#include "xfs_error.h"
#include "xfs_itable.h"
#include "xfs_rw.h"
-#include "xfs_acl.h"
#include "xfs_attr.h"
#include "xfs_buf_item.h"
#include "xfs_trans_space.h"
#include "xfs_utils.h"
#include "xfs_iomap.h"
+#include "xfs_trace.h"
-#if defined(XFS_RW_TRACE)
-void
-xfs_iomap_enter_trace(
- int tag,
- xfs_inode_t *ip,
- xfs_off_t offset,
- ssize_t count)
-{
- if (!ip->i_rwtrace)
- 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 *)((unsigned long)((offset >> 32) & 0xffffffff)),
- (void *)((unsigned long)(offset & 0xffffffff)),
- (void *)((unsigned long)count),
- (void *)((unsigned long)((ip->i_new_size >> 32) & 0xffffffff)),
- (void *)((unsigned long)(ip->i_new_size & 0xffffffff)),
- (void *)((unsigned long)current_pid()),
- (void *)NULL,
- (void *)NULL,
- (void *)NULL,
- (void *)NULL,
- (void *)NULL,
- (void *)NULL);
-}
-
-void
-xfs_iomap_map_trace(
- int tag,
- xfs_inode_t *ip,
- xfs_off_t offset,
- ssize_t count,
- xfs_iomap_t *iomapp,
- xfs_bmbt_irec_t *imapp,
- int flags)
-{
- if (!ip->i_rwtrace)
- 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 *)((unsigned long)((offset >> 32) & 0xffffffff)),
- (void *)((unsigned long)(offset & 0xffffffff)),
- (void *)((unsigned long)count),
- (void *)((unsigned long)flags),
- (void *)((unsigned long)((iomapp->iomap_offset >> 32) & 0xffffffff)),
- (void *)((unsigned long)(iomapp->iomap_offset & 0xffffffff)),
- (void *)((unsigned long)(iomapp->iomap_delta)),
- (void *)((unsigned long)(iomapp->iomap_bsize)),
- (void *)((unsigned long)(iomapp->iomap_bn)),
- (void *)(__psint_t)(imapp->br_startoff),
- (void *)((unsigned long)(imapp->br_blockcount)),
- (void *)(__psint_t)(imapp->br_startblock));
-}
-#else
-#define xfs_iomap_enter_trace(tag, io, offset, count)
-#define xfs_iomap_map_trace(tag, io, offset, count, iomapp, imapp, flags)
-#endif
#define XFS_WRITEIO_ALIGN(mp,off) (((off) >> mp->m_writeio_log) \
<< mp->m_writeio_log)
#define XFS_STRAT_WRITE_IMAPS 2
#define XFS_WRITE_IMAPS XFS_BMAP_MAX_NMAP
-STATIC int
-xfs_imap_to_bmap(
- xfs_inode_t *ip,
- xfs_off_t offset,
- xfs_bmbt_irec_t *imap,
- xfs_iomap_t *iomapp,
- int imaps, /* Number of imap entries */
- int iomaps, /* Number of iomap entries */
- int flags)
-{
- xfs_mount_t *mp = ip->i_mount;
- int pbm;
- xfs_fsblock_t start_block;
-
-
- for (pbm = 0; imaps && pbm < iomaps; imaps--, iomapp++, imap++, pbm++) {
- iomapp->iomap_offset = XFS_FSB_TO_B(mp, imap->br_startoff);
- iomapp->iomap_delta = offset - iomapp->iomap_offset;
- iomapp->iomap_bsize = XFS_FSB_TO_B(mp, imap->br_blockcount);
- iomapp->iomap_flags = flags;
-
- if (XFS_IS_REALTIME_INODE(ip)) {
- iomapp->iomap_flags |= IOMAP_REALTIME;
- iomapp->iomap_target = mp->m_rtdev_targp;
- } else {
- iomapp->iomap_target = mp->m_ddev_targp;
- }
- start_block = imap->br_startblock;
- if (start_block == HOLESTARTBLOCK) {
- iomapp->iomap_bn = IOMAP_DADDR_NULL;
- iomapp->iomap_flags |= IOMAP_HOLE;
- } else if (start_block == DELAYSTARTBLOCK) {
- iomapp->iomap_bn = IOMAP_DADDR_NULL;
- iomapp->iomap_flags |= IOMAP_DELAY;
