[Deepin-Kernel-SIG] [linux 6.6-y] [Upstream] Unified cross-architecture kernel-mode FPU API

[Avenger-285714]

This series unifies the kernel-mode FPU API across several architectures
by wrapping the existing functions (where needed) in consistently-named
functions placed in a consistent header location, with mostly the same
semantics: they can be called from preemptible or non-preemptible task
context, and are not assumed to be reentrant. Architectures are also
expected to provide CFLAGS adjustments for compiling FPU-dependent code.
For the moment, SIMD/vector units are out of scope for this common API.

This allows us to remove the ifdeffery and duplicated Makefile logic at
each FPU user. It then implements the common API on RISC-V, and converts
a couple of users to the new API: the AMDGPU DRM driver, and the FPU
self test.

The underlying goal of this series is to allow using newer AMD GPUs
(e.g. Navi) on RISC-V boards such as SiFive's HiFive Unmatched. Those
GPUs need CONFIG_DRM_AMD_DC_FP to initialize, which requires kernel-mode
FPU support.

Link: https://lore.kernel.org/all/20240329072441.591471-2-samuel.holland@sifive.com/T/#u

Summary by Sourcery

Unify and centralize the kernel-mode floating-point API and build flags across all architectures by introducing consistent kernel_fpu_begin/end/available interfaces, consolidating FPU-related CFLAGS into CC_FLAGS_FPU and CC_FLAGS_NO_FPU, implementing the new API on RISC-V, LoongArch, PowerPC, ARM, ARM64, and x86, and converting existing FPU users (the AMDGPU DRM driver and the FPU self-test) to the new interface; add core documentation for floating-point usage and split the FPU self-test into glue and implementation modules.

New Features:

• Define a unified kernel-mode FPU API (kernel_fpu_begin, kernel_fpu_end, kernel_fpu_available) with a common header
• Introduce CC_FLAGS_FPU and CC_FLAGS_NO_FPU build variables for enabling or disabling FPU support project-wide
• Implement the new FPU API on RISC-V, LoongArch, PowerPC, ARM, ARM64, and x86 architectures
• Convert the AMDGPU DRM driver and the FPU self-test to use the unified FPU API
• Refactor the FPU self-test into separate glue and implementation files, guarded by kernel_fpu_available

Enhancements:

• Remove per-architecture conditional FPU flags in scattered Makefiles and apply centralized flags consistently
• Add include/linux/fpu.h as an umbrella header and update existing consumers to use it

Documentation:

• Add Documentation/core-api/floating-point.rst describing floating-point compilation and usage

Tests:

• Modify the FPU self-test to only initialize when kernel_fpu_available returns true

[sourcery-ai]

Reviewer's Guide

This series centralizes and unifies kernel-mode FPU support across architectures by defining a common API and build flags, providing per-architecture implementations, migrating existing drivers and tests to the new interface, and documenting floating-point compilation requirements.

Class diagram for the unified kernel-mode FPU API

classDiagram
    class FPU_API {
        +bool kernel_fpu_available()
        +void kernel_fpu_begin()
        +void kernel_fpu_end()
    }
    class arch_x86_fpu_h {
        +bool kernel_fpu_available() = true
        +void kernel_fpu_begin() (from asm/fpu/api.h)
        +void kernel_fpu_end() (from asm/fpu/api.h)
    }
    class arch_arm_fpu_h {
        +bool kernel_fpu_available() = cpu_has_neon()
        +void kernel_fpu_begin() = kernel_neon_begin()
        +void kernel_fpu_end() = kernel_neon_end()
    }
    class arch_arm64_fpu_h {
        +bool kernel_fpu_available() = cpu_has_neon()
        +void kernel_fpu_begin() = kernel_neon_begin()
        +void kernel_fpu_end() = kernel_neon_end()
    }
    class arch_loongarch_fpu_h {
        +bool kernel_fpu_available() = cpu_has_fpu
        +void kernel_fpu_begin()
        +void kernel_fpu_end()
    }
    class arch_powerpc_fpu_h {
        +bool kernel_fpu_available() = !cpu_has_feature(CPU_FTR_FPU_UNAVAILABLE)
        +void kernel_fpu_begin()
        +void kernel_fpu_end()
    }
    class arch_riscv_fpu_h {
        +bool kernel_fpu_available() = has_fpu()
        +void kernel_fpu_begin()
        +void kernel_fpu_end()
    }
    FPU_API <|.. arch_x86_fpu_h
    FPU_API <|.. arch_arm_fpu_h
    FPU_API <|.. arch_arm64_fpu_h
    FPU_API <|.. arch_loongarch_fpu_h
    FPU_API <|.. arch_powerpc_fpu_h
    FPU_API <|.. arch_riscv_fpu_h

