /* Disable bad-block debugging code */
		if (bh->b_req && !bh->b_lock &&
		    !bh->b_dirt && !bh->b_uptodate)
			printk ("Warning (IO error) - orphaned block %08x on %04x\n",
				bh->b_blocknr, bh->b_dev);
#endif
		if (bh->b_lock)
		{
			/* Buffer is locked; skip it unless wait is
			   requested AND pass > 0. */
			if (!wait || !pass) {
				retry = 1;
				continue;
			}
			wait_on_buffer (bh);
		}
		/* If an unlocked buffer is not uptodate, there has been 
		   an IO error. Skip it. */
		if (wait && bh->b_req && !bh->b_lock &&
		    !bh->b_dirt && !bh->b_uptodate)
		{
			err = 1;
			continue;
		}
		/* Don't write clean buffers.  Don't write ANY buffers
		   on the third pass. */
		if (!bh->b_dirt || pass>=2)
			continue;
		bh->b_count++;
		ll_rw_block(WRITE, 1, &bh);
		bh->b_count--;
		retry = 1;
	}
	/* If we are waiting for the sync to succeed, and if any dirty
	   blocks were written, then repeat; on the second pass, only
	   wait for buffers being written (do not pass to write any
	   more buffers on the second pass). */
	if (wait && retry && ++pass<=2)
		goto repeat;
	return err;
}

void sync_dev(dev_t dev)
{
	sync_buffers(dev, 0);
	sync_supers(dev);
	sync_inodes(dev);
	sync_buffers(dev, 0);
}

int fsync_dev(dev_t dev)
{
	sync_buffers(dev, 0);
	sync_supers(dev);
	sync_inodes(dev);
	return sync_buffers(dev, 1);
}

asmlinkage int sys_sync(void)
{
	sync_dev(0);
	return 0;
}

int file_fsync (struct inode *inode, struct file *filp)
{
	return fsync_dev(inode->i_dev);
}

asmlinkage int sys_fsync(unsigned int fd)
{
	struct file * file;
	struct inode * inode;

	if (fd>=NR_OPEN || !(file=current->filp[fd]) || !(inode=file->f_inode))
		return -EBADF;
	if (!file->f_op || !file->f_op->fsync)
		return -EINVAL;
	if (file->f_op->fsync(inode,file))
		return -EIO;
	return 0;
}

void invalidate_buffers(dev_t dev)
{
	int i;
	struct buffer_head * bh;

	bh = free_list;
	for (i = nr_buffers*2 ; --i > 0 ; bh = bh->b_next_free) {
		if (bh->b_dev != dev)
			continue;
		wait_on_buffer(bh);
		if (bh->b_dev == dev)
			bh->b_uptodate = bh->b_dirt = bh->b_req = 0;
	}
}

/*
 * 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(dev_t dev)
{
	int i;
	struct buffer_head * bh;

	switch(MAJOR(dev)){
	case FLOPPY_MAJOR:
		if (!(bh = getblk(dev,0,1024)))
			return;
		i = floppy_change(bh);
		brelse(bh);
		break;

#if defined(CONFIG_BLK_DEV_SD) && defined(CONFIG_SCSI)
         case SCSI_DISK_MAJOR:
		i = check_scsidisk_media_change(dev, 0);
		break;
#endif

#if defined(CONFIG_BLK_DEV_SR) && defined(CONFIG_SCSI)
	 case SCSI_CDROM_MAJOR:
		i = check_cdrom_media_change(dev, 0);
		break;
#endif

#if defined(CONFIG_CDU31A)
         case CDU31A_CDROM_MAJOR:
		i = check_cdu31a_media_change(dev, 0);
		break;
#endif

#if defined(CONFIG_MCD)
         case MITSUMI_CDROM_MAJOR:
		i = check_mcd_media_change(dev, 0);
		break;
#endif

         default:
		return;
	};

	if (!i)	return;

	printk("VFS: Disk change detected on device %d/%d\n",
					MAJOR(dev), MINOR(dev));
	for (i=0 ; i<NR_SUPER ; i++)
		if (super_blocks[i].s_dev == dev)
			put_super(super_blocks[i].s_dev);
	invalidate_inodes(dev);
	invalidate_buffers(dev);

#if defined(CONFIG_BLK_DEV_SD) && defined(CONFIG_SCSI)
/* This is trickier for a removable hardisk, because we have to invalidate
   all of the partitions that lie on the disk. */
	if (MAJOR(dev) == SCSI_DISK_MAJOR)
		revalidate_scsidisk(dev, 0);
#endif
}

