xfs_lrw.c

linux 内核源代码

/*
 * Copyright (c) 2000-2003,2005 Silicon Graphics, Inc.
 * All Rights Reserved.
 *
 * This program is free software; you can redistribute it and/or
 * modify it under the terms of the GNU General Public License as
 * published by the Free Software Foundation.
 *
 * This program is distributed in the hope that it would be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 * GNU General Public License for more details.
 *
 * You should have received a copy of the GNU General Public License
 * along with this program; if not, write the Free Software Foundation,
 * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
 */
#include "xfs.h"
#include "xfs_fs.h"
#include "xfs_bit.h"
#include "xfs_log.h"
#include "xfs_inum.h"
#include "xfs_trans.h"
#include "xfs_sb.h"
#include "xfs_ag.h"
#include "xfs_dir2.h"
#include "xfs_alloc.h"
#include "xfs_dmapi.h"
#include "xfs_quota.h"
#include "xfs_mount.h"
#include "xfs_bmap_btree.h"
#include "xfs_alloc_btree.h"
#include "xfs_ialloc_btree.h"
#include "xfs_dir2_sf.h"
#include "xfs_attr_sf.h"
#include "xfs_dinode.h"
#include "xfs_inode.h"
#include "xfs_bmap.h"
#include "xfs_btree.h"
#include "xfs_ialloc.h"
#include "xfs_rtalloc.h"
#include "xfs_error.h"
#include "xfs_itable.h"
#include "xfs_rw.h"
#include "xfs_acl.h"
#include "xfs_attr.h"
#include "xfs_inode_item.h"
#include "xfs_buf_item.h"
#include "xfs_utils.h"
#include "xfs_iomap.h"
#include "xfs_vnodeops.h"

#include <linux/capability.h>
#include <linux/writeback.h>


#if defined(XFS_RW_TRACE)
void
xfs_rw_enter_trace(
	int			tag,
	xfs_iocore_t		*io,
	void			*data,
	size_t			segs,
	loff_t			offset,
	int			ioflags)
{
	xfs_inode_t	*ip = XFS_IO_INODE(io);

	if (ip->i_rwtrace == NULL)
		return;
	ktrace_enter(ip->i_rwtrace,
		(void *)(unsigned long)tag,
		(void *)ip,
		(void *)((unsigned long)((ip->i_d.di_size >> 32) & 0xffffffff)),
		(void *)((unsigned long)(ip->i_d.di_size & 0xffffffff)),
		(void *)data,
		(void *)((unsigned long)segs),
		(void *)((unsigned long)((offset >> 32) & 0xffffffff)),
		(void *)((unsigned long)(offset & 0xffffffff)),
		(void *)((unsigned long)ioflags),
		(void *)((unsigned long)((io->io_new_size >> 32) & 0xffffffff)),
		(void *)((unsigned long)(io->io_new_size & 0xffffffff)),
		(void *)((unsigned long)current_pid()),
		(void *)NULL,
		(void *)NULL,
		(void *)NULL,
		(void *)NULL);
}

void
xfs_inval_cached_trace(
	xfs_iocore_t	*io,
	xfs_off_t	offset,
	xfs_off_t	len,
	xfs_off_t	first,
	xfs_off_t	last)
{
	xfs_inode_t	*ip = XFS_IO_INODE(io);

	if (ip->i_rwtrace == NULL)
		return;
	ktrace_enter(ip->i_rwtrace,
		(void *)(__psint_t)XFS_INVAL_CACHED,
		(void *)ip,
		(void *)((unsigned long)((offset >> 32) & 0xffffffff)),
		(void *)((unsigned long)(offset & 0xffffffff)),
		(void *)((unsigned long)((len >> 32) & 0xffffffff)),
		(void *)((unsigned long)(len & 0xffffffff)),
		(void *)((unsigned long)((first >> 32) & 0xffffffff)),
		(void *)((unsigned long)(first & 0xffffffff)),
		(void *)((unsigned long)((last >> 32) & 0xffffffff)),
		(void *)((unsigned long)(last & 0xffffffff)),
		(void *)((unsigned long)current_pid()),
		(void *)NULL,
		(void *)NULL,
		(void *)NULL,
		(void *)NULL,
		(void *)NULL);
}
#endif

