/*
* Try to retrieve the next record to the left/right from the current one.
*/
STATIC int
xfs_ialloc_next_rec(
struct xfs_btree_cur *cur,
xfs_inobt_rec_incore_t *rec,
int *done,
int left)
{
int error;
int i;
if (left)
error = xfs_btree_decrement(cur, 0, &i);
else
error = xfs_btree_increment(cur, 0, &i);
if (error)
return error;
*done = !i;
if (i) {
error = xfs_inobt_get_rec(cur, rec, &i);
if (error)
return error;
XFS_WANT_CORRUPTED_RETURN(i == 1);
}
return 0;
}
STATIC int
xfs_trim_extents(
struct xfs_mount *mp,
xfs_agnumber_t agno,
xfs_fsblock_t start,
xfs_fsblock_t len,
xfs_fsblock_t minlen,
__uint64_t *blocks_trimmed)
{
struct block_device *bdev = mp->m_ddev_targp->bt_bdev;
struct xfs_btree_cur *cur;
struct xfs_buf *agbp;
struct xfs_perag *pag;
int error;
int i;
pag = xfs_perag_get(mp, agno);
error = xfs_alloc_read_agf(mp, NULL, agno, 0, &agbp);
if (error || !agbp)
goto out_put_perag;
cur = xfs_allocbt_init_cursor(mp, NULL, agbp, agno, XFS_BTNUM_CNT);
/*
* Force out the log. This means any transactions that might have freed
* space before we took the AGF buffer lock are now on disk, and the
* volatile disk cache is flushed.
*/
xfs_log_force(mp, XFS_LOG_SYNC);
/*
* Look up the longest btree in the AGF and start with it.
*/
error = xfs_alloc_lookup_le(cur, 0,
XFS_BUF_TO_AGF(agbp)->agf_longest, &i);
if (error)
goto out_del_cursor;
/*
* Loop until we are done with all extents that are large
* enough to be worth discarding.
*/
while (i) {
xfs_agblock_t fbno;
xfs_extlen_t flen;
error = xfs_alloc_get_rec(cur, &fbno, &flen, &i);
if (error)
goto out_del_cursor;
XFS_WANT_CORRUPTED_GOTO(i == 1, out_del_cursor);
ASSERT(flen <= XFS_BUF_TO_AGF(agbp)->agf_longest);
/*
* Too small? Give up.
*/
if (flen < minlen) {
trace_xfs_discard_toosmall(mp, agno, fbno, flen);
goto out_del_cursor;
}
/*
* If the extent is entirely outside of the range we are
* supposed to discard skip it. Do not bother to trim
* down partially overlapping ranges for now.
*/
if (XFS_AGB_TO_FSB(mp, agno, fbno) + flen < start ||
XFS_AGB_TO_FSB(mp, agno, fbno) >= start + len) {
trace_xfs_discard_exclude(mp, agno, fbno, flen);
goto next_extent;
}
/*
* If any blocks in the range are still busy, skip the
* discard and try again the next time.
*/
if (xfs_alloc_busy_search(mp, agno, fbno, flen)) {
trace_xfs_discard_busy(mp, agno, fbno, flen);
goto next_extent;
}
trace_xfs_discard_extent(mp, agno, fbno, flen);
error = -blkdev_issue_discard(bdev,
XFS_AGB_TO_DADDR(mp, agno, fbno),
XFS_FSB_TO_BB(mp, flen),
GFP_NOFS, 0);
if (error)
goto out_del_cursor;
*blocks_trimmed += flen;
next_extent:
error = xfs_btree_decrement(cur, 0, &i);
if (error)
goto out_del_cursor;
}
out_del_cursor:
xfs_btree_del_cursor(cur, error ? XFS_BTREE_ERROR : XFS_BTREE_NOERROR);
xfs_buf_relse(agbp);
out_put_perag:
xfs_perag_put(pag);
return error;
}
/*
* When we allocate a new block, the first thing we do is add a reference to
* the extent in the rmap btree. This takes the form of a [agbno, length,
* owner, offset] record. Flags are encoded in the high bits of the offset
* field.
*/
STATIC int
xfs_rmap_map(
struct xfs_btree_cur *cur,
xfs_agblock_t bno,
xfs_extlen_t len,
bool unwritten,
struct xfs_owner_info *oinfo)
{
struct xfs_mount *mp = cur->bc_mp;
struct xfs_rmap_irec ltrec;
struct xfs_rmap_irec gtrec;
int have_gt;
int have_lt;
int error = 0;
int i;
uint64_t owner;
uint64_t offset;
unsigned int flags = 0;
bool ignore_off;
xfs_owner_info_unpack(oinfo, &owner, &offset, &flags);
ASSERT(owner != 0);
ignore_off = XFS_RMAP_NON_INODE_OWNER(owner) ||
(flags & XFS_RMAP_BMBT_BLOCK);
if (unwritten)
flags |= XFS_RMAP_UNWRITTEN;
trace_xfs_rmap_map(mp, cur->bc_private.a.agno, bno, len,
unwritten, oinfo);
/*
* For the initial lookup, look for an exact match or the left-adjacent
* record for our insertion point. This will also give us the record for
* start block contiguity tests.
