pub unsafe trait Allocator {
fn allocate(&self, layout: Layout) -> Result<NonNull<[u8]>, AllocError>;
unsafe fn deallocate(&self, ptr: NonNull<u8>, layout: Layout);
fn allocate_zeroed(
&self,
layout: Layout
) -> Result<NonNull<[u8]>, AllocError> { ... }
unsafe fn grow(
&self,
ptr: NonNull<u8>,
old_layout: Layout,
new_layout: Layout
) -> Result<NonNull<[u8]>, AllocError> { ... }
unsafe fn grow_zeroed(
&self,
ptr: NonNull<u8>,
old_layout: Layout,
new_layout: Layout
) -> Result<NonNull<[u8]>, AllocError> { ... }
unsafe fn shrink(
&self,
ptr: NonNull<u8>,
old_layout: Layout,
new_layout: Layout
) -> Result<NonNull<[u8]>, AllocError> { ... }
fn by_ref(&self) -> &Self
where
Self: Sized,
{ ... }
}
allocator_api
#32838)Expand description
An implementation of Allocator
can allocate, grow, shrink, and deallocate arbitrary blocks of
data described via Layout
.
Allocator
is designed to be implemented on ZSTs, references, or smart pointers because having
an allocator like MyAlloc([u8; N])
cannot be moved, without updating the pointers to the
allocated memory.
Unlike GlobalAlloc
, zero-sized allocations are allowed in Allocator
. If an underlying
allocator does not support this (like jemalloc) or return a null pointer (such as
libc::malloc
), this must be caught by the implementation.
Currently allocated memory
Some of the methods require that a memory block be currently allocated via an allocator. This means that:
-
the starting address for that memory block was previously returned by
allocate
,grow
, orshrink
, and -
the memory block has not been subsequently deallocated, where blocks are either deallocated directly by being passed to
deallocate
or were changed by being passed togrow
orshrink
that returnsOk
. Ifgrow
orshrink
have returnedErr
, the passed pointer remains valid.
Memory fitting
Some of the methods require that a layout fit a memory block. What it means for a layout to “fit” a memory block means (or equivalently, for a memory block to “fit” a layout) is that the following conditions must hold:
-
The block must be allocated with the same alignment as
layout.align()
, and -
The provided
layout.size()
must fall in the rangemin ..= max
, where:
Safety
-
Memory blocks returned from an allocator must point to valid memory and retain their validity until the instance and all of its clones are dropped,
-
cloning or moving the allocator must not invalidate memory blocks returned from this allocator. A cloned allocator must behave like the same allocator, and
-
any pointer to a memory block which is currently allocated may be passed to any other method of the allocator.
Required Methods§
sourcefn allocate(&self, layout: Layout) -> Result<NonNull<[u8]>, AllocError>
fn allocate(&self, layout: Layout) -> Result<NonNull<[u8]>, AllocError>
allocator_api
#32838)Attempts to allocate a block of memory.
On success, returns a NonNull<[u8]>
meeting the size and alignment guarantees of layout
.
The returned block may have a larger size than specified by layout.size()
, and may or may
not have its contents initialized.
Errors
Returning Err
indicates that either memory is exhausted or layout
does not meet
allocator’s size or alignment constraints.
Implementations are encouraged to return Err
on memory exhaustion rather than panicking or
aborting, but this is not a strict requirement. (Specifically: it is legal to implement
this trait atop an underlying native allocation library that aborts on memory exhaustion.)
Clients wishing to abort computation in response to an allocation error are encouraged to
call the handle_alloc_error
function, rather than directly invoking panic!
or similar.
sourceunsafe fn deallocate(&self, ptr: NonNull<u8>, layout: Layout)
unsafe fn deallocate(&self, ptr: NonNull<u8>, layout: Layout)
allocator_api
#32838)Deallocates the memory referenced by ptr
.
Safety
ptr
must denote a block of memory currently allocated via this allocator, andlayout
must fit that block of memory.
Provided Methods§
sourcefn allocate_zeroed(&self, layout: Layout) -> Result<NonNull<[u8]>, AllocError>
fn allocate_zeroed(&self, layout: Layout) -> Result<NonNull<[u8]>, AllocError>
allocator_api
#32838)Behaves like allocate
, but also ensures that the returned memory is zero-initialized.
Errors
Returning Err
indicates that either memory is exhausted or layout
does not meet
allocator’s size or alignment constraints.
Implementations are encouraged to return Err
on memory exhaustion rather than panicking or
aborting, but this is not a strict requirement. (Specifically: it is legal to implement
this trait atop an underlying native allocation library that aborts on memory exhaustion.)
Clients wishing to abort computation in response to an allocation error are encouraged to
call the handle_alloc_error
function, rather than directly invoking panic!
or similar.
sourceunsafe fn grow(
&self,
ptr: NonNull<u8>,
old_layout: Layout,
new_layout: Layout
) -> Result<NonNull<[u8]>, AllocError>
unsafe fn grow(
&self,
ptr: NonNull<u8>,
old_layout: Layout,
new_layout: Layout
) -> Result<NonNull<[u8]>, AllocError>
allocator_api
#32838)Attempts to extend the memory block.
Returns a new NonNull<[u8]>
containing a pointer and the actual size of the allocated
memory. The pointer is suitable for holding data described by new_layout
. To accomplish
this, the allocator may extend the allocation referenced by ptr
to fit the new layout.
If this returns Ok
, then ownership of the memory block referenced by ptr
has been
transferred to this allocator. Any access to the old ptr
is Undefined Behavior, even if the
allocation was grown in-place. The newly returned pointer is the only valid pointer
for accessing this memory now.
If this method returns Err
, then ownership of the memory block has not been transferred to
this allocator, and the contents of the memory block are unaltered.
Safety
ptr
must denote a block of memory currently allocated via this allocator.old_layout
must fit that block of memory (Thenew_layout
argument need not fit it.).new_layout.size()
must be greater than or equal toold_layout.size()
.
Note that new_layout.align()
need not be the same as old_layout.align()
.
Errors
Returns Err
if the new layout does not meet the allocator’s size and alignment
constraints of the allocator, or if growing otherwise fails.
Implementations are encouraged to return Err
on memory exhaustion rather than panicking or
aborting, but this is not a strict requirement. (Specifically: it is legal to implement
this trait atop an underlying native allocation library that aborts on memory exhaustion.)
Clients wishing to abort computation in response to an allocation error are encouraged to
call the handle_alloc_error
function, rather than directly invoking panic!
or similar.
sourceunsafe fn grow_zeroed(
&self,
ptr: NonNull<u8>,
old_layout: Layout,
new_layout: Layout
) -> Result<NonNull<[u8]>, AllocError>
unsafe fn grow_zeroed(
&self,
ptr: NonNull<u8>,
old_layout: Layout,
new_layout: Layout
) -> Result<NonNull<[u8]>, AllocError>
allocator_api
#32838)Behaves like grow
, but also ensures that the new contents are set to zero before being
returned.
The memory block will contain the following contents after a successful call to
grow_zeroed
:
- Bytes
0..old_layout.size()
are preserved from the original allocation. - Bytes
old_layout.size()..old_size
will either be preserved or zeroed, depending on the allocator implementation.old_size
refers to the size of the memory block prior to thegrow_zeroed
call, which may be larger than the size that was originally requested when it was allocated. - Bytes
old_size..new_size
are zeroed.new_size
refers to the size of the memory block returned by thegrow_zeroed
call.
Safety
ptr
must denote a block of memory currently allocated via this allocator.old_layout
must fit that block of memory (Thenew_layout
argument need not fit it.).new_layout.size()
must be greater than or equal toold_layout.size()
.
Note that new_layout.align()
need not be the same as old_layout.align()
.
Errors
Returns Err
if the new layout does not meet the allocator’s size and alignment
constraints of the allocator, or if growing otherwise fails.
Implementations are encouraged to return Err
on memory exhaustion rather than panicking or
aborting, but this is not a strict requirement. (Specifically: it is legal to implement
this trait atop an underlying native allocation library that aborts on memory exhaustion.)
Clients wishing to abort computation in response to an allocation error are encouraged to
call the handle_alloc_error
function, rather than directly invoking panic!
or similar.
sourceunsafe fn shrink(
&self,
ptr: NonNull<u8>,
old_layout: Layout,
new_layout: Layout
) -> Result<NonNull<[u8]>, AllocError>
unsafe fn shrink(
&self,
ptr: NonNull<u8>,
old_layout: Layout,
new_layout: Layout
) -> Result<NonNull<[u8]>, AllocError>
allocator_api
#32838)Attempts to shrink the memory block.
Returns a new NonNull<[u8]>
containing a pointer and the actual size of the allocated
memory. The pointer is suitable for holding data described by new_layout
. To accomplish
this, the allocator may shrink the allocation referenced by ptr
to fit the new layout.
If this returns Ok
, then ownership of the memory block referenced by ptr
has been
transferred to this allocator. Any access to the old ptr
is Undefined Behavior, even if the
allocation was shrunk in-place. The newly returned pointer is the only valid pointer
for accessing this memory now.
If this method returns Err
, then ownership of the memory block has not been transferred to
this allocator, and the contents of the memory block are unaltered.
Safety
ptr
must denote a block of memory currently allocated via this allocator.old_layout
must fit that block of memory (Thenew_layout
argument need not fit it.).new_layout.size()
must be smaller than or equal toold_layout.size()
.
Note that new_layout.align()
need not be the same as old_layout.align()
.
Errors
Returns Err
if the new layout does not meet the allocator’s size and alignment
constraints of the allocator, or if shrinking otherwise fails.
Implementations are encouraged to return Err
on memory exhaustion rather than panicking or
aborting, but this is not a strict requirement. (Specifically: it is legal to implement
this trait atop an underlying native allocation library that aborts on memory exhaustion.)
Clients wishing to abort computation in response to an allocation error are encouraged to
call the handle_alloc_error
function, rather than directly invoking panic!
or similar.