#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_iocore_t *io,
- xfs_off_t offset,
- ssize_t count)
-{
- xfs_inode_t *ip = XFS_IO_INODE(io);
-
- 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)((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 *)NULL,
- (void *)NULL);
-}
-
-void
-xfs_iomap_map_trace(
- int tag,
- xfs_iocore_t *io,
- xfs_off_t offset,
- ssize_t count,
- xfs_iomap_t *iomapp,
- xfs_bmbt_irec_t *imapp,
- int flags)
-{
- xfs_inode_t *ip = XFS_IO_INODE(io);
-
- 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_iocore_t *io,
- 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;
- xfs_fsize_t nisize;
- int pbm;
- xfs_fsblock_t start_block;
-
- mp = io->io_mount;
- nisize = XFS_SIZE(mp, io);
- if (io->io_new_size > nisize)
- nisize = io->io_new_size;
-
- 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 (io->io_flags & XFS_IOCORE_RT) {
- 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_IO(io, start_block);
- if (ISUNWRITTEN(imap))
- iomapp->iomap_flags |= IOMAP_UNWRITTEN;
- }
-
- if ((iomapp->iomap_offset + iomapp->iomap_bsize) >= nisize) {
- iomapp->iomap_flags |= IOMAP_EOF;
- }
-
- 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_iocore_t *io,
- 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 = io->io_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);
- switch (flags &
- (BMAPI_READ | BMAPI_WRITE | BMAPI_ALLOCATE |
- BMAPI_UNWRITTEN | BMAPI_DEVICE)) {
+ 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, io, offset, count);
- lockmode = XFS_LCK_MAP_SHARED(mp, io);
+ lockmode = xfs_ilock_map_shared(ip);
bmapi_flags = XFS_BMAPI_ENTIRE;
break;
case BMAPI_WRITE:
- xfs_iomap_enter_trace(XFS_IOMAP_WRITE_ENTER, io, offset, count);
- lockmode = XFS_ILOCK_EXCL|XFS_EXTSIZE_WR;
+ lockmode = XFS_ILOCK_EXCL;
if (flags & BMAPI_IGNSTATE)
bmapi_flags |= XFS_BMAPI_IGSTATE|XFS_BMAPI_ENTIRE;
- XFS_ILOCK(mp, io, lockmode);
+ xfs_ilock(ip, lockmode);
break;
case BMAPI_ALLOCATE:
- xfs_iomap_enter_trace(XFS_IOMAP_ALLOC_ENTER, io, offset, count);
- lockmode = XFS_ILOCK_SHARED|XFS_EXTSIZE_RD;
+ lockmode = XFS_ILOCK_SHARED;
bmapi_flags = XFS_BMAPI_ENTIRE;
+
/* Attempt non-blocking lock */
if (flags & BMAPI_TRYLOCK) {
- if (!XFS_ILOCK_NOWAIT(mp, io, lockmode))
+ if (!xfs_ilock_nowait(ip, lockmode))
return XFS_ERROR(EAGAIN);
} else {
- XFS_ILOCK(mp, io, lockmode);
+ xfs_ilock(ip, lockmode);
}
break;
- case BMAPI_UNWRITTEN:
- goto phase2;
- case BMAPI_DEVICE:
- lockmode = XFS_LCK_MAP_SHARED(mp, io);
- iomapp->iomap_target = io->io_flags & XFS_IOCORE_RT ?
