xfs_iomap.c

来自「Linux Kernel 2.6.9 for OMAP1710」· C语言 代码 · 共 1,001 行 · 第 1/2 页

	if (error) {
		goto error0;
	}

	error = xfs_trans_commit(tp, XFS_TRANS_RELEASE_LOG_RES, NULL);
	if (error) {
		goto error_out;
	}

	/* copy any maps to caller's array and return any error. */
	if (nimaps == 0) {
		error = (ENOSPC);
		goto error_out;
	}

	*ret_imap = imap[0];
	*nmaps = 1;
	if ( !(io->io_flags & XFS_IOCORE_RT)  && !ret_imap->br_startblock) {
                cmn_err(CE_PANIC,"Access to block zero:  fs <%s> inode: %lld "
                        "start_block : %llx start_off : %llx blkcnt : %llx "
                        "extent-state : %x \n",
                        (ip->i_mount)->m_fsname,
                        (long long)ip->i_ino,
                        ret_imap->br_startblock, ret_imap->br_startoff,
                        ret_imap->br_blockcount,ret_imap->br_state);
        }
	return 0;

 error0:	/* Cancel bmap, unlock inode, and cancel trans */
	xfs_bmap_cancel(&free_list);

 error1:	/* Just cancel transaction */
	xfs_trans_cancel(tp, XFS_TRANS_RELEASE_LOG_RES | XFS_TRANS_ABORT);
	*nmaps = 0;	/* nothing set-up here */

error_out:
	return XFS_ERROR(error);
}

int
xfs_iomap_write_delay(
	xfs_inode_t	*ip,
	loff_t		offset,
	size_t		count,
	int		ioflag,
	xfs_bmbt_irec_t *ret_imap,
	int		*nmaps)
{
	xfs_mount_t	*mp = ip->i_mount;
	xfs_iocore_t	*io = &ip->i_iocore;
	xfs_fileoff_t	offset_fsb;
	xfs_fileoff_t	last_fsb;
	xfs_fsize_t	isize;
	xfs_fsblock_t	firstblock;
	int		nimaps;
	int		error;
	xfs_bmbt_irec_t imap[XFS_WRITE_IMAPS];
	int		aeof;
	int		fsynced = 0;

	ASSERT(ismrlocked(&ip->i_lock, MR_UPDATE) != 0);

	/*
	 * Make sure that the dquots are there. This doesn't hold
	 * the ilock across a disk read.
	 */

	error = XFS_QM_DQATTACH(mp, ip, XFS_QMOPT_ILOCKED);
	if (error)
		return XFS_ERROR(error);

retry:
	isize = ip->i_d.di_size;
	if (io->io_new_size > isize) {
		isize = io->io_new_size;
	}

	aeof = 0;
	offset_fsb = XFS_B_TO_FSBT(mp, offset);
	last_fsb = XFS_B_TO_FSB(mp, ((xfs_ufsize_t)(offset + count)));
	/*
	 * If the caller is doing a write at the end of the file,
	 * then extend the allocation (and the buffer used for the write)
	 * out to the file system's write iosize.  We clean up any extra
	 * space left over when the file is closed in xfs_inactive().
	 *
	 * For sync writes, we are flushing delayed allocate space to
	 * try to make additional space available for allocation near
	 * the filesystem full boundary - preallocation hurts in that
	 * situation, of course.
	 */
	if (!(ioflag & BMAPI_SYNC) && ((offset + count) > ip->i_d.di_size)) {
		xfs_off_t	aligned_offset;
		xfs_filblks_t   count_fsb;
		unsigned int	iosize;
		xfs_fileoff_t	ioalign;
		int		n;
		xfs_fileoff_t   start_fsb;

		/*
		 * If there are any real blocks past eof, then don't
		 * do any speculative allocation.
		 */
		start_fsb = XFS_B_TO_FSBT(mp,
					((xfs_ufsize_t)(offset + count - 1)));
		count_fsb = XFS_B_TO_FSB(mp, (xfs_ufsize_t)XFS_MAXIOFFSET(mp));
		while (count_fsb > 0) {
			nimaps = XFS_WRITE_IMAPS;
			error = XFS_BMAPI(mp, NULL, io, start_fsb, count_fsb,
					0, &firstblock, 0, imap, &nimaps, NULL);
			if (error) {
				return error;
			}
			for (n = 0; n < nimaps; n++) {
				if ( !(io->io_flags & XFS_IOCORE_RT)  && 
					!imap[n].br_startblock) {
					cmn_err(CE_PANIC,"Access to block "
						"zero:  fs <%s> inode: %lld "
						"start_block : %llx start_off "
						": %llx blkcnt : %llx "
						"extent-state : %x \n",
						(ip->i_mount)->m_fsname,
						(long long)ip->i_ino,
						imap[n].br_startblock,
						imap[n].br_startoff,
						imap[n].br_blockcount,
						imap[n].br_state);
        			}
				if ((imap[n].br_startblock != HOLESTARTBLOCK) &&
				    (imap[n].br_startblock != DELAYSTARTBLOCK)) {
					goto write_map;
				}
				start_fsb += imap[n].br_blockcount;
				count_fsb -= imap[n].br_blockcount;
			}
		}
		iosize = mp->m_writeio_blocks;
		aligned_offset = XFS_WRITEIO_ALIGN(mp, (offset + count - 1));
		ioalign = XFS_B_TO_FSBT(mp, aligned_offset);
		last_fsb = ioalign + iosize;
		aeof = 1;
	}
write_map:
	nimaps = XFS_WRITE_IMAPS;
	firstblock = NULLFSBLOCK;

