/*
* Invalidate buffers for per-AG btree blocks we're dumping. This function
* is not intended for use with file data repairs; we have bunmapi for that.
*/
int
xrep_invalidate_blocks(
struct xfs_scrub *sc,
struct xfs_bitmap *bitmap)
{
struct xfs_bitmap_range *bmr;
struct xfs_bitmap_range *n;
struct xfs_buf *bp;
xfs_fsblock_t fsbno;
/*
* For each block in each extent, see if there's an incore buffer for
* exactly that block; if so, invalidate it. The buffer cache only
* lets us look for one buffer at a time, so we have to look one block
* at a time. Avoid invalidating AG headers and post-EOFS blocks
* because we never own those; and if we can't TRYLOCK the buffer we
* assume it's owned by someone else.
*/
for_each_xfs_bitmap_block(fsbno, bmr, n, bitmap) {
/* Skip AG headers and post-EOFS blocks */
if (!xfs_verify_fsbno(sc->mp, fsbno))
continue;
bp = xfs_buf_incore(sc->mp->m_ddev_targp,
XFS_FSB_TO_DADDR(sc->mp, fsbno),
XFS_FSB_TO_BB(sc->mp, 1), XBF_TRYLOCK);
if (bp) {
xfs_trans_bjoin(sc->tp, bp);
xfs_trans_binval(sc->tp, bp);
}
}
return 0;
}
/*
* Indiscriminately delete the entire attribute fork
*
* Recurse (gasp!) through the attribute nodes until we find leaves.
* We're doing a depth-first traversal in order to invalidate everything.
*/
int
xfs_attr3_root_inactive(
struct xfs_trans **trans,
struct xfs_inode *dp)
{
struct xfs_da_blkinfo *info;
struct xfs_buf *bp;
xfs_daddr_t blkno;
int error;
/*
* Read block 0 to see what we have to work with.
* We only get here if we have extents, since we remove
* the extents in reverse order the extent containing
* block 0 must still be there.
*/
error = xfs_da3_node_read(*trans, dp, 0, -1, &bp, XFS_ATTR_FORK);
if (error)
return error;
blkno = bp->b_bn;
/*
* Invalidate the tree, even if the "tree" is only a single leaf block.
* This is a depth-first traversal!
*/
info = bp->b_addr;
switch (info->magic) {
case cpu_to_be16(XFS_DA_NODE_MAGIC):
case cpu_to_be16(XFS_DA3_NODE_MAGIC):
error = xfs_attr3_node_inactive(trans, dp, bp, 1);
break;
case cpu_to_be16(XFS_ATTR_LEAF_MAGIC):
case cpu_to_be16(XFS_ATTR3_LEAF_MAGIC):
error = xfs_attr3_leaf_inactive(trans, dp, bp);
break;
default:
error = -EIO;
xfs_trans_brelse(*trans, bp);
break;
}
if (error)
return error;
/*
* Invalidate the incore copy of the root block.
*/
error = xfs_da_get_buf(*trans, dp, 0, blkno, &bp, XFS_ATTR_FORK);
if (error)
return error;
xfs_trans_binval(*trans, bp); /* remove from cache */
/*
* Commit the invalidate and start the next transaction.
*/
error = xfs_trans_roll(trans, dp);
return error;
}
/*
* Free a symlink that has blocks associated with it.
*/
STATIC int
xfs_inactive_symlink_rmt(
struct xfs_inode *ip)
{
xfs_buf_t *bp;
int committed;
int done;
int error;
xfs_fsblock_t first_block;
xfs_bmap_free_t free_list;
int i;
xfs_mount_t *mp;
xfs_bmbt_irec_t mval[XFS_SYMLINK_MAPS];
int nmaps;
int size;
xfs_trans_t *tp;
mp = ip->i_mount;
ASSERT(ip->i_df.if_flags & XFS_IFEXTENTS);
/*
* We're freeing a symlink that has some
* blocks allocated to it. Free the
* blocks here. We know that we've got
* either 1 or 2 extents and that we can
* free them all in one bunmapi call.
*/
ASSERT(ip->i_d.di_nextents > 0 && ip->i_d.di_nextents <= 2);
tp = xfs_trans_alloc(mp, XFS_TRANS_INACTIVE);
error = xfs_trans_reserve(tp, &M_RES(mp)->tr_itruncate, 0, 0);
if (error) {
xfs_trans_cancel(tp, 0);
return error;
}
xfs_ilock(ip, XFS_ILOCK_EXCL);
xfs_trans_ijoin(tp, ip, 0);
/*
* Lock the inode, fix the size, and join it to the transaction.
* Hold it so in the normal path, we still have it locked for
* the second transaction. In the error paths we need it
* held so the cancel won't rele it, see below.
*/
size = (int)ip->i_d.di_size;
ip->i_d.di_size = 0;
xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
/*
* Find the block(s) so we can inval and unmap them.
*/
done = 0;
xfs_bmap_init(&free_list, &first_block);
nmaps = ARRAY_SIZE(mval);
error = xfs_bmapi_read(ip, 0, xfs_symlink_blocks(mp, size),
mval, &nmaps, 0);
if (error)
goto error_trans_cancel;
/*
* Invalidate the block(s). No validation is done.