- mp->m_rtdev_targp : mp->m_ddev_targp;
- error = 0;
- *niomaps = 1;
- goto out;
default:
BUG();
}
end_fsb = XFS_B_TO_FSB(mp, (xfs_ufsize_t)offset + count);
offset_fsb = XFS_B_TO_FSBT(mp, offset);
- error = XFS_BMAPI(mp, NULL, io, offset_fsb,
+ 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;
-phase2:
- switch (flags & (BMAPI_WRITE|BMAPI_ALLOCATE|BMAPI_UNWRITTEN)) {
+ 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, io,
- 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(mp, io, offset,
- count, flags, &imap, &nimaps, nimaps);
+ error = xfs_iomap_write_direct(ip, offset, count, flags,
+ imap, nimaps);
} else {
- error = XFS_IOMAP_WRITE_DELAY(mp, io, offset, count,
- flags, &imap, &nimaps);
+ error = xfs_iomap_write_delay(ip, offset, count, flags,
+ imap, nimaps);
}
if (!error) {
- xfs_iomap_map_trace(XFS_IOMAP_ALLOC_MAP, io,
- 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(mp, io, lockmode);
+ xfs_iunlock(ip, lockmode);
lockmode = 0;
- if (nimaps && !ISNULLSTARTBLOCK(imap.br_startblock)) {
- xfs_iomap_map_trace(XFS_IOMAP_WRITE_MAP, io,
- 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(mp, io, offset, count,
- &imap, &nimaps);
- break;
- case BMAPI_UNWRITTEN:
- lockmode = 0;
- error = XFS_IOMAP_WRITE_UNWRITTEN(mp, io, offset, count);
- nimaps = 0;
+ error = xfs_iomap_write_allocate(ip, offset, count,
+ imap, nimaps);
break;
}
- if (nimaps) {
- *niomaps = xfs_imap_to_bmap(io, offset, &imap,
- iomapp, nimaps, *niomaps, iomap_flags);
- } else if (niomaps) {
- *niomaps = 0;
- }
+ ASSERT(*nimaps <= 1);
out:
if (lockmode)
- XFS_IUNLOCK(mp, io, lockmode);
+ xfs_iunlock(ip, lockmode);
return XFS_ERROR(error);
}
STATIC int
xfs_iomap_eof_align_last_fsb(
xfs_mount_t *mp,
- xfs_iocore_t *io,
- xfs_fsize_t isize,
+ xfs_inode_t *ip,
xfs_extlen_t extsize,
xfs_fileoff_t *last_fsb)
{
xfs_extlen_t align;
int eof, error;
- if (io->io_flags & XFS_IOCORE_RT)
+ if (XFS_IS_REALTIME_INODE(ip))
;
/*
* If mounted with the "-o swalloc" option, roundup the allocation
* stripe width and we are allocating past the allocation eof.
*/
else if (mp->m_swidth && (mp->m_flags & XFS_MOUNT_SWALLOC) &&
- (isize >= XFS_FSB_TO_B(mp, mp->m_swidth)))
+ (ip->i_size >= XFS_FSB_TO_B(mp, mp->m_swidth)))
new_last_fsb = roundup_64(*last_fsb, mp->m_swidth);
/*
* 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 (mp->m_dalign && (isize >= XFS_FSB_TO_B(mp, mp->m_dalign)))
+ else if (mp->m_dalign && (ip->i_size >= XFS_FSB_TO_B(mp, mp->m_dalign)))
new_last_fsb = roundup_64(*last_fsb, mp->m_dalign);
/*
}
if (new_last_fsb) {
- error = XFS_BMAP_EOF(mp, io, new_last_fsb, XFS_DATA_FORK, &eof);
+ error = xfs_bmap_eof(ip, new_last_fsb, XFS_DATA_FORK, &eof);
if (error)
return error;
if (eof)
}
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);
- xfs_flush_inode(ip);
- 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_device(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_iocore_t *io = &ip->i_iocore;
xfs_fileoff_t offset_fsb;
xfs_fileoff_t last_fsb;
xfs_filblks_t count_fsb, resaligned;
xfs_fsblock_t firstfsb;
xfs_extlen_t extsz, temp;
- xfs_fsize_t isize;
int nimaps;
int bmapi_flag;
int quota_flag;
* 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);
rt = XFS_IS_REALTIME_INODE(ip);
- if (unlikely(rt)) {
- if (!(extsz = ip->i_d.di_extsize))
- extsz = mp->m_sb.sb_rextsize;
- } else {
- extsz = ip->i_d.di_extsize;
- }
-
- isize = ip->i_size;
- if (io->io_new_size > isize)
- isize = io->io_new_size;
+ extsz = xfs_get_extsz_hint(ip);
- offset_fsb = XFS_B_TO_FSBT(mp, offset);
- last_fsb = XFS_B_TO_FSB(mp, ((xfs_ufsize_t)(offset + count)));
- if ((offset + count) > isize) {
- error = xfs_iomap_eof_align_last_fsb(mp, io, isize, extsz,
- &last_fsb);
+ offset_fsb = XFS_B_TO_FSBT(mp, offset);
+ last_fsb = XFS_B_TO_FSB(mp, ((xfs_ufsize_t)(offset + count)));
+ if ((offset + count) > ip->i_size) {
+ error = xfs_iomap_eof_align_last_fsb(mp, ip, extsz, &last_fsb);
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;
