wehub-resource-sync
50 lines
1.5 KiB
Python
50 lines
1.5 KiB
Python
# SPDX-License-Identifier: Apache-2.0
|
|
"""Platform-abstraction base classes.
|
|
|
|
This module defines abstract base classes for platform-specific capabilities.
|
|
:class:`PinMemoryBackend` is the first; future platform abstractions should
|
|
be added here as well.
|
|
"""
|
|
|
|
|
|
class PinMemoryBackend:
|
|
"""Base class for host-memory pinning per platform.
|
|
|
|
The default implementation is a no-op that always returns ``False``,
|
|
so platforms that do not support pinning do not need to subclass this.
|
|
"""
|
|
|
|
def pin_memory(self, ptr: int, size: int, flags: int = 0) -> bool:
|
|
"""Pin a host memory region for DMA access.
|
|
|
|
Args:
|
|
ptr: Raw pointer (data_ptr) to the memory region.
|
|
size: Size in bytes of the region to pin.
|
|
flags: Platform-specific registration flags (e.g.
|
|
``cudaHostRegisterDefault = 0``).
|
|
|
|
Returns:
|
|
True if pinning succeeded, False otherwise.
|
|
"""
|
|
return False
|
|
|
|
def unpin_memory(self, ptr: int) -> bool:
|
|
"""Unpin a previously pinned host memory region.
|
|
|
|
Args:
|
|
ptr: Raw pointer (data_ptr) to the memory region.
|
|
|
|
Returns:
|
|
True if unpinning succeeded, False otherwise.
|
|
"""
|
|
return False
|
|
|
|
@property
|
|
def is_pin_supported(self) -> bool:
|
|
"""Whether the current platform supports memory pinning.
|
|
|
|
Returns:
|
|
True if pinning is supported, False otherwise.
|
|
"""
|
|
return False
|