On Thu, Jul 09, 2026 at 02:36:44PM -0700, Deborah Brouwer wrote:
> From: Boris Brezillon <[email protected]>
>
> Add GPU virtual address space management using the DRM GPUVM framework.
> Each virtual memory (VM) space is backed by ARM64 LPAE Stage 1 page tables
> and can be mapped into hardware address space (AS) slots for GPU execution.
>
> The implementation provides memory isolation and virtual address
> allocation. VMs support mapping GEM buffer objects with configurable
> protection flags (readonly, noexec, uncached) and handle both 4KB and 2MB
> page sizes. A new_dummy_object() helper is provided to create a dummy GEM
> object for use as a GPUVM root.
>
> The vm module integrates with the MMU for address space activation and
> provides map/unmap/remap operations with page table synchronization.
>
> Signed-off-by: Boris Brezillon <[email protected]>
> Co-developed-by: Daniel Almeida <[email protected]>
> Signed-off-by: Daniel Almeida <[email protected]>
> Co-developed-by: Deborah Brouwer <[email protected]>
> Signed-off-by: Deborah Brouwer <[email protected]>
I'm seeing some build failures on 32-bit arm:
error[E0308]: mismatched types
--> /home/runner/work/linux/linux/linux/drivers/gpu/drm/tyr/vm.rs:572:45
|
572 | let mut sgt_entry_length: u64 = sgt_entry.dma_len();
| --- ^^^^^^^^^^^^^^^^^^^ expected
`u64`, found `u32`
| |
| expected due to this
|
help: you can convert a `u32` to a `u64`
|
572 | let mut sgt_entry_length: u64 = sgt_entry.dma_len().into();
| +++++++
error[E0308]: mismatched types
--> /home/runner/work/linux/linux/linux/drivers/gpu/drm/tyr/vm.rs:581:26
|
581 | paddr += skip;
| ^^^^ expected `u32`, found `u64`
error[E0277]: cannot add-assign `u64` to `u32`
--> /home/runner/work/linux/linux/linux/drivers/gpu/drm/tyr/vm.rs:581:23
|
581 | paddr += skip;
| ^^ no implementation for `u32 += u64`
|
= help: the trait `core::ops::AddAssign<u64>` is not implemented for `u32`
help: `u32` implements trait `core::ops::AddAssign<Rhs>`
-->
/home/runner/.rustup/toolchains/stable-x86_64-unknown-linux-gnu/lib/rustlib/src/rust/library/core/src/internal_macros.rs:61:9
|
61 | impl const $imp<&$u> for $t {
| ^^^^^^^^^^^^^^^^^^^^^^^^^^^ `core::ops::AddAssign<&u32>`
|
::: /home/runner/work/linux/linux/linux/rust/zerocopy/src/byteorder.rs:316:9
|
316 | impl<O: ByteOrder> core::ops::$trait_assign<$name<O>> for
$native {
|
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
`core::ops::AddAssign<zerocopy::byteorder::U32<O>>`
...
720 | / define_type!(
721 | | A,
722 | | "A 32-bit unsigned integer",
723 | | U32,
... |
735 | | [U64, U128]
736 | | );
| |_- in this macro invocation
|
:::
/home/runner/.rustup/toolchains/stable-x86_64-unknown-linux-gnu/lib/rustlib/src/rust/library/core/src/ops/arith.rs:786:9
|
786 | impl const AddAssign for $t {
| ^^^^^^^^^^^^^^^^^^^^^^^^^^^ `core::ops::AddAssign`
CC drivers/base/firmware_loader/builtin/main.o
...
799 | add_assign_impl! { usize u8 u16 u32 u64 u128 isize i8 i16 i32 i64 i128
f16 f32 f64 f128 }
|
-----------------------------------------------------------------------------------------
in this macro invocation
= note: this error originates in the macro `impl_ops_traits` which comes
from the expansion of the macro `add_assign_impl` (in Nightly builds, run with
-Z macro-backtrace for more info)
error[E0308]: mismatched types
--> /home/runner/work/linux/linux/linux/drivers/gpu/drm/tyr/vm.rs:596:65
|
596 | let segment_mapped = match pt_map(context.pt, iova, paddr,
len, prot) {
| ------ ^^^^^
expected `u64`, found `u32`
| |
| arguments to this function are
incorrect
|
note: function defined here
--> /home/runner/work/linux/linux/linux/drivers/gpu/drm/tyr/vm.rs:731:4
|
731 | fn pt_map(
| ^^^^^^
...
734 | paddr: u64,
| ----------
help: you can convert a `u32` to a `u64`
|
596 | let segment_mapped = match pt_map(context.pt, iova,
paddr.into(), len, prot) {
|
+++++++
CC net/core/tso.o
error: aborting due to 4 previous errors
