xfs_inode_item.c

linux 内核源代码

 * failed to get the inode flush lock but did get the inode locked SHARED.
 * Here we're trying to see if the inode buffer is incore, and if so whether it's
 * marked delayed write. If that's the case, we'll initiate a bawrite on that
 * buffer to expedite the process.
 *
 * We aren't holding the AIL_LOCK (or the flush lock) when this gets called,
 * so it is inherently race-y.
 */
STATIC void
xfs_inode_item_pushbuf(
	xfs_inode_log_item_t	*iip)
{
	xfs_inode_t	*ip;
	xfs_mount_t	*mp;
	xfs_buf_t	*bp;
	uint		dopush;

	ip = iip->ili_inode;

	ASSERT(ismrlocked(&(ip->i_lock), MR_ACCESS));

	/*
	 * The ili_pushbuf_flag keeps others from
	 * trying to duplicate our effort.
	 */
	ASSERT(iip->ili_pushbuf_flag != 0);
	ASSERT(iip->ili_push_owner == current_pid());

	/*
	 * If flushlock isn't locked anymore, chances are that the
	 * inode flush completed and the inode was taken off the AIL.
	 * So, just get out.
	 */
	if (!issemalocked(&(ip->i_flock)) ||
	    ((iip->ili_item.li_flags & XFS_LI_IN_AIL) == 0)) {
		iip->ili_pushbuf_flag = 0;
		xfs_iunlock(ip, XFS_ILOCK_SHARED);
		return;
	}

	mp = ip->i_mount;
	bp = xfs_incore(mp->m_ddev_targp, iip->ili_format.ilf_blkno,
		    iip->ili_format.ilf_len, XFS_INCORE_TRYLOCK);

	if (bp != NULL) {
		if (XFS_BUF_ISDELAYWRITE(bp)) {
			/*
			 * We were racing with iflush because we don't hold
			 * the AIL_LOCK or the flush lock. However, at this point,
			 * we have the buffer, and we know that it's dirty.
			 * So, it's possible that iflush raced with us, and
			 * this item is already taken off the AIL.
			 * If not, we can flush it async.
			 */
			dopush = ((iip->ili_item.li_flags & XFS_LI_IN_AIL) &&
				  issemalocked(&(ip->i_flock)));
			iip->ili_pushbuf_flag = 0;
			xfs_iunlock(ip, XFS_ILOCK_SHARED);
			xfs_buftrace("INODE ITEM PUSH", bp);
			if (XFS_BUF_ISPINNED(bp)) {
				xfs_log_force(mp, (xfs_lsn_t)0,
					      XFS_LOG_FORCE);
			}
			if (dopush) {
				xfs_bawrite(mp, bp);
			} else {
				xfs_buf_relse(bp);
			}
		} else {
			iip->ili_pushbuf_flag = 0;
			xfs_iunlock(ip, XFS_ILOCK_SHARED);
			xfs_buf_relse(bp);
		}
		return;
	}
	/*
	 * We have to be careful about resetting pushbuf flag too early (above).
	 * Even though in theory we can do it as soon as we have the buflock,
	 * we don't want others to be doing work needlessly. They'll come to
	 * this function thinking that pushing the buffer is their
	 * responsibility only to find that the buffer is still locked by
	 * another doing the same thing
	 */
	iip->ili_pushbuf_flag = 0;
	xfs_iunlock(ip, XFS_ILOCK_SHARED);
	return;
}


/*
 * This is called to asynchronously write the inode associated with this
 * inode log item out to disk. The inode will already have been locked by
 * a successful call to xfs_inode_item_trylock().
 */
STATIC void
xfs_inode_item_push(
	xfs_inode_log_item_t	*iip)
{
	xfs_inode_t	*ip;

	ip = iip->ili_inode;

	ASSERT(ismrlocked(&(ip->i_lock), MR_ACCESS));
	ASSERT(issemalocked(&(ip->i_flock)));
	/*
	 * Since we were able to lock the inode's flush lock and
	 * we found it on the AIL, the inode must be dirty.  This
	 * is because the inode is removed from the AIL while still
	 * holding the flush lock in xfs_iflush_done().  Thus, if
	 * we found it in the AIL and were able to obtain the flush
	 * lock without sleeping, then there must not have been
	 * anyone in the process of flushing the inode.
	 */
	ASSERT(XFS_FORCED_SHUTDOWN(ip->i_mount) ||
	       iip->ili_format.ilf_fields != 0);