/*
 *	xfs_iozero
 *
 *	xfs_iozero clears the specified range of buffer supplied,
 *	and marks all the affected blocks as valid and modified.  If
 *	an affected block is not allocated, it will be allocated.  If
 *	an affected block is not completely overwritten, and is not
 *	valid before the operation, it will be read from disk before
 *	being partially zeroed.
 */
STATIC int
xfs_iozero(
	struct inode		*ip,	/* inode			*/
	loff_t			pos,	/* offset in file		*/
	size_t			count)	/* size of data to zero		*/
{
	struct page		*page;
	struct address_space	*mapping;
	int			status;

	mapping = ip->i_mapping;
	do {
		unsigned offset, bytes;
		void *fsdata;

		offset = (pos & (PAGE_CACHE_SIZE -1)); /* Within page */
		bytes = PAGE_CACHE_SIZE - offset;
		if (bytes > count)
			bytes = count;

		status = pagecache_write_begin(NULL, mapping, pos, bytes,
					AOP_FLAG_UNINTERRUPTIBLE,
					&page, &fsdata);
		if (status)
			break;

		zero_user_page(page, offset, bytes, KM_USER0);

		status = pagecache_write_end(NULL, mapping, pos, bytes, bytes,
					page, fsdata);
		WARN_ON(status <= 0); /* can't return less than zero! */
		pos += bytes;
		count -= bytes;
		status = 0;
	} while (count);

	return (-status);
}

ssize_t			/* bytes read, or (-)  error */
xfs_read(
	xfs_inode_t		*ip,
	struct kiocb		*iocb,
	const struct iovec	*iovp,
	unsigned int		segs,
	loff_t			*offset,
	int			ioflags)
{
	struct file		*file = iocb->ki_filp;
	struct inode		*inode = file->f_mapping->host;
	bhv_vnode_t		*vp = XFS_ITOV(ip);
	xfs_mount_t		*mp = ip->i_mount;
	size_t			size = 0;
	ssize_t			ret = 0;
	xfs_fsize_t		n;
	unsigned long		seg;


	XFS_STATS_INC(xs_read_calls);

	/* START copy & waste from filemap.c */
	for (seg = 0; seg < segs; seg++) {
		const struct iovec *iv = &iovp[seg];

		/*
		 * If any segment has a negative length, or the cumulative
		 * length ever wraps negative then return -EINVAL.
		 */
		size += iv->iov_len;
		if (unlikely((ssize_t)(size|iv->iov_len) < 0))
			return XFS_ERROR(-EINVAL);
	}
	/* END copy & waste from filemap.c */

	if (unlikely(ioflags & IO_ISDIRECT)) {
		xfs_buftarg_t	*target =
			(ip->i_d.di_flags & XFS_DIFLAG_REALTIME) ?
				mp->m_rtdev_targp : mp->m_ddev_targp;
		if ((*offset & target->bt_smask) ||
		    (size & target->bt_smask)) {
			if (*offset == ip->i_size) {
				return (0);
			}
			return -XFS_ERROR(EINVAL);
		}
	}

	n = XFS_MAXIOFFSET(mp) - *offset;
	if ((n <= 0) || (size == 0))
		return 0;

	if (n < size)
		size = n;

	if (XFS_FORCED_SHUTDOWN(mp))
		return -EIO;

	if (unlikely(ioflags & IO_ISDIRECT))
		mutex_lock(&inode->i_mutex);
	xfs_ilock(ip, XFS_IOLOCK_SHARED);

	if (DM_EVENT_ENABLED(ip, DM_EVENT_READ) && !(ioflags & IO_INVIS)) {
		bhv_vrwlock_t locktype = VRWLOCK_READ;
		int dmflags = FILP_DELAY_FLAG(file) | DM_SEM_FLAG_RD(ioflags);

		ret = -XFS_SEND_DATA(mp, DM_EVENT_READ, vp, *offset, size,
					dmflags, &locktype);
		if (ret) {
			xfs_iunlock(ip, XFS_IOLOCK_SHARED);
			if (unlikely(ioflags & IO_ISDIRECT))
				mutex_unlock(&inode->i_mutex);
			return ret;
		}
	}

	if (unlikely(ioflags & IO_ISDIRECT)) {
		if (VN_CACHED(vp))
			ret = xfs_flushinval_pages(ip,
					ctooff(offtoct(*offset)),
					-1, FI_REMAPF_LOCKED);
		mutex_unlock(&inode->i_mutex);
		if (ret) {
			xfs_iunlock(ip, XFS_IOLOCK_SHARED);
			return ret;
		}
	}

