xfs_trans_buf.c

linux 内核源代码

		  xfs_buf_t	*bp,
		  uint		first,
		  uint		last)
{
	xfs_buf_log_item_t	*bip;
	xfs_log_item_desc_t	*lidp;

	ASSERT(XFS_BUF_ISBUSY(bp));
	ASSERT(XFS_BUF_FSPRIVATE2(bp, xfs_trans_t *) == tp);
	ASSERT(XFS_BUF_FSPRIVATE(bp, void *) != NULL);
	ASSERT((first <= last) && (last < XFS_BUF_COUNT(bp)));
	ASSERT((XFS_BUF_IODONE_FUNC(bp) == NULL) ||
	       (XFS_BUF_IODONE_FUNC(bp) == xfs_buf_iodone_callbacks));

	/*
	 * Mark the buffer as needing to be written out eventually,
	 * and set its iodone function to remove the buffer's buf log
	 * item from the AIL and free it when the buffer is flushed
	 * to disk.  See xfs_buf_attach_iodone() for more details
	 * on li_cb and xfs_buf_iodone_callbacks().
	 * If we end up aborting this transaction, we trap this buffer
	 * inside the b_bdstrat callback so that this won't get written to
	 * disk.
	 */
	XFS_BUF_DELAYWRITE(bp);
	XFS_BUF_DONE(bp);

	bip = XFS_BUF_FSPRIVATE(bp, xfs_buf_log_item_t *);
	ASSERT(atomic_read(&bip->bli_refcount) > 0);
	XFS_BUF_SET_IODONE_FUNC(bp, xfs_buf_iodone_callbacks);
	bip->bli_item.li_cb = (void(*)(xfs_buf_t*,xfs_log_item_t*))xfs_buf_iodone;

	/*
	 * If we invalidated the buffer within this transaction, then
	 * cancel the invalidation now that we're dirtying the buffer
	 * again.  There are no races with the code in xfs_buf_item_unpin(),
	 * because we have a reference to the buffer this entire time.
	 */
	if (bip->bli_flags & XFS_BLI_STALE) {
		xfs_buf_item_trace("BLOG UNSTALE", bip);
		bip->bli_flags &= ~XFS_BLI_STALE;
		ASSERT(XFS_BUF_ISSTALE(bp));
		XFS_BUF_UNSTALE(bp);
		bip->bli_format.blf_flags &= ~XFS_BLI_CANCEL;
	}

	lidp = xfs_trans_find_item(tp, (xfs_log_item_t*)bip);
	ASSERT(lidp != NULL);

	tp->t_flags |= XFS_TRANS_DIRTY;
	lidp->lid_flags |= XFS_LID_DIRTY;
	lidp->lid_flags &= ~XFS_LID_BUF_STALE;
	bip->bli_flags |= XFS_BLI_LOGGED;
	xfs_buf_item_log(bip, first, last);
	xfs_buf_item_trace("BLOG", bip);
}


/*
 * This called to invalidate a buffer that is being used within
 * a transaction.  Typically this is because the blocks in the
 * buffer are being freed, so we need to prevent it from being
 * written out when we're done.  Allowing it to be written again
 * might overwrite data in the free blocks if they are reallocated
 * to a file.
 *
 * We prevent the buffer from being written out by clearing the
 * B_DELWRI flag.  We can't always
 * get rid of the buf log item at this point, though, because
 * the buffer may still be pinned by another transaction.  If that
 * is the case, then we'll wait until the buffer is committed to
 * disk for the last time (we can tell by the ref count) and
 * free it in xfs_buf_item_unpin().  Until it is cleaned up we
 * will keep the buffer locked so that the buffer and buf log item
 * are not reused.
 */
void
xfs_trans_binval(
	xfs_trans_t	*tp,
	xfs_buf_t	*bp)
{
	xfs_log_item_desc_t	*lidp;
	xfs_buf_log_item_t	*bip;

	ASSERT(XFS_BUF_ISBUSY(bp));
	ASSERT(XFS_BUF_FSPRIVATE2(bp, xfs_trans_t *) == tp);
	ASSERT(XFS_BUF_FSPRIVATE(bp, void *) != NULL);

	bip = XFS_BUF_FSPRIVATE(bp, xfs_buf_log_item_t *);
	lidp = xfs_trans_find_item(tp, (xfs_log_item_t*)bip);
	ASSERT(lidp != NULL);
	ASSERT(atomic_read(&bip->bli_refcount) > 0);

	if (bip->bli_flags & XFS_BLI_STALE) {
		/*
		 * If the buffer is already invalidated, then
		 * just return.
		 */
		ASSERT(!(XFS_BUF_ISDELAYWRITE(bp)));
		ASSERT(XFS_BUF_ISSTALE(bp));
		ASSERT(!(bip->bli_flags & (XFS_BLI_LOGGED | XFS_BLI_DIRTY)));
		ASSERT(!(bip->bli_format.blf_flags & XFS_BLI_INODE_BUF));
		ASSERT(bip->bli_format.blf_flags & XFS_BLI_CANCEL);
		ASSERT(lidp->lid_flags & XFS_LID_DIRTY);
		ASSERT(tp->t_flags & XFS_TRANS_DIRTY);
		xfs_buftrace("XFS_BINVAL RECUR", bp);
		xfs_buf_item_trace("BINVAL RECUR", bip);
		return;
	}

