Struct L2TableWriteGuard

pub(super) struct L2TableWriteGuard<'a> {
    table: &'a L2Table,
    _lock: MutexGuard<'a, ()>,
}

Write guard for an L2 table.

Fields

table: &'a L2Table

Referenced L2 table.

_lock: MutexGuard<'a, ()>

Held guard mutex on that L2 table.

Implementations

impl L2TableWriteGuard<'_>

pub fn get_mapping(&self, lookup_cluster: GuestCluster) -> Result<L2Mapping>

Look up a cluster mapping.

pub fn map_cluster( &mut self, index: usize, host_cluster: HostCluster, ) -> Option<(HostCluster, ClusterCount)>

Enter the given raw data cluster mapping into the L2 table.

If the previous entry pointed to an allocated cluster, return the old allocation so its refcount can be decreased (offset of the first cluster and number of clusters – compressed clusters can span across host cluster boundaries).

If the allocation is reused, None is returned, so this function only returns Some(_) if some cluster is indeed leaked.

pub fn zero_cluster( &mut self, index: usize, keep_allocation: bool, ) -> Result<Option<(HostCluster, ClusterCount)>>

Make the given index a zero mapping.

If keep_allocation is true, keep the zero cluster pre-allocated if there is a pre-existing single-cluster allocation (i.e. data cluster or pre-allocated zero cluster). Otherwise, the existing mapping is discarded.

If a previous mapping is discarded, return the old allocation so its refcount can be decreased (offset of the first cluster and number of clusters – compressed clusters can span across host cluster boundaries).

pub fn discard_cluster( &mut self, index: usize, ) -> Option<(HostCluster, ClusterCount)>

Remove the given mapping, leaving it empty.

If a previous mapping is discarded, return the old allocation so its refcount can be decreased (offset of the first cluster and number of clusters – compressed clusters can span across host cluster boundaries).

Trait Implementations

impl<'a> Debug for L2TableWriteGuard<'a>

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.