	xfs_rw_enter_trace(XFS_READ_ENTER, &ip->i_iocore,
				(void *)iovp, segs, *offset, ioflags);

	iocb->ki_pos = *offset;
	ret = generic_file_aio_read(iocb, iovp, segs, *offset);
	if (ret == -EIOCBQUEUED && !(ioflags & IO_ISAIO))
		ret = wait_on_sync_kiocb(iocb);
	if (ret > 0)
		XFS_STATS_ADD(xs_read_bytes, ret);

	xfs_iunlock(ip, XFS_IOLOCK_SHARED);
	return ret;
}

ssize_t
xfs_splice_read(
	xfs_inode_t		*ip,
	struct file		*infilp,
	loff_t			*ppos,
	struct pipe_inode_info	*pipe,
	size_t			count,
	int			flags,
	int			ioflags)
{
	bhv_vnode_t		*vp = XFS_ITOV(ip);
	xfs_mount_t		*mp = ip->i_mount;
	ssize_t			ret;

	XFS_STATS_INC(xs_read_calls);
	if (XFS_FORCED_SHUTDOWN(ip->i_mount))
		return -EIO;

	xfs_ilock(ip, XFS_IOLOCK_SHARED);

	if (DM_EVENT_ENABLED(ip, DM_EVENT_READ) && !(ioflags & IO_INVIS)) {
		bhv_vrwlock_t locktype = VRWLOCK_READ;
		int error;

		error = XFS_SEND_DATA(mp, DM_EVENT_READ, vp, *ppos, count,
					FILP_DELAY_FLAG(infilp), &locktype);
		if (error) {
			xfs_iunlock(ip, XFS_IOLOCK_SHARED);
			return -error;
		}
	}
	xfs_rw_enter_trace(XFS_SPLICE_READ_ENTER, &ip->i_iocore,
			   pipe, count, *ppos, ioflags);
	ret = generic_file_splice_read(infilp, ppos, pipe, count, flags);
	if (ret > 0)
		XFS_STATS_ADD(xs_read_bytes, ret);

	xfs_iunlock(ip, XFS_IOLOCK_SHARED);
	return ret;
}

ssize_t
xfs_splice_write(
	xfs_inode_t		*ip,
	struct pipe_inode_info	*pipe,
	struct file		*outfilp,
	loff_t			*ppos,
	size_t			count,
	int			flags,
	int			ioflags)
{
	bhv_vnode_t		*vp = XFS_ITOV(ip);
	xfs_mount_t		*mp = ip->i_mount;
	xfs_iocore_t		*io = &ip->i_iocore;
	ssize_t			ret;
	struct inode		*inode = outfilp->f_mapping->host;
	xfs_fsize_t		isize, new_size;

	XFS_STATS_INC(xs_write_calls);
	if (XFS_FORCED_SHUTDOWN(ip->i_mount))
		return -EIO;

	xfs_ilock(ip, XFS_IOLOCK_EXCL);

	if (DM_EVENT_ENABLED(ip, DM_EVENT_WRITE) && !(ioflags & IO_INVIS)) {
		bhv_vrwlock_t locktype = VRWLOCK_WRITE;
		int error;

		error = XFS_SEND_DATA(mp, DM_EVENT_WRITE, vp, *ppos, count,
					FILP_DELAY_FLAG(outfilp), &locktype);
		if (error) {
			xfs_iunlock(ip, XFS_IOLOCK_EXCL);
			return -error;
		}
	}

	new_size = *ppos + count;

	xfs_ilock(ip, XFS_ILOCK_EXCL);
	if (new_size > ip->i_size)
		io->io_new_size = new_size;
	xfs_iunlock(ip, XFS_ILOCK_EXCL);

	xfs_rw_enter_trace(XFS_SPLICE_WRITE_ENTER, &ip->i_iocore,
			   pipe, count, *ppos, ioflags);
	ret = generic_file_splice_write(pipe, outfilp, ppos, count, flags);
	if (ret > 0)
		XFS_STATS_ADD(xs_write_bytes, ret);

	isize = i_size_read(inode);
	if (unlikely(ret < 0 && ret != -EFAULT && *ppos > isize))
		*ppos = isize;

