Feat: sort changelog and implement DeviceCopy derive in cust_core

Author: RDambrosio016

crates/cust/CHANGELOG.md

@@ -4,6 +4,13 @@ Notable changes to this project will be documented in this file.
 
 ## [Unreleased]
 
+### TLDR
+
+This release is gigantic, so here are the main things you need to worry about:
+
+`Context::create_and_push(FLAGS, device)` -> `Context::new(device)`.
+`Module::from_str(PTX)` -> `Module::from_ptx(PTX, &[])`.
+
 ### Context handling overhaul
 
 The way that contexts are handled in cust has been completely overhauled, it now
@@ -17,63 +24,79 @@ overall simplifying the context handling APIs. This does mean that the API chang
 The old context handling is fully present in `cust::context::legacy` for anyone who needs it for specific reasons. If you use `quick_init` you don't need to worry about
 any breaking changes, the API is the same.
 
-- `Stream::add_callback` now internally uses `cuLaunchHostFunc` anticipating the deprecation and removal of `cuStreamAddCallback` per the driver docs. This does however mean that the function no longer takes a device status as a parameter and does not execute on context error.
+### `cust_core`
+
+`DeviceCopy` has now been split into its own crate, `cust_core`. The crate is `#![no_std]`, which allows you to
+pull in `cust_core` in GPU crates for deriving `DeviceCopy` without cfg shenanigans.
+
+### Removed
+
+- Deleted `DeviceBox::wrap`, use `DeviceBox::from_raw`.
+- Deleted `DeviceSlice::as_ptr` and `DeviceSlice::as_mut_ptr`. Use `DeviceSlice::as_device_ptr` then `DevicePointer::as_(mut)_ptr`.
+- Deleted `DeviceSlice::chunks` and consequently `DeviceChunks`.
+- Deleted `DeviceSlice::chunks_mut` and consequently `DeviceChunksMut`.
+- Deleted `DeviceSlice::from_slice` and `DeviceSlice::from_slice_mut` because it was unsound.
+- Deleted `DevicePointer::as_raw_mut` (use `DevicePointer::as_mut_ptr`).
+- Deleted `DevicePointer::wrap` (use `DevicePointer::from_raw`).
+- `DeviceSlice` no longer implements `Index` and `IndexMut`, switching away from `[T]` made this impossible to implement.
+Instead you can now use `DeviceSlice::index` which behaves the same.
+- `vek` is no longer re-exported.
+
+### Deprecated
+
+- Deprecated `Module::from_str`, use `Module::from_ptx` and pass `&[]` for options.
+`ModuleJitOption::MaxRegisters` does not seem to work currently, but NVIDIA is looking into it.
+- Deprecated `Module::load_from_string`, use `Module::from_ptx_cstr`.
+
+### Added 
+
 - Added `cust::memory::LockedBox`, same as `LockedBuffer` except for single elements.
 - Added `cust::memory::cuda_malloc_async`.
 - Added `cust::memory::cuda_free_async`.
 - Added `impl AsyncCopyDestination<LockedBox<T>> for DeviceBox<T>` for async HtoD memcpy.
 - Added the `bytemuck` feature which is enabled by default.
-- `zeroed` functions on `DeviceBox` and others are no longer unsafe and instead now require `T: Zeroable`. The functions are only available with the `bytemuck` feature.
 - Added `zeroed_async` to `DeviceBox`.
 - Added `drop_async` to `DeviceBox`.
 - Added `new_async` to `DeviceBox`.
-- `Linker::complete` now only returns the built cubin, and not the cubin and a duration.
-- `Stream`, `Module`, `Linker`, `Function`, `Event`, `UnifiedBox`, `ArrayObject`, `LockedBuffer`, `LockedBox`, `DeviceSlice`, `DeviceBuffer`, and `DeviceBox` all now impl `Send` and `Sync`, this makes