	/*
	 * If mounted with the "-o swalloc" option, roundup the allocation
	 * request to a stripe width boundary if the file size is >=
	 * stripe width and we are allocating past the allocation eof.
	 */
	if (!(io->io_flags & XFS_IOCORE_RT) && mp->m_swidth 
	    && (mp->m_flags & XFS_MOUNT_SWALLOC)
	    && (isize >= XFS_FSB_TO_B(mp, mp->m_swidth)) && aeof) {
		int eof;
		xfs_fileoff_t new_last_fsb;

		new_last_fsb = roundup_64(last_fsb, mp->m_swidth);
		error = xfs_bmap_eof(ip, new_last_fsb, XFS_DATA_FORK, &eof);
		if (error) {
			return error;
		}
		if (eof) {
			last_fsb = new_last_fsb;
		}
	/*
	 * Roundup the allocation request to a stripe unit (m_dalign) boundary
	 * if the file size is >= stripe unit size, and we are allocating past
	 * the allocation eof.
	 */
	} else if (!(io->io_flags & XFS_IOCORE_RT) && mp->m_dalign &&
		   (isize >= XFS_FSB_TO_B(mp, mp->m_dalign)) && aeof) {
		int eof;
		xfs_fileoff_t new_last_fsb;
		new_last_fsb = roundup_64(last_fsb, mp->m_dalign);
		error = xfs_bmap_eof(ip, new_last_fsb, XFS_DATA_FORK, &eof);
		if (error) {
			return error;
		}
		if (eof) {
			last_fsb = new_last_fsb;
		}
	/*
	 * Round up the allocation request to a real-time extent boundary
	 * if the file is on the real-time subvolume.
	 */
	} else if (io->io_flags & XFS_IOCORE_RT && aeof) {
		int eof;
		xfs_fileoff_t new_last_fsb;

		new_last_fsb = roundup_64(last_fsb, mp->m_sb.sb_rextsize);
		error = XFS_BMAP_EOF(mp, io, new_last_fsb, XFS_DATA_FORK, &eof);
		if (error) {
			return error;
		}
		if (eof)
			last_fsb = new_last_fsb;
	}
	error = xfs_bmapi(NULL, ip, offset_fsb,
			  (xfs_filblks_t)(last_fsb - offset_fsb),
			  XFS_BMAPI_DELAY | XFS_BMAPI_WRITE |
			  XFS_BMAPI_ENTIRE, &firstblock, 1, imap,
			  &nimaps, NULL);
	/*
	 * This can be EDQUOT, if nimaps == 0
	 */
	if (error && (error != ENOSPC)) {
		return XFS_ERROR(error);
	}
	/*
	 * If bmapi returned us nothing, and if we didn't get back EDQUOT,
	 * then we must have run out of space.
	 */
	if (nimaps == 0) {
		xfs_iomap_enter_trace(XFS_IOMAP_WRITE_NOSPACE,
					io, offset, count);
		if (xfs_flush_space(ip, &fsynced, &ioflag))
			return XFS_ERROR(ENOSPC);

		error = 0;
		goto retry;
	}

	*ret_imap = imap[0];
	*nmaps = 1;
	if ( !(io->io_flags & XFS_IOCORE_RT)  && !ret_imap->br_startblock) {
		cmn_err(CE_PANIC,"Access to block zero:  fs <%s> inode: %lld "
                        "start_block : %llx start_off : %llx blkcnt : %llx "
                        "extent-state : %x \n",
                        (ip->i_mount)->m_fsname,
                        (long long)ip->i_ino,
                        ret_imap->br_startblock, ret_imap->br_startoff,
                        ret_imap->br_blockcount,ret_imap->br_state);
	}
	return 0;
}