	/*
	 * Write out the inode.  The completion routine ('iflush_done') will
	 * pull it from the AIL, mark it clean, unlock the flush lock.
	 */
	(void) xfs_iflush(ip, XFS_IFLUSH_ASYNC);
	xfs_iunlock(ip, XFS_ILOCK_SHARED);

	return;
}

/*
 * XXX rcc - this one really has to do something.  Probably needs
 * to stamp in a new field in the incore inode.
 */
/* ARGSUSED */
STATIC void
xfs_inode_item_committing(
	xfs_inode_log_item_t	*iip,
	xfs_lsn_t		lsn)
{
	iip->ili_last_lsn = lsn;
	return;
}

/*
 * This is the ops vector shared by all buf log items.
 */
static struct xfs_item_ops xfs_inode_item_ops = {
	.iop_size	= (uint(*)(xfs_log_item_t*))xfs_inode_item_size,
	.iop_format	= (void(*)(xfs_log_item_t*, xfs_log_iovec_t*))
					xfs_inode_item_format,
	.iop_pin	= (void(*)(xfs_log_item_t*))xfs_inode_item_pin,
	.iop_unpin	= (void(*)(xfs_log_item_t*, int))xfs_inode_item_unpin,
	.iop_unpin_remove = (void(*)(xfs_log_item_t*, xfs_trans_t*))
					xfs_inode_item_unpin_remove,
	.iop_trylock	= (uint(*)(xfs_log_item_t*))xfs_inode_item_trylock,
	.iop_unlock	= (void(*)(xfs_log_item_t*))xfs_inode_item_unlock,
	.iop_committed	= (xfs_lsn_t(*)(xfs_log_item_t*, xfs_lsn_t))
					xfs_inode_item_committed,
	.iop_push	= (void(*)(xfs_log_item_t*))xfs_inode_item_push,
	.iop_pushbuf	= (void(*)(xfs_log_item_t*))xfs_inode_item_pushbuf,
	.iop_committing = (void(*)(xfs_log_item_t*, xfs_lsn_t))
					xfs_inode_item_committing
};


/*
 * Initialize the inode log item for a newly allocated (in-core) inode.
 */
void
xfs_inode_item_init(
	xfs_inode_t	*ip,
	xfs_mount_t	*mp)
{
	xfs_inode_log_item_t	*iip;

	ASSERT(ip->i_itemp == NULL);
	iip = ip->i_itemp = kmem_zone_zalloc(xfs_ili_zone, KM_SLEEP);

	iip->ili_item.li_type = XFS_LI_INODE;
	iip->ili_item.li_ops = &xfs_inode_item_ops;
	iip->ili_item.li_mountp = mp;
	iip->ili_inode = ip;

	/*
	   We have zeroed memory. No need ...
	   iip->ili_extents_buf = NULL;
	   iip->ili_pushbuf_flag = 0;
	 */

	iip->ili_format.ilf_type = XFS_LI_INODE;
	iip->ili_format.ilf_ino = ip->i_ino;
	iip->ili_format.ilf_blkno = ip->i_blkno;
	iip->ili_format.ilf_len = ip->i_len;
	iip->ili_format.ilf_boffset = ip->i_boffset;
}

/*
 * Free the inode log item and any memory hanging off of it.
 */
void
xfs_inode_item_destroy(
	xfs_inode_t	*ip)
{
#ifdef XFS_TRANS_DEBUG
	if (ip->i_itemp->ili_root_size != 0) {
		kmem_free(ip->i_itemp->ili_orig_root,
			  ip->i_itemp->ili_root_size);
	}
#endif
	kmem_zone_free(xfs_ili_zone, ip->i_itemp);
}


/*
 * This is the inode flushing I/O completion routine.  It is called
 * from interrupt level when the buffer containing the inode is
 * flushed to disk.  It is responsible for removing the inode item
 * from the AIL if it has not been re-logged, and unlocking the inode's
 * flush lock.
 */
/*ARGSUSED*/
void
xfs_iflush_done(
	xfs_buf_t		*bp,
	xfs_inode_log_item_t	*iip)
{
	xfs_inode_t	*ip;
	SPLDECL(s);

	ip = iip->ili_inode;

	/*
	 * We only want to pull the item from the AIL if it is
	 * actually there and its location in the log has not
	 * changed since we started the flush.  Thus, we only bother
	 * if the ili_logged flag is set and the inode's lsn has not
	 * changed.  First we check the lsn outside
	 * the lock since it's cheaper, and then we recheck while
	 * holding the lock before removing the inode from the AIL.
	 */
	if (iip->ili_logged &&
	    (iip->ili_item.li_lsn == iip->ili_flush_lsn)) {
		AIL_LOCK(ip->i_mount, s);
		if (iip->ili_item.li_lsn == iip->ili_flush_lsn) {
			/*
			 * xfs_trans_delete_ail() drops the AIL lock.
			 */
			xfs_trans_delete_ail(ip->i_mount,
					     (xfs_log_item_t*)iip, s);
		} else {
			AIL_UNLOCK(ip->i_mount, s);
		}
	}

	iip->ili_logged = 0;