- } else {
- iomapp->iomap_bn = xfs_fsb_to_db(ip, start_block);
- if (ISUNWRITTEN(imap))
- iomapp->iomap_flags |= IOMAP_UNWRITTEN;
- }
-
- offset += iomapp->iomap_bsize - iomapp->iomap_delta;
- }
- return pbm; /* Return the number filled */
-}
+STATIC int xfs_iomap_write_direct(struct xfs_inode *, xfs_off_t, size_t,
+ int, struct xfs_bmbt_irec *, int *);
+STATIC int xfs_iomap_write_delay(struct xfs_inode *, xfs_off_t, size_t, int,
+ struct xfs_bmbt_irec *, int *);
+STATIC int xfs_iomap_write_allocate(struct xfs_inode *, xfs_off_t, size_t,
+ struct xfs_bmbt_irec *, int *);
int
xfs_iomap(
- xfs_inode_t *ip,
- xfs_off_t offset,
- ssize_t count,
- int flags,
- xfs_iomap_t *iomapp,
- int *niomaps)
+ struct xfs_inode *ip,
+ xfs_off_t offset,
+ ssize_t count,
+ int flags,
+ struct xfs_bmbt_irec *imap,
+ int *nimaps,
+ int *new)
{
- xfs_mount_t *mp = ip->i_mount;
- xfs_fileoff_t offset_fsb, end_fsb;
- int error = 0;
- int lockmode = 0;
- xfs_bmbt_irec_t imap;
- int nimaps = 1;
- int bmapi_flags = 0;
- int iomap_flags = 0;
+ struct xfs_mount *mp = ip->i_mount;
+ xfs_fileoff_t offset_fsb, end_fsb;
+ int error = 0;
+ int lockmode = 0;
+ int bmapi_flags = 0;
ASSERT((ip->i_d.di_mode & S_IFMT) == S_IFREG);
+ *new = 0;
+
if (XFS_FORCED_SHUTDOWN(mp))
return XFS_ERROR(EIO);
+ trace_xfs_iomap_enter(ip, offset, count, flags, NULL);
+
switch (flags & (BMAPI_READ | BMAPI_WRITE | BMAPI_ALLOCATE)) {
case BMAPI_READ:
- xfs_iomap_enter_trace(XFS_IOMAP_READ_ENTER, ip, offset, count);
lockmode = xfs_ilock_map_shared(ip);
bmapi_flags = XFS_BMAPI_ENTIRE;
break;
case BMAPI_WRITE:
- xfs_iomap_enter_trace(XFS_IOMAP_WRITE_ENTER, ip, offset, count);
lockmode = XFS_ILOCK_EXCL;
if (flags & BMAPI_IGNSTATE)
bmapi_flags |= XFS_BMAPI_IGSTATE|XFS_BMAPI_ENTIRE;
xfs_ilock(ip, lockmode);
break;
case BMAPI_ALLOCATE:
- xfs_iomap_enter_trace(XFS_IOMAP_ALLOC_ENTER, ip, offset, count);
lockmode = XFS_ILOCK_SHARED;
bmapi_flags = XFS_BMAPI_ENTIRE;
error = xfs_bmapi(NULL, ip, offset_fsb,
(xfs_filblks_t)(end_fsb - offset_fsb),
- bmapi_flags, NULL, 0, &imap,
- &nimaps, NULL, NULL);
+ bmapi_flags, NULL, 0, imap,
+ nimaps, NULL, NULL);
if (error)
goto out;
switch (flags & (BMAPI_WRITE|BMAPI_ALLOCATE)) {
case BMAPI_WRITE:
/* If we found an extent, return it */
- if (nimaps &&
- (imap.br_startblock != HOLESTARTBLOCK) &&
- (imap.br_startblock != DELAYSTARTBLOCK)) {
- xfs_iomap_map_trace(XFS_IOMAP_WRITE_MAP, ip,
- offset, count, iomapp, &imap, flags);
+ if (*nimaps &&
+ (imap->br_startblock != HOLESTARTBLOCK) &&
+ (imap->br_startblock != DELAYSTARTBLOCK)) {
+ trace_xfs_iomap_found(ip, offset, count, flags, imap);
break;
}
if (flags & (BMAPI_DIRECT|BMAPI_MMAP)) {
error = xfs_iomap_write_direct(ip, offset, count, flags,
- &imap, &nimaps, nimaps);
+ imap, nimaps);
} else {
error = xfs_iomap_write_delay(ip, offset, count, flags,
- &imap, &nimaps);
+ imap, nimaps);
}
if (!error) {
- xfs_iomap_map_trace(XFS_IOMAP_ALLOC_MAP, ip,
- offset, count, iomapp, &imap, flags);
+ trace_xfs_iomap_alloc(ip, offset, count, flags, imap);
}