/*
 * Pass in a delayed allocate extent, convert it to real extents;
 * return to the caller the extent we create which maps on top of
 * the originating callers request.
 *
 * Called without a lock on the inode.
 */
int
xfs_iomap_write_allocate(
	xfs_inode_t	*ip,
	xfs_bmbt_irec_t *map,
	int		*retmap)
{
	xfs_mount_t	*mp = ip->i_mount;
	xfs_iocore_t    *io = &ip->i_iocore;
	xfs_fileoff_t	offset_fsb, last_block;
	xfs_fileoff_t	end_fsb, map_start_fsb;
	xfs_fsblock_t	first_block;
	xfs_bmap_free_t	free_list;
	xfs_filblks_t	count_fsb;
	xfs_bmbt_irec_t	imap[XFS_STRAT_WRITE_IMAPS];
	xfs_trans_t	*tp;
	int		i, nimaps, committed;
	int		error = 0;
	int		nres;

	*retmap = 0;

	/*
	 * Make sure that the dquots are there.
	 */
	if ((error = XFS_QM_DQATTACH(mp, ip, 0)))
		return XFS_ERROR(error);

	offset_fsb = map->br_startoff;
	count_fsb = map->br_blockcount;
	map_start_fsb = offset_fsb;

	XFS_STATS_ADD(xs_xstrat_bytes, XFS_FSB_TO_B(mp, count_fsb));

	while (count_fsb != 0) {
		/*
		 * Set up a transaction with which to allocate the
		 * backing store for the file.  Do allocations in a
		 * loop until we get some space in the range we are
		 * interested in.  The other space that might be allocated
		 * is in the delayed allocation extent on which we sit
		 * but before our buffer starts.
		 */

		nimaps = 0;
		while (nimaps == 0) {
			tp = xfs_trans_alloc(mp, XFS_TRANS_STRAT_WRITE);
			nres = XFS_EXTENTADD_SPACE_RES(mp, XFS_DATA_FORK);
			error = xfs_trans_reserve(tp, nres,
					XFS_WRITE_LOG_RES(mp),
					0, XFS_TRANS_PERM_LOG_RES,
					XFS_WRITE_LOG_COUNT);
			if (error == ENOSPC) {
				error = xfs_trans_reserve(tp, 0,
						XFS_WRITE_LOG_RES(mp),
						0,
						XFS_TRANS_PERM_LOG_RES,
						XFS_WRITE_LOG_COUNT);
			}
			if (error) {
				xfs_trans_cancel(tp, 0);
				return XFS_ERROR(error);
			}
			xfs_ilock(ip, XFS_ILOCK_EXCL);
			xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL);
			xfs_trans_ihold(tp, ip);

			XFS_BMAP_INIT(&free_list, &first_block);

			nimaps = XFS_STRAT_WRITE_IMAPS;
			/*
			 * Ensure we don't go beyond eof - it is possible
			 * the extents changed since we did the read call,
			 * we dropped the ilock in the interim.
			 */

			end_fsb = XFS_B_TO_FSB(mp, ip->i_d.di_size);
			xfs_bmap_last_offset(NULL, ip, &last_block,
				XFS_DATA_FORK);
			last_block = XFS_FILEOFF_MAX(last_block, end_fsb);
			if ((map_start_fsb + count_fsb) > last_block) {
				count_fsb = last_block - map_start_fsb;
				if (count_fsb == 0) {
					error = EAGAIN;
					goto trans_cancel;
				}
			}

			/* Go get the actual blocks */
			error = xfs_bmapi(tp, ip, map_start_fsb, count_fsb,
					XFS_BMAPI_WRITE, &first_block, 1,
					imap, &nimaps, &free_list);
			if (error)
				goto trans_cancel;

			error = xfs_bmap_finish(&tp, &free_list,
					first_block, &committed);
			if (error)
				goto trans_cancel;

			error = xfs_trans_commit(tp,
					XFS_TRANS_RELEASE_LOG_RES, NULL);
			if (error)
				goto error0;

			xfs_iunlock(ip, XFS_ILOCK_EXCL);
		}

		/*
		 * See if we were able to allocate an extent that
		 * covers at least part of the callers request
		 */

		for (i = 0; i < nimaps; i++) {
			if ( !(io->io_flags & XFS_IOCORE_RT)  && 
				!imap[i].br_startblock) {
				cmn_err(CE_PANIC,"Access to block zero:  "
					"fs <%s> inode: %lld "
					"start_block : %llx start_off : %llx " 
					"blkcnt : %llx extent-state : %x \n",
					(ip->i_mount)->m_fsname,
					(long long)ip->i_ino,
					imap[i].br_startblock,
					imap[i].br_startoff,
				        imap[i].br_blockcount,imap[i].br_state);
                        }
			if ((map->br_startoff >= imap[i].br_startoff) &&
			    (map->br_startoff < (imap[i].br_startoff +
						 imap[i].br_blockcount))) {
				*map = imap[i];
				*retmap = 1;
				XFS_STATS_INC(xs_xstrat_quick);
				return 0;
			}
			count_fsb -= imap[i].br_blockcount;
		}