*/
error = xfs_rmap_lookup_le(cur, bno, len, owner, offset, flags,
&have_lt);
if (error)
goto out_error;
XFS_WANT_CORRUPTED_GOTO(mp, have_lt == 1, out_error);
error = xfs_rmap_get_rec(cur, <rec, &have_lt);
if (error)
goto out_error;
XFS_WANT_CORRUPTED_GOTO(mp, have_lt == 1, out_error);
trace_xfs_rmap_lookup_le_range_result(cur->bc_mp,
cur->bc_private.a.agno, ltrec.rm_startblock,
ltrec.rm_blockcount, ltrec.rm_owner,
ltrec.rm_offset, ltrec.rm_flags);
if (!xfs_rmap_is_mergeable(<rec, owner, flags))
have_lt = 0;
XFS_WANT_CORRUPTED_GOTO(mp,
have_lt == 0 ||
ltrec.rm_startblock + ltrec.rm_blockcount <= bno, out_error);
/*
* Increment the cursor to see if we have a right-adjacent record to our
* insertion point. This will give us the record for end block
* contiguity tests.
*/
error = xfs_btree_increment(cur, 0, &have_gt);
if (error)
goto out_error;
if (have_gt) {
error = xfs_rmap_get_rec(cur, >rec, &have_gt);
if (error)
goto out_error;
XFS_WANT_CORRUPTED_GOTO(mp, have_gt == 1, out_error);
XFS_WANT_CORRUPTED_GOTO(mp, bno + len <= gtrec.rm_startblock,
out_error);
trace_xfs_rmap_find_right_neighbor_result(cur->bc_mp,
cur->bc_private.a.agno, gtrec.rm_startblock,
gtrec.rm_blockcount, gtrec.rm_owner,
gtrec.rm_offset, gtrec.rm_flags);
if (!xfs_rmap_is_mergeable(>rec, owner, flags))
have_gt = 0;
}
/*
* Note: cursor currently points one record to the right of ltrec, even
* if there is no record in the tree to the right.
*/
if (have_lt &&
ltrec.rm_startblock + ltrec.rm_blockcount == bno &&
(ignore_off || ltrec.rm_offset + ltrec.rm_blockcount == offset)) {
/*
* left edge contiguous, merge into left record.
*
* ltbno ltlen
* orig: |ooooooooo|
* adding: |aaaaaaaaa|
* result: |rrrrrrrrrrrrrrrrrrr|
* bno len
*/
ltrec.rm_blockcount += len;
if (have_gt &&
bno + len == gtrec.rm_startblock &&
(ignore_off || offset + len == gtrec.rm_offset) &&
(unsigned long)ltrec.rm_blockcount + len +
gtrec.rm_blockcount <= XFS_RMAP_LEN_MAX) {
/*
* right edge also contiguous, delete right record
* and merge into left record.
*
* ltbno ltlen gtbno gtlen
* orig: |ooooooooo| |ooooooooo|
* adding: |aaaaaaaaa|
* result: |rrrrrrrrrrrrrrrrrrrrrrrrrrrrr|
*/
ltrec.rm_blockcount += gtrec.rm_blockcount;
trace_xfs_rmap_delete(mp, cur->bc_private.a.agno,
gtrec.rm_startblock,
gtrec.rm_blockcount,
gtrec.rm_owner,
gtrec.rm_offset,
gtrec.rm_flags);
error = xfs_btree_delete(cur, &i);
if (error)
goto out_error;
XFS_WANT_CORRUPTED_GOTO(mp, i == 1, out_error);
}
/* point the cursor back to the left record and update */
error = xfs_btree_decrement(cur, 0, &have_gt);
if (error)
goto out_error;
error = xfs_rmap_update(cur, <rec);
if (error)
goto out_error;
} else if (have_gt &&
bno + len == gtrec.rm_startblock &&
(ignore_off || offset + len == gtrec.rm_offset)) {
/*
* right edge contiguous, merge into right record.
*
* gtbno gtlen
* Orig: |ooooooooo|
* adding: |aaaaaaaaa|
* Result: |rrrrrrrrrrrrrrrrrrr|
* bno len
*/
gtrec.rm_startblock = bno;
gtrec.rm_blockcount += len;
if (!ignore_off)
gtrec.rm_offset = offset;
error = xfs_rmap_update(cur, >rec);
if (error)
goto out_error;
} else {
/*
* no contiguous edge with identical owner, insert
* new record at current cursor position.