#define _hashfn(dev,block) (((unsigned)(dev^block))%NR_HASH)
#define hash(dev,block) hash_table[_hashfn(dev,block)]

static inline void remove_from_hash_queue(struct buffer_head * bh)
{
	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;
	bh->b_next = bh->b_prev = NULL;
}

static inline void remove_from_free_list(struct buffer_head * bh)
{
	if (!(bh->b_prev_free) || !(bh->b_next_free))
		panic("VFS: 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;
	bh->b_next_free = bh->b_prev_free = NULL;
}

static inline void remove_from_queues(struct buffer_head * bh)
{
	remove_from_hash_queue(bh);
	remove_from_free_list(bh);
}

static inline void put_first_free(struct buffer_head * bh)
{
	if (!bh || (bh == free_list))
		return;
	remove_from_free_list(bh);
/* add to front 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;
	free_list = bh;
}

static inline void put_last_free(struct buffer_head * bh)
{
	if (!bh)
		return;
	if (bh == free_list) {
		free_list = bh->b_next_free;
		return;
	}
	remove_from_free_list(bh);
/* add to back 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;
}

static inline void insert_into_queues(struct buffer_head * bh)
{
/* 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;
	if (bh->b_next)
		bh->b_next->b_prev = bh;
}

static struct buffer_head * find_buffer(dev_t dev, int block, int size)
{		
	struct buffer_head * tmp;

	for (tmp = hash(dev,block) ; tmp != NULL ; tmp = tmp->b_next)
		if (tmp->b_dev==dev && tmp->b_blocknr==block)
			if (tmp->b_size == size)
				return tmp;
			else {
				printk("VFS: Wrong blocksize on device %d/%d\n",
							MAJOR(dev), MINOR(dev));
				return NULL;
			}
	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(dev_t dev, int block, int size)
{
	struct buffer_head * bh;

	for (;;) {
		if (!(bh=find_buffer(dev,block,size)))
			return NULL;
		bh->b_count++;
		wait_on_buffer(bh);
		if (bh->b_dev == dev && bh->b_blocknr == block && bh->b_size == size)
			return bh;
		bh->b_count--;
	}
}

void set_blocksize(dev_t dev, int size)
{
	int i;
	struct buffer_head * bh, *bhnext;

	if (!blksize_size[MAJOR(dev)])
		return;

	switch(size) {
		default: panic("Invalid blocksize passed to set_blocksize");
		case 512: case 1024: case 2048: case 4096:;
	}

	if (blksize_size[MAJOR(dev)][MINOR(dev)] == 0 && size == BLOCK_SIZE) {
		blksize_size[MAJOR(dev)][MINOR(dev)] = size;
		return;
	}
	if (blksize_size[MAJOR(dev)][MINOR(dev)] == size)
		return;
	sync_buffers(dev, 2);
	blksize_size[MAJOR(dev)][MINOR(dev)] = size;

  /* We need to be quite careful how we do this - we are moving entries
     around on the free list, and we can get in a loop if we are not careful.*/

	bh = free_list;
	for (i = nr_buffers*2 ; --i > 0 ; bh = bhnext) {
		bhnext = bh->b_next_free; 
		if (bh->b_dev != dev)
			continue;
		if (bh->b_size == size)
			continue;

		wait_on_buffer(bh);
		if (bh->b_dev == dev && bh->b_size != size)
			bh->b_uptodate = bh->b_dirt = 0;
		remove_from_hash_queue(bh);
/*    put_first_free(bh); */
	}
}

/*
 * 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.
 *
 * 14.02.92: changed it to sync dirty buffers a bit: better performance
 * when the filesystem starts to get full of dirty blocks (I hope).
 */
#define BADNESS(bh) (((bh)->b_dirt<<1)+(bh)->b_lock)
struct buffer_head * getblk(dev_t dev, int block, int size)
{
	struct buffer_head * bh, * tmp;
	int buffers;
	static int grow_size = 0;

repeat:
	bh = get_hash_table(dev, block, size);
	if (bh) {
		if (bh->b_uptodate && !bh->b_dirt)
			put_last_free(bh);
		return bh;
	}
	grow_size -= size;
	if (nr_free_pages > min_free_pages && grow_size <= 0) {
		if (grow_buffers(GFP_BUFFER, size))
			grow_size = PAGE_SIZE;
	}
	buffers = nr_buffers;
	bh = NULL;

	for (tmp = free_list; buffers-- > 0 ; tmp = tmp->b_next_free) {
		if (tmp->b_count || tmp->b_size != size)
			continue;
		if (mem_map[MAP_NR((unsigned long) tmp->b_data)] != 1)
			continue;
		if (!bh || BADNESS(tmp)<BADNESS(bh)) {
			bh = tmp;
			if (!BADNESS(tmp))
				break;
		}
#if 0
		if (tmp->b_dirt) {
			tmp->b_count++;
			ll_rw_block(WRITEA, 1, &tmp);
			tmp->b_count--;
		}
#endif
	}

	if (!bh) {
		if (nr_free_pages > 5)
			if (grow_buffers(GFP_BUFFER, size))
				goto repeat;
		if (!grow_buffers(GFP_ATOMIC, size))
			sleep_on(&buffer_wait);
		goto repeat;
	}

	wait_on_buffer(bh);
	if (bh->b_count || bh->b_size != size)
		goto repeat;
	if (bh->b_dirt) {
		sync_buffers(0,0);
		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,size))
		goto repeat;
/* OK, FINALLY we know that this buffer is the only one of its kind, */
/* and that it's unused (b_count=0), unlocked (b_lock=0), and clean */
	bh->b_count=1;
	bh->b_dirt=0;
	bh->b_uptodate=0;
	bh->b_req=0;
	remove_from_queues(bh);