	/*
	 * Clear the dirty bit in the buffer and set the STALE flag
	 * in the buf log item.  The STALE flag will be used in
	 * xfs_buf_item_unpin() to determine if it should clean up
	 * when the last reference to the buf item is given up.
	 * We set the XFS_BLI_CANCEL flag in the buf log format structure
	 * and log the buf item.  This will be used at recovery time
	 * to determine that copies of the buffer in the log before
	 * this should not be replayed.
	 * We mark the item descriptor and the transaction dirty so
	 * that we'll hold the buffer until after the commit.
	 *
	 * Since we're invalidating the buffer, we also clear the state
	 * about which parts of the buffer have been logged.  We also
	 * clear the flag indicating that this is an inode buffer since
	 * the data in the buffer will no longer be valid.
	 *
	 * We set the stale bit in the buffer as well since we're getting
	 * rid of it.
	 */
	XFS_BUF_UNDELAYWRITE(bp);
	XFS_BUF_STALE(bp);
	bip->bli_flags |= XFS_BLI_STALE;
	bip->bli_flags &= ~(XFS_BLI_LOGGED | XFS_BLI_DIRTY);
	bip->bli_format.blf_flags &= ~XFS_BLI_INODE_BUF;
	bip->bli_format.blf_flags |= XFS_BLI_CANCEL;
	memset((char *)(bip->bli_format.blf_data_map), 0,
	      (bip->bli_format.blf_map_size * sizeof(uint)));
	lidp->lid_flags |= XFS_LID_DIRTY|XFS_LID_BUF_STALE;
	tp->t_flags |= XFS_TRANS_DIRTY;
	xfs_buftrace("XFS_BINVAL", bp);
	xfs_buf_item_trace("BINVAL", bip);
}

/*
 * This call is used to indicate that the buffer contains on-disk
 * inodes which must be handled specially during recovery.  They
 * require special handling because only the di_next_unlinked from
 * the inodes in the buffer should be recovered.  The rest of the
 * data in the buffer is logged via the inodes themselves.
 *
 * All we do is set the XFS_BLI_INODE_BUF flag in the buffer's log
 * format structure so that we'll know what to do at recovery time.
 */
/* ARGSUSED */
void
xfs_trans_inode_buf(
	xfs_trans_t	*tp,
	xfs_buf_t	*bp)
{
	xfs_buf_log_item_t	*bip;

	ASSERT(XFS_BUF_ISBUSY(bp));
	ASSERT(XFS_BUF_FSPRIVATE2(bp, xfs_trans_t *) == tp);
	ASSERT(XFS_BUF_FSPRIVATE(bp, void *) != NULL);

	bip = XFS_BUF_FSPRIVATE(bp, xfs_buf_log_item_t *);
	ASSERT(atomic_read(&bip->bli_refcount) > 0);

	bip->bli_format.blf_flags |= XFS_BLI_INODE_BUF;
}

/*
 * This call is used to indicate that the buffer is going to
 * be staled and was an inode buffer. This means it gets
 * special processing during unpin - where any inodes 
 * associated with the buffer should be removed from ail.
 * There is also special processing during recovery,
 * any replay of the inodes in the buffer needs to be
 * prevented as the buffer may have been reused.
 */
void
xfs_trans_stale_inode_buf(
	xfs_trans_t	*tp,
	xfs_buf_t	*bp)
{
	xfs_buf_log_item_t	*bip;

	ASSERT(XFS_BUF_ISBUSY(bp));
	ASSERT(XFS_BUF_FSPRIVATE2(bp, xfs_trans_t *) == tp);
	ASSERT(XFS_BUF_FSPRIVATE(bp, void *) != NULL);

	bip = XFS_BUF_FSPRIVATE(bp, xfs_buf_log_item_t *);
	ASSERT(atomic_read(&bip->bli_refcount) > 0);

	bip->bli_flags |= XFS_BLI_STALE_INODE;
	bip->bli_item.li_cb = (void(*)(xfs_buf_t*,xfs_log_item_t*))
		xfs_buf_iodone;
}



/*
 * Mark the buffer as being one which contains newly allocated
 * inodes.  We need to make sure that even if this buffer is
 * relogged as an 'inode buf' we still recover all of the inode
 * images in the face of a crash.  This works in coordination with
 * xfs_buf_item_committed() to ensure that the buffer remains in the
 * AIL at its original location even after it has been relogged.
 */
/* ARGSUSED */
void
xfs_trans_inode_alloc_buf(
	xfs_trans_t	*tp,
	xfs_buf_t	*bp)
{
	xfs_buf_log_item_t	*bip;