*/
cur->bc_rec.r.rm_startblock = bno;
cur->bc_rec.r.rm_blockcount = len;
cur->bc_rec.r.rm_owner = owner;
cur->bc_rec.r.rm_offset = offset;
cur->bc_rec.r.rm_flags = flags;
trace_xfs_rmap_insert(mp, cur->bc_private.a.agno, bno, len,
owner, offset, flags);
error = xfs_btree_insert(cur, &i);
if (error)
goto out_error;
XFS_WANT_CORRUPTED_GOTO(mp, i == 1, out_error);
}
trace_xfs_rmap_map_done(mp, cur->bc_private.a.agno, bno, len,
unwritten, oinfo);
out_error:
if (error)
trace_xfs_rmap_map_error(mp, cur->bc_private.a.agno,
error, _RET_IP_);
return error;
}
/*
* Convert an unwritten extent to a real extent or vice versa.
* Does not handle overlapping extents.
*/
STATIC int
xfs_rmap_convert(
struct xfs_btree_cur *cur,
xfs_agblock_t bno,
xfs_extlen_t len,
bool unwritten,
struct xfs_owner_info *oinfo)
{
struct xfs_mount *mp = cur->bc_mp;
struct xfs_rmap_irec r[4]; /* neighbor extent entries */
/* left is 0, right is 1, prev is 2 */
/* new is 3 */
uint64_t owner;
uint64_t offset;
uint64_t new_endoff;
unsigned int oldext;
unsigned int newext;
unsigned int flags = 0;
int i;
int state = 0;
int error;
xfs_owner_info_unpack(oinfo, &owner, &offset, &flags);
ASSERT(!(XFS_RMAP_NON_INODE_OWNER(owner) ||
(flags & (XFS_RMAP_ATTR_FORK | XFS_RMAP_BMBT_BLOCK))));
oldext = unwritten ? XFS_RMAP_UNWRITTEN : 0;
new_endoff = offset + len;
trace_xfs_rmap_convert(mp, cur->bc_private.a.agno, bno, len,
unwritten, oinfo);
/*
* For the initial lookup, look for an exact match or the left-adjacent
* record for our insertion point. This will also give us the record for
* start block contiguity tests.
*/
error = xfs_rmap_lookup_le(cur, bno, len, owner, offset, oldext, &i);
if (error)
goto done;
XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
error = xfs_rmap_get_rec(cur, &PREV, &i);
if (error)
goto done;
XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
trace_xfs_rmap_lookup_le_range_result(cur->bc_mp,
cur->bc_private.a.agno, PREV.rm_startblock,
PREV.rm_blockcount, PREV.rm_owner,
PREV.rm_offset, PREV.rm_flags);
ASSERT(PREV.rm_offset <= offset);
ASSERT(PREV.rm_offset + PREV.rm_blockcount >= new_endoff);
ASSERT((PREV.rm_flags & XFS_RMAP_UNWRITTEN) == oldext);
newext = ~oldext & XFS_RMAP_UNWRITTEN;
/*
* Set flags determining what part of the previous oldext allocation
* extent is being replaced by a newext allocation.
*/
if (PREV.rm_offset == offset)
state |= RMAP_LEFT_FILLING;
if (PREV.rm_offset + PREV.rm_blockcount == new_endoff)
state |= RMAP_RIGHT_FILLING;
/*
* Decrement the cursor to see if we have a left-adjacent record to our
* insertion point. This will give us the record for end block
* contiguity tests.
*/
error = xfs_btree_decrement(cur, 0, &i);
if (error)
goto done;
if (i) {
state |= RMAP_LEFT_VALID;
error = xfs_rmap_get_rec(cur, &LEFT, &i);
if (error)
goto done;
XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
XFS_WANT_CORRUPTED_GOTO(mp,
LEFT.rm_startblock + LEFT.rm_blockcount <= bno,
done);
trace_xfs_rmap_find_left_neighbor_result(cur->bc_mp,
cur->bc_private.a.agno, LEFT.rm_startblock,
LEFT.rm_blockcount, LEFT.rm_owner,
LEFT.rm_offset, LEFT.rm_flags);
if (LEFT.rm_startblock + LEFT.rm_blockcount == bno &&
LEFT.rm_offset + LEFT.rm_blockcount == offset &&
xfs_rmap_is_mergeable(&LEFT, owner, newext))
state |= RMAP_LEFT_CONTIG;
}
/*
* Increment the cursor to see if we have a right-adjacent record to our
* insertion point. This will give us the record for end block
* contiguity tests.