- iomap_flags = IOMAP_NEW;
+ *new = 1;
break;
case BMAPI_ALLOCATE:
/* If we found an extent, return it */
xfs_iunlock(ip, lockmode);
lockmode = 0;
- if (nimaps && !isnullstartblock(imap.br_startblock)) {
- xfs_iomap_map_trace(XFS_IOMAP_WRITE_MAP, ip,
- offset, count, iomapp, &imap, flags);
+ if (*nimaps && !isnullstartblock(imap->br_startblock)) {
+ trace_xfs_iomap_found(ip, offset, count, flags, imap);
break;
}
error = xfs_iomap_write_allocate(ip, offset, count,
- &imap, &nimaps);
+ imap, nimaps);
break;
}
- if (nimaps) {
- *niomaps = xfs_imap_to_bmap(ip, offset, &imap,
- iomapp, nimaps, *niomaps, iomap_flags);
- } else if (niomaps) {
- *niomaps = 0;
- }
+ ASSERT(*nimaps <= 1);
out:
if (lockmode)
return XFS_ERROR(error);
}
-
STATIC int
xfs_iomap_eof_align_last_fsb(
xfs_mount_t *mp,
}
STATIC int
-xfs_flush_space(
- xfs_inode_t *ip,
- int *fsynced,
- int *ioflags)
-{
- switch (*fsynced) {
- case 0:
- if (ip->i_delayed_blks) {
- xfs_iunlock(ip, XFS_ILOCK_EXCL);
- delay(1);
- xfs_ilock(ip, XFS_ILOCK_EXCL);
- *fsynced = 1;
- } else {
- *ioflags |= BMAPI_SYNC;
- *fsynced = 2;
- }
- return 0;
- case 1:
- *fsynced = 2;
- *ioflags |= BMAPI_SYNC;
- return 0;
- case 2:
- xfs_iunlock(ip, XFS_ILOCK_EXCL);
- xfs_flush_inodes(ip);
- xfs_ilock(ip, XFS_ILOCK_EXCL);
- *fsynced = 3;
- return 0;
- }
- return 1;
-}
-
-STATIC int
xfs_cmn_err_fsblock_zero(
xfs_inode_t *ip,
xfs_bmbt_irec_t *imap)
return EFSCORRUPTED;
}
-int
+STATIC int
xfs_iomap_write_direct(
xfs_inode_t *ip,
xfs_off_t offset,
size_t count,
int flags,
xfs_bmbt_irec_t *ret_imap,
- int *nmaps,
- int found)
+ int *nmaps)
{
xfs_mount_t *mp = ip->i_mount;
xfs_fileoff_t offset_fsb;
* Make sure that the dquots are there. This doesn't hold
* the ilock across a disk read.
*/
- error = XFS_QM_DQATTACH(ip->i_mount, ip, XFS_QMOPT_ILOCKED);
+ error = xfs_qm_dqattach_locked(ip, 0);
if (error)
return XFS_ERROR(error);
if (error)
goto error_out;
} else {
- if (found && (ret_imap->br_startblock == HOLESTARTBLOCK))
+ if (*nmaps && (ret_imap->br_startblock == HOLESTARTBLOCK))
last_fsb = MIN(last_fsb, (xfs_fileoff_t)
ret_imap->br_blockcount +
ret_imap->br_startoff);
if (error)
goto error_out;
- error = XFS_TRANS_RESERVE_QUOTA_NBLKS(mp, tp, ip,
- qblocks, 0, quota_flag);
+ error = xfs_trans_reserve_quota_nblks(tp, ip, qblocks, 0, quota_flag);
if (error)
goto error1;
error0: /* Cancel bmap, unlock inode, unreserve quota blocks, cancel trans */
xfs_bmap_cancel(&free_list);
- XFS_TRANS_UNRESERVE_QUOTA_NBLKS(mp, tp, ip, qblocks, 0, quota_flag);
+ xfs_trans_unreserve_quota_nblks(tp, ip, qblocks, 0, quota_flag);
error1: /* Just cancel transaction */
xfs_trans_cancel(tp, XFS_TRANS_RELEASE_LOG_RES | XFS_TRANS_ABORT);
}
/*
- * If the caller is doing a write at the end of the file,
- * then extend the allocation 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 the caller is doing a write at the end of the file, then extend the
+ * allocation out to the file system's write iosize. We clean up any extra
+ * space left over when the file is closed in xfs_inactive().
