Skip to content

Add O_DIRECT fd on Linux to bypass page cache on aligned reads #212

Add O_DIRECT fd on Linux to bypass page cache on aligned reads

Add O_DIRECT fd on Linux to bypass page cache on aligned reads #212

Workflow file for this run

name: HMLL Continuous Integration
on:
push:
branches: [ "main" ]
pull_request:
branches: [ "main" ]
concurrency:
group: ${{ github.workflow }}-${{ github.ref }}
cancel-in-progress: true
jobs:
integration-linux:
name: Linux test-suite
uses: ./.github/workflows/build.yml
with:
os: ubuntu-latest
compiler-cc: gcc
compiler-cxx: g++
enable-rust: true
enable-python: true
enable-sanitizers: true
# If this passes, the logic is likely fine everywhere
integration-compat:
name: Compat (${{ matrix.os }}, ${{ matrix.variant }})
needs: [integration-linux] # Optional: Don't run if integration fails
strategy:
fail-fast: false
matrix:
include:
- os: windows-latest
variant: "MSVC"
- os: macos-latest
variant: "Clang"
- os: ubuntu-latest
variant: "Clang Static"
cc: clang
cxx: clang++
static: true
- os: ubuntu-24.04-arm
variant: "ARM64"
uses: ./.github/workflows/build.yml
with:
os: ${{ matrix.os }}
compiler-cc: ${{ matrix.cc }}
compiler-cxx: ${{ matrix.cxx }}
static-build: ${{ matrix.static || false }}
enable-rust: false
enable-python: false
cuda:
name: CUDA Integration
strategy:
fail-fast: false
matrix:
os: [ ubuntu-latest, windows-latest ]
uses: ./.github/workflows/build.yml
with:
os: ${{ matrix.os }}
enable-cuda: true
run-tests: false