	ASSERT(XFS_BUF_ISBUSY(bp));
	ASSERT(XFS_BUF_FSPRIVATE2(bp, xfs_trans_t *) == tp);
	ASSERT(XFS_BUF_FSPRIVATE(bp, void *) != NULL);

	bip = XFS_BUF_FSPRIVATE(bp, xfs_buf_log_item_t *);
	ASSERT(atomic_read(&bip->bli_refcount) > 0);

	bip->bli_flags |= XFS_BLI_INODE_ALLOC_BUF;
}


/*
 * Similar to xfs_trans_inode_buf(), this marks the buffer as a cluster of
 * dquots. However, unlike in inode buffer recovery, dquot buffers get
 * recovered in their entirety. (Hence, no XFS_BLI_DQUOT_ALLOC_BUF flag).
 * The only thing that makes dquot buffers different from regular
 * buffers is that we must not replay dquot bufs when recovering
 * if a _corresponding_ quotaoff has happened. We also have to distinguish
 * between usr dquot bufs and grp dquot bufs, because usr and grp quotas
 * can be turned off independently.
 */
/* ARGSUSED */
void
xfs_trans_dquot_buf(
	xfs_trans_t	*tp,
	xfs_buf_t	*bp,
	uint		type)
{
	xfs_buf_log_item_t	*bip;

	ASSERT(XFS_BUF_ISBUSY(bp));
	ASSERT(XFS_BUF_FSPRIVATE2(bp, xfs_trans_t *) == tp);
	ASSERT(XFS_BUF_FSPRIVATE(bp, void *) != NULL);
	ASSERT(type == XFS_BLI_UDQUOT_BUF ||
	       type == XFS_BLI_PDQUOT_BUF ||
	       type == XFS_BLI_GDQUOT_BUF);

	bip = XFS_BUF_FSPRIVATE(bp, xfs_buf_log_item_t *);
	ASSERT(atomic_read(&bip->bli_refcount) > 0);

	bip->bli_format.blf_flags |= type;
}

/*
 * Check to see if a buffer matching the given parameters is already
 * a part of the given transaction.  Only check the first, embedded
 * chunk, since we don't want to spend all day scanning large transactions.
 */
STATIC xfs_buf_t *
xfs_trans_buf_item_match(
	xfs_trans_t	*tp,
	xfs_buftarg_t	*target,
	xfs_daddr_t	blkno,
	int		len)
{
	xfs_log_item_chunk_t	*licp;
	xfs_log_item_desc_t	*lidp;
	xfs_buf_log_item_t	*blip;
	xfs_buf_t		*bp;
	int			i;

	bp = NULL;
	len = BBTOB(len);
	licp = &tp->t_items;
	if (!XFS_LIC_ARE_ALL_FREE(licp)) {
		for (i = 0; i < licp->lic_unused; i++) {
			/*
			 * Skip unoccupied slots.
			 */
			if (XFS_LIC_ISFREE(licp, i)) {
				continue;
			}

			lidp = XFS_LIC_SLOT(licp, i);
			blip = (xfs_buf_log_item_t *)lidp->lid_item;
			if (blip->bli_item.li_type != XFS_LI_BUF) {
				continue;
			}

			bp = blip->bli_buf;
			if ((XFS_BUF_TARGET(bp) == target) &&
			    (XFS_BUF_ADDR(bp) == blkno) &&
			    (XFS_BUF_COUNT(bp) == len)) {
				/*
				 * We found it.  Break out and
				 * return the pointer to the buffer.
				 */
				break;
			} else {
				bp = NULL;
			}
		}
	}
	return bp;
}

/*
 * Check to see if a buffer matching the given parameters is already
 * a part of the given transaction.  Check all the chunks, we
 * want to be thorough.
 */
STATIC xfs_buf_t *
xfs_trans_buf_item_match_all(
	xfs_trans_t	*tp,
	xfs_buftarg_t	*target,
	xfs_daddr_t	blkno,
	int		len)
{
	xfs_log_item_chunk_t	*licp;
	xfs_log_item_desc_t	*lidp;
	xfs_buf_log_item_t	*blip;
	xfs_buf_t		*bp;
	int			i;

	bp = NULL;
	len = BBTOB(len);
	for (licp = &tp->t_items; licp != NULL; licp = licp->lic_next) {
		if (XFS_LIC_ARE_ALL_FREE(licp)) {
			ASSERT(licp == &tp->t_items);
			ASSERT(licp->lic_next == NULL);
			return NULL;
		}
		for (i = 0; i < licp->lic_unused; i++) {
			/*
			 * Skip unoccupied slots.
			 */
			if (XFS_LIC_ISFREE(licp, i)) {
				continue;
			}

			lidp = XFS_LIC_SLOT(licp, i);
			blip = (xfs_buf_log_item_t *)lidp->lid_item;
			if (blip->bli_item.li_type != XFS_LI_BUF) {
				continue;
			}

			bp = blip->bli_buf;
			if ((XFS_BUF_TARGET(bp) == target) &&
			    (XFS_BUF_ADDR(bp) == blkno) &&
			    (XFS_BUF_COUNT(bp) == len)) {
				/*
				 * We found it.  Break out and
				 * return the pointer to the buffer.
				 */
				return bp;
			}
		}
	}
	return NULL;
}