	if (*ppos > ip->i_size) {
		xfs_ilock(ip, XFS_ILOCK_EXCL);
		if (*ppos > ip->i_size)
			ip->i_size = *ppos;
		xfs_iunlock(ip, XFS_ILOCK_EXCL);
	}

	if (io->io_new_size) {
		xfs_ilock(ip, XFS_ILOCK_EXCL);
		io->io_new_size = 0;
		if (ip->i_d.di_size > ip->i_size)
			ip->i_d.di_size = ip->i_size;
		xfs_iunlock(ip, XFS_ILOCK_EXCL);
	}
	xfs_iunlock(ip, XFS_IOLOCK_EXCL);
	return ret;
}

/*
 * This routine is called to handle zeroing any space in the last
 * block of the file that is beyond the EOF.  We do this since the
 * size is being increased without writing anything to that block
 * and we don't want anyone to read the garbage on the disk.
 */
STATIC int				/* error (positive) */
xfs_zero_last_block(
	struct inode	*ip,
	xfs_iocore_t	*io,
	xfs_fsize_t	offset,
	xfs_fsize_t	isize)
{
	xfs_fileoff_t	last_fsb;
	xfs_mount_t	*mp = io->io_mount;
	int		nimaps;
	int		zero_offset;
	int		zero_len;
	int		error = 0;
	xfs_bmbt_irec_t	imap;

	ASSERT(ismrlocked(io->io_lock, MR_UPDATE) != 0);

	zero_offset = XFS_B_FSB_OFFSET(mp, isize);
	if (zero_offset == 0) {
		/*
		 * There are no extra bytes in the last block on disk to
		 * zero, so return.
		 */
		return 0;
	}

	last_fsb = XFS_B_TO_FSBT(mp, isize);
	nimaps = 1;
	error = XFS_BMAPI(mp, NULL, io, last_fsb, 1, 0, NULL, 0, &imap,
			  &nimaps, NULL, NULL);
	if (error) {
		return error;
	}
	ASSERT(nimaps > 0);
	/*
	 * If the block underlying isize is just a hole, then there
	 * is nothing to zero.
	 */
	if (imap.br_startblock == HOLESTARTBLOCK) {
		return 0;
	}
	/*
	 * Zero the part of the last block beyond the EOF, and write it
	 * out sync.  We need to drop the ilock while we do this so we
	 * don't deadlock when the buffer cache calls back to us.
	 */
	XFS_IUNLOCK(mp, io, XFS_ILOCK_EXCL| XFS_EXTSIZE_RD);

	zero_len = mp->m_sb.sb_blocksize - zero_offset;
	if (isize + zero_len > offset)
		zero_len = offset - isize;
	error = xfs_iozero(ip, isize, zero_len);

	XFS_ILOCK(mp, io, XFS_ILOCK_EXCL|XFS_EXTSIZE_RD);
	ASSERT(error >= 0);
	return error;
}

/*
 * Zero any on disk space between the current EOF and the new,
 * larger EOF.  This handles the normal case of zeroing the remainder
 * of the last block in the file and the unusual case of zeroing blocks
 * out beyond the size of the file.  This second case only happens
 * with fixed size extents and when the system crashes before the inode
 * size was updated but after blocks were allocated.  If fill is set,
 * then any holes in the range are filled and zeroed.  If not, the holes
 * are left alone as holes.
 */

int					/* error (positive) */
xfs_zero_eof(
	bhv_vnode_t	*vp,
	xfs_iocore_t	*io,
	xfs_off_t	offset,		/* starting I/O offset */
	xfs_fsize_t	isize)		/* current inode size */
{
	struct inode	*ip = vn_to_inode(vp);
	xfs_fileoff_t	start_zero_fsb;
	xfs_fileoff_t	end_zero_fsb;
	xfs_fileoff_t	zero_count_fsb;
	xfs_fileoff_t	last_fsb;
	xfs_fileoff_t	zero_off;
	xfs_fsize_t	zero_len;
	xfs_mount_t	*mp = io->io_mount;
	int		nimaps;
	int		error = 0;
	xfs_bmbt_irec_t	imap;

	ASSERT(ismrlocked(io->io_lock, MR_UPDATE));
	ASSERT(ismrlocked(io->io_iolock, MR_UPDATE));
	ASSERT(offset > isize);