*/
error = xfs_btree_increment(cur, 0, &i);
if (error)
goto done;
XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
error = xfs_btree_increment(cur, 0, &i);
if (error)
goto done;
if (i) {
state |= RMAP_RIGHT_VALID;
error = xfs_rmap_get_rec(cur, &RIGHT, &i);
if (error)
goto done;
XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
XFS_WANT_CORRUPTED_GOTO(mp, bno + len <= RIGHT.rm_startblock,
done);
trace_xfs_rmap_find_right_neighbor_result(cur->bc_mp,
cur->bc_private.a.agno, RIGHT.rm_startblock,
RIGHT.rm_blockcount, RIGHT.rm_owner,
RIGHT.rm_offset, RIGHT.rm_flags);
if (bno + len == RIGHT.rm_startblock &&
offset + len == RIGHT.rm_offset &&
xfs_rmap_is_mergeable(&RIGHT, owner, newext))
state |= RMAP_RIGHT_CONTIG;
}
/* check that left + prev + right is not too long */
if ((state & (RMAP_LEFT_FILLING | RMAP_LEFT_CONTIG |
RMAP_RIGHT_FILLING | RMAP_RIGHT_CONTIG)) ==
(RMAP_LEFT_FILLING | RMAP_LEFT_CONTIG |
RMAP_RIGHT_FILLING | RMAP_RIGHT_CONTIG) &&
(unsigned long)LEFT.rm_blockcount + len +
RIGHT.rm_blockcount > XFS_RMAP_LEN_MAX)
state &= ~RMAP_RIGHT_CONTIG;
trace_xfs_rmap_convert_state(mp, cur->bc_private.a.agno, state,
_RET_IP_);
/* reset the cursor back to PREV */
error = xfs_rmap_lookup_le(cur, bno, len, owner, offset, oldext, &i);
if (error)
goto done;
XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
/*
* Switch out based on the FILLING and CONTIG state bits.
*/
switch (state & (RMAP_LEFT_FILLING | RMAP_LEFT_CONTIG |
RMAP_RIGHT_FILLING | RMAP_RIGHT_CONTIG)) {
case RMAP_LEFT_FILLING | RMAP_LEFT_CONTIG |
RMAP_RIGHT_FILLING | RMAP_RIGHT_CONTIG:
/*
* Setting all of a previous oldext extent to newext.
* The left and right neighbors are both contiguous with new.
*/
error = xfs_btree_increment(cur, 0, &i);
if (error)
goto done;
XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
trace_xfs_rmap_delete(mp, cur->bc_private.a.agno,
RIGHT.rm_startblock, RIGHT.rm_blockcount,
RIGHT.rm_owner, RIGHT.rm_offset,
RIGHT.rm_flags);
error = xfs_btree_delete(cur, &i);
if (error)
goto done;
XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
error = xfs_btree_decrement(cur, 0, &i);
if (error)
goto done;
XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
trace_xfs_rmap_delete(mp, cur->bc_private.a.agno,
PREV.rm_startblock, PREV.rm_blockcount,
PREV.rm_owner, PREV.rm_offset,
PREV.rm_flags);
error = xfs_btree_delete(cur, &i);
if (error)
goto done;
XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
error = xfs_btree_decrement(cur, 0, &i);
if (error)
goto done;
XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
NEW = LEFT;
NEW.rm_blockcount += PREV.rm_blockcount + RIGHT.rm_blockcount;
error = xfs_rmap_update(cur, &NEW);
if (error)
goto done;
break;
case RMAP_LEFT_FILLING | RMAP_RIGHT_FILLING | RMAP_LEFT_CONTIG:
/*
* Setting all of a previous oldext extent to newext.
* The left neighbor is contiguous, the right is not.
*/
trace_xfs_rmap_delete(mp, cur->bc_private.a.agno,
PREV.rm_startblock, PREV.rm_blockcount,
PREV.rm_owner, PREV.rm_offset,
PREV.rm_flags);
error = xfs_btree_delete(cur, &i);
if (error)
goto done;
XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
error = xfs_btree_decrement(cur, 0, &i);
if (error)
goto done;
XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
NEW = LEFT;
NEW.rm_blockcount += PREV.rm_blockcount;
error = xfs_rmap_update(cur, &NEW);
if (error)
goto done;
break;
case RMAP_LEFT_FILLING | RMAP_RIGHT_FILLING | RMAP_RIGHT_CONTIG:
/*
* Setting all of a previous oldext extent to newext.
* The right neighbor is contiguous, the left is not.
*/
error = xfs_btree_increment(cur, 0, &i);
if (error)
goto done;
XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
trace_xfs_rmap_delete(mp, cur->bc_private.a.agno,
RIGHT.rm_startblock, RIGHT.rm_blockcount,
RIGHT.rm_owner, RIGHT.rm_offset,
RIGHT.rm_flags);
error = xfs_btree_delete(cur, &i);
if (error)
goto done;
XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
error = xfs_btree_decrement(cur, 0, &i);
if (error)
goto done;
XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
NEW = PREV;
NEW.rm_blockcount = len + RIGHT.rm_blockcount;
NEW.rm_flags = newext;
error = xfs_rmap_update(cur, &NEW);
if (error)
goto done;
break;
case RMAP_LEFT_FILLING | RMAP_RIGHT_FILLING:
/*
* Setting all of a previous oldext extent to newext.
* Neither the left nor right neighbors are contiguous with
* the new one.
*/
NEW = PREV;
NEW.rm_flags = newext;
error = xfs_rmap_update(cur, &NEW);
if (error)
goto done;
break;
case RMAP_LEFT_FILLING | RMAP_LEFT_CONTIG:
/*
* Setting the first part of a previous oldext extent to newext.
* The left neighbor is contiguous.
*/
NEW = PREV;
NEW.rm_offset += len;
NEW.rm_startblock += len;
NEW.rm_blockcount -= len;
error = xfs_rmap_update(cur, &NEW);
if (error)
goto done;
error = xfs_btree_decrement(cur, 0, &i);
if (error)
goto done;
NEW = LEFT;
NEW.rm_blockcount += len;
error = xfs_rmap_update(cur, &NEW);
if (error)
goto done;
break;
case RMAP_LEFT_FILLING:
/*
* Setting the first part of a previous oldext extent to newext.
* The left neighbor is not contiguous.
*/
NEW = PREV;
NEW.rm_startblock += len;
NEW.rm_offset += len;
NEW.rm_blockcount -= len;
error = xfs_rmap_update(cur, &NEW);
if (error)
goto done;
NEW.rm_startblock = bno;
NEW.rm_owner = owner;
NEW.rm_offset = offset;
NEW.rm_blockcount = len;
NEW.rm_flags = newext;
cur->bc_rec.r = NEW;
trace_xfs_rmap_insert(mp, cur->bc_private.a.agno, bno,
len, owner, offset, newext);
error = xfs_btree_insert(cur, &i);
if (error)
goto done;
XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
break;
case RMAP_RIGHT_FILLING | RMAP_RIGHT_CONTIG:
/*
* Setting the last part of a previous oldext extent to newext.
* The right neighbor is contiguous with the new allocation.
*/
NEW = PREV;
NEW.rm_blockcount -= len;
error = xfs_rmap_update(cur, &NEW);
if (error)
goto done;
error = xfs_btree_increment(cur, 0, &i);
if (error)
goto done;
NEW = RIGHT;
NEW.rm_offset = offset;
NEW.rm_startblock = bno;
NEW.rm_blockcount += len;
error = xfs_rmap_update(cur, &NEW);
if (error)
goto done;
break;
case RMAP_RIGHT_FILLING:
/*
* Setting the last part of a previous oldext extent to newext.
* The right neighbor is not contiguous.
*/
NEW = PREV;
NEW.rm_blockcount -= len;
error = xfs_rmap_update(cur, &NEW);
if (error)
goto done;
error = xfs_rmap_lookup_eq(cur, bno, len, owner, offset,
oldext, &i);
if (error)
goto done;
XFS_WANT_CORRUPTED_GOTO(mp, i == 0, done);
NEW.rm_startblock = bno;
NEW.rm_owner = owner;
NEW.rm_offset = offset;
NEW.rm_blockcount = len;
NEW.rm_flags = newext;
cur->bc_rec.r = NEW;
trace_xfs_rmap_insert(mp, cur->bc_private.a.agno, bno,
len, owner, offset, newext);
error = xfs_btree_insert(cur, &i);
if (error)
goto done;
XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
break;
case 0:
/*
* Setting the middle part of a previous oldext extent to
* newext. Contiguity is impossible here.
* One extent becomes three extents.
*/
/* new right extent - oldext */
NEW.rm_startblock = bno + len;
NEW.rm_owner = owner;
NEW.rm_offset = new_endoff;
NEW.rm_blockcount = PREV.rm_offset + PREV.rm_blockcount -
new_endoff;
NEW.rm_flags = PREV.rm_flags;
error = xfs_rmap_update(cur, &NEW);
if (error)
goto done;
/* new left extent - oldext */
NEW = PREV;
NEW.rm_blockcount = offset - PREV.rm_offset;
cur->bc_rec.r = NEW;
trace_xfs_rmap_insert(mp, cur->bc_private.a.agno,
NEW.rm_startblock, NEW.rm_blockcount,
NEW.rm_owner, NEW.rm_offset,
NEW.rm_flags);
error = xfs_btree_insert(cur, &i);
if (error)
goto done;
XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
/*
* Reset the cursor to the position of the new extent
* we are about to insert as we can't trust it after
* the previous insert.
*/
error = xfs_rmap_lookup_eq(cur, bno, len, owner, offset,
oldext, &i);
if (error)
goto done;
XFS_WANT_CORRUPTED_GOTO(mp, i == 0, done);
/* new middle extent - newext */
cur->bc_rec.r.rm_flags &= ~XFS_RMAP_UNWRITTEN;
cur->bc_rec.r.rm_flags |= newext;
trace_xfs_rmap_insert(mp, cur->bc_private.a.agno, bno, len,
owner, offset, newext);
error = xfs_btree_insert(cur, &i);
if (error)
goto done;
XFS_WANT_CORRUPTED_GOTO(mp, i == 1, done);
break;
case RMAP_LEFT_FILLING | RMAP_LEFT_CONTIG | RMAP_RIGHT_CONTIG:
case RMAP_RIGHT_FILLING | RMAP_LEFT_CONTIG | RMAP_RIGHT_CONTIG:
case RMAP_LEFT_FILLING | RMAP_RIGHT_CONTIG:
case RMAP_RIGHT_FILLING | RMAP_LEFT_CONTIG:
case RMAP_LEFT_CONTIG | RMAP_RIGHT_CONTIG:
case RMAP_LEFT_CONTIG:
case RMAP_RIGHT_CONTIG:
/*
* These cases are all impossible.
*/
ASSERT(0);
}
trace_xfs_rmap_convert_done(mp, cur->bc_private.a.agno, bno, len,
unwritten, oinfo);
done:
if (error)
trace_xfs_rmap_convert_error(cur->bc_mp,
cur->bc_private.a.agno, error, _RET_IP_);
return error;
}
/*
* Allocate an inode on disk.
* Mode is used to tell whether the new inode will need space, and whether
* it is a directory.
*
* The arguments IO_agbp and alloc_done are defined to work within
* the constraint of one allocation per transaction.
* xfs_dialloc() is designed to be called twice if it has to do an
* allocation to make more free inodes. On the first call,
* IO_agbp should be set to NULL. If an inode is available,
* i.e., xfs_dialloc() did not need to do an allocation, an inode
* number is returned. In this case, IO_agbp would be set to the
* current ag_buf and alloc_done set to false.
* If an allocation needed to be done, xfs_dialloc would return
* the current ag_buf in IO_agbp and set alloc_done to true.
* The caller should then commit the current transaction, allocate a new
* transaction, and call xfs_dialloc() again, passing in the previous
* value of IO_agbp. IO_agbp should be held across the transactions.
* Since the agbp is locked across the two calls, the second call is
* guaranteed to have a free inode available.
*
* Once we successfully pick an inode its number is returned and the
* on-disk data structures are updated. The inode itself is not read
* in, since doing so would break ordering constraints with xfs_reclaim.
*/
int
xfs_dialloc(
xfs_trans_t *tp, /* transaction pointer */
xfs_ino_t parent, /* parent inode (directory) */
mode_t mode, /* mode bits for new inode */
int okalloc, /* ok to allocate more space */
xfs_buf_t **IO_agbp, /* in/out ag header's buffer */
boolean_t *alloc_done, /* true if we needed to replenish
inode freelist */
xfs_ino_t *inop) /* inode number allocated */
{
xfs_agnumber_t agcount; /* number of allocation groups */
xfs_buf_t *agbp; /* allocation group header's buffer */
xfs_agnumber_t agno; /* allocation group number */
xfs_agi_t *agi; /* allocation group header structure */
xfs_btree_cur_t *cur; /* inode allocation btree cursor */
int error; /* error return value */
int i; /* result code */
int ialloced; /* inode allocation status */
int noroom = 0; /* no space for inode blk allocation */
xfs_ino_t ino; /* fs-relative inode to be returned */
/* REFERENCED */
int j; /* result code */
xfs_mount_t *mp; /* file system mount structure */
int offset; /* index of inode in chunk */
xfs_agino_t pagino; /* parent's a.g. relative inode # */
xfs_agnumber_t pagno; /* parent's allocation group number */
xfs_inobt_rec_incore_t rec; /* inode allocation record */
xfs_agnumber_t tagno; /* testing allocation group number */
xfs_btree_cur_t *tcur; /* temp cursor */
xfs_inobt_rec_incore_t trec; /* temp inode allocation record */
if (*IO_agbp == NULL) {
/*
* We do not have an agbp, so select an initial allocation
* group for inode allocation.
*/
agbp = xfs_ialloc_ag_select(tp, parent, mode, okalloc);
/*
* Couldn't find an allocation group satisfying the
* criteria, give up.
*/
if (!agbp) {
*inop = NULLFSINO;
return 0;
}
agi = XFS_BUF_TO_AGI(agbp);
ASSERT(be32_to_cpu(agi->agi_magicnum) == XFS_AGI_MAGIC);
} else {
/*
* Continue where we left off before. In this case, we
* know that the allocation group has free inodes.
*/
agbp = *IO_agbp;
agi = XFS_BUF_TO_AGI(agbp);
ASSERT(be32_to_cpu(agi->agi_magicnum) == XFS_AGI_MAGIC);
ASSERT(be32_to_cpu(agi->agi_freecount) > 0);
}
mp = tp->t_mountp;
agcount = mp->m_sb.sb_agcount;
agno = be32_to_cpu(agi->agi_seqno);
tagno = agno;
pagno = XFS_INO_TO_AGNO(mp, parent);
pagino = XFS_INO_TO_AGINO(mp, parent);
/*
* If we have already hit the ceiling of inode blocks then clear
* okalloc so we scan all available agi structures for a free
* inode.
*/
if (mp->m_maxicount &&
mp->m_sb.sb_icount + XFS_IALLOC_INODES(mp) > mp->m_maxicount) {
noroom = 1;
okalloc = 0;
}
/*
* Loop until we find an allocation group that either has free inodes
* or in which we can allocate some inodes. Iterate through the
* allocation groups upward, wrapping at the end.
*/
*alloc_done = B_FALSE;
while (!agi->agi_freecount) {
/*
* Don't do anything if we're not supposed to allocate
* any blocks, just go on to the next ag.
*/
if (okalloc) {
/*
* Try to allocate some new inodes in the allocation
* group.
*/
if ((error = xfs_ialloc_ag_alloc(tp, agbp, &ialloced))) {
xfs_trans_brelse(tp, agbp);
if (error == ENOSPC) {
*inop = NULLFSINO;
return 0;
} else
return error;
}
if (ialloced) {
/*
* We successfully allocated some inodes, return
* the current context to the caller so that it
* can commit the current transaction and call
* us again where we left off.
*/
ASSERT(be32_to_cpu(agi->agi_freecount) > 0);
*alloc_done = B_TRUE;
*IO_agbp = agbp;
*inop = NULLFSINO;
return 0;
}
}
/*
* If it failed, give up on this ag.
*/
xfs_trans_brelse(tp, agbp);
/*
* Go on to the next ag: get its ag header.
*/
nextag:
if (++tagno == agcount)
tagno = 0;
if (tagno == agno) {
*inop = NULLFSINO;
return noroom ? ENOSPC : 0;
}
down_read(&mp->m_peraglock);
if (mp->m_perag[tagno].pagi_inodeok == 0) {
up_read(&mp->m_peraglock);
goto nextag;
}
error = xfs_ialloc_read_agi(mp, tp, tagno, &agbp);
up_read(&mp->m_peraglock);
if (error)
goto nextag;
agi = XFS_BUF_TO_AGI(agbp);
ASSERT(be32_to_cpu(agi->agi_magicnum) == XFS_AGI_MAGIC);
}
/*
* Here with an allocation group that has a free inode.
* Reset agno since we may have chosen a new ag in the
* loop above.
*/
agno = tagno;
*IO_agbp = NULL;
cur = xfs_inobt_init_cursor(mp, tp, agbp, be32_to_cpu(agi->agi_seqno));
/*
* If pagino is 0 (this is the root inode allocation) use newino.
* This must work because we've just allocated some.
*/
if (!pagino)
pagino = be32_to_cpu(agi->agi_newino);
#ifdef DEBUG
if (cur->bc_nlevels == 1) {
int freecount = 0;
if ((error = xfs_inobt_lookup_ge(cur, 0, 0, 0, &i)))
goto error0;
XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
do {
if ((error = xfs_inobt_get_rec(cur, &rec.ir_startino,
&rec.ir_freecount, &rec.ir_free, &i)))
goto error0;
XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
freecount += rec.ir_freecount;
if ((error = xfs_btree_increment(cur, 0, &i)))
goto error0;
} while (i == 1);
ASSERT(freecount == be32_to_cpu(agi->agi_freecount) ||
XFS_FORCED_SHUTDOWN(mp));
}
#endif
/*
* If in the same a.g. as the parent, try to get near the parent.
*/
if (pagno == agno) {
if ((error = xfs_inobt_lookup_le(cur, pagino, 0, 0, &i)))
goto error0;
if (i != 0 &&
(error = xfs_inobt_get_rec(cur, &rec.ir_startino,
&rec.ir_freecount, &rec.ir_free, &j)) == 0 &&
j == 1 &&
rec.ir_freecount > 0) {
/*
* Found a free inode in the same chunk
* as parent, done.
*/
}
/*
* In the same a.g. as parent, but parent's chunk is full.
*/
else {
int doneleft; /* done, to the left */
int doneright; /* done, to the right */
if (error)
goto error0;
ASSERT(i == 1);
ASSERT(j == 1);
/*
* Duplicate the cursor, search left & right
* simultaneously.
*/
if ((error = xfs_btree_dup_cursor(cur, &tcur)))
goto error0;
/*
* Search left with tcur, back up 1 record.
*/
if ((error = xfs_btree_decrement(tcur, 0, &i)))
goto error1;
doneleft = !i;
if (!doneleft) {
if ((error = xfs_inobt_get_rec(tcur,
&trec.ir_startino,
&trec.ir_freecount,
&trec.ir_free, &i)))
goto error1;
XFS_WANT_CORRUPTED_GOTO(i == 1, error1);
}
/*
* Search right with cur, go forward 1 record.
*/
if ((error = xfs_btree_increment(cur, 0, &i)))
goto error1;
doneright = !i;
if (!doneright) {
if ((error = xfs_inobt_get_rec(cur,
&rec.ir_startino,
&rec.ir_freecount,
&rec.ir_free, &i)))
goto error1;
XFS_WANT_CORRUPTED_GOTO(i == 1, error1);
}
/*
* Loop until we find the closest inode chunk
* with a free one.
*/
while (!doneleft || !doneright) {
int useleft; /* using left inode
chunk this time */
/*
* Figure out which block is closer,
* if both are valid.