/*
* Issue the xfs_bmapi() call to allocate the blocks
*/
- XFS_BMAP_INIT(&free_list, &firstfsb);
+ xfs_bmap_init(&free_list, &firstfsb);
nimaps = 1;
- error = XFS_BMAPI(mp, tp, io, offset_fsb, count_fsb, bmapi_flag,
+ error = xfs_bmapi(tp, ip, offset_fsb, count_fsb, bmapi_flag,
&firstfsb, 0, &imap, &nimaps, &free_list, NULL);
if (error)
goto error0;
goto error_out;
}
- if (unlikely(!imap.br_startblock && !(io->io_flags & XFS_IOCORE_RT))) {
+ if (!(imap.br_startblock || XFS_IS_REALTIME_INODE(ip))) {
error = xfs_cmn_err_fsblock_zero(ip, &imap);
goto error_out;
}
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(
xfs_mount_t *mp,
- xfs_iocore_t *io,
- xfs_fsize_t isize,
+ xfs_inode_t *ip,
xfs_off_t offset,
size_t count,
int ioflag,
int n, error, imaps;
*prealloc = 0;
- if ((ioflag & BMAPI_SYNC) || (offset + count) <= isize)
+ if ((offset + count) <= ip->i_size)
return 0;
/*
while (count_fsb > 0) {
imaps = nimaps;
firstblock = NULLFSBLOCK;
- error = XFS_BMAPI(mp, NULL, io, start_fsb, count_fsb, 0,
+ error = xfs_bmapi(NULL, ip, start_fsb, count_fsb, 0,
&firstblock, 0, imap, &imaps, NULL, NULL);
if (error)
return error;
return 0;
}
-int
+STATIC int
xfs_iomap_write_delay(
xfs_inode_t *ip,
xfs_off_t offset,
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_off_t aligned_offset;
xfs_fileoff_t ioalign;
xfs_fsblock_t firstblock;
xfs_extlen_t extsz;
- xfs_fsize_t isize;
int nimaps;
xfs_bmbt_irec_t imap[XFS_WRITE_IMAPS];
- int prealloc, fsynced = 0;
+ int prealloc, flushed = 0;
int error;
- ASSERT(ismrlocked(&ip->i_lock, MR_UPDATE) != 0);
+ 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);
- if (XFS_IS_REALTIME_INODE(ip)) {
- if (!(extsz = ip->i_d.di_extsize))
- extsz = mp->m_sb.sb_rextsize;
- } else {
- extsz = ip->i_d.di_extsize;
- }
-
+ extsz = xfs_get_extsz_hint(ip);
offset_fsb = XFS_B_TO_FSBT(mp, offset);
-retry:
- isize = ip->i_size;
- if (io->io_new_size > isize)
- isize = io->io_new_size;
-
- error = xfs_iomap_eof_want_preallocate(mp, io, isize, offset, count,
+ 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 (prealloc || extsz) {
- error = xfs_iomap_eof_align_last_fsb(mp, io, isize, extsz,
- &last_fsb);
+ error = xfs_iomap_eof_align_last_fsb(mp, ip, extsz, &last_fsb);
if (error)
return error;
}
nimaps = XFS_WRITE_IMAPS;
firstblock = NULLFSBLOCK;
- error = XFS_BMAPI(mp, NULL, io, offset_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,
/*
* 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,
- io, 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;
}
- if (unlikely(!imap[0].br_startblock && !(io->io_flags & XFS_IOCORE_RT)))
+ if (!(imap[0].br_startblock || XFS_IS_REALTIME_INODE(ip)))
return xfs_cmn_err_fsblock_zero(ip, &imap[0]);
*ret_imap = imap[0];
* the originating callers request.
*
* Called without a lock on the inode.
+ *
+ * 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,
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_bmbt_irec_t imap;
xfs_trans_t *tp;
- int i, nimaps, committed;
+ int nimaps, committed;
int error = 0;
int nres;
/*
* 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);
xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL);
xfs_trans_ihold(tp, ip);
- XFS_BMAP_INIT(&free_list, &first_block);
+ 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.
+ * it is possible that the extents have changed since
+ * we did the read call as we dropped the ilock for a
+ * while. We have to be careful about truncates or hole
+ * punchs here - we are not allowed to allocate
+ * non-delalloc blocks here.
+ *
+ * The only protection against truncation is the pages
+ * for the range we are being asked to convert are
+ * locked and hence a truncate will block on them
+ * first.
+ *
+ * As a result, if we go beyond the range we really
+ * need and hit an delalloc extent boundary followed by
+ * a hole while we have excess blocks in the map, we
+ * will fill the hole incorrectly and overrun the
+ * transaction reservation.