*/
for (i = 0; i < nmaps; i++) {
bp = xfs_trans_get_buf(tp, mp->m_ddev_targp,
XFS_FSB_TO_DADDR(mp, mval[i].br_startblock),
XFS_FSB_TO_BB(mp, mval[i].br_blockcount), 0);
if (!bp) {
error = ENOMEM;
goto error_bmap_cancel;
}
xfs_trans_binval(tp, bp);
}
/*
* Unmap the dead block(s) to the free_list.
*/
error = xfs_bunmapi(tp, ip, 0, size, XFS_BMAPI_METADATA, nmaps,
&first_block, &free_list, &done);
if (error)
goto error_bmap_cancel;
ASSERT(done);
/*
* Commit the first transaction. This logs the EFI and the inode.
*/
error = xfs_bmap_finish(&tp, &free_list, &committed);
if (error)
goto error_bmap_cancel;
/*
* The transaction must have been committed, since there were
* actually extents freed by xfs_bunmapi. See xfs_bmap_finish.
* The new tp has the extent freeing and EFDs.
*/
ASSERT(committed);
/*
* The first xact was committed, so add the inode to the new one.
* Mark it dirty so it will be logged and moved forward in the log as
* part of every commit.
*/
xfs_trans_ijoin(tp, ip, 0);
xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
/*
* Commit the transaction containing extent freeing and EFDs.
*/
error = xfs_trans_commit(tp, XFS_TRANS_RELEASE_LOG_RES);
if (error) {
ASSERT(XFS_FORCED_SHUTDOWN(mp));
goto error_unlock;
}
/*
* Remove the memory for extent descriptions (just bookkeeping).
*/
if (ip->i_df.if_bytes)
xfs_idata_realloc(ip, -ip->i_df.if_bytes, XFS_DATA_FORK);
ASSERT(ip->i_df.if_bytes == 0);
xfs_iunlock(ip, XFS_ILOCK_EXCL);
return 0;
error_bmap_cancel:
xfs_bmap_cancel(&free_list);
error_trans_cancel:
xfs_trans_cancel(tp, XFS_TRANS_RELEASE_LOG_RES | XFS_TRANS_ABORT);
error_unlock:
xfs_iunlock(ip, XFS_ILOCK_EXCL);
return error;
}
/*
* Remove the given block from the directory.
* This routine is used for data and free blocks, leaf/node are done
* by xfs_da_shrink_inode.
*/
int
xfs_dir2_shrink_inode(
xfs_da_args_t *args,
xfs_dir2_db_t db,
struct xfs_buf *bp)
{
xfs_fileoff_t bno; /* directory file offset */
xfs_dablk_t da; /* directory file offset */
int done; /* bunmap is finished */
xfs_inode_t *dp;
int error;
xfs_mount_t *mp;
xfs_trans_t *tp;
trace_xfs_dir2_shrink_inode(args, db);
dp = args->dp;
mp = dp->i_mount;
tp = args->trans;
da = xfs_dir2_db_to_da(args->geo, db);
/*
* Unmap the fsblock(s).
*/
if ((error = xfs_bunmapi(tp, dp, da, args->geo->fsbcount,
XFS_BMAPI_METADATA, 0, args->firstblock, args->flist,
&done))) {
/*
* ENOSPC actually can happen if we're in a removename with
* no space reservation, and the resulting block removal
* would cause a bmap btree split or conversion from extents
* to btree. This can only happen for un-fragmented
* directory blocks, since you need to be punching out
* the middle of an extent.
* In this case we need to leave the block in the file,
* and not binval it.
* So the block has to be in a consistent empty state
* and appropriately logged.
* We don't free up the buffer, the caller can tell it
* hasn't happened since it got an error back.
*/
return error;
}
ASSERT(done);
/*
* Invalidate the buffer from the transaction.
*/
xfs_trans_binval(tp, bp);
/*
* If it's not a data block, we're done.
*/
if (db >= xfs_dir2_byte_to_db(args->geo, XFS_DIR2_LEAF_OFFSET))
return 0;
/*
* If the block isn't the last one in the directory, we're done.
*/
if (dp->i_d.di_size > xfs_dir2_db_off_to_byte(args->geo, db + 1, 0))
return 0;
bno = da;
if ((error = xfs_bmap_last_before(tp, dp, &bno, XFS_DATA_FORK))) {
/*
* This can't really happen unless there's kernel corruption.
*/
return error;
}
if (db == args->geo->datablk)
ASSERT(bno == 0);
else
ASSERT(bno > 0);
/*
* Set the size to the new last block.
*/
dp->i_d.di_size = XFS_FSB_TO_B(mp, bno);
xfs_trans_log_inode(tp, dp, XFS_ILOG_CORE);
return 0;
}
/*
* Look at all the extents for this logical region,
* invalidate any buffers that are incore/in transactions.
*/
STATIC int
xfs_attr3_leaf_freextent(
struct xfs_trans **trans,
struct xfs_inode *dp,
xfs_dablk_t blkno,
int blkcnt)
{
struct xfs_bmbt_irec map;
struct xfs_buf *bp;
xfs_dablk_t tblkno;
xfs_daddr_t dblkno;
int tblkcnt;
int dblkcnt;
int nmap;
int error;
/*
* Roll through the "value", invalidating the attribute value's
* blocks.
*/
tblkno = blkno;
tblkcnt = blkcnt;
while (tblkcnt > 0) {
/*
* Try to remember where we decided to put the value.
*/
nmap = 1;
error = xfs_bmapi_read(dp, (xfs_fileoff_t)tblkno, tblkcnt,
&map, &nmap, XFS_BMAPI_ATTRFORK);
if (error) {
return error;
}
ASSERT(nmap == 1);
ASSERT(map.br_startblock != DELAYSTARTBLOCK);
/*
* If it's a hole, these are already unmapped
* so there's nothing to invalidate.
*/
if (map.br_startblock != HOLESTARTBLOCK) {
dblkno = XFS_FSB_TO_DADDR(dp->i_mount,
map.br_startblock);
dblkcnt = XFS_FSB_TO_BB(dp->i_mount,
map.br_blockcount);
bp = xfs_trans_get_buf(*trans,
dp->i_mount->m_ddev_targp,
dblkno, dblkcnt, 0);
if (!bp)
return -ENOMEM;
xfs_trans_binval(*trans, bp);
/*
* Roll to next transaction.
*/
error = xfs_trans_roll(trans, dp);
if (error)
return error;
}
tblkno += map.br_blockcount;
tblkcnt -= map.br_blockcount;
}
return 0;
}
/*
* Recurse (gasp!) through the attribute nodes until we find leaves.
* We're doing a depth-first traversal in order to invalidate everything.
*/
STATIC int
xfs_attr3_node_inactive(
struct xfs_trans **trans,
struct xfs_inode *dp,
struct xfs_buf *bp,
int level)
{
xfs_da_blkinfo_t *info;
xfs_da_intnode_t *node;
xfs_dablk_t child_fsb;
xfs_daddr_t parent_blkno, child_blkno;
int error, i;
struct xfs_buf *child_bp;
struct xfs_da_node_entry *btree;
struct xfs_da3_icnode_hdr ichdr;
/*
* Since this code is recursive (gasp!) we must protect ourselves.
*/
if (level > XFS_DA_NODE_MAXDEPTH) {
xfs_trans_brelse(*trans, bp); /* no locks for later trans */
return -EIO;
}
node = bp->b_addr;
dp->d_ops->node_hdr_from_disk(&ichdr, node);
parent_blkno = bp->b_bn;
if (!ichdr.count) {
xfs_trans_brelse(*trans, bp);
return 0;
}
btree = dp->d_ops->node_tree_p(node);
child_fsb = be32_to_cpu(btree[0].before);
xfs_trans_brelse(*trans, bp); /* no locks for later trans */
/*
* If this is the node level just above the leaves, simply loop
* over the leaves removing all of them. If this is higher up
* in the tree, recurse downward.
*/
for (i = 0; i < ichdr.count; i++) {
/*
* Read the subsidiary block to see what we have to work with.
* Don't do this in a transaction. This is a depth-first
* traversal of the tree so we may deal with many blocks
* before we come back to this one.
*/
error = xfs_da3_node_read(*trans, dp, child_fsb, -2, &child_bp,
XFS_ATTR_FORK);
if (error)
return error;
if (child_bp) {
/* save for re-read later */
child_blkno = XFS_BUF_ADDR(child_bp);
/*
* Invalidate the subtree, however we have to.
*/
info = child_bp->b_addr;
switch (info->magic) {
case cpu_to_be16(XFS_DA_NODE_MAGIC):
case cpu_to_be16(XFS_DA3_NODE_MAGIC):
error = xfs_attr3_node_inactive(trans, dp,
child_bp, level + 1);
break;
case cpu_to_be16(XFS_ATTR_LEAF_MAGIC):
case cpu_to_be16(XFS_ATTR3_LEAF_MAGIC):
error = xfs_attr3_leaf_inactive(trans, dp,
child_bp);
break;
default:
error = -EIO;
xfs_trans_brelse(*trans, child_bp);
break;
}
if (error)
return error;
/*
* Remove the subsidiary block from the cache
* and from the log.
*/
error = xfs_da_get_buf(*trans, dp, 0, child_blkno,
&child_bp, XFS_ATTR_FORK);
if (error)
return error;
xfs_trans_binval(*trans, child_bp);
}
/*
* If we're not done, re-read the parent to get the next
* child block number.
*/
if (i + 1 < ichdr.count) {
error = xfs_da3_node_read(*trans, dp, 0, parent_blkno,
&bp, XFS_ATTR_FORK);
if (error)
return error;
child_fsb = be32_to_cpu(btree[i + 1].before);
xfs_trans_brelse(*trans, bp);
}
/*
* Atomically commit the whole invalidate stuff.
*/
error = xfs_trans_roll(trans, dp);
if (error)
return error;
}
return 0;
}