Loading

Class diagram for updated AMDGPU FPU usage

classDiagram
    class dc_fpu {
        +void dc_fpu_begin(const char *function_name, int line)
        +void dc_fpu_end(const char *function_name, int line)
    }
    class FPU_API {
        +bool kernel_fpu_available()
        +void kernel_fpu_begin()
        +void kernel_fpu_end()
    }
    dc_fpu ..> FPU_API : uses

Loading

Class diagram for FPU self-test refactor

classDiagram
    class test_fpu_glue {
        <<file>>
    }
    class test_fpu_impl {
        <<file>>
    }
    class FPU_API {
        +bool kernel_fpu_available()
        +void kernel_fpu_begin()
        +void kernel_fpu_end()
    }
    test_fpu_glue ..> test_fpu_impl : delegates
    test_fpu_impl ..> FPU_API : uses

Loading

File-Level Changes

Change	Details	Files
Introduce a unified kernel-mode FPU API header and architecture-specific wrappers	Added include/linux/fpu.h as the generic API entry point Created arch/{x86,arm,arm64,loongarch,powerpc,riscv}/include/asm/fpu.h to define kernel_fpu_* Replaced arch-specific #includes in user code with <linux/fpu.h>	include/linux/fpu.h arch/x86/include/asm/fpu.h arch/arm/include/asm/fpu.h arch/arm64/include/asm/fpu.h arch/loongarch/include/asm/fpu.h arch/powerpc/include/asm/fpu.h arch/riscv/include/asm/fpu.h drivers/gpu/drm/amd/display/amdgpu_dm/dc_fpu.c lib/test_fpu_glue.c
Define and apply CC_FLAGS_FPU and CC_FLAGS_NO_FPU build flags	Export CC_FLAGS_FPU and CC_FLAGS_NO_FPU in the top-level Makefile Replaced scattered per-architecture FPU flag logic with CC_FLAGS_FPU/NO_FPU Adjusted Makefiles for drivers/gpu, lib/raid6, lib, and multiple arch subdirs	Makefile drivers/gpu/drm/amd/display/dc/dml/Makefile lib/raid6/Makefile lib/Makefile arch/x86/Makefile arch/arm/Makefile arch/arm64/Makefile arch/loongarch/Makefile arch/powerpc/Makefile arch/arm/lib/Makefile arch/arm64/lib/Makefile
Implement per-architecture kernel-mode FPU routines for RISC-V	Created arch/riscv/kernel/kernel_mode_fpu.c with begin/end definitions Updated arch/riscv/kernel/Makefile to build kernel_mode_fpu.o Removed inline kernel_fpu_* definitions from arch/riscv/include/asm/switch_to.h	arch/riscv/kernel/kernel_mode_fpu.c arch/riscv/kernel/Makefile arch/riscv/include/asm/switch_to.h
Migrate existing FPU users to the unified API	Updated AMDGPU dc_fpu.c to call kernel_fpu_available(), kernel_fpu_begin()/end() Split test_fpu into glue and implementation files using the new API Added lib/test_fpu_impl.c and header lib/test_fpu.h	drivers/gpu/drm/amd/display/amdgpu_dm/dc_fpu.c lib/test_fpu_glue.c lib/test_fpu_impl.c lib/test_fpu.h
Document floating-point compilation requirements	Added Documentation/core-api/floating-point.rst describing CC_FLAGS_FPU/NO_FPU usage	Documentation/core-api/floating-point.rst

---

Tips and commands

Interacting with Sourcery

• Trigger a new review: Comment @sourcery-ai review on the pull request.
• Continue discussions: Reply directly to Sourcery's review comments.
• Generate a GitHub issue from a review comment: Ask Sourcery to create an
  issue from a review comment by replying to it. You can also reply to a
  review comment with @sourcery-ai issue to create an issue from it.
• Generate a pull request title: Write @sourcery-ai anywhere in the pull
  request title to generate a title at any time. You can also comment
  @sourcery-ai title on the pull request to (re-)generate the title at any time.
• Generate a pull request summary: Write @sourcery-ai summary anywhere in
  the pull request body to generate a PR summary at any time exactly where you
  want it. You can also comment @sourcery-ai summary on the pull request to
  (re-)generate the summary at any time.
• Generate reviewer's guide: Comment @sourcery-ai guide on the pull
  request to (re-)generate the reviewer's guide at any time.
• Resolve all Sourcery comments: Comment @sourcery-ai resolve on the
  pull request to resolve all Sourcery comments. Useful if you've already
  addressed all the comments and don't want to see them anymore.
• Dismiss all Sourcery reviews: Comment @sourcery-ai dismiss on the pull
  request to dismiss all existing Sourcery reviews. Especially useful if you
  want to start fresh with a new review - don't forget to comment
  @sourcery-ai review to trigger a new review!