+ *
+ * Using a single map prevents this as we are forced to
+ * check each map we look for overlap with the desired
+ * range and abort as soon as we find it. Also, given
+ * that we only return a single map, having one beyond
+ * what we can return is probably a bit silly.
+ *
+ * We also need to check that we don't go beyond EOF;
+ * this is a truncate optimisation as a truncate sets
+ * the new file size before block on the pages we
+ * currently have locked under writeback. Because they
+ * are about to be tossed, we don't need to write them
+ * back....
*/
-
+ nimaps = 1;
end_fsb = XFS_B_TO_FSB(mp, ip->i_size);
- xfs_bmap_last_offset(NULL, ip, &last_block,
- XFS_DATA_FORK);
+ error = xfs_bmap_last_offset(NULL, ip, &last_block,
+ XFS_DATA_FORK);
+ if (error)
+ goto trans_cancel;
+
last_block = XFS_FILEOFF_MAX(last_block, end_fsb);
if ((map_start_fsb + count_fsb) > last_block) {
count_fsb = last_block - map_start_fsb;
}
/* Go get the actual blocks */
- error = XFS_BMAPI(mp, tp, io, map_start_fsb, count_fsb,
+ error = xfs_bmapi(tp, ip, map_start_fsb, count_fsb,
XFS_BMAPI_WRITE, &first_block, 1,
- imap, &nimaps, &free_list, NULL);
+ &imap, &nimaps, &free_list, NULL);
if (error)
goto trans_cancel;
* 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 (unlikely(!imap[i].br_startblock &&
- !(io->io_flags & XFS_IOCORE_RT)))
- return xfs_cmn_err_fsblock_zero(ip, &imap[i]);
- if ((offset_fsb >= imap[i].br_startoff) &&
- (offset_fsb < (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;
+ if (!(imap.br_startblock || XFS_IS_REALTIME_INODE(ip)))
+ return xfs_cmn_err_fsblock_zero(ip, &imap);
+
+ if ((offset_fsb >= imap.br_startoff) &&
+ (offset_fsb < (imap.br_startoff +
+ imap.br_blockcount))) {
+ *map = imap;
+ *retmap = 1;
+ XFS_STATS_INC(xs_xstrat_quick);
+ return 0;
}
- /* So far we have not mapped the requested part of the
+ /*
+ * So far we have not mapped the requested part of the
* file, just surrounding data, try again.
*/
- nimaps--;
- map_start_fsb = imap[nimaps].br_startoff +
- imap[nimaps].br_blockcount;
+ count_fsb -= imap.br_blockcount;
+ map_start_fsb = imap.br_startoff + imap.br_blockcount;
}
trans_cancel:
size_t count)
{
xfs_mount_t *mp = ip->i_mount;
- xfs_iocore_t *io = &ip->i_iocore;
xfs_fileoff_t offset_fsb;
xfs_filblks_t count_fsb;
xfs_filblks_t numblks_fsb;
int committed;
int error;
- xfs_iomap_enter_trace(XFS_IOMAP_UNWRITTEN,
- &ip->i_iocore, 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);
count_fsb = (xfs_filblks_t)(count_fsb - offset_fsb);
+ /*
+ * Reserve enough blocks in this transaction for two complete extent
+ * btree splits. We may be converting the middle part of an unwritten
+ * extent and in this case we will insert two new extents in the btree
+ * each of which could cause a full split.
+ *
+ * This reservation amount will be used in the first call to
+ * xfs_bmbt_split() to select an AG with enough space to satisfy the
+ * rest of the operation.
+ */
resblks = XFS_DIOSTRAT_SPACE_RES(mp, 0) << 1;
do {
* 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,
/*
* Modify the unwritten extent state of the buffer.
*/
- XFS_BMAP_INIT(&free_list, &firstfsb);
+ xfs_bmap_init(&free_list, &firstfsb);
nimaps = 1;
- error = XFS_BMAPI(mp, tp, io, offset_fsb, count_fsb,
+ error = xfs_bmapi(tp, ip, offset_fsb, count_fsb,
XFS_BMAPI_WRITE|XFS_BMAPI_CONVERT, &firstfsb,
1, &imap, &nimaps, &free_list, NULL);
if (error)
if (error)
return XFS_ERROR(error);
- if (unlikely(!imap.br_startblock &&
- !(io->io_flags & XFS_IOCORE_RT)))
+ if (!(imap.br_startblock || XFS_IS_REALTIME_INODE(ip)))
return xfs_cmn_err_fsblock_zero(ip, &imap);
if ((numblks_fsb = imap.br_blockcount) == 0) {