	/*
	 * Clear the ili_last_fields bits now that we know that the
	 * data corresponding to them is safely on disk.
	 */
	iip->ili_last_fields = 0;

	/*
	 * Release the inode's flush lock since we're done with it.
	 */
	xfs_ifunlock(ip);

	return;
}

/*
 * This is the inode flushing abort routine.  It is called
 * from xfs_iflush when the filesystem is shutting down to clean
 * up the inode state.
 * It is responsible for removing the inode item
 * from the AIL if it has not been re-logged, and unlocking the inode's
 * flush lock.
 */
void
xfs_iflush_abort(
	xfs_inode_t		*ip)
{
	xfs_inode_log_item_t	*iip;
	xfs_mount_t		*mp;
	SPLDECL(s);

	iip = ip->i_itemp;
	mp = ip->i_mount;
	if (iip) {
		if (iip->ili_item.li_flags & XFS_LI_IN_AIL) {
			AIL_LOCK(mp, s);
			if (iip->ili_item.li_flags & XFS_LI_IN_AIL) {
				/*
				 * xfs_trans_delete_ail() drops the AIL lock.
				 */
				xfs_trans_delete_ail(mp, (xfs_log_item_t *)iip,
					s);
			} else
				AIL_UNLOCK(mp, s);
		}
		iip->ili_logged = 0;
		/*
		 * Clear the ili_last_fields bits now that we know that the
		 * data corresponding to them is safely on disk.
		 */
		iip->ili_last_fields = 0;
		/*
		 * Clear the inode logging fields so no more flushes are
		 * attempted.
		 */
		iip->ili_format.ilf_fields = 0;
	}
	/*
	 * Release the inode's flush lock since we're done with it.
	 */
	xfs_ifunlock(ip);
}

void
xfs_istale_done(
	xfs_buf_t		*bp,
	xfs_inode_log_item_t	*iip)
{
	xfs_iflush_abort(iip->ili_inode);
}

/*
 * convert an xfs_inode_log_format struct from either 32 or 64 bit versions
 * (which can have different field alignments) to the native version
 */
int
xfs_inode_item_format_convert(
	xfs_log_iovec_t		*buf,
	xfs_inode_log_format_t	*in_f)
{
	if (buf->i_len == sizeof(xfs_inode_log_format_32_t)) {
		xfs_inode_log_format_32_t *in_f32;

		in_f32 = (xfs_inode_log_format_32_t *)buf->i_addr;
		in_f->ilf_type = in_f32->ilf_type;
		in_f->ilf_size = in_f32->ilf_size;
		in_f->ilf_fields = in_f32->ilf_fields;
		in_f->ilf_asize = in_f32->ilf_asize;
		in_f->ilf_dsize = in_f32->ilf_dsize;
		in_f->ilf_ino = in_f32->ilf_ino;
		/* copy biggest field of ilf_u */
		memcpy(in_f->ilf_u.ilfu_uuid.__u_bits,
		       in_f32->ilf_u.ilfu_uuid.__u_bits,
		       sizeof(uuid_t));
		in_f->ilf_blkno = in_f32->ilf_blkno;
		in_f->ilf_len = in_f32->ilf_len;
		in_f->ilf_boffset = in_f32->ilf_boffset;
		return 0;
	} else if (buf->i_len == sizeof(xfs_inode_log_format_64_t)){
		xfs_inode_log_format_64_t *in_f64;

		in_f64 = (xfs_inode_log_format_64_t *)buf->i_addr;
		in_f->ilf_type = in_f64->ilf_type;
		in_f->ilf_size = in_f64->ilf_size;
		in_f->ilf_fields = in_f64->ilf_fields;
		in_f->ilf_asize = in_f64->ilf_asize;
		in_f->ilf_dsize = in_f64->ilf_dsize;
		in_f->ilf_ino = in_f64->ilf_ino;
		/* copy biggest field of ilf_u */
		memcpy(in_f->ilf_u.ilfu_uuid.__u_bits,
		       in_f64->ilf_u.ilfu_uuid.__u_bits,
		       sizeof(uuid_t));
		in_f->ilf_blkno = in_f64->ilf_blkno;
		in_f->ilf_len = in_f64->ilf_len;
		in_f->ilf_boffset = in_f64->ilf_boffset;
		return 0;
	}
	return EFSCORRUPTED;
}