*/
STATIC int
xfs_iomap_eof_want_preallocate(
int n, error, imaps;
*prealloc = 0;
- if ((ioflag & BMAPI_SYNC) || (offset + count) <= ip->i_size)
+ if ((offset + count) <= ip->i_size)
return 0;
/*
return 0;
}
-int
+STATIC int
xfs_iomap_write_delay(
xfs_inode_t *ip,
xfs_off_t offset,
xfs_extlen_t extsz;
int nimaps;
xfs_bmbt_irec_t imap[XFS_WRITE_IMAPS];
- int prealloc, fsynced = 0;
+ int prealloc, flushed = 0;
int error;
ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
* 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);
+ error = xfs_qm_dqattach_locked(ip, 0);
if (error)
return XFS_ERROR(error);
extsz = xfs_get_extsz_hint(ip);
offset_fsb = XFS_B_TO_FSBT(mp, offset);
-retry:
error = xfs_iomap_eof_want_preallocate(mp, ip, offset, count,
ioflag, imap, XFS_WRITE_IMAPS, &prealloc);
if (error)
return error;
+retry:
if (prealloc) {
aligned_offset = XFS_WRITEIO_ALIGN(mp, (offset + count - 1));
ioalign = XFS_B_TO_FSBT(mp, aligned_offset);
/*
* If bmapi returned us nothing, and if we didn't get back EDQUOT,
- * then we must have run out of space - flush delalloc, and retry..
+ * then we must have run out of space - flush all other inodes with
+ * delalloc blocks and retry without EOF preallocation.
*/
if (nimaps == 0) {
- xfs_iomap_enter_trace(XFS_IOMAP_WRITE_NOSPACE,
- ip, offset, count);
- if (xfs_flush_space(ip, &fsynced, &ioflag))
+ trace_xfs_delalloc_enospc(ip, offset, count);
+ if (flushed)
return XFS_ERROR(ENOSPC);
+ xfs_iunlock(ip, XFS_ILOCK_EXCL);
+ xfs_flush_inodes(ip);
+ xfs_ilock(ip, XFS_ILOCK_EXCL);
+
+ flushed = 1;
error = 0;
+ prealloc = 0;
goto retry;
}
* We no longer bother to look at the incoming map - all we have to
* guarantee is that whatever we allocate fills the required range.
*/
-int
+STATIC int
xfs_iomap_write_allocate(
xfs_inode_t *ip,
xfs_off_t offset,
/*
* Make sure that the dquots are there.
*/
- if ((error = XFS_QM_DQATTACH(mp, ip, 0)))
+ error = xfs_qm_dqattach(ip, 0);
+ if (error)
return XFS_ERROR(error);
offset_fsb = XFS_B_TO_FSBT(mp, offset);
int committed;
int error;
- xfs_iomap_enter_trace(XFS_IOMAP_UNWRITTEN, ip, offset, count);
+ trace_xfs_unwritten_convert(ip, offset, count);
offset_fsb = XFS_B_TO_FSBT(mp, offset);
count_fsb = XFS_B_TO_FSB(mp, (xfs_ufsize_t)offset + count);
* 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.
+ *
+ * Note that we open code the transaction allocation here
+ * to pass KM_NOFS--we can't risk to recursing back into
+ * the filesystem here as we might be asked to write out
+ * the same inode that we complete here and might deadlock
+ * on the iolock.
*/
- tp = xfs_trans_alloc(mp, XFS_TRANS_STRAT_WRITE);
+ xfs_wait_for_freeze(mp, SB_FREEZE_TRANS);
+ tp = _xfs_trans_alloc(mp, XFS_TRANS_STRAT_WRITE, KM_NOFS);
tp->t_flags |= XFS_TRANS_RESERVE;
error = xfs_trans_reserve(tp, resblks,
XFS_WRITE_LOG_RES(mp), 0,