-it much easier to write multigpu code. The CUDA API is fully thread-safe except for graph objects.
-- Features such as `vek` for implementing DeviceCopy are now `impl_cratename`, e.g. `impl_vek`, `impl_half`, etc.
-- `DevicePointer::as_raw` now returns a `CUdeviceptr` instead of a `*const T`.
 - Added `DevicePointer::as_ptr` and `DevicePointer::as_mut_ptr` for returning `*const T` or `*mut T`.
 - Added mint integration behind `impl_mint`.
 - Added half integration behind `impl_half`.
 - Added glam integration behind `impl_glam`.
-- `num-complex` integration is now behind `impl_num_complex`, not `num-complex`.
 - Added experimental linux external memory import APIs through `cust::external::ExternalMemory`.
-- `vek` is no longer re-exported.
-- `DeviceBox` now requires `T: DeviceCopy` (previously it didn't but almost all its methods did).
-- `DeviceBox::from_raw` now takes a `CUdeviceptr` instead of a `*mut T`.
-- `DeviceBox::as_device_ptr` now requires `&self` instead of `&mut self`.
-- Deleted `DeviceBox::wrap`, use `DeviceBox::from_raw`.
-- `DeviceBuffer` now requires `T: DeviceCopy`.
-- `DeviceBuffer` is now `repr(C)` and is represented by a `DevicePointer<T>` and a `usize`.
 - Added `DeviceBuffer::as_slice`.
-- `DeviceSlice` now requires `T: DeviceCopy`.
-- `DeviceSlice` is now represented as a `DevicePointer<T>` and a `usize` (and is repr(C)) instead of `[T]` which was definitely unsound.
-- `DeviceSlice::as_ptr` and `DeviceSlice::as_ptr_mut` now both return a `DevicePointer<T>`.
-- Deleted `DeviceSlice::as_ptr` and `DeviceSlice::as_mut_ptr`. Use `DeviceSlice::as_device_ptr` then `DevicePointer::as_(mut)_ptr`.
-- Deleted `DeviceSlice::chunks` and consequently `DeviceChunks`.
-- Deleted `DeviceSlice::chunks_mut` and consequently `DeviceChunksMut`.
-- Deleted `DeviceSlice::from_slice` and `DeviceSlice::from_slice_mut` because it was unsound.
-- `DeviceSlice` no longer implements `Index` and `IndexMut`, switching away from `[T]` made this impossible to implement.
-Instead you can now use `DeviceSlice::index` which behaves the same.
-- `DeviceSlice` is now `Clone` and `Copy`.
 - Added `DeviceVariable`, a simple wrapper around `DeviceBox<T>` and `T` which allows easy management of a CPU and GPU version of a type.
 - Added `DeviceMemory`, a trait describing any region of GPU memory that can be described with a pointer + a length.
 - Added `memcpy_htod`, a wrapper around `cuMemcpyHtoD_v2`.
 - Added `mem_get_info` to query the amount of free and total memory.
-- `DevicePointer::as_raw` now returns a `CUdeviceptr`, not a `*const T` (use `DevicePointer::as_ptr`).
-- Deleted `DevicePointer::as_raw_mut` (use `DevicePointer::as_mut_ptr`).
 - Added `DevicePointer::as_ptr` and `DevicePointer::as_mut_ptr` for `*const T` and `*mut T`.
 - Added `DevicePointer::from_raw` for `CUdeviceptr -> DevicePointer<T>` with a safe function.
-- Deleted `DevicePointer::wrap` (use `DevicePointer::from_raw`).
 - Added dependency on `cust_core` for `DeviceCopy`.
 - Added dependency on `goblin` for verifying cubins and fatbins (impossible to implement safe module loading without it).
-- Deprecated `Module::from_str`, use `Module::from_ptx` and pass `&[]` for options.