		/* So far we have not mapped the requested part of the
		 * file, just surrounding data, try again.
		 */
		nimaps--;
		offset_fsb = imap[nimaps].br_startoff +
			     imap[nimaps].br_blockcount;
		map_start_fsb = offset_fsb;
	}

trans_cancel:
	xfs_bmap_cancel(&free_list);
	xfs_trans_cancel(tp, XFS_TRANS_RELEASE_LOG_RES | XFS_TRANS_ABORT);
error0:
	xfs_iunlock(ip, XFS_ILOCK_EXCL);
	return XFS_ERROR(error);
}

int
xfs_iomap_write_unwritten(
	xfs_inode_t	*ip,
	loff_t		offset,
	size_t		count)
{
	xfs_mount_t	*mp = ip->i_mount;
	xfs_iocore_t    *io = &ip->i_iocore;
	xfs_trans_t	*tp;
	xfs_fileoff_t	offset_fsb;
	xfs_filblks_t	count_fsb;
	xfs_filblks_t	numblks_fsb;
	xfs_bmbt_irec_t	imap;
	int		committed;
	int		error;
	int		nres;
	int		nimaps;
	xfs_fsblock_t	firstfsb;
	xfs_bmap_free_t	free_list;

	xfs_iomap_enter_trace(XFS_IOMAP_UNWRITTEN,
				&ip->i_iocore, offset, count);

	offset_fsb = XFS_B_TO_FSBT(mp, offset);
	count_fsb = XFS_B_TO_FSB(mp, (xfs_ufsize_t)offset + count);
	count_fsb = (xfs_filblks_t)(count_fsb - offset_fsb);

	do {
		nres = XFS_DIOSTRAT_SPACE_RES(mp, 0);

		/*
		 * set up a transaction to convert the range of extents
		 * from unwritten to real. Do allocations in a loop until
		 * we have covered the range passed in.
		 */

		tp = xfs_trans_alloc(mp, XFS_TRANS_STRAT_WRITE);
		error = xfs_trans_reserve(tp, nres,
				XFS_WRITE_LOG_RES(mp), 0,
				XFS_TRANS_PERM_LOG_RES,
				XFS_WRITE_LOG_COUNT);
		if (error) {
			xfs_trans_cancel(tp, 0);
			goto error0;
		}

		xfs_ilock(ip, XFS_ILOCK_EXCL);
		xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL);
		xfs_trans_ihold(tp, ip);

		/*
		 * Modify the unwritten extent state of the buffer.
		 */
		XFS_BMAP_INIT(&free_list, &firstfsb);
		nimaps = 1;
		error = xfs_bmapi(tp, ip, offset_fsb, count_fsb,
				  XFS_BMAPI_WRITE, &firstfsb,
				  1, &imap, &nimaps, &free_list);
		if (error)
			goto error_on_bmapi_transaction;

		error = xfs_bmap_finish(&(tp), &(free_list),
				firstfsb, &committed);
		if (error)
			goto error_on_bmapi_transaction;

		error = xfs_trans_commit(tp, XFS_TRANS_RELEASE_LOG_RES, NULL);
		xfs_iunlock(ip, XFS_ILOCK_EXCL);
		if (error)
			goto error0;
		
		if ( !(io->io_flags & XFS_IOCORE_RT)  && !imap.br_startblock) {
			cmn_err(CE_PANIC,"Access to block zero:  fs <%s> "
				"inode: %lld start_block : %llx start_off : "
				"%llx blkcnt : %llx extent-state : %x \n",
				(ip->i_mount)->m_fsname,
				(long long)ip->i_ino,
				imap.br_startblock,imap.br_startoff,
				imap.br_blockcount,imap.br_state);
        	}

		if ((numblks_fsb = imap.br_blockcount) == 0) {
			/*
			 * The numblks_fsb value should always get
			 * smaller, otherwise the loop is stuck.
			 */
			ASSERT(imap.br_blockcount);
			break;
		}
		offset_fsb += numblks_fsb;
		count_fsb -= numblks_fsb;
	} while (count_fsb > 0);

	return 0;

error_on_bmapi_transaction:
	xfs_bmap_cancel(&free_list);
	xfs_trans_cancel(tp, (XFS_TRANS_RELEASE_LOG_RES | XFS_TRANS_ABORT));
	xfs_iunlock(ip, XFS_ILOCK_EXCL);
error0:
	return XFS_ERROR(error);
}