*/
if (!doneleft && !doneright)
useleft =
pagino -
(trec.ir_startino +
XFS_INODES_PER_CHUNK - 1) <
rec.ir_startino - pagino;
else
useleft = !doneleft;
/*
* If checking the left, does it have
* free inodes?
*/
if (useleft && trec.ir_freecount) {
/*
* Yes, set it up as the chunk to use.
*/
rec = trec;
xfs_btree_del_cursor(cur,
XFS_BTREE_NOERROR);
cur = tcur;
break;
}
/*
* If checking the right, does it have
* free inodes?
*/
if (!useleft && rec.ir_freecount) {
/*
* Yes, it's already set up.
*/
xfs_btree_del_cursor(tcur,
XFS_BTREE_NOERROR);
break;
}
/*
* If used the left, get another one
* further left.
*/
if (useleft) {
if ((error = xfs_btree_decrement(tcur, 0,
&i)))
goto error1;
doneleft = !i;
if (!doneleft) {
if ((error = xfs_inobt_get_rec(
tcur,
&trec.ir_startino,
&trec.ir_freecount,
&trec.ir_free, &i)))
goto error1;
XFS_WANT_CORRUPTED_GOTO(i == 1,
error1);
}
}
/*
* If used the right, get another one
* further right.
*/
else {
if ((error = xfs_btree_increment(cur, 0,
&i)))
goto error1;
doneright = !i;
if (!doneright) {
if ((error = xfs_inobt_get_rec(
cur,
&rec.ir_startino,
&rec.ir_freecount,
&rec.ir_free, &i)))
goto error1;
XFS_WANT_CORRUPTED_GOTO(i == 1,
error1);
}
}
}
ASSERT(!doneleft || !doneright);
}
}
/*
* In a different a.g. from the parent.
* See if the most recently allocated block has any free.
*/
else if (be32_to_cpu(agi->agi_newino) != NULLAGINO) {
if ((error = xfs_inobt_lookup_eq(cur,
be32_to_cpu(agi->agi_newino), 0, 0, &i)))
goto error0;
if (i == 1 &&
(error = xfs_inobt_get_rec(cur, &rec.ir_startino,
&rec.ir_freecount, &rec.ir_free, &j)) == 0 &&
j == 1 &&
rec.ir_freecount > 0) {
/*
* The last chunk allocated in the group still has
* a free inode.
*/
}
/*
* None left in the last group, search the whole a.g.
*/
else {
if (error)
goto error0;
if ((error = xfs_inobt_lookup_ge(cur, 0, 0, 0, &i)))
goto error0;
ASSERT(i == 1);
for (;;) {
if ((error = xfs_inobt_get_rec(cur,
&rec.ir_startino,
&rec.ir_freecount, &rec.ir_free,
&i)))
goto error0;
XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
if (rec.ir_freecount > 0)
break;
if ((error = xfs_btree_increment(cur, 0, &i)))
goto error0;
XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
}
}
}
offset = xfs_ialloc_find_free(&rec.ir_free);
ASSERT(offset >= 0);
ASSERT(offset < XFS_INODES_PER_CHUNK);
ASSERT((XFS_AGINO_TO_OFFSET(mp, rec.ir_startino) %
XFS_INODES_PER_CHUNK) == 0);
ino = XFS_AGINO_TO_INO(mp, agno, rec.ir_startino + offset);
XFS_INOBT_CLR_FREE(&rec, offset);
rec.ir_freecount--;
if ((error = xfs_inobt_update(cur, rec.ir_startino, rec.ir_freecount,
rec.ir_free)))
goto error0;
be32_add_cpu(&agi->agi_freecount, -1);
xfs_ialloc_log_agi(tp, agbp, XFS_AGI_FREECOUNT);
down_read(&mp->m_peraglock);
mp->m_perag[tagno].pagi_freecount--;
up_read(&mp->m_peraglock);
#ifdef DEBUG
if (cur->bc_nlevels == 1) {
int freecount = 0;
if ((error = xfs_inobt_lookup_ge(cur, 0, 0, 0, &i)))
goto error0;
do {
if ((error = xfs_inobt_get_rec(cur, &rec.ir_startino,
&rec.ir_freecount, &rec.ir_free, &i)))
goto error0;
XFS_WANT_CORRUPTED_GOTO(i == 1, error0);
freecount += rec.ir_freecount;
if ((error = xfs_btree_increment(cur, 0, &i)))
goto error0;
} while (i == 1);
ASSERT(freecount == be32_to_cpu(agi->agi_freecount) ||
XFS_FORCED_SHUTDOWN(mp));
}
#endif
xfs_btree_del_cursor(cur, XFS_BTREE_NOERROR);
xfs_trans_mod_sb(tp, XFS_TRANS_SB_IFREE, -1);
*inop = ino;
return 0;
error1:
xfs_btree_del_cursor(tcur, XFS_BTREE_ERROR);
error0:
xfs_btree_del_cursor(cur, XFS_BTREE_ERROR);
return error;
}