 - Added `ModuleJitOption`, `JitFallback`, `JitTarget`, and `OptLevel` for specifying options when loading a module. Note that
-`ModuleJitOption::MaxRegisters` does not seem to work currently, but NVIDIA is looking into it.
 - Added `Module::from_fatbin` and `Module::from_fatbin_unchecked`.
 - Added `Module::from_cubin` and `Module::from_cubin_unchecked`.
 - Added `Module::from_ptr` and `Module::from_ptx_cstr`.
-- Deprecated `Module::load_from_string`, use `Module::from_ptx_cstr`.
+- `Stream`, `Module`, `Linker`, `Function`, `Event`, `UnifiedBox`, `ArrayObject`, `LockedBuffer`, `LockedBox`, `DeviceSlice`, `DeviceBuffer`, and `DeviceBox` all now impl `Send` and `Sync`, this makes
+it much easier to write multigpu code. The CUDA API is fully thread-safe except for graph objects.
+
+### Changed 
+
+- `zeroed` functions on `DeviceBox` and others are no longer unsafe and instead now require `T: Zeroable`. The functions are only available with the `bytemuck` feature.
+- `Stream::add_callback` now internally uses `cuLaunchHostFunc` anticipating the deprecation and removal of `cuStreamAddCallback` per the driver docs. This does however mean that the function no longer takes a device status as a parameter and does not execute on context error.
+- `Linker::complete` now only returns the built cubin, and not the cubin and a duration.
+- Features such as `vek` for implementing DeviceCopy are now `impl_cratename`, e.g. `impl_vek`, `impl_half`, etc.
+- `DevicePointer::as_raw` now returns a `CUdeviceptr` instead of a `*const T`.
+- `num-complex` integration is now behind `impl_num_complex`, not `num-complex`.
+- `DeviceBox` now requires `T: DeviceCopy` (previously it didn't but almost all its methods did).
+- `DeviceBox::from_raw` now takes a `CUdeviceptr` instead of a `*mut T`.
+- `DeviceBox::as_device_ptr` now requires `&self` instead of `&mut self`.
+- `DeviceBuffer` now requires `T: DeviceCopy`.
+- `DeviceBuffer` is now `repr(C)` and is represented by a `DevicePointer<T>` and a `usize`.
+- `DeviceSlice` now requires `T: DeviceCopy`.
+- `DeviceSlice` is now represented as a `DevicePointer<T>` and a `usize` (and is repr(C)) instead of `[T]` which was definitely unsound.
+- `DeviceSlice::as_ptr` and `DeviceSlice::as_ptr_mut` now both return a `DevicePointer<T>`.
+- `DeviceSlice` is now `Clone` and `Copy`.
+- `DevicePointer::as_raw` now returns a `CUdeviceptr`, not a `*const T` (use `DevicePointer::as_ptr`).
 
 ## 0.2.2 - 12/5/21
 

crates/cust/src/lib.rs

@@ -75,9 +75,9 @@ mod surface;
 mod texture;
 pub mod util;
 
-pub use cust_raw as sys;
-
+pub use cust_core;
 pub use cust_derive::DeviceCopy;
+pub use cust_raw as sys;
 
 use crate::context::{Context, ContextFlags};
 use crate::device::Device;

crates/cust/src/memory/mod.rs

@@ -89,7 +89,8 @@ pub use self::unified::*;
 
 use crate::error::*;
 
-pub use cust_core::DeviceCopy;
+pub use crate::DeviceCopy;
+pub use cust_core::_hidden::DeviceCopy;
 
 use std::ffi::c_void;
 

crates/cust_core/Cargo.toml

@@ -9,6 +9,7 @@ glam = { version = "0.20", features=["cuda", "libm"], default-features=false, op
 mint = { version = "^0.5", optional = true }
 half = { version = "1.8", optional = true }
 num-complex = { version = "0.4", optional = true }
+cust_derive = { path = "../cust_derive", version = "0.1" }
 
 [features]
 default = ["vek", "glam", "mint"]