Customizing Your Experience

Access your dashboard to:

• Enable or disable review features such as the Sourcery-generated pull request
  summary, the reviewer's guide, and others.
• Change the review language.
• Add, remove or edit custom review instructions.
• Adjust other review settings.

Getting Help

• Contact our support team for questions or feedback.
• Visit our documentation for detailed guides and information.
• Keep in touch with the Sourcery team by following us on X/Twitter, LinkedIn or GitHub.

[Copilot]

Pull Request Overview

This PR introduces a unified cross-architecture kernel-mode FPU API to standardize floating-point operations across different CPU architectures. The series replaces architecture-specific FPU handling with a consistent API defined in linux/fpu.h, enabling better code reuse and simplifying FPU-dependent code compilation.

• Unified FPU API across x86, ARM, ARM64, PowerPC, LoongArch, and RISC-V architectures
• Standardized compilation flags (CC_FLAGS_FPU/CC_FLAGS_NO_FPU) for FPU code
• Updated AMD GPU driver and FPU self-test to use the new common API

Reviewed Changes

Copilot reviewed 37 out of 37 changed files in this pull request and generated 4 comments.

Show a summary per file

File	Description
include/linux/fpu.h	New unified FPU header with compilation unit check
arch/*/include/asm/fpu.h	Architecture-specific FPU API implementations
arch/*/Makefile	Standardized FPU compilation flags definitions
arch/*/Kconfig	Added ARCH_HAS_KERNEL_FPU_SUPPORT configuration
drivers/gpu/drm/amd/display/	Updated AMD GPU driver to use unified FPU API
lib/test_fpu*	Refactored FPU self-test with separated implementation
lib/raid6/Makefile	Updated to use standardized FPU compilation flags
arch/Kconfig	Added new kernel FPU support configuration option
Documentation/core-api/floating-point.rst	New documentation for FPU API usage

[Copilot]

The error message should use 'FPU' instead of 'FP' for consistency with the rest of the codebase and the filename.

Suggested change

	#error FP code must be compiled separately. See Documentation/core-api/floating-point.rst.
	#error FPU code must be compiled separately. See Documentation/core-api/floating-point.rst.

[Copilot]

Calling task_pt_regs(current) twice (here and in kernel_fpu_end) is inefficient. Consider storing the result in a local variable.

Suggested change

	fstate_save(current, task_pt_regs(current));
	struct pt_regs *regs = task_pt_regs(current);
	fstate_save(current, regs);

[Copilot]

Using BUG_ON() in this context may be too harsh as it crashes the kernel. Consider using WARN_ON_ONCE() and returning early, or checking kernel_fpu_available() at module initialization time instead.

Suggested change

	BUG_ON(!kernel_fpu_available());
	if (WARN_ON_ONCE(!kernel_fpu_available())) {
	put_cpu_ptr(&fpu_recursion_depth);
	return;
	}

[Copilot]

The availability check in test_fpu_init() should be moved earlier, before creating debugfs entries, to avoid creating debugfs files when FPU is not available.

[sourcery-ai]

Hey @Avenger-285714 - I've reviewed your changes and they look great!

Prompt for AI Agents

Please address the comments from this code review:
## Individual Comments

### Comment 1
<location> `arch/arm/include/asm/fpu.h:11` </location>
<code_context>
+
+#include <asm/neon.h>
+
+#define kernel_fpu_available()	cpu_has_neon()
+#define kernel_fpu_begin()	kernel_neon_begin()
+#define kernel_fpu_end()	kernel_neon_end()
</code_context>

<issue_to_address>
kernel_fpu_available() is tied to NEON; may not cover all ARM FPU cases.

cpu_has_neon() may not accurately reflect FPU support on all ARM CPUs, potentially leading to incorrect detection of FPU availability.
</issue_to_address>

---

Sourcery is free for open source - if you like our reviews please consider sharing them ✨

• X
• Mastodon
• LinkedIn
• Facebook

_{Help me be more useful! Please click 👍 or 👎 on each comment and I'll use the feedback to improve your reviews.}

[sourcery-ai]

issue (bug_risk): kernel_fpu_available() is tied to NEON; may not cover all ARM FPU cases.

cpu_has_neon() may not accurately reflect FPU support on all ARM CPUs, potentially leading to incorrect detection of FPU availability.

[deepin-ci-robot]

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: opsiff

The full list of commands accepted by this bot can be found here.

The pull request process is described here

Needs approval from an approver in each of these files:

• deepin/OWNERS [opsiff]

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment