Update prebuilt objects [skip actions]

add shaffenmeisters PR and add web config

.github/workflows/CrossBuild.yml

@@ -0,0 +1,162 @@
+# This is a basic workflow to help you get started with Actions
+name: Cross-Build
+
+on:
+  push:
+    branches: 
+      - 'master-cmake'
+      - '!**4.3'
+  pull_request:
+    branches:
+      - 'master-cmake'
+      - '!**4.3'
+jobs:
+  job1:
+    name: Build Number
+    runs-on: ubuntu-latest
+    outputs:
+      build_number: ${{ steps.buildnumber.outputs.build_number }}
+    steps:
+    - name: Generate common build number
+      id: buildnumber
+      uses: einaregilsson/build-number@v3
+      with:
+        token: ${{secrets.github_token}}
+  build:
+    runs-on: ubuntu-latest
+    needs: job1
+    strategy:
+      max-parallel: 1
+      matrix:
+        node: [I2S-4MFlash, SqueezeAmp, Muse]
+        depth: [16, 32]
+        exclude:
+          - node: Muse
+            depth: 32
+    steps:
+    - name: Set target name
+      run: |
+        echo "TARGET_BUILD_NAME=${{ matrix.node }}" >> $GITHUB_ENV
+        echo "build_version_prefix=1." >> $GITHUB_ENV
+    - uses: actions/checkout@v2
+      with:
+        fetch-depth: 15
+        submodules: true
+    - name: Cache build
+      id: cache-build
+      uses: actions/cache@v2
+      with:
+        path: |
+          build
+          /var/lib/docker
+        key: ${{ runner.os }}-${{ matrix.node }}-${{ matrix.depth }}
+    - name: Set build parameters
+      run: |
+        git update-index --chmod=+x ./server_certs/getcert.sh
+        cd server_certs;./getcert.sh;cat github.pem;cd ..
+        shopt -s nocasematch
+        branch_name="${GITHUB_REF//refs\/heads\//}"
+        branch_name="${branch_name//[^a-zA-Z0-9\-~!@_\.]/}"
+        BUILD_NUMBER=${{ needs.job1.outputs.build_number }}
+        echo "BUILD_NUMBER=${BUILD_NUMBER}" >> $GITHUB_ENV
+        tag="${TARGET_BUILD_NAME}.${{matrix.depth}}.${BUILD_NUMBER}.${branch_name}"
+        echo "tag=${tag}" >> $GITHUB_ENV
+        last_commit="$(git log --pretty=format:'%s' --max-count=1)"
+        if [[ "$last_commit" =~ .*"Release".* ]]; then echo "release_flag=1" >> $GITHUB_ENV; else echo "release_flag=0" >> $GITHUB_ENV; fi
+        name="1.${BUILD_NUMBER}-${{matrix.depth}}#v4.0#${TARGET_BUILD_NAME}#${branch_name}"
+        artifact_prefix="squeezelite-esp32-${branch_name}-${TARGET_BUILD_NAME}-${{matrix.depth}}-${build_version_prefix}${BUILD_NUMBER}"
+        artifact_file_name="${artifact_prefix}.zip"
+        artifact_bin_file_name="${artifact_prefix}.bin"
+        echo "name=${name}" >> $GITHUB_ENV
+        echo "last_commit=${last_commit}" >> $GITHUB_ENV
+        echo "artifact_file_name=${artifact_file_name}" >> $GITHUB_ENV
+        echo "PROJECT_VER=${TARGET_BUILD_NAME}-${{ steps.buildnumber.outputs.build_number }} " >> $GITHUB_ENV
+        echo "artifact_bin_file_name=${artifact_bin_file_name}" >> $GITHUB_ENV
+        description=""
+        description=${description}$'------------------------------\n### Revision Log\n\n'
+        description="$description$(git log --pretty=format:'%h %s (%cI) <%an>' --abbrev-commit --max-count=15 | sed --r 's/(^[\*]+)/\\\1/g') "        
+        echo 'description<<~EOD' >> $GITHUB_ENV
+        echo ${description}>> $GITHUB_ENV
+        echo '~EOD' >> $GITHUB_ENV        
+        echo ####### 
+        echo ####### Environment
+        echo ####### 
+        env
+        echo ####### 
+        echo ####### GITHUB ENV
+        echo ####### 
+        cat $GITHUB_ENV
+    - name: Build the firmware
+      run: |
+        env | grep "artifact\|tag\|GITHUB\|version\|NUMBER\|TARGET" >${TARGET_BUILD_NAME}-env.txt
+        echo "${tag}" >version.txt
+        docker pull sle118/idf:release-v4.0 
+        docker info 
+        docker run --env-file=${TARGET_BUILD_NAME}-env.txt  -v $PWD:/project -w /project sle118/idf:release-v4.0 /bin/bash -c "cp build-scripts/${TARGET_BUILD_NAME}-sdkconfig.defaults sdkconfig  && idf.py build -DDEPTH=${{ matrix.depth }} -DBUILD_NUMBER=${BUILD_NUMBER}-${{ matrix.depth }} && zip -r build_output.zip build && zip build/${artifact_file_name} partitions*.csv build/*.bin build/bootloader/bootloader.bin build/partition_table/partition-table.bin build/flash_project_args build/size_*.txt"
+    # - name: Build Mock firmware
+    #   run: |
+    #     mkdir -p build
+    #     cd build
+    #     mkdir -p partition_table
+    #     mkdir -p bootloader
+    #     echo "mock content"> squeezelite.bin
+    #     echo "mock content"> recovery.bin
+    #     echo "mock content"> ./bootloader/bootloader.bin
+    #     echo "mock content"> ./partition_table/partition-table.bin 
+    #     echo "mock content"> flash_project_args 
+    #     echo "mock content"> size_comp1.txt
+    #     echo "mock content"> size_comp2.txt
+    #     echo "mock content"> ../partitions.csv
+    - uses: actions/upload-artifact@v2
+      with:
+        name: ${{ env.artifact_file_name }}
+        path: |
+          build/*.bin
+          build/bootloader/bootloader.bin
+          build/partition_table/partition-table.bin 
+          build/flash_project_args 
+          build/size_comp1.txt
+          build/size_comp2.txt
+          partitions.csv
+          sdkconfig
+          server_certs/github.pem
+          build_output.zip
+    - uses: actions/upload-artifact@v2
+      with:
+        name: ${{ env.artifact_bin_file_name }}
+        path: |
+          build/squeezelite.bin
+    - name: Create Release
+      if: env.release_flag == 1     
+      id: create_release
+      uses: actions/create-release@v1
+      env:
+        GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }} # This token is provided by Actions, you do not need to create your own token
+      with:
+        tag_name: ${{ env.tag }}
+        release_name: ${{ env.name }}
+        body: ${{ env.description }}
+        draft: false
+        prerelease: true
+    - name: Upload Release Asset - Squeezelite binary file
+      if: env.release_flag == 1         
+      id: upload-release-asset 
+      uses: actions/upload-release-asset@v1
+      env:
+        GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+      with:
+        upload_url: ${{ steps.create_release.outputs.upload_url }} # This pulls from the CREATE RELEASE step above, referencing it's ID to get its outputs object, which include a `upload_url`. See this blog post for more info: https://jasonet.co/posts/new-features-of-github-actions/#passing-data-to-future-steps 
+        asset_path: build/squeezelite.bin
+        asset_name: ${{ env.artifact_bin_file_name }}
+        asset_content_type: application/octet-stream        
+    - name: Upload Release Asset - Zip file
+      if: env.release_flag == 1         
+      id: upload-release-asset-zip
+      uses: actions/upload-release-asset@v1
+      env:
+        GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+      with:
+        upload_url: ${{ steps.create_release.outputs.upload_url }} # This pulls from the CREATE RELEASE step above, referencing it's ID to get its outputs object, which include a `upload_url`. See this blog post for more info: https://jasonet.co/posts/new-features-of-github-actions/#passing-data-to-future-steps 
+        asset_path: build/${{ env.artifact_file_name }}
+        asset_name: ${{ env.artifact_file_name }}
+        asset_content_type: application/octet-stream                

.github/workflows/Platform_build.yml

@@ -0,0 +1,232 @@
+name: Platform Build
+on: 
+  push:
+    branches: 
+      - '**4.3'
+  workflow_dispatch:
+    inputs:
+      ui_build:
+        description: 'Force Rebuilding the UI. When not forced, the system will check for [ui-build] in the last commit message to trigger a ui rebuild'
+        required: true
+        type: boolean
+      release_build:
+        description: 'Force a Release build. When not forced, the system will check for release word in the last commit message to trigger a release'
+        required: true
+        type: boolean
+
+jobs:
+  bootstrap:
+    name: Global setup
+    runs-on: ubuntu-latest
+    container:
+      image: sle118/squeezelite-esp32-idfv43
+    outputs:
+      build_number: ${{ steps.buildnumber.outputs.build_number }}
+      ui_build: ${{ steps.build_flags.outputs.ui_build }}
+      release_flag: ${{ steps.build_flags.outputs.release_flag }}      
+      mock: ${{ steps.build_flags.outputs.mock }}            
+    steps:
+    - uses: actions/checkout@v2
+      with:
+        fetch-depth: 0
+        submodules: true
+    - name: Generate common build number
+      id: buildnumber
+      uses: einaregilsson/build-number@v3
+      with:
+        token: ${{secrets.github_token}}
+    - name: Set build flags
+      id: build_flags
+      run: |
+        git config --global --add safe.directory /__w/squeezelite-esp32/squeezelite-esp32   
+        [ ${{github.event.inputs.ui_build}} ] && ui_build_option="--ui_build" || ui_build_option="" 
+        [ ${{github.event.inputs.release_build}} ] && release_build_option="--force" || release_build_option="" 
+        echo "ui_build_option=$ui_build_option" >> $GITHUB_ENV
+        echo "release_build_option=$release_build_option" >> $GITHUB_ENV
+        echo "Dumping environment"
+        env
+        . /opt/esp/python_env/idf4.3_py3.8_env/bin/activate
+        # build_flags support the following options
+        # --mock - to mock the compilation part - this is to be used for testing only
+        # --force - to force a release build even if the last commit message doesn't contain the word "release"
+        # --ui_build - to force a ui_build even if the last commit message doesn't contain "[ui-build]"
+        build_tools.py build_flags $ui_build_option $release_build_option 
+    - name: Show Build Flags        
+      run: |
+          echo "Running with the following options" 
+          echo "Web Build Flag=${{steps.build_flags.outputs.ui_build}}"
+          echo "Mock flag=${{steps.build_flags.outputs.mock}}"
+          echo "Release Flag=${{steps.build_flags.outputs.release_flag}}"      
+    - name: Refresh certificates
+      if: ${{ steps.build_flags.outputs.release_flag }}
+      run: |
+        git update-index --chmod=+x ./server_certs/getcert.sh
+        cd server_certs;./getcert.sh;cat github.pem;cd ..        
+    - name: Setup Node.js dependencies
+      if: ${{ !env.ACT }}
+      uses: actions/setup-node@v3
+      with:
+        node-version: 16
+        cache: 'npm'
+        cache-dependency-path: components/wifi-manager/webapp/package.json   
+    - name: Build Web Application
+      if: ${{ steps.build_flags.outputs.ui_build == 1 }} 
+      run: |
+        cd components/wifi-manager/webapp/ 
+        npm install 
+        npm run-script build
+    - name: Update repository with prebuilt items
+      if:  ${{ steps.build_flags.outputs.ui_build == 1 || steps.build_flags.outputs.release_flag == 1 }} 
+      run: |
+        git config user.name github-actions
+        git config user.email github-actions@github.com
+        git add server_certs 
+        git add components/wifi-manager/webapp/*.h 
+        git add components/wifi-manager/webapp/*.c 
+        git add components/wifi-manager/webapp/*.cmake 
+        git add components/wifi-manager/webapp/dist/*
+        git commit -m "Update prebuilt objects [skip actions]"
+        git push  https://${{secrets.github_token}}@github.com/sle118/squeezelite-esp32.git
+    - name: Locally store commonly built objects
+      uses: actions/upload-artifact@v3
+      with:
+        name: prebuilt_objects
+        path: |
+          server_certs 
+          components/wifi-manager/webapp/*.h
+          components/wifi-manager/webapp/*.c
+          components/wifi-manager/webapp/dist/*
+          components/wifi-manager/webapp/*.cmake 
+  build:
+    runs-on: ubuntu-latest
+    container:
+      image: sle118/squeezelite-esp32-idfv43
+    needs: [bootstrap]
+    strategy:
+      matrix:
+        node: [I2S-4MFlash, SqueezeAmp, Muse]
+        depth: [16, 32]
+        exclude:
+          - node: Muse
+            depth: 32      
+          - node: bootstrap
+            depth: 32
+    steps:
+    - uses: actions/checkout@v2
+      with:
+        fetch-depth: 0
+        submodules: true
+    - name: Show Build Flags        
+      run: |
+          echo "Running with the following options" 
+          echo "Web Build Flag=${{needs.bootstrap.outputs.ui_build}}"
+          echo "Mock flag=${{needs.bootstrap.outputs.mock}}"
+          echo "Release Flag=${{needs.bootstrap.outputs.release_flag}}"
+          echo Environment File name: $GITHUB_ENV        
+    - name: Set build parameters
+      run: |
+        . /opt/esp/python_env/idf4.3_py3.8_env/bin/activate
+        git config --global --add safe.directory /__w/squeezelite-esp32/squeezelite-esp32   
+        git status
+        build_tools.py environment --build ${{ needs.bootstrap.outputs.build_number }} --env_file "$GITHUB_ENV" --node "${{matrix.node}}" --depth ${{matrix.depth}} --major 2 --docker sle118/squeezelite-esp32-idfv43 
+        
+    - uses: actions/download-artifact@master
+      name: Restore common objects
+      with:
+        name: prebuilt_objects
+    - name: Build the firmware
+      if: ${{ needs.bootstrap.outputs.mock == 0 }} 
+      run: |
+        . ${IDF_PYTHON_ENV_PATH}/bin/activate
+        chmod +x ./components/spotify/cspot/bell/nanopb/generator/protoc
+        chmod +x ./components/spotify/cspot/bell/nanopb/generator/protoc-gen-nanopb
+        chmod +x ./components/spotify/cspot/bell/nanopb/generator/*.py
+        chmod +x ./components/spotify/cspot/bell/nanopb/generator/*.py2
+        chmod +x ./components/spotify/cspot/bell/nanopb/generator/proto/*.py      
+        echo "Copying target sdkconfig"
+        cp build-scripts/${TARGET_BUILD_NAME}-sdkconfig.defaults sdkconfig
+        echo "Building project"
+        idf.py build -DDEPTH=${DEPTH} -DBUILD_NUMBER=${BUILD_NUMBER}-${DEPTH} 
+    - name: Build Mock firmware
+      if: ${{ needs.bootstrap.outputs.mock == 1 }} 
+      run: |
+        mkdir -p build
+        cd build
+        mkdir -p partition_table
+        mkdir -p bootloader
+        echo \\"mock content\\"> ./squeezelite.bin
+        echo \"mock content\"> ./recovery.bin
+        echo \"mock content\"> ./bootloader/bootloader.bin
+        echo \"mock content\"> ./partition_table/partition-table.bin 
+        echo \"mock content\"> ./ota_data_initial.bin
+        echo \"mock content\"> ./flash_project_args 
+        echo \"mock content\"> ./size_comp1.txt
+        echo \"mock content\"> ./size_comp2.txt
+        echo \"mock content\"> ./partitions.csv        
+        echo {  \"write_flash_args\" : [ \"--flash_mode\", \"dio\", \"--flash_size\", \"detect\", \"--flash_freq\", \"80m\" ],  \"flash_settings\" : {      \"flash_mode\": \"dio\",      \"flash_size\": \"detect\",      \"flash_freq\": \"80m\"  },  \"flash_files\" : {      \"0x8000\" : \"partition_table/partition-table.bin\",      \"0xd000\" : \"ota_data_initial.bin\",      \"0x1000\" : \"bootloader/bootloader.bin\",      \"0x10000\" : \"recovery.bin\",      \"0x150000\" : \"squeezelite.bin\"  },  \"partition_table\" : { \"offset\" : \"0x8000\", \"file\" : \"partition_table/partition-table.bin\" },  \"otadata\" : { \"offset\" : \"0xd000\", \"file\" : \"ota_data_initial.bin\" },  \"bootloader\" : { \"offset\" : \"0x1000\", \"file\" : \"bootloader/bootloader.bin\" },  \"app\" : { \"offset\" : \"0x10000\", \"file\" : \"recovery.bin\" },  \"squeezelite\" : { \"offset\" : \"0x150000\", \"file\" : \"squeezelite.bin\" },  \"extra_esptool_args\" : {      \"after\"  : \"hard_reset\",      \"before\" : \"default_reset\"  }  }  > ./flasher_args.json
+    - name: Create Release Artifact Zip 
+      if: ${{ needs.bootstrap.outputs.release_flag == 1 && needs.bootstrap.outputs.mock == 0 }}
+      run: |
+        if [ -z "${artifact_file_name}" ]
+        then
+            echo "No artifact file name set.  Will not generate zip file."
+        else
+            echo "Generating build artifact zip file"
+            zip -r build_output.zip build
+            zip build/${artifact_file_name} partitions*.csv components/ build/*.bin build/bootloader/bootloader.bin build/partition_table/partition-table.bin build/flash_project_args build/size_*.txt
+        fi
+    - name: Upload Build Artifacts
+      uses: actions/upload-artifact@v3
+      if: ${{ needs.bootstrap.outputs.mock == 0 }}
+      with:
+        name: ${{ env.artifact_prefix }}
+        path: |
+          build/flash_project_args 
+          build/size_comp1.txt
+          build/size_comp2.txt
+          partitions.csv
+          sdkconfig
+          server_certs/github.pem
+          build/*.bin 
+          build/bootloader/bootloader.bin 
+          build/partition_table/partition-table.bin
+          build_output.zip
+    - name: Create Release
+      if: ${{ needs.bootstrap.outputs.release_flag == 1 && needs.bootstrap.outputs.mock == 0 }}
+      id: create_release
+      uses: actions/create-release@v1
+      env:
+        GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }} # This token is provided by Actions, you do not need to create your own token
+      with:
+        tag_name: ${{ env.tag }}
+        release_name: ${{ env.name }}
+        body: ${{ env.description }}
+        draft: false
+        prerelease: true
+    - name: Upload Release Asset - Squeezelite binary file
+      if: ${{ needs.bootstrap.outputs.release_flag == 1 && needs.bootstrap.outputs.mock == 0 }}
+      id: upload-release-asset 
+      uses: actions/upload-release-asset@v1
+      env:
+        GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+      with:
+        upload_url: ${{ steps.create_release.outputs.upload_url }} # This pulls from the CREATE RELEASE step above, referencing it's ID to get its outputs object, which include a `upload_url`. See this blog post for more info: https://jasonet.co/posts/new-features-of-github-actions/#passing-data-to-future-steps 
+        asset_path: build/squeezelite.bin
+        asset_name: ${{ env.artifact_bin_file_name }}
+        asset_content_type: application/octet-stream        
+    - name: Upload Release Asset - Zip file
+      if: ${{ needs.bootstrap.outputs.release_flag == 1 && needs.bootstrap.outputs.mock == 0 }}
+      id: upload-release-asset-zip
+      uses: actions/upload-release-asset@v1
+      env:
+        GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+      with:
+        upload_url: ${{ steps.create_release.outputs.upload_url }} # This pulls from the CREATE RELEASE step above, referencing it's ID to get its outputs object, which include a `upload_url`. See this blog post for more info: https://jasonet.co/posts/new-features-of-github-actions/#passing-data-to-future-steps 
+        asset_path: build/${{ env.artifact_file_name }}
+        asset_name: ${{ env.artifact_file_name }}
+        asset_content_type: application/octet-stream     
+  update_web_installer:
+    name: Update Web Installer After Release
+    needs: [ bootstrap, build ]
+    if:  ${{( always() && !cancelled() ) && needs.bootstrap.outputs.release_flag == 1 && needs.bootstrap.outputs.mock == 0 }}
+    uses: ./.github/workflows/web_deploy.yml

.github/workflows/codeql-analysis.yml

@@ -0,0 +1,67 @@
+# For most projects, this workflow file will not need changing; you simply need
+# to commit it to your repository.
+#
+# You may wish to alter this file to override the set of languages analyzed,
+# or to provide custom queries or build logic.
+#
+# ******** NOTE ********
+# We have attempted to detect the languages in your repository. Please check
+# the `language` matrix defined below to confirm you have the correct set of
+# supported CodeQL languages.
+#
+name: "CodeQL"
+
+on:
+  push:
+    branches: [ master-cmake ]
+  pull_request:
+    # The branches below must be a subset of the branches above
+    branches: [ master-cmake ]
+  schedule:
+    - cron: '19 12 * * 4'
+
+jobs:
+  analyze:
+    name: Analyze
+    runs-on: ubuntu-latest
+
+    strategy:
+      fail-fast: false
+      matrix:
+        language: [ 'cpp', 'javascript', 'python']
+        # CodeQL supports [ 'cpp', 'csharp', 'go', 'java', 'javascript', 'python' ]
+        # Learn more:
+        # https://docs.github.com/en/free-pro-team@latest/github/finding-security-vulnerabilities-and-errors-in-your-code/configuring-code-scanning#changing-the-languages-that-are-analyzed
+
+    steps:
+    - name: Checkout repository
+      uses: actions/checkout@v2
+
+    # Initializes the CodeQL tools for scanning.
+    - name: Initialize CodeQL
+      uses: github/codeql-action/init@v1
+      with:
+        languages: ${{ matrix.language }}
+        # If you wish to specify custom queries, you can do so here or in a config file.
+        # By default, queries listed here will override any specified in a config file.
+        # Prefix the list here with "+" to use these queries and those in the config file.
+        # queries: ./path/to/local/query, your-org/your-repo/queries@main
+
+    # Autobuild attempts to build any compiled languages  (C/C++, C#, or Java).
+    # If this step fails, then you should remove it and run the build manually (see below)
+    - name: Autobuild
+      uses: github/codeql-action/autobuild@v1
+
+    # ℹ️ Command-line programs to run using the OS shell.
+    # 📚 https://git.io/JvXDl
+
+    # ✏️ If the Autobuild fails above, remove it and uncomment the following three lines
+    #    and modify them (or add more) to build your code if your project
+    #    uses a compiled language
+
+    #- run: |
+    #   make bootstrap
+    #   make release
+
+    - name: Perform CodeQL Analysis
+      uses: github/codeql-action/analyze@v1

.github/workflows/esp-idf-v4.3-build.yml

@@ -0,0 +1,167 @@
+# This is a basic workflow to help you get started with Actions
+name: ESP-IDF v4.3.1
+on:
+  push:
+    branches: 
+      - '**4.3'
+  pull_request:
+    branches:
+      - '**4.3'
+jobs:
+  job1:
+    name: Build Number
+    runs-on: ubuntu-latest
+    outputs:
+      build_number: ${{ steps.buildnumber.outputs.build_number }}
+    steps:
+    - name: Generate common build number
+      id: buildnumber
+      uses: einaregilsson/build-number@v3
+      with:
+        token: ${{secrets.github_token}}
+  build:
+    runs-on: ubuntu-latest
+    needs: job1
+    strategy:
+      max-parallel: 1
+      matrix:
+        node: [I2S-4MFlash, SqueezeAmp, Muse]
+        depth: [16, 32]
+        exclude:
+         - node: Muse
+           depth: 32
+    steps:
+    - name: Set target name
+      run: |
+        echo "TARGET_BUILD_NAME=${{ matrix.node }}" >> $GITHUB_ENV
+        echo "build_version_prefix=2." >> $GITHUB_ENV
+    - uses: actions/checkout@v2
+      with:
+        fetch-depth: 15
+        submodules: true
+    - name: Cache build
+      id: cache-build
+      uses: actions/cache@v2
+      with:
+        path: |
+          ~/build 
+          ./
+        key: ${{ runner.os }}-${{ matrix.node }}
+    - name: Set build parameters
+      run: |
+        git update-index --chmod=+x ./server_certs/getcert.sh
+        git update-index --chmod=+x ./buildFirmware.sh
+        git update-index --chmod=+x ./components/spotify/cspot/bell/nanopb/generator/protoc
+        git update-index --chmod=+x ./components/spotify/cspot/bell/nanopb/generator/protoc-gen-nanopb
+        git update-index --chmod=+x ./components/spotify/cspot/bell/nanopb/generator/*.py
+        git update-index --chmod=+x ./components/spotify/cspot/bell/nanopb/generator/*.py2
+        git update-index --chmod=+x ./components/spotify/cspot/bell/nanopb/generator/proto/*.py
+        
+
+        cd server_certs;./getcert.sh;cat github.pem;cd ..
+        shopt -s nocasematch
+        branch_name="${GITHUB_REF//refs\/heads\//}"
+        branch_name="${branch_name//[^a-zA-Z0-9\-~!@_\.]/}"
+        BUILD_NUMBER=${{ needs.job1.outputs.build_number }}
+        echo "BUILD_NUMBER=${BUILD_NUMBER}" >> $GITHUB_ENV
+        echo "DOCKER_IMAGE_NAME=sle118/squeezelite-esp32-idfv4-master" >> $GITHUB_ENV
+        tag="${TARGET_BUILD_NAME}.${{matrix.depth}}.${build_version_prefix}${BUILD_NUMBER}.${branch_name}"
+        echo "tag=${tag}" >> $GITHUB_ENV
+        last_commit="$(git log --pretty=format:'%s' --max-count=1)"
+        if [[ "$last_commit" =~ .*"Release".* ]]; then echo "release_flag=1" >> $GITHUB_ENV; else echo "release_flag=0" >> $GITHUB_ENV; fi
+        name="${build_version_prefix}${BUILD_NUMBER}-${{matrix.depth}}#v4.0#${TARGET_BUILD_NAME}#${branch_name}"
+        artifact_prefix="squeezelite-esp32-${branch_name}-${TARGET_BUILD_NAME}-${{matrix.depth}}-${build_version_prefix}${BUILD_NUMBER}"
+        artifact_file_name="${artifact_prefix}.zip"
+        artifact_bin_file_name="${artifact_prefix}.bin"
+        echo "name=${name}" >> $GITHUB_ENV
+        echo "last_commit=${last_commit}" >> $GITHUB_ENV
+        echo "artifact_file_name=${artifact_file_name}" >> $GITHUB_ENV
+        echo "PROJECT_VER=${TARGET_BUILD_NAME}-${{ steps.buildnumber.outputs.build_number }} " >> $GITHUB_ENV
+        echo "artifact_bin_file_name=${artifact_bin_file_name}" >> $GITHUB_ENV
+        description=$'### Revision Log\n'
+        githist="$(git log --pretty=format:'%h %s (%cI) <%an>' --abbrev-commit --max-count=15 | sed --r 's/(^[\*]+)/\\\1/g')"
+        description="$description$githist"
+        echo 'description<<~EOD' >> $GITHUB_ENV
+        echo ${description}>> $GITHUB_ENV
+        echo '~EOD' >> $GITHUB_ENV        
+        echo ####### 
+        echo ####### Environment
+        echo ####### 
+        env
+        echo ####### 
+        echo ####### GITHUB ENV
+        echo ####### 
+        cat $GITHUB_ENV
+    - name: Build the firmware
+      run: |
+        env | grep "artifact\|tag\|GITHUB\|version\|NUMBER\|TARGET" >${TARGET_BUILD_NAME}-env.txt
+        echo pulling custom docker image  ${DOCKER_IMAGE_NAME}
+        docker pull ${DOCKER_IMAGE_NAME}
+        docker run --env-file=${TARGET_BUILD_NAME}-env.txt  -v $PWD:/project -w /project ${DOCKER_IMAGE_NAME} /bin/bash -c "./buildFirmware.sh"
+    # - name: Build Mock firmware
+    #   run: |
+    #     mkdir -p build
+    #     cd build
+    #     mkdir -p partition_table
+    #     mkdir -p bootloader
+    #     echo "mock content"> squeezelite.bin
+    #     echo "mock content"> recovery.bin
+    #     echo "mock content"> ./bootloader/bootloader.bin
+    #     echo "mock content"> ./partition_table/partition-table.bin 
+    #     echo "mock content"> flash_project_args 
+    #     echo "mock content"> size_comp1.txt
+    #     echo "mock content"> size_comp2.txt
+    #     echo "mock content"> ../partitions.csv
+    - uses: actions/upload-artifact@v2
+      with:
+        name: ${{ env.artifact_file_name }}
+        path: |
+          build/*.bin
+          build/bootloader/bootloader.bin
+          build/partition_table/partition-table.bin 
+          build/flash_project_args 
+          build/size_comp1.txt
+          build/size_comp2.txt
+          partitions.csv
+          sdkconfig
+          server_certs/github.pem
+          build_output.zip
+    - uses: actions/upload-artifact@v2
+      with:
+        name: ${{ env.artifact_bin_file_name }}
+        path: |
+          build/squeezelite.bin
+    - name: Create Release
+      if: env.release_flag == 1     
+      id: create_release
+      uses: actions/create-release@v1
+      env:
+        GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }} # This token is provided by Actions, you do not need to create your own token
+      with:
+        tag_name: ${{ env.tag }}
+        release_name: ${{ env.name }}
+        body: ${{ env.description }}
+        draft: false
+        prerelease: true
+    - name: Upload Release Asset - Squeezelite binary file
+      if: env.release_flag == 1         
+      id: upload-release-asset 
+      uses: actions/upload-release-asset@v1
+      env:
+        GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+      with:
+        upload_url: ${{ steps.create_release.outputs.upload_url }} # This pulls from the CREATE RELEASE step above, referencing it's ID to get its outputs object, which include a `upload_url`. See this blog post for more info: https://jasonet.co/posts/new-features-of-github-actions/#passing-data-to-future-steps 
+        asset_path: build/squeezelite.bin
+        asset_name: ${{ env.artifact_bin_file_name }}
+        asset_content_type: application/octet-stream        
+    - name: Upload Release Asset - Zip file
+      if: env.release_flag == 1         
+      id: upload-release-asset-zip
+      uses: actions/upload-release-asset@v1
+      env:
+        GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+      with:
+        upload_url: ${{ steps.create_release.outputs.upload_url }} # This pulls from the CREATE RELEASE step above, referencing it's ID to get its outputs object, which include a `upload_url`. See this blog post for more info: https://jasonet.co/posts/new-features-of-github-actions/#passing-data-to-future-steps 
+        asset_path: build/${{ env.artifact_file_name }}
+        asset_name: ${{ env.artifact_file_name }}
+        asset_content_type: application/octet-stream          

.github/workflows/web_deploy.yml

@@ -0,0 +1,23 @@
+name: Update Web Installer
+on:
+  workflow_call:
+  workflow_dispatch:
+jobs:
+  update_web_installer:
+    name: Update Web Installer After Release
+    runs-on: ubuntu-latest
+    container:
+      image: sle118/squeezelite-esp32-idfv43
+    env:
+      WEB_INSTALLER: ${{ secrets.WEB_INSTALLER }}
+    steps:
+      - uses: actions/checkout@v2
+        with:
+          fetch-depth: 0
+          submodules: true    
+      - name: Update Web Installer Project
+        run: |
+          . /opt/esp/python_env/idf4.3_py3.8_env/bin/activate
+          git config --global --add safe.directory /__w/squeezelite-esp32/squeezelite-esp32   
+          build_tools.py manifest --flash_file "/build/flash_project_args" --outdir "./bin_files" --manif_name "manifest" --max_count 3
+          build_tools.py pushinstaller --source "./bin_files" --manif_name "manifest" --target "web-installer" --url "https://github.com/sle118/squeezelite-esp32-installer.git" --artifacts "docs/artifacts" --web_installer_branch "main" --token "${{env.WEB_INSTALLER}}"

.gitignore

@@ -1,71 +1,30 @@
-# Prerequisites
-*.d
-
-# Compiled Object files
-*.slo
-*.lo
-*.o
-*.obj
-
-# Precompiled Headers
-*.gch
-*.pch
-
-# Compiled Dynamic libraries
-*.so
-*.dylib
-*.dll
-
-# Fortran module files
-*.mod
-*.smod
-
-# Compiled Static libraries
-*.lai
-*.la
-*.lib
-
-# Executables
-*.exe
-*.out
-*.app
-
-# Build files with potential private info
 build/
+.vscode/c_cpp_properties.json
+.vscode/launch.json
+.vscode/settings.json
+.vscode/tasks.json
+alltags.txt
+components/wifi-manager/network_manager_handlers.multi
+esp32.code-workspace
 sdkconfig.old
+test/.vscode/c_cpp_properties.json
+test/.vscode/launch.json
+test/.vscode/settings.json
+test/.vscode/tasks.json
+test/sdkconfig
+components/wifi-manager/UML-State-Machine-in-C
+*.bak
 
-# =========================
-# Operating System Files
-# =========================
+envfile.txt
+artifacts
+web-installer
 
-# Windows
-# =========================
+squeezelite-esp32.code-workspace
 
-# Windows thumbnail cache files
-Thumbs.db
-ehthumbs.db
-ehthumbs_vista.db
+esp-idf-vscode-generated.gdb
 
-# Folder config file
-Desktop.ini
+debug.log
 
-# Recycle Bin used on file shares
-$RECYCLE.BIN/
+components/wifi-manager/esp32_improv.cpp.txt
 
-# Windows Installer files
-*.cab
-*.msi
-*.msm
-*.msp
-
-# Windows shortcuts
-*.lnk
-*.save
-libs/
-
-/cdump.cmd
-/_*
-squeezelite-esp32-jsonblob.zip
-/flash_cmd.txt
-/writeSequeezeEsp.bat
-/writeSequeezeEsp.sh
+components/wifi-manager/esp32_improv.h.txt

.gitmodules

@@ -5,3 +5,6 @@
 [submodule "components/esp-dsp"]
 	path = components/esp-dsp
 	url = https://github.com/philippe44/esp-dsp.git
+[submodule "components/wifi-manager/UML-State-Machine-in-C"]
+	path = components/wifi-manager/UML-State-Machine-in-C
+	url = https://github.com/kiishor/UML-State-Machine-in-C

.pydevproject

@@ -1,5 +0,0 @@
-<?xml version="1.0" encoding="UTF-8" standalone="no"?>
-<?eclipse-pydev version="1.0"?><pydev_project>
-    <pydev_property name="org.python.pydev.PYTHON_PROJECT_INTERPRETER">Default</pydev_property>
-    <pydev_property name="org.python.pydev.PYTHON_PROJECT_VERSION">python interpreter</pydev_property>
-</pydev_project>

CMakeLists.txt

@@ -1,15 +1,128 @@
 cmake_minimum_required(VERSION 3.5)
 set(EXTRA_COMPONENT_DIRS components/platform_console/app_recovery components/platform_console/app_squeezelite  )
 include($ENV{IDF_PATH}/tools/cmake/project.cmake)
+add_definitions(-DMODEL_NAME=SqueezeESP32)
 
+# State machine hierarchy enabled and logging enabled
+add_definitions(-DSTATE_MACHINE_LOGGER=1)
+add_definitions(-DHIERARCHICAL_STATES=1)
+
+# Uncomment line below to get memory usage trace details
+#add_definitions(-DENABLE_MEMTRACE=1)
+#uncomment line below to get network ethernet debug logs 
+#add_definitions(-DNETWORK_ETHERNET_LOG_LEVEL=ESP_LOG_DEBUG)
+#uncomment line below to get network status debug logs
+#add_definitions(-DNETWORK_STATUS_LOG_LEVEL=ESP_LOG_DEBUG)
+#add_definitions(-DNETWORK_HANDLERS_LOG_LEVEL=ESP_LOG_DEBUG) 
+#add_definitions(-DNETWORK_WIFI_LOG_LEVEL=ESP_LOG_DEBUG)
+#add_definitions(-DNETWORK_MANAGER_LOG_LEVEL=ESP_LOG_DEBUG)
+#add_definitions(-DNETWORK_HTTP_SERVER_LOG_LEVEL=ESP_LOG_DEBUG)
+
+if(NOT DEFINED DEPTH)
+	set(DEPTH "16")
+endif()	
 message(STATUS  "Building RECOVERY")
 project(recovery)
 set_property(TARGET recovery.elf PROPERTY RECOVERY_PREFIX app_recovery )
-
 include(squeezelite.cmake)
-
-
 set(PROJECT_VER $ENV{PROJECT_VER})
-   
-#target_compile_definitions(__idf_squeezelite-ota PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_VERBOSE)
-#target_compile_definitions(__idf_driver_bt PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_VERBOSE)
+
+#target_compile_definitions(__idf_esp_eth PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_INFO)
+
+#target_compile_definitions(__idf_services PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+
+#target_compile_definitions(__idf_driver PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+#target_compile_definitions(__idf_wifi-manager PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+#target_compile_definitions(__idf_esp_wifi PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+#target_compile_definitions(__idf_platform_console PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+
+#target_compile_definitions(__idf_app_recovery PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_INFO)
+# target_compile_definitions(__idf_esp_eth PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_INFO)
+# target_compile_definitions(__idf_esp_event PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_INFO)
+# target_compile_definitions(__idf_esp_netif PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+
+# target_compile_definitions(__idf_freertos PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+#target_compile_definitions(__idf_bt PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_ERROR)
+# target_compile_definitions(__idf_mdns PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+# target_compile_definitions(__idf_tcpip_adapter PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+# target_compile_definitions(__idf_tcp_transport PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+
+#target_compile_definitions(__idf_app_squeezelite PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+# target_compile_definitions(__idf_app_trace PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+# target_compile_definitions(__idf_app_update PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+# target_compile_definitions(__idf_asio PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+# target_compile_definitions(__idf_audio PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+# target_compile_definitions(__idf_bootloader_support PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+
+# target_compile_definitions(__idf_cbor PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+# target_compile_definitions(__idf_cmock PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+# target_compile_definitions(__idf_coap PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+# target_compile_definitions(__idf_console PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+# target_compile_definitions(__idf_cxx PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+# target_compile_definitions(__idf_display PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+# target_compile_definitions(__idf_driver PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+# target_compile_definitions(__idf_driver_bt PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+# target_compile_definitions(__idf_efuse PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+# target_compile_definitions(__idf_esp-dsp PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+# target_compile_definitions(__idf_esp-tls PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+# target_compile_definitions(__idf_esp32 PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+# target_compile_definitions(__idf_espcoredump PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+# target_compile_definitions(__idf_esp_adc_cal PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+# target_compile_definitions(__idf_esp_common PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+# target_compile_definitions(__idf_esp_gdbstub PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+# target_compile_definitions(__idf_esp_hid PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+# target_compile_definitions(__idf_esp_https_ota PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+# target_compile_definitions(__idf_esp_http_client PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+# target_compile_definitions(__idf_esp_http_server PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+# target_compile_definitions(__idf_esp_hw_support PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+# target_compile_definitions(__idf_esp_ipc PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+# target_compile_definitions(__idf_esp_local_ctrl PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+# target_compile_definitions(__idf_esp_pm PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+# target_compile_definitions(__idf_esp_ringbuf PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+# target_compile_definitions(__idf_esp_rom PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+# target_compile_definitions(__idf_esp_serial_slave_link PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+# target_compile_definitions(__idf_esp_system PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+# target_compile_definitions(__idf_esp_timer PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+# target_compile_definitions(__idf_esp_websocket_client PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+# target_compile_definitions(__idf_expat PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+# target_compile_definitions(__idf_fatfs PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+# target_compile_definitions(__idf_freemodbus PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+# target_compile_definitions(__idf_hal PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+# target_compile_definitions(__idf_heap PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+# target_compile_definitions(__idf_jsmn PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+# target_compile_definitions(__idf_json PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+# target_compile_definitions(__idf_libsodium PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+# target_compile_definitions(__idf_log PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+# target_compile_definitions(__idf_lwip PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+# target_compile_definitions(__idf_main PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+# target_compile_definitions(__idf_mbedtls PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+# target_compile_definitions(mbedcrypto PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+# target_compile_definitions(mbedtls PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+# target_compile_definitions(mbedx509 PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+# target_compile_definitions(__idf_mqtt PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+# target_compile_definitions(__idf_newlib PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+# target_compile_definitions(__idf_nghttp PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+# target_compile_definitions(__idf_nvs_flash PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+# target_compile_definitions(__idf_openssl PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+# target_compile_definitions(__idf_perfmon PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+# target_compile_definitions(__idf_platform_config PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+# target_compile_definitions(__idf_protobuf-c PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+# target_compile_definitions(__idf_protocomm PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+# target_compile_definitions(__idf_pthread PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+# target_compile_definitions(__idf_raop PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+# target_compile_definitions(__idf_sdmmc PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+
+# target_compile_definitions(__idf_soc PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+# target_compile_definitions(__idf_spiffs PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+# target_compile_definitions(__idf_spi_flash PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+# target_compile_definitions(__idf_squeezelite PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+# target_compile_definitions(__idf_squeezelite-ota PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+# target_compile_definitions(__idf_telnet PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+# target_compile_definitions(__idf_tools PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+# target_compile_definitions(__idf_ulp PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+# target_compile_definitions(__idf_unity PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+# target_compile_definitions(__idf_vfs PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+# target_compile_definitions(__idf_wear_levelling PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+# target_compile_definitions(__idf_wifi_provisioning PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+# target_compile_definitions(__idf_wpa_supplicant PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)
+# target_compile_definitions(__idf_xtensa PRIVATE -DLOG_LOCAL_LEVEL=ESP_LOG_DEBUG)

Dockerfile

@@ -1,35 +1,156 @@
-FROM ubuntu:18.04
+FROM ubuntu:20.04
 
-RUN apt-get update && apt-get install -y git wget libncurses-dev flex bison gperf \
-  python python-pip python-setuptools python-serial python-click \
-  python-cryptography python-future python-pyparsing \
-  python-pyelftools cmake ninja-build ccache libusb-1.0
 
-RUN mkdir /workspace
-WORKDIR /workspace
+ARG DEBIAN_FRONTEND=noninteractive
+ENV GCC_TOOLS_BASE=/opt/esp/tools/xtensa-esp32-elf/esp-2021r2-8.4.0/xtensa-esp32-elf/bin/xtensa-esp32-elf-
+# To build the image for a branch or a tag of IDF, pass --build-arg IDF_CLONE_BRANCH_OR_TAG=name.
+# To build the image with a specific commit ID of IDF, pass --build-arg IDF_CHECKOUT_REF=commit-id.
+# It is possibe to combine both, e.g.:
+#   IDF_CLONE_BRANCH_OR_TAG=release/vX.Y
+#   IDF_CHECKOUT_REF=<some commit on release/vX.Y branch>.
+# The following commit contains the ldgen fix: eab738c79e063b3d6f4c345ea5e1d4f8caef725b
+# to build an image using that commit: docker build . --build-arg IDF_CHECKOUT_REF=eab738c79e063b3d6f4c345ea5e1d4f8caef725b -t sle118/squeezelite-esp32-idfv43
+# Docker build for release 4.3.2 as of 2022/02/28
+# docker build . --build-arg IDF_CHECKOUT_REF=8bf14a9238329954c7c5062eeeda569529aedf75 -t sle118/squeezelite-esp32-idfv43
+# To run the image interactive (windows): 
+# docker run --rm -v %cd%:/project -w /project -it sle118/squeezelite-esp32-idfv43
+# To run the image interactive (linux):
+# docker run --rm -v `pwd`:/project -w /project -it sle118/squeezelite-esp32-idfv4-master
+# to build the web app inside of the interactive session
+# pushd components/wifi-manager/webapp/ && npm install && npm run-script build && popd
+#
+# to run the docker with netwotrk port published on the host:
+# docker run --rm -p 5000:5000/tcp -v %cd%:/project -w /project -it sle118/squeezelite-esp32-idfv43
 
-# Download and checkout known good esp-idf commit
-RUN git clone --recursive https://github.com/espressif/esp-idf.git esp-idf
-RUN cd esp-idf && git checkout 4dac7c7df885adaa86a5c79f2adeaf8d68667349
-RUN git clone https://github.com/sle118/squeezelite-esp32.git
+ARG IDF_CLONE_URL=https://github.com/espressif/esp-idf.git
+ARG IDF_CLONE_BRANCH_OR_TAG=master
+ARG IDF_CHECKOUT_REF=8bf14a9238329954c7c5062eeeda569529aedf75
 
-# Download GCC 5.2.0
-RUN wget https://dl.espressif.com/dl/xtensa-esp32-elf-linux64-1.22.0-80-g6c4433a-5.2.0.tar.gz
-RUN tar -xzf xtensa-esp32-elf-linux64-1.22.0-80-g6c4433a-5.2.0.tar.gz
-RUN rm xtensa-esp32-elf-linux64-1.22.0-80-g6c4433a-5.2.0.tar.gz
+ENV IDF_PATH=/opt/esp/idf
+ENV IDF_TOOLS_PATH=/opt/esp
 
-# Patch I2S in esp-idf
-RUN cp /workspace/squeezelite-esp32/idf-patch/i2s.c /workspace/esp-idf/components/driver/i2s.c
-RUN rm -r /workspace/squeezelite-esp32
-RUN mkdir /workspace/squeezelite-esp32
+# We need libpython2.7 due to GDB tools
+# we also need npm 8 for the webapp to work
+RUN : \
+  && apt-get update \
+  && apt-get install -y \
+    apt-utils \
+    bison \
+    ca-certificates \
+    ccache \
+    check \
+    curl \
+    flex \
+    git \
+    gperf \
+    lcov \
+    libffi-dev \
+    libncurses-dev \
+    libpython2.7 \
+    libusb-1.0-0-dev \
+    make \
+    ninja-build \
+    python3 \
+    python3-pip \
+    unzip \
+    wget \
+    xz-utils \
+    zip \
+    	npm \
+	nodejs \
+  && apt-get autoremove -y \
+  && rm -rf /var/lib/apt/lists/* \
+  && update-alternatives --install /usr/bin/python python /usr/bin/python3 10 \
+  && python -m pip install --upgrade \
+    pip \
+    virtualenv \
+  && cd /opt \  
+  && git clone https://github.com/HBehrens/puncover.git \
+  && cd puncover \
+  && python setup.py -q install \
+  && echo IDF_CHECKOUT_REF=$IDF_CHECKOUT_REF IDF_CLONE_BRANCH_OR_TAG=$IDF_CLONE_BRANCH_OR_TAG \
+  && git clone --recursive \
+      ${IDF_CLONE_BRANCH_OR_TAG:+-b $IDF_CLONE_BRANCH_OR_TAG} \
+      $IDF_CLONE_URL $IDF_PATH \
+	&& if [ -n "$IDF_CHECKOUT_REF" ]; then \
+      cd $IDF_PATH \
+  &&  git checkout $IDF_CHECKOUT_REF \
+  &&  git submodule update --init --recursive; \
+    fi \
+  && update-ca-certificates --fresh \
+  && $IDF_PATH/tools/idf_tools.py --non-interactive install required \
+  && $IDF_PATH/tools/idf_tools.py --non-interactive install cmake \
+  && $IDF_PATH/tools/idf_tools.py --non-interactive install-python-env \
+  && :
+RUN : \
+  echo Installing pygit2  ******************************************************** \
+  && . /opt/esp/python_env/idf4.3_py3.8_env/bin/activate \
+  && ln -sf /opt/esp/python_env/idf4.3_py3.8_env/bin/python  /usr/local/bin/python \
+  && pip install pygit2 requests \
+  && pip show pygit2 \ 
+  && python --version \  
+  && pip --version \
+  && pip3 install protobuf grpcio-tools \
+  && rm -rf $IDF_TOOLS_PATH/dist \
+  && :
 
-# Setup PATH to use esp-idf and gcc-5.2.0
-RUN touch /root/.bashrc && \
- echo export PATH="\$PATH:/workspace/xtensa-esp32-elf/bin" >> /root/.bashrc && \
- echo export IDF_PATH=/workspace/esp-idf >> /root/.bashrc
+COPY docker/patches $IDF_PATH
 
-# OPTIONAL: Install vim for text editing in Bash
-RUN apt-get update && apt-get install -y vim
+#set idf environment variabies
+ENV PATH /opt/esp/idf/components/esptool_py/esptool:/opt/esp/idf/components/espcoredump:/opt/esp/idf/components/partition_table:/opt/esp/idf/components/app_update:/opt/esp/tools/xtensa-esp32-elf/esp-2021r2-8.4.0/xtensa-esp32-elf/bin:/opt/esp/tools/xtensa-esp32s2-elf/esp-2021r2-8.4.0/xtensa-esp32s2-elf/bin:/opt/esp/tools/xtensa-esp32s3-elf/esp-2021r2-8.4.0/xtensa-esp32s3-elf/bin:/opt/esp/tools/riscv32-esp-elf/esp-2021r2-8.4.0/riscv32-esp-elf/bin:/opt/esp/tools/esp32ulp-elf/2.28.51-esp-20191205/esp32ulp-elf-binutils/bin:/opt/esp/tools/esp32s2ulp-elf/2.28.51-esp-20191205/esp32s2ulp-elf-binutils/bin:/opt/esp/tools/cmake/3.16.4/bin:/opt/esp/tools/openocd-esp32/v0.10.0-esp32-20211111/openocd-esp32/bin:/opt/esp/python_env/idf4.3_py3.8_env/bin:/opt/esp/idf/tools:$PATH
+ENV GCC_TOOLS_BASE="/opt/esp/tools/xtensa-esp32-elf/esp-2021r2-8.4.0/xtensa-esp32-elf/bin/xtensa-esp32-elf-"
+ENV IDF_PATH="/opt/esp/idf"
+ENV IDF_PYTHON_ENV_PATH="/opt/esp/python_env/idf4.3_py3.8_env"
+ENV IDF_TOOLS_EXPORT_CMD="/opt/esp/idf/export.sh"
+ENV IDF_TOOLS_INSTALL_CMD="/opt/esp/idf/install.sh"
+ENV IDF_TOOLS_PATH="/opt/esp"
+ENV NODE_PATH="/v8/lib/node_modules"
+ENV NODE_VERSION="8"
+ENV OPENOCD_SCRIPTS="/opt/esp/tools/openocd-esp32/v0.10.0-esp32-20211111/openocd-esp32/share/openocd/scripts"
+# Ccache is installed, enable it by default
 
-WORKDIR /workspace/squeezelite-esp32
-CMD ["bash"]
+ENV IDF_CCACHE_ENABLE=1
+COPY docker/entrypoint.sh /opt/esp/entrypoint.sh
+COPY components/wifi-manager/webapp/package.json /opt
+
+ENV NODE_VERSION 8
+
+SHELL ["/bin/bash", "--login", "-c"]
+# Install nvm with node and npm
+# RUN wget -qO- https://raw.githubusercontent.com/nvm-sh/nvm/v0.39.1/install.sh | bash \
+#     && export NVM_DIR="$([ -z "${XDG_CONFIG_HOME-}" ] && printf %s "${HOME}/.nvm" || printf %s "${XDG_CONFIG_HOME}/nvm")" \
+#     && [ -s "$NVM_DIR/nvm.sh" ] && \. "$NVM_DIR/nvm.sh" \
+#     && nvm install $NODE_VERSION \
+#     && nvm alias default $NODE_VERSION \
+#     && nvm use default \
+#     && echo installing nodejs version 16  \
+#     && curl -sL https://deb.nodesource.com/setup_16.x | bash - \
+#     && echo installing node modules  \
+#     && cd /opt \
+#     && nvm use default \
+#     && npm install -g \  
+#     && :    
+
+RUN : \
+  && curl -fsSL https://deb.nodesource.com/setup_16.x | bash - \
+  && apt-get install -y nodejs jq \
+  && echo installing dev node modules globally \
+  && cd /opt \
+  && cat ./package.json | jq '.devDependencies | keys[] as $k | "\($k)@\(.[$k])"' | xargs -t npm install --global \
+  && echo installing npm global packages \
+  && npm i -g npm \
+  && node --version \
+  && npm install -g \  
+  && :      
+
+ENV NODE_PATH $NVM_DIR/v$NODE_VERSION/lib/node_modules
+ENV PATH $IDF_PYTHON_ENV_PATH:$NVM_DIR/v$NODE_VERSION/bin:$PATH
+COPY ./docker/build_tools.py /usr/sbin/build_tools.py
+RUN : \
+  && echo Changing permissions ********************************************************  \
+  && chmod +x /opt/esp/entrypoint.sh \
+  && chmod +x /usr/sbin/build_tools.py \  
+  && :
+
+ENTRYPOINT [ "/opt/esp/entrypoint.sh" ]
+CMD [ "/bin/bash" ]

README.md

@@ -1,20 +1,91 @@
+![Cross-Build](https://github.com/sle118/squeezelite-esp32/workflows/Cross-Build/badge.svg?branch=master-cmake)
+![ESP-IDF v4.3.1](https://github.com/sle118/squeezelite-esp32/actions/workflows/esp-idf-v4.3-build.yml/badge.svg?branch=master-v4.3)
 # Squeezelite-esp32
-## Supported Hardware 
-### SqueezeAMP
-Works with the SqueezeAMP see [here](https://forums.slimdevices.com/showthread.php?110926-pre-ANNOUNCE-SqueezeAMP-and-SqueezeliteESP32) and [here](https://github.com/philippe44/SqueezeAMP/blob/master/README.md). Add repository https://raw.githubusercontent.com/sle118/squeezelite-esp32/master/plugin/repo.xml to LMS if you want to have a display
+## What is this
+Squeezelite-esp32 is an audio software suite made to run on espressif's ESP32 wifi (b/g/n) and bluetooth chipset. It offers the following capabilities
 
-Use the `squeezelite-esp32-SqueezeAmp-sdkconfig.defaults` configuration file.
+- Stream your local music and connect to all major on-line music providers (Spotify, Deezer, Tidal, Qobuz) using [Logitech Media Server - a.k.a LMS](https://forums.slimdevices.com/) and enjoy multi-room audio synchronization. LMS can be extended by numerous plugins and can be controlled using a Web browser or dedicated applications (iPhone, Android). It can also send audio to UPnP, Sonos, ChromeCast and AirPlay speakers/devices.
+- Stream from a Bluetooth device (iPhone, Android)
+- Stream from an AirPlay controller (iPhone, iTunes ...) and enjoy synchronization multiroom as well (although it's AirPlay 1 only)
+
+Depending on the hardware connected to the ESP32, you can send audio to a local DAC, to SPDIF or to a Bluetooth speaker. The bare minimum required hardware is a WROVER module with 4MB of Flash and 4MB of PSRAM (https://www.espressif.com/en/products/modules/esp32). With that module standalone, just apply power and you can stream to a Bluetooth speaker. You can also send audio to most I2S DAC as well as to SPDIF receivers using just a cable or an optical transducer.
+
+But squeezelite-esp32 is highly extensible and you can add
+
+- Buttons and Rotary Encoder and map/combine them to various functions (play, pause, volume, next ...)
+- IR receiver (no pullup resistor or capacitor needed, just the 38kHz receiver)
+- Monochrome, GrayScale or Color displays using SPI or I2C (supported drivers are SH1106, SSD1306, SSD1322, SSD1326/7, SSD1351, ST7735, ST7789 and ILI9341).
+- Ethernet using a Microchip LAN8720 with RMII interface or Davicom DM9051 over SPI.
+
+Other features include
+
+ - Resampling 
+ - 10-bands equalizer
+ - Automatic initial setup using any WiFi device 
+ - Full web interface for further configuration/management
+ - Firmware over-the-air update 
+
+To control the equalizer or use the display on LMS, a new player model is required and this is provided through a plugin that is part of LMS' 3rd party repositories
+
+## Performances 
+*(opinions presented here so I = @philippe44)*
+The main build of squeezelite-esp32 is a 16 bits internal core with all calculations in 32 bits or float precision. This is a design choice I've made to preserve CPU performances (it is already stretching a lot the esp32 chipset) and optimize memory usage as we only have 4MB of usable RAM. Some might correctly comment that the WROVER module have 8MB of RAM, but the processor is only able to address 4MB and the remaining 4MB must be paginated by smaller blocks and I don't have patience to that. 
+
+Now, when I did the porting of squeezelite to esp32, I've also made the core 16 or 32 bits compatible at compile-time. So far, it works in 32 bits but less tests have been done. You can chose to compile it in 32 bits mode. I'm not very interested above 16 bits samples because it does not bring anything (I have an engineering background in theory of information). 
+
+| Capability                 |16 bits|32 bits| comment         						          |	
+|----------------------------|-------|-------|--------------------------------------------------------------------|
+| max sampling rate          |  192k |  96k  | 192k is very challenging, especially when combined with display    |
+| max bit depth              |  16   |  24   | 24 bits are truncated in 16 bits mode                              | 
+| spdif			     |16 bits|20 bits|		                                                          |
+| mp3, aac, opus, ogg/vorbis |  48k  |  48k  |		                                                          |
+| alac, flac, ogg/flac       |  96k  |  96k  | 		                                                          |
+| pcm, wav, aif              |  192k |  96k  |		                                                          |
+| equalizer                  |   Y   |   N   | 48kHz max (after resampling) - equalization skipped on >48k tracks |
+| resampling                 |   Y   |   N   |		                                                          |
+| cross-fade                 |  10s  |  <5s  | depends on buffer size and sampling rate		                  |
+
+The esp32 must run at 240 MHz, with Quad-SPI I/O at 80 MHz and a clock of 40 Mhz. Still, it's a lot to run, especially knowing that it has a serial Flash and PSRAM, so kudos to Espressif for their chipset optimization. Now, to have all the decoding, resampling, equalizing, gain, display, spectrum/vu is a very (very) delicate equilibrium between use of internal /external RAM, tasks priorities and buffer handling. It is not perfect and the more you push the system to the limit, the higher the risk that some files would not play (see below). In general, the display will always have the lowest priority and you'll notice slowdown in scrolling and VU/Spectrum refresh rates. Now, even display thread has some critical section and impacts the capabilities. For example, a 16 bits-depth color display with low SPI speed might prevent 24/96 flac to work but still work with pcm 24/96
+
+In 16 bits mode, although 192 kHz is reported as max rate, it's highly recommended to limit reported sampling rate to 96k (-Z 96000). Note that some high-speed 24/96k on-line streams might stutter because of TCP/IP stack performances. It is usually due to the fact that the server sends small packets of data and the esp32 cannot receive encoded audio fast enough, regardless of task priority settings (I've tried to tweak that a fair bit). The best option in that case is to let LMS proxy the stream as it will provide larger chunks and a "smoother" stream that can then be handled.
+
+Note as well that some codecs consume more CPU than others or have not been optimized as much. I've done my best to tweak these, but that level of optimization includes writing some assembly which is painful. One very demanding codec is AAC when files are encoded with SBR. It allows reconstruction of upper part of spectrum and thus higher sampling rate, but the codec spec is such that this is optional, you can decode simply lower band and accept lower sampling rate - See the AAC_DISABLE_SBR option below.
+## Supported Hardware
+Any esp32-based hardware with at least 4MB of flash and 4MB of PSRAM will be capable of running squeezelite-esp32 and there are various boards that include such chip. A few are mentionned below, but any should work. You can find various help & instructions [here](https://forums.slimdevices.com/showthread.php?112697-ANNOUNCE-Squeezelite-ESP32-(dedicated-thread))
+
+**For the sake of clarity, WROOM modules DO NOT work as they don't include PSRAM. Some designs might add it externally, but it's (very) unlikely.**
+### Raw WROVER module
+Per above description, a [WROVER module](https://www.espressif.com/en/products/modules/esp32) is enough to run Squeezelite-esp32, but that requires a bit of tinkering to extend it to have analogue audio or hardware buttons (e.g.) 
+
+Please note that when sending to a Bluetooth speaker (source), only 44.1 kHz can be used, so you either let LMS do the resampling, but you must make sure it only sends 44.1kHz tracks or enable internal resampling (using -R) option. If you connect a DAC, choice of sample rates will depends on its capabilities. See below for more details.
+
+Most DAC will work out-of-the-box with simply an I2S connection, but some require specific commands to be sent using I2C. See DAC option below to understand how to send these dedicated commands. There is build-in support for TAS575x, TAS5780, TAS5713 and AC101 DAC.
+### SqueezeAMP
+This is the main hardware companion of Squeezelite-esp32 and has been developped together. Details on capabilities can be found [here](https://forums.slimdevices.com/showthread.php?110926-pre-ANNOUNCE-SqueezeAMP-and-SqueezeliteESP32) and [here](https://github.com/philippe44/SqueezeAMP).
+
+if you want to rebuild, use the `squeezelite-esp32-SqueezeAmp-sdkconfig.defaults` configuration file.
+
+NB: You can use the pre-build binaries SqueezeAMP4MBFlash which has all the hardware I/O set properly. You can also use the generic binary I2S4MBFlash in which case the NVS parameters shall be set to get the exact same behavior
+- set_GPIO: 12=green,13=red,34=jack,2=spkfault
+- batt_config: channel=7,scale=20.24
+- dac_config: model=TAS57xx,bck=33,ws=25,do=32,sda=27,scl=26,mute=14:0
+- spdif_config: bck=33,ws=25,do=15
 
 ### ESP32-A1S
-Works with [ESP32-A1S](https://docs.ai-thinker.com/esp32-a1s) module that includes audio codec and headset output. You still need to use a demo board like [this](https://www.aliexpress.com/item/4000765857347.html?spm=2114.12010615.8148356.11.5d963cd0j669ns) or an external amplifier if you want direct speaker connection. 
+Works with [ESP32-A1S](https://docs.ai-thinker.com/esp32-a1s) module that includes audio codec and headset output. You still need to use a demo board like [this](https://www.aliexpress.com/item/4001060963585.html) or an external amplifier if you want direct speaker connection. Note that there is a version with AC101 codec and another one with ES8388 (see below)
 
-The board showed above has the following IO set
+The board shown above has the following IO set
 - amplifier: GPIO21
 - key2: GPIO13, key3: GPIO19, key4: GPIO23, key5: GPIO18, key6: GPIO5 (to be confirmed with dip switches)
-- key1: not sure, something with GPIO36
+- key1: not sure, using GPIO36 in a matrix
 - jack insertion: GPIO39 (inserted low)
-- LED: GPIO22 (active low)
-(note that GPIO need pullups)
+- D4 -> GPIO22 used for green LED (active low)
+- D5 -> GPIO19 (muxed with key3)
+- The IO connector also brings GPIO5, GPIO18, GPIO19, GPIO21, GPIO22 and GPIO23 (don't forget it's muxed with keys!)
+- The JTAG connector uses GPIO 12, 13, 14 and 15 (see dip switch) but these are also used for SD-card (and GPIO13 is key2 as well)
+- It's always possible to re-use GPIOO (download at boot) and GPIO1/GPIO3 which are RX/TX of UART0 but you'll lose trace
+
+(note that some GPIO need pullups)
 
 So a possible config would be
 - set_GPIO: 21=amp,22=green:0,39=jack:0
@@ -22,7 +93,18 @@ So a possible config would be
 ```
 [{"gpio":5,"normal":{"pressed":"ACTRLS_TOGGLE"}},{"gpio":18,"pull":true,"shifter_gpio":5,"normal":{"pressed":"ACTRLS_VOLUP"}, "shifted":{"pressed":"ACTRLS_NEXT"}}, {"gpio":23,"pull":true,"shifter_gpio":5,"normal":{"pressed":"ACTRLS_VOLDOWN"},"shifted":{"pressed":"ACTRLS_PREV"}}]
 ```
+for AC101
+- dac_config: model=AC101,bck=27,ws=26,do=25,di=35,sda=33,scl=32
+ 
+for ES8388
+- dac_config: model=ES8388,bck=5,ws=25,do=26,sda=18,scl=23,i2c=16
+### T-WATCH2020 by LilyGo
+This is a fun [smartwatch](http://www.lilygo.cn/prod_view.aspx?TypeId=50036&Id=1290&FId=t3:50036:3) based on ESP32. It has a 240x240 ST7789 screen and onboard audio. Not very useful to listen to anything but it works. This is an example of a device that requires an I2C set of commands for its dac (see below). There is a build-option if you decide to rebuild everything by yourself, otherwise the I2S default option works with the following parameters
 
+- dac_config: model=I2S,bck=26,ws=25,do=33,i2c=106,sda=21,scl=22
+- dac_controlset: { "init": [ {"reg":41, "val":128}, {"reg":18, "val":255} ], "poweron": [ {"reg":18, "val":64, "mode":"or"} ], "poweroff": [ {"reg":18, "val":191, "mode":"and"} ] }
+- spi_config: dc=27,data=19,clk=18
+- display_config: SPI,driver=ST7789,width=240,height=240,cs=5,back=12,speed=16000000,HFlip,VFlip
 ### ESP32-WROVER + I2S DAC
 Squeezelite-esp32 requires esp32 chipset and 4MB PSRAM. ESP32-WROVER meets these requirements. To get an audio output an I2S DAC can be used. Cheap PCM5102 I2S DACs work others may also work. PCM5012 DACs can be hooked up via:
 
@@ -41,6 +123,10 @@ XMT - 3.3V
 
 Use the `squeezelite-esp32-I2S-4MFlash-sdkconfig.defaults` configuration file.
 
+### SqueezeAmpToo !
+
+And the super cool project https://github.com/rochuck/squeeze-amp-too
+
 ## Configuration
 To access NVS, in the webUI, go to credits and select "shows nvs editor". Go into the NVS editor tab to change NFS parameters. In syntax description below \<\> means a value while \[\] describe optional parameters. 
 
@@ -49,16 +135,31 @@ The NVS parameter "i2c_config" set the i2c's gpio used for generic purpose (e.g.
 ```
 sda=<gpio>,scl=<gpio>[,port=0|1][,speed=<speed>]
 ```
+<strong>Please note that you can not use the same GPIO or port as the DAC</strong>
 ### SPI
-The NVS parameter "spi_config" set the spi's gpio used for generic purpose (e.g. display). Leave it blank to disable SPI usage. The DC parameter is needed for displays. Syntax is
-```
-data=<gpio>,clk=<gpio>[,dc=<gpio>][,host=1|2]
+The esp32 has 4 SPI sub-systems, one is unaccessible so numbering is 0..2 and SPI0 is reserved for Flash/PSRAM. The NVS parameter "spi_config" set the spi's gpio used for generic purpose (e.g. display). Leave it blank to disable SPI usage. The DC parameter is needed for displays. Syntax is
 ```
+data|mosi=<gpio>,clk=<gpio>[,dc=<gpio>][,host=1|2][,miso=<gpio>]
+``` 
+Default "host" is 1. The "miso" parameter is only used when SPI bus is to be shared with other peripheral (e.g. ethernet, see below), otherwise it can be omitted. Note that "data" can also be named "mosi". 
 ### DAC/I2S
-The NVS parameter "dac_config" set the gpio used for i2s communication with your DAC. You can define the defaults at compile time but nvs parameter takes precedence except for SqueezeAMP and A1S where these are forced at runtime. If your DAC also requires i2c, then you must go the re-compile route. Syntax is
+The NVS parameter "dac_config" set the gpio used for i2s communication with your DAC. You can define the defaults at compile time but nvs parameter takes precedence except for SqueezeAMP and A1S where these are forced at runtime. Syntax is
 ```
-bck=<gpio>,ws=<gpio>,do=<gpio>
+bck=<gpio>,ws=<gpio>,do=<gpio>[,mck][,mute=<gpio>[:0|1][,model=TAS57xx|TAS5713|AC101|I2S][,sda=<gpio>,scl=gpio[,i2c=<addr>]]
 ```
+if "model" is not set or is not recognized, then default "I2S" is used. The option "mck" is used for some codecs that require a master clock (although they should not). Only GPIO0 can be used as MCLK and be aware that this cannot coexit with RMII Ethernet (see ethernet section below). I2C parameters are optional and only needed if your DAC requires an I2C control (See 'dac_controlset' below). Note that "i2c" parameters are decimal, hex notation is not allowed.
+
+So far, TAS57xx, TAS5713, AC101, WM8978 and ES8388 are recognized models where the proper init sequence/volume/power controls are sent. For other codecs that might require an I2C commands, please use the parameter "dac_controlset" that allows definition of simple commands to be sent over i2c for init, power on and off using a JSON syntax:
+```
+{ init: [ {"reg":<register>,"val":<value>,"mode":<nothing>|"or"|"and"}, ... {{"reg":<register>,"val":<value>,"mode":<nothing>|"or"|"and"} ],
+  poweron: [ {"reg":<register>,"val":<value>,"mode":<nothing>|"or"|"and"}, ... {{"reg":<register>,"val":<value>,"mode":<nothing>|"or"|"and"} ],
+  poweroff: [ {"reg":<register>,"val":<value>,"mode":<nothing>|"or"|"and"}, ... {{"reg":<register>,"val":<value>,"mode":<nothing>|"or"|"and"} ] }
+```
+This is standard JSON notation, so if you are not familiar with it, Google is your best friend. Be aware that the '...' means you can have as many entries as you want, it's not part of the syntax. Every section is optional, but it does not make sense to set i2c in the 'dac_config' parameter and not setting anything here. The parameter 'mode' allows to *or* the register with the value or to *and* it. Don't set 'mode' if you simply want to write. **Note that all values must be decimal**. You can use a validator like [this](https://jsonlint.com) to verify your syntax
+
+NB: For specific builds (all except I2S), all this is ignored. For know codecs, the built-in sequences can be overwritten using dac_controlset
+
+<strong>Please note that you can not use the same GPIO or port as the I2C</strong>
 ### SPDIF
 The NVS parameter "spdif_config" sets the i2s's gpio needed for SPDIF. 
 
@@ -70,16 +171,43 @@ Leave it blank to disable SPDIF usage, you can also define them at compile time
 ```
 bck=<gpio>,ws=<gpio>,do=<gpio>
 ```
+NB: For well-known configuration, this is ignored
+
+To optimize speed, a bit-manipulation trick is used and as a result, the bit depth is limited to 20 bits, even in 32 bits mode. As said before, this is more than enough for any human ear. In theory, it could be extended up to 23 bits but I don't see the need. Now, you can also get SPDIF using a specialized chip that offers a I2S interface like a DAC but spits out SPDIF (optical and coax). Refers to DAC chapter then.
+
+If you want coax, you can also use a poor-man's trick to generate signal from a 3.3V GPIO. All that does is dividing the 3.3V to generate a 0.6V peak-to-peak and then remove DC
+```
+                          100nF
+GPIO  ----210ohm-----------||---- coax S/PDIF signal out
+                    |
+                  110ohm
+                    |
+Ground -------------------------- coax signal ground
+```
 ### Display
 The NVS parameter "display_config" sets the parameters for an optional display. Syntax is
 ```
-I2C,width=<pixels>,height=<pixels>[address=<i2c_address>][,HFlip][,VFlip][driver=SSD1306|SSD1326|SH1106]
-SPI,width=<pixels>,height=<pixels>,cs=<gpio>[,speed=<speed>][,HFlip][,VFlip][driver=SSD1306|SSD1326|SH1106]
+I2C,width=<pixels>,height=<pixels>[address=<i2c_address>][,reset=<gpio>][,HFlip][,VFlip][driver=SSD1306|SSD1326[:1|4]|SSD1327|SH1106]
+SPI,width=<pixels>,height=<pixels>,cs=<gpio>[,back=<gpio>][,reset=<gpio>][,speed=<speed>][,HFlip][,VFlip][driver=SSD1306|SSD1322|SSD1326[:1|4]|SSD1327|SH1106|SSD1675|ST7735|ST7789|ILI9341[:16|18][,rotate]][,mode=<mode>]
 ```
+- back: a LED backlight used by some older devices (ST7735). It is PWM controlled for brightness
+- reset: some display have a reset pin that is should normally be pulled up if unused
 - VFlip and HFlip are optional can be used to change display orientation
+- rotate: for non-square *drivers*, move to portrait mode. Note that *width* and *height* must be inverted then
 - Default speed is 8000000 (8MHz) but SPI can work up to 26MHz or even 40MHz
+- mode:  Default mode = 0.  Some display modules use different transaction line timings.  Check the module documentation if a non-standard mode is required.
+- SH1106 is 128x64 monochrome I2C/SPI [here]((https://www.waveshare.com/wiki/1.3inch_OLED_HAT))
+- SSD1306 is 128x32 monochrome I2C/SPI [here](https://www.buydisplay.com/i2c-blue-0-91-inch-oled-display-module-128x32-arduino-raspberry-pi)
+- SSD1322 is 256x64 grayscale 16-levels SPI in multiple sizes [here](https://www.buydisplay.com/oled-display/oled-display-module?resolution=159) - it is very nice
+- SSD1326 is 256x32 monochrome or grayscale 16-levels SPI [here](https://www.aliexpress.com/item/32833603664.html?spm=a2g0o.productlist.0.0.2d19776cyQvsBi&algo_pvid=c7a3db92-e019-4095-8a28-dfdf0a087f98&algo_expid=c7a3db92-e019-4095-8a28-dfdf0a087f98-1&btsid=0ab6f81e15955375483301352e4208&ws_ab_test=searchweb0_0,searchweb201602_,searchweb201603_)
+- SSD1327 is 128x128 16-level grayscale SPI [here](https://www.amazon.com/gp/product/B079N1LLG8/ref=ox_sc_act_title_1?smid=A1N6DLY3NQK2VM&psc=1) - artwork can be up to 96x96 with vertical vu-meter/spectrum
+- SSD1351 is 128x128 65k/262k color SPI [here](https://www.waveshare.com/product/displays/lcd-oled/lcd-oled-3/1.5inch-rgb-oled-module.htm)
+- SSD1675 is an e-ink paper and is experimental as e-ink is really not suitable for LMS du to its very low refresh rate
+- ST7735 is a 128x160 65k color SPI [here](https://www.waveshare.com/product/displays/lcd-oled/lcd-oled-3/1.8inch-lcd-module.htm). This needs a backlight control
+- ST7789 is a 240x320 65k (262k not enabled) color SPI [here](https://www.waveshare.com/product/displays/lcd-oled/lcd-oled-3/2inch-lcd-module.htm). It also exist with 240x240 displays. See **rotate** for use in portrait mode
+- ILI9341 is another 240x320 65k (262k capable) color SPI. I've not used it much, the driver it has been provided by one external contributor to the project
 
-Currently 128x32/64 I2C and SPI display like [this](https://www.buydisplay.com/i2c-blue-0-91-inch-oled-display-module-128x32-arduino-raspberry-pi) and [this](https://www.waveshare.com/wiki/1.3inch_OLED_HAT) are supported
+You can tweak how the vu-meter and spectrum analyzer are displayed, as well as size of artwork through a dedicated menu in player's settings (don't forget to add the plugin).
 
 The NVS parameter "metadata_config" sets how metadata is displayed for AirPlay and Bluetooth. Syntax is
 ```
@@ -91,8 +219,6 @@ The NVS parameter "metadata_config" sets how metadata is displayed for AirPlay a
 
 - 'format' can contain free text and any of the 3 keywords %artist%, %album%, %title%. Using that format string, the keywords are replaced by their value to build the string to be displayed. Note that the plain text following a keyword that happens to be empty during playback of a track will be removed. For example, if you have set format=%artist% - %title% and there is no artist in the metadata then only <title> will be displayed not " - <title>".
 
-You can install the excellent plugin "Music Information Screen" which is super useful to tweak the layout for these small displays.
-
 ### Infrared
 You can use any IR receiver compatible with NEC protocol (38KHz). Vcc, GND and output are the only pins that need to be connected, no pullup, no filtering capacitor, it's a straight connection.
 
@@ -103,11 +229,11 @@ In AirPlay and Bluetooth mode, only these native remotes are supported, I've not
 See "set GPIO" below to set the GPIO associated to infrared receiver (option "ir"). 
 
 ### Set GPIO
-The parameter "set_GPIO" is use to assign GPIO to various functions.
+The parameter "set_GPIO" is used to assign GPIO to various functions.
 
 GPIO can be set to GND provide or Vcc at boot. This is convenient to power devices that consume less than 40mA from the side connector. Be careful because there is no conflict checks being made wrt which GPIO you're changing, so you might damage your board or create a conflict here. 
 
-The \<amp\> parameter can use used to assign a GPIO that will be set to 1 when playback starts. It will be reset to 0 when squeezelite becomes idle. The idle timeout is set on the squeezelite command line through -C \<timeout\>
+The \<amp\> parameter can use used to assign a GPIO that will be set to active level (default 1) when playback starts. It will be reset when squeezelite becomes idle. The idle timeout is set on the squeezelite command line through -C \<timeout\>
 
 If you have an audio jack that supports insertion (use :0 or :1 to set the level when inserted), you can specify which GPIO it's connected to. Using the parameter jack_mutes_amp allows to mute the amp when headset (e.g.) is inserted.
 
@@ -118,24 +244,41 @@ The \<ir\> parameter set the GPIO associated to an IR receiver. No need to add p
 Syntax is:
 
 ```
-<gpio>=Vcc|GND|amp|ir|jack[:0|1]|green[:0|1]|red[:0|1]|spkfault[:0|1][,<repeated sequence for next GPIO>]
+<gpio>=Vcc|GND|amp[:1|0]|ir|jack[:0|1]|green[:0|1]|red[:0|1]|spkfault[:0|1][,<repeated sequence for next GPIO>]
 ```
-You can define the defaults for jack, spkfault leds at compile time but nvs parameter takes precedence except for SqueezeAMP where these are forced at runtime.
+You can define the defaults for jack, spkfault leds at compile time but nvs parameter takes precedence except for well-known configurations where these are forced at runtime.
+**Note that gpio 36 and 39 are input only and cannot use interrupt. When set to jack or speaker fault, a 100ms polling checks their value but that's expensive**
+### LED 
+See §**set_GPIO** for how to set the green and red LEDs. In addition, their brightness can be controlled using the "led_brigthness" parameter. The syntax is
+```
+[green=0..100][,red=0..100]
+```
+NB: For well-known configuration, this is ignored
 ### Rotary Encoder
 One rotary encoder is supported, quadrature shift with press. Such encoders usually have 2 pins for encoders (A and B), and common C that must be set to ground and an optional SW pin for press. A, B and SW must be pulled up, so automatic pull-up is provided by ESP32, but you can add your own resistors. A bit of filtering on A and B (~470nF) helps for debouncing which is not made by software. 
 
-Encoder is normally hard-coded to respectively knob left, right and push on LMS and to volume down/up/play toggle on BT and AirPlay. Using the option 'volume' makes it hard-coded to volume down/up/play toggle all the time (even in LMS). The option 'longpress' allows an alternate mode when SW is long-pressed. In that mode, left is previous, right is next and press is toggle. Every long press on SW alternates between modes.
+Encoder is normally hard-coded to respectively knob left, right and push on LMS and to volume down/up/play toggle on BT and AirPlay. Using the option 'volume' makes it hard-coded to volume down/up/play toggle all the time (even in LMS). The option 'longpress' allows an alternate mode when SW is long-pressed. In that mode, left is previous, right is next and press is toggle. Every long press on SW alternates between modes (the main mode actual behavior depends on 'volume').
+
+There is also the possibility to use 'knobonly' option (exclusive with 'volume' and 'longpress'). This mode attempts to offer a single knob full navigation which is a bit contorded due to LMS UI's principles. Left, Right and Press obey to LMS's navigation rules and especially Press always goes to lower submenu item, even when navigating in the Music Library. That causes a challenge as there is no 'Play', 'Back' or 'Pause' button. Workaround are as of below:
+- longpress is 'Play'
+- double press is 'Back' (Left in LMS's terminology). 
+- a quick left-right movement on the encoder is 'Pause' 
+
+The speed of double click (or left-right) can be set using the optional parameter of 'knobonly'. This is not a perfect solution, and other ideas are welcome. Be aware that the longer you set double click speed, the less responsive the interface will be. The reason is that I need to wait for that delay before deciding if it's a single or double click. It can also make menu navigation "hesitations" being easily interpreted as 'Pause'
 
 Use parameter rotary_config with the following syntax:
 
 ```
-A=<gpio>,B=<gpio>[,SW=gpio>[,volume][,longpress]]
+A=<gpio>,B=<gpio>[,SW=gpio>[[,knobonly[=<ms>]]|[[,volume][,longpress]]]]
 ```
 
 HW note: all gpio used for rotary have internal pull-up so normally there is no need to provide Vcc to the encoder. Nevertheless if the encoder board you're using also has its own pull-up that are stronger than ESP32's ones (which is likely the case), then there will be crosstalk between gpio, so you must bring Vcc. Look at your board schematic and you'll understand that these board pull-up create a "winning" pull-down when any other pin is grounded. 
 
-See also the "IMPORTANT NOTE" on the "Buttons" section
+The SW gpio is optional, you can re-affect it to a pure button if you prefer but the volume, longpress and knobonly options make little sense as the missing switch plays an important role in these modes. You could still have the "volume" mode, but you won't be able to use it for *anything* expect volume up and down. So be aware that the use of syntax [] is a bit misleading hereabove.
 
+See also the "IMPORTANT NOTE" on the "Buttons" section and remember that when 'lms_ctrls_raw' (see below) is activated, none of these knobonly,volume,longpress options apply, raw button codes (not actions) are simply sent to LMS
+
+**Note that gpio 36 and 39 are input only and cannot use interrupt, so they cannot be set to A or B. When using them for SW, a 100ms polling is used which is expensive**
 ### Buttons
 Buttons are described using a JSON string with the following syntax
 ```
@@ -170,10 +313,11 @@ Where (all parameters are optionals except gpio)
 Where \<action\> is either the name of another configuration to load (remap) or one amongst
 
 ```
-ACTRLS_NONE, ACTRLS_VOLUP, ACTRLS_VOLDOWN, ACTRLS_TOGGLE, ACTRLS_PLAY, 
+ACTRLS_NONE, ACTRLS_POWER, ACTRLS_VOLUP, ACTRLS_VOLDOWN, ACTRLS_TOGGLE, ACTRLS_PLAY, 
 ACTRLS_PAUSE, ACTRLS_STOP, ACTRLS_REW, ACTRLS_FWD, ACTRLS_PREV, ACTRLS_NEXT, 
-BCTRLS_UP, BCTRLS_DOWN, BCTRLS_LEFT, BCTRLS_RIGHT,
-KNOB_LEFT, KNOB_RIGHT, KNOB_PUSH
+BCTRLS_UP, BCTRLS_DOWN, BCTRLS_LEFT, BCTRLS_RIGHT, 
+BCTRLS_PS1, BCTRLS_PS2, BCTRLS_PS3, BCTRLS_PS4, BCTRLS_PS5, BCTRLS_PS6,
+KNOB_LEFT, KNOB_RIGHT, KNOB_PUSH,	
 ```
 				
 One you've created such a string, use it to fill a new NVS parameter with any name below 16(?) characters. You can have as many of these configs as you can. Then set the config parameter "actrls_config" with the name of your default config
@@ -225,86 +369,94 @@ Below is a difficult but functional 2-buttons interface for your decoding pleasu
 
 The benefit of the "raw" mode is that you can build a player which is as close as possible to a Boom (e.g.) but you can't use the remapping function nor longress or shift logics to do your own mapping when you have a limited set of buttons. In 'raw' mode, all you really need to define is the mapping between the gpio and the button. As far as LMS is concerned, any other option in these JSON payloads does not matter. Now, when you use BT or AirPlay, the full JSON construct described above fully applies, so the shift, longpress, remapping options still work. 
 
+**Be aware that when using non "raw" mode, the CLI (Command Line Interface) of LMS is used and *must* be available without password**
+
 There is no good or bad option, it's your choice. Use the NVS parameter "lms_ctrls_raw" to change that option
+	
+**Note that gpio 36 and 39 are input only and cannot use interrupt. When using them for a button, a 100ms polling is started which is expensive. Long press is also likely to not work very well**
+### Ethernet (coming soon)
+Wired ethernet is supported by esp32 with various options but squeezelite is only supporting a Microchip LAN8720 with a RMII interface like [this](https://www.aliexpress.com/item/32858432526.html) or Davicom DM9051 over SPI like [that](https://www.amazon.com/dp/B08JLFWX9Z).
+	
+#### RMII (LAN8720)	
+- RMII PHY wiring is fixed and can not be changed
 
+| GPIO   | RMII Signal | Notes        |
+| ------ | ----------- | ------------ |
+| GPIO21 | TX_EN       | EMAC_TX_EN   |
+| GPIO19 | TX0         | EMAC_TXD0    |
+| GPIO22 | TX1         | EMAC_TXD1    |
+| GPIO25 | RX0         | EMAC_RXD0    |
+| GPIO26 | RX1         | EMAC_RXD1    |
+| GPIO27 | CRS_DV      | EMAC_RX_DRV  |
+
+- SMI (Serial Management Interface) wiring is not fixed and you can change it either in the configuration or using "eth_config" parameter with the following syntax:
+```
+model=lan8720,mdc=<gpio>,mdio=<gpio>[,rst=<gpio>]
+```
+Connecting a reset pin for the LAN8720 is optional but recommended to avoid that GPIO0 (50MHz input clock) locks the esp32 in download mode at boot time.
+- Clock
+	
+The APLL of the esp32 is required for the audio codec, so we **need** a LAN8720 that provides a 50MHz clock. That clock **must** be connected to GPIO0, there is no alternative. This means that if your DAC requires an MCLK, then you are out of luck. It is not possible to have both to work together. There might be some workaround using CLK_OUT2 and GPIO3, but I don't have time for this.
+#### SPI (DM9051)
+Ethernet over SPI is supported as well and requires less GPIOs but is obvsiously slower. Another benefit is that the SPI bus can be shared with the display, but it's also possible to have a dedicated SPI interface. The esp32 has 4 SPI sub-systems, one is unaccessible so numbering is 0..2 and SPI0 is reserved for Flash/PSRAM. The "eth_config" parameter syntax becomes:
+```
+model=dm9051,cs=<gpio>,speed=<clk_in_Hz>,intr=<gpio>[,host=<-1|1|2>][,rst=<gpio>][,mosi=<gpio>,miso=<gpio>,clk=<gpio>]
+```
+- To use the system SPI, shared with display (see spi_config) "host" must be set to -1. Any other value will reserve the SPI interface (careful of conflict with spi_config). The default "host" is 2 to avoid conflicting wiht default "spi_config" settings.
+- When not using system SPI, "mosi" for data out, "miso" for data in and "clk" **must** be set
+- The esp32 has a special I/O multiplexer for faster speed (up to 80 MHz) but that requires using specific GPIOs, which depends on SPI bus (See [here](https://docs.espressif.com/projects/esp-idf/en/latest/esp32/api-reference/peripherals/spi_master.html) for more details)
+
+| Pin Name | SPI2 | SPI3 |
+| -------- | ---- | ---- |		
+| CS0*     |  15  |  5   |
+| SCLK	   |  14  |  18  |
+| MISO	   |  12  |  19  |
+| MOSI	   |  13  |  23  |
+	
+** THIS IS NOT AVAILABLE YET, SO MORE TO COME ON HOW TO USE WIRED ETHERNET***
 ### Battery / ADC
-The NVS parameter "bat_config" sets the ADC1 channel used to measure battery/DC voltage. Scale is a float ratio applied to every sample of the 12 bits ADC. A measure is taken every 10s and an average is made every 5 minutes (not a sliding window). Syntax is
+The NVS parameter "bat_config" sets the ADC1 channel used to measure battery/DC voltage. The "atten" value attenuates the input voltage to the ADC input (the read value maintains a 0-1V rage) where: 0=no attenuation(0..800mV), 1=2.5dB attenuation(0..1.1V), 2=6dB attenuation(0..1.35V), 3=11dB attenuation(0..2.6V). Scale is a float ratio applied to every sample of the 12 bits ADC. A measure is taken every 10s and an average is made every 5 minutes (not a sliding window). Syntax is
 ```
-channel=0..7,scale=<scale>
+channel=0..7,scale=<scale>,cells=<2|3>[,atten=<0|1|2|3>]
 ```
-NB: Set parameter to empty to disable battery reading
-## Setting up ESP-IDF
-### Docker
-#### **************** todo: Docker scripts needs some rework.
-
-You can use docker to build squeezelite-esp32  
-First you need to build the Docker container:
-```
-docker build -t esp-idf .
-```
-Then you need to run the container:
-```
-docker run -i -t -v `pwd`:/workspace/squeezelite-esp32 esp-idf
-```
-The above command will mount this repo into the docker container and start a bash terminal
-for you to then follow the below build steps
-
-### Manual Install of ESP-IDF
-<<<<<<< HEAD
-Follow the instructions from https://docs.espressif.com/projects/esp-idf/en/v4.0/get-started/index.html to install the esp-idf v4.0. This is the currently supported release of the espressif software development system. 
-=======
-<strong>Currently the master branch of this project requires this [IDF](https://github.com/espressif/esp-idf/tree/28f1cdf5ed7149d146ad5019c265c8bc3bfa2ac9) with gcc 5.2 (toolschain dated 20181001)
-If you want to use a more recent version of gcc and IDF (4.0 stable), move to cmake-master branch</strong>
-
-You can install IDF manually on Linux or Windows (using the Subsystem for Linux) following the instructions at: https://www.instructables.com/id/ESP32-Development-on-Windows-Subsystem-for-Linux/
-You also need to use esp-dsp recent version or at least make sure you have this patch https://github.com/espressif/esp-dsp/pull/12/commits/8b082c1071497d49346ee6ed55351470c1cb4264
->>>>>>> refs/remotes/origin/master
-
-## Building Squeezelite-esp32
-MOST IMPORTANT: create the right default config file
-- make defconfig
-(Note: You can also copy over config files from the build-scripts folder to ./sdkconfig)
-Then adapt the config file to your wifi/BT/I2C device (can also be done on the command line)
-- make menuconfig
-Then
-
-```
-idf.py -p PORT [-b BAUD] flash
-idf.py -p PORT [-b BAUD] monitor
-
-```
-
+NB: Set parameter to empty to disable battery reading. For well-known configuration, this is ignored (except for SqueezeAMP where number of cells is required)
 # Configuration
-1/ setup WiFi
+## Setup WiFi
 - Boot the esp, look for a new wifi access point showing up and connect to it. Default build ssid and passwords are "squeezelite"/"squeezelite". 
 - Once connected, navigate to 192.168.4.1 
 - Wait for the list of access points visible from the device to populate in the web page.
 - Choose an access point and enter any credential as needed
 - Once connection is established, note down the address the device received; this is the address you will use to configure it going forward 
 
-2/ setup squeezelite command line (optional)
+## Setup squeezelite command line (optional)
 
 At this point, the device should have disabled its built-in access point and should be connected to a known WiFi network.
 - navigate to the address that was noted in step #1
-- Using the list of predefined options, hoose the mode in which you want squeezelite to start
+- Using the list of predefined options, choose the mode in which you want squeezelite to start
 - Generate the command
 - Add or change any additional command line option (for example player name, etc)
 - Activate squeezelite execution: this tells the device to automatiaclly run the command at start
 - Update the configuration
 - click on the "start toggle" button. This will force a reboot. 
-- The toggle switch should be set to 'ON' to ensure that squeezelite is active after booting
+- The toggle switch should be set to 'ON' to ensure that squeezelite is active after booting (you might have to fiddle with it a few times)
+- You can enable accessto  NVS parameters under 'credits'
 
-3/ Updating Squeezelite
+## Monitor
+
+In addition of the esp-idf serial link monitor option, you can also enable a telnet server (see NVS parameters) where you'll have access to a ton of logs of what's happening inside the WROVER.
+
+## Update Squeezelite
 - From the firmware tab, click on "Check for Updates"
 - Look for updated binaries
 - Select a line
 - Click on "Flash!"
 - The system will reboot into recovery mode (if not already in that mode), wipe the squeezelite partition and download/flash the selected version 
+- You can choose a local file or have a local webserver
 
-3/ Recovery
+## Recovery
 - From the firmware tab, click on the "Recovery" button. This will reboot the ESP32 into recovery, where additional configuration options are available from the NVS editor
 
-# Additional command line notes, configured from the http configuration
+## Additional configuration notes (from the Web UI)
 The squeezelite options are very similar to the regular Linux ones. Differences are :
 
 	- the output is -o ["BT -n '<sinkname>' "] | [I2S]
@@ -320,12 +472,48 @@ See squeezlite command line, but keys options are
 	- Z <rate> : tell LMS what is the max sample rate supported before LMS resamples
 	- R (see above)
 	- r "<minrate>-<maxrate>"
+	- C <sec> : set timeout to switch off amp gpio
+	- W : activate WAV and AIFF header parsing
+# Building everything yourself
+## Setting up ESP-IDF
+### Docker
+A simple alternative to building the project's binaries is to leverage the same docker image that is being used on the GitHub Actions to build our releases. The instructions below assume that you have cloned  the squeezelite-esp32 code that you want to build locally and that you have opened a command line/bash session in the folder that contains the code. 
+Pull the most recent docker image for the environment: 
+```
+docker pull sle118/squeezelite-esp32-idfv4-master
+```
+Then run the container interactively :
+```
+for windows:
+docker run -v %cd%:/project -w /project -it sle118/squeezelite-esp32-idfv4-master
+for linux:
+docker run -it -v `pwd`:/workspace/squeezelite-esp32 sle118/squeezelite-esp32-idfv4-master
+```
+The above command will mount this repo into the docker container and start a bash terminal. From there, simply run idf.py build to build, etc. Note that at the time of writing these lines, flashing is not possible for docker running under windows https://github.com/docker/for-win/issues/1018.
 
-## Additional misc notes to do you build
-- as of this writing, ESP-IDF has a bug int he way the PLL values are calculated for i2s, so you *must* use the i2s.c file in the patch directory
+### Manual Install of ESP-IDF
+You can install IDF manually on Linux or Windows (using the Subsystem for Linux) following the instructions at: https://www.instructables.com/id/ESP32-Development-on-Windows-Subsystem-for-Linux/ or see here https://docs.espressif.com/projects/esp-idf/en/latest/esp32/get-started/windows-setup.html for a direct install. 
+
+**Use the esp-idf 4.0 https://github.com/espressif/esp-idf/tree/release/v4.0 and a recent add esp-dsp (after 08/2020)**
+
+## Building Squeezelite-esp32
+When initially cloning the repo, make sure you do it recursively. For example: `git clone --recursive https://github.com/sle118/squeezelite-esp32.git`
+	
+Don't forget to choose one of the config files in build_scripts/ and rename it sdkconfig.defaults or sdkconfig as many important WiFi/BT options are set there. **The codecs libraries will not be rebuilt by these scripts (it's a tedious process - see below)**
+
+Create and tweak your config using `idf.py menuconfig` then build binaries using `idf.py all`. It will build the recovery and the application (squeezelite). then use `idf.py flash` to write everything. Otherwise, if you just want to download squeezelite, do (assuming you have set ESPPORT (e.g. COM10) and ESPBAUD (e.g. 921600)
+```
+<path_to_your_python>/python.exe <path_to_your_esptool>/esptool.py -p %ESPPORT% -b %ESPBAUD% --before default_reset --after hard_reset write_flash --flash_mode dio --flash_size detect --flash_freq 80m 0x150000 build/squeezelite.bin
+```
+Use `idf.py monitor` to monitor the application (see esp-idf documentation)
+
+Note: You can use `idf.py build -DDEPTH=32` to build the 32 bits version and add the `-DVERSION=<your_version>` to add a custom version name (it will be 0.0-<your_version>). If you want to change the whole version string, see squeezelite.h. You can also disable the SBR extension of AAC codecs as it consumes a lot of CPU and might overload the esp32. Use `-DAAC_DISABLE_SBR=1` for that
+
+If you have already cloned the repository and you are getting compile errors on one of the submodules (e.g. telnet), run the following git command in the root of the repository location: `git submodule update --init --recursive`
+### Rebuild codecs (highly recommended to NOT try that)
 - for codecs libraries, add -mlongcalls if you want to rebuild them, but you should not (use the provided ones in codecs/lib). if you really want to rebuild them, open an issue
 - libmad, libflac (no esp's version), libvorbis (tremor - not esp's version), alac work
-- libfaad does not really support real time, but if you want to try
+- libfaad does not really support real time, but if you want to try (but using helixaac is a better option)
 	- -O3 -DFIXED_POINT -DSMALL_STACK
 	- change ac_link in configure and case ac_files, remove ''
 	- compiler but in cfft.c and cffti1, must disable optimization using
@@ -337,20 +525,8 @@ See squeezlite command line, but keys options are
 	- per mad & few others, edit configure and change $ac_link to add -c (faking link)
 	- change ac_files to remove ''
 	- add DEPS_CFLAGS and DEPS_LIBS to avoid pkg-config to be required
-	- stack consumption can be very high with some codec variants, so set NONTHREADSAFE_PSEUDOSTACK and GLOBAL_STACK_SIZE=32000 and unset VAR_ARRAYS in config.h
-- better use helixaac			
-- set IDF_PATH=/home/esp-idf
-- set ESPPORT=COM9
-- update flash partition size
-- other compiler #define
-	- use no resampling or set RESAMPLE (soxr) or set RESAMPLE16 for fast fixed 16 bits resampling
-	- use LOOPBACK (mandatory)
-	- use BYTES_PER_FRAME=4 (8 is not fully functionnal)
-	- LINKALL (mandatory)
-	- NO_FAAD unless you want to us faad, which currently overloads the CPU
-	- TREMOR_ONLY (mandatory)
-- When initially cloning the repo, make sure you do it recursively. For example: 
-	- git clone --recursive https://github.com/sle118/squeezelite-esp32.git
-- If you have already cloned the repository and you are getting compile errors on one of the submodules (e.g. telnet), run the following git command in the root of the repository location
-	-  git submodule update --init --recursive
+	- stack consumption can be very high with some codec variants, so set NONTHREADSAFE_PSEUDOSTACK and GLOBAL_STACK_SIZE=48000 and unset VAR_ARRAYS in config.h
+- libmad has been patched to avoid using a lot of stack and is not provided here. There is an issue with sync detection in 1.15.1b from where the original stack patch was done but since a few fixes have been made wrt sync detection. This 1.15.1b-10 found on debian fixes the issue where mad thinks it has reached sync but has not and so returns a wrong sample rate. It comes at the expense of 8KB (!) of code where a simple check in squeezelite/mad.c that next_frame[0] is 0xff and next_frame[1] & 0xf0 is 0xf0 does the trick ...
 
+# Footnotes
+(1) SPDIF is made by tricking the I2S bus but this consumes a fair bit of CPU as it multiplies by four the throughput on the i2s bus. To optimize some computation, the parity of the spdif frames must always be 0, so at least one bit has to be available to force it. As SPDIF samples are 20+4 bits length maximum, the LSB is used for that purpose, so the bit 24 is randomly toggling. It does not matter for 16 bits samples but it has been chosen to truncate the last 4 bits for 24 bits samples. I'm sure that some smart dude can further optimize spdif_convert() and use the user bit instead. You're welcome to do a PR but, as said above, I (philippe44) am not interested by 24 bits mental illness :-) and I've already made an effort to provide 20 bits which already way more what's needed :-)

buildFirmware.sh

@@ -0,0 +1,47 @@
+#!/bin/bash
+
+
+
+echo "Build process started"
+echo "Setting up build name and build number"
+if [ -z "${TARGET_BUILD_NAME}" ]
+then
+    export TARGET_BUILD_NAME="I2S-4MFlash"
+    echo "TARGET_BUILD_NAME is not set. Defaulting to ${TARGET_BUILD_NAME}"
+fi
+if [ -z "${BUILD_NUMBER}" ]
+then
+    export BUILD_NUMBER="500"
+    echo "BUILD_NUMBER is not set. Defaulting to ${BUILD_NUMBER}"
+fi
+if [ -z "$DEPTH" ]
+then
+    export DEPTH="16"
+    echo "DEPTH is not set. Defaulting to ${DEPTH}"
+fi
+if [ -z "$tag" ]
+then
+    branch_name="$(git rev-parse --abbrev-ref HEAD)"
+    branch_name="${branch_name//[^a-zA-Z0-9\-~!@_\.]/}"
+    app_name="${TARGET_BUILD_NAME}.${DEPTH}.dev-$(git log --pretty=format:'%h' --max-count=1).${branch_name}" 
+    echo "${app_name}">version.txt
+    echo "app_name is not set. Defaulting to ${app_name}"
+else
+    echo "${tag}" >version.txt
+fi
+
+echo "Copying target sdkconfig"
+cp build-scripts/${TARGET_BUILD_NAME}-sdkconfig.defaults sdkconfig
+echo "Building project"
+idf.py build -DDEPTH=${DEPTH} -DBUILD_NUMBER=${BUILD_NUMBER}-${DEPTH} 
+echo "Generating size report"
+idf.py size-components >build/size_components.txt
+idf.py size-components-squeezelite build/size_components_squeezelite.txt
+if [ -z "${artifact_file_name}" ]
+then
+    echo "No artifact file name set.  Will not generate zip file."
+else
+    echo "Generating build artifact zip file"
+    zip -r build_output.zip build
+    zip build/${artifact_file_name} partitions*.csv components/ build/*.bin build/bootloader/bootloader.bin build/partition_table/partition-table.bin build/flash_project_args build/size_*.txt
+fi

components/_override/CMakeLists.txt

@@ -0,0 +1,21 @@
+set(lib_dir ${build_dir}/esp-idf)
+
+set(driver i2s.c i2s_hal.c spi_bus_lock.c)
+string(REPLACE ".c" ".c.obj" driver_obj "${driver}")
+
+idf_component_register( SRCS ${driver}
+					REQUIRES driver           
+					INCLUDE_DIRS ${IDF_PATH}/components/driver
+					PRIV_INCLUDE_DIRS ${IDF_PATH}/components/driver/include/driver
+)
+ 
+# CMake is just a pile of crap
+message("!! overriding ${driver} !!")
+message("CAREFUL, LIBRARIES STRIPPING FROM DUPLICATED COMPONENTS DEPENDS ON THIS BEING REBUILD")
+
+add_custom_command(
+			TARGET ${COMPONENT_LIB}
+			PRE_LINK
+			COMMAND xtensa-esp32-elf-ar -d ${lib_dir}/driver/libdriver.a ${driver_obj}
+			VERBATIM
+)

components/_override/i2s_hal.c

@@ -0,0 +1,275 @@
+// Copyright 2015-2019 Espressif Systems (Shanghai) PTE LTD
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+// The HAL layer for I2S (common part)
+
+#include "soc/soc.h"
+#include "soc/soc_caps.h"
+#include "hal/i2s_hal.h"
+
+#define I2S_TX_PDM_FP_DEF  960   // Set to the recommended value(960) in TRM
+#define I2S_RX_PDM_DSR_DEF 0
+
+void i2s_hal_set_tx_mode(i2s_hal_context_t *hal, i2s_channel_t ch, i2s_bits_per_sample_t bits)
+{
+    if (bits <= I2S_BITS_PER_SAMPLE_16BIT) {
+        i2s_ll_set_tx_fifo_mod(hal->dev, (ch == I2S_CHANNEL_STEREO) ? 0 : 1);
+    } else {
+        i2s_ll_set_tx_fifo_mod(hal->dev, (ch == I2S_CHANNEL_STEREO) ? 2 : 3);
+    }
+    i2s_ll_set_tx_chan_mod(hal->dev, (ch == I2S_CHANNEL_STEREO) ? 0 : 1);
+#if SOC_I2S_SUPPORTS_DMA_EQUAL
+    i2s_ll_set_tx_dma_equal(hal->dev, (ch == I2S_CHANNEL_STEREO) ? 0 : 1);
+#endif
+}
+
+void i2s_hal_set_rx_mode(i2s_hal_context_t *hal, i2s_channel_t ch, i2s_bits_per_sample_t bits)
+{
+    if (bits <= I2S_BITS_PER_SAMPLE_16BIT) {
+        i2s_ll_set_rx_fifo_mod(hal->dev, (ch == I2S_CHANNEL_STEREO) ? 0 : 1);
+    } else {
+        i2s_ll_set_rx_fifo_mod(hal->dev, (ch == I2S_CHANNEL_STEREO) ? 2 : 3);
+    }
+    i2s_ll_set_rx_chan_mod(hal->dev, (ch == I2S_CHANNEL_STEREO) ? 0 : 1);
+#if SOC_I2S_SUPPORTS_DMA_EQUAL
+    i2s_ll_set_rx_dma_equal(hal->dev, (ch == I2S_CHANNEL_STEREO) ? 0 : 1);
+#endif
+}
+
+void i2s_hal_set_in_link(i2s_hal_context_t *hal, uint32_t bytes_num, uint32_t addr)
+{
+    i2s_ll_set_in_link_addr(hal->dev, addr);
+    i2s_ll_set_rx_eof_num(hal->dev, bytes_num);
+}
+
+#if SOC_I2S_SUPPORTS_PDM
+void i2s_hal_tx_pdm_cfg(i2s_hal_context_t *hal, uint32_t fp, uint32_t fs)
+{
+    i2s_ll_tx_pdm_cfg(hal->dev, fp, fs);
+}
+
+void i2s_hal_get_tx_pdm(i2s_hal_context_t *hal, uint32_t *fp, uint32_t *fs)
+{
+    i2s_ll_get_tx_pdm(hal->dev, fp, fs);
+}
+
+void i2s_hal_rx_pdm_cfg(i2s_hal_context_t *hal, uint32_t dsr)
+{
+    i2s_ll_rx_pdm_cfg(hal->dev, dsr);
+}
+
+void i2s_hal_get_rx_pdm(i2s_hal_context_t *hal, uint32_t *dsr)
+{
+    i2s_ll_get_rx_pdm(hal->dev, dsr);
+}
+#endif
+
+void i2s_hal_set_clk_div(i2s_hal_context_t *hal, int div_num, int div_a, int div_b, int tx_bck_div, int rx_bck_div)
+{
+    i2s_ll_set_clkm_div_num(hal->dev, div_num);
+    i2s_ll_set_clkm_div_a(hal->dev, div_a);
+    i2s_ll_set_clkm_div_b(hal->dev, div_b);
+    i2s_ll_set_tx_bck_div_num(hal->dev, tx_bck_div);
+    i2s_ll_set_rx_bck_div_num(hal->dev, rx_bck_div);
+}
+
+void i2s_hal_set_tx_bits_mod(i2s_hal_context_t *hal, i2s_bits_per_sample_t bits)
+{
+    i2s_ll_set_tx_bits_mod(hal->dev, bits);
+}
+
+void i2s_hal_set_rx_bits_mod(i2s_hal_context_t *hal, i2s_bits_per_sample_t bits)
+{
+    i2s_ll_set_rx_bits_mod(hal->dev, bits);
+}
+
+void i2s_hal_reset(i2s_hal_context_t *hal)
+{
+    // Reset I2S TX/RX module first, and then, reset DMA and FIFO.
+    i2s_ll_reset_tx(hal->dev);
+    i2s_ll_reset_rx(hal->dev);
+    i2s_ll_reset_dma_in(hal->dev);
+    i2s_ll_reset_dma_out(hal->dev);
+    i2s_ll_reset_rx_fifo(hal->dev);
+    i2s_ll_reset_tx_fifo(hal->dev);
+}
+
+void i2s_hal_start_tx(i2s_hal_context_t *hal)
+{
+    i2s_ll_start_out_link(hal->dev);
+    i2s_ll_start_tx(hal->dev);
+}
+
+void i2s_hal_start_rx(i2s_hal_context_t *hal)
+{
+    i2s_ll_start_in_link(hal->dev);
+    i2s_ll_start_rx(hal->dev);
+}
+
+void i2s_hal_stop_tx(i2s_hal_context_t *hal)
+{
+    i2s_ll_stop_out_link(hal->dev);
+    i2s_ll_stop_tx(hal->dev);
+}
+
+void i2s_hal_stop_rx(i2s_hal_context_t *hal)
+{
+    i2s_ll_stop_in_link(hal->dev);
+    i2s_ll_stop_rx(hal->dev);
+}
+
+void i2s_hal_format_config(i2s_hal_context_t *hal, const i2s_config_t *i2s_config)
+{
+    switch (i2s_config->communication_format) {
+        case I2S_COMM_FORMAT_STAND_MSB:
+            if (i2s_config->mode & I2S_MODE_TX) {
+                i2s_ll_set_tx_format_msb_align(hal->dev);
+            }
+            if (i2s_config->mode & I2S_MODE_RX) {
+                i2s_ll_set_rx_format_msb_align(hal->dev);
+            }
+            break;
+        case I2S_COMM_FORMAT_STAND_PCM_SHORT:
+            if (i2s_config->mode & I2S_MODE_TX) {
+                i2s_ll_set_tx_pcm_long(hal->dev);
+            }
+            if (i2s_config->mode & I2S_MODE_RX) {
+                i2s_ll_set_rx_pcm_long(hal->dev);
+            }
+            break;
+        case I2S_COMM_FORMAT_STAND_PCM_LONG:
+            if (i2s_config->mode & I2S_MODE_TX) {
+                i2s_ll_set_tx_pcm_short(hal->dev);
+            }
+            if (i2s_config->mode & I2S_MODE_RX) {
+                i2s_ll_set_rx_pcm_short(hal->dev);
+            }
+            break;
+        default: //I2S_COMM_FORMAT_STAND_I2S
+            if (i2s_config->mode & I2S_MODE_TX) {
+                i2s_ll_set_tx_format_philip(hal->dev);
+            }
+            if (i2s_config->mode & I2S_MODE_RX) {
+                i2s_ll_set_rx_format_philip(hal->dev);
+            }
+            break;
+    }
+}
+
+void i2s_hal_config_param(i2s_hal_context_t *hal, const i2s_config_t *i2s_config)
+{
+    //reset i2s
+    i2s_ll_reset_tx(hal->dev);
+    i2s_ll_reset_rx(hal->dev);
+
+    //reset dma
+    i2s_ll_reset_dma_in(hal->dev);
+    i2s_ll_reset_dma_out(hal->dev);
+
+    i2s_ll_enable_dma(hal->dev);
+
+    i2s_ll_set_lcd_en(hal->dev, 0);
+    i2s_ll_set_camera_en(hal->dev, 0);
+
+    i2s_ll_set_dscr_en(hal->dev, 0);
+
+    i2s_ll_set_tx_chan_mod(hal->dev, i2s_config->channel_format < I2S_CHANNEL_FMT_ONLY_RIGHT ? i2s_config->channel_format : (i2s_config->channel_format >> 1)); // 0-two channel;1-right;2-left;3-righ;4-left
+    i2s_ll_set_tx_fifo_mod(hal->dev, i2s_config->channel_format < I2S_CHANNEL_FMT_ONLY_RIGHT ? 0 : 1); // 0-right&left channel;1-one channel
+    i2s_ll_set_tx_mono(hal->dev, 0);
+
+    i2s_ll_set_rx_chan_mod(hal->dev, i2s_config->channel_format < I2S_CHANNEL_FMT_ONLY_RIGHT ? i2s_config->channel_format : (i2s_config->channel_format >> 1)); // 0-two channel;1-right;2-left;3-righ;4-left
+    i2s_ll_set_rx_fifo_mod(hal->dev, i2s_config->channel_format < I2S_CHANNEL_FMT_ONLY_RIGHT ? 0 : 1); // 0-right&left channel;1-one channel
+    i2s_ll_set_rx_mono(hal->dev, 0);
+
+    i2s_ll_set_dscr_en(hal->dev, 1); //connect dma to fifo
+
+    i2s_ll_stop_tx(hal->dev);
+    i2s_ll_stop_rx(hal->dev);
+
+    if (i2s_config->mode & I2S_MODE_TX) {
+		int order = i2s_config->bits_per_sample == 32 ? 0 : 1;
+        i2s_ll_set_tx_msb_right(hal->dev, order);
+        i2s_ll_set_tx_right_first(hal->dev, ~order);
+
+        i2s_ll_set_tx_slave_mod(hal->dev, 0); // Master
+        i2s_ll_set_tx_fifo_mod_force_en(hal->dev, 1);
+
+        if (i2s_config->mode & I2S_MODE_SLAVE) {
+            i2s_ll_set_tx_slave_mod(hal->dev, 1); //TX Slave
+        }
+    }
+
+    if (i2s_config->mode & I2S_MODE_RX) {
+        i2s_ll_set_rx_msb_right(hal->dev, 0);
+        i2s_ll_set_rx_right_first(hal->dev, 0);
+        i2s_ll_set_rx_slave_mod(hal->dev, 0); // Master
+        i2s_ll_set_rx_fifo_mod_force_en(hal->dev, 1);
+
+        if (i2s_config->mode & I2S_MODE_SLAVE) {
+            i2s_ll_set_rx_slave_mod(hal->dev, 1); //RX Slave
+        }
+    }
+
+#if SOC_I2S_SUPPORTS_PDM
+    if (!(i2s_config->mode & I2S_MODE_PDM)) {
+        i2s_ll_set_rx_pdm_en(hal->dev, 0);
+        i2s_ll_set_tx_pdm_en(hal->dev, 0);
+    } else {
+        if (i2s_config->mode & I2S_MODE_TX) {
+            i2s_ll_tx_pdm_cfg(hal->dev, I2S_TX_PDM_FP_DEF, i2s_config->sample_rate/100);
+        }
+        if(i2s_config->mode & I2S_MODE_RX) {
+            i2s_ll_rx_pdm_cfg(hal->dev, I2S_RX_PDM_DSR_DEF);
+        }
+        // PDM mode have nothing to do with communication format configuration.
+        return;
+    }
+#endif
+
+#if SOC_I2S_SUPPORTS_ADC_DAC
+    if (i2s_config->mode & (I2S_MODE_DAC_BUILT_IN | I2S_MODE_ADC_BUILT_IN)) {
+        if (i2s_config->mode & I2S_MODE_DAC_BUILT_IN) {
+            i2s_ll_build_in_dac_ena(hal->dev);
+        }
+        if (i2s_config->mode & I2S_MODE_ADC_BUILT_IN) {
+            i2s_ll_build_in_adc_ena(hal->dev);
+            i2s_ll_set_rx_chan_mod(hal->dev, 1);
+            i2s_ll_set_rx_fifo_mod(hal->dev, 1);
+            i2s_ll_set_rx_mono(hal->dev, 0);
+        }
+        // Buildin ADC and DAC have nothing to do with communication format configuration.
+        return;
+    }
+#endif
+
+    i2s_hal_format_config(hal, i2s_config);
+}
+
+void i2s_hal_enable_master_mode(i2s_hal_context_t *hal)
+{
+    i2s_ll_set_tx_slave_mod(hal->dev, 0); //MASTER Slave
+    i2s_ll_set_rx_slave_mod(hal->dev, 1); //RX Slave
+}
+
+void i2s_hal_enable_slave_mode(i2s_hal_context_t *hal)
+{
+    i2s_ll_set_tx_slave_mod(hal->dev, 1); //TX Slave
+    i2s_ll_set_rx_slave_mod(hal->dev, 1); //RX Slave
+}
+
+void i2s_hal_init(i2s_hal_context_t *hal, int i2s_num)
+{
+    //Get hardware instance.
+    hal->dev = I2S_LL_GET_HW(i2s_num);
+}

components/_override/spi_bus_lock.c

@@ -0,0 +1,849 @@
+// Copyright 2015-2020 Espressif Systems (Shanghai) PTE LTD
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+#include "freertos/FreeRTOS.h"
+#include "freertos/semphr.h"
+#include <stdatomic.h>
+#include "sdkconfig.h"
+#include "spi_common_internal.h"
+#include "esp_intr_alloc.h"
+#include "soc/soc_caps.h"
+#include "stdatomic.h"
+#include "esp_log.h"
+#include <strings.h>
+#include "esp_heap_caps.h"
+
+/*
+ * This lock is designed to solve the conflicts between SPI devices (used in tasks) and
+ * the background operations (ISR or cache access).
+ *
+ * There are N (device/task) + 1 (BG) acquiring processer candidates that may touch the bus.
+ *
+ * The core of the lock is a `status` atomic variable, which is always available. No intermediate
+ * status is allowed. The atomic operations (mainly `atomic_fetch_and`, `atomic_fetch_or`)
+ * atomically read the status, and bitwisely write status value ORed / ANDed with given masks.
+ *
+ * Definitions of the status:
+ * - [30]    WEAK_BG_FLAG, active when the BG is the cache
+ * - [29:20] LOCK bits, active when corresponding device is asking for acquiring
+ * - [19:10] PENDING bits, active when the BG acknowledges the REQ bits, but hasn't fully handled them.
+ * - [ 9: 0] REQ bits, active when corresponding device is requesting for BG operations.
+ *
+ *   The REQ bits together PENDING bits are called BG bits, which represent the actual BG request
+ *   state of devices. Either one of REQ or PENDING being active indicates the device has pending BG
+ *   requests. Reason of having two bits instead of one is in the appendix below.
+ *
+ * Acquiring processer means the current processor (task or ISR) allowed to touch the critical
+ * resources, or the SPI bus.
+ *
+ * States of the lock:
+ * - STATE_IDLE: There's no acquiring processor. No device is acquiring the bus, and no BG
+ *   operation is in progress.
+ *
+ * - STATE_ACQ: The acquiring processor is a device task. This means one of the devices is
+ *   acquiring the bus.
+ *
+ * - STATE_BG: The acquiring processor is the ISR, and there is no acquiring device.
+ *
+ * - STATE_BG_ACQ: The acquiring processor is the ISR, and there is an acquiring device.
+ *
+ *
+ * Whenever a bit is written to the status, it means the a device on a task is trying to acquire
+ * the lock (either for the task, or the ISR). When there is no LOCK bits or BG bits active, the
+ * caller immediately become the acquiring processor. Otherwise, the task has to block, and the ISR
+ * will not be invoked until scheduled by the current acquiring processor.
+ *
+ * The acquiring processor is responsible to assign the next acquiring processor by calling the
+ * scheduler, usually after it finishes some requests, and cleared the corresponding status bit.
+ * But there is one exception, when the last bit is cleared from the status, after which there is
+ * no other LOCK bits or BG bits active, the acquiring processor lost its role immediately, and
+ * don't need to call the scheduler to assign the next acquiring processor.
+ *
+ * The acquiring processor may also choose to assign a new acquiring device when there is no, by
+ * calling `spi_bus_lock_bg_rotate_acq_dev` in the ISR. But the acquiring processor, in this case,
+ * is still the ISR, until it calls the scheduler.
+ *
+ *
+ * Transition of the FSM:
+ *
+ * - STATE_IDLE: no acquiring device, nor acquiring processor, no LOCK or BG bits active
+ *   -> STATE_BG: by `req_core`
+ *   -> STATE_ACQ: by `acquire_core`
+ *
+ * - STATE_BG:
+ *      * No acquiring device, the ISR is the acquiring processor, there is BG bits active, but no LOCK
+ *        bits
+ *      * The BG operation should be enabled while turning into this state.
+ *
+ *   -> STATE_IDLE: by `bg_exit_core` after `clear_pend_core` for all BG bits
+ *   -> STATE_BG_ACQ: by `schedule_core`, when there is new LOCK bit set (by `acquire_core`)
+ *
+ * - STATE_BG_ACQ:
+ *      * There is acquiring device, the ISR is the acquiring processor, there may be BG bits active for
+ *        the acquiring device.
+ *      * The BG operation should be enabled while turning into this state.
+ *
+ *   -> STATE_ACQ: by `bg_exit_core` after `clear_pend_core` for all BG bits for the acquiring
+ *                 device.
+ *
+ *                 Should not go to the STATE_ACQ (unblock the acquiring task) until all requests of the
+ *                 acquiring device are finished. This is to preserve the sequence of foreground (polling) and
+ *                 background operations of the device. The background operations queued before the acquiring
+ *                 should be completed first.
+ *
+ * - STATE_ACQ:
+ *      * There is acquiring device, the task is the acquiring processor, there is no BG bits active for
+ *        the acquiring device.
+ *      * The acquiring task (if blocked at `spi_bus_lock_acquire_start` or `spi_bus_lock_wait_bg_done`)
+ *        should be resumed while turning into this state.
+ *
+ *   -> STATE_BG_ACQ: by `req_core`
+ *   -> STATE_BG_ACQ (other device): by `acquire_end_core`, when there is LOCK bit for another
+ *                    device, and the new acquiring device has active BG bits.
+ *   -> STATE_ACQ (other device): by `acquire_end_core`, when there is LOCK bit for another devices,
+ *                    but the new acquiring device has no active BG bits.
+ *   -> STATE_BG: by `acquire_end_core` when there is no LOCK bit active, but there are active BG
+ *                bits.
+ *   -> STATE_IDLE: by `acquire_end_core` when there is no LOCK bit, nor BG bit active.
+ *
+ * The `req_core` used in the task is a little special. It asks for acquiring processor for the
+ * ISR. When it succeed for the first time, it will invoke the ISR (hence passing the acquiring
+ * role to the BG). Otherwise it will not block, the ISR will be automatically be invoked by other
+ * acquiring processor. The caller of `req_core` will never become acquiring processor by this
+ * function.
+ *
+ *
+ * Appendix: The design, that having both request bit and pending bit, is to solve the
+ * concurrency issue between tasks and the bg, when the task can queue several requests,
+ * however the request bit cannot represent the number of requests queued.
+ *
+ * Here's the workflow of task and ISR work concurrently:
+ * - Task: (a) Write to Queue -> (b) Write request bit
+ *   The Task have to write request bit (b) after the data is prepared in the queue (a),
+ *   otherwise the BG may fail to read from the queue when it sees the request bit set.
+ *
+ * - BG: (c) Read queue -> (d) Clear request bit
+ *   Since the BG cannot know the number of requests queued, it have to repeatedly check the
+ *   queue (c), until it find the data is empty, and then clear the request bit (d).
+ *
+ * The events are possible to happen in the order: (c) -> (a) -> (b) -> (d). This may cause a false
+ * clear of the request bit. And there will be data prepared in the queue, but the request bit is
+ * inactive.
+ *
+ * (e) move REQ bits to PEND bits, happen before (c) is introduced to solve this problem. In this
+ * case (d) is changed to clear the PEND bit. Even if (e) -> (c) -> (a) -> (b) -> (d), only PEND
+ * bit is cleared, while the REQ bit is still active.
+ */
+
+struct spi_bus_lock_dev_t;
+typedef struct spi_bus_lock_dev_t spi_bus_lock_dev_t;
+
+typedef struct spi_bus_lock_t spi_bus_lock_t;
+
+
+#define MAX_DEV_NUM     10
+
+// Bit 29-20: lock bits, Bit 19-10: pending bits
+// Bit 9-0: request bits, Bit 30:
+#define LOCK_SHIFT      20
+#define PENDING_SHIFT   10
+#define REQ_SHIFT       0
+
+#define WEAK_BG_FLAG    BIT(30)    /**< The bus is permanently requested by background operations.
+                                     * This flag is weak, will not prevent acquiring of devices. But will help the BG to be re-enabled again after the bus is release.
+                                     */
+
+// get the bit mask wher bit [high-1, low] are all 1'b1 s.
+#define BIT1_MASK(high, low)   ((UINT32_MAX << (high)) ^ (UINT32_MAX << (low)))
+
+#define LOCK_BIT(mask)      ((mask) << LOCK_SHIFT)
+#define REQUEST_BIT(mask)   ((mask) << REQ_SHIFT)
+#define PENDING_BIT(mask)   ((mask) << PENDING_SHIFT)
+#define DEV_MASK(id)        (LOCK_BIT(1<<id) | PENDING_BIT(1<<id) | REQUEST_BIT(1<<id))
+#define ID_DEV_MASK(mask)   (ffs(mask) - 1)
+
+#define REQ_MASK            BIT1_MASK(REQ_SHIFT+MAX_DEV_NUM, REQ_SHIFT)
+#define PEND_MASK           BIT1_MASK(PENDING_SHIFT+MAX_DEV_NUM, PENDING_SHIFT)
+#define BG_MASK             BIT1_MASK(REQ_SHIFT+MAX_DEV_NUM*2, REQ_SHIFT)
+#define LOCK_MASK           BIT1_MASK(LOCK_SHIFT+MAX_DEV_NUM, LOCK_SHIFT)
+
+#define DEV_REQ_MASK(dev)   ((dev)->mask & REQ_MASK)
+#define DEV_PEND_MASK(dev)  ((dev)->mask & PEND_MASK)
+#define DEV_BG_MASK(dev)    ((dev)->mask & BG_MASK)
+
+struct spi_bus_lock_t {
+    /**
+     * The core of the lock. These bits are status of the lock, which should be always available.
+     * No intermediate status is allowed. This is realized by atomic operations, mainly
+     * `atomic_fetch_and`, `atomic_fetch_or`, which atomically read the status, and bitwise write
+     * status value ORed / ANDed with given masks.
+     *
+     * The request bits together pending bits represent the actual bg request state of one device.
+     * Either one of them being active indicates the device has pending bg requests.
+     *
+     * Whenever a bit is written to the status, it means the a device on a task is trying to
+     * acquire the lock. But this will succeed only when no LOCK or BG bits active.
+     *
+     * The acquiring processor is responsible to call the scheduler to pass its role to other tasks
+     * or the BG, unless it clear the last bit in the status register.
+     */
+    //// Critical resources, they are only writable by acquiring processor, and stable only when read by the acquiring processor.
+    atomic_uint_fast32_t    status;
+    spi_bus_lock_dev_t* volatile acquiring_dev;   ///< The acquiring device
+    bool                volatile acq_dev_bg_active;    ///< BG is the acquiring processor serving the acquiring device, used for the wait_bg to skip waiting quickly.
+    bool                volatile in_isr;         ///< ISR is touching HW
+    //// End of critical resources
+
+    atomic_intptr_t     dev[DEV_NUM_MAX];     ///< Child locks.
+    bg_ctrl_func_t      bg_enable;      ///< Function to enable background operations.
+    bg_ctrl_func_t      bg_disable;     ///< Function to disable background operations
+    void*               bg_arg;            ///< Argument for `bg_enable` and `bg_disable` functions.
+
+    spi_bus_lock_dev_t* last_dev;       ///< Last used device, to decide whether to refresh all registers.
+    int                 periph_cs_num;  ///< Number of the CS pins the HW has.
+
+    //debug information
+    int                 host_id;        ///< Host ID, for debug information printing
+    uint32_t            new_req;        ///< Last int_req when `spi_bus_lock_bg_start` is called. Debug use.
+};
+
+struct spi_bus_lock_dev_t {
+    SemaphoreHandle_t   semphr;     ///< Binray semaphore to notify the device it claimed the bus
+    spi_bus_lock_t*     parent;     ///< Pointer to parent spi_bus_lock_t
+    uint32_t            mask;       ///< Bitwise OR-ed mask of the REQ, PEND, LOCK bits of this device
+};
+
+portMUX_TYPE s_spinlock = portMUX_INITIALIZER_UNLOCKED;
+
+DRAM_ATTR static const char TAG[] = "bus_lock";
+
+#define LOCK_CHECK(a, str, ret_val, ...) \
+    if (!(a)) { \
+        ESP_LOGE(TAG,"%s(%d): "str, __FUNCTION__, __LINE__, ##__VA_ARGS__); \
+        return (ret_val); \
+    }
+
+static inline int mask_get_id(uint32_t mask);
+static inline int dev_lock_get_id(spi_bus_lock_dev_t *dev_lock);
+
+/*******************************************************************************
+ * atomic operations to the status
+ ******************************************************************************/
+SPI_MASTER_ISR_ATTR static inline uint32_t lock_status_fetch_set(spi_bus_lock_t *lock, uint32_t set)
+{
+    return atomic_fetch_or(&lock->status, set);
+}
+
+IRAM_ATTR static inline uint32_t lock_status_fetch_clear(spi_bus_lock_t *lock, uint32_t clear)
+{
+    return atomic_fetch_and(&lock->status, ~clear);
+}
+
+IRAM_ATTR static inline uint32_t lock_status_fetch(spi_bus_lock_t *lock)
+{
+    return atomic_load(&lock->status);
+}
+
+SPI_MASTER_ISR_ATTR static inline void lock_status_init(spi_bus_lock_t *lock)
+{
+    atomic_store(&lock->status, 0);
+}
+
+// return the remaining status bits
+IRAM_ATTR static inline uint32_t lock_status_clear(spi_bus_lock_t* lock, uint32_t clear)
+{
+    //the fetch and clear should be atomic, avoid missing the all '0' status when all bits are clear.
+    uint32_t state = lock_status_fetch_clear(lock, clear);
+    return state & (~clear);
+}
+
+/*******************************************************************************
+ * Schedule service
+ *
+ * The modification to the status bits may cause rotating of the acquiring processor. It also have
+ * effects to `acquired_dev` (the acquiring device), `in_isr` (HW used in BG), and
+ * `acq_dev_bg_active` (wait_bg_end can be skipped) members of the lock structure.
+ *
+ * Most of them should be atomic, and special attention should be paid to the operation
+ * sequence.
+ ******************************************************************************/
+SPI_MASTER_ISR_ATTR static inline void resume_dev_in_isr(spi_bus_lock_dev_t *dev_lock, BaseType_t *do_yield)
+{
+    xSemaphoreGiveFromISR(dev_lock->semphr, do_yield);
+}
+
+IRAM_ATTR static inline void resume_dev(const spi_bus_lock_dev_t *dev_lock)
+{
+    xSemaphoreGive(dev_lock->semphr);
+}
+
+SPI_MASTER_ISR_ATTR static inline void bg_disable(spi_bus_lock_t *lock)
+{
+    BUS_LOCK_DEBUG_EXECUTE_CHECK(lock->bg_disable);
+    lock->bg_disable(lock->bg_arg);
+}
+
+IRAM_ATTR static inline void bg_enable(spi_bus_lock_t* lock)
+{
+    BUS_LOCK_DEBUG_EXECUTE_CHECK(lock->bg_enable);
+    lock->bg_enable(lock->bg_arg);
+}
+
+// Set the REQ bit. If we become the acquiring processor, invoke the ISR and pass that to it.
+// The caller will never become the acquiring processor after this function returns.
+SPI_MASTER_ATTR static inline void req_core(spi_bus_lock_dev_t *dev_handle)
+{
+    spi_bus_lock_t *lock = dev_handle->parent;
+
+    // Though `acquired_dev` is critical resource, `dev_handle == lock->acquired_dev`
+    // is a stable statement unless `acquire_start` or `acquire_end` is called by current
+    // device.
+    if (dev_handle == lock->acquiring_dev){
+        // Set the REQ bit and check BG bits if we are the acquiring processor.
+        // If the BG bits were not active before, invoke the BG again.
+
+        // Avoid competitive risk against the `clear_pend_core`, `acq_dev_bg_active` should be set before
+        // setting REQ bit.
+        lock->acq_dev_bg_active = true;
+        uint32_t status = lock_status_fetch_set(lock, DEV_REQ_MASK(dev_handle));
+        if ((status & DEV_BG_MASK(dev_handle)) == 0) {
+            bg_enable(lock); //acquiring processor passed to BG
+        }
+    } else {
+        uint32_t status = lock_status_fetch_set(lock, DEV_REQ_MASK(dev_handle));
+        if (status == 0) {
+            bg_enable(lock); //acquiring processor passed to BG
+        }
+    }
+}
+
+//Set the LOCK bit. Handle related stuff and return true if we become the acquiring processor.
+SPI_MASTER_ISR_ATTR static inline bool acquire_core(spi_bus_lock_dev_t *dev_handle)
+{
+    spi_bus_lock_t* lock = dev_handle->parent;
+	portENTER_CRITICAL_SAFE(&s_spinlock);
+    uint32_t status = lock_status_fetch_set(lock, dev_handle->mask & LOCK_MASK);
+	portEXIT_CRITICAL_SAFE(&s_spinlock);
+
+    // Check all bits except WEAK_BG
+    if ((status & (BG_MASK | LOCK_MASK)) == 0) {
+        //succeed at once
+        lock->acquiring_dev = dev_handle;
+        BUS_LOCK_DEBUG_EXECUTE_CHECK(!lock->acq_dev_bg_active);
+        if (status & WEAK_BG_FLAG) {
+            //Mainly to disable the cache (Weak_BG), that is not able to disable itself
+            bg_disable(lock);
+        }
+        return true;
+    }
+    return false;
+}
+
+/**
+ * Find the next acquiring processor according to the status. Will directly change
+ * the acquiring device if new one found.
+ *
+ * Cases:
+ * - BG should still be the acquiring processor (Return false):
+ *     1. Acquiring device has active BG bits: out_desired_dev = new acquiring device
+ *     2. No acquiring device, but BG active: out_desired_dev = randomly pick one device with active BG bits
+ * - BG should yield to the task (Return true):
+ *     3. Acquiring device has no active BG bits: out_desired_dev = new acquiring device
+ *     4. No acquiring device while no active BG bits: out_desired_dev=NULL
+ *
+ * Acquiring device task need to be resumed only when case 3.
+ *
+ * This scheduling can happen in either task or ISR, so `in_isr` or `bg_active` not touched.
+ *
+ * @param lock
+ * @param status Current status
+ * @param out_desired_dev Desired device to work next, see above.
+ *
+ * @return False if BG should still be the acquiring processor, otherwise True (yield to task).
+ */
+IRAM_ATTR static inline bool
+schedule_core(spi_bus_lock_t *lock, uint32_t status, spi_bus_lock_dev_t **out_desired_dev)
+{
+    spi_bus_lock_dev_t* desired_dev = NULL;
+    uint32_t lock_bits = (status & LOCK_MASK) >> LOCK_SHIFT;
+    uint32_t bg_bits = status & BG_MASK;
+    bg_bits = ((bg_bits >> REQ_SHIFT) | (bg_bits >> PENDING_SHIFT)) & REQ_MASK;
+
+    bool bg_yield;
+    if (lock_bits) {
+        int dev_id = mask_get_id(lock_bits);
+        desired_dev = (spi_bus_lock_dev_t *)atomic_load(&lock->dev[dev_id]);
+        BUS_LOCK_DEBUG_EXECUTE_CHECK(desired_dev);
+
+        lock->acquiring_dev = desired_dev;
+        bg_yield = ((bg_bits & desired_dev->mask) == 0);
+        lock->acq_dev_bg_active = !bg_yield;
+    } else {
+        lock->acq_dev_bg_active = false;
+        if (bg_bits) {
+            int dev_id = mask_get_id(bg_bits);
+            desired_dev = (spi_bus_lock_dev_t *)atomic_load(&lock->dev[dev_id]);
+            BUS_LOCK_DEBUG_EXECUTE_CHECK(desired_dev);
+
+            lock->acquiring_dev = NULL;
+            bg_yield = false;
+        } else {
+            desired_dev = NULL;
+            lock->acquiring_dev = NULL;
+            bg_yield = true;
+        }
+    }
+    *out_desired_dev = desired_dev;
+    return bg_yield;
+}
+
+//Clear the LOCK bit and trigger a rescheduling.
+IRAM_ATTR static inline void acquire_end_core(spi_bus_lock_dev_t *dev_handle)
+{
+    spi_bus_lock_t* lock = dev_handle->parent;
+    //uint32_t status = lock_status_clear(lock, dev_handle->mask & LOCK_MASK);
+    spi_bus_lock_dev_t* desired_dev = NULL;
+	
+    portENTER_CRITICAL_SAFE(&s_spinlock);
+    uint32_t status = lock_status_clear(lock, dev_handle->mask & LOCK_MASK);
+    bool invoke_bg = !schedule_core(lock, status, &desired_dev);
+    portEXIT_CRITICAL_SAFE(&s_spinlock);
+
+    if (invoke_bg) {
+        bg_enable(lock);
+    } else if (desired_dev) {
+        resume_dev(desired_dev);
+    } else if (status & WEAK_BG_FLAG) {
+        bg_enable(lock);
+    }
+}
+
+// Move the REQ bits to corresponding PEND bits. Must be called by acquiring processor.
+// Have no side effects on the acquiring device/processor.
+SPI_MASTER_ISR_ATTR static inline void update_pend_core(spi_bus_lock_t *lock, uint32_t status)
+{
+    uint32_t active_req_bits = status & REQ_MASK;
+#if PENDING_SHIFT > REQ_SHIFT
+    uint32_t pending_mask = active_req_bits << (PENDING_SHIFT - REQ_SHIFT);
+#else
+    uint32_t pending_mask = active_req_bits >> (REQ_SHIFT - PENDING_SHIFT);
+#endif
+    // We have to set the PEND bits and then clear the REQ bits, since BG bits are using bitwise OR logic,
+    // this will not influence the effectiveness of the BG bits of every device.
+    lock_status_fetch_set(lock, pending_mask);
+    lock_status_fetch_clear(lock, active_req_bits);
+}
+
+// Clear the PEND bit (not REQ bit!) of a device, return the suggestion whether we can try to quit the ISR.
+// Lost the acquiring processor immediately when the BG bits for active device are inactive, indiciating by the return value.
+// Can be called only when ISR is acting as the acquiring processor.
+SPI_MASTER_ISR_ATTR static inline bool clear_pend_core(spi_bus_lock_dev_t *dev_handle)
+{
+    bool finished;
+    spi_bus_lock_t *lock = dev_handle->parent;
+    uint32_t pend_mask = DEV_PEND_MASK(dev_handle);
+    BUS_LOCK_DEBUG_EXECUTE_CHECK(lock_status_fetch(lock) & pend_mask);
+
+    uint32_t status = lock_status_clear(lock, pend_mask);
+
+    if (lock->acquiring_dev == dev_handle) {
+        finished = ((status & DEV_REQ_MASK(dev_handle)) == 0);
+        if (finished) {
+            lock->acq_dev_bg_active = false;
+        }
+    } else {
+        finished = (status == 0);
+    }
+    return finished;
+}
+
+// Return true if the ISR has already touched the HW, which means previous operations should
+// be terminated first, before we use the HW again. Otherwise return false.
+// In either case `in_isr` will be marked as true, until call to `bg_exit_core` with `wip=false` successfully.
+SPI_MASTER_ISR_ATTR static inline bool bg_entry_core(spi_bus_lock_t *lock)
+{
+    BUS_LOCK_DEBUG_EXECUTE_CHECK(!lock->acquiring_dev || lock->acq_dev_bg_active);
+    /*
+     * The interrupt is disabled at the entry of ISR to avoid competitive risk as below:
+     *
+     * The `esp_intr_enable` will be called (b) after new BG request is queued (a) in the task;
+     * while `esp_intr_disable` should be called (c) if we check and found the sending queue is empty (d).
+     * If (c) happens after (d), if things happens in this sequence:
+     * (d) -> (a) -> (b) -> (c), the interrupt will be disabled while there's pending BG request in the queue.
+     *
+     * To avoid this, interrupt is disabled here, and re-enabled later if required. (c) -> (d) -> (a) -> (b) -> revert (c) if !d
+     */
+    bg_disable(lock);
+    if (lock->in_isr) {
+        return false;
+    } else {
+        lock->in_isr = true;
+        return true;
+    }
+}
+
+// Handle the conditions of status and interrupt, avoiding the ISR being disabled when there is any new coming BG requests.
+// When called with `wip=true`, means the ISR is performing some operations. Will enable the interrupt again and exit unconditionally.
+// When called with `wip=false`, will only return `true` when there is no coming BG request. If return value is `false`, the ISR should try again.
+// Will not change acquiring device.
+SPI_MASTER_ISR_ATTR static inline bool bg_exit_core(spi_bus_lock_t *lock, bool wip, BaseType_t *do_yield)
+{
+    //See comments in `bg_entry_core`, re-enable interrupt disabled in entry if we do need the interrupt
+    if (wip) {
+        bg_enable(lock);
+        BUS_LOCK_DEBUG_EXECUTE_CHECK(!lock->acquiring_dev || lock->acq_dev_bg_active);
+        return true;
+    }
+
+    bool ret;
+    uint32_t status = lock_status_fetch(lock);
+    if (lock->acquiring_dev) {
+        if (status & DEV_BG_MASK(lock->acquiring_dev)) {
+            BUS_LOCK_DEBUG_EXECUTE_CHECK(lock->acq_dev_bg_active);
+            ret = false;
+        } else {
+            // The request may happen any time, even after we fetched the status.
+            // The value of `acq_dev_bg_active` is random.
+            resume_dev_in_isr(lock->acquiring_dev, do_yield);
+            ret = true;
+        }
+    } else {
+        BUS_LOCK_DEBUG_EXECUTE_CHECK(!lock->acq_dev_bg_active);
+        ret = !(status & BG_MASK);
+    }
+    if (ret) {
+        //when successfully exit, but no transaction done, mark BG as inactive
+        lock->in_isr = false;
+    }
+    return ret;
+}
+
+IRAM_ATTR static inline void dev_wait_prepare(spi_bus_lock_dev_t *dev_handle)
+{
+    xSemaphoreTake(dev_handle->semphr, 0);
+}
+
+SPI_MASTER_ISR_ATTR static inline esp_err_t dev_wait(spi_bus_lock_dev_t *dev_handle, TickType_t wait)
+{
+    BaseType_t ret = xSemaphoreTake(dev_handle->semphr, wait);
+
+    if (ret == pdFALSE) return ESP_ERR_TIMEOUT;
+    return ESP_OK;
+}
+
+/*******************************************************************************
+ * Initialization & Deinitialization
+ ******************************************************************************/
+esp_err_t spi_bus_init_lock(spi_bus_lock_handle_t *out_lock, const spi_bus_lock_config_t *config)
+{
+    spi_bus_lock_t* lock = (spi_bus_lock_t*)calloc(sizeof(spi_bus_lock_t), 1);
+    if (lock == NULL) {
+        return ESP_ERR_NO_MEM;
+    }
+
+    lock_status_init(lock);
+    lock->acquiring_dev = NULL;
+    lock->last_dev = NULL;
+    lock->periph_cs_num = config->cs_num;
+    lock->host_id = config->host_id;
+
+    *out_lock = lock;
+    return ESP_OK;
+}
+
+void spi_bus_deinit_lock(spi_bus_lock_handle_t lock)
+{
+    for (int i = 0; i < DEV_NUM_MAX; i++) {
+        assert(atomic_load(&lock->dev[i]) == (intptr_t)NULL);
+    }
+    free(lock);
+}
+
+static int try_acquire_free_dev(spi_bus_lock_t *lock, bool cs_required)
+{
+    if (cs_required) {
+        int i;
+        for (i = 0; i < lock->periph_cs_num; i++) {
+            intptr_t null = (intptr_t) NULL;
+            //use 1 to occupy the slot, actual setup comes later
+            if (atomic_compare_exchange_strong(&lock->dev[i], &null, (intptr_t) 1)) {
+                break;
+            }
+        }
+        return ((i == lock->periph_cs_num)? -1: i);
+    } else {
+        int i;
+        for (i = DEV_NUM_MAX - 1; i >= 0; i--) {
+            intptr_t null = (intptr_t) NULL;
+            //use 1 to occupy the slot, actual setup comes later
+            if (atomic_compare_exchange_strong(&lock->dev[i], &null, (intptr_t) 1)) {
+                break;
+            }
+        }
+        return i;
+    }
+}
+
+esp_err_t spi_bus_lock_register_dev(spi_bus_lock_handle_t lock, spi_bus_lock_dev_config_t *config,
+                                    spi_bus_lock_dev_handle_t *out_dev_handle)
+{
+    if (lock == NULL) return ESP_ERR_INVALID_ARG;
+    int id = try_acquire_free_dev(lock, config->flags & SPI_BUS_LOCK_DEV_FLAG_CS_REQUIRED);
+    if (id == -1) return ESP_ERR_NOT_SUPPORTED;
+
+    spi_bus_lock_dev_t* dev_lock = (spi_bus_lock_dev_t*)heap_caps_calloc(sizeof(spi_bus_lock_dev_t), 1, MALLOC_CAP_INTERNAL | MALLOC_CAP_8BIT);
+    if (dev_lock == NULL) {
+        return ESP_ERR_NO_MEM;
+    }
+    dev_lock->semphr = xSemaphoreCreateBinary();
+    if (dev_lock->semphr == NULL) {
+        free(dev_lock);
+        atomic_store(&lock->dev[id], (intptr_t)NULL);
+        return ESP_ERR_NO_MEM;
+    }
+    dev_lock->parent = lock;
+    dev_lock->mask = DEV_MASK(id);
+
+    ESP_LOGV(TAG, "device registered on bus %d slot %d.", lock->host_id, id);
+    atomic_store(&lock->dev[id], (intptr_t)dev_lock);
+    *out_dev_handle = dev_lock;
+    return ESP_OK;
+}
+
+void spi_bus_lock_unregister_dev(spi_bus_lock_dev_handle_t dev_handle)
+{
+    int id = dev_lock_get_id(dev_handle);
+
+    spi_bus_lock_t* lock = dev_handle->parent;
+    BUS_LOCK_DEBUG_EXECUTE_CHECK(atomic_load(&lock->dev[id]) == (intptr_t)dev_handle);
+
+    if (lock->last_dev == dev_handle) lock->last_dev = NULL;
+
+    atomic_store(&lock->dev[id], (intptr_t)NULL);
+    if (dev_handle->semphr) {
+        vSemaphoreDelete(dev_handle->semphr);
+    }
+
+    free(dev_handle);
+}
+
+IRAM_ATTR static inline int mask_get_id(uint32_t mask)
+{
+    return ID_DEV_MASK(mask);
+}
+
+IRAM_ATTR static inline int dev_lock_get_id(spi_bus_lock_dev_t *dev_lock)
+{
+    return mask_get_id(dev_lock->mask);
+}
+
+void spi_bus_lock_set_bg_control(spi_bus_lock_handle_t lock, bg_ctrl_func_t bg_enable, bg_ctrl_func_t bg_disable, void *arg)
+{
+    lock->bg_enable = bg_enable;
+    lock->bg_disable = bg_disable;
+    lock->bg_arg = arg;
+}
+
+IRAM_ATTR int spi_bus_lock_get_dev_id(spi_bus_lock_dev_handle_t dev_handle)
+{
+    return (dev_handle? dev_lock_get_id(dev_handle): -1);
+}
+
+//will be called when cache disabled
+IRAM_ATTR bool spi_bus_lock_touch(spi_bus_lock_dev_handle_t dev_handle)
+{
+    spi_bus_lock_dev_t* last_dev = dev_handle->parent->last_dev;
+    dev_handle->parent->last_dev = dev_handle;
+    if (last_dev != dev_handle) {
+        int last_dev_id = (last_dev? dev_lock_get_id(last_dev): -1);
+        ESP_DRAM_LOGV(TAG, "SPI dev changed from %d to %d",
+                    last_dev_id, dev_lock_get_id(dev_handle));
+    }
+    return (dev_handle != last_dev);
+}
+
+/*******************************************************************************
+ * Acquiring service
+ ******************************************************************************/
+IRAM_ATTR esp_err_t spi_bus_lock_acquire_start(spi_bus_lock_dev_t *dev_handle, TickType_t wait)
+{
+    LOCK_CHECK(wait == portMAX_DELAY, "timeout other than portMAX_DELAY not supported", ESP_ERR_INVALID_ARG);
+
+    spi_bus_lock_t* lock = dev_handle->parent;
+
+    // Clear the semaphore before checking
+    dev_wait_prepare(dev_handle);
+    if (!acquire_core(dev_handle)) {
+        //block until becoming the acquiring processor (help by previous acquiring processor)
+        esp_err_t err = dev_wait(dev_handle, wait);
+        //TODO: add timeout handling here.
+        if (err != ESP_OK) return err;
+    }
+
+    ESP_DRAM_LOGV(TAG, "dev %d acquired.", dev_lock_get_id(dev_handle));
+    BUS_LOCK_DEBUG_EXECUTE_CHECK(lock->acquiring_dev == dev_handle);
+
+    //When arrives at here, requests of this device should already be handled
+    uint32_t status = lock_status_fetch(lock);
+    (void) status;
+    BUS_LOCK_DEBUG_EXECUTE_CHECK((status & DEV_BG_MASK(dev_handle)) == 0);
+
+    return ESP_OK;
+}
+
+IRAM_ATTR esp_err_t spi_bus_lock_acquire_end(spi_bus_lock_dev_t *dev_handle)
+{
+    //release the bus
+    spi_bus_lock_t* lock = dev_handle->parent;
+    LOCK_CHECK(lock->acquiring_dev == dev_handle, "Cannot release a lock that hasn't been acquired.", ESP_ERR_INVALID_STATE);
+
+    acquire_end_core(dev_handle);
+
+    ESP_LOGV(TAG, "dev %d released.", dev_lock_get_id(dev_handle));
+    return ESP_OK;
+}
+
+SPI_MASTER_ISR_ATTR spi_bus_lock_dev_handle_t spi_bus_lock_get_acquiring_dev(spi_bus_lock_t *lock)
+{
+    return lock->acquiring_dev;
+}
+
+/*******************************************************************************
+ * BG (background operation) service
+ ******************************************************************************/
+SPI_MASTER_ISR_ATTR bool spi_bus_lock_bg_entry(spi_bus_lock_t* lock)
+{
+    return bg_entry_core(lock);
+}
+
+SPI_MASTER_ISR_ATTR bool spi_bus_lock_bg_exit(spi_bus_lock_t* lock, bool wip, BaseType_t* do_yield)
+{
+    return bg_exit_core(lock, wip, do_yield);
+}
+
+SPI_MASTER_ATTR esp_err_t spi_bus_lock_bg_request(spi_bus_lock_dev_t *dev_handle)
+{
+    req_core(dev_handle);
+    return ESP_OK;
+}
+
+IRAM_ATTR esp_err_t spi_bus_lock_wait_bg_done(spi_bus_lock_dev_handle_t dev_handle, TickType_t wait)
+{
+    spi_bus_lock_t *lock = dev_handle->parent;
+    LOCK_CHECK(lock->acquiring_dev == dev_handle, "Cannot wait for a device that is not acquired", ESP_ERR_INVALID_STATE);
+    LOCK_CHECK(wait == portMAX_DELAY, "timeout other than portMAX_DELAY not supported", ESP_ERR_INVALID_ARG);
+
+    // If no BG bits active, skip quickly. This is ensured by `spi_bus_lock_wait_bg_done`
+    // cannot be executed with `bg_request` on the same device concurrently.
+    if (lock_status_fetch(lock) & DEV_BG_MASK(dev_handle)) {
+        // Clear the semaphore before checking
+        dev_wait_prepare(dev_handle);
+        if (lock_status_fetch(lock) & DEV_BG_MASK(dev_handle)) {
+            //block until becoming the acquiring processor (help by previous acquiring processor)
+            esp_err_t err = dev_wait(dev_handle, wait);
+            //TODO: add timeout handling here.
+            if (err != ESP_OK) return err;
+        }
+    }
+
+    BUS_LOCK_DEBUG_EXECUTE_CHECK(!lock->acq_dev_bg_active);
+    BUS_LOCK_DEBUG_EXECUTE_CHECK((lock_status_fetch(lock) & DEV_BG_MASK(dev_handle)) == 0);
+    return ESP_OK;
+}
+
+SPI_MASTER_ISR_ATTR bool spi_bus_lock_bg_clear_req(spi_bus_lock_dev_t *dev_handle)
+{
+    bool finished = clear_pend_core(dev_handle);
+    ESP_EARLY_LOGV(TAG, "dev %d served from bg.", dev_lock_get_id(dev_handle));
+    return finished;
+}
+
+SPI_MASTER_ISR_ATTR bool spi_bus_lock_bg_check_dev_acq(spi_bus_lock_t *lock,
+                                                       spi_bus_lock_dev_handle_t *out_dev_lock)
+{
+    BUS_LOCK_DEBUG_EXECUTE_CHECK(!lock->acquiring_dev);
+    uint32_t status = lock_status_fetch(lock);
+    return schedule_core(lock, status, out_dev_lock);
+}
+
+SPI_MASTER_ISR_ATTR bool spi_bus_lock_bg_check_dev_req(spi_bus_lock_dev_t *dev_lock)
+{
+    spi_bus_lock_t* lock = dev_lock->parent;
+    uint32_t status = lock_status_fetch(lock);
+    uint32_t dev_status = status & dev_lock->mask;
+
+    // move REQ bits of all device to corresponding PEND bits.
+    // To reduce executing time, only done when the REQ bit of the calling device is set.
+    if (dev_status & REQ_MASK) {
+        update_pend_core(lock, status);
+        return true;
+    } else {
+        return dev_status & PEND_MASK;
+    }
+}
+
+SPI_MASTER_ISR_ATTR bool spi_bus_lock_bg_req_exist(spi_bus_lock_t *lock)
+{
+    uint32_t status = lock_status_fetch(lock);
+    return status & BG_MASK;
+}
+
+/*******************************************************************************
+ * Static variables of the locks of the main flash
+ ******************************************************************************/
+#if CONFIG_SPI_FLASH_SHARE_SPI1_BUS
+static spi_bus_lock_dev_t lock_main_flash_dev;
+
+static spi_bus_lock_t main_spi_bus_lock = {
+    /*
+     * the main bus cache is permanently required, this flag is set here and never clear so that the
+     * cache will always be enabled if acquiring devices yield.
+     */
+    .status = ATOMIC_VAR_INIT(WEAK_BG_FLAG),
+    .acquiring_dev = NULL,
+    .dev = {ATOMIC_VAR_INIT((intptr_t)&lock_main_flash_dev)},
+    .new_req = 0,
+    .periph_cs_num = SOC_SPI_PERIPH_CS_NUM(0),
+};
+const spi_bus_lock_handle_t g_main_spi_bus_lock = &main_spi_bus_lock;
+
+esp_err_t spi_bus_lock_init_main_bus(void)
+{
+    spi_bus_main_set_lock(g_main_spi_bus_lock);
+    return ESP_OK;
+}
+
+static StaticSemaphore_t main_flash_semphr;
+
+static spi_bus_lock_dev_t lock_main_flash_dev = {
+    .semphr = NULL,
+    .parent = &main_spi_bus_lock,
+    .mask = DEV_MASK(0),
+};
+const spi_bus_lock_dev_handle_t g_spi_lock_main_flash_dev = &lock_main_flash_dev;
+
+esp_err_t spi_bus_lock_init_main_dev(void)
+{
+    g_spi_lock_main_flash_dev->semphr = xSemaphoreCreateBinaryStatic(&main_flash_semphr);
+    if (g_spi_lock_main_flash_dev->semphr == NULL) {
+        return ESP_ERR_NO_MEM;
+    }
+    return ESP_OK;
+}
+#else //CONFIG_SPI_FLASH_SHARE_SPI1_BUS
+
+//when the dev lock is not initialized, point to NULL
+const spi_bus_lock_dev_handle_t g_spi_lock_main_flash_dev = NULL;
+
+#endif

components/codecs/CMakeLists.txt

@@ -1,24 +1,28 @@
-idf_component_register(SRC_DIRS . 
-					   INCLUDE_DIRS .  ./inc  inc/alac inc/FLAC  inc/helix-aac inc/mad inc/ogg inc/opus inc/opusfile inc/resample16 inc/soxr inc/vorbis
-					   PRIV_REQUIRES newlib
+idf_component_register(
+	   INCLUDE_DIRS . ./inc inc/alac inc/FLAC inc/helix-aac inc/mad inc/ogg inc/opus inc/opusfile inc/resample16 inc/soxr inc/vorbis
 )
 
+if (DEFINED AAC_DISABLE_SBR)
+add_prebuilt_library(libhelix-aac 	lib/libhelix-aac.a ) 
+else ()
+add_prebuilt_library(libhelix-aac 	lib/libhelix-aac-sbr.a ) 
+endif()
+
 add_prebuilt_library(libmad 		lib/libmad.a)
 add_prebuilt_library(libFLAC 		lib/libFLAC.a ) 
-add_prebuilt_library(libhelix-aac 	lib/libhelix-aac.a ) 
 add_prebuilt_library(libvorbisidec 	lib/libvorbisidec.a ) 
 add_prebuilt_library(libogg 		lib/libogg.a )
 add_prebuilt_library(libalac 		lib/libalac.a ) 
-add_prebuilt_library(libremple16 	lib/libresample16.a ) 
+add_prebuilt_library(libresample16 	lib/libresample16.a ) 
 add_prebuilt_library(libopusfile 	lib/libopusfile.a ) 
 add_prebuilt_library(libopus 		lib/libopus.a ) 
-target_link_libraries(${COMPONENT_LIB} PRIVATE libmad)
-target_link_libraries(${COMPONENT_LIB} PRIVATE libFLAC)
-target_link_libraries(${COMPONENT_LIB} PRIVATE libhelix-aac)
-target_link_libraries(${COMPONENT_LIB} PRIVATE libvorbisidec)
 
-target_link_libraries(${COMPONENT_LIB} PRIVATE libogg)
-target_link_libraries(${COMPONENT_LIB} PRIVATE libalac)
-target_link_libraries(${COMPONENT_LIB} PRIVATE libremple16)
-target_link_libraries(${COMPONENT_LIB} PRIVATE libopusfile)
-target_link_libraries(${COMPONENT_LIB} PRIVATE libopus)
+target_link_libraries(${COMPONENT_LIB} INTERFACE libmad)
+target_link_libraries(${COMPONENT_LIB} INTERFACE libFLAC)
+target_link_libraries(${COMPONENT_LIB} INTERFACE libhelix-aac)
+target_link_libraries(${COMPONENT_LIB} INTERFACE libvorbisidec)
+target_link_libraries(${COMPONENT_LIB} INTERFACE libogg)
+target_link_libraries(${COMPONENT_LIB} INTERFACE libalac)
+target_link_libraries(${COMPONENT_LIB} INTERFACE libresample16)
+target_link_libraries(${COMPONENT_LIB} INTERFACE libopusfile)
+target_link_libraries(${COMPONENT_LIB} INTERFACE libopus)

components/codecs/inc/alac/alac_wrapper.h

@@ -19,7 +19,7 @@ extern "C" {
 
 struct alac_codec_s *alac_create_decoder(int magic_cookie_size, unsigned char *magic_cookie,
 								unsigned char *sample_size, unsigned *sample_rate,
-								unsigned char *channels);
+								unsigned char *channels, unsigned int *block_size);
 void alac_delete_decoder(struct alac_codec_s *codec);
 bool alac_to_pcm(struct alac_codec_s *codec, unsigned char* input,
 				 unsigned char *output, char channels, unsigned *out_frames);

components/codecs/link_helper.c

@@ -1,3 +0,0 @@
-void dummy_obj() {
-	return;
-}

components/display/CMakeLists.txt

@@ -1,11 +1,16 @@
+# the JPEG library is in ROM but seems to fail randomly (PSRAM issue?)
+set(TJPGD tjpgd)
 
 idf_component_register(SRC_DIRS . core core/ifaces  fonts
 						INCLUDE_DIRS . fonts core
 						REQUIRES platform_config tools esp_common
-						PRIV_REQUIRES services freertos driver           
-)
+						PRIV_REQUIRES services freertos driver ${TJPGD}
+						EMBED_FILES note.jpg )		
+
+if (NOT TJPGD)
+add_compile_definitions(TJPGD_ROM)												
+endif()
 
 set_source_files_properties(display.c
     PROPERTIES COMPILE_FLAGS
-   -Wno-format-overflow
-)
+   -Wno-format-overflow )

components/display/ILI9341.c

@@ -1,6 +1,7 @@
 /**
  * Copyright (c) 2017-2018 Tara Keeling
  *				 2020 Philippe G.
+ *				 2021 Mumpf and Harry1999
  * 
  * This software is released under the MIT License.
  * https://opensource.org/licenses/MIT
@@ -10,356 +11,337 @@
 #include <string.h>
 #include <stdint.h>
 #include <stdbool.h>
-#include "freertos/FreeRTOS.h"
-#include "freertos/task.h"
-#include "driver/gpio.h"
-
 #include <esp_heap_caps.h>
 #include <esp_log.h>
 
 #include "gds.h"
 #include "gds_private.h"
 
-//#define SHADOW_BUFFER
-#define PAGE_BLOCK	1024
+#define SHADOW_BUFFER
+#define USE_IRAM
+#define PAGE_BLOCK		2048
+#define ENABLE_WRITE	0x2c
+#define MADCTL_MX  0x40
+#define TFT_RGB_BGR  0x08
 
 #define min(a,b) (((a) < (b)) ? (a) : (b))
 
 static char TAG[] = "ILI9341";
 
-
-#define L1_CMD_NOP 0X00
-#define L1_CMD_SOFTWARE_RESET 0X01
-#define L1_CMD_READ_DISPLAY_IDENTIFICATION_INFORMATION 0X04
-#define L1_CMD_READ_DISPLAY_STATUS 0X09
-#define L1_CMD_READ_DISPLAY_POWER_MODE 0X0A
-#define L1_CMD_READ_DISPLAY_MADCTL 0X0B
-#define L1_CMD_READ_DISPLAY_PIXEL_FORMAT 0X0C
-#define L1_CMD_READ_DISPLAY_IMAGE_FORMAT 0X0D
-#define L1_CMD_READ_DISPLAY_SIGNAL_MODE 0X0E
-#define L1_CMD_READ_DISPLAY_SELF_DIAGNOSTIC_RESULT 0X0F
-#define L1_CMD_ENTER_SLEEP_MODE 0X10
-#define L1_CMD_SLEEP_OUT 0X11
-#define L1_CMD_PARTIAL_MODE_ON 0X12
-#define L1_CMD_NORMAL_DISPLAY_MODE_ON 0X13
-#define L1_CMD_DISPLAY_INVERSION_OFF 0X20
-#define L1_CMD_DISPLAY_INVERSION_ON 0X21
-#define L1_CMD_GAMMA_SET 0X26
-#define L1_CMD_DISPLAY_OFF 0X28
-#define L1_CMD_DISPLAY_ON 0X29
-#define L1_CMD_COLUMN_ADDRESS_SET 0X2A
-#define L1_CMD_PAGE_ADDRESS_SET 0X2B
-#define L1_CMD_MEMORY_WRITE 0X2C
-#define L1_CMD_COLOR_SET 0X2D
-#define L1_CMD_MEMORY_READ 0X2E
-#define L1_CMD_PARTIAL_AREA 0X30
-#define L1_CMD_VERTICAL_SCROLLING_DEFINITION 0X33
-#define L1_CMD_TEARING_EFFECT_LINE_OFF 0X34
-#define L1_CMD_TEARING_EFFECT_LINE_ON 0X35
-#define L1_CMD_MEMORY_ACCESS_CONTROL 0X36
-#define L1_CMD_VERTICAL_SCROLLING_START_ADDRESS 0X37
-#define L1_CMD_IDLE_MODE_OFF 0X38
-#define L1_CMD_IDLE_MODE_ON 0X39
-#define L1_CMD_COLMOD_PIXEL_FORMAT_SET 0X3A
-#define L1_CMD_WRITE_MEMORY_CONTINUE 0X3C
-#define L1_CMD_READ_MEMORY_CONTINUE 0X3E
-#define L1_CMD_SET_TEAR_SCANLINE 0X44
-#define L1_CMD_GET_SCANLINE 0X45
-#define L1_CMD_WRITE_DISPLAY_BRIGHTNESS 0X51
-#define L1_CMD_READ_DISPLAY_BRIGHTNESS 0X52
-#define L1_CMD_WRITE_CTRL_DISPLAY 0X53
-#define L1_CMD_READ_CTRL_DISPLAY 0X54
-#define L1_CMD_WRITE_CONTENT_ADAPTIVE_BRIGHTNESS_CONTROL 0X55
-#define L1_CMD_READ_CONTENT_ADAPTIVE_BRIGHTNESS_CONTROL 0X56
-#define L1_CMD_WRITE_CABC_MINIMUM_BRIGHTNESS 0X5E
-#define L1_CMD_READ_CABC_MINIMUM_BRIGHTNESS 0X5F
-#define L1_CMD_READ_ID1 0XDA
-#define L1_CMD_READ_ID2 0XDB
-#define L1_CMD_READ_ID3 0XDC
-
-#define L2_CMD_RGB_INTERFACE_SIGNAL_CONTROL 0XB0
-#define L2_CMD_FRAME_RATE_CONTROL_IN_NORMAL_MODE_FULL_COLORS 0XB1
-#define L2_CMD_FRAME_RATE_CONTROL_IN_IDLE_MODE_8_COLORS 0XB2
-#define L2_CMD_FRAME_RATE_CONTROL_IN_PARTIAL_MODE_FULL_COLORS 0XB3
-#define L2_CMD_DISPLAY_INVERSION_CONTROL 0XB4
-#define L2_CMD_BLANKING_PORCH_CONTROL 0XB5
-#define L2_CMD_DISPLAY_FUNCTION_CONTROL 0XB6
-#define L2_CMD_ENTRY_MODE_SET 0XB7
-#define L2_CMD_BACKLIGHT_CONTROL_1 0XB8
-#define L2_CMD_BACKLIGHT_CONTROL_2 0XB9
-#define L2_CMD_BACKLIGHT_CONTROL_3 0XBA
-#define L2_CMD_BACKLIGHT_CONTROL_4 0XBB
-#define L2_CMD_BACKLIGHT_CONTROL_5 0XBC
-#define L2_CMD_BACKLIGHT_CONTROL_7 0XBE
-#define L2_CMD_BACKLIGHT_CONTROL_8 0XBF
-#define L2_CMD_POWER_CONTROL_1 0XC0
-#define L2_CMD_POWER_CONTROL_2 0XC1
-#define L2_CMD_VCOM_CONTROL_1 0XC5
-#define L2_CMD_VCOM_CONTROL_2 0XC7
-#define L2_CMD_NV_MEMORY_WRITE 0XD0
-#define L2_CMD_NV_MEMORY_PROTECTION_KEY 0XD1
-#define L2_CMD_NV_MEMORY_STATUS_READ 0XD2
-#define L2_CMD_READ_ID4 0XD3
-#define L2_CMD_POSITIVE_GAMMA_CORRECTION 0XE0
-#define L2_CMD_NEGATIVE_GAMMA_CORRECTION 0XE1
-#define L2_CMD_DIGITAL_GAMMA_CONTROL_1 0XE2
-#define L2_CMD_DIGITAL_GAMMA_CONTROL_2 0XE3
-#define L2_CMD_INTERFACE_CONTROL 0XF6
-
-
-
-/*
- The LCD needs a bunch of command/argument values to be initialized. They are stored in this struct.
-*/
-typedef struct {
-    uint8_t cmd;
-    uint8_t data[16];
-    uint8_t databytes; //No of data in data; bit 7 = delay after set; 0xFF = end of cmds.
-} lcd_init_cmd_t;
-
-
-
-static const lcd_init_cmd_t ili_init_cmds[]={
-    /* Power contorl B, power control = 0, DC_ENA = 1 */
-    {0xCF, {0x00, 0x83, 0X30}, 3},
-    /* Power on sequence control,
-     * cp1 keeps 1 frame, 1st frame enable
-     * vcl = 0, ddvdh=3, vgh=1, vgl=2
-     * DDVDH_ENH=1
-     */
-    {0xED, {0x64, 0x03, 0X12, 0X81}, 4},
-    /* Driver timing control A,
-     * non-overlap=default +1
-     * EQ=default - 1, CR=default
-     * pre-charge=default - 1
-     */
-    {0xE8, {0x85, 0x01, 0x79}, 3},
-    /* Power control A, Vcore=1.6V, DDVDH=5.6V */
-    {0xCB, {0x39, 0x2C, 0x00, 0x34, 0x02}, 5},
-    /* Pump ratio control, DDVDH=2xVCl */
-    {0xF7, {0x20}, 1},
-    /* Driver timing control, all=0 unit */
-    {0xEA, {0x00, 0x00}, 2},
-    /* Power control 1, GVDD=4.75V */
-    {0xC0, {0x26}, 1},
-    /* Power control 2, DDVDH=VCl*2, VGH=VCl*7, VGL=-VCl*3 */
-    {0xC1, {0x11}, 1},
-    /* VCOM control 1, VCOMH=4.025V, VCOML=-0.950V */
-    {0xC5, {0x35, 0x3E}, 2},
-    /* VCOM control 2, VCOMH=VMH-2, VCOML=VML-2 */
-    {0xC7, {0xBE}, 1},
-    /* Memory access contorl, MX=MY=0, MV=1, ML=0, BGR=1, MH=0 */
-    {0x36, {0x28}, 1},
-    /* Pixel format, 16bits/pixel for RGB/MCU interface */
-    {0x3A, {0x55}, 1},
-    /* Frame rate control, f=fosc, 70Hz fps */
-    {0xB1, {0x00, 0x1B}, 2},
-    /* Enable 3G, disabled */
-    {0xF2, {0x08}, 1},
-    /* Gamma set, curve 1 */
-    {0x26, {0x01}, 1},
-    /* Positive gamma correction */
-    {0xE0, {0x1F, 0x1A, 0x18, 0x0A, 0x0F, 0x06, 0x45, 0X87, 0x32, 0x0A, 0x07, 0x02, 0x07, 0x05, 0x00}, 15},
-    /* Negative gamma correction */
-    {0XE1, {0x00, 0x25, 0x27, 0x05, 0x10, 0x09, 0x3A, 0x78, 0x4D, 0x05, 0x18, 0x0D, 0x38, 0x3A, 0x1F}, 15},
-    /* Column address set, SC=0, EC=0xEF */
-    {0x2A, {0x00, 0x00, 0x00, 0xEF}, 4},
-    /* Page address set, SP=0, EP=0x013F */
-    {0x2B, {0x00, 0x00, 0x01, 0x3f}, 4},
-    /* Memory write */
-    {0x2C, {0}, 0},
-    /* Entry mode set, Low vol detect disabled, normal display */
-    {0xB7, {0x07}, 1},
-    /* Display function control */
-    {0xB6, {0x0A, 0x82, 0x27, 0x00}, 4},
-    /* Sleep out */
-    {0x11, {0}, 0x80},
-    /* Display on */
-    {0x29, {0}, 0x80},
-    {0, {0}, 0xff},
-};
-
-//To speed up transfers, every SPI transfer sends a bunch of lines. This define specifies how many. More means more memory use,
-//but less overhead for setting up / finishing transfers. Make sure 240 is dividable by this.
-#define PARALLEL_LINES 16
+enum { ILI9341, ILI9341_24 };	//ILI9341_24 for future use...
 
 struct PrivateSpace {
 	uint8_t *iRAM, *Shadowbuffer;
-	uint8_t ReMap, PageSize;
-	uint8_t Offset;
+	struct {
+		uint16_t Height, Width;
+	} Offset;
+	uint8_t MADCtl, PageSize;
+	uint8_t Model;
 };
 
 // Functions are not declared to minimize # of lines
 
-static void WriteDataByte( struct GDS_Device* Device, uint8_t Data ) {
-	Device->WriteData( Device, &Data, 1);
+static void WriteByte( struct GDS_Device* Device, uint8_t Data ) {
+	Device->WriteData( Device, &Data, 1 );
 }
 
-static void SetColumnAddress( struct GDS_Device* Device, uint8_t Start, uint8_t End ) {
-	Device->WriteCommand( Device, L1_CMD_COLUMN_ADDRESS_SET );
-	Device->WriteData( Device, &Start, 1 );
-	Device->WriteData( Device, &End, 1 );
-}
-static void SetRowAddress( struct GDS_Device* Device, uint8_t Start, uint8_t End ) {
-	Device->WriteCommand( Device, L1_CMD_PAGE_ADDRESS_SET );
-	Device->WriteData( Device, &Start, 1 );
-	Device->WriteData( Device, &End, 1 );
+static void SetColumnAddress( struct GDS_Device* Device, uint16_t Start, uint16_t End ) {
+	uint32_t Addr = __builtin_bswap16(Start) | (__builtin_bswap16(End) << 16);
+	Device->WriteCommand( Device, 0x2A );
+	Device->WriteData( Device, (uint8_t*) &Addr, 4 );
 }
 
+static void SetRowAddress( struct GDS_Device* Device, uint16_t Start, uint16_t End ) {
+	uint32_t Addr = __builtin_bswap16(Start) | (__builtin_bswap16(End) << 16);
+	Device->WriteCommand( Device, 0x2B );
+	Device->WriteData( Device, (uint8_t*) &Addr, 4 );
+}
 
-
-static void Update( struct GDS_Device* Device ) {
+static void Update16( struct GDS_Device* Device ) {
 	struct PrivateSpace *Private = (struct PrivateSpace*) Device->Private;
 		
-	//SetColumnAddress( Device, Private->Offset, Private->Offset + Device->Width / 4 - 1);
-	SetColumnAddress( Device, Private->Offset, Private->Offset + Device->Width - 1);
-	
 #ifdef SHADOW_BUFFER
-	uint16_t *optr = (uint16_t*) Private->Shadowbuffer, *iptr = (uint16_t*) Device->Framebuffer;
-	bool dirty = false;
+	uint32_t *optr = (uint32_t*) Private->Shadowbuffer, *iptr = (uint32_t*) Device->Framebuffer;
+	int FirstCol = Device->Width / 2, LastCol = 0, FirstRow = -1, LastRow = 0;  
 	
-	for (int r = 0, page = 0; r < Device->Height; r++) {
-		// look for change and update shadow (cheap optimization = width always / by 2)
-		for (int c = Device->Width / 2 / 2; --c >= 0;) {
+	for (int r = 0; r < Device->Height; r++) {
+		// look for change and update shadow (cheap optimization = width is always a multiple of 2)
+		for (int c = 0; c < Device->Width / 2; c++, iptr++, optr++) {
 			if (*optr != *iptr) {
-				dirty = true;
 				*optr = *iptr;
+				if (c < FirstCol) FirstCol = c;	
+				if (c > LastCol) LastCol = c;
+				if (FirstRow < 0) FirstRow = r;
+				LastRow = r;
 			}
-			iptr++; optr++;
 		}
+
+		// wait for a large enough window - careful that window size might increase by more than a line at once !
+		if (FirstRow < 0 || ((LastCol - FirstCol + 1) * (r - FirstRow + 1) * 4 < PAGE_BLOCK && r != Device->Height - 1)) continue;
 		
-		// one line done, check for page boundary
-		if (++page == Private->PageSize) {
-			if (dirty) {
-				uint16_t *optr = (uint16_t*) Private->iRAM, *iptr = (uint16_t*) (Private->Shadowbuffer + (r - page + 1) * Device->Width / 2);
-				SetRowAddress( Device, r - page + 1, r );
-				for (int i = page * Device->Width / 2 / 2; --i >= 0; iptr++) *optr++ = (*iptr >> 8) | (*iptr << 8);
-				//memcpy(Private->iRAM, Private->Shadowbuffer + (r - page + 1) * Device->Width / 2, page * Device->Width / 2 );
-				Device->WriteCommand( Device, 0x5c );
-				Device->WriteData( Device, Private->iRAM, Device->Width * page / 2 );
-				dirty = false;
-			}	
-			page = 0;
+		FirstCol *= 2;
+		LastCol = LastCol * 2 + 1;
+		SetRowAddress( Device, FirstRow + Private->Offset.Height, LastRow + Private->Offset.Height);
+		SetColumnAddress( Device, FirstCol + Private->Offset.Width, LastCol + Private->Offset.Width );
+		Device->WriteCommand( Device, ENABLE_WRITE );
+			
+		int ChunkSize = (LastCol - FirstCol + 1) * 2;
+			
+		// own use of IRAM has not proven to be much better than letting SPI do its copy
+		if (Private->iRAM) {
+			uint8_t *optr = Private->iRAM;
+			for (int i = FirstRow; i <= LastRow; i++) {
+				memcpy(optr, Private->Shadowbuffer + (i * Device->Width + FirstCol) * 2, ChunkSize);
+				optr += ChunkSize;
+				if (optr - Private->iRAM <= (PAGE_BLOCK - ChunkSize) && i < LastRow) continue;
+				Device->WriteData(Device, Private->iRAM, optr - Private->iRAM);
+				optr = Private->iRAM;
+			}
+		} else for (int i = FirstRow; i <= LastRow; i++) {
+			Device->WriteData( Device, Private->Shadowbuffer + (i * Device->Width + FirstCol) * 2, ChunkSize );
 		}	
+
+		FirstCol = Device->Width / 2; LastCol = 0;
+		FirstRow = -1;
 	}	
 #else
-	for (int r = 0; r < Device->Height; r += Private->PageSize) {
-		SetRowAddress( Device, r, r + Private->PageSize - 1 );
-		Device->WriteCommand( Device, L1_CMD_MEMORY_WRITE );
+	// always update by full lines
+	SetColumnAddress( Device, Private->Offset.Width, Device->Width - 1);
+	
+	for (int r = 0; r < Device->Height; r += min(Private->PageSize, Device->Height - r)) {
+		int Height = min(Private->PageSize, Device->Height - r);
+		
+		SetRowAddress( Device, Private->Offset.Height + r, Private->Offset.Height + r + Height - 1 );
+		Device->WriteCommand(Device, ENABLE_WRITE);
+		
 		if (Private->iRAM) {
-			uint16_t *optr = (uint16_t*) Private->iRAM, *iptr = (uint16_t*) (Device->Framebuffer + r * Device->Width / 2);
-			for (int i = Private->PageSize * Device->Width / 2 / 2; --i >= 0; iptr++) *optr++ = (*iptr >> 8) | (*iptr << 8);
-			//memcpy(Private->iRAM, Device->Framebuffer + r * Device->Width / 2, Private->PageSize * Device->Width / 2 );
-			Device->WriteData( Device, Private->iRAM, Private->PageSize * Device->Width / 2 );
+			memcpy(Private->iRAM, Device->Framebuffer + r * Device->Width * 2, Height * Device->Width * 2 );
+			Device->WriteData( Device, Private->iRAM, Height * Device->Width * 2 );
 		} else	{
-			Device->WriteData( Device, Device->Framebuffer + r * Device->Width / 2, Private->PageSize * Device->Width / 2 );
+			Device->WriteData( Device, Device->Framebuffer + r * Device->Width * 2, Height * Device->Width * 2 );
 		}	
 	}	
 #endif	
 }
 
-//Bit 	Name 						Description
-//---	---------------------------	------------------------------------------------------
-//MY  	Row Address Order 	 		MCU to memory write/read direction.
-//MX 	Column Address Order 		MCU to memory write/read direction.
-//MV 	Row / Column Exchange 		MCU to memory write/read direction.
-//ML 	Vertical Refresh Order 		LCD vertical refresh direction control.
-//BGR 	RGB-BGR Order 				Color selector switch control
-//									(0=RGB color filter panel, 1=BGR color filter panel)
-//MH 	Horizontal Refresh ORDER 	LCD horizontal refreshing direction control.
-// Bits 17-0
-//		XX XX XX XX XX XX XX XX XX XX MY MX MV ML BGR MH 0 0
-typedef enum  {
-	MAC_BIT_MH=2,
-	MAC_BIT_BGR,
-	MAC_BIT_ML,
-	MAC_BIT_MV,
-	MAC_BIT_MX,
-	MAC_BIT_MY,
-} mac_bits;
-
-uint16_t set_mac_bit(mac_bits bit, uint16_t val){
-	return (1 << bit) | val;
-}
-uint16_t unset_mac_bit(mac_bits bit, uint16_t val){
-	return ~(1 << bit) & val;
-}
-
-static void SetLayout( struct GDS_Device* Device, bool HFlip, bool VFlip, bool Rotate ) { 
+static void Update24( struct GDS_Device* Device ) {
 	struct PrivateSpace *Private = (struct PrivateSpace*) Device->Private;
-	Private->ReMap = HFlip ? (Private->ReMap & ~(1 << MAC_BIT_MX)) : (Private->ReMap | (1 << MAC_BIT_MX));
-	Private->ReMap = VFlip ? (Private->ReMap | (1 << MAC_BIT_MY)) : (Private->ReMap & ~(1 << MAC_BIT_MY));
-	Device->WriteCommand( Device, L1_CMD_MEMORY_ACCESS_CONTROL );
-	Device->WriteData( Device, &Private->ReMap, 1 );
-	WriteDataByte(Device,0x00);
+		
+#ifdef SHADOW_BUFFER
+	uint16_t *optr = (uint16_t*) Private->Shadowbuffer, *iptr = (uint16_t*) Device->Framebuffer;
+	int FirstCol = (Device->Width * 3) / 2, LastCol = 0, FirstRow = -1, LastRow = 0;  
+
+	for (int r = 0; r < Device->Height; r++) {
+		// look for change and update shadow (cheap optimization = width always / by 2)
+		for (int c = 0; c < (Device->Width * 3) / 2; c++, optr++, iptr++) {
+			if (*optr != *iptr) {
+				*optr = *iptr;
+				if (c < FirstCol) FirstCol = c;	
+				if (c > LastCol) LastCol = c;
+				if (FirstRow < 0) FirstRow = r;
+				LastRow = r;
+			}
+		}
+
+		// do we have enough to send (cols are divided by 3/2)
+		if (FirstRow < 0 || ((((LastCol - FirstCol + 1) * 2 ) / 3) * (r - FirstRow + 1) * 4 < PAGE_BLOCK && r != Device->Height - 1)) continue;
+		
+		FirstCol = (FirstCol * 2) / 3;
+		LastCol = (LastCol * 2 + 1 ) / 3; 
+		SetRowAddress( Device, FirstRow + Private->Offset.Height, LastRow + Private->Offset.Height);
+		SetColumnAddress( Device, FirstCol + Private->Offset.Width, LastCol + Private->Offset.Width );
+		Device->WriteCommand( Device, ENABLE_WRITE );
+			
+		int ChunkSize = (LastCol - FirstCol + 1) * 3;
+					
+		// own use of IRAM has not proven to be much better than letting SPI do its copy
+		if (Private->iRAM) {
+			uint8_t *optr = Private->iRAM;
+			for (int i = FirstRow; i <= LastRow; i++) {
+				memcpy(optr, Private->Shadowbuffer + (i * Device->Width + FirstCol) * 3, ChunkSize);
+				optr += ChunkSize;
+				if (optr - Private->iRAM <= (PAGE_BLOCK - ChunkSize) && i < LastRow) continue;
+				Device->WriteData(Device, Private->iRAM, optr - Private->iRAM);
+				optr = Private->iRAM;
+			}	
+		} else for (int i = FirstRow; i <= LastRow; i++) {
+			Device->WriteData( Device, Private->Shadowbuffer + (i * Device->Width + FirstCol) * 3, ChunkSize );
+		}	
+
+		FirstCol = (Device->Width * 3) / 2; LastCol = 0;
+		FirstRow = -1;
+	}	
+#else
+	// always update by full lines
+	SetColumnAddress( Device, Private->Offset.Width, Device->Width - 1);
+	
+	for (int r = 0; r < Device->Height; r += min(Private->PageSize, Device->Height - r)) {
+		int Height = min(Private->PageSize, Device->Height - r);
+		
+		SetRowAddress( Device, Private->Offset.Height + r, Private->Offset.Height + r + Height - 1 );
+		Device->WriteCommand(Device, ENABLE_WRITE);
+		
+		if (Private->iRAM) {
+			memcpy(Private->iRAM, Device->Framebuffer + r * Device->Width * 3, Height * Device->Width * 3 );
+			Device->WriteData( Device, Private->iRAM, Height * Device->Width * 3 );
+		} else	{
+			Device->WriteData( Device, Device->Framebuffer + r * Device->Width * 3, Height * Device->Width * 3 );
+		}	
+	}	
+#endif	
+}
+
+static void SetLayout( struct GDS_Device* Device, struct GDS_Layout *Layout ) { 
+	struct PrivateSpace *Private = (struct PrivateSpace*) Device->Private;
+	ESP_LOGI(TAG, "SetLayout 197 HFlip=%d VFlip=%d Rotate=%d (1=true)", Layout->HFlip, Layout->VFlip, Layout->Rotate);
+	//        D/CX RDX WRX D17-8 D7 D6 D5 D4 D3  D2 D1 D0 HEX
+	//Command   0   1   ↑    XX  0  0  1  1  0   1  1  0  36h
+	//Parameter 1   1   ↑    XX  MY MX MV ML BGR MH 0  0  00
+	//Orientation 0: MADCtl = 0x80  =   1000 0000 (MY=1)
+	if ((Device->Height)>(Device->Width)){		//Resolution = 320x240
+		Private->MADCtl = (1 << 7);				// 0x80 = default (no Rotation an no Flip)
+		if (Layout->HFlip) {							//Flip Horizontal
+			int a = Private->MADCtl;
+			Private->MADCtl = (a ^ (1 << 7));
+		}
+		if (Layout->Rotate) {							//Rotate 180 degr.
+			int a = Private->MADCtl;
+			a = (a ^ (1 << 7));
+			Private->MADCtl = (a ^ (1 << 6));
+		}
+		if (Layout->VFlip) {							//Flip Vertical
+			int a = Private->MADCtl;
+			Private->MADCtl = (a ^ (1 << 6));
+		}
+	} else {									//Resolution = 240x320
+		Private->MADCtl = (1 << 5);				// 0x20 = default (no Rotation an no Flip)
+		if (Layout->HFlip) {							//Flip Horizontal
+			int a = Private->MADCtl;
+			Private->MADCtl = (a ^ (1 << 6));
+		}
+		if (Layout->Rotate) {							//Rotate 180 degr.
+			int a = Private->MADCtl;
+			a = (a ^ (1 << 7));
+			Private->MADCtl = (a ^ (1 << 6));
+		}
+		if (Layout->VFlip) {							//Flip Vertical
+			int a = Private->MADCtl;
+			Private->MADCtl = (a ^ (1 << 7));
+		}
+	}
+	
+	Private->MADCtl = Layout->ColorSwap ? (Private->MADCtl | (1 << 3)) : (Private->MADCtl & ~(1 << 3));
+
+	ESP_LOGI(TAG, "SetLayout 255 Private->MADCtl=%hhu", Private->MADCtl);
+
+	Device->WriteCommand( Device, 0x36 );
+	WriteByte( Device, Private->MADCtl );
+	
+	Device->WriteCommand( Device, Layout->Invert ? 0x21 : 0x20 );
+	
+
+#ifdef SHADOW_BUFFER
+	// force a full refresh (almost ...)
+	memset(Private->Shadowbuffer, 0xAA, Device->FramebufferSize);
+#endif	
 }	
 
-static void DisplayOn( struct GDS_Device* Device ) { Device->WriteCommand( Device, L1_CMD_DISPLAY_ON ); }
-static void DisplayOff( struct GDS_Device* Device ) { Device->WriteCommand( Device, L1_CMD_DISPLAY_OFF ); }
+static void DisplayOn( struct GDS_Device* Device ) { Device->WriteCommand( Device, 0x29 ); }	//DISPON =0x29
+static void DisplayOff( struct GDS_Device* Device ) { Device->WriteCommand( Device, 0x28 ); }	//DISPOFF=0x28
 
 static void SetContrast( struct GDS_Device* Device, uint8_t Contrast ) {
-    Device->WriteCommand( Device, L1_CMD_WRITE_DISPLAY_BRIGHTNESS );
-    uint8_t loc_contrast =  (uint8_t)((float)Contrast/5.0f*  255.0f);
-    Device->WriteData( Device, &loc_contrast , 1 );
-    WriteDataByte(Device,0x00);
+	Device->WriteCommand( Device, 0x51 );
+	WriteByte( Device, Contrast );
+	
+	Device->SetContrast = NULL;
+	GDS_SetContrast( Device, Contrast );
+	Device->SetContrast = SetContrast;	// 0x00 value means the lowest brightness and 0xFF value means the highest brightness.
 }
 
 static bool Init( struct GDS_Device* Device ) {
 	struct PrivateSpace *Private = (struct PrivateSpace*) Device->Private;
+	int Depth = (Device->Depth + 8 - 1) / 8;
 	
+	Private->PageSize = min(8, PAGE_BLOCK / (Device->Width * Depth));
 
-//	Private->Offset = (480 - Device->Width) / 4 / 2;
-	
-	// find a page size that is not too small is an integer of height
-	Private->PageSize = min(8, PAGE_BLOCK / (Device->Width / 2));
-	Private->PageSize = Device->Height / (Device->Height / Private->PageSize) ;
-	
 #ifdef SHADOW_BUFFER	
-//	Private->Shadowbuffer = malloc( Device->FramebufferSize );
-//	memset(Private->Shadowbuffer, 0xFF, Device->FramebufferSize);
+	Private->Shadowbuffer = malloc( Device->FramebufferSize );	
+	memset(Private->Shadowbuffer, 0xFF, Device->FramebufferSize);
+#endif
+#ifdef USE_IRAM
+	Private->iRAM = heap_caps_malloc( (Private->PageSize + 1) * Device->Width * Depth, MALLOC_CAP_INTERNAL | MALLOC_CAP_DMA );
 #endif
-	Private->iRAM =NULL;
-	//Private->iRAM =heap_caps_malloc(320*PARALLEL_LINES*sizeof(uint16_t), MALLOC_CAP_DMA);
 
-
-	//ESP_LOGI(TAG, "ILI9341 with offset %u, page %u, iRAM %p", Private->Offset, Private->PageSize, Private->iRAM);
-	ESP_LOGI(TAG, "ILI9341 ");
-
-	// need to be off and disable display RAM
-	Device->DisplayOff( Device );
-	int cmd=0;
-	//Send all the commands
-	while (ili_init_cmds[cmd].databytes!=0xff) {
-		Device->WriteCommand( Device, ili_init_cmds[cmd].cmd );
-		Device->WriteData(Device,ili_init_cmds[cmd].data,ili_init_cmds[cmd].databytes&0x1F);
-		if (ili_init_cmds[cmd].databytes&0x80) {
-			vTaskDelay(100 / portTICK_RATE_MS);
-		}
-		cmd++;
-	}
+	ESP_LOGI(TAG, "ILI9341 with bit default-depth %u, page %u, iRAM %p", Device->Depth, Private->PageSize, Private->iRAM);
 	
+	// Sleepout + Booster
+	Device->WriteCommand( Device, 0x11 );
+			
+	// set flip modes & contrast
+	GDS_SetContrast( Device, 0x7f );
+	struct GDS_Layout Layout = { };
+	Device->SetLayout( Device, &Layout );
+	
+	// set screen depth (16/18) *** INTERFACE PIXEL FORMAT: 0x66=18 bit; 0x55=16 bit
+	Device->WriteCommand( Device, 0x3A );
+	if (Private->Model == ILI9341_24) WriteByte( Device, Device->Depth == 24 ? 0x66 : 0x55 );
+	else WriteByte( Device, Device->Depth == 24 ? 0x66 : 0x55 );
+
+	ESP_LOGI(TAG, "ILI9341_Init 312 device-depth %u, 0x66/0x55=0x%X", Device->Depth, Device->Depth == 24 ? 0x66 : 0x55);
+
+	// no Display Inversion (INVOFF=0x20 INVON=0x21)
+	Device->WriteCommand( Device, 0x20 );
+
+	//Gamma Correction: Enable next two line and enabel one of the Test0x Section... or build you own 15 Parameter...
+	Device->WriteCommand( Device, 0xF2 ); WriteByte( Device, 0x03 );	// 3Gamma Function: Disable = default (0x02), Enable (0x03)
+	Device->WriteCommand( Device, 0x26 ); WriteByte( Device, 0x01 );	// Gamma curve selected (0x01, 0x02, 0x04, 0x08) - A maximum of 4 fixed gamma curves can be selected
+	//Gamma Correction Test01
+	Device->WriteCommand( Device, 0xE0 );								// Positive Gamma Correction (15 Parameter)
+	WriteByte( Device, 0x0F ); WriteByte( Device, 0x31 ); WriteByte( Device, 0x2B ); WriteByte( Device, 0x0C ); WriteByte( Device, 0x0E );
+	WriteByte( Device, 0x08 ); WriteByte( Device, 0x4E ); WriteByte( Device, 0xF1 ); WriteByte( Device, 0x37 ); WriteByte( Device, 0x07 );
+	WriteByte( Device, 0x10 ); WriteByte( Device, 0x03 ); WriteByte( Device, 0x0E ); WriteByte( Device, 0x09 ); WriteByte( Device, 0x00 );
+	Device->WriteCommand( Device, 0xE1 ); 								// Negative Gamma Correction (15 Parameter)
+	WriteByte( Device, 0x00 ); WriteByte( Device, 0x0E ); WriteByte( Device, 0x14 ); WriteByte( Device, 0x03 ); WriteByte( Device, 0x11 );
+	WriteByte( Device, 0x07 ); WriteByte( Device, 0x31 ); WriteByte( Device, 0xC1 ); WriteByte( Device, 0x48 ); WriteByte( Device, 0x08 );
+	WriteByte( Device, 0x0F ); WriteByte( Device, 0x0C ); WriteByte( Device, 0x31 ); WriteByte( Device, 0x36 ); WriteByte( Device, 0x0F );
+	 
 	// gone with the wind
 	Device->DisplayOn( Device );
 	Device->Update( Device );
-	
+
 	return true;
 }	
 
-static const struct GDS_Device ILI9341 = {
-	.DisplayOn = DisplayOn, .DisplayOff = DisplayOff, .SetContrast = SetContrast,
+static const struct GDS_Device ILI9341_X = {
+	.DisplayOn = DisplayOn, .DisplayOff = DisplayOff,
 	.SetLayout = SetLayout,
-	.Update = Update, .Init = Init,
-};	
+	.Update = Update16, .Init = Init,
+	.Mode = GDS_RGB565, .Depth = 16,
+};		
 
 struct GDS_Device* ILI9341_Detect(char *Driver, struct GDS_Device* Device) {
-	if (!strcasestr(Driver, "ILI9341")) return NULL;
+	uint8_t Model;
+	int Depth=16;		// 16bit colordepth
+	
+	if (strcasestr(Driver, "ILI9341")) Model = ILI9341;
+	else if (strcasestr(Driver, "ILI9341_24")) Model = ILI9341_24;	//for future use...
+	else return NULL;
 		
 	if (!Device) Device = calloc(1, sizeof(struct GDS_Device));
-	
-	*Device = ILI9341;
-	Device->Depth = 4;
 		
+	*Device = ILI9341_X;	
+	sscanf(Driver, "%*[^:]:%u", &Depth);		// NVS-Parameter driver=ILI9341[:16|18]
+	struct PrivateSpace* Private = (struct PrivateSpace*) Device->Private;
+	Private->Model = Model;
+		ESP_LOGI(TAG, "ILI9341_Detect 391 Driver= %s   Depth=%d", Driver, Depth);
+
+	if (Depth == 18) {
+		Device->Mode = GDS_RGB888;
+		Device->Depth = 24;
+		Device->Update = Update24;
+	} 	
+	
+	if (Model == ILI9341_24) Device->SetContrast = SetContrast;
+
 	return Device;
-}
-
-
+}

components/display/SH1106.c

@@ -73,9 +73,10 @@ static void Update( struct GDS_Device* Device ) {
 #endif	
 }
 
-static void SetLayout( struct GDS_Device* Device, bool HFlip, bool VFlip, bool Rotate ) {
-	Device->WriteCommand( Device, HFlip ? 0xA1 : 0xA0 );
-	Device->WriteCommand( Device, VFlip ? 0xC8 : 0xC0 );
+static void SetLayout( struct GDS_Device* Device, struct GDS_Layout *Layout ) { 
+	Device->WriteCommand( Device, Layout->HFlip ? 0xA1 : 0xA0 );
+	Device->WriteCommand( Device, Layout->VFlip ? 0xC8 : 0xC0 );
+    Device->WriteCommand( Device, Layout->Invert ? 0xA7 : 0xA6 );
 }	
 
 static void DisplayOn( struct GDS_Device* Device ) { Device->WriteCommand( Device, 0xAF ); }
@@ -86,6 +87,10 @@ static void SetContrast( struct GDS_Device* Device, uint8_t Contrast ) {
     Device->WriteCommand( Device, Contrast );
 }
 
+static void SPIParams(int Speed, uint8_t *mode, uint16_t *CS_pre, uint8_t *CS_post) {
+	*CS_post = Speed / (8*1000*1000);
+}
+
 static bool Init( struct GDS_Device* Device ) {
 #ifdef SHADOW_BUFFER	
 	struct PrivateSpace *Private = (struct PrivateSpace*) Device->Private;
@@ -120,7 +125,8 @@ static bool Init( struct GDS_Device* Device ) {
     Device->WriteCommand( Device, 0x40 + 0x00 );
 	Device->SetContrast( Device, 0x7F );
 	// set flip modes
-	Device->SetLayout( Device, false, false, false );
+	struct GDS_Layout Layout = { };
+	Device->SetLayout( Device, &Layout );
 	// no Display Inversion
     Device->WriteCommand( Device, 0xA6 );
 	// set Clocks
@@ -140,6 +146,7 @@ static const struct GDS_Device SH1106 = {
 	.SetLayout = SetLayout,
 	.Update = Update, .Init = Init,
 	.Depth = 1,
+	.SPIParams = SPIParams,
 #if !defined SHADOW_BUFFER && defined USE_IRAM	
 	.Alloc = GDS_ALLOC_IRAM_SPI;
 #endif		

components/display/SSD1306.c

@@ -85,9 +85,10 @@ static void Update( struct GDS_Device* Device ) {
 #endif	
 }
 
-static void SetLayout( struct GDS_Device* Device, bool HFlip, bool VFlip, bool Rotate ) { 
-	Device->WriteCommand( Device, HFlip ? 0xA1 : 0xA0 );
-	Device->WriteCommand( Device, VFlip ? 0xC8 : 0xC0 );
+static void SetLayout( struct GDS_Device* Device, struct GDS_Layout *Layout ) { 
+	Device->WriteCommand( Device, Layout->HFlip ? 0xA1 : 0xA0 );
+	Device->WriteCommand( Device, Layout->VFlip ? 0xC8 : 0xC0 );
+	Device->WriteCommand( Device, Layout->Invert ? 0xA7 : 0xA6 );
 }
 	
 static void DisplayOn( struct GDS_Device* Device ) { Device->WriteCommand( Device, 0xAF ); }
@@ -132,7 +133,8 @@ static bool Init( struct GDS_Device* Device ) {
     Device->WriteCommand( Device, 0x40 + 0x00 );
 	Device->SetContrast( Device, 0x7F );
 	// set flip modes
-	Device->SetLayout( Device, false, false, false);
+	struct GDS_Layout Layout = { };
+	Device->SetLayout( Device, &Layout );
 	// no Display Inversion
     Device->WriteCommand( Device, 0xA6 );
 	// set Clocks

components/display/SSD1322.c

@@ -96,13 +96,14 @@ static void Update( struct GDS_Device* Device ) {
 #endif	
 }
 
-static void SetLayout( struct GDS_Device* Device, bool HFlip, bool VFlip, bool Rotate ) { 
+static void SetLayout( struct GDS_Device* Device, struct GDS_Layout *Layout ) { 
 	struct PrivateSpace *Private = (struct PrivateSpace*) Device->Private;
-	Private->ReMap = HFlip ? (Private->ReMap & ~(1 << 1)) : (Private->ReMap | (1 << 1));
-	Private->ReMap = VFlip ? (Private->ReMap | (1 << 4)) : (Private->ReMap & ~(1 << 4));
+	Private->ReMap = Layout->HFlip ? (Private->ReMap & ~(1 << 1)) : (Private->ReMap | (1 << 1));
+	Private->ReMap = Layout->VFlip ? (Private->ReMap | (1 << 4)) : (Private->ReMap & ~(1 << 4));
 	Device->WriteCommand( Device, 0xA0 );
 	Device->WriteData( Device, &Private->ReMap, 1 );
 	WriteDataByte( Device, 0x11 );		
+	Device->WriteCommand( Device, Layout->Invert ? 0xA7 : 0xA6 );
 }	
 
 static void DisplayOn( struct GDS_Device* Device ) { Device->WriteCommand( Device, 0xAF ); }
@@ -145,25 +146,31 @@ static bool Init( struct GDS_Device* Device ) {
 	
 	// set flip modes
 	Private->ReMap = 0;
-	Device->SetLayout( Device, false, false, false);
+	struct GDS_Layout Layout = { };
+	Device->SetLayout( Device, &Layout );
+	
+	// set Display Enhancement
+    Device->WriteCommand( Device, 0xB4 );
+	WriteDataByte( Device, 0xA0 );
+	WriteDataByte( Device, 0xB5 );
 	
 	// set Clocks
     Device->WriteCommand( Device, 0xB3 );
-	WriteDataByte( Device, 0x91 );
+	WriteDataByte( Device, 0xB2 ); // 0x91 seems to be common but is too slow for 5.5'
 	
 	// set MUX
 	Device->WriteCommand( Device, 0xCA );
 	WriteDataByte( Device, Device->Height - 1 );
 	
-	// phase 1 & 2 period (needed?)	
+	// phase 1 & 2 period
 	Device->WriteCommand( Device, 0xB1 );
-	WriteDataByte( Device, 0xE2 );
+	WriteDataByte( Device, 0xE3 );	// 0xE2 was recommended
 	
-	// set pre-charge V (needed?°)
+	// set pre-charge V 
 	Device->WriteCommand( Device, 0xBB );
-	WriteDataByte( Device, 0x1F );
+	WriteDataByte( Device, 0x0F); // 0x1F causes column interferences 
 	
-	// set COM deselect voltage (needed?)
+	// set COM deselect voltage
 	Device->WriteCommand( Device, 0xBE );
 	WriteDataByte( Device, 0x07 );
 	

components/display/SSD132x.c

@@ -222,17 +222,18 @@ static void DrawBitmapCBR(struct GDS_Device* Device, uint8_t *Data, int Width, i
 	}
 }
 
-static void SetLayout( struct GDS_Device* Device, bool HFlip, bool VFlip, bool Rotate ) { 
+static void SetLayout( struct GDS_Device* Device, struct GDS_Layout *Layout ) { 
 	struct PrivateSpace *Private = (struct PrivateSpace*) Device->Private;
 	if (Private->Model == SSD1326) {
-		Private->ReMap = HFlip ? (Private->ReMap | ((1 << 0) | (1 << 2))) : (Private->ReMap & ~((1 << 0) | (1 << 2)));
-		Private->ReMap = HFlip ? (Private->ReMap | (1 << 1)) : (Private->ReMap & ~(1 << 1));		
+		Private->ReMap = Layout->HFlip ? (Private->ReMap | ((1 << 0) | (1 << 2))) : (Private->ReMap & ~((1 << 0) | (1 << 2)));
+		Private->ReMap = Layout->HFlip ? (Private->ReMap | (1 << 1)) : (Private->ReMap & ~(1 << 1));		
 	} else {
-		Private->ReMap = VFlip ? (Private->ReMap | ((1 << 0) | (1 << 1))) : (Private->ReMap & ~((1 << 0) | (1 << 1)));
-		Private->ReMap = VFlip ? (Private->ReMap | (1 << 4)) : (Private->ReMap & ~(1 << 4));
+		Private->ReMap = Layout->VFlip ? (Private->ReMap | ((1 << 0) | (1 << 1))) : (Private->ReMap & ~((1 << 0) | (1 << 1)));
+		Private->ReMap = Layout->VFlip ? (Private->ReMap | (1 << 4)) : (Private->ReMap & ~(1 << 4));
 	}	
 	Device->WriteCommand( Device, 0xA0 );
 	Device->WriteCommand( Device, Private->ReMap );
+    Device->WriteCommand( Device, Layout->Invert ? 0xA7 : 0xA6 );	
 }	
 
 static void DisplayOn( struct GDS_Device* Device ) { Device->WriteCommand( Device, 0xAF ); }
@@ -288,7 +289,8 @@ static bool Init( struct GDS_Device* Device ) {
 	Device->WriteCommand( Device, 0x00 );
 	Device->SetContrast( Device, 0x7F );
 	// set flip modes
-	Device->SetLayout( Device, false, false, false );
+	struct GDS_Layout Layout = { };
+	Device->SetLayout( Device, &Layout );
 	// no Display Inversion
     Device->WriteCommand( Device, 0xA6 );
 	// set Clocks

components/display/SSD1351.c

@@ -181,12 +181,13 @@ static void Update24( struct GDS_Device* Device ) {
 #endif	
 }
 
-static void SetLayout( struct GDS_Device* Device, bool HFlip, bool VFlip, bool Rotate ) { 
+static void SetLayout( struct GDS_Device* Device, struct GDS_Layout *Layout ) { 
 	struct PrivateSpace *Private = (struct PrivateSpace*) Device->Private;
-	Private->ReMap = HFlip ? (Private->ReMap & ~(1 << 1)) : (Private->ReMap | (1 << 1));
-	Private->ReMap = VFlip ? (Private->ReMap | (1 << 4)) : (Private->ReMap & ~(1 << 4));
+	Private->ReMap = Layout->HFlip ? (Private->ReMap & ~(1 << 1)) : (Private->ReMap | (1 << 1));
+	Private->ReMap = Layout->VFlip ? (Private->ReMap | (1 << 4)) : (Private->ReMap & ~(1 << 4));
 	Device->WriteCommand( Device, 0xA0 );
 	WriteByte( Device, Private->ReMap );
+    Device->WriteCommand( Device, Layout->Invert ? 0xA7 : 0xA6 );	
 }	
 
 static void DisplayOn( struct GDS_Device* Device ) { Device->WriteCommand( Device, 0xAF ); }
@@ -239,7 +240,8 @@ static bool Init( struct GDS_Device* Device ) {
 	
 	// set flip modes & contrast
 	Device->SetContrast( Device, 0x7F );
-	Device->SetLayout( Device, false, false, false );
+	struct GDS_Layout Layout = { };
+	Device->SetLayout( Device, &Layout );
 	
 	// set Adressing Mode Horizontal
 	Private->ReMap |= (0 << 2);

components/display/SSD1675.c

@@ -243,7 +243,7 @@ struct GDS_Device* SSD1675_Detect(char *Driver, struct GDS_Device* Device) {
 	char *p;
 	struct PrivateSpace* Private = (struct PrivateSpace*) Device->Private;
 	Private->ReadyPin = -1;
-	if ((p = strcasestr(Driver, "ready")) != NULL) Private->ReadyPin = atoi(strchr(p, '=') + 1);
+	if ((p = strcasestr(Driver, "ready")) && (p = strchr(p, '='))) Private->ReadyPin = atoi(p + 1);
 	
 	ESP_LOGI(TAG, "SSD1675 driver with ready GPIO %d", Private->ReadyPin);
 	

components/display/ST77xx.c

@@ -188,20 +188,24 @@ static void Update24( struct GDS_Device* Device ) {
 #endif	
 }
 
-static void SetLayout( struct GDS_Device* Device, bool HFlip, bool VFlip, bool Rotate ) { 
+static void SetLayout( struct GDS_Device* Device, struct GDS_Layout *Layout ) { 
 	struct PrivateSpace *Private = (struct PrivateSpace*) Device->Private;
 	
-	Private->MADCtl = HFlip ? (Private->MADCtl | (1 << 7)) : (Private->MADCtl & ~(1 << 7));
-	Private->MADCtl = VFlip ? (Private->MADCtl | (1 << 6)) : (Private->MADCtl & ~(1 << 6));
-	Private->MADCtl = Rotate ? (Private->MADCtl | (1 << 5)) : (Private->MADCtl & ~(1 << 5));
-	
+	if (Private->Model == ST7789) {
+		if (Layout->Rotate) Private->Offset.Width += Layout->HFlip ? 320 - Device->Width : 0;
+		else Private->Offset.Height += Layout->HFlip ? 320 - Device->Height : 0;
+		Device->WriteCommand( Device, Layout->Invert ? 0x20 : 0x21 );			
+	} else {
+		Device->WriteCommand( Device, Layout->Invert ? 0x21 : 0x20 );	
+	}		
+
+	Private->MADCtl = Layout->HFlip ? (Private->MADCtl | (1 << 7)) : (Private->MADCtl & ~(1 << 7));
+	Private->MADCtl = Layout->VFlip ? (Private->MADCtl | (1 << 6)) : (Private->MADCtl & ~(1 << 6));
+	Private->MADCtl = Layout->Rotate ? (Private->MADCtl | (1 << 5)) : (Private->MADCtl & ~(1 << 5));
+	Private->MADCtl = Layout->ColorSwap ? (Private->MADCtl & ~(1 << 3)) : (Private->MADCtl | (1 << 3));	
+
 	Device->WriteCommand( Device, 0x36 );
 	WriteByte( Device, Private->MADCtl );
-	
-	if (Private->Model == ST7789) {
-		if (Rotate) Private->Offset.Width = HFlip ? 320 - Device->Width : 0;
-		else Private->Offset.Height = HFlip ? 320 - Device->Height : 0;
-	}
 
 #ifdef SHADOW_BUFFER
 	// force a full refresh (almost ...)
@@ -235,28 +239,26 @@ static bool Init( struct GDS_Device* Device ) {
 	Private->iRAM = heap_caps_malloc( (Private->PageSize + 1) * Device->Width * Depth, MALLOC_CAP_INTERNAL | MALLOC_CAP_DMA );
 #endif
 
-	ESP_LOGI(TAG, "ST77xx with bit depth %u, page %u, iRAM %p", Device->Depth, Private->PageSize, Private->iRAM);
+	ESP_LOGI(TAG, "ST77xx with bit depth %u, offsets %hu:%hu, page %u, iRAM %p", Device->Depth, Private->Offset.Height, Private->Offset.Width, Private->PageSize, Private->iRAM);
 	
 	// Sleepout + Booster
 	Device->WriteCommand( Device, 0x11 );
-		
+	
 	// need BGR & Address Mode
 	Private->MADCtl = 1 << 3;
 	Device->WriteCommand( Device, 0x36 );
 	WriteByte( Device, Private->MADCtl );		
-		
+	
 	// set flip modes & contrast
 	GDS_SetContrast( Device, 0x7f );
-	Device->SetLayout( Device, false, false, false );
+	struct GDS_Layout Layout = { };
+	Device->SetLayout( Device, &Layout );
 	
 	// set screen depth (16/18)
 	Device->WriteCommand( Device, 0x3A );
 	if (Private->Model == ST7789) WriteByte( Device, Device->Depth == 24 ? 0x066 : 0x55 );
 	else WriteByte( Device, Device->Depth == 24 ? 0x06 : 0x05 );
 	
-	// no Display Inversion
-    Device->WriteCommand( Device, Private->Model == ST7735 ? 0x20 : 0x21 );	
-		
 	// gone with the wind
 	Device->DisplayOn( Device );
 	Device->Update( Device );
@@ -283,8 +285,14 @@ struct GDS_Device* ST77xx_Detect(char *Driver, struct GDS_Device* Device) {
 		
 	*Device = ST77xx;	
 	sscanf(Driver, "%*[^:]:%u", &Depth);
+
 	struct PrivateSpace* Private = (struct PrivateSpace*) Device->Private;
 	Private->Model = Model;
+
+	if (Model == ST7735) {
+		sscanf(Driver, "%*[^:]%*[^x]%*[^=]=%hu", &Private->Offset.Height);		
+		sscanf(Driver, "%*[^:]%*[^y]%*[^=]=%hu", &Private->Offset.Width);		
+	}
 	
 	if (Depth == 18) {
 		Device->Mode = GDS_RGB666;

components/display/core/gds.c

@@ -235,12 +235,13 @@ void GDS_SetContrast( struct GDS_Device* Device, uint8_t Contrast ) {
 		ledc_update_duty( LEDC_HIGH_SPEED_MODE, Device->Backlight.Channel );		
 	}
 }
-	
-void GDS_SetLayout( struct GDS_Device* Device, bool HFlip, bool VFlip, bool Rotate ) { if (Device->SetLayout) Device->SetLayout( Device, HFlip, VFlip, Rotate ); }
+
+void GDS_SetLayout( struct GDS_Device* Device, struct GDS_Layout *Layout ) { if (Device->SetLayout) Device->SetLayout( Device, Layout ); }
 void GDS_SetDirty( struct GDS_Device* Device ) { Device->Dirty = true; }
-int	GDS_GetWidth( struct GDS_Device* Device ) { return Device->Width; }
-int	GDS_GetHeight( struct GDS_Device* Device ) { return Device->Height; }
-int	GDS_GetDepth( struct GDS_Device* Device ) { return Device->Depth; }
-int	GDS_GetMode( struct GDS_Device* Device ) { return Device->Mode; }
+int	 GDS_GetWidth( struct GDS_Device* Device ) { return Device ? Device->Width : 0; }
+void GDS_SetTextWidth( struct GDS_Device* Device, int TextWidth ) { Device->TextWidth = Device && TextWidth && TextWidth < Device->Width ? TextWidth : Device->Width; }
+int	 GDS_GetHeight( struct GDS_Device* Device ) { return Device ? Device->Height : 0; }
+int	 GDS_GetDepth( struct GDS_Device* Device ) { return Device ? Device->Depth : 0; }
+int	 GDS_GetMode( struct GDS_Device* Device ) { return Device ? Device->Mode : 0; }
 void GDS_DisplayOn( struct GDS_Device* Device ) { if (Device->DisplayOn) Device->DisplayOn( Device ); }
 void GDS_DisplayOff( struct GDS_Device* Device ) { if (Device->DisplayOff) Device->DisplayOff( Device ); }

components/display/core/gds.h

@@ -26,6 +26,12 @@ struct GDS_BacklightPWM {
 	int Channel, Timer, Max;
 	bool Init;
 };
+struct GDS_Layout {
+	bool HFlip, VFlip;
+	bool Rotate;
+	bool Invert; 
+	bool ColorSwap;
+};
 
 typedef struct GDS_Device* GDS_DetectFunc(char *Driver, struct GDS_Device *Device);
 
@@ -35,9 +41,10 @@ void 	GDS_SetContrast( struct GDS_Device* Device, uint8_t Contrast );
 void 	GDS_DisplayOn( struct GDS_Device* Device );
 void 	GDS_DisplayOff( struct GDS_Device* Device ); 
 void 	GDS_Update( struct GDS_Device* Device );
-void 	GDS_SetLayout( struct GDS_Device* Device, bool HFlip, bool VFlip, bool Rotate );
+void 	GDS_SetLayout( struct GDS_Device* Device, struct GDS_Layout* Layout);
 void 	GDS_SetDirty( struct GDS_Device* Device );
 int 	GDS_GetWidth( struct GDS_Device* Device );
+void 	GDS_SetTextWidth( struct GDS_Device* Device, int TextWidth );
 int 	GDS_GetHeight( struct GDS_Device* Device );
 int 	GDS_GetDepth( struct GDS_Device* Device );
 int 	GDS_GetMode( struct GDS_Device* Device );

components/display/core/gds_default_if.h

@@ -11,7 +11,7 @@ bool GDS_I2CInit( int PortNumber, int SDA, int SCL, int speed );
 bool GDS_I2CAttachDevice( struct GDS_Device* Device, int Width, int Height, int I2CAddress, int RSTPin, int BacklightPin );
 
 bool GDS_SPIInit( int SPI, int DC );
-bool GDS_SPIAttachDevice( struct GDS_Device* Device, int Width, int Height, int CSPin, int RSTPin, int Speed, int BacklightPin );
+bool GDS_SPIAttachDevice( struct GDS_Device* Device, int Width, int Height, int CSPin, int RSTPin, int Speed, int BacklightPin, int Mode );
 
 #ifdef __cplusplus
 }

components/display/core/gds_font.c

@@ -73,13 +73,13 @@ void GDS_FontDrawChar( struct GDS_Device* Device, char Character, int x, int y,
         CharStartY+= OffsetY;
 
         /* Do not attempt to draw if this character is entirely offscreen */
-        if ( CharEndX < 0 || CharStartX >= Device->Width || CharEndY < 0 || CharStartY >= Device->Height ) {
+        if ( CharEndX < 0 || CharStartX >= Device->TextWidth || CharEndY < 0 || CharStartY >= Device->Height ) {
             ClipDebug( x, y );
             return;
         }
 
         /* Do not attempt to draw past the end of the screen */
-        CharEndX = ( CharEndX >= Device->Width ) ? Device->Width - 1 : CharEndX;
+        CharEndX = ( CharEndX >= Device->TextWidth ) ? Device->TextWidth - 1 : CharEndX;
         CharEndY = ( CharEndY >= Device->Height ) ? Device->Height - 1 : CharEndY;
 		Device->Dirty = true;
 
@@ -98,12 +98,14 @@ void GDS_FontDrawChar( struct GDS_Device* Device, char Character, int x, int y,
     }
 }
 
-bool GDS_SetFont( struct GDS_Device* Display, const struct GDS_FontDef* Font ) {
+const struct GDS_FontDef* GDS_SetFont( struct GDS_Device* Display, const struct GDS_FontDef* Font ) {
+	const struct GDS_FontDef* OldFont = Display->Font;
+
     Display->FontForceProportional = false;
     Display->FontForceMonospace = false;
     Display->Font = Font;
 
-    return true;
+    return OldFont;
 }
 
 void GDS_FontForceProportional( struct GDS_Device* Display, bool Force ) {
@@ -144,7 +146,7 @@ int GDS_FontGetCharWidth( struct GDS_Device* Display, char Character ) {
 }
 
 int GDS_FontGetMaxCharsPerRow( struct GDS_Device* Display ) {
-    return Display->Width / Display->Font->Width;
+    return Display->TextWidth / Display->Font->Width;
 }
 
 int GDS_FontGetMaxCharsPerColumn( struct GDS_Device* Display ) {
@@ -208,7 +210,7 @@ void GDS_FontGetAnchoredStringCoords( struct GDS_Device* Display, int* OutX, int
     switch ( Anchor ) {
         case TextAnchor_East: {
             *OutY = ( Display->Height / 2 ) - ( StringHeight / 2 );
-            *OutX = ( Display->Width - StringWidth );
+            *OutX = ( Display->TextWidth - StringWidth );
 
             break;
         }
@@ -219,19 +221,19 @@ void GDS_FontGetAnchoredStringCoords( struct GDS_Device* Display, int* OutX, int
             break;
         }
         case TextAnchor_North: {
-            *OutX = ( Display->Width / 2 ) - ( StringWidth / 2 );
+            *OutX = ( Display->TextWidth / 2 ) - ( StringWidth / 2 );
             *OutY = 0;
 
             break;
         }
         case TextAnchor_South: {
-            *OutX = ( Display->Width / 2 ) - ( StringWidth / 2 );
+            *OutX = ( Display->TextWidth / 2 ) - ( StringWidth / 2 );
             *OutY = ( Display->Height - StringHeight );
             
             break;
         }
         case TextAnchor_NorthEast: {
-            *OutX = ( Display->Width - StringWidth );
+            *OutX = ( Display->TextWidth - StringWidth );
             *OutY = 0;
 
             break;
@@ -244,7 +246,7 @@ void GDS_FontGetAnchoredStringCoords( struct GDS_Device* Display, int* OutX, int
         }
         case TextAnchor_SouthEast: {
             *OutY = ( Display->Height - StringHeight );
-            *OutX = ( Display->Width - StringWidth );
+            *OutX = ( Display->TextWidth - StringWidth );
 
             break;
         }
@@ -256,7 +258,7 @@ void GDS_FontGetAnchoredStringCoords( struct GDS_Device* Display, int* OutX, int
         }
         case TextAnchor_Center: {
             *OutY = ( Display->Height / 2 ) - ( StringHeight / 2 );
-            *OutX = ( Display->Width / 2 ) - ( StringWidth / 2 );
+            *OutX = ( Display->TextWidth / 2 ) - ( StringWidth / 2 );
 
             break;
         }

components/display/core/gds_font.h

@@ -46,7 +46,7 @@ typedef enum {
     TextAnchor_Center
 } TextAnchor;
 
-bool GDS_SetFont( struct GDS_Device* Display, const struct GDS_FontDef* Font );
+const struct GDS_FontDef* GDS_SetFont( struct GDS_Device* Display, const struct GDS_FontDef* Font );
 
 void GDS_FontForceProportional( struct GDS_Device* Display, bool Force );
 void GDS_FontForceMonospace( struct GDS_Device* Display, bool Force );
@@ -59,7 +59,8 @@ int GDS_FontGetMaxCharsPerColumn( struct GDS_Device* Display );
 
 int GDS_FontGetCharWidth( struct GDS_Device* Display, char Character );
 int GDS_FontGetCharHeight( struct GDS_Device* Display );
-int GDS_FontMeasureString( struct GDS_Device* Display, const char* Text );\
+int GDS_FontMeasureString( struct GDS_Device* Display, const char* Text );
+int GDS_FontMeasureStringLine( struct GDS_Device* Display, int Line, const char* Text );
 
 void GDS_FontDrawChar( struct GDS_Device* Display, char Character, int x, int y, int Color );
 void GDS_FontDrawString( struct GDS_Device* Display, int x, int y, const char* Text, int Color );

components/display/core/gds_image.c

@@ -8,14 +8,18 @@
  
 #include <string.h>
 #include "math.h"
+#ifdef TJPGD_ROM
 #include "esp32/rom/tjpgd.h"
+#else
+#include "tjpgd.h"
+#endif
 #include "esp_log.h"
 
 #include "gds.h"
 #include "gds_private.h"
 #include "gds_image.h"
 
-const char TAG[] = "ImageDec";
+const static char TAG[] = "ImageDec";
 
 #define SCRATCH_SIZE	3100
 
@@ -167,7 +171,7 @@ static unsigned OutHandlerDirect(JDEC *Decoder, void *Bitmap, JRECT *Frame) {
 static void* DecodeJPEG(uint8_t *Source, int *Width, int *Height, float Scale, bool SizeOnly, int RGB_Mode) {
     JDEC Decoder;
     JpegCtx Context;
-	char *Scratch = calloc(SCRATCH_SIZE, 1);
+	char *Scratch = malloc(SCRATCH_SIZE);
 	
     if (!Scratch) {
         ESP_LOGE(TAG, "Cannot allocate workspace");
@@ -372,7 +376,7 @@ bool GDS_DrawJPEG(struct GDS_Device* Device, uint8_t *Source, int x, int y, int
     JDEC Decoder;
     JpegCtx Context;
 	bool Ret = false;
-	char *Scratch = calloc(SCRATCH_SIZE, 1);
+	char *Scratch = malloc(SCRATCH_SIZE);
 	
     if (!Scratch) {
         ESP_LOGE(TAG, "Cannot allocate workspace");

components/display/core/gds_private.h

@@ -95,7 +95,7 @@ struct GDS_Device {
 		const struct GDS_FontDef* Font;
 	} Lines[MAX_LINES];
 	
-	uint16_t Width;
+	uint16_t Width, TextWidth;
     uint16_t Height;
 	uint8_t Depth, Mode;
 	
@@ -117,13 +117,15 @@ struct GDS_Device {
 	void (*SetContrast)( struct GDS_Device* Device, uint8_t Contrast );
 	void (*DisplayOn)( struct GDS_Device* Device );
 	void (*DisplayOff)( struct GDS_Device* Device );
-	void (*SetLayout)( struct GDS_Device* Device, bool HFlip, bool VFlip, bool Rotate );
+	void (*SetLayout)( struct GDS_Device* Device, struct GDS_Layout *Layout );
 	// must provide for depth other than 1 (vertical) and 4 (may provide for optimization)
 	void (*DrawPixelFast)( struct GDS_Device* Device, int X, int Y, int Color );
 	void (*DrawBitmapCBR)(struct GDS_Device* Device, uint8_t *Data, int Width, int Height, int Color );
 	// may provide for optimization
 	void (*DrawRGB)( struct GDS_Device* Device, uint8_t *Image,int x, int y, int Width, int Height, int RGB_Mode );
 	void (*ClearWindow)( struct GDS_Device* Device, int x1, int y1, int x2, int y2, int Color );
+	// may provide for tweaking
+	void (*SPIParams)(int Speed, uint8_t *mode, uint16_t *CS_pre, uint8_t *CS_post);
 		    
 	// interface-specific methods	
     WriteCommandProc WriteCommand;

components/display/core/gds_text.c

@@ -26,19 +26,20 @@ static char TAG[] = "gds";
  */
 static const struct GDS_FontDef *GuessFont( struct GDS_Device *Device, int FontType) {
 	switch(FontType) {
+	case GDS_FONT_DEFAULT:	
+		return Device->Font;		
 	case GDS_FONT_LINE_1:	
 		return &Font_line_1;
 	case GDS_FONT_LINE_2:	
 		return &Font_line_2;
-	case GDS_FONT_SMALL:	
-		return &Font_droid_sans_fallback_11x13;	
 	case GDS_FONT_MEDIUM:			
-	default:		
-		return &Font_droid_sans_fallback_15x17;	
+		//return &Font_droid_sans_fallback_15x17;	
+	case GDS_FONT_SMALL:	
+	default:
+		return &Font_droid_sans_fallback_11x13;	
 #ifdef USE_LARGE_FONTS
 	case GDS_FONT_LARGE:	
 		return &Font_droid_sans_fallback_24x28;
-		break;		
 	case GDS_FONT_SEGMENT:			
 		if (Device->Height == 32) return &Font_Tarable7Seg_16x32;
 		else return &Font_Tarable7Seg_32x64;
@@ -46,8 +47,8 @@ static const struct GDS_FontDef *GuessFont( struct GDS_Device *Device, int FontT
 	case GDS_FONT_LARGE:	
 	case GDS_FONT_SEGMENT:			
 		ESP_LOGW(TAG, "large fonts disabled");
-		return &Font_droid_sans_fallback_15x17;
-		break;		
+		//return &Font_droid_sans_fallback_15x17;
+		return &Font_droid_sans_fallback_11x13;	
 #endif	
 	}
 }
@@ -99,13 +100,13 @@ bool GDS_TextLine(struct GDS_Device* Device, int N, int Pos, int Attr, char *Tex
 	Width = GDS_FontMeasureString( Device, Text );
 	
 	// adjusting position, erase only EoL for rigth-justified
-	if (Pos == GDS_TEXT_RIGHT) X = Device->Width - Width - 1;
-	else if (Pos == GDS_TEXT_CENTER) X = (Device->Width - Width) / 2;
+	if (Pos == GDS_TEXT_RIGHT) X = Device->TextWidth - Width - 1;
+	else if (Pos == GDS_TEXT_CENTER) X = (Device->TextWidth - Width) / 2;
 	
 	// erase if requested
 	if (Attr & GDS_TEXT_CLEAR) {
 		int Y_min = max(0, Device->Lines[N].Y), Y_max = max(0, Device->Lines[N].Y + Device->Lines[N].Font->Height);
-		for (int c = (Attr & GDS_TEXT_CLEAR_EOL) ? X : 0; c < Device->Width; c++) 
+		for (int c = (Attr & GDS_TEXT_CLEAR_EOL) ? X : 0; c < Device->TextWidth; c++) 
 			for (int y = Y_min; y < Y_max; y++)
 				DrawPixelFast( Device, c, y, GDS_COLOR_BLACK );
 	}
@@ -118,7 +119,20 @@ bool GDS_TextLine(struct GDS_Device* Device, int N, int Pos, int Attr, char *Tex
 	Device->Dirty = true;
 	if (Attr & GDS_TEXT_UPDATE) GDS_Update( Device );
 		
-	return Width + X < Device->Width;
+	return Width + X < Device->TextWidth;
+}
+
+/****************************************************************************************
+ * 
+ */
+int GDS_GetTextWidth(struct GDS_Device* Device, int N, int Attr, char *Text) {
+	const struct GDS_FontDef *Font = GDS_SetFont( Device, Device->Lines[N-1].Font );	
+
+	if (Attr & GDS_TEXT_MONOSPACE) GDS_FontForceMonospace( Device, true );
+	int Width = GDS_FontMeasureString( Device, Text );
+	GDS_SetFont( Device, Font );
+
+	return Width;
 }
 
 /****************************************************************************************
@@ -132,7 +146,7 @@ int GDS_TextStretch(struct GDS_Device* Device, int N, char *String, int Max) {
 	
 	// we might already fit
 	GDS_SetFont( Device, Device->Lines[N].Font );	
-	if (GDS_FontMeasureString( Device, String ) <= Device->Width) return 0;
+	if (GDS_FontMeasureString( Device, String ) <= Device->TextWidth) return 0;
 		
 	// add some space for better visual 
 	strncat(String, Space, Max-Len);
@@ -143,7 +157,7 @@ int GDS_TextStretch(struct GDS_Device* Device, int N, char *String, int Max) {
 	Boundary = GDS_FontMeasureString( Device, String );
 			
 	// add a full display width	
-	while (Len < Max && GDS_FontMeasureString( Device, String ) - Boundary < Device->Width) {
+	while (Len < Max && GDS_FontMeasureString( Device, String ) - Boundary < Device->TextWidth) {
 		String[Len++] = String[Extra++];
 		String[Len] = '\0';
 	}

components/display/core/gds_text.h

@@ -31,5 +31,6 @@ struct GDS_Device;
 bool 	GDS_TextSetFontAuto(struct GDS_Device* Device, int N, int FontType, int Space);
 bool 	GDS_TextSetFont(struct GDS_Device* Device, int N, const struct GDS_FontDef *Font, int Space);
 bool 	GDS_TextLine(struct GDS_Device* Device, int N, int Pos, int Attr, char *Text);
+int		GDS_GetTextWidth(struct GDS_Device* Device, int N, int Attr, char *Text);
 int 	GDS_TextStretch(struct GDS_Device* Device, int N, char *String, int Max);
 void 	GDS_TextPos(struct GDS_Device* Device, int FontType, int Where, int Attr, char *Text, ...);

components/display/core/ifaces/default_if_i2c.c

@@ -35,7 +35,7 @@ static bool I2CDefaultWriteData( struct GDS_Device* Device, const uint8_t* Data,
 bool GDS_I2CInit( int PortNumber, int SDA, int SCL, int Speed ) {
 	I2CPortNumber = PortNumber;
 	
-	I2CWait = pdMS_TO_TICKS( Speed ? Speed / 4000 : 100 );
+	I2CWait = pdMS_TO_TICKS( Speed ? (250 * 250000) / Speed : 250 );
 	
 	if (SDA != -1 && SCL != -1) {
 		i2c_config_t Config = { 0 };
@@ -75,7 +75,7 @@ bool GDS_I2CAttachDevice( struct GDS_Device* Device, int Width, int Height, int
     Device->RSTPin = RSTPin;
 	Device->Backlight.Pin = BacklightPin;	
 	Device->IF = GDS_IF_I2C;
-	Device->Width = Width;
+	Device->Width = Device->TextWidth = Width;
 	Device->Height = Height;
 	
 	if ( RSTPin >= 0 ) {

components/display/core/ifaces/default_if_spi.c

@@ -34,8 +34,8 @@ bool GDS_SPIInit( int SPI, int DC ) {
     return true;
 }
 
-bool GDS_SPIAttachDevice( struct GDS_Device* Device, int Width, int Height, int CSPin, int RSTPin, int BackLightPin, int Speed ) {
-    spi_device_interface_config_t SPIDeviceConfig;
+bool GDS_SPIAttachDevice( struct GDS_Device* Device, int Width, int Height, int CSPin, int RSTPin, int BackLightPin, int Speed, int Mode ) {
+    spi_device_interface_config_t SPIDeviceConfig = { };
     spi_device_handle_t SPIDevice;
 
     NullCheck( Device, return false );
@@ -43,15 +43,16 @@ bool GDS_SPIAttachDevice( struct GDS_Device* Device, int Width, int Height, int
 	if (CSPin >= 0) {
 		ESP_ERROR_CHECK_NONFATAL( gpio_set_direction( CSPin, GPIO_MODE_OUTPUT ), return false );
 		ESP_ERROR_CHECK_NONFATAL( gpio_set_level( CSPin, 0 ), return false );
-	}	
+	}
 	
-    memset( &SPIDeviceConfig, 0, sizeof( spi_device_interface_config_t ) );
-
     SPIDeviceConfig.clock_speed_hz = Speed > 0 ? Speed : SPI_MASTER_FREQ_8M;
     SPIDeviceConfig.spics_io_num = CSPin;
     SPIDeviceConfig.queue_size = 1;
+    SPIDeviceConfig.mode = Mode;
 	SPIDeviceConfig.flags = SPI_DEVICE_NO_DUMMY;
-
+	if (Device->SPIParams) Device->SPIParams(SPIDeviceConfig.clock_speed_hz, &SPIDeviceConfig.mode, 
+											 &SPIDeviceConfig.cs_ena_pretrans, &SPIDeviceConfig.cs_ena_posttrans);
+	
     ESP_ERROR_CHECK_NONFATAL( spi_bus_add_device( SPIHost, &SPIDeviceConfig, &SPIDevice ), return false );
 	
 	Device->WriteCommand = SPIDefaultWriteCommand;
@@ -61,7 +62,7 @@ bool GDS_SPIAttachDevice( struct GDS_Device* Device, int Width, int Height, int
     Device->CSPin = CSPin;
 	Device->Backlight.Pin = BackLightPin;	
 	Device->IF = GDS_IF_SPI;
-	Device->Width = Width;
+	Device->Width = Device->TextWidth = Width;
 	Device->Height = Height;
 	
 	if ( RSTPin >= 0 ) {
@@ -74,16 +75,23 @@ bool GDS_SPIAttachDevice( struct GDS_Device* Device, int Width, int Height, int
 }
 
 static bool SPIDefaultWriteBytes( spi_device_handle_t SPIHandle, int WriteMode, const uint8_t* Data, size_t DataLength ) {
-    spi_transaction_t SPITransaction = { 0 };
+    spi_transaction_t SPITransaction = { };
 
     NullCheck( SPIHandle, return false );
     NullCheck( Data, return false );
 
     if ( DataLength > 0 ) {
 		gpio_set_level( DCPin, WriteMode );
-		
+
 		SPITransaction.length = DataLength * 8;
-		SPITransaction.tx_buffer = Data;
+		
+		if (DataLength <= 4) {
+			SPITransaction.flags = SPI_TRANS_USE_TXDATA;
+			SPITransaction.tx_data[0] = *Data++; SPITransaction.tx_data[1] = *Data++;
+			SPITransaction.tx_data[2] = *Data++; SPITransaction.tx_data[3] = *Data;
+		} else {
+			SPITransaction.tx_buffer = Data;
+		}	
             
 		// only do polling as we don't have contention on SPI (otherwise DMA for transfers > 16 bytes)		
 		ESP_ERROR_CHECK_NONFATAL( spi_device_polling_transmit(SPIHandle, &SPITransaction), return false );

components/display/display.c

@@ -20,6 +20,7 @@
 #include "gds_draw.h"
 #include "gds_text.h"
 #include "gds_font.h"
+#include "gds_image.h"
 
 static const char *TAG = "display";
 
@@ -30,6 +31,9 @@ static const char *TAG = "display";
 #define SCROLLABLE_SIZE			384
 #define HEADER_SIZE				64
 #define	DEFAULT_SLEEP			3600
+#define ARTWORK_BORDER			1
+
+extern const uint8_t default_artwork[]   asm("_binary_note_jpg_start");
 
 static EXT_RAM_ATTR struct {
 	TaskHandle_t task;
@@ -41,7 +45,19 @@ static EXT_RAM_ATTR struct {
 	int offset, boundary;
 	char *metadata_config;
 	bool timer, refresh;
-	uint32_t elapsed, duration;
+	uint32_t elapsed;
+	struct {
+		uint32_t value;
+		char string[8]; // H:MM:SS
+		bool visible;
+	} duration;
+	struct {
+		bool enable, active;
+		bool fit;
+		bool updated;
+		int tick;
+		int offset;
+	}  artwork;
 	TickType_t tick;
 } displayer;
 
@@ -52,15 +68,16 @@ static const char *known_drivers[] = {"SH1106",
 		"SSD1327",
 		"SSD1675",
 		"SSD1351",
-		"ST77xx",
+		"ST7735",
+		"ST7789",
 		"ILI9341",
 		NULL
 	};
 static void displayer_task(void *args);
 
 struct GDS_Device *display;   
-extern GDS_DetectFunc SSD1306_Detect, SSD132x_Detect, SH1106_Detect, SSD1675_Detect, SSD1322_Detect, SSD1351_Detect, ST77xx_Detect;
-GDS_DetectFunc *drivers[] = { SH1106_Detect, SSD1306_Detect, SSD132x_Detect, SSD1675_Detect, SSD1322_Detect, SSD1351_Detect, ST77xx_Detect, NULL };
+extern GDS_DetectFunc SSD1306_Detect, SSD132x_Detect, SH1106_Detect, SSD1675_Detect, SSD1322_Detect, SSD1351_Detect, ST77xx_Detect, ILI9341_Detect;
+GDS_DetectFunc *drivers[] = { SH1106_Detect, SSD1306_Detect, SSD132x_Detect, SSD1675_Detect, SSD1322_Detect, SSD1351_Detect, ST77xx_Detect, ILI9341_Detect, NULL };
 
 /****************************************************************************************
  * 
@@ -70,11 +87,11 @@ void display_init(char *welcome) {
 	char *config = config_alloc_get_str("display_config", CONFIG_DISPLAY_CONFIG, "N/A");
 	
 	int width = -1, height = -1, backlight_pin = -1;
-	char *p, *drivername = strstr(config, "driver");
+	char *drivername = strstr(config, "driver");
 
-	if ((p = strcasestr(config, "width")) != NULL) width = atoi(strchr(p, '=') + 1);
-	if ((p = strcasestr(config, "height")) != NULL) height = atoi(strchr(p, '=') + 1);
-	if ((p = strcasestr(config, "back")) != NULL) backlight_pin = atoi(strchr(p, '=') + 1);	
+	PARSE_PARAM(config, "width", '=', width);
+	PARSE_PARAM(config, "height", '=', height);
+	PARSE_PARAM(config, "back", '=', backlight_pin);
 		
 	// query drivers to see if we have a match
 	ESP_LOGI(TAG, "Trying to configure display with %s", config);
@@ -83,33 +100,34 @@ void display_init(char *welcome) {
 		display = GDS_AutoDetect(drivername, drivers, &PWMConfig);
 	} else {
 		display = GDS_AutoDetect(drivername, drivers, NULL);
-	}	
-		
+	}
+
 	// so far so good
 	if (display && width > 0 && height > 0) {
 		int RST_pin = -1;
-		if ((p = strcasestr(config, "reset")) != NULL) RST_pin = atoi(strchr(p, '=') + 1);
+		PARSE_PARAM(config, "reset", '=', RST_pin);
 		
 		// Detect driver interface
-		if (strstr(config, "I2C") && i2c_system_port != -1) {
+		if (strcasestr(config, "I2C") && i2c_system_port != -1) {
 			int address = 0x3C;
 				
-			if ((p = strcasestr(config, "address")) != NULL) address = atoi(strchr(p, '=') + 1);
+			PARSE_PARAM(config, "address", '=', address);
 				
 			init = true;
 			GDS_I2CInit( i2c_system_port, -1, -1, i2c_system_speed ) ;
 			GDS_I2CAttachDevice( display, width, height, address, RST_pin, backlight_pin );
 		
 			ESP_LOGI(TAG, "Display is I2C on port %u", address);
-		} else if (strstr(config, "SPI") && spi_system_host != -1) {
-			int CS_pin = -1, speed = 0;
+		} else if (strcasestr(config, "SPI") && spi_system_host != -1) {
+			int CS_pin = -1, speed = 0, mode = 0;
 		
-			if ((p = strcasestr(config, "cs")) != NULL) CS_pin = atoi(strchr(p, '=') + 1);
-			if ((p = strcasestr(config, "speed")) != NULL) speed = atoi(strchr(p, '=') + 1);
+			PARSE_PARAM(config, "cs", '=', CS_pin);
+			PARSE_PARAM(config, "speed", '=', speed);
+			PARSE_PARAM(config, "mode", '=', mode);
 		
 			init = true;
 			GDS_SPIInit( spi_system_host, spi_system_dc_gpio );
-			GDS_SPIAttachDevice( display, width, height, CS_pin, RST_pin, backlight_pin, speed );
+			GDS_SPIAttachDevice( display, width, height, CS_pin, RST_pin, backlight_pin, speed, mode );
 				
 			ESP_LOGI(TAG, "Display is SPI host %u with cs:%d", spi_system_host, CS_pin);
 		} else {
@@ -124,23 +142,38 @@ void display_init(char *welcome) {
 	if (init) {
 		static DRAM_ATTR StaticTask_t xTaskBuffer __attribute__ ((aligned (4)));
 		static EXT_RAM_ATTR StackType_t xStack[DISPLAYER_STACK_SIZE] __attribute__ ((aligned (4)));
-		
-		GDS_SetLayout( display, strcasestr(config, "HFlip"), strcasestr(config, "VFlip"), strcasestr(config, "rotate"));
-		GDS_SetFont(display, &Font_droid_sans_fallback_15x17 );
-		GDS_TextPos(display, GDS_FONT_MEDIUM, GDS_TEXT_CENTERED, GDS_TEXT_CLEAR | GDS_TEXT_UPDATE, welcome);
+		struct GDS_Layout Layout = {
+			.HFlip = strcasestr(config, "HFlip"), 
+			.VFlip = strcasestr(config, "VFlip"), 
+			.Rotate = strcasestr(config, "rotate"), 
+			.Invert = strcasestr(config, "invert"),
+			.ColorSwap = strcasestr(config, "cswap"),
+		};	
+
+		GDS_SetLayout(display, &Layout);
+		GDS_SetFont(display, &Font_line_2);
+		GDS_TextPos(display, GDS_FONT_DEFAULT, GDS_TEXT_CENTERED, GDS_TEXT_CLEAR | GDS_TEXT_UPDATE, welcome);
 
 		// start the task that will handle scrolling & counting
 		displayer.mutex = xSemaphoreCreateMutex();
 		displayer.by = 2;
 		displayer.pause = 3600;
 		displayer.speed = 33;
-		displayer.task = xTaskCreateStatic( (TaskFunction_t) displayer_task, "displayer_thread", DISPLAYER_STACK_SIZE, NULL, ESP_TASK_PRIO_MIN + 1, xStack, &xTaskBuffer);
+		displayer.task = xTaskCreateStatic( (TaskFunction_t) displayer_task, "common_displayer", DISPLAYER_STACK_SIZE, NULL, ESP_TASK_PRIO_MIN + 1, xStack, &xTaskBuffer);
 		
 		// set lines for "fixed" text mode
 		GDS_TextSetFontAuto(display, 1, GDS_FONT_LINE_1, -3);
 		GDS_TextSetFontAuto(display, 2, GDS_FONT_LINE_2, -3);
 		
 		displayer.metadata_config = config_alloc_get(NVS_TYPE_STR, "metadata_config");
+		
+		// leave room for artwork is display is horizontal-style
+		if (strcasestr(displayer.metadata_config, "artwork")) {
+			displayer.artwork.enable = true;
+			displayer.artwork.fit = true;
+			if (height <= 64 && width > height * 2) displayer.artwork.offset = width - height - ARTWORK_BORDER;
+			PARSE_PARAM(displayer.metadata_config, "artwork", ':', displayer.artwork.fit);
+		}	
 	}
 	
 	free(config);
@@ -192,18 +225,39 @@ static void displayer_task(void *args) {
 		
 		// handler elapsed track time
 		if (displayer.timer && displayer.state == DISPLAYER_ACTIVE) {
-			char counter[16];
+			char line[19] = "-", *_line = line + 1; // [-]H:MM:SS / H:MM:SS
 			TickType_t tick = xTaskGetTickCount();
 			uint32_t elapsed = (tick - displayer.tick) * portTICK_PERIOD_MS;
-			
+
 			if (elapsed >= 1000) {
 				xSemaphoreTake(displayer.mutex, portMAX_DELAY);
 				displayer.tick = tick;
-				displayer.elapsed += elapsed / 1000;
-				xSemaphoreGive(displayer.mutex);				
-				if (displayer.elapsed < 3600) snprintf(counter, 16, "%5u:%02u", displayer.elapsed / 60, displayer.elapsed % 60);
-				else snprintf(counter, 16, "%2u:%02u:%02u", displayer.elapsed / 3600, (displayer.elapsed % 3600) / 60, displayer.elapsed % 60);
-				GDS_TextLine(display, 1, GDS_TEXT_RIGHT, (GDS_TEXT_CLEAR | GDS_TEXT_CLEAR_EOL) | GDS_TEXT_UPDATE, counter);
+				elapsed = displayer.elapsed += elapsed / 1000;
+				xSemaphoreGive(displayer.mutex);
+
+				// when we have duration but no space, display remaining time
+				if (displayer.duration.value && !displayer.duration.visible) elapsed = displayer.duration.value - elapsed;
+
+				if (elapsed < 3600) sprintf(_line, "%u:%02u", elapsed / 60, elapsed % 60);
+				else sprintf(_line, "%u:%02u:%02u", (elapsed / 3600) % 100, (elapsed % 3600) / 60, elapsed % 60);
+
+				// concatenate if we have room for elapsed / duration
+				if (displayer.duration.visible) {
+					strcat(_line, "/");
+					strcat(_line, displayer.duration.string);
+				} else if (displayer.duration.value) {
+					_line--;
+				}
+
+				// just re-write the whole line it's easier
+				GDS_TextLine(display, 1, GDS_TEXT_LEFT, GDS_TEXT_CLEAR, displayer.header);	
+				GDS_TextLine(display, 1, GDS_TEXT_RIGHT, GDS_TEXT_UPDATE, _line);
+				
+				// if we have not received artwork after 5s, display a default icon
+				if (displayer.artwork.active && !displayer.artwork.updated && tick - displayer.artwork.tick > pdMS_TO_TICKS(5000)) {
+					ESP_LOGI(TAG, "no artwork received, setting default");
+					displayer_artwork((uint8_t*) default_artwork);
+				}	
 				timer_sleep = 1000;
 			} else timer_sleep = max(1000 - elapsed, 0);	
 		} else timer_sleep = DEFAULT_SLEEP;
@@ -216,6 +270,32 @@ static void displayer_task(void *args) {
 	}
 }	
 
+/****************************************************************************************
+ * 
+ */
+void displayer_artwork(uint8_t *data) {
+	if (!displayer.artwork.active) return;
+	
+	int x = displayer.artwork.offset ? displayer.artwork.offset + ARTWORK_BORDER : 0;
+	int y = x ? 0 : 32;
+	GDS_ClearWindow(display, x, y, -1, -1, GDS_COLOR_BLACK);
+	if (data) {
+		displayer.artwork.updated = true;
+		GDS_DrawJPEG(display, data, x, y, GDS_IMAGE_CENTER | (displayer.artwork.fit ? GDS_IMAGE_FIT : 0));
+	} else {
+		displayer.artwork.updated = false;
+		displayer.artwork.tick = xTaskGetTickCount();
+	}	
+	
+}
+
+/****************************************************************************************
+ * 
+ */
+bool displayer_can_artwork(void) {
+	return displayer.artwork.active;
+}
+
 /****************************************************************************************
  * 
  */
@@ -278,8 +358,8 @@ void displayer_metadata(char *artist, char *album, char *title) {
 	}
 	
 	// get optional scroll speed & pause
-	if ((p = strcasestr(displayer.metadata_config, "speed")) != NULL) sscanf(p, "%*[^=]=%d", &displayer.speed);
-	if ((p = strcasestr(displayer.metadata_config, "pause")) != NULL) sscanf(p, "%*[^=]=%d", &displayer.pause);
+	PARSE_PARAM(displayer.metadata_config, "speed", '=', displayer.speed);
+	PARSE_PARAM(displayer.metadata_config, "pause", '=', displayer.pause);
 	
 	displayer.offset = 0;	
 	utf8_decode(displayer.string);
@@ -317,9 +397,27 @@ void displayer_timer(enum displayer_time_e mode, int elapsed, int duration) {
 	
 	xSemaphoreTake(displayer.mutex, portMAX_DELAY);
 
-	if (elapsed >= 0) displayer.elapsed = elapsed / 1000;	
-	if (duration >= 0) displayer.duration = duration / 1000;
 	if (displayer.timer) displayer.tick = xTaskGetTickCount();
+	if (elapsed >= 0) displayer.elapsed = elapsed / 1000;	
+	if (duration > 0) {
+		displayer.duration.visible = true;
+		displayer.duration.value = duration / 1000;
+
+		if (displayer.duration.value > 3600) sprintf(displayer.duration.string, "%u:%02u:%02u", (displayer.duration.value / 3600) % 10,
+													(displayer.duration.value % 3600) / 60, displayer.duration.value % 60);
+		else sprintf(displayer.duration.string, "%u:%02u", displayer.duration.value / 60, displayer.duration.value % 60);
+
+		char *buf;
+		asprintf(&buf, "%s %s/%s", displayer.header, displayer.duration.string, displayer.duration.string);
+		if (GDS_GetTextWidth(display, 1, 0, buf) > GDS_GetWidth(display)) {
+			ESP_LOGW(TAG, "Can't fit duration %s (%d) on screen using elapsed only", buf, GDS_GetTextWidth(display, 1, 0, buf));
+			displayer.duration.visible = false;
+		}
+		free(buf);
+	} else if (!duration) {
+		displayer.duration.visible = false;
+		displayer.duration.value = 0;
+	}
 		
 	xSemaphoreGive(displayer.mutex);
 }	
@@ -338,25 +436,31 @@ void displayer_control(enum displayer_cmd_e cmd, ...) {
 	switch(cmd) {
 	case DISPLAYER_ACTIVATE: {	
 		char *header = va_arg(args, char*);
+		displayer.artwork.active = displayer.artwork.enable && va_arg(args, int);
 		strncpy(displayer.header, header, HEADER_SIZE);
 		displayer.header[HEADER_SIZE] = '\0';
 		displayer.state = DISPLAYER_ACTIVE;
 		displayer.timer = false;
 		displayer.refresh = true;
 		displayer.string[0] = '\0';
-		displayer.elapsed = displayer.duration = 0;
+		displayer.elapsed = displayer.duration.value = 0;
+		displayer.duration.visible = false;
 		displayer.offset = displayer.boundary = 0;
 		display_bus(&displayer, DISPLAY_BUS_TAKE);
+		if (displayer.artwork.active) GDS_SetTextWidth(display, displayer.artwork.offset);
 		vTaskResume(displayer.task);
 		break;
 	}	
 	case DISPLAYER_SUSPEND:		
 		// task will display the line 2 from beginning and suspend
 		displayer.state = DISPLAYER_IDLE;
+		displayer_artwork(NULL);
 		display_bus(&displayer, DISPLAY_BUS_GIVE);
 		break;		
 	case DISPLAYER_SHUTDOWN:
 		// let the task self-suspend (we might be doing i2c_write)
+		GDS_SetTextWidth(display, 0);
+		displayer_artwork(NULL);
 		displayer.state = DISPLAYER_DOWN;
 		display_bus(&displayer, DISPLAY_BUS_GIVE);
 		break;
@@ -381,14 +485,14 @@ void displayer_control(enum displayer_cmd_e cmd, ...) {
 /****************************************************************************************
  *
  */
-bool display_is_valid_driver(char * driver){
+bool display_is_valid_driver(const char * driver){
 	return display_conf_get_driver_name(driver)!=NULL;
 }
 
 /****************************************************************************************
  *
  */
-const char *display_conf_get_driver_name(char * driver){
+const char *display_conf_get_driver_name(const char * driver){
 	for(uint8_t i=0;known_drivers[i]!=NULL && strlen(known_drivers[i])>0;i++ ){
 		if(strcasestr(driver,known_drivers[i])){
 			return known_drivers[i];
@@ -396,3 +500,27 @@ const char *display_conf_get_driver_name(char * driver){
 	}
 	return NULL;
 }
+
+/****************************************************************************************
+ *
+ */
+char * display_get_supported_drivers(void){
+	int total_size = 1;
+	char * supported_drivers=NULL;
+	const char * separator = "|";
+	int separator_len = strlen(separator);
+
+	for(uint8_t i=0;known_drivers[i]!=NULL && strlen(known_drivers[i])>0;i++ ){
+		total_size += strlen(known_drivers[i])+separator_len;
+	}
+	total_size+=2;
+	supported_drivers = malloc(total_size);
+	memset(supported_drivers,0x00,total_size);
+	strcat(supported_drivers,"<");
+	for(uint8_t i=0;known_drivers[i]!=NULL && strlen(known_drivers[i])>0;i++ ){
+		supported_drivers = strcat(supported_drivers,known_drivers[i]);
+		supported_drivers = strcat(supported_drivers,separator);
+	}
+	strcat(supported_drivers,">");
+	return supported_drivers;
+}

components/display/display.h

@@ -10,6 +10,7 @@
 
 #include "gds.h"
 
+
 /* 
  The displayer is not thread-safe and the caller must ensure use its own 
  mutexes if it wants something better. Especially, text() line() and draw()
@@ -31,10 +32,13 @@ enum displayer_time_e 	{ DISPLAYER_ELAPSED, DISPLAYER_REMAINING };
 
 enum display_bus_cmd_e { DISPLAY_BUS_TAKE, DISPLAY_BUS_GIVE };
 bool (*display_bus)(void *from, enum display_bus_cmd_e cmd);
-const char *display_conf_get_driver_name(char * driver);
-bool display_is_valid_driver(char * driver);
+const char *display_conf_get_driver_name(const char * driver);
+bool display_is_valid_driver(const char * driver);
 
 void displayer_scroll(char *string, int speed, int pause);
 void displayer_control(enum displayer_cmd_e cmd, ...);
 void displayer_metadata(char *artist, char *album, char *title);
+void displayer_artwork(uint8_t *data);
 void displayer_timer(enum displayer_time_e mode, int elapsed, int duration);
+bool displayer_can_artwork(void);
+char * display_get_supported_drivers(void);

components/display/fonts/font_line_1.c

@@ -26,7 +26,7 @@ static const uint8_t Square721_BT11x14[] = {
         0x02, 0x00, 0x00, 0x00, 0x0C, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,  // Code for char ,
         0x04, 0x00, 0x00, 0x80, 0x00, 0x80, 0x00, 0x80, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,  // Code for char -
         0x02, 0x00, 0x00, 0x00, 0x04, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,  // Code for char .
-        0x04, 0x00, 0x0C, 0x80, 0x03, 0x70, 0x00, 0x08, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,  // Code for char /
+        0x06, 0x00, 0x00, 0x00, 0x00, 0x00, 0x0C, 0x80, 0x03, 0x70, 0x00, 0x08, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,  // Code for char /
         0x08, 0x00, 0x00, 0x00, 0x00, 0xF0, 0x03, 0x08, 0x04, 0x08, 0x04, 0x08, 0x04, 0x08, 0x04, 0xF0, 0x03, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,  // Code for char 0
         0x08, 0x00, 0x00, 0x00, 0x00, 0x20, 0x00, 0x10, 0x04, 0x08, 0x04, 0xF8, 0x07, 0x00, 0x04, 0x00, 0x04, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,  // Code for char 1
         0x08, 0x00, 0x00, 0x00, 0x00, 0x30, 0x07, 0x08, 0x05, 0x88, 0x04, 0x88, 0x04, 0x88, 0x04, 0x70, 0x04, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,  // Code for char 2

components/driver_bt/bt_app_core.c

@@ -6,27 +6,27 @@
    CONDITIONS OF ANY KIND, either express or implied.
 */
 
-#include "bt_app_core.h"
-
 #include <stdint.h>
-#include "esp_system.h"
 #include <string.h>
 #include <stdbool.h>
 #include "esp_log.h"
-#include "freertos/xtensa_api.h"
-#include "freertos/FreeRTOSConfig.h"
 #include "freertos/FreeRTOS.h"
 #include "freertos/queue.h"
 #include "freertos/task.h"
+#include "esp_bt.h"
+#include "esp_bt_main.h"
+#include "esp_gap_bt_api.h"
+#include "bt_app_core.h"
+#include "tools.h"
 
-static const char * TAG = "btappcore";
+static const char *TAG = "btappcore";
 
 static void bt_app_task_handler(void *arg);
 static bool bt_app_send_msg(bt_app_msg_t *msg);
 static void bt_app_work_dispatched(bt_app_msg_t *msg);
 
-static xQueueHandle s_bt_app_task_queue = NULL;
-static xTaskHandle s_bt_app_task_handle = NULL;
+static xQueueHandle s_bt_app_task_queue;
+static bool running;
 
 bool bt_app_work_dispatch(bt_app_cb_t p_cback, uint16_t event, void *p_params, int param_len, bt_app_copy_cb_t p_copy_cback)
 {
@@ -42,8 +42,7 @@ bool bt_app_work_dispatch(bt_app_cb_t p_cback, uint16_t event, void *p_params, i
     if (param_len == 0) {
         return bt_app_send_msg(&msg);
     } else if (p_params && param_len > 0) {
-        if ((msg.param = malloc(param_len)) != NULL) {
-            memcpy(msg.param, p_params, param_len);
+        if ((msg.param = clone_obj_psram(p_params, param_len)) != NULL) {
             /* check if caller has provided a copy callback to do the deep copy */
             if (p_copy_cback) {
                 p_copy_cback(&msg, msg.param, p_params);
@@ -78,9 +77,58 @@ static void bt_app_work_dispatched(bt_app_msg_t *msg)
 static void bt_app_task_handler(void *arg)
 {
     bt_app_msg_t msg;
-    for (;;) {
+	esp_err_t err;
+	
+	s_bt_app_task_queue = xQueueCreate(10, sizeof(bt_app_msg_t));
+	
+	esp_bt_controller_mem_release(ESP_BT_MODE_BLE);
+    esp_bt_controller_config_t bt_cfg = BT_CONTROLLER_INIT_CONFIG_DEFAULT();
+	if (esp_bt_controller_get_status() == ESP_BT_CONTROLLER_STATUS_IDLE ) {
+        if ((err = esp_bt_controller_init(&bt_cfg)) != ESP_OK) {
+            ESP_LOGE(TAG, "%s initialize controller failed: %s\n", __func__, esp_err_to_name(err));
+            goto exit;
+        }
+
+        if ((err = esp_bt_controller_enable(ESP_BT_MODE_CLASSIC_BT)) != ESP_OK) {
+            ESP_LOGE(TAG, "%s enable controller failed: %s\n", __func__, esp_err_to_name(err));
+            goto exit;
+        }
+
+        if ((err = esp_bluedroid_init()) != ESP_OK) {
+            ESP_LOGE(TAG, "%s initialize bluedroid failed: %s\n", __func__, esp_err_to_name(err));
+            goto exit;
+        }
+
+        if ((err = esp_bluedroid_enable()) != ESP_OK) {
+            ESP_LOGE(TAG, "%s enable bluedroid failed: %s\n", __func__, esp_err_to_name(err));
+            goto exit;
+        }
+    }
+	
+	/* Bluetooth device name, connection mode and profile set up */
+	bt_app_work_dispatch((bt_av_hdl_stack_evt_t*) arg, BT_APP_EVT_STACK_UP, NULL, 0, NULL);
+	
+#if (CONFIG_BT_SSP_ENABLED)
+    /* Set default parameters for Secure Simple Pairing */
+    esp_bt_sp_param_t param_type = ESP_BT_SP_IOCAP_MODE;
+    esp_bt_io_cap_t iocap = ESP_BT_IO_CAP_IO;
+    esp_bt_gap_set_security_param(param_type, &iocap, sizeof(uint8_t));
+#endif
+	
+    /*
+     * Set default parameters for Legacy Pairing
+     * Use variable pin, input pin code when pairing
+     */
+    esp_bt_pin_type_t pin_type = ESP_BT_PIN_TYPE_VARIABLE;
+    esp_bt_pin_code_t pin_code;
+    esp_bt_gap_set_pin(pin_type, 0, pin_code);
+	
+	running = true;
+	
+	while (running) {
         if (pdTRUE == xQueueReceive(s_bt_app_task_queue, &msg, (portTickType)portMAX_DELAY)) {
         	ESP_LOGV(TAG,"%s, sig 0x%x, 0x%x", __func__, msg.sig, msg.event);
+			
             switch (msg.sig) {
             case BT_APP_SIG_WORK_DISPATCH:
                 bt_app_work_dispatched(&msg);
@@ -88,36 +136,46 @@ static void bt_app_task_handler(void *arg)
             default:
                 ESP_LOGW(TAG,"%s, unhandled sig: %d", __func__, msg.sig);
                 break;
-            } // switch (msg.sig)
+            }
 
             if (msg.param) {
                 free(msg.param);
             }
-        }
-        else
-        {
+        } else {
         	ESP_LOGW(TAG,"No messaged received from queue.");
         }
     }
+	
+	ESP_LOGD(TAG, "bt_app_task shutting down");
+	
+	if (esp_bluedroid_disable() != ESP_OK) goto exit;
+	// this disable has a sleep timer BTA_DISABLE_DELAY in bt_target.h and 
+	// if we don't wait for it then disable crashes... don't know why
+	vTaskDelay(2*200 / portTICK_PERIOD_MS);	
+	
+    ESP_LOGD(TAG, "esp_bluedroid_disable called successfully");
+    if (esp_bluedroid_deinit() != ESP_OK) goto exit;
+	
+    ESP_LOGD(TAG, "esp_bluedroid_deinit called successfully");
+    if (esp_bt_controller_disable() != ESP_OK) goto exit;
+	
+    ESP_LOGD(TAG, "esp_bt_controller_disable called successfully");
+    if (esp_bt_controller_deinit() != ESP_OK) goto exit;
+	
+	ESP_LOGD(TAG, "bt stopped successfully");	
+
+exit:
+	vQueueDelete(s_bt_app_task_queue);
+	running = false;		
+    vTaskDelete(NULL);
 }
 
-void bt_app_task_start_up(void)
+void bt_app_task_start_up(bt_av_hdl_stack_evt_t* handler)
 {
-
-    s_bt_app_task_queue = xQueueCreate(10, sizeof(bt_app_msg_t));
-    assert(s_bt_app_task_queue!=NULL);
-    assert(xTaskCreate(bt_app_task_handler, "BtAppT", 4096, NULL, configMAX_PRIORITIES - 3, &s_bt_app_task_handle)==pdPASS);
-    return;
+    xTaskCreate(bt_app_task_handler, "BtAppT", 4096, handler, configMAX_PRIORITIES - 3, NULL);
 }
 
 void bt_app_task_shut_down(void)
 {
-    if (s_bt_app_task_handle) {
-        vTaskDelete(s_bt_app_task_handle);
-        s_bt_app_task_handle = NULL;
-    }
-    if (s_bt_app_task_queue) {
-        vQueueDelete(s_bt_app_task_queue);
-        s_bt_app_task_queue = NULL;
-    }
+	running = false;
 }

components/driver_bt/bt_app_core.h

@@ -15,9 +15,33 @@
 #include <stdio.h>
 
 #define BT_APP_CORE_TAG                   "BT_APP_CORE"
-
 #define BT_APP_SIG_WORK_DISPATCH          (0x01)
 
+enum {
+    BT_APP_EVT_STACK_UP = 0,
+};
+
+/* A2DP global state */
+enum {
+    APP_AV_STATE_IDLE,
+    APP_AV_STATE_DISCOVERING,
+    APP_AV_STATE_DISCOVERED,
+    APP_AV_STATE_UNCONNECTED,
+    APP_AV_STATE_CONNECTING,
+    APP_AV_STATE_CONNECTED,
+    APP_AV_STATE_DISCONNECTING,
+};
+
+/* sub states of APP_AV_STATE_CONNECTED */
+enum {
+    APP_AV_MEDIA_STATE_IDLE,
+    APP_AV_MEDIA_STATE_STARTING,
+    APP_AV_MEDIA_STATE_STARTED,
+    APP_AV_MEDIA_STATE_STOPPING,
+	APP_AV_MEDIA_STATE_WAIT_DISCONNECT
+};
+extern int bt_app_source_get_a2d_state();
+extern int bt_app_source_get_media_state();
 /**
  * @brief     handler for the dispatched work
  */
@@ -36,12 +60,17 @@ typedef struct {
  */
 typedef void (* bt_app_copy_cb_t) (bt_app_msg_t *msg, void *p_dest, void *p_src);
 
+/**
+ * @brief     callback for startup event
+ */
+typedef void bt_av_hdl_stack_evt_t(uint16_t event, void *p_param);
+
 /**
  * @brief     work dispatcher for the application task
  */
 bool bt_app_work_dispatch(bt_app_cb_t p_cback, uint16_t event, void *p_params, int param_len, bt_app_copy_cb_t p_copy_cback);
 
-void bt_app_task_start_up(void);
+void bt_app_task_start_up(bt_av_hdl_stack_evt_t* handler);
 
 void bt_app_task_shut_down(void);
 

components/driver_bt/bt_app_sink.c

@@ -25,7 +25,7 @@
 #include "platform_config.h"
 #include "freertos/FreeRTOS.h"
 #include "freertos/task.h"
-#include "trace.h"
+#include "tools.h"
 #include "audio_controls.h"
 #include "sys/lock.h"
 #include "display.h"
@@ -37,19 +37,15 @@
 #define APP_RC_CT_TL_RN_PLAYBACK_CHANGE  (3)
 #define APP_RC_CT_TL_RN_PLAY_POS_CHANGE  (4)
 
-#define BT_AV_TAG               "BT_AV"
-#define BT_RC_TG_TAG            "RCTG"
-#define BT_RC_CT_TAG            "RCCT"
+static const char BT_AV_TAG[] = "BT_AV";
+static const char BT_RC_TG_TAG[] = "RCTG";
+static const char BT_RC_CT_TAG[] = "RCCT";
 
 #ifndef CONFIG_BT_NAME
 #define CONFIG_BT_NAME	"ESP32-BT"
 #endif
 
-/* event for handler "bt_av_hdl_stack_up */
-enum {
-    BT_APP_EVT_STACK_UP = 0,
-};
-char * bt_name = NULL;
+static char * bt_name = NULL;
 
 static bool (*bt_app_a2d_cmd_cb)(bt_sink_cmd_t cmd, ...);
 static void (*bt_app_a2d_data_cb)(const uint8_t *data, uint32_t len);
@@ -62,14 +58,15 @@ static void bt_av_hdl_a2d_evt(uint16_t event, void *p_param);
 static void bt_av_hdl_avrc_ct_evt(uint16_t event, void *p_param);
 /* avrc TG event handler */
 static void bt_av_hdl_avrc_tg_evt(uint16_t event, void *p_param);
-static void volume_set_by_local_host(uint8_t volume);
+static void volume_set_by_local_host(int value, bool is_step);
+static void bt_av_notify_evt_handler(uint8_t event_id, esp_avrc_rn_param_t *event_parameter);
 
 static const char *s_a2d_conn_state_str[] = {"Disconnected", "Connecting", "Connected", "Disconnecting"};
 static const char *s_a2d_audio_state_str[] = {"Suspended", "Stopped", "Started"};
 static esp_avrc_rn_evt_cap_mask_t s_avrc_peer_rn_cap;
 
 static _lock_t s_volume_lock;
-static uint8_t s_volume = 0;
+static int s_volume, abs_volume, sink_volume;
 static bool s_volume_notify;
 static enum { AUDIO_IDLE, AUDIO_CONNECTED, AUDIO_PLAYING } s_audio = AUDIO_IDLE;
 
@@ -89,16 +86,14 @@ static EXT_RAM_ATTR struct {
 
 static void bt_volume_up(bool pressed) {
 	if (!pressed) return;
-	// volume UP/DOWN buttons are not supported by iPhone/Android
-	volume_set_by_local_host(s_volume < 127-3 ? s_volume + 3 : 127);
+	volume_set_by_local_host(+3, true);
 	(*bt_app_a2d_cmd_cb)(BT_SINK_VOLUME, s_volume);
 	ESP_LOGD(BT_AV_TAG, "BT volume up %u", s_volume);
 }
 
 static void bt_volume_down(bool pressed) {
 	if (!pressed) return;
-	// volume UP/DOWN buttons are not supported by iPhone/Android
-	volume_set_by_local_host(s_volume > 3 ? s_volume - 3 : 0);
+	volume_set_by_local_host(-3, true);
 	(*bt_app_a2d_cmd_cb)(BT_SINK_VOLUME, s_volume);
 }
 
@@ -134,12 +129,14 @@ static void bt_next(bool pressed) {
 }
 
 const static actrls_t controls = {
+	NULL, // power
 	bt_volume_up, bt_volume_down,	// volume up, volume down
 	bt_toggle, bt_play,	// toggle, play
 	bt_pause, bt_stop,	// pause, stop
 	NULL, NULL,			// rew, fwd
 	bt_prev, bt_next,	// prev, next
 	NULL, NULL, NULL, NULL, // left, right, up, down
+	NULL, NULL, NULL, NULL, NULL, NULL, // pre1-6
 	bt_volume_down, bt_volume_up, bt_toggle// knob left, knob_right, knob push
 };
 
@@ -175,7 +172,7 @@ static bool cmd_handler(bt_sink_cmd_t cmd, ...) {
 	// now handle events for display
 	switch(cmd) {
 	case BT_SINK_AUDIO_STARTED:
-		displayer_control(DISPLAYER_ACTIVATE, "BLUETOOTH");
+		displayer_control(DISPLAYER_ACTIVATE, "BLUETOOTH", false);
 		break;
 	case BT_SINK_AUDIO_STOPPED:
 		displayer_control(DISPLAYER_SUSPEND);
@@ -217,6 +214,10 @@ void bt_app_a2d_cb(esp_a2d_cb_event_t event, esp_a2d_cb_param_t *param)
         bt_app_work_dispatch(bt_av_hdl_a2d_evt, event, param, sizeof(esp_a2d_cb_param_t), NULL);
         break;
     }
+    case ESP_A2D_PROF_STATE_EVT: {
+        ESP_LOGI(BT_AV_TAG, "Bluetooth Init complete");
+        break;
+    }    
     default:
         ESP_LOGE(BT_AV_TAG, "Invalid A2DP event: %d", event);
         break;
@@ -283,6 +284,8 @@ static void bt_av_hdl_a2d_evt(uint16_t event, void *p_param)
             esp_bt_gap_set_scan_mode(ESP_BT_CONNECTABLE, ESP_BT_GENERAL_DISCOVERABLE);
 			(*bt_app_a2d_cmd_cb)(BT_SINK_DISCONNECTED);
         } else if (a2d->conn_stat.state == ESP_A2D_CONNECTION_STATE_CONNECTED){
+			abs_volume = -1;
+			s_volume = sink_volume;
             esp_bt_gap_set_scan_mode(ESP_BT_NON_CONNECTABLE, ESP_BT_NON_DISCOVERABLE);
 			(*bt_app_a2d_cmd_cb)(BT_SINK_CONNECTED);
         }
@@ -294,6 +297,9 @@ static void bt_av_hdl_a2d_evt(uint16_t event, void *p_param)
 
         if (ESP_A2D_AUDIO_STATE_STARTED == a2d->audio_stat.state) {
 			s_audio = AUDIO_CONNECTED;
+
+			// send memorized volume for devices that can't do absolute volume
+			(*bt_app_a2d_cmd_cb)(BT_SINK_VOLUME, s_volume);
 			
 			// verify that we can take control
 			if ((*bt_app_a2d_cmd_cb)(BT_SINK_AUDIO_STARTED, s_sample_rate)) {
@@ -378,7 +384,7 @@ static void bt_av_play_pos_changed(void)
     }
 }
 
-void bt_av_notify_evt_handler(uint8_t event_id, esp_avrc_rn_param_t *event_parameter)
+static void bt_av_notify_evt_handler(uint8_t event_id, esp_avrc_rn_param_t *event_parameter)
 {
     switch (event_id) {
     case ESP_AVRC_RN_TRACK_CHANGE:
@@ -487,20 +493,29 @@ static void bt_av_hdl_avrc_ct_evt(uint16_t event, void *p_param)
 
 static void volume_set_by_controller(uint8_t volume)
 {
-    ESP_LOGD(BT_RC_TG_TAG, "Volume is set by remote controller %d%%\n", (uint32_t)volume * 100 / 0x7f);
+	// do not modified NVS volume
     _lock_acquire(&s_volume_lock);
-    s_volume = volume;
+    s_volume = abs_volume = (volume * 100) / 127;
     _lock_release(&s_volume_lock);
-	(*bt_app_a2d_cmd_cb)(BT_SINK_VOLUME, volume);
+	(*bt_app_a2d_cmd_cb)(BT_SINK_VOLUME, s_volume);
 }
 
-static void volume_set_by_local_host(uint8_t volume)
+static void volume_set_by_local_host(int value, bool is_step)
 {
-    ESP_LOGD(BT_RC_TG_TAG, "Volume is set locally to: %d%%", (uint32_t)volume * 100 / 0x7f);
-    _lock_acquire(&s_volume_lock);
-    s_volume = volume;
-    _lock_release(&s_volume_lock);
+	_lock_acquire(&s_volume_lock);
+	s_volume = is_step ? s_volume + value : value;
+	if (s_volume > 127) s_volume = 127;
+	else if (s_volume < 0) s_volume = 0;	
+	if (abs_volume >= 0) abs_volume = s_volume;
+	else sink_volume = s_volume;
+	_lock_release(&s_volume_lock);
 
+	// volume has been set by controller, do not store it in NVS	
+	if (abs_volume < 0) {
+		char p[4];
+		config_set_value(NVS_TYPE_STR, "bt_sink_volume", itoa(s_volume, p, 10));					
+	}
+	
     if (s_volume_notify) {
         esp_avrc_rn_param_t rn_param;
         rn_param.volume = s_volume;
@@ -525,7 +540,7 @@ static void bt_av_hdl_avrc_tg_evt(uint16_t event, void *p_param)
         break;
     }
     case ESP_AVRC_TG_SET_ABSOLUTE_VOLUME_CMD_EVT: {
-        ESP_LOGD(BT_RC_TG_TAG, "AVRC set absolute volume: %d%%", (int)rc->set_abs_vol.volume * 100/ 0x7f);
+        ESP_LOGD(BT_RC_TG_TAG, "AVRC set absolute volume: %d%%", (rc->set_abs_vol.volume * 100) / 127);
         volume_set_by_controller(rc->set_abs_vol.volume);
         break;
     }
@@ -551,47 +566,16 @@ static void bt_av_hdl_avrc_tg_evt(uint16_t event, void *p_param)
 
 void bt_sink_init(bt_cmd_vcb_t cmd_cb, bt_data_cb_t data_cb)
 {
-	esp_err_t err;
-	
 	bt_app_a2d_cmd_cb = cmd_handler;
 	cmd_handler_chain = cmd_cb;
   	bt_app_a2d_data_cb = data_cb;
 	
-    ESP_ERROR_CHECK(esp_bt_controller_mem_release(ESP_BT_MODE_BLE));
-
-    esp_bt_controller_config_t bt_cfg = BT_CONTROLLER_INIT_CONFIG_DEFAULT();
-    if ((err = esp_bt_controller_init(&bt_cfg)) != ESP_OK) {
-        ESP_LOGE(BT_AV_TAG, "%s initialize controller failed: %s\n", __func__, esp_err_to_name(err));
-        return;
-    }
-
-    if ((err = esp_bt_controller_enable(ESP_BT_MODE_CLASSIC_BT)) != ESP_OK) {
-        ESP_LOGE(BT_AV_TAG, "%s enable controller failed: %s\n", __func__, esp_err_to_name(err));
-        return;
-    }
-
-    if ((err = esp_bluedroid_init()) != ESP_OK) {
-        ESP_LOGE(BT_AV_TAG, "%s initialize bluedroid failed: %s\n", __func__, esp_err_to_name(err));
-        return;
-    }
-
-    if ((err = esp_bluedroid_enable()) != ESP_OK) {
-        ESP_LOGE(BT_AV_TAG, "%s enable bluedroid failed: %s\n", __func__, esp_err_to_name(err));
-        return;
-    }
-
-    /* create application task */
-    bt_app_task_start_up();
-
-    /* Bluetooth device name, connection mode and profile set up */
-    bt_app_work_dispatch(bt_av_hdl_stack_evt, BT_APP_EVT_STACK_UP, NULL, 0, NULL);
-
-#if (CONFIG_BT_SSP_ENABLED == true)
-    /* Set default parameters for Secure Simple Pairing */
-    esp_bt_sp_param_t param_type = ESP_BT_SP_IOCAP_MODE;
-    esp_bt_io_cap_t iocap = ESP_BT_IO_CAP_IO;
-    esp_bt_gap_set_security_param(param_type, &iocap, sizeof(uint8_t));
-#endif
+	// create task and run event loop
+    bt_app_task_start_up(bt_av_hdl_stack_evt);
+	
+	char *item = config_alloc_get_default(NVS_TYPE_STR, "bt_sink_volume", "127", 0);
+	sink_volume = atol(item);
+	free(item);
 
     /*
      * Set default parameters for Legacy Pairing
@@ -617,30 +601,17 @@ void bt_sink_init(bt_cmd_vcb_t cmd_cb, bt_data_cb_t data_cb)
     		bError=true;
     	}
     	esp_pin_code[i]= pin_code[i];
-
     }
-    if(bError){
-    	esp_pin_code[0]='1';
-    	esp_pin_code[1]='2';
-    	esp_pin_code[2]='3';
-    	esp_pin_code[3]='4';
-    }
-    esp_bt_gap_set_pin(pin_type, strlen(pin_code), esp_pin_code);
+	
+    if (bError) memcpy(esp_pin_code, "1234", 4);
+	esp_bt_gap_set_pin(pin_type, strlen(pin_code), esp_pin_code);
+	
+	free(pin_code);
 }
 
 void bt_sink_deinit(void)
 {
-	/* this still does not work, can't figure out how to stop properly this BT stack */
 	bt_app_task_shut_down();
-	ESP_LOGD(BT_AV_TAG, "bt_app_task shutdown successfully");
-	if (esp_bluedroid_disable() != ESP_OK) return;
-    ESP_LOGD(BT_AV_TAG, "esp_bluedroid_disable called successfully");
-    if (esp_bluedroid_deinit() != ESP_OK) return;
-    ESP_LOGD(BT_AV_TAG, "esp_bluedroid_deinit called successfully");
-    if (esp_bt_controller_disable() != ESP_OK) return;
-    ESP_LOGD(BT_AV_TAG, "esp_bt_controller_disable called successfully");
-    if (esp_bt_controller_deinit() != ESP_OK) return;
-	ESP_LOGD(BT_AV_TAG, "bt stopped successfully");
 }
 
 static void bt_app_gap_cb(esp_bt_gap_cb_event_t event, esp_bt_gap_cb_param_t *param)

components/esp_http_server/CMakeLists.txt

@@ -0,0 +1,10 @@
+idf_build_get_property(prefix IDF_PATH)
+string(CONCAT prefix "${prefix}" "/components/esp_http_server")
+
+idf_component_register(
+		SRC_DIRS "${prefix}/src" "${prefix}/src/util"	
+		INCLUDE_DIRS "${prefix}/include"
+        PRIV_INCLUDE_DIRS "." "${prefix}/src/port/esp32" "${prefix}/src/util"
+        REQUIRES nghttp # for http_parser.h
+        PRIV_REQUIRES lwip mbedtls esp_timer
+) 

components/esp_http_server/Kconfig

@@ -0,0 +1 @@
+source "$IDF_PATH/components/esp_http_server/Kconfig"

components/esp_http_server/osal.h

@@ -0,0 +1,68 @@
+// Copyright 2018 Espressif Systems (Shanghai) PTE LTD
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+#ifndef _OSAL_H_
+#define _OSAL_H_
+
+#include <freertos/FreeRTOS.h>
+#include <freertos/task.h>
+#include <unistd.h>
+#include <stdint.h>
+#include <esp_timer.h>
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+#define OS_SUCCESS ESP_OK
+#define OS_FAIL    ESP_FAIL
+
+typedef TaskHandle_t othread_t;
+
+static inline int httpd_os_thread_create(othread_t *thread,
+                                 const char *name, uint16_t stacksize, int prio,
+                                 void (*thread_routine)(void *arg), void *arg,
+                                 BaseType_t core_id)
+{
+	StaticTask_t *xTaskBuffer = (StaticTask_t*) heap_caps_malloc(sizeof(StaticTask_t), (MALLOC_CAP_INTERNAL|MALLOC_CAP_8BIT));
+	StackType_t *xStack = heap_caps_malloc(stacksize,(MALLOC_CAP_SPIRAM|MALLOC_CAP_8BIT));
+
+	*thread = xTaskCreateStaticPinnedToCore(thread_routine, name, stacksize, arg, prio, xStack,xTaskBuffer,core_id);
+    if (*thread) {
+        return OS_SUCCESS;
+    }
+    return OS_FAIL;
+}
+
+/* Only self delete is supported */
+static inline void httpd_os_thread_delete(void)
+{
+    vTaskDelete(xTaskGetCurrentTaskHandle());
+}
+
+static inline void httpd_os_thread_sleep(int msecs)
+{
+    vTaskDelay(msecs / portTICK_RATE_MS);
+}
+
+static inline othread_t httpd_os_thread_handle(void)
+{
+    return xTaskGetCurrentTaskHandle();
+}
+
+#ifdef __cplusplus
+}
+#endif
+
+#endif /* ! _OSAL_H_ */

components/heap/CMakeLists.txt

@@ -0,0 +1,49 @@
+set(srcs
+    "heap_caps.c"
+    "heap_caps_init.c"
+    "multi_heap.c"
+    "heap_tlsf.c")
+
+if(NOT CONFIG_HEAP_POISONING_DISABLED)
+    list(APPEND srcs "multi_heap_poisoning.c")
+endif()
+
+if(CONFIG_HEAP_TASK_TRACKING)
+    list(APPEND srcs "heap_task_info.c")
+endif()
+
+if(CONFIG_HEAP_TRACING_STANDALONE)
+    list(APPEND srcs "heap_trace_standalone.c")
+    set_source_files_properties(heap_trace_standalone.c
+        PROPERTIES COMPILE_FLAGS
+        -Wno-frame-address)
+endif()
+
+idf_component_register(SRCS "${srcs}"
+                    INCLUDE_DIRS include
+                    LDFRAGMENTS linker.lf
+                    PRIV_REQUIRES soc)
+
+if(CONFIG_HEAP_TRACING)
+    set(WRAP_FUNCTIONS
+        calloc
+        malloc
+        free
+        realloc
+        heap_caps_malloc
+        heap_caps_free
+        heap_caps_realloc
+        heap_caps_malloc_default
+        heap_caps_realloc_default)
+
+    foreach(wrap ${WRAP_FUNCTIONS})
+        target_link_libraries(${COMPONENT_LIB} INTERFACE "-Wl,--wrap=${wrap}")
+    endforeach()
+endif()
+
+if(NOT CMAKE_BUILD_EARLY_EXPANSION)
+    idf_build_get_property(build_components BUILD_COMPONENTS)
+    if(freertos IN_LIST build_components)
+        target_compile_options(${COMPONENT_TARGET} PRIVATE "-DMULTI_HEAP_FREERTOS")
+    endif()
+endif()

components/heap/Kconfig

@@ -0,0 +1,74 @@
+menu "Heap memory debugging"
+
+    choice HEAP_CORRUPTION_DETECTION
+        prompt "Heap corruption detection"
+        default HEAP_POISONING_DISABLED
+        help
+            Enable heap poisoning features to detect heap corruption caused by out-of-bounds access to heap memory.
+
+            See the "Heap Memory Debugging" page of the IDF documentation
+            for a description of each level of heap corruption detection.
+
+        config HEAP_POISONING_DISABLED
+            bool "Basic (no poisoning)"
+        config HEAP_POISONING_LIGHT
+            bool "Light impact"
+        config HEAP_POISONING_COMPREHENSIVE
+            bool "Comprehensive"
+    endchoice
+
+    choice HEAP_TRACING_DEST
+        bool "Heap tracing"
+        default HEAP_TRACING_OFF
+        help
+            Enables the heap tracing API defined in esp_heap_trace.h.
+
+            This function causes a moderate increase in IRAM code side and a minor increase in heap function
+            (malloc/free/realloc) CPU overhead, even when the tracing feature is not used.
+            So it's best to keep it disabled unless tracing is being used.
+
+        config HEAP_TRACING_OFF
+            bool "Disabled"
+        config HEAP_TRACING_STANDALONE
+            bool "Standalone"
+            select HEAP_TRACING
+        config HEAP_TRACING_TOHOST
+            bool "Host-based"
+            select HEAP_TRACING
+    endchoice
+
+    config HEAP_TRACING
+        bool
+        default F
+        help
+            Enables/disables heap tracing API.
+
+    config HEAP_TRACING_STACK_DEPTH
+        int "Heap tracing stack depth"
+        range 0 0 if IDF_TARGET_ARCH_RISCV # Disabled for RISC-V due to `__builtin_return_address` limitation
+        default 0 if IDF_TARGET_ARCH_RISCV
+        range 0 10
+        default 2
+        depends on HEAP_TRACING
+        help
+            Number of stack frames to save when tracing heap operation callers.
+
+            More stack frames uses more memory in the heap trace buffer (and slows down allocation), but
+            can provide useful information.
+
+    config HEAP_TASK_TRACKING
+        bool "Enable heap task tracking"
+        depends on !HEAP_POISONING_DISABLED
+        help
+            Enables tracking the task responsible for each heap allocation.
+
+            This function depends on heap poisoning being enabled and adds four more bytes of overhead for each block
+            allocated.
+
+    config HEAP_ABORT_WHEN_ALLOCATION_FAILS
+        bool "Abort if memory allocation fails"
+        default n
+        help
+            When enabled, if a memory allocation operation fails it will cause a system abort.
+
+endmenu

components/heap/component.mk

@@ -0,0 +1,32 @@
+#
+# Component Makefile
+#
+
+COMPONENT_OBJS := heap_caps_init.o heap_caps.o multi_heap.o heap_tlsf.o
+
+ifndef CONFIG_HEAP_POISONING_DISABLED
+COMPONENT_OBJS += multi_heap_poisoning.o
+
+ifdef CONFIG_HEAP_TASK_TRACKING
+COMPONENT_OBJS += heap_task_info.o
+endif
+endif
+
+ifdef CONFIG_HEAP_TRACING_STANDALONE
+
+COMPONENT_OBJS += heap_trace_standalone.o
+
+endif
+
+ifdef CONFIG_HEAP_TRACING
+
+WRAP_FUNCTIONS = calloc malloc free realloc heap_caps_malloc heap_caps_free heap_caps_realloc heap_caps_malloc_default heap_caps_realloc_default
+WRAP_ARGUMENT := -Wl,--wrap=
+
+COMPONENT_ADD_LDFLAGS = -l$(COMPONENT_NAME) $(addprefix $(WRAP_ARGUMENT),$(WRAP_FUNCTIONS))
+
+endif
+
+COMPONENT_ADD_LDFRAGMENTS += linker.lf
+
+CFLAGS += -DMULTI_HEAP_FREERTOS

components/heap/heap_caps.c

@@ -0,0 +1,609 @@
+// Copyright 2015-2016 Espressif Systems (Shanghai) PTE LTD
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+#include <stdbool.h>
+#include <string.h>
+#include <assert.h>
+#include <stdio.h>
+#include <sys/param.h>
+#include "esp_attr.h"
+#include "esp_heap_caps.h"
+#include "multi_heap.h"
+#include "esp_log.h"
+#include "heap_private.h"
+#include "esp_system.h"
+
+/*
+This file, combined with a region allocator that supports multiple heaps, solves the problem that the ESP32 has RAM
+that's slightly heterogeneous. Some RAM can be byte-accessed, some allows only 32-bit accesses, some can execute memory,
+some can be remapped by the MMU to only be accessed by a certain PID etc. In order to allow the most flexible memory
+allocation possible, this code makes it possible to request memory that has certain capabilities. The code will then use
+its knowledge of how the memory is configured along with a priority scheme to allocate that memory in the most sane way
+possible. This should optimize the amount of RAM accessible to the code without hardwiring addresses.
+*/
+
+static esp_alloc_failed_hook_t alloc_failed_callback;
+
+/*
+  This takes a memory chunk in a region that can be addressed as both DRAM as well as IRAM. It will convert it to
+  IRAM in such a way that it can be later freed. It assumes both the address as well as the length to be word-aligned.
+  It returns a region that's 1 word smaller than the region given because it stores the original Dram address there.
+*/
+IRAM_ATTR static void *dram_alloc_to_iram_addr(void *addr, size_t len)
+{
+    uintptr_t dstart = (uintptr_t)addr; //First word
+    uintptr_t dend = dstart + len - 4; //Last word
+    assert(esp_ptr_in_diram_dram((void *)dstart));
+    assert(esp_ptr_in_diram_dram((void *)dend));
+    assert((dstart & 3) == 0);
+    assert((dend & 3) == 0);
+#if SOC_DIRAM_INVERTED // We want the word before the result to hold the DRAM address
+    uint32_t *iptr = esp_ptr_diram_dram_to_iram((void *)dend);
+#else
+    uint32_t *iptr = esp_ptr_diram_dram_to_iram((void *)dstart);
+#endif
+    *iptr = dstart;
+    return iptr + 1;
+}
+
+
+static void heap_caps_alloc_failed(size_t requested_size, uint32_t caps, const char *function_name)
+{
+    if (alloc_failed_callback) {
+        alloc_failed_callback(requested_size, caps, function_name);
+    }
+
+    #ifdef CONFIG_HEAP_ABORT_WHEN_ALLOCATION_FAILS
+    esp_system_abort("Memory allocation failed");
+    #endif
+}
+
+esp_err_t heap_caps_register_failed_alloc_callback(esp_alloc_failed_hook_t callback)
+{
+    if (callback == NULL) {
+        return ESP_ERR_INVALID_ARG;
+    }
+
+    alloc_failed_callback = callback;
+
+    return ESP_OK;
+}
+
+bool heap_caps_match(const heap_t *heap, uint32_t caps)
+{
+    return heap->heap != NULL && ((get_all_caps(heap) & caps) == caps);
+}
+
+/*
+Routine to allocate a bit of memory with certain capabilities. caps is a bitfield of MALLOC_CAP_* bits.
+*/
+IRAM_ATTR void *heap_caps_malloc( size_t size, uint32_t caps )
+{
+    void *ret = NULL;
+
+    if (size > HEAP_SIZE_MAX) {
+        // Avoids int overflow when adding small numbers to size, or
+        // calculating 'end' from start+size, by limiting 'size' to the possible range
+        heap_caps_alloc_failed(size, caps, __func__);
+
+        return NULL;
+    }
+
+    if (caps & MALLOC_CAP_EXEC) {
+        //MALLOC_CAP_EXEC forces an alloc from IRAM. There is a region which has both this as well as the following
+        //caps, but the following caps are not possible for IRAM.  Thus, the combination is impossible and we return
+        //NULL directly, even although our heap capabilities (based on soc_memory_tags & soc_memory_regions) would
+        //indicate there is a tag for this.
+        if ((caps & MALLOC_CAP_8BIT) || (caps & MALLOC_CAP_DMA)) {
+            heap_caps_alloc_failed(size, caps, __func__);
+
+            return NULL;
+        }
+        caps |= MALLOC_CAP_32BIT; // IRAM is 32-bit accessible RAM
+    }
+
+    if (caps & MALLOC_CAP_32BIT) {
+        /* 32-bit accessible RAM should allocated in 4 byte aligned sizes
+         * (Future versions of ESP-IDF should possibly fail if an invalid size is requested)
+         */
+        size = (size + 3) & (~3); // int overflow checked above
+    }
+
+    for (int prio = 0; prio < SOC_MEMORY_TYPE_NO_PRIOS; prio++) {
+        //Iterate over heaps and check capabilities at this priority
+        heap_t *heap;
+        SLIST_FOREACH(heap, &registered_heaps, next) {
+            if (heap->heap == NULL) {
+                continue;
+            }
+            if ((heap->caps[prio] & caps) != 0) {
+                //Heap has at least one of the caps requested. If caps has other bits set that this prio
+                //doesn't cover, see if they're available in other prios.
+                if ((get_all_caps(heap) & caps) == caps) {
+                    //This heap can satisfy all the requested capabilities. See if we can grab some memory using it.
+                    if ((caps & MALLOC_CAP_EXEC) && esp_ptr_in_diram_dram((void *)heap->start)) {
+                        //This is special, insofar that what we're going to get back is a DRAM address. If so,
+                        //we need to 'invert' it (lowest address in DRAM == highest address in IRAM and vice-versa) and
+                        //add a pointer to the DRAM equivalent before the address we're going to return.
+                        ret = multi_heap_malloc(heap->heap, size + 4);  // int overflow checked above
+
+                        if (ret != NULL) {
+                            return dram_alloc_to_iram_addr(ret, size + 4);  // int overflow checked above
+                        }
+                    } else {
+                        //Just try to alloc, nothing special.
+                        ret = multi_heap_malloc(heap->heap, size);
+                        if (ret != NULL) {
+                            return ret;
+                        }
+                    }
+                }
+            }
+        }
+    }
+
+    heap_caps_alloc_failed(size, caps, __func__);
+
+    //Nothing usable found.
+    return NULL;
+}
+
+
+#define MALLOC_DISABLE_EXTERNAL_ALLOCS -1
+//Dual-use: -1 (=MALLOC_DISABLE_EXTERNAL_ALLOCS) disables allocations in external memory, >=0 sets the limit for allocations preferring internal memory.
+static int malloc_alwaysinternal_limit=MALLOC_DISABLE_EXTERNAL_ALLOCS;
+
+void heap_caps_malloc_extmem_enable(size_t limit)
+{
+    malloc_alwaysinternal_limit=limit;
+}
+
+/*
+ Default memory allocation implementation. Should return standard 8-bit memory. malloc() essentially resolves to this function.
+*/
+IRAM_ATTR void *heap_caps_malloc_default( size_t size )
+{
+    if (malloc_alwaysinternal_limit==MALLOC_DISABLE_EXTERNAL_ALLOCS) {
+        return heap_caps_malloc( size, MALLOC_CAP_DEFAULT | MALLOC_CAP_INTERNAL);
+    } else {
+        void *r;
+        if (size <= (size_t)malloc_alwaysinternal_limit) {
+            r=heap_caps_malloc( size, MALLOC_CAP_DEFAULT | MALLOC_CAP_INTERNAL );
+        } else {
+            r=heap_caps_malloc( size, MALLOC_CAP_DEFAULT | MALLOC_CAP_SPIRAM );
+        }
+        if (r==NULL) {
+            //try again while being less picky
+            r=heap_caps_malloc( size, MALLOC_CAP_DEFAULT );
+        }
+        return r;
+    }
+}
+
+/*
+ Same for realloc()
+ Note: keep the logic in here the same as in heap_caps_malloc_default (or merge the two as soon as this gets more complex...)
+ */
+IRAM_ATTR void *heap_caps_realloc_default( void *ptr, size_t size )
+{
+    if (malloc_alwaysinternal_limit==MALLOC_DISABLE_EXTERNAL_ALLOCS) {
+        return heap_caps_realloc( ptr, size, MALLOC_CAP_DEFAULT | MALLOC_CAP_INTERNAL );
+    } else {
+        void *r;
+        if (size <= (size_t)malloc_alwaysinternal_limit) {
+            r=heap_caps_realloc( ptr, size, MALLOC_CAP_DEFAULT | MALLOC_CAP_INTERNAL );
+        } else {
+            r=heap_caps_realloc( ptr, size, MALLOC_CAP_DEFAULT | MALLOC_CAP_SPIRAM );
+        }
+        if (r==NULL && size>0) {
+            //We needed to allocate memory, but we didn't. Try again while being less picky.
+            r=heap_caps_realloc( ptr, size, MALLOC_CAP_DEFAULT );
+        }
+        return r;
+    }
+}
+
+/*
+ Memory allocation as preference in decreasing order.
+ */
+IRAM_ATTR void *heap_caps_malloc_prefer( size_t size, size_t num, ... )
+{
+    va_list argp;
+    va_start( argp, num );
+    void *r = NULL;
+    while (num--) {
+        uint32_t caps = va_arg( argp, uint32_t );
+        r = heap_caps_malloc( size, caps );
+        if (r != NULL) {
+            break;
+        }
+    }
+    va_end( argp );
+    return r;
+}
+
+/*
+ Memory reallocation as preference in decreasing order.
+ */
+IRAM_ATTR void *heap_caps_realloc_prefer( void *ptr, size_t size, size_t num, ... )
+{
+    va_list argp;
+    va_start( argp, num );
+    void *r = NULL;
+    while (num--) {
+        uint32_t caps = va_arg( argp, uint32_t );
+        r = heap_caps_realloc( ptr, size, caps );
+        if (r != NULL || size == 0) {
+            break;
+        }
+    }
+    va_end( argp );
+    return r;
+}
+
+/*
+ Memory callocation as preference in decreasing order.
+ */
+IRAM_ATTR void *heap_caps_calloc_prefer( size_t n, size_t size, size_t num, ... )
+{
+    va_list argp;
+    va_start( argp, num );
+    void *r = NULL;
+    while (num--) {
+        uint32_t caps = va_arg( argp, uint32_t );
+        r = heap_caps_calloc( n, size, caps );
+        if (r != NULL) break;
+    }
+    va_end( argp );
+    return r;
+}
+
+/* Find the heap which belongs to ptr, or return NULL if it's
+   not in any heap.
+
+   (This confirms if ptr is inside the heap's region, doesn't confirm if 'ptr'
+   is an allocated block or is some other random address inside the heap.)
+*/
+IRAM_ATTR static heap_t *find_containing_heap(void *ptr )
+{
+    intptr_t p = (intptr_t)ptr;
+    heap_t *heap;
+    SLIST_FOREACH(heap, &registered_heaps, next) {
+        if (heap->heap != NULL && p >= heap->start && p < heap->end) {
+            return heap;
+        }
+    }
+    return NULL;
+}
+
+IRAM_ATTR void heap_caps_free( void *ptr)
+{
+    if (ptr == NULL) {
+        return;
+    }
+
+    if (esp_ptr_in_diram_iram(ptr)) {
+        //Memory allocated here is actually allocated in the DRAM alias region and
+        //cannot be de-allocated as usual. dram_alloc_to_iram_addr stores a pointer to
+        //the equivalent DRAM address, though; free that.
+        uint32_t *dramAddrPtr = (uint32_t *)ptr;
+        ptr = (void *)dramAddrPtr[-1];
+    }
+
+    heap_t *heap = find_containing_heap(ptr);
+    assert(heap != NULL && "free() target pointer is outside heap areas");
+    multi_heap_free(heap->heap, ptr);
+}
+
+IRAM_ATTR void *heap_caps_realloc( void *ptr, size_t size, uint32_t caps)
+{
+    bool ptr_in_diram_case = false;
+    heap_t *heap = NULL;
+    void *dram_ptr = NULL;
+
+    if (ptr == NULL) {
+        return heap_caps_malloc(size, caps);
+    }
+
+    if (size == 0) {
+        heap_caps_free(ptr);
+        return NULL;
+    }
+
+    if (size > HEAP_SIZE_MAX) {
+        heap_caps_alloc_failed(size, caps, __func__);
+
+        return NULL;
+    }
+
+    //The pointer to memory may be aliased, we need to
+    //recover the corresponding address before to manage a new allocation:
+    if(esp_ptr_in_diram_iram((void *)ptr)) {
+        uint32_t *dram_addr = (uint32_t *)ptr;
+        dram_ptr  = (void *)dram_addr[-1];
+
+        heap = find_containing_heap(dram_ptr);
+        assert(heap != NULL && "realloc() pointer is outside heap areas");
+
+        //with pointers that reside on diram space, we avoid using
+        //the realloc implementation due to address translation issues,
+        //instead force a malloc/copy/free
+        ptr_in_diram_case = true;
+
+    } else {
+        heap = find_containing_heap(ptr);
+        assert(heap != NULL && "realloc() pointer is outside heap areas");
+    }
+
+    // are the existing heap's capabilities compatible with the
+    // requested ones?
+    bool compatible_caps = (caps & get_all_caps(heap)) == caps;
+
+    if (compatible_caps && !ptr_in_diram_case) {
+        // try to reallocate this memory within the same heap
+        // (which will resize the block if it can)
+        void *r = multi_heap_realloc(heap->heap, ptr, size);
+        if (r != NULL) {
+            return r;
+        }
+    }
+
+    // if we couldn't do that, try to see if we can reallocate
+    // in a different heap with requested capabilities.
+    void *new_p = heap_caps_malloc(size, caps);
+    if (new_p != NULL) {
+        size_t old_size = 0;
+
+        //If we're dealing with aliased ptr, information regarding its containing
+        //heap can only be obtained with translated address.
+        if(ptr_in_diram_case) {
+            old_size = multi_heap_get_allocated_size(heap->heap, dram_ptr);
+        } else {
+            old_size = multi_heap_get_allocated_size(heap->heap, ptr);
+        }
+
+        assert(old_size > 0);
+        memcpy(new_p, ptr, MIN(size, old_size));
+        heap_caps_free(ptr);
+        return new_p;
+    }
+
+    heap_caps_alloc_failed(size, caps, __func__);
+
+    return NULL;
+}
+
+IRAM_ATTR void *heap_caps_calloc( size_t n, size_t size, uint32_t caps)
+{
+    void *result;
+    size_t size_bytes;
+
+    if (__builtin_mul_overflow(n, size, &size_bytes)) {
+        return NULL;
+    }
+
+    result = heap_caps_malloc(size_bytes, caps);
+    if (result != NULL) {
+        bzero(result, size_bytes);
+    }
+    return result;
+}
+
+size_t heap_caps_get_total_size(uint32_t caps)
+{
+    size_t total_size = 0;
+    heap_t *heap;
+    SLIST_FOREACH(heap, &registered_heaps, next) {
+        if (heap_caps_match(heap, caps)) {
+            total_size += (heap->end - heap->start);
+        }
+    }
+    return total_size;
+}
+
+size_t heap_caps_get_free_size( uint32_t caps )
+{
+    size_t ret = 0;
+    heap_t *heap;
+    SLIST_FOREACH(heap, &registered_heaps, next) {
+        if (heap_caps_match(heap, caps)) {
+            ret += multi_heap_free_size(heap->heap);
+        }
+    }
+    return ret;
+}
+
+size_t heap_caps_get_minimum_free_size( uint32_t caps )
+{
+    size_t ret = 0;
+    heap_t *heap;
+    SLIST_FOREACH(heap, &registered_heaps, next) {
+        if (heap_caps_match(heap, caps)) {
+            ret += multi_heap_minimum_free_size(heap->heap);
+        }
+    }
+    return ret;
+}
+
+size_t heap_caps_get_largest_free_block( uint32_t caps )
+{
+    multi_heap_info_t info;
+    heap_caps_get_info(&info, caps);
+    return info.largest_free_block;
+}
+
+void heap_caps_get_info( multi_heap_info_t *info, uint32_t caps )
+{
+    bzero(info, sizeof(multi_heap_info_t));
+
+    heap_t *heap;
+    SLIST_FOREACH(heap, &registered_heaps, next) {
+        if (heap_caps_match(heap, caps)) {
+            multi_heap_info_t hinfo;
+            multi_heap_get_info(heap->heap, &hinfo);
+
+            info->total_free_bytes += hinfo.total_free_bytes;
+            info->total_allocated_bytes += hinfo.total_allocated_bytes;
+            info->largest_free_block = MAX(info->largest_free_block,
+                                           hinfo.largest_free_block);
+            info->minimum_free_bytes += hinfo.minimum_free_bytes;
+            info->allocated_blocks += hinfo.allocated_blocks;
+            info->free_blocks += hinfo.free_blocks;
+            info->total_blocks += hinfo.total_blocks;
+        }
+    }
+}
+
+void heap_caps_print_heap_info( uint32_t caps )
+{
+    multi_heap_info_t info;
+    printf("Heap summary for capabilities 0x%08X:\n", caps);
+    heap_t *heap;
+    SLIST_FOREACH(heap, &registered_heaps, next) {
+        if (heap_caps_match(heap, caps)) {
+            multi_heap_get_info(heap->heap, &info);
+
+            printf("  At 0x%08x len %d free %d allocated %d min_free %d\n",
+                   heap->start, heap->end - heap->start, info.total_free_bytes, info.total_allocated_bytes, info.minimum_free_bytes);
+            printf("    largest_free_block %d alloc_blocks %d free_blocks %d total_blocks %d\n",
+                   info.largest_free_block, info.allocated_blocks,
+                   info.free_blocks, info.total_blocks);
+        }
+    }
+    printf("  Totals:\n");
+    heap_caps_get_info(&info, caps);
+
+    printf("    free %d allocated %d min_free %d largest_free_block %d\n", info.total_free_bytes, info.total_allocated_bytes, info.minimum_free_bytes, info.largest_free_block);
+}
+
+bool heap_caps_check_integrity(uint32_t caps, bool print_errors)
+{
+    bool all_heaps = caps & MALLOC_CAP_INVALID;
+    bool valid = true;
+
+    heap_t *heap;
+    SLIST_FOREACH(heap, &registered_heaps, next) {
+        if (heap->heap != NULL
+            && (all_heaps || (get_all_caps(heap) & caps) == caps)) {
+            valid = multi_heap_check(heap->heap, print_errors) && valid;
+        }
+    }
+
+    return valid;
+}
+
+bool heap_caps_check_integrity_all(bool print_errors)
+{
+    return heap_caps_check_integrity(MALLOC_CAP_INVALID, print_errors);
+}
+
+bool heap_caps_check_integrity_addr(intptr_t addr, bool print_errors)
+{
+    heap_t *heap = find_containing_heap((void *)addr);
+    if (heap == NULL) {
+        return false;
+    }
+    return multi_heap_check(heap->heap, print_errors);
+}
+
+void heap_caps_dump(uint32_t caps)
+{
+    bool all_heaps = caps & MALLOC_CAP_INVALID;
+    heap_t *heap;
+    SLIST_FOREACH(heap, &registered_heaps, next) {
+        if (heap->heap != NULL
+            && (all_heaps || (get_all_caps(heap) & caps) == caps)) {
+            multi_heap_dump(heap->heap);
+        }
+    }
+}
+
+void heap_caps_dump_all(void)
+{
+    heap_caps_dump(MALLOC_CAP_INVALID);
+}
+
+size_t heap_caps_get_allocated_size( void *ptr )
+{
+    heap_t *heap = find_containing_heap(ptr);
+    size_t size = multi_heap_get_allocated_size(heap->heap, ptr);
+    return size;
+}
+
+IRAM_ATTR void *heap_caps_aligned_alloc(size_t alignment, size_t size, uint32_t caps)
+{
+    void *ret = NULL;
+
+    if(!alignment) {
+        return NULL;
+    }
+
+    //Alignment must be a power of two:
+    if((alignment & (alignment - 1)) != 0) {
+        return NULL;
+    }
+
+    if (size > HEAP_SIZE_MAX) {
+        // Avoids int overflow when adding small numbers to size, or
+        // calculating 'end' from start+size, by limiting 'size' to the possible range
+        heap_caps_alloc_failed(size, caps, __func__);
+
+        return NULL;
+    }
+
+    for (int prio = 0; prio < SOC_MEMORY_TYPE_NO_PRIOS; prio++) {
+        //Iterate over heaps and check capabilities at this priority
+        heap_t *heap;
+        SLIST_FOREACH(heap, &registered_heaps, next) {
+            if (heap->heap == NULL) {
+                continue;
+            }
+            if ((heap->caps[prio] & caps) != 0) {
+                //Heap has at least one of the caps requested. If caps has other bits set that this prio
+                //doesn't cover, see if they're available in other prios.
+                if ((get_all_caps(heap) & caps) == caps) {
+                    //Just try to alloc, nothing special.
+                    ret = multi_heap_aligned_alloc(heap->heap, size, alignment);
+                    if (ret != NULL) {
+                        return ret;
+                    }
+                }
+            }
+        }
+    }
+
+    heap_caps_alloc_failed(size, caps, __func__);
+
+    //Nothing usable found.
+    return NULL;
+}
+
+IRAM_ATTR void heap_caps_aligned_free(void *ptr)
+{
+    heap_caps_free(ptr);
+}
+
+void *heap_caps_aligned_calloc(size_t alignment, size_t n, size_t size, uint32_t caps)
+{
+    size_t size_bytes;
+    if (__builtin_mul_overflow(n, size, &size_bytes)) {
+        return NULL;
+    }
+
+    void *ptr = heap_caps_aligned_alloc(alignment,size_bytes, caps);
+    if(ptr != NULL) {
+        memset(ptr, 0, size_bytes);
+    }
+
+    return ptr;
+}

components/heap/heap_caps_init.c

@@ -0,0 +1,241 @@
+// Copyright 2015-2016 Espressif Systems (Shanghai) PTE LTD
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+#include "heap_private.h"
+#include <assert.h>
+#include <string.h>
+#include <sys/lock.h>
+
+#include "esp_log.h"
+#include "multi_heap.h"
+#include "multi_heap_platform.h"
+#include "esp_heap_caps_init.h"
+#include "soc/soc_memory_layout.h"
+
+static const char *TAG = "heap_init";
+
+/* Linked-list of registered heaps */
+struct registered_heap_ll registered_heaps;
+
+static void register_heap(heap_t *region)
+{
+    size_t heap_size = region->end - region->start;
+    assert(heap_size <= HEAP_SIZE_MAX);
+    region->heap = multi_heap_register((void *)region->start, heap_size);
+    if (region->heap != NULL) {
+        ESP_EARLY_LOGD(TAG, "New heap initialised at %p", region->heap);
+    }
+}
+
+void heap_caps_enable_nonos_stack_heaps(void)
+{
+    heap_t *heap;
+    SLIST_FOREACH(heap, &registered_heaps, next) {
+        // Assume any not-yet-registered heap is
+        // a nonos-stack heap
+        if (heap->heap == NULL) {
+            register_heap(heap);
+            if (heap->heap != NULL) {
+                multi_heap_set_lock(heap->heap, &heap->heap_mux);
+            }
+        }
+    }
+}
+
+/* Initialize the heap allocator to use all of the memory not
+   used by static data or reserved for other purposes
+ */
+void heap_caps_init(void)
+{
+    /* Get the array of regions that we can use for heaps
+       (with reserved memory removed already.)
+     */
+    size_t num_regions = soc_get_available_memory_region_max_count();
+    soc_memory_region_t regions[num_regions];
+    num_regions = soc_get_available_memory_regions(regions);
+
+    //The heap allocator will treat every region given to it as separate. In order to get bigger ranges of contiguous memory,
+    //it's useful to coalesce adjacent regions that have the same type.
+    for (size_t i = 1; i < num_regions; i++) {
+        soc_memory_region_t *a = &regions[i - 1];
+        soc_memory_region_t *b = &regions[i];
+        if (b->start == (intptr_t)(a->start + a->size) && b->type == a->type ) {
+            a->type = -1;
+            b->start = a->start;
+            b->size += a->size;
+        }
+    }
+
+    /* Count the heaps left after merging */
+    size_t num_heaps = 0;
+    for (size_t i = 0; i < num_regions; i++) {
+        if (regions[i].type != -1) {
+            num_heaps++;
+        }
+    }
+
+    /* Start by allocating the registered heap data on the stack.
+
+       Once we have a heap to copy it to, we will copy it to a heap buffer.
+    */
+    heap_t temp_heaps[num_heaps];
+    size_t heap_idx = 0;
+
+    ESP_EARLY_LOGI(TAG, "Initializing. RAM available for dynamic allocation:");
+    for (size_t i = 0; i < num_regions; i++) {
+        soc_memory_region_t *region = &regions[i];
+        const soc_memory_type_desc_t *type = &soc_memory_types[region->type];
+        heap_t *heap = &temp_heaps[heap_idx];
+        if (region->type == -1) {
+            continue;
+        }
+        heap_idx++;
+        assert(heap_idx <= num_heaps);
+
+        memcpy(heap->caps, type->caps, sizeof(heap->caps));
+        heap->start = region->start;
+        heap->end = region->start + region->size;
+        MULTI_HEAP_LOCK_INIT(&heap->heap_mux);
+        if (type->startup_stack) {
+            /* Will be registered when OS scheduler starts */
+            heap->heap = NULL;
+        } else {
+            register_heap(heap);
+        }
+        SLIST_NEXT(heap, next) = NULL;
+
+        ESP_EARLY_LOGI(TAG, "At %08X len %08X (%d KiB): %s",
+                       region->start, region->size, region->size / 1024, type->name);
+    }
+
+    assert(heap_idx == num_heaps);
+
+    /* Allocate the permanent heap data that we'll use as a linked list at runtime.
+
+       Allocate this part of data contiguously, even though it's a linked list... */
+    assert(SLIST_EMPTY(&registered_heaps));
+
+    heap_t *heaps_array = NULL;
+    for (size_t i = 0; i < num_heaps; i++) {
+        if (heap_caps_match(&temp_heaps[i], MALLOC_CAP_8BIT|MALLOC_CAP_INTERNAL)) {
+            /* use the first DRAM heap which can fit the data */
+            heaps_array = multi_heap_malloc(temp_heaps[i].heap, sizeof(heap_t) * num_heaps);
+            if (heaps_array != NULL) {
+                break;
+            }
+        }
+    }
+    assert(heaps_array != NULL); /* if NULL, there's not enough free startup heap space */
+
+    memcpy(heaps_array, temp_heaps, sizeof(heap_t)*num_heaps);
+
+    /* Iterate the heaps and set their locks, also add them to the linked list. */
+    for (size_t i = 0; i < num_heaps; i++) {
+        if (heaps_array[i].heap != NULL) {
+            multi_heap_set_lock(heaps_array[i].heap, &heaps_array[i].heap_mux);
+        }
+        if (i == 0) {
+            SLIST_INSERT_HEAD(&registered_heaps, &heaps_array[0], next);
+        } else {
+            SLIST_INSERT_AFTER(&heaps_array[i-1], &heaps_array[i], next);
+        }
+    }
+}
+
+esp_err_t heap_caps_add_region(intptr_t start, intptr_t end)
+{
+    if (start == 0) {
+        return ESP_ERR_INVALID_ARG;
+    }
+
+    for (size_t i = 0; i < soc_memory_region_count; i++) {
+        const soc_memory_region_t *region = &soc_memory_regions[i];
+        // Test requested start only as 'end' may be in a different region entry, assume 'end' has same caps
+        if (region->start <= start && (intptr_t)(region->start + region->size) > start) {
+            const uint32_t *caps = soc_memory_types[region->type].caps;
+            return heap_caps_add_region_with_caps(caps, start, end);
+        }
+    }
+
+    return ESP_ERR_NOT_FOUND;
+}
+
+esp_err_t heap_caps_add_region_with_caps(const uint32_t caps[], intptr_t start, intptr_t end)
+{
+    esp_err_t err = ESP_FAIL;
+    if (caps == NULL || start == 0 || end == 0 || end <= start) {
+        return ESP_ERR_INVALID_ARG;
+    }
+
+    //Check if region overlaps the start and/or end of an existing region. If so, the
+    //region is invalid (or maybe added twice)
+    /*
+     *  assume that in on region, start must be less than end (cannot equal to) !!
+     *  Specially, the 4th scenario can be allowed. For example, allocate memory from heap,
+     *  then change the capability and call this function to create a new region for special
+     *  application.
+     *  In the following chart, 'start = start' and 'end = end' is contained in 3rd scenario.
+     *  This all equal scenario is incorrect because the same region cannot be add twice. For example,
+     *  add the .bss memory to region twice, if not do the check, it will cause exception.
+     *
+     *  the existing heap region                                  s(tart)                e(nd)
+     *                                                            |----------------------|
+     *  1.add region  [Correct]   (s1<s && e1<=s)           |-----|
+     *  2.add region  [Incorrect] (s2<=s && s<e2<=e)        |---------------|
+     *  3.add region  [Incorrect] (s3<=s && e<e3)           |-------------------------------------|
+     *  4 add region  [Correct]   (s<s4<e && s<e4<=e)                  |-------|
+     *  5.add region  [Incorrect] (s<s5<e && e<e5)                     |----------------------------|
+     *  6.add region  [Correct]   (e<=s6 && e<e6)                                        |----|
+     */
+
+    heap_t *heap;
+    SLIST_FOREACH(heap, &registered_heaps, next) {
+        if ((start <= heap->start && end > heap->start)
+                || (start < heap->end && end > heap->end)) {
+            return ESP_FAIL;
+        }
+    }
+
+    heap_t *p_new = heap_caps_malloc(sizeof(heap_t), MALLOC_CAP_INTERNAL|MALLOC_CAP_8BIT);
+    if (p_new == NULL) {
+        err = ESP_ERR_NO_MEM;
+        goto done;
+    }
+    memcpy(p_new->caps, caps, sizeof(p_new->caps));
+    p_new->start = start;
+    p_new->end = end;
+    MULTI_HEAP_LOCK_INIT(&p_new->heap_mux);
+    p_new->heap = multi_heap_register((void *)start, end - start);
+    SLIST_NEXT(p_new, next) = NULL;
+    if (p_new->heap == NULL) {
+        err = ESP_ERR_INVALID_SIZE;
+        goto done;
+    }
+    multi_heap_set_lock(p_new->heap, &p_new->heap_mux);
+
+    /* (This insertion is atomic to registered_heaps, so
+       we don't need to worry about thread safety for readers,
+       only for writers. */
+    static multi_heap_lock_t registered_heaps_write_lock = MULTI_HEAP_LOCK_STATIC_INITIALIZER;
+    MULTI_HEAP_LOCK(&registered_heaps_write_lock);
+    SLIST_INSERT_HEAD(&registered_heaps, p_new, next);
+    MULTI_HEAP_UNLOCK(&registered_heaps_write_lock);
+
+    err = ESP_OK;
+
+ done:
+    if (err != ESP_OK) {
+        free(p_new);
+    }
+    return err;
+}

components/heap/heap_private.h

@@ -0,0 +1,77 @@
+// Copyright 2015-2016 Espressif Systems (Shanghai) PTE LTD
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+#pragma once
+
+#include <stdlib.h>
+#include <stdint.h>
+#include <soc/soc_memory_layout.h>
+#include "multi_heap.h"
+#include "multi_heap_platform.h"
+#include "sys/queue.h"
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+/* Some common heap registration data structures used
+   for heap_caps_init.c to share heap information with heap_caps.c
+*/
+
+#define HEAP_SIZE_MAX (SOC_MAX_CONTIGUOUS_RAM_SIZE)
+
+/* Type for describing each registered heap */
+typedef struct heap_t_ {
+    uint32_t caps[SOC_MEMORY_TYPE_NO_PRIOS]; ///< Capabilities for the type of memory in this heap (as a prioritised set). Copied from soc_memory_types so it's in RAM not flash.
+    intptr_t start;
+    intptr_t end;
+    multi_heap_lock_t heap_mux;
+    multi_heap_handle_t heap;
+    SLIST_ENTRY(heap_t_) next;
+} heap_t;
+
+/* All registered heaps.
+
+   Forms a single linked list, even though most entries are contiguous.
+   This means at the expense of 4 bytes per heap, new heaps can be
+   added at runtime in a fast & thread-safe way.
+*/
+extern SLIST_HEAD(registered_heap_ll, heap_t_) registered_heaps;
+
+bool heap_caps_match(const heap_t *heap, uint32_t caps);
+
+/* return all possible capabilities (across all priorities) for a given heap */
+inline static IRAM_ATTR uint32_t get_all_caps(const heap_t *heap)
+{
+    if (heap->heap == NULL) {
+        return 0;
+    }
+    uint32_t all_caps = 0;
+    for (int prio = 0; prio < SOC_MEMORY_TYPE_NO_PRIOS; prio++) {
+        all_caps |= heap->caps[prio];
+    }
+    return all_caps;
+}
+
+/*
+ Because we don't want to add _another_ known allocation method to the stack of functions to trace wrt memory tracing,
+ these are declared private. The newlib malloc()/realloc() implementation also calls these, so they are declared
+ separately in newlib/syscalls.c.
+*/
+void *heap_caps_realloc_default(void *p, size_t size);
+void *heap_caps_malloc_default(size_t size);
+
+
+#ifdef __cplusplus
+}
+#endif

components/heap/heap_task_info.c

@@ -0,0 +1,129 @@
+// Copyright 2018 Espressif Systems (Shanghai) PTE LTD
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+#include <freertos/FreeRTOS.h>
+#include <freertos/task.h>
+#include <multi_heap.h>
+#include "multi_heap_internal.h"
+#include "heap_private.h"
+#include "esp_heap_task_info.h"
+
+#ifdef CONFIG_HEAP_TASK_TRACKING
+
+/*
+ * Return per-task heap allocation totals and lists of blocks.
+ *
+ * For each task that has allocated memory from the heap, return totals for
+ * allocations within regions matching one or more sets of capabilities.
+ *
+ * Optionally also return an array of structs providing details about each
+ * block allocated by one or more requested tasks, or by all tasks.
+ *
+ * Returns the number of block detail structs returned.
+ */
+size_t heap_caps_get_per_task_info(heap_task_info_params_t *params)
+{
+    heap_t *reg;
+    heap_task_block_t *blocks = params->blocks;
+    size_t count = *params->num_totals;
+    size_t remaining = params->max_blocks;
+
+    // Clear out totals for any prepopulated tasks.
+    if (params->totals) {
+        for (size_t i = 0; i < count; ++i) {
+            for (size_t type = 0; type < NUM_HEAP_TASK_CAPS; ++type) {
+                params->totals[i].size[type] = 0;
+                params->totals[i].count[type] = 0;
+            }
+        }
+    }
+
+    SLIST_FOREACH(reg, &registered_heaps, next) {
+        multi_heap_handle_t heap = reg->heap;
+        if (heap == NULL) {
+            continue;
+        }
+
+        // Find if the capabilities of this heap region match on of the desired
+        // sets of capabilities.
+        uint32_t caps = get_all_caps(reg);
+        uint32_t type;
+        for (type = 0; type < NUM_HEAP_TASK_CAPS; ++type) {
+            if ((caps & params->mask[type]) == params->caps[type]) {
+                break;
+            }
+        }
+        if (type == NUM_HEAP_TASK_CAPS) {
+            continue;
+        }
+
+        multi_heap_block_handle_t b = multi_heap_get_first_block(heap);
+        multi_heap_internal_lock(heap);
+        for ( ; b ; b = multi_heap_get_next_block(heap, b)) {
+            if (multi_heap_is_free(b)) {
+                continue;
+            }
+            void *p = multi_heap_get_block_address(b);  // Safe, only arithmetic
+            size_t bsize = multi_heap_get_allocated_size(heap, p); // Validates
+            TaskHandle_t btask = (TaskHandle_t)multi_heap_get_block_owner(b);
+
+            // Accumulate per-task allocation totals.
+            if (params->totals) {
+                size_t i;
+                for (i = 0; i < count; ++i) {
+                    if (params->totals[i].task == btask) {
+                        break;
+                    }
+                }
+                if (i < count) {
+                    params->totals[i].size[type] += bsize;
+                    params->totals[i].count[type] += 1;
+                }
+                else {
+                    if (count < params->max_totals) {
+                        params->totals[count].task = btask;
+                        params->totals[count].size[type] = bsize;
+                        params->totals[i].count[type] = 1;
+                        ++count;
+                    }
+                }
+            }
+
+            // Return details about allocated blocks for selected tasks.
+            if (blocks && remaining > 0) {
+                if (params->tasks) {
+                    size_t i;
+                    for (i = 0; i < params->num_tasks; ++i) {
+                        if (btask == params->tasks[i]) {
+                            break;
+                        }
+                    }
+                    if (i == params->num_tasks) {
+                        continue;
+                    }
+                }
+                blocks->task = btask;
+                blocks->address = p;
+                blocks->size = bsize;
+                ++blocks;
+                --remaining;
+            }
+        }
+        multi_heap_internal_unlock(heap);
+    }
+    *params->num_totals = count;
+    return params->max_blocks - remaining;
+}
+
+#endif // CONFIG_HEAP_TASK_TRACKING

components/heap/heap_tlsf.h

@@ -0,0 +1,119 @@
+/*
+** Two Level Segregated Fit memory allocator, version 3.1.
+** Written by Matthew Conte
+**	http://tlsf.baisoku.org
+**
+** Based on the original documentation by Miguel Masmano:
+**	http://www.gii.upv.es/tlsf/main/docs
+**
+** This implementation was written to the specification
+** of the document, therefore no GPL restrictions apply.
+**
+** Copyright (c) 2006-2016, Matthew Conte
+** All rights reserved.
+**
+** Redistribution and use in source and binary forms, with or without
+** modification, are permitted provided that the following conditions are met:
+**     * Redistributions of source code must retain the above copyright
+**       notice, this list of conditions and the following disclaimer.
+**     * Redistributions in binary form must reproduce the above copyright
+**       notice, this list of conditions and the following disclaimer in the
+**       documentation and/or other materials provided with the distribution.
+**     * Neither the name of the copyright holder nor the
+**       names of its contributors may be used to endorse or promote products
+**       derived from this software without specific prior written permission.
+**
+** THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
+** ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
+** WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
+** DISCLAIMED. IN NO EVENT SHALL MATTHEW CONTE BE LIABLE FOR ANY
+** DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
+** (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+** LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
+** ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+** (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
+** SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+*/
+
+#pragma once
+#include <assert.h>
+#include <limits.h>
+#include <stddef.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <stddef.h>
+#include "heap_tlsf_config.h"
+
+#if defined(__cplusplus)
+extern "C" {
+#endif
+
+/*
+** Cast and min/max macros.
+*/
+#define tlsf_cast(t, exp)	((t) (exp))
+#define tlsf_min(a, b)		((a) < (b) ? (a) : (b))
+#define tlsf_max(a, b)		((a) > (b) ? (a) : (b))
+
+/* A type used for casting when doing pointer arithmetic. */
+typedef ptrdiff_t tlsfptr_t;
+
+typedef struct block_header_t
+{
+	/* Points to the previous physical block. */
+	struct block_header_t* prev_phys_block;
+
+	/* The size of this block, excluding the block header. */
+	size_t size;
+
+	/* Next and previous free blocks. */
+	struct block_header_t* next_free;
+	struct block_header_t* prev_free;
+} block_header_t;
+
+#include "heap_tlsf_block_functions.h"
+
+/* tlsf_t: a TLSF structure. Can contain 1 to N pools. */
+/* pool_t: a block of memory that TLSF can manage. */
+typedef void* tlsf_t;
+typedef void* pool_t;
+
+/* Create/destroy a memory pool. */
+tlsf_t tlsf_create(void* mem, size_t max_bytes);
+tlsf_t tlsf_create_with_pool(void* mem, size_t pool_bytes, size_t max_bytes);
+pool_t tlsf_get_pool(tlsf_t tlsf);
+
+/* Add/remove memory pools. */
+pool_t tlsf_add_pool(tlsf_t tlsf, void* mem, size_t bytes);
+void tlsf_remove_pool(tlsf_t tlsf, pool_t pool);
+
+/* malloc/memalign/realloc/free replacements. */
+void* tlsf_malloc(tlsf_t tlsf, size_t size);
+void* tlsf_memalign(tlsf_t tlsf, size_t align, size_t size);
+void* tlsf_memalign_offs(tlsf_t tlsf, size_t align, size_t size, size_t offset);
+void* tlsf_realloc(tlsf_t tlsf, void* ptr, size_t size);
+void tlsf_free(tlsf_t tlsf, void* ptr);
+
+/* Returns internal block size, not original request size */
+size_t tlsf_block_size(void* ptr);
+
+/* Overheads/limits of internal structures. */
+size_t tlsf_size(tlsf_t tlsf);
+size_t tlsf_align_size(void);
+size_t tlsf_block_size_min(void);
+size_t tlsf_block_size_max(tlsf_t tlsf);
+size_t tlsf_pool_overhead(void);
+size_t tlsf_alloc_overhead(void);
+size_t tlsf_fit_size(tlsf_t tlsf, size_t size);
+
+/* Debugging. */
+typedef void (*tlsf_walker)(void* ptr, size_t size, int used, void* user);
+void tlsf_walk_pool(pool_t pool, tlsf_walker walker, void* user);
+/* Returns nonzero if any internal consistency check fails. */
+int tlsf_check(tlsf_t tlsf);
+int tlsf_check_pool(pool_t pool);
+
+#if defined(__cplusplus)
+};
+#endif

components/heap/heap_tlsf_block_functions.h

@@ -0,0 +1,174 @@
+/*
+** Two Level Segregated Fit memory allocator, version 3.1.
+** Written by Matthew Conte
+**	http://tlsf.baisoku.org
+**
+** Based on the original documentation by Miguel Masmano:
+**	http://www.gii.upv.es/tlsf/main/docs
+**
+** This implementation was written to the specification
+** of the document, therefore no GPL restrictions apply.
+**
+** Copyright (c) 2006-2016, Matthew Conte
+** All rights reserved.
+**
+** Redistribution and use in source and binary forms, with or without
+** modification, are permitted provided that the following conditions are met:
+**     * Redistributions of source code must retain the above copyright
+**       notice, this list of conditions and the following disclaimer.
+**     * Redistributions in binary form must reproduce the above copyright
+**       notice, this list of conditions and the following disclaimer in the
+**       documentation and/or other materials provided with the distribution.
+**     * Neither the name of the copyright holder nor the
+**       names of its contributors may be used to endorse or promote products
+**       derived from this software without specific prior written permission.
+**
+** THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
+** ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
+** WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
+** DISCLAIMED. IN NO EVENT SHALL MATTHEW CONTE BE LIABLE FOR ANY
+** DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
+** (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+** LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
+** ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+** (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
+** SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+*/
+
+#pragma once
+
+/*
+** Data structures and associated constants.
+*/
+
+/*
+** Since block sizes are always at least a multiple of 4, the two least
+** significant bits of the size field are used to store the block status:
+** - bit 0: whether block is busy or free
+** - bit 1: whether previous block is busy or free
+*/
+#define block_header_free_bit  (1 << 0)
+#define block_header_prev_free_bit  (1 << 1)
+
+/*
+** The size of the block header exposed to used blocks is the size field.
+** The prev_phys_block field is stored *inside* the previous free block.
+*/
+#define block_header_overhead  (sizeof(size_t))
+
+/* User data starts directly after the size field in a used block. */
+#define block_start_offset (offsetof(block_header_t, size) + sizeof(size_t))
+
+/*
+** A free block must be large enough to store its header minus the size of
+** the prev_phys_block field, and no larger than the number of addressable
+** bits for FL_INDEX.
+** The block_size_max macro returns the maximum block for the minimum pool
+** use tlsf_block_size_max for a value specific to the pool
+*/
+#define block_size_min  (sizeof(block_header_t) - sizeof(block_header_t*))
+#define block_size_max  (tlsf_cast(size_t, 1) << FL_INDEX_MAX_MIN)
+
+/*
+** block_header_t member functions.
+*/
+static inline __attribute__((__always_inline__)) size_t block_size(const block_header_t* block)
+{
+	return block->size & ~(block_header_free_bit | block_header_prev_free_bit);
+}
+
+static inline __attribute__((__always_inline__)) void block_set_size(block_header_t* block, size_t size)
+{
+	const size_t oldsize = block->size;
+	block->size = size | (oldsize & (block_header_free_bit | block_header_prev_free_bit));
+}
+
+static inline __attribute__((__always_inline__)) int block_is_last(const block_header_t* block)
+{
+	return block_size(block) == 0;
+}
+
+static inline __attribute__((__always_inline__)) int block_is_free(const block_header_t* block)
+{
+	return tlsf_cast(int, block->size & block_header_free_bit);
+}
+
+static inline __attribute__((__always_inline__)) void block_set_free(block_header_t* block)
+{
+	block->size |= block_header_free_bit;
+}
+
+static inline __attribute__((__always_inline__)) void block_set_used(block_header_t* block)
+{
+	block->size &= ~block_header_free_bit;
+}
+
+static inline __attribute__((__always_inline__)) int block_is_prev_free(const block_header_t* block)
+{
+	return tlsf_cast(int, block->size & block_header_prev_free_bit);
+}
+
+static inline __attribute__((__always_inline__)) void block_set_prev_free(block_header_t* block)
+{
+	block->size |= block_header_prev_free_bit;
+}
+
+static inline __attribute__((__always_inline__)) void block_set_prev_used(block_header_t* block)
+{
+	block->size &= ~block_header_prev_free_bit;
+}
+
+static inline __attribute__((__always_inline__)) block_header_t* block_from_ptr(const void* ptr)
+{
+	return tlsf_cast(block_header_t*,
+		tlsf_cast(unsigned char*, ptr) - block_start_offset);
+}
+
+static inline __attribute__((__always_inline__)) void* block_to_ptr(const block_header_t* block)
+{
+	return tlsf_cast(void*,
+		tlsf_cast(unsigned char*, block) + block_start_offset);
+}
+
+/* Return location of next block after block of given size. */
+static inline __attribute__((__always_inline__)) block_header_t* offset_to_block(const void* ptr, size_t size)
+{
+	return tlsf_cast(block_header_t*, tlsf_cast(tlsfptr_t, ptr) + size);
+}
+
+/* Return location of previous block. */
+static inline __attribute__((__always_inline__)) block_header_t* block_prev(const block_header_t* block)
+{
+	return block->prev_phys_block;
+}
+
+/* Return location of next existing block. */
+static inline __attribute__((__always_inline__)) block_header_t* block_next(const block_header_t* block)
+{
+	block_header_t* next = offset_to_block(block_to_ptr(block),
+		block_size(block) - block_header_overhead);
+	return next;
+}
+
+/* Link a new block with its physical neighbor, return the neighbor. */
+static inline __attribute__((__always_inline__)) block_header_t* block_link_next(block_header_t* block)
+{
+	block_header_t* next = block_next(block);
+	next->prev_phys_block = block;
+	return next;
+}
+
+static inline __attribute__((__always_inline__)) void block_mark_as_free(block_header_t* block)
+{
+	/* Link the block to the next block, first. */
+	block_header_t* next = block_link_next(block);
+	block_set_prev_free(next);
+	block_set_free(block);
+}
+
+static inline __attribute__((__always_inline__)) void block_mark_as_used(block_header_t* block)
+{
+	block_header_t* next = block_next(block);
+	block_set_prev_used(next);
+	block_set_used(block);
+}

components/heap/heap_tlsf_config.h

@@ -0,0 +1,66 @@
+/*
+** Two Level Segregated Fit memory allocator, version 3.1.
+** Written by Matthew Conte
+**	http://tlsf.baisoku.org
+**
+** Based on the original documentation by Miguel Masmano:
+**	http://www.gii.upv.es/tlsf/main/docs
+**
+** This implementation was written to the specification
+** of the document, therefore no GPL restrictions apply.
+**
+** Copyright (c) 2006-2016, Matthew Conte
+** All rights reserved.
+**
+** Redistribution and use in source and binary forms, with or without
+** modification, are permitted provided that the following conditions are met:
+**     * Redistributions of source code must retain the above copyright
+**       notice, this list of conditions and the following disclaimer.
+**     * Redistributions in binary form must reproduce the above copyright
+**       notice, this list of conditions and the following disclaimer in the
+**       documentation and/or other materials provided with the distribution.
+**     * Neither the name of the copyright holder nor the
+**       names of its contributors may be used to endorse or promote products
+**       derived from this software without specific prior written permission.
+**
+** THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
+** ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
+** WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
+** DISCLAIMED. IN NO EVENT SHALL MATTHEW CONTE BE LIABLE FOR ANY
+** DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
+** (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+** LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
+** ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+** (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
+** SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+*/
+
+#pragma once
+
+enum tlsf_config
+{
+	/* log2 of number of linear subdivisions of block sizes. Larger
+	** values require more memory in the control structure. Values of
+	** 4 or 5 are typical, 3 is for very small pools.
+	*/
+	SL_INDEX_COUNT_LOG2_MIN = 3, 
+
+	/* All allocation sizes and addresses are aligned to 4 bytes. */
+	ALIGN_SIZE_LOG2 = 2,
+	ALIGN_SIZE = (1 << ALIGN_SIZE_LOG2),
+
+	/*
+	** We support allocations of sizes up to (1 << FL_INDEX_MAX) bits.
+	** However, because we linearly subdivide the second-level lists, and
+	** our minimum size granularity is 4 bytes, it doesn't make sense to
+	** create first-level lists for sizes smaller than SL_INDEX_COUNT * 4,
+	** or (1 << (SL_INDEX_COUNT_LOG2 + 2)) bytes, as there we will be
+	** trying to split size ranges into more slots than we have available.
+	** Instead, we calculate the minimum threshold size, and place all
+	** blocks below that size into the 0th first-level list.
+	** Values below are the absolute minimum to accept a pool addition
+	*/
+	FL_INDEX_MAX_MIN = 14, // For a less than 16kB pool
+	SL_INDEX_COUNT_MIN = (1 << SL_INDEX_COUNT_LOG2_MIN),
+	FL_INDEX_COUNT_MIN = (FL_INDEX_MAX_MIN - (SL_INDEX_COUNT_LOG2_MIN + ALIGN_SIZE_LOG2) + 1),
+};

components/heap/heap_trace_standalone.c

@@ -0,0 +1,255 @@
+// Copyright 2015-2016 Espressif Systems (Shanghai) PTE LTD
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+#include <string.h>
+#include <sdkconfig.h>
+
+#define HEAP_TRACE_SRCFILE /* don't warn on inclusion here */
+#include "esp_heap_trace.h"
+#undef HEAP_TRACE_SRCFILE
+
+#include "esp_attr.h"
+#include "freertos/FreeRTOS.h"
+#include "freertos/task.h"
+
+
+#define STACK_DEPTH CONFIG_HEAP_TRACING_STACK_DEPTH
+
+#if CONFIG_HEAP_TRACING_STANDALONE
+
+static portMUX_TYPE trace_mux = portMUX_INITIALIZER_UNLOCKED;
+static bool tracing;
+static heap_trace_mode_t mode;
+
+/* Buffer used for records, starting at offset 0
+*/
+static heap_trace_record_t *buffer;
+static size_t total_records;
+
+/* Count of entries logged in the buffer.
+
+   Maximum total_records
+*/
+static size_t count;
+
+/* Actual number of allocations logged */
+static size_t total_allocations;
+
+/* Actual number of frees logged */
+static size_t total_frees;
+
+/* Has the buffer overflowed and lost trace entries? */
+static bool has_overflowed = false;
+
+esp_err_t heap_trace_init_standalone(heap_trace_record_t *record_buffer, size_t num_records)
+{
+    if (tracing) {
+        return ESP_ERR_INVALID_STATE;
+    }
+    buffer = record_buffer;
+    total_records = num_records;
+    memset(buffer, 0, num_records * sizeof(heap_trace_record_t));
+    return ESP_OK;
+}
+
+esp_err_t heap_trace_start(heap_trace_mode_t mode_param)
+{
+    if (buffer == NULL || total_records == 0) {
+        return ESP_ERR_INVALID_STATE;
+    }
+
+    portENTER_CRITICAL(&trace_mux);
+
+    tracing = false;
+    mode = mode_param;
+    count = 0;
+    total_allocations = 0;
+    total_frees = 0;
+    has_overflowed = false;
+    heap_trace_resume();
+
+    portEXIT_CRITICAL(&trace_mux);
+    return ESP_OK;
+}
+
+static esp_err_t set_tracing(bool enable)
+{
+    if (tracing == enable) {
+        return ESP_ERR_INVALID_STATE;
+    }
+    tracing = enable;
+    return ESP_OK;
+}
+
+esp_err_t heap_trace_stop(void)
+{
+    return set_tracing(false);
+}
+
+esp_err_t heap_trace_resume(void)
+{
+    return set_tracing(true);
+}
+
+size_t heap_trace_get_count(void)
+{
+    return count;
+}
+
+esp_err_t heap_trace_get(size_t index, heap_trace_record_t *record)
+{
+    if (record == NULL) {
+        return ESP_ERR_INVALID_STATE;
+    }
+    esp_err_t result = ESP_OK;
+
+    portENTER_CRITICAL(&trace_mux);
+    if (index >= count) {
+        result = ESP_ERR_INVALID_ARG; /* out of range for 'count' */
+    } else {
+        memcpy(record, &buffer[index], sizeof(heap_trace_record_t));
+    }
+    portEXIT_CRITICAL(&trace_mux);
+    return result;
+}
+
+
+void heap_trace_dump(void)
+{
+    size_t delta_size = 0;
+    size_t delta_allocs = 0;
+    printf("%u allocations trace (%u entry buffer)\n",
+           count, total_records);
+    size_t start_count = count;
+    for (int i = 0; i < count; i++) {
+        heap_trace_record_t *rec = &buffer[i];
+
+        if (rec->address != NULL) {
+            printf("%d bytes (@ %p) allocated CPU %d ccount 0x%08x caller ",
+                   rec->size, rec->address, rec->ccount & 1, rec->ccount & ~3);
+            for (int j = 0; j < STACK_DEPTH && rec->alloced_by[j] != 0; j++) {
+                printf("%p%s", rec->alloced_by[j],
+                       (j < STACK_DEPTH - 1) ? ":" : "");
+            }
+
+            if (mode != HEAP_TRACE_ALL || STACK_DEPTH == 0 || rec->freed_by[0] == NULL) {
+                delta_size += rec->size;
+                delta_allocs++;
+                printf("\n");
+            } else {
+                printf("\nfreed by ");
+                for (int j = 0; j < STACK_DEPTH; j++) {
+                    printf("%p%s", rec->freed_by[j],
+                           (j < STACK_DEPTH - 1) ? ":" : "\n");
+                }
+            }
+        }
+    }
+    if (mode == HEAP_TRACE_ALL) {
+        printf("%u bytes alive in trace (%u/%u allocations)\n",
+               delta_size, delta_allocs, heap_trace_get_count());
+    } else {
+        printf("%u bytes 'leaked' in trace (%u allocations)\n", delta_size, delta_allocs);
+    }
+    printf("total allocations %u total frees %u\n", total_allocations, total_frees);
+    if (start_count != count) { // only a problem if trace isn't stopped before dumping
+        printf("(NB: New entries were traced while dumping, so trace dump may have duplicate entries.)\n");
+    }
+    if (has_overflowed) {
+        printf("(NB: Buffer has overflowed, so trace data is incomplete.)\n");
+    }
+}
+
+/* Add a new allocation to the heap trace records */
+static IRAM_ATTR void record_allocation(const heap_trace_record_t *record)
+{
+    if (!tracing || record->address == NULL) {
+        return;
+    }
+
+    portENTER_CRITICAL(&trace_mux);
+    if (tracing) {
+        if (count == total_records) {
+            has_overflowed = true;
+            /* Move the whole buffer back one slot.
+
+               This is a bit slow, compared to treating this buffer as a ringbuffer and rotating a head pointer.
+
+               However, ringbuffer code gets tricky when we remove elements in mid-buffer (for leak trace mode) while
+               trying to keep track of an item count that may overflow.
+            */
+            memmove(&buffer[0], &buffer[1], sizeof(heap_trace_record_t) * (total_records -1));
+            count--;
+        }
+        // Copy new record into place
+        memcpy(&buffer[count], record, sizeof(heap_trace_record_t));
+        count++;
+        total_allocations++;
+    }
+    portEXIT_CRITICAL(&trace_mux);
+}
+
+// remove a record, used when freeing
+static void remove_record(int index);
+
+/* record a free event in the heap trace log
+
+   For HEAP_TRACE_ALL, this means filling in the freed_by pointer.
+   For HEAP_TRACE_LEAKS, this means removing the record from the log.
+*/
+static IRAM_ATTR void record_free(void *p, void **callers)
+{
+    if (!tracing || p == NULL) {
+        return;
+    }
+
+    portENTER_CRITICAL(&trace_mux);
+    if (tracing && count > 0) {
+        total_frees++;
+        /* search backwards for the allocation record matching this free */
+        int i;
+        for (i = count - 1; i >= 0; i--) {
+            if (buffer[i].address == p) {
+                break;
+            }
+        }
+
+        if (i >= 0) {
+            if (mode == HEAP_TRACE_ALL) {
+                memcpy(buffer[i].freed_by, callers, sizeof(void *) * STACK_DEPTH);
+            } else { // HEAP_TRACE_LEAKS
+                // Leak trace mode, once an allocation is freed we remove it from the list
+                remove_record(i);
+            }
+        }
+    }
+    portEXIT_CRITICAL(&trace_mux);
+}
+
+/* remove the entry at 'index' from the ringbuffer of saved records */
+static IRAM_ATTR void remove_record(int index)
+{
+    if (index < count - 1) {
+        // Remove the buffer entry from the list
+        memmove(&buffer[index], &buffer[index+1],
+                sizeof(heap_trace_record_t) * (total_records - index - 1));
+    } else {
+        // For last element, just zero it out to avoid ambiguity
+        memset(&buffer[index], 0, sizeof(heap_trace_record_t));
+    }
+    count--;
+}
+
+#include "heap_trace.inc"
+
+#endif /*CONFIG_HEAP_TRACING_STANDALONE*/

components/heap/include/esp_heap_caps.h

@@ -0,0 +1,402 @@
+// Copyright 2015-2016 Espressif Systems (Shanghai) PTE LTD
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+#pragma once
+
+#include <stdint.h>
+#include <stdlib.h>
+#include "multi_heap.h"
+#include <sdkconfig.h>
+#include "esp_err.h"
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+/**
+ * @brief Flags to indicate the capabilities of the various memory systems
+ */
+#define MALLOC_CAP_EXEC             (1<<0)  ///< Memory must be able to run executable code
+#define MALLOC_CAP_32BIT            (1<<1)  ///< Memory must allow for aligned 32-bit data accesses
+#define MALLOC_CAP_8BIT             (1<<2)  ///< Memory must allow for 8/16/...-bit data accesses
+#define MALLOC_CAP_DMA              (1<<3)  ///< Memory must be able to accessed by DMA
+#define MALLOC_CAP_PID2             (1<<4)  ///< Memory must be mapped to PID2 memory space (PIDs are not currently used)
+#define MALLOC_CAP_PID3             (1<<5)  ///< Memory must be mapped to PID3 memory space (PIDs are not currently used)
+#define MALLOC_CAP_PID4             (1<<6)  ///< Memory must be mapped to PID4 memory space (PIDs are not currently used)
+#define MALLOC_CAP_PID5             (1<<7)  ///< Memory must be mapped to PID5 memory space (PIDs are not currently used)
+#define MALLOC_CAP_PID6             (1<<8)  ///< Memory must be mapped to PID6 memory space (PIDs are not currently used)
+#define MALLOC_CAP_PID7             (1<<9)  ///< Memory must be mapped to PID7 memory space (PIDs are not currently used)
+#define MALLOC_CAP_SPIRAM           (1<<10) ///< Memory must be in SPI RAM
+#define MALLOC_CAP_INTERNAL         (1<<11) ///< Memory must be internal; specifically it should not disappear when flash/spiram cache is switched off
+#define MALLOC_CAP_DEFAULT          (1<<12) ///< Memory can be returned in a non-capability-specific memory allocation (e.g. malloc(), calloc()) call
+#define MALLOC_CAP_IRAM_8BIT        (1<<13) ///< Memory must be in IRAM and allow unaligned access
+#define MALLOC_CAP_RETENTION        (1<<14)
+
+#define MALLOC_CAP_INVALID          (1<<31) ///< Memory can't be used / list end marker
+
+/**
+ * @brief callback called when a allocation operation fails, if registered
+ * @param size in bytes of failed allocation
+ * @param caps capabillites requested of failed allocation
+ * @param function_name function which generated the failure
+ */
+typedef void (*esp_alloc_failed_hook_t) (size_t size, uint32_t caps, const char * function_name);
+
+/**
+ * @brief registers a callback function to be invoked if a memory allocation operation fails
+ * @param callback caller defined callback to be invoked
+ * @return ESP_OK if callback was registered.
+ */
+esp_err_t heap_caps_register_failed_alloc_callback(esp_alloc_failed_hook_t callback);
+
+/**
+ * @brief Allocate a chunk of memory which has the given capabilities
+ *
+ * Equivalent semantics to libc malloc(), for capability-aware memory.
+ *
+ * In IDF, ``malloc(p)`` is equivalent to ``heap_caps_malloc(p, MALLOC_CAP_8BIT)``.
+ *
+ * @param size Size, in bytes, of the amount of memory to allocate
+ * @param caps        Bitwise OR of MALLOC_CAP_* flags indicating the type
+ *                    of memory to be returned
+ *
+ * @return A pointer to the memory allocated on success, NULL on failure
+ */
+void *heap_caps_malloc(size_t size, uint32_t caps);
+
+
+/**
+ * @brief Free memory previously allocated via heap_caps_malloc() or heap_caps_realloc().
+ *
+ * Equivalent semantics to libc free(), for capability-aware memory.
+ *
+ *  In IDF, ``free(p)`` is equivalent to ``heap_caps_free(p)``.
+ *
+ * @param ptr Pointer to memory previously returned from heap_caps_malloc() or heap_caps_realloc(). Can be NULL.
+ */
+void heap_caps_free( void *ptr);
+
+/**
+ * @brief Reallocate memory previously allocated via heap_caps_malloc() or heap_caps_realloc().
+ *
+ * Equivalent semantics to libc realloc(), for capability-aware memory.
+ *
+ * In IDF, ``realloc(p, s)`` is equivalent to ``heap_caps_realloc(p, s, MALLOC_CAP_8BIT)``.
+ *
+ * 'caps' parameter can be different to the capabilities that any original 'ptr' was allocated with. In this way,
+ * realloc can be used to "move" a buffer if necessary to ensure it meets a new set of capabilities.
+ *
+ * @param ptr Pointer to previously allocated memory, or NULL for a new allocation.
+ * @param size Size of the new buffer requested, or 0 to free the buffer.
+ * @param caps        Bitwise OR of MALLOC_CAP_* flags indicating the type
+ *                    of memory desired for the new allocation.
+ *
+ * @return Pointer to a new buffer of size 'size' with capabilities 'caps', or NULL if allocation failed.
+ */
+void *heap_caps_realloc( void *ptr, size_t size, uint32_t caps);
+
+/**
+ * @brief Allocate a aligned chunk of memory which has the given capabilities
+ *
+ * Equivalent semantics to libc aligned_alloc(), for capability-aware memory.
+ * @param alignment  How the pointer received needs to be aligned
+ *                   must be a power of two
+ * @param size Size, in bytes, of the amount of memory to allocate
+ * @param caps        Bitwise OR of MALLOC_CAP_* flags indicating the type
+ *                    of memory to be returned
+ *
+ * @return A pointer to the memory allocated on success, NULL on failure
+ *
+ *
+ */
+void *heap_caps_aligned_alloc(size_t alignment, size_t size, uint32_t caps);
+
+/**
+ * @brief Used to deallocate memory previously allocated with heap_caps_aligned_alloc
+ *
+ * @param ptr Pointer to the memory allocated
+ * @note This function is deprecated, plase consider using heap_caps_free() instead
+ */
+void __attribute__((deprecated))  heap_caps_aligned_free(void *ptr);
+
+/**
+ * @brief Allocate a aligned chunk of memory which has the given capabilities. The initialized value in the memory is set to zero.
+ *
+ * @param alignment  How the pointer received needs to be aligned
+ *                   must be a power of two
+ * @param n    Number of continuing chunks of memory to allocate
+ * @param size Size, in bytes, of a chunk of memory to allocate
+ * @param caps        Bitwise OR of MALLOC_CAP_* flags indicating the type
+ *                    of memory to be returned
+ *
+ * @return A pointer to the memory allocated on success, NULL on failure
+ *
+ */
+void *heap_caps_aligned_calloc(size_t alignment, size_t n, size_t size, uint32_t caps);
+
+
+/**
+ * @brief Allocate a chunk of memory which has the given capabilities. The initialized value in the memory is set to zero.
+ *
+ * Equivalent semantics to libc calloc(), for capability-aware memory.
+ *
+ * In IDF, ``calloc(p)`` is equivalent to ``heap_caps_calloc(p, MALLOC_CAP_8BIT)``.
+ *
+ * @param n    Number of continuing chunks of memory to allocate
+ * @param size Size, in bytes, of a chunk of memory to allocate
+ * @param caps        Bitwise OR of MALLOC_CAP_* flags indicating the type
+ *                    of memory to be returned
+ *
+ * @return A pointer to the memory allocated on success, NULL on failure
+ */
+void *heap_caps_calloc(size_t n, size_t size, uint32_t caps);
+
+/**
+ * @brief Get the total size of all the regions that have the given capabilities
+ *
+ * This function takes all regions capable of having the given capabilities allocated in them
+ * and adds up the total space they have.
+ *
+ * @param caps        Bitwise OR of MALLOC_CAP_* flags indicating the type
+ *                    of memory
+ *
+ * @return total size in bytes
+ */
+
+size_t heap_caps_get_total_size(uint32_t caps);
+
+/**
+ * @brief Get the total free size of all the regions that have the given capabilities
+ *
+ * This function takes all regions capable of having the given capabilities allocated in them
+ * and adds up the free space they have.
+ *
+ * Note that because of heap fragmentation it is probably not possible to allocate a single block of memory
+ * of this size. Use heap_caps_get_largest_free_block() for this purpose.
+
+ * @param caps        Bitwise OR of MALLOC_CAP_* flags indicating the type
+ *                    of memory
+ *
+ * @return Amount of free bytes in the regions
+ */
+size_t heap_caps_get_free_size( uint32_t caps );
+
+
+/**
+ * @brief Get the total minimum free memory of all regions with the given capabilities
+ *
+ * This adds all the low water marks of the regions capable of delivering the memory
+ * with the given capabilities.
+ *
+ * Note the result may be less than the global all-time minimum available heap of this kind, as "low water marks" are
+ * tracked per-region. Individual regions' heaps may have reached their "low water marks" at different points in time. However
+ * this result still gives a "worst case" indication for all-time minimum free heap.
+ *
+ * @param caps        Bitwise OR of MALLOC_CAP_* flags indicating the type
+ *                    of memory
+ *
+ * @return Amount of free bytes in the regions
+ */
+size_t heap_caps_get_minimum_free_size( uint32_t caps );
+
+/**
+ * @brief Get the largest free block of memory able to be allocated with the given capabilities.
+ *
+ * Returns the largest value of ``s`` for which ``heap_caps_malloc(s, caps)`` will succeed.
+ *
+ * @param caps        Bitwise OR of MALLOC_CAP_* flags indicating the type
+ *                    of memory
+ *
+ * @return Size of largest free block in bytes.
+ */
+size_t heap_caps_get_largest_free_block( uint32_t caps );
+
+
+/**
+ * @brief Get heap info for all regions with the given capabilities.
+ *
+ * Calls multi_heap_info() on all heaps which share the given capabilities.  The information returned is an aggregate
+ * across all matching heaps.  The meanings of fields are the same as defined for multi_heap_info_t, except that
+ * ``minimum_free_bytes`` has the same caveats described in heap_caps_get_minimum_free_size().
+ *
+ * @param info        Pointer to a structure which will be filled with relevant
+ *                    heap metadata.
+ * @param caps        Bitwise OR of MALLOC_CAP_* flags indicating the type
+ *                    of memory
+ *
+ */
+void heap_caps_get_info( multi_heap_info_t *info, uint32_t caps );
+
+
+/**
+ * @brief Print a summary of all memory with the given capabilities.
+ *
+ * Calls multi_heap_info on all heaps which share the given capabilities, and
+ * prints a two-line summary for each, then a total summary.
+ *
+ * @param caps        Bitwise OR of MALLOC_CAP_* flags indicating the type
+ *                    of memory
+ *
+ */
+void heap_caps_print_heap_info( uint32_t caps );
+
+/**
+ * @brief Check integrity of all heap memory in the system.
+ *
+ * Calls multi_heap_check on all heaps. Optionally print errors if heaps are corrupt.
+ *
+ * Calling this function is equivalent to calling heap_caps_check_integrity
+ * with the caps argument set to MALLOC_CAP_INVALID.
+ *
+ * @param print_errors Print specific errors if heap corruption is found.
+ *
+ * @return True if all heaps are valid, False if at least one heap is corrupt.
+ */
+bool heap_caps_check_integrity_all(bool print_errors);
+
+/**
+ * @brief Check integrity of all heaps with the given capabilities.
+ *
+ * Calls multi_heap_check on all heaps which share the given capabilities. Optionally
+ * print errors if the heaps are corrupt.
+ *
+ * See also heap_caps_check_integrity_all to check all heap memory
+ * in the system and heap_caps_check_integrity_addr to check memory
+ * around a single address.
+ *
+ * @param caps        Bitwise OR of MALLOC_CAP_* flags indicating the type
+ *                    of memory
+ * @param print_errors Print specific errors if heap corruption is found.
+ *
+ * @return True if all heaps are valid, False if at least one heap is corrupt.
+ */
+bool heap_caps_check_integrity(uint32_t caps, bool print_errors);
+
+/**
+ * @brief Check integrity of heap memory around a given address.
+ *
+ * This function can be used to check the integrity of a single region of heap memory,
+ * which contains the given address.
+ *
+ * This can be useful if debugging heap integrity for corruption at a known address,
+ * as it has a lower overhead than checking all heap regions. Note that if the corrupt
+ * address moves around between runs (due to timing or other factors) then this approach
+ * won't work and you should call heap_caps_check_integrity or
+ * heap_caps_check_integrity_all instead.
+ *
+ * @note The entire heap region around the address is checked, not only the adjacent
+ * heap blocks.
+ *
+ * @param addr Address in memory. Check for corruption in region containing this address.
+ * @param print_errors Print specific errors if heap corruption is found.
+ *
+ * @return True if the heap containing the specified address is valid,
+ * False if at least one heap is corrupt or the address doesn't belong to a heap region.
+ */
+bool heap_caps_check_integrity_addr(intptr_t addr, bool print_errors);
+
+/**
+ * @brief Enable malloc() in external memory and set limit below which
+ *        malloc() attempts are placed in internal memory.
+ *
+ * When external memory is in use, the allocation strategy is to initially try to
+ * satisfy smaller allocation requests with internal memory and larger requests
+ * with external memory. This sets the limit between the two, as well as generally
+ * enabling allocation in external memory.
+ *
+ * @param limit       Limit, in bytes.
+ */
+void heap_caps_malloc_extmem_enable(size_t limit);
+
+/**
+ * @brief Allocate a chunk of memory as preference in decreasing order.
+ *
+ * @attention The variable parameters are bitwise OR of MALLOC_CAP_* flags indicating the type of memory.
+ *            This API prefers to allocate memory with the first parameter. If failed, allocate memory with
+ *            the next parameter. It will try in this order until allocating a chunk of memory successfully
+ *            or fail to allocate memories with any of the parameters.
+ *
+ * @param size Size, in bytes, of the amount of memory to allocate
+ * @param num Number of variable paramters
+ *
+ * @return A pointer to the memory allocated on success, NULL on failure
+ */
+void *heap_caps_malloc_prefer( size_t size, size_t num, ... );
+
+/**
+ * @brief Allocate a chunk of memory as preference in decreasing order.
+ *
+ * @param ptr Pointer to previously allocated memory, or NULL for a new allocation.
+ * @param size Size of the new buffer requested, or 0 to free the buffer.
+ * @param num Number of variable paramters
+ *
+ * @return Pointer to a new buffer of size 'size', or NULL if allocation failed.
+ */
+void *heap_caps_realloc_prefer( void *ptr, size_t size, size_t num, ... );
+
+/**
+ * @brief Allocate a chunk of memory as preference in decreasing order.
+ *
+ * @param n    Number of continuing chunks of memory to allocate
+ * @param size Size, in bytes, of a chunk of memory to allocate
+ * @param num  Number of variable paramters
+ *
+ * @return A pointer to the memory allocated on success, NULL on failure
+ */
+void *heap_caps_calloc_prefer( size_t n, size_t size, size_t num, ... );
+
+/**
+ * @brief Dump the full structure of all heaps with matching capabilities.
+ *
+ * Prints a large amount of output to serial (because of locking limitations,
+ * the output bypasses stdout/stderr). For each (variable sized) block
+ * in each matching heap, the following output is printed on a single line:
+ *
+ * - Block address (the data buffer returned by malloc is 4 bytes after this
+ *   if heap debugging is set to Basic, or 8 bytes otherwise).
+ * - Data size (the data size may be larger than the size requested by malloc,
+ *   either due to heap fragmentation or because of heap debugging level).
+ * - Address of next block in the heap.
+ * - If the block is free, the address of the next free block is also printed.
+ *
+ * @param caps        Bitwise OR of MALLOC_CAP_* flags indicating the type
+ *                    of memory
+ */
+void heap_caps_dump(uint32_t caps);
+
+/**
+ * @brief Dump the full structure of all heaps.
+ *
+ * Covers all registered heaps. Prints a large amount of output to serial.
+ *
+ * Output is the same as for heap_caps_dump.
+ *
+ */
+void heap_caps_dump_all(void);
+
+/**
+ * @brief Return the size that a particular pointer was allocated with.
+ *
+ * @param ptr Pointer to currently allocated heap memory. Must be a pointer value previously
+ * returned by heap_caps_malloc,malloc,calloc, etc. and not yet freed.
+ *
+ * @note The app will crash with an assertion failure if the pointer is not valid.
+ *
+ * @return Size of the memory allocated at this block.
+ *
+ */
+size_t heap_caps_get_allocated_size( void *ptr );
+
+#ifdef __cplusplus
+}
+#endif

components/heap/include/esp_heap_caps_init.h

@@ -0,0 +1,92 @@
+// Copyright 2017 Espressif Systems (Shanghai) PTE LTD
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+#pragma once
+
+#include "esp_err.h"
+#include "esp_heap_caps.h"
+#include "soc/soc_memory_layout.h"
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+/**
+ * @brief Initialize the capability-aware heap allocator.
+ *
+ * This is called once in the IDF startup code. Do not call it
+ * at other times.
+ */
+void heap_caps_init(void);
+
+/**
+ * @brief Enable heap(s) in memory regions where the startup stacks are located.
+ *
+ * On startup, the pro/app CPUs have a certain memory region they use as stack, so we
+ * cannot do allocations in the regions these stack frames are. When FreeRTOS is
+ * completely started, they do not use that memory anymore and heap(s) there can
+ * be enabled.
+ */
+void heap_caps_enable_nonos_stack_heaps(void);
+
+/**
+ * @brief Add a region of memory to the collection of heaps at runtime.
+ *
+ * Most memory regions are defined in soc_memory_layout.c for the SoC,
+ * and are registered via heap_caps_init(). Some regions can't be used
+ * immediately and are later enabled via heap_caps_enable_nonos_stack_heaps().
+ *
+ * Call this function to add a region of memory to the heap at some later time.
+ *
+ * This function does not consider any of the "reserved" regions or other data in soc_memory_layout, caller needs to
+ * consider this themselves.
+ *
+ * All memory within the region specified by start & end parameters must be otherwise unused.
+ *
+ * The capabilities of the newly registered memory will be determined by the start address, as looked up in the regions
+ * specified in soc_memory_layout.c.
+ *
+ * Use heap_caps_add_region_with_caps() to register a region with custom capabilities.
+ *
+ * @param start Start address of new region.
+ * @param end End address of new region.
+ *
+ * @return ESP_OK on success, ESP_ERR_INVALID_ARG if a parameter is invalid, ESP_ERR_NOT_FOUND if the
+ * specified start address doesn't reside in a known region, or any error returned by heap_caps_add_region_with_caps().
+ */
+esp_err_t heap_caps_add_region(intptr_t start, intptr_t end);
+
+
+/**
+ * @brief Add a region of memory to the collection of heaps at runtime, with custom capabilities.
+ *
+ * Similar to heap_caps_add_region(), only custom memory capabilities are specified by the caller.
+ *
+ * @param caps Ordered array of capability masks for the new region, in order of priority. Must have length
+ * SOC_MEMORY_TYPE_NO_PRIOS. Does not need to remain valid after the call returns.
+ * @param start Start address of new region.
+ * @param end End address of new region.
+ *
+ * @return
+ *         - ESP_OK on success
+ *         - ESP_ERR_INVALID_ARG if a parameter is invalid
+ *         - ESP_ERR_NO_MEM if no memory to register new heap.
+ *         - ESP_ERR_INVALID_SIZE if the memory region is too small to fit a heap
+ *         - ESP_FAIL if region overlaps the start and/or end of an existing region
+ */
+esp_err_t heap_caps_add_region_with_caps(const uint32_t caps[], intptr_t start, intptr_t end);
+
+
+#ifdef __cplusplus
+}
+#endif

components/heap/include/esp_heap_task_info.h

@@ -0,0 +1,98 @@
+// Copyright 2018 Espressif Systems (Shanghai) PTE LTD
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+#pragma once
+
+#ifdef CONFIG_HEAP_TASK_TRACKING
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+// This macro controls how much space is provided for partitioning the per-task
+// heap allocation info according to one or more sets of heap capabilities.
+#define NUM_HEAP_TASK_CAPS 4
+
+/** @brief Structure to collect per-task heap allocation totals partitioned by selected caps */
+typedef struct {
+    TaskHandle_t task;                ///< Task to which these totals belong
+    size_t size[NUM_HEAP_TASK_CAPS];  ///< Total allocations partitioned by selected caps
+    size_t count[NUM_HEAP_TASK_CAPS]; ///< Number of blocks partitioned by selected caps
+} heap_task_totals_t;
+
+/** @brief Structure providing details about a block allocated by a task */
+typedef struct {
+    TaskHandle_t task;                ///< Task that allocated the block
+    void *address;                    ///< User address of allocated block
+    uint32_t size;                    ///< Size of the allocated block
+} heap_task_block_t;
+
+/** @brief Structure to provide parameters to heap_caps_get_per_task_info
+ *
+ * The 'caps' and 'mask' arrays allow partitioning the per-task heap allocation
+ * totals by selected sets of heap region capabilities so that totals for
+ * multiple regions can be accumulated in one scan.  The capabilities flags for
+ * each region ANDed with mask[i] are compared to caps[i] in order; the
+ * allocations in that region are added to totals->size[i] and totals->count[i]
+ * for the first i that matches.  To collect the totals without any
+ * partitioning, set mask[0] and caps[0] both to zero.  The allocation totals
+ * are returned in the 'totals' array of heap_task_totals_t structs.  To allow
+ * easily comparing the totals array between consecutive calls, that array can
+ * be left populated from one call to the next so the order of tasks is the
+ * same even if some tasks have freed their blocks or have been deleted.  The
+ * number of blocks prepopulated is given by num_totals, which is updated upon
+ * return.  If there are more tasks with allocations than the capacity of the
+ * totals array (given by max_totals), information for the excess tasks will be
+ * not be collected.  The totals array pointer can be NULL if the totals are
+ * not desired.
+ *
+ * The 'tasks' array holds a list of handles for tasks whose block details are
+ * to be returned in the 'blocks' array of heap_task_block_t structs.  If the
+ * tasks array pointer is NULL, block details for all tasks will be returned up
+ * to the capacity of the buffer array, given by max_blocks.  The function
+ * return value tells the number of blocks filled into the array.  The blocks
+ * array pointer can be NULL if block details are not desired, or max_blocks
+ * can be set to zero.
+ */
+typedef struct {
+    int32_t caps[NUM_HEAP_TASK_CAPS]; ///< Array of caps for partitioning task totals
+    int32_t mask[NUM_HEAP_TASK_CAPS]; ///< Array of masks under which caps must match
+    TaskHandle_t *tasks;              ///< Array of tasks whose block info is returned
+    size_t num_tasks;                 ///< Length of tasks array
+    heap_task_totals_t *totals;       ///< Array of structs to collect task totals
+    size_t *num_totals;               ///< Number of task structs currently in array
+    size_t max_totals;                ///< Capacity of array of task totals structs
+    heap_task_block_t *blocks;        ///< Array of task block details structs
+    size_t max_blocks;                ///< Capacity of array of task block info structs
+} heap_task_info_params_t;
+
+/**
+ * @brief Return per-task heap allocation totals and lists of blocks.
+ *
+ * For each task that has allocated memory from the heap, return totals for
+ * allocations within regions matching one or more sets of capabilities.
+ *
+ * Optionally also return an array of structs providing details about each
+ * block allocated by one or more requested tasks, or by all tasks.
+ *
+ * @param params Structure to hold all the parameters for the function
+ * (@see heap_task_info_params_t).
+ * @return Number of block detail structs returned (@see heap_task_block_t).
+ */
+extern size_t heap_caps_get_per_task_info(heap_task_info_params_t *params);
+
+#ifdef __cplusplus
+}
+#endif
+
+#endif // CONFIG_HEAP_TASK_TRACKING

components/heap/include/esp_heap_trace.h

@@ -0,0 +1,154 @@
+// Copyright 2015-2016 Espressif Systems (Shanghai) PTE LTD
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+#pragma once
+
+#include "sdkconfig.h"
+#include <stdint.h>
+#include <esp_err.h>
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+#if !defined(CONFIG_HEAP_TRACING) && !defined(HEAP_TRACE_SRCFILE)
+#warning "esp_heap_trace.h is included but heap tracing is disabled in menuconfig, functions are no-ops"
+#endif
+
+#ifndef CONFIG_HEAP_TRACING_STACK_DEPTH
+#define CONFIG_HEAP_TRACING_STACK_DEPTH 0
+#endif
+
+typedef enum {
+    HEAP_TRACE_ALL,
+    HEAP_TRACE_LEAKS,
+} heap_trace_mode_t;
+
+/**
+ * @brief Trace record data type. Stores information about an allocated region of memory.
+ */
+typedef struct {
+    uint32_t ccount; ///< CCOUNT of the CPU when the allocation was made. LSB (bit value 1) is the CPU number (0 or 1).
+    void *address;   ///< Address which was allocated
+    size_t size;     ///< Size of the allocation
+    void *alloced_by[CONFIG_HEAP_TRACING_STACK_DEPTH]; ///< Call stack of the caller which allocated the memory.
+    void *freed_by[CONFIG_HEAP_TRACING_STACK_DEPTH];   ///< Call stack of the caller which freed the memory (all zero if not freed.)
+} heap_trace_record_t;
+
+/**
+ * @brief Initialise heap tracing in standalone mode.
+ *
+ * This function must be called before any other heap tracing functions.
+ *
+ * To disable heap tracing and allow the buffer to be freed, stop tracing and then call heap_trace_init_standalone(NULL, 0);
+ *
+ * @param record_buffer Provide a buffer to use for heap trace data. Must remain valid any time heap tracing is enabled, meaning
+ * it must be allocated from internal memory not in PSRAM.
+ * @param num_records Size of the heap trace buffer, as number of record structures.
+ * @return
+ *  - ESP_ERR_NOT_SUPPORTED Project was compiled without heap tracing enabled in menuconfig.
+ *  - ESP_ERR_INVALID_STATE Heap tracing is currently in progress.
+ *  - ESP_OK Heap tracing initialised successfully.
+ */
+esp_err_t heap_trace_init_standalone(heap_trace_record_t *record_buffer, size_t num_records);
+
+/**
+ * @brief Initialise heap tracing in host-based mode.
+ *
+ * This function must be called before any other heap tracing functions.
+ *
+ * @return
+ *  - ESP_ERR_INVALID_STATE Heap tracing is currently in progress.
+ *  - ESP_OK Heap tracing initialised successfully.
+ */
+esp_err_t heap_trace_init_tohost(void);
+
+/**
+ * @brief Start heap tracing. All heap allocations & frees will be traced, until heap_trace_stop() is called.
+ *
+ * @note heap_trace_init_standalone() must be called to provide a valid buffer, before this function is called.
+ *
+ * @note Calling this function while heap tracing is running will reset the heap trace state and continue tracing.
+ *
+ * @param mode Mode for tracing.
+ * - HEAP_TRACE_ALL means all heap allocations and frees are traced.
+ * - HEAP_TRACE_LEAKS means only suspected memory leaks are traced. (When memory is freed, the record is removed from the trace buffer.)
+ * @return
+ * - ESP_ERR_NOT_SUPPORTED Project was compiled without heap tracing enabled in menuconfig.
+ * - ESP_ERR_INVALID_STATE A non-zero-length buffer has not been set via heap_trace_init_standalone().
+ * - ESP_OK Tracing is started.
+ */
+esp_err_t heap_trace_start(heap_trace_mode_t mode);
+
+/**
+ * @brief Stop heap tracing.
+ *
+ * @return
+ * - ESP_ERR_NOT_SUPPORTED Project was compiled without heap tracing enabled in menuconfig.
+ * - ESP_ERR_INVALID_STATE Heap tracing was not in progress.
+ * - ESP_OK Heap tracing stopped..
+ */
+esp_err_t heap_trace_stop(void);
+
+/**
+ * @brief Resume heap tracing which was previously stopped.
+ *
+ * Unlike heap_trace_start(), this function does not clear the
+ * buffer of any pre-existing trace records.
+ *
+ * The heap trace mode is the same as when heap_trace_start() was
+ * last called (or HEAP_TRACE_ALL if heap_trace_start() was never called).
+ *
+ * @return
+ * - ESP_ERR_NOT_SUPPORTED Project was compiled without heap tracing enabled in menuconfig.
+ * - ESP_ERR_INVALID_STATE Heap tracing was already started.
+ * - ESP_OK Heap tracing resumed.
+ */
+esp_err_t heap_trace_resume(void);
+
+/**
+ * @brief Return number of records in the heap trace buffer
+ *
+ * It is safe to call this function while heap tracing is running.
+ */
+size_t heap_trace_get_count(void);
+
+/**
+ * @brief Return a raw record from the heap trace buffer
+ *
+ * @note It is safe to call this function while heap tracing is running, however in HEAP_TRACE_LEAK mode record indexing may
+ * skip entries unless heap tracing is stopped first.
+ *
+ * @param index Index (zero-based) of the record to return.
+ * @param[out] record Record where the heap trace record will be copied.
+ * @return
+ * - ESP_ERR_NOT_SUPPORTED Project was compiled without heap tracing enabled in menuconfig.
+ * - ESP_ERR_INVALID_STATE Heap tracing was not initialised.
+ * - ESP_ERR_INVALID_ARG Index is out of bounds for current heap trace record count.
+ * - ESP_OK Record returned successfully.
+ */
+esp_err_t heap_trace_get(size_t index, heap_trace_record_t *record);
+
+/**
+ * @brief Dump heap trace record data to stdout
+ *
+ * @note It is safe to call this function while heap tracing is running, however in HEAP_TRACE_LEAK mode the dump may skip
+ * entries unless heap tracing is stopped first.
+ *
+ *
+ */
+void heap_trace_dump(void);
+
+#ifdef __cplusplus
+}
+#endif

components/heap/include/heap_trace.inc

@@ -0,0 +1,200 @@
+// Copyright 2015-2016 Espressif Systems (Shanghai) PTE LTD
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+#include <string.h>
+#include <sdkconfig.h>
+#include "soc/soc_memory_layout.h"
+#include "esp_attr.h"
+
+/* Encode the CPU ID in the LSB of the ccount value */
+inline static uint32_t get_ccount(void)
+{
+    uint32_t ccount = cpu_hal_get_cycle_count() & ~3;
+#ifndef CONFIG_FREERTOS_UNICORE
+    ccount |= xPortGetCoreID();
+#endif
+    return ccount;
+}
+
+/* Architecture-specific return value of __builtin_return_address which
+ * should be interpreted as an invalid address.
+ */
+#ifdef __XTENSA__
+#define HEAP_ARCH_INVALID_PC  0x40000000
+#else
+#define HEAP_ARCH_INVALID_PC  0x00000000
+#endif
+
+// Caller is 2 stack frames deeper than we care about
+#define STACK_OFFSET  2
+
+#define TEST_STACK(N) do {                                              \
+        if (STACK_DEPTH == N) {                                         \
+            return;                                                     \
+        }                                                               \
+        callers[N] = __builtin_return_address(N+STACK_OFFSET);          \
+        if (!esp_ptr_executable(callers[N])                             \
+            || callers[N] == (void*) HEAP_ARCH_INVALID_PC) {            \
+            callers[N] = 0;                                             \
+            return;                                                     \
+        }                                                               \
+    } while(0)
+
+/* Static function to read the call stack for a traced heap call.
+
+   Calls to __builtin_return_address are "unrolled" via TEST_STACK macro as gcc requires the
+   argument to be a compile-time constant.
+*/
+static IRAM_ATTR __attribute__((noinline)) void get_call_stack(void **callers)
+{
+    bzero(callers, sizeof(void *) * STACK_DEPTH);
+    TEST_STACK(0);
+    TEST_STACK(1);
+    TEST_STACK(2);
+    TEST_STACK(3);
+    TEST_STACK(4);
+    TEST_STACK(5);
+    TEST_STACK(6);
+    TEST_STACK(7);
+    TEST_STACK(8);
+    TEST_STACK(9);
+}
+
+_Static_assert(STACK_DEPTH >= 0 && STACK_DEPTH <= 10, "CONFIG_HEAP_TRACING_STACK_DEPTH must be in range 0-10");
+
+
+typedef enum {
+    TRACE_MALLOC_CAPS,
+    TRACE_MALLOC_DEFAULT
+} trace_malloc_mode_t;
+
+
+void *__real_heap_caps_malloc(size_t size, uint32_t caps);
+void *__real_heap_caps_malloc_default( size_t size );
+void *__real_heap_caps_realloc_default( void *ptr, size_t size );
+
+/* trace any 'malloc' event */
+static IRAM_ATTR __attribute__((noinline)) void *trace_malloc(size_t size, uint32_t caps, trace_malloc_mode_t mode)
+{
+    uint32_t ccount = get_ccount();
+    void *p;
+
+    if ( mode == TRACE_MALLOC_CAPS ) {
+        p = __real_heap_caps_malloc(size, caps);
+    } else { //TRACE_MALLOC_DEFAULT
+        p = __real_heap_caps_malloc_default(size);
+    }
+
+    heap_trace_record_t rec = {
+        .address = p,
+        .ccount = ccount,
+        .size = size,
+    };
+    get_call_stack(rec.alloced_by);
+    record_allocation(&rec);
+    return p;
+}
+
+void __real_heap_caps_free(void *p);
+
+/* trace any 'free' event */
+static IRAM_ATTR __attribute__((noinline)) void trace_free(void *p)
+{
+    void *callers[STACK_DEPTH];
+    get_call_stack(callers);
+    record_free(p, callers);
+
+    __real_heap_caps_free(p);
+}
+
+void * __real_heap_caps_realloc(void *p, size_t size, uint32_t caps);
+
+/* trace any 'realloc' event */
+static IRAM_ATTR __attribute__((noinline)) void *trace_realloc(void *p, size_t size, uint32_t caps, trace_malloc_mode_t mode)
+{
+    void *callers[STACK_DEPTH];
+    uint32_t ccount = get_ccount();
+    void *r;
+
+    /* trace realloc as free-then-alloc */
+    get_call_stack(callers);
+    record_free(p, callers);
+
+    if (mode == TRACE_MALLOC_CAPS ) {
+        r = __real_heap_caps_realloc(p, size, caps);
+    } else { //TRACE_MALLOC_DEFAULT
+        r = __real_heap_caps_realloc_default(p, size);
+    }
+    /* realloc with zero size is a free */
+    if (size != 0) {
+        heap_trace_record_t rec = {
+            .address = r,
+            .ccount = ccount,
+            .size = size,
+        };
+        memcpy(rec.alloced_by, callers, sizeof(void *) * STACK_DEPTH);
+        record_allocation(&rec);
+    }
+    return r;
+}
+
+/* Note: this changes the behaviour of libc malloc/realloc/free a bit,
+   as they no longer go via the libc functions in ROM. But more or less
+   the same in the end. */
+
+IRAM_ATTR void *__wrap_malloc(size_t size)
+{
+    return trace_malloc(size, 0, TRACE_MALLOC_DEFAULT);
+}
+
+IRAM_ATTR void __wrap_free(void *p)
+{
+    trace_free(p);
+}
+
+IRAM_ATTR void *__wrap_realloc(void *p, size_t size)
+{
+    return trace_realloc(p, size, 0, TRACE_MALLOC_DEFAULT);
+}
+
+IRAM_ATTR void *__wrap_calloc(size_t nmemb, size_t size)
+{
+    size = size * nmemb;
+    void *result = trace_malloc(size, 0, TRACE_MALLOC_DEFAULT);
+    if (result != NULL) {
+        memset(result, 0, size);
+    }
+    return result;
+}
+
+IRAM_ATTR void *__wrap_heap_caps_malloc(size_t size, uint32_t caps)
+{
+    return trace_malloc(size, caps, TRACE_MALLOC_CAPS);
+}
+
+void __wrap_heap_caps_free(void *p) __attribute__((alias("__wrap_free")));
+
+IRAM_ATTR void *__wrap_heap_caps_realloc(void *p, size_t size, uint32_t caps)
+{
+    return trace_realloc(p, size, caps, TRACE_MALLOC_CAPS);
+}
+
+IRAM_ATTR void *__wrap_heap_caps_malloc_default( size_t size )
+{
+    return trace_malloc(size, 0, TRACE_MALLOC_DEFAULT);
+}
+
+IRAM_ATTR void *__wrap_heap_caps_realloc_default( void *ptr, size_t size )
+{
+    return trace_realloc(ptr, size, 0, TRACE_MALLOC_DEFAULT);
+}

components/heap/include/multi_heap.h

@@ -0,0 +1,190 @@
+// Copyright 2015-2016 Espressif Systems (Shanghai) PTE LTD
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+#pragma once
+#include <stdint.h>
+#include <stdlib.h>
+#include <stdbool.h>
+
+/* multi_heap is a heap implementation for handling multiple
+   heterogenous heaps in a single program.
+
+   Any contiguous block of memory can be registered as a heap.
+*/
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+/** @brief Opaque handle to a registered heap */
+typedef struct multi_heap_info *multi_heap_handle_t;
+
+/**
+ * @brief allocate a chunk of memory with specific alignment
+ *
+ * @param heap  Handle to a registered heap.
+ * @param size  size in bytes of memory chunk
+ * @param alignment  how the memory must be aligned
+ *
+ * @return pointer to the memory allocated, NULL on failure
+ */
+void *multi_heap_aligned_alloc(multi_heap_handle_t heap, size_t size, size_t alignment);
+
+/** @brief malloc() a buffer in a given heap
+ *
+ * Semantics are the same as standard malloc(), only the returned buffer will be allocated in the specified heap.
+ *
+ * @param heap Handle to a registered heap.
+ * @param size Size of desired buffer.
+ *
+ * @return Pointer to new memory, or NULL if allocation fails.
+ */
+void *multi_heap_malloc(multi_heap_handle_t heap, size_t size);
+
+/** @brief free() a buffer aligned in a given heap.
+ *
+ * @param heap Handle to a registered heap.
+ * @param p NULL, or a pointer previously returned from multi_heap_aligned_alloc() for the same heap.
+ * @note This function is deprecated, consider using  multi_heap_free() instead
+ */
+void __attribute__((deprecated)) multi_heap_aligned_free(multi_heap_handle_t heap, void *p);
+
+/** @brief free() a buffer in a given heap.
+ *
+ * Semantics are the same as standard free(), only the argument 'p' must be NULL or have been allocated in the specified heap.
+ *
+ * @param heap Handle to a registered heap.
+ * @param p NULL, or a pointer previously returned from multi_heap_malloc() or multi_heap_realloc() for the same heap.
+ */
+void multi_heap_free(multi_heap_handle_t heap, void *p);
+
+/** @brief realloc() a buffer in a given heap.
+ *
+ * Semantics are the same as standard realloc(), only the argument 'p' must be NULL or have been allocated in the specified heap.
+ *
+ * @param heap Handle to a registered heap.
+ * @param p NULL, or a pointer previously returned from multi_heap_malloc() or multi_heap_realloc() for the same heap.
+ * @param size Desired new size for buffer.
+ *
+ * @return New buffer of 'size' containing contents of 'p', or NULL if reallocation failed.
+ */
+void *multi_heap_realloc(multi_heap_handle_t heap, void *p, size_t size);
+
+
+/** @brief Return the size that a particular pointer was allocated with.
+ *
+ * @param heap Handle to a registered heap.
+ * @param p Pointer, must have been previously returned from multi_heap_malloc() or multi_heap_realloc() for the same heap.
+ *
+ * @return Size of the memory allocated at this block. May be more than the original size argument, due
+ * to padding and minimum block sizes.
+ */
+size_t multi_heap_get_allocated_size(multi_heap_handle_t heap, void *p);
+
+
+/** @brief Register a new heap for use
+ *
+ * This function initialises a heap at the specified address, and returns a handle for future heap operations.
+ *
+ * There is no equivalent function for deregistering a heap - if all blocks in the heap are free, you can immediately start using the memory for other purposes.
+ *
+ * @param start Start address of the memory to use for a new heap.
+ * @param size Size (in bytes) of the new heap.
+ *
+ * @return Handle of a new heap ready for use, or NULL if the heap region was too small to be initialised.
+ */
+multi_heap_handle_t multi_heap_register(void *start, size_t size);
+
+
+/** @brief Associate a private lock pointer with a heap
+ *
+ * The lock argument is supplied to the MULTI_HEAP_LOCK() and MULTI_HEAP_UNLOCK() macros, defined in multi_heap_platform.h.
+ *
+ * The lock in question must be recursive.
+ *
+ * When the heap is first registered, the associated lock is NULL.
+ *
+ * @param heap Handle to a registered heap.
+ * @param lock Optional pointer to a locking structure to associate with this heap.
+ */
+void multi_heap_set_lock(multi_heap_handle_t heap, void* lock);
+
+/** @brief Dump heap information to stdout
+ *
+ * For debugging purposes, this function dumps information about every block in the heap to stdout.
+ *
+ * @param heap Handle to a registered heap.
+ */
+void multi_heap_dump(multi_heap_handle_t heap);
+
+/** @brief Check heap integrity
+ *
+ * Walks the heap and checks all heap data structures are valid. If any errors are detected, an error-specific message
+ * can be optionally printed to stderr. Print behaviour can be overriden at compile time by defining
+ * MULTI_CHECK_FAIL_PRINTF in multi_heap_platform.h.
+ *
+ * @param heap Handle to a registered heap.
+ * @param print_errors If true, errors will be printed to stderr.
+ * @return true if heap is valid, false otherwise.
+ */
+bool multi_heap_check(multi_heap_handle_t heap, bool print_errors);
+
+/** @brief Return free heap size
+ *
+ * Returns the number of bytes available in the heap.
+ *
+ * Equivalent to the total_free_bytes member returned by multi_heap_get_heap_info().
+ *
+ * Note that the heap may be fragmented, so the actual maximum size for a single malloc() may be lower. To know this
+ * size, see the largest_free_block member returned by multi_heap_get_heap_info().
+ *
+ * @param heap Handle to a registered heap.
+ * @return Number of free bytes.
+ */
+size_t multi_heap_free_size(multi_heap_handle_t heap);
+
+/** @brief Return the lifetime minimum free heap size
+ *
+ * Equivalent to the minimum_free_bytes member returned by multi_heap_get_info().
+ *
+ * Returns the lifetime "low water mark" of possible values returned from multi_free_heap_size(), for the specified
+ * heap.
+ *
+ * @param heap Handle to a registered heap.
+ * @return Number of free bytes.
+ */
+size_t multi_heap_minimum_free_size(multi_heap_handle_t heap);
+
+/** @brief Structure to access heap metadata via multi_heap_get_info */
+typedef struct {
+    size_t total_free_bytes;      ///<  Total free bytes in the heap. Equivalent to multi_free_heap_size().
+    size_t total_allocated_bytes; ///<  Total bytes allocated to data in the heap.
+    size_t largest_free_block;    ///<  Size of largest free block in the heap. This is the largest malloc-able size.
+    size_t minimum_free_bytes;    ///<  Lifetime minimum free heap size. Equivalent to multi_minimum_free_heap_size().
+    size_t allocated_blocks;      ///<  Number of (variable size) blocks allocated in the heap.
+    size_t free_blocks;           ///<  Number of (variable size) free blocks in the heap.
+    size_t total_blocks;          ///<  Total number of (variable size) blocks in the heap.
+} multi_heap_info_t;
+
+/** @brief Return metadata about a given heap
+ *
+ * Fills a multi_heap_info_t structure with information about the specified heap.
+ *
+ * @param heap Handle to a registered heap.
+ * @param info Pointer to a structure to fill with heap metadata.
+ */
+void multi_heap_get_info(multi_heap_handle_t heap, multi_heap_info_t *info);
+
+#ifdef __cplusplus
+}
+#endif

components/heap/linker.lf

@@ -0,0 +1,7 @@
+[mapping:heap]
+archive: libheap.a
+entries:
+    heap_tlsf (noflash)
+    multi_heap (noflash)
+    if HEAP_POISONING_DISABLED = n:
+        multi_heap_poisoning (noflash)

components/heap/multi_heap.c

@@ -0,0 +1,376 @@
+// Copyright 2015-2016 Espressif Systems (Shanghai) PTE LTD
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+#include <stdint.h>
+#include <stdlib.h>
+#include <stdbool.h>
+#include <assert.h>
+#include <string.h>
+#include <stddef.h>
+#include <stdio.h>
+#include <sys/cdefs.h>
+#include "heap_tlsf.h"
+#include <multi_heap.h>
+#include "multi_heap_internal.h"
+
+/* Note: Keep platform-specific parts in this header, this source
+   file should depend on libc only */
+#include "multi_heap_platform.h"
+
+/* Defines compile-time configuration macros */
+#include "multi_heap_config.h"
+
+#ifndef MULTI_HEAP_POISONING
+/* if no heap poisoning, public API aliases directly to these implementations */
+void *multi_heap_malloc(multi_heap_handle_t heap, size_t size)
+    __attribute__((alias("multi_heap_malloc_impl")));
+
+void *multi_heap_aligned_alloc(multi_heap_handle_t heap, size_t size, size_t alignment)
+    __attribute__((alias("multi_heap_aligned_alloc_impl")));
+
+void multi_heap_aligned_free(multi_heap_handle_t heap, void *p)
+    __attribute__((alias("multi_heap_free_impl")));
+
+void multi_heap_free(multi_heap_handle_t heap, void *p)
+    __attribute__((alias("multi_heap_free_impl")));
+
+void *multi_heap_realloc(multi_heap_handle_t heap, void *p, size_t size)
+    __attribute__((alias("multi_heap_realloc_impl")));
+
+size_t multi_heap_get_allocated_size(multi_heap_handle_t heap, void *p)
+    __attribute__((alias("multi_heap_get_allocated_size_impl")));
+
+multi_heap_handle_t multi_heap_register(void *start, size_t size)
+    __attribute__((alias("multi_heap_register_impl")));
+
+void multi_heap_get_info(multi_heap_handle_t heap, multi_heap_info_t *info)
+    __attribute__((alias("multi_heap_get_info_impl")));
+
+size_t multi_heap_free_size(multi_heap_handle_t heap)
+    __attribute__((alias("multi_heap_free_size_impl")));
+
+size_t multi_heap_minimum_free_size(multi_heap_handle_t heap)
+    __attribute__((alias("multi_heap_minimum_free_size_impl")));
+
+void *multi_heap_get_block_address(multi_heap_block_handle_t block)
+    __attribute__((alias("multi_heap_get_block_address_impl")));
+
+void *multi_heap_get_block_owner(multi_heap_block_handle_t block)
+{
+    return NULL;
+}
+
+#endif
+
+#define ALIGN(X) ((X) & ~(sizeof(void *)-1))
+#define ALIGN_UP(X) ALIGN((X)+sizeof(void *)-1)
+#define ALIGN_UP_BY(num, align) (((num) + ((align) - 1)) & ~((align) - 1))
+
+
+typedef struct multi_heap_info {
+    void *lock;
+    size_t free_bytes;
+    size_t minimum_free_bytes;
+    size_t pool_size;
+    tlsf_t heap_data;
+} heap_t;
+
+/* Return true if this block is free. */
+static inline bool is_free(const block_header_t *block)
+{
+    return ((block->size & 0x01) != 0);
+}
+
+/* Data size of the block (excludes this block's header) */
+static inline size_t block_data_size(const block_header_t *block)
+{
+    return (block->size & ~0x03);
+}
+
+/* Check a block is valid for this heap. Used to verify parameters. */
+static void assert_valid_block(const heap_t *heap, const block_header_t *block)
+{
+    pool_t pool = tlsf_get_pool(heap->heap_data);
+    void *ptr = block_to_ptr(block);
+
+    MULTI_HEAP_ASSERT((ptr >= pool) &&
+                    (ptr < pool + heap->pool_size),
+                    (uintptr_t)ptr);
+}
+
+void *multi_heap_get_block_address_impl(multi_heap_block_handle_t block)
+{
+    void *ptr = block_to_ptr(block);
+    return (ptr);
+}
+
+size_t multi_heap_get_allocated_size_impl(multi_heap_handle_t heap, void *p)
+{
+    return tlsf_block_size(p);
+}
+
+multi_heap_handle_t multi_heap_register_impl(void *start_ptr, size_t size)
+{
+    assert(start_ptr);
+    if(size < (tlsf_size(NULL) + tlsf_block_size_min() + sizeof(heap_t))) {
+        //Region too small to be a heap.
+        return NULL;
+    }
+
+    heap_t *result = (heap_t *)start_ptr;
+    size -= sizeof(heap_t);
+
+    result->heap_data = tlsf_create_with_pool(start_ptr + sizeof(heap_t), size, 0);
+    if(!result->heap_data) {
+        return NULL;
+    }
+
+    result->lock = NULL;
+    result->free_bytes = size - tlsf_size(result->heap_data);
+    result->pool_size = size;
+    result->minimum_free_bytes = result->free_bytes;
+    return result;
+}
+
+void multi_heap_set_lock(multi_heap_handle_t heap, void *lock)
+{
+    heap->lock = lock;
+}
+
+void inline multi_heap_internal_lock(multi_heap_handle_t heap)
+{
+    MULTI_HEAP_LOCK(heap->lock);
+}
+
+void inline multi_heap_internal_unlock(multi_heap_handle_t heap)
+{
+    MULTI_HEAP_UNLOCK(heap->lock);
+}
+
+multi_heap_block_handle_t multi_heap_get_first_block(multi_heap_handle_t heap)
+{
+    assert(heap != NULL);
+    pool_t pool = tlsf_get_pool(heap->heap_data);
+    block_header_t* block = offset_to_block(pool, -(int)block_header_overhead);
+
+    return (multi_heap_block_handle_t)block;
+}
+
+multi_heap_block_handle_t multi_heap_get_next_block(multi_heap_handle_t heap, multi_heap_block_handle_t block)
+{
+    assert(heap != NULL);
+    assert_valid_block(heap, block);
+    block_header_t* next = block_next(block);
+
+    if(block_data_size(next) == 0) {
+        //Last block:
+        return NULL;
+    } else {
+        return (multi_heap_block_handle_t)next;
+    }
+
+}
+
+bool multi_heap_is_free(multi_heap_block_handle_t block)
+{
+    return is_free(block);
+}
+
+void *multi_heap_malloc_impl(multi_heap_handle_t heap, size_t size)
+{
+    if (size == 0 || heap == NULL) {
+        return NULL;
+    }
+
+
+    multi_heap_internal_lock(heap);
+    void *result = tlsf_malloc(heap->heap_data, size);
+    if(result) {
+        heap->free_bytes -= tlsf_block_size(result);
+        if (heap->free_bytes < heap->minimum_free_bytes) {
+            heap->minimum_free_bytes = heap->free_bytes;
+        }
+    }
+    multi_heap_internal_unlock(heap);
+
+    return result;
+}
+
+void multi_heap_free_impl(multi_heap_handle_t heap, void *p)
+{
+
+    if (heap == NULL || p == NULL) {
+        return;
+    }
+
+    assert_valid_block(heap, p);
+
+    multi_heap_internal_lock(heap);
+    heap->free_bytes += tlsf_block_size(p);
+    tlsf_free(heap->heap_data, p);
+    multi_heap_internal_unlock(heap);
+}
+
+void *multi_heap_realloc_impl(multi_heap_handle_t heap, void *p, size_t size)
+{
+    assert(heap != NULL);
+
+    if (p == NULL) {
+        return multi_heap_malloc_impl(heap, size);
+    }
+
+    assert_valid_block(heap, p);
+
+    if (heap == NULL) {
+        return NULL;
+    }
+
+    multi_heap_internal_lock(heap);
+    size_t previous_block_size =  tlsf_block_size(p);
+    void *result = tlsf_realloc(heap->heap_data, p, size);
+    if(result) {
+        heap->free_bytes += previous_block_size;
+        heap->free_bytes -= tlsf_block_size(result);
+        if (heap->free_bytes < heap->minimum_free_bytes) {
+            heap->minimum_free_bytes = heap->free_bytes;
+        }
+    }
+
+    multi_heap_internal_unlock(heap);
+
+    return result;
+}
+
+void *multi_heap_aligned_alloc_impl_offs(multi_heap_handle_t heap, size_t size, size_t alignment, size_t offset)
+{
+    if(heap == NULL) {
+        return NULL;
+    }
+
+    if(!size) {
+        return NULL;
+    }
+
+    //Alignment must be a power of two:
+    if(((alignment & (alignment - 1)) != 0) ||(!alignment)) {
+        return NULL;
+    }
+
+    multi_heap_internal_lock(heap);
+    void *result = tlsf_memalign_offs(heap->heap_data, alignment, size, offset);
+    if(result) {
+        heap->free_bytes -= tlsf_block_size(result);
+        if(heap->free_bytes < heap->minimum_free_bytes) {
+            heap->minimum_free_bytes = heap->free_bytes;
+        }
+    }
+    multi_heap_internal_unlock(heap);
+
+    return result;
+}
+
+
+void *multi_heap_aligned_alloc_impl(multi_heap_handle_t heap, size_t size, size_t alignment)
+{
+    return multi_heap_aligned_alloc_impl_offs(heap, size, alignment, 0);
+}
+
+bool multi_heap_check(multi_heap_handle_t heap, bool print_errors)
+{
+    (void)print_errors;
+    bool valid = true;
+    assert(heap != NULL);
+
+    multi_heap_internal_lock(heap);
+    if(tlsf_check(heap->heap_data)) {
+        valid = false;
+    }
+
+    if(tlsf_check_pool(tlsf_get_pool(heap->heap_data))) {
+        valid = false;
+    }
+
+    multi_heap_internal_unlock(heap);
+    return valid;
+}
+
+static void multi_heap_dump_tlsf(void* ptr, size_t size, int used, void* user)
+{
+    (void)user;
+    MULTI_HEAP_STDERR_PRINTF("Block %p data, size: %d bytes, Free: %s \n",
+                            (void *)ptr,
+                            size,
+                            used ? "No" : "Yes");
+}
+
+void multi_heap_dump(multi_heap_handle_t heap)
+{
+    assert(heap != NULL);
+
+    multi_heap_internal_lock(heap);
+    MULTI_HEAP_STDERR_PRINTF("Showing data for heap: %p \n", (void *)heap);
+    tlsf_walk_pool(tlsf_get_pool(heap->heap_data), multi_heap_dump_tlsf, NULL);
+    multi_heap_internal_unlock(heap);
+}
+
+size_t multi_heap_free_size_impl(multi_heap_handle_t heap)
+{
+    if (heap == NULL) {
+        return 0;
+    }
+
+    return heap->free_bytes;
+}
+
+size_t multi_heap_minimum_free_size_impl(multi_heap_handle_t heap)
+{
+    if (heap == NULL) {
+        return 0;
+    }
+
+    return heap->minimum_free_bytes;
+}
+
+static void multi_heap_get_info_tlsf(void* ptr, size_t size, int used, void* user)
+{
+    multi_heap_info_t *info = user;
+
+    if(used) {
+        info->allocated_blocks++;
+    } else {
+        info->free_blocks++;
+
+        if(size > info->largest_free_block ) {
+            info->largest_free_block = size;
+        }
+    }
+
+    info->total_blocks++;
+}
+
+void multi_heap_get_info_impl(multi_heap_handle_t heap, multi_heap_info_t *info)
+{
+    memset(info, 0, sizeof(multi_heap_info_t));
+
+    if (heap == NULL) {
+        return;
+    }
+
+    multi_heap_internal_lock(heap);
+    tlsf_walk_pool(tlsf_get_pool(heap->heap_data), multi_heap_get_info_tlsf, info);
+    info->total_allocated_bytes = (heap->pool_size - tlsf_size(heap->heap_data)) - heap->free_bytes;
+    info->minimum_free_bytes = heap->minimum_free_bytes;
+    info->total_free_bytes = heap->free_bytes;
+	info->largest_free_block = tlsf_fit_size(heap->heap_data, info->largest_free_block);
+    multi_heap_internal_unlock(heap);
+}

components/heap/multi_heap_config.h

@@ -0,0 +1,31 @@
+// Copyright 2015-2016 Espressif Systems (Shanghai) PTE LTD
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+#pragma once
+
+#ifdef ESP_PLATFORM
+#include "sdkconfig.h"
+#include "soc/soc.h"
+#include "soc/soc_caps.h"
+#endif
+
+/* Configuration macros for multi-heap */
+
+#ifdef CONFIG_HEAP_POISONING_LIGHT
+#define MULTI_HEAP_POISONING
+#endif
+
+#ifdef CONFIG_HEAP_POISONING_COMPREHENSIVE
+#define MULTI_HEAP_POISONING
+#define MULTI_HEAP_POISONING_SLOW
+#endif

components/heap/multi_heap_internal.h

@@ -0,0 +1,76 @@
+// Copyright 2015-2016 Espressif Systems (Shanghai) PTE LTD
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+#pragma once
+
+/* Opaque handle to a heap block */
+typedef const struct block_header_t *multi_heap_block_handle_t;
+
+/* Internal definitions for the "implementation" of the multi_heap API,
+   as defined in multi_heap.c.
+
+   If heap poisioning is disabled, these are aliased directly to the public API.
+
+   If heap poisoning is enabled, wrapper functions call each of these.
+*/
+
+void *multi_heap_malloc_impl(multi_heap_handle_t heap, size_t size);
+
+/* Allocate a memory region of minimum `size` bytes, aligned on `alignment`. */
+void *multi_heap_aligned_alloc_impl(multi_heap_handle_t heap, size_t size, size_t alignment);
+
+/* Allocate a memory region of minimum `size` bytes, where memory's `offset` is aligned on `alignment`. */
+void *multi_heap_aligned_alloc_impl_offs(multi_heap_handle_t heap, size_t size, size_t alignment, size_t offset);
+
+void multi_heap_free_impl(multi_heap_handle_t heap, void *p);
+void *multi_heap_realloc_impl(multi_heap_handle_t heap, void *p, size_t size);
+multi_heap_handle_t multi_heap_register_impl(void *start, size_t size);
+void multi_heap_get_info_impl(multi_heap_handle_t heap, multi_heap_info_t *info);
+size_t multi_heap_free_size_impl(multi_heap_handle_t heap);
+size_t multi_heap_minimum_free_size_impl(multi_heap_handle_t heap);
+size_t multi_heap_get_allocated_size_impl(multi_heap_handle_t heap, void *p);
+void *multi_heap_get_block_address_impl(multi_heap_block_handle_t block);
+
+/* Some internal functions for heap poisoning use */
+
+/* Check an allocated block's poison bytes are correct. Called by multi_heap_check(). */
+bool multi_heap_internal_check_block_poisoning(void *start, size_t size, bool is_free, bool print_errors);
+
+/* Fill a region of memory with the free or malloced pattern.
+   Called when merging blocks, to overwrite the old block header.
+*/
+void multi_heap_internal_poison_fill_region(void *start, size_t size, bool is_free);
+
+/* Allow heap poisoning to lock/unlock the heap to avoid race conditions
+   if multi_heap_check() is running concurrently.
+*/
+void multi_heap_internal_lock(multi_heap_handle_t heap);
+
+void multi_heap_internal_unlock(multi_heap_handle_t heap);
+
+/* Some internal functions for heap debugging code to use */
+
+/* Get the handle to the first (fixed free) block in a heap */
+multi_heap_block_handle_t multi_heap_get_first_block(multi_heap_handle_t heap);
+
+/* Get the handle to the next block in a heap, with validation */
+multi_heap_block_handle_t multi_heap_get_next_block(multi_heap_handle_t heap, multi_heap_block_handle_t block);
+
+/* Test if a heap block is free */
+bool multi_heap_is_free(const multi_heap_block_handle_t block);
+
+/* Get the data address of a heap block */
+void *multi_heap_get_block_address(multi_heap_block_handle_t block);
+
+/* Get the owner identification for a heap block */
+void *multi_heap_get_block_owner(multi_heap_block_handle_t block);

components/heap/multi_heap_platform.h

@@ -0,0 +1,108 @@
+// Copyright 2015-2016 Espressif Systems (Shanghai) PTE LTD
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+#pragma once
+
+#ifdef MULTI_HEAP_FREERTOS
+
+#include "freertos/FreeRTOS.h"
+
+#include "sdkconfig.h"
+#include "esp_rom_sys.h"
+#if CONFIG_IDF_TARGET_ESP32
+#include "esp32/rom/ets_sys.h" // will be removed in idf v5.0
+#elif CONFIG_IDF_TARGET_ESP32S2
+#include "esp32s2/rom/ets_sys.h"
+#endif
+#include <assert.h>
+
+typedef portMUX_TYPE multi_heap_lock_t;
+
+/* Because malloc/free can happen inside an ISR context,
+   we need to use portmux spinlocks here not RTOS mutexes */
+#define MULTI_HEAP_LOCK(PLOCK) do {                         \
+        if((PLOCK) != NULL) {                               \
+            portENTER_CRITICAL((PLOCK));                    \
+        }                                                   \
+    } while(0)
+
+
+#define MULTI_HEAP_UNLOCK(PLOCK) do {                       \
+        if ((PLOCK) != NULL) {                              \
+            portEXIT_CRITICAL((PLOCK));                     \
+        }                                                   \
+    } while(0)
+
+#define MULTI_HEAP_LOCK_INIT(PLOCK) do {                    \
+        vPortCPUInitializeMutex((PLOCK));                   \
+    } while(0)
+
+#define MULTI_HEAP_LOCK_STATIC_INITIALIZER     portMUX_INITIALIZER_UNLOCKED
+
+/* Not safe to use std i/o while in a portmux critical section,
+   can deadlock, so we use the ROM equivalent functions. */
+
+#define MULTI_HEAP_PRINTF esp_rom_printf
+#define MULTI_HEAP_STDERR_PRINTF(MSG, ...) esp_rom_printf(MSG, __VA_ARGS__)
+
+inline static void multi_heap_assert(bool condition, const char *format, int line, intptr_t address)
+{
+    /* Can't use libc assert() here as it calls printf() which can cause another malloc() for a newlib lock.
+
+       Also, it's useful to be able to print the memory address where corruption was detected.
+    */
+#ifndef NDEBUG
+    if(!condition) {
+#ifndef CONFIG_COMPILER_OPTIMIZATION_ASSERTIONS_SILENT
+        esp_rom_printf(format, line, address);
+#endif  // CONFIG_COMPILER_OPTIMIZATION_ASSERTIONS_SILENT
+        abort();
+    }
+#else // NDEBUG
+    (void) condition;
+#endif // NDEBUG
+}
+
+#define MULTI_HEAP_ASSERT(CONDITION, ADDRESS) \
+    multi_heap_assert((CONDITION), "CORRUPT HEAP: multi_heap.c:%d detected at 0x%08x\n", \
+                      __LINE__, (intptr_t)(ADDRESS))
+
+#ifdef CONFIG_HEAP_TASK_TRACKING
+#include <freertos/task.h>
+#define MULTI_HEAP_BLOCK_OWNER TaskHandle_t task;
+#define MULTI_HEAP_SET_BLOCK_OWNER(HEAD) (HEAD)->task = xTaskGetCurrentTaskHandle()
+#define MULTI_HEAP_GET_BLOCK_OWNER(HEAD) ((HEAD)->task)
+#else
+#define MULTI_HEAP_BLOCK_OWNER
+#define MULTI_HEAP_SET_BLOCK_OWNER(HEAD)
+#define MULTI_HEAP_GET_BLOCK_OWNER(HEAD) (NULL)
+#endif
+
+#else // MULTI_HEAP_FREERTOS
+
+#include <assert.h>
+
+#define MULTI_HEAP_PRINTF printf
+#define MULTI_HEAP_STDERR_PRINTF(MSG, ...) fprintf(stderr, MSG, __VA_ARGS__)
+#define MULTI_HEAP_LOCK(PLOCK)  (void) (PLOCK)
+#define MULTI_HEAP_UNLOCK(PLOCK)  (void) (PLOCK)
+#define MULTI_HEAP_LOCK_INIT(PLOCK)  (void) (PLOCK)
+#define MULTI_HEAP_LOCK_STATIC_INITIALIZER  0
+
+#define MULTI_HEAP_ASSERT(CONDITION, ADDRESS) assert((CONDITION) && "Heap corrupt")
+
+#define MULTI_HEAP_BLOCK_OWNER
+#define MULTI_HEAP_SET_BLOCK_OWNER(HEAD)
+#define MULTI_HEAP_GET_BLOCK_OWNER(HEAD) (NULL)
+
+#endif // MULTI_HEAP_FREERTOS

components/heap/multi_heap_poisoning.c

@@ -0,0 +1,426 @@
+// Copyright 2015-2016 Espressif Systems (Shanghai) PTE LTD
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+
+//     http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+#include <stdint.h>
+#include <stdlib.h>
+#include <stdbool.h>
+#include <assert.h>
+#include <string.h>
+#include <stddef.h>
+#include <stdio.h>
+#include <sys/param.h>
+#include <multi_heap.h>
+#include "multi_heap_internal.h"
+
+/* Note: Keep platform-specific parts in this header, this source
+   file should depend on libc only */
+#include "multi_heap_platform.h"
+
+/* Defines compile-time configuration macros */
+#include "multi_heap_config.h"
+
+#ifdef MULTI_HEAP_POISONING
+
+/* Alias MULTI_HEAP_POISONING_SLOW to SLOW for better readabilty */
+#ifdef SLOW
+#error "external header has defined SLOW"
+#endif
+#ifdef MULTI_HEAP_POISONING_SLOW
+#define SLOW 1
+#endif
+
+#define MALLOC_FILL_PATTERN 0xce
+#define FREE_FILL_PATTERN 0xfe
+
+#define HEAD_CANARY_PATTERN 0xABBA1234
+#define TAIL_CANARY_PATTERN 0xBAAD5678
+
+
+#define ALIGN_UP(num, align) (((num) + ((align) - 1)) & ~((align) - 1))
+
+typedef struct {
+    uint32_t head_canary;
+    MULTI_HEAP_BLOCK_OWNER
+    size_t alloc_size;
+} poison_head_t;
+
+typedef struct {
+    uint32_t tail_canary;
+} poison_tail_t;
+
+#define POISON_OVERHEAD (sizeof(poison_head_t) + sizeof(poison_tail_t))
+
+/* Given a "poisoned" region with pre-data header 'head', and actual data size 'alloc_size', fill in the head and tail
+   region checks.
+
+   Returns the pointer to the actual usable data buffer (ie after 'head')
+*/
+static uint8_t *poison_allocated_region(poison_head_t *head, size_t alloc_size)
+{
+    uint8_t *data = (uint8_t *)(&head[1]); /* start of data ie 'real' allocated buffer */
+    poison_tail_t *tail = (poison_tail_t *)(data + alloc_size);
+    head->alloc_size = alloc_size;
+    head->head_canary = HEAD_CANARY_PATTERN;
+    MULTI_HEAP_SET_BLOCK_OWNER(head);
+
+    uint32_t tail_canary = TAIL_CANARY_PATTERN;
+    if ((intptr_t)tail % sizeof(void *) == 0) {
+        tail->tail_canary = tail_canary;
+    } else {
+        /* unaligned tail_canary */
+        memcpy(&tail->tail_canary, &tail_canary, sizeof(uint32_t));
+    }
+
+    return data;
+}
+
+/* Given a pointer to some allocated data, check the head & tail poison structures (before & after it) that were
+   previously injected by poison_allocated_region().
+
+   Returns a pointer to the poison header structure, or NULL if the poison structures are corrupt.
+*/
+static poison_head_t *verify_allocated_region(void *data, bool print_errors)
+{
+    poison_head_t *head = (poison_head_t *)((intptr_t)data - sizeof(poison_head_t));
+    poison_tail_t *tail = (poison_tail_t *)((intptr_t)data + head->alloc_size);
+
+    /* check if the beginning of the data was overwritten */
+    if (head->head_canary != HEAD_CANARY_PATTERN) {
+        if (print_errors) {
+            MULTI_HEAP_STDERR_PRINTF("CORRUPT HEAP: Bad head at %p. Expected 0x%08x got 0x%08x\n", &head->head_canary,
+                   HEAD_CANARY_PATTERN, head->head_canary);
+        }
+        return NULL;
+    }
+
+    /* check if the end of the data was overrun */
+    uint32_t canary;
+    if ((intptr_t)tail % sizeof(void *) == 0) {
+        canary = tail->tail_canary;
+    } else {
+        /* tail is unaligned */
+        memcpy(&canary, &tail->tail_canary, sizeof(canary));
+    }
+    if (canary != TAIL_CANARY_PATTERN) {
+        if (print_errors) {
+            MULTI_HEAP_STDERR_PRINTF("CORRUPT HEAP: Bad tail at %p. Expected 0x%08x got 0x%08x\n", &tail->tail_canary,
+                   TAIL_CANARY_PATTERN, canary);
+        }
+        return NULL;
+    }
+
+    return head;
+}
+
+#ifdef SLOW
+/* Go through a region that should have the specified fill byte 'pattern',
+   verify it.
+
+   if expect_free is true, expect FREE_FILL_PATTERN otherwise MALLOC_FILL_PATTERN.
+
+   if swap_pattern is true, swap patterns in the buffer (ie replace MALLOC_FILL_PATTERN with FREE_FILL_PATTERN, and vice versa.)
+
+   Returns true if verification checks out.
+*/
+static bool verify_fill_pattern(void *data, size_t size, bool print_errors, bool expect_free, bool swap_pattern)
+{
+    const uint32_t FREE_FILL_WORD = (FREE_FILL_PATTERN << 24) | (FREE_FILL_PATTERN << 16) | (FREE_FILL_PATTERN << 8) | FREE_FILL_PATTERN;
+    const uint32_t MALLOC_FILL_WORD = (MALLOC_FILL_PATTERN << 24) | (MALLOC_FILL_PATTERN << 16) | (MALLOC_FILL_PATTERN << 8) | MALLOC_FILL_PATTERN;
+
+    const uint32_t EXPECT_WORD = expect_free ? FREE_FILL_WORD : MALLOC_FILL_WORD;
+    const uint32_t REPLACE_WORD = expect_free ? MALLOC_FILL_WORD : FREE_FILL_WORD;
+    bool valid = true;
+
+    /* Use 4-byte operations as much as possible */
+    if ((intptr_t)data % 4 == 0) {
+        uint32_t *p = data;
+        while (size >= 4) {
+            if (*p != EXPECT_WORD) {
+                if (print_errors) {
+                    MULTI_HEAP_STDERR_PRINTF("CORRUPT HEAP: Invalid data at %p. Expected 0x%08x got 0x%08x\n", p, EXPECT_WORD, *p);
+                }
+                valid = false;
+#ifndef NDEBUG
+                /* If an assertion is going to fail as soon as we're done verifying the pattern, leave the rest of the
+                   buffer contents as-is for better post-mortem analysis
+                */
+                swap_pattern = false;
+#endif
+            }
+            if (swap_pattern) {
+                *p = REPLACE_WORD;
+            }
+            p++;
+            size -= 4;
+        }
+        data = p;
+    }
+
+    uint8_t *p = data;
+    for (size_t i = 0; i < size; i++) {
+        if (p[i] != (uint8_t)EXPECT_WORD) {
+            if (print_errors) {
+                MULTI_HEAP_STDERR_PRINTF("CORRUPT HEAP: Invalid data at %p. Expected 0x%02x got 0x%02x\n", p, (uint8_t)EXPECT_WORD, *p);
+            }
+            valid = false;
+#ifndef NDEBUG
+            swap_pattern = false; // same as above
+#endif
+        }
+        if (swap_pattern) {
+            p[i] = (uint8_t)REPLACE_WORD;
+        }
+    }
+    return valid;
+}
+#endif
+
+void *multi_heap_aligned_alloc(multi_heap_handle_t heap, size_t size, size_t alignment)
+{
+    if (!size) {
+        return NULL;
+    }
+
+    if (size > SIZE_MAX  - POISON_OVERHEAD) {
+        return NULL;
+    }
+
+    multi_heap_internal_lock(heap);
+    poison_head_t *head = multi_heap_aligned_alloc_impl_offs(heap, size + POISON_OVERHEAD,
+                                                             alignment, sizeof(poison_head_t));
+    uint8_t *data = NULL;
+    if (head != NULL) {
+        data = poison_allocated_region(head, size);
+#ifdef SLOW
+        /* check everything we got back is FREE_FILL_PATTERN & swap for MALLOC_FILL_PATTERN */
+        bool ret = verify_fill_pattern(data, size, true, true, true);
+        assert( ret );
+#endif
+    } else {
+        multi_heap_internal_unlock(heap);
+        return NULL;
+    }
+
+    multi_heap_internal_unlock(heap);
+
+    return data;
+}
+
+void *multi_heap_malloc(multi_heap_handle_t heap, size_t size)
+{
+    if (!size) {
+        return NULL;
+    }
+
+    if(size > SIZE_MAX - POISON_OVERHEAD) {
+        return NULL;
+    }
+
+    multi_heap_internal_lock(heap);
+    poison_head_t *head = multi_heap_malloc_impl(heap, size + POISON_OVERHEAD);
+    uint8_t *data = NULL;
+    if (head != NULL) {
+        data = poison_allocated_region(head, size);
+#ifdef SLOW
+        /* check everything we got back is FREE_FILL_PATTERN & swap for MALLOC_FILL_PATTERN */
+        bool ret = verify_fill_pattern(data, size, true, true, true);
+        assert( ret );
+#endif
+    }
+
+    multi_heap_internal_unlock(heap);
+    return data;
+}
+
+void multi_heap_free(multi_heap_handle_t heap, void *p)
+{
+    if (p == NULL) {
+        return;
+    }
+    multi_heap_internal_lock(heap);
+
+    poison_head_t *head = verify_allocated_region(p, true);
+    assert(head != NULL);
+
+    #ifdef SLOW
+    /* replace everything with FREE_FILL_PATTERN, including the poison head/tail */
+    memset(head, FREE_FILL_PATTERN,
+           head->alloc_size + POISON_OVERHEAD);
+    #endif
+    multi_heap_free_impl(heap, head);
+
+    multi_heap_internal_unlock(heap);
+}
+
+void multi_heap_aligned_free(multi_heap_handle_t heap, void *p)
+{
+    multi_heap_free(heap, p);
+}
+
+void *multi_heap_realloc(multi_heap_handle_t heap, void *p, size_t size)
+{
+    poison_head_t *head = NULL;
+    poison_head_t *new_head;
+    void *result = NULL;
+
+    if(size > SIZE_MAX - POISON_OVERHEAD) {
+        return NULL;
+    }
+    if (p == NULL) {
+        return multi_heap_malloc(heap, size);
+    }
+    if (size == 0) {
+        multi_heap_free(heap, p);
+        return NULL;
+    }
+
+    /* p != NULL, size != 0 */
+    head = verify_allocated_region(p, true);
+    assert(head != NULL);
+
+    multi_heap_internal_lock(heap);
+
+#ifndef SLOW
+    new_head = multi_heap_realloc_impl(heap, head, size + POISON_OVERHEAD);
+    if (new_head != NULL) {
+        /* For "fast" poisoning, we only overwrite the head/tail of the new block so it's safe
+           to poison, so no problem doing this even if realloc resized in place.
+        */
+        result = poison_allocated_region(new_head, size);
+    }
+#else // SLOW
+    /* When slow poisoning is enabled, it becomes very fiddly to try and correctly fill memory when resizing in place
+       (where the buffer may be moved (including to an overlapping address with the old buffer), grown, or shrunk in
+       place.)
+
+       For now we just malloc a new buffer, copy, and free. :|
+
+       Note: If this ever changes, multi_heap defrag realloc test should be enabled.
+    */
+    size_t orig_alloc_size = head->alloc_size;
+
+    new_head = multi_heap_malloc_impl(heap, size + POISON_OVERHEAD);
+    if (new_head != NULL) {
+        result = poison_allocated_region(new_head, size);
+        memcpy(result, p, MIN(size, orig_alloc_size));
+        multi_heap_free(heap, p);
+    }
+#endif
+
+    multi_heap_internal_unlock(heap);
+
+    return result;
+}
+
+void *multi_heap_get_block_address(multi_heap_block_handle_t block)
+{
+    char *head = multi_heap_get_block_address_impl(block);
+    return head + sizeof(poison_head_t);
+}
+
+void *multi_heap_get_block_owner(multi_heap_block_handle_t block)
+{
+    return MULTI_HEAP_GET_BLOCK_OWNER((poison_head_t*)multi_heap_get_block_address_impl(block));
+}
+
+multi_heap_handle_t multi_heap_register(void *start, size_t size)
+{
+#ifdef SLOW
+    if (start != NULL) {
+        memset(start, FREE_FILL_PATTERN, size);
+    }
+#endif
+    return multi_heap_register_impl(start, size);
+}
+
+static inline void subtract_poison_overhead(size_t *arg) {
+    if (*arg > POISON_OVERHEAD) {
+        *arg -= POISON_OVERHEAD;
+    } else {
+        *arg = 0;
+    }
+}
+
+size_t multi_heap_get_allocated_size(multi_heap_handle_t heap, void *p)
+{
+    poison_head_t *head = verify_allocated_region(p, true);
+    assert(head != NULL);
+    size_t result = multi_heap_get_allocated_size_impl(heap, head);
+    return result;
+}
+
+void multi_heap_get_info(multi_heap_handle_t heap, multi_heap_info_t *info)
+{
+    multi_heap_get_info_impl(heap, info);
+    /* don't count the heap poison head & tail overhead in the allocated bytes size */
+    info->total_allocated_bytes -= info->allocated_blocks * POISON_OVERHEAD;
+    /* trim largest_free_block to account for poison overhead */
+    subtract_poison_overhead(&info->largest_free_block);
+    /* similarly, trim total_free_bytes so there's no suggestion that
+       a block this big may be available. */
+    subtract_poison_overhead(&info->total_free_bytes);
+    subtract_poison_overhead(&info->minimum_free_bytes);
+}
+
+size_t multi_heap_free_size(multi_heap_handle_t heap)
+{
+    size_t r = multi_heap_free_size_impl(heap);
+    subtract_poison_overhead(&r);
+    return r;
+}
+
+size_t multi_heap_minimum_free_size(multi_heap_handle_t heap)
+{
+    size_t r = multi_heap_minimum_free_size_impl(heap);
+    subtract_poison_overhead(&r);
+    return r;
+}
+
+/* Internal hooks used by multi_heap to manage poisoning, while keeping some modularity */
+
+bool multi_heap_internal_check_block_poisoning(void *start, size_t size, bool is_free, bool print_errors)
+{
+    if (is_free) {
+#ifdef SLOW
+        return verify_fill_pattern(start, size, print_errors, true, false);
+#else
+        return true; /* can only verify empty blocks in SLOW mode */
+#endif
+    } else {
+        void *data = (void *)((intptr_t)start + sizeof(poison_head_t));
+        poison_head_t *head = verify_allocated_region(data, print_errors);
+        if (head != NULL && head->alloc_size > size - POISON_OVERHEAD) {
+            /* block can be bigger than alloc_size, for reasons of alignment & fragmentation,
+               but block can never be smaller than head->alloc_size... */
+            if (print_errors) {
+                MULTI_HEAP_STDERR_PRINTF("CORRUPT HEAP: Size at %p expected <=0x%08x got 0x%08x\n", &head->alloc_size,
+                       size - POISON_OVERHEAD, head->alloc_size);
+            }
+            return false;
+        }
+        return head != NULL;
+    }
+}
+
+void multi_heap_internal_poison_fill_region(void *start, size_t size, bool is_free)
+{
+    memset(start, is_free ? FREE_FILL_PATTERN : MALLOC_FILL_PATTERN, size);
+}
+
+#else // !MULTI_HEAP_POISONING
+
+#ifdef MULTI_HEAP_POISONING_SLOW
+#error "MULTI_HEAP_POISONING_SLOW requires MULTI_HEAP_POISONING"
+#endif
+
+#endif  // MULTI_HEAP_POISONING

components/heap/test/CMakeLists.txt

@@ -0,0 +1,3 @@
+idf_component_register(SRC_DIRS "."
+                    PRIV_INCLUDE_DIRS "."
+                    PRIV_REQUIRES cmock test_utils heap)

components/heap/test/component.mk

@@ -0,0 +1,5 @@
+#
+#Component Makefile
+#
+
+COMPONENT_ADD_LDFLAGS = -Wl,--whole-archive -l$(COMPONENT_NAME) -Wl,--no-whole-archive

components/heap/test/test_aligned_alloc_caps.c

@@ -0,0 +1,147 @@
+/*
+ Tests for the capabilities-based memory allocator.
+*/
+
+#include <esp_types.h>
+#include <stdio.h>
+#include "unity.h"
+#include "esp_attr.h"
+#include "esp_heap_caps.h"
+#include "esp_spi_flash.h"
+#include <stdlib.h>
+#include <sys/param.h>
+#include <string.h>
+#include <malloc.h>
+
+TEST_CASE("Capabilities aligned allocator test", "[heap]")
+{
+    uint32_t alignments = 0;
+
+    printf("[ALIGNED_ALLOC] Allocating from default CAP: \n");
+
+    for(;alignments <= 1024; alignments++) {
+        uint8_t *buf = (uint8_t *)memalign(alignments, (alignments + 137));
+        if(((alignments & (alignments - 1)) != 0) || (!alignments)) {
+            TEST_ASSERT( buf == NULL );
+            //printf("[ALIGNED_ALLOC] alignment: %u is not a power of two, don't allow allocation \n", aligments);
+        } else {
+            TEST_ASSERT( buf != NULL );
+            printf("[ALIGNED_ALLOC] alignment required: %u \n", alignments);
+            printf("[ALIGNED_ALLOC] address of allocated memory: %p \n\n", (void *)buf);
+            //Address of obtained block must be aligned with selected value
+            TEST_ASSERT(((intptr_t)buf & (alignments - 1)) == 0);
+
+            //Write some data, if it corrupts memory probably the heap
+            //canary verification will fail:
+            memset(buf, 0xA5, (alignments + 137));
+
+            free(buf);
+        }
+    }
+
+    //Alloc from a non permitted area:
+    uint32_t *not_permitted_buf = (uint32_t *)heap_caps_aligned_alloc(alignments, (alignments + 137), MALLOC_CAP_EXEC | MALLOC_CAP_32BIT);
+    TEST_ASSERT( not_permitted_buf == NULL );
+
+#if CONFIG_ESP32_SPIRAM_SUPPORT || CONFIG_ESP32S2_SPIRAM_SUPPORT
+    alignments = 0;
+    printf("[ALIGNED_ALLOC] Allocating from external memory: \n");
+
+    for(;alignments <= 1024 * 1024; alignments++) {
+        //Now try to take aligned memory from IRAM:
+        uint8_t *buf = (uint8_t *)heap_caps_aligned_alloc(alignments, 10*1024, MALLOC_CAP_SPIRAM);
+        if(((alignments & (alignments - 1)) != 0) || (!alignments)) {
+            TEST_ASSERT( buf == NULL );
+            //printf("[ALIGNED_ALLOC] alignment: %u is not a power of two, don't allow allocation \n", aligments);
+        } else {
+            TEST_ASSERT( buf != NULL );
+            printf("[ALIGNED_ALLOC] alignment required: %u \n", alignments);
+            printf("[ALIGNED_ALLOC] address of allocated memory: %p \n\n", (void *)buf);
+            //Address of obtained block must be aligned with selected value
+            TEST_ASSERT(((intptr_t)buf & (alignments - 1)) == 0);
+
+            //Write some data, if it corrupts memory probably the heap
+            //canary verification will fail:
+            memset(buf, 0xA5, (10*1024));
+            heap_caps_free(buf);
+        }
+    }
+#endif
+
+}
+
+TEST_CASE("Capabilities aligned calloc test", "[heap]")
+{
+    uint32_t alignments = 0;
+
+    printf("[ALIGNED_ALLOC] Allocating from default CAP: \n");
+
+    for(;alignments <= 1024; alignments++) {
+        uint8_t *buf = (uint8_t *)heap_caps_aligned_calloc(alignments, 1, (alignments + 137), MALLOC_CAP_DEFAULT);
+        if(((alignments & (alignments - 1)) != 0) || (!alignments)) {
+            TEST_ASSERT( buf == NULL );
+            //printf("[ALIGNED_ALLOC] alignment: %u is not a power of two, don't allow allocation \n", aligments);
+        } else {
+            TEST_ASSERT( buf != NULL );
+            printf("[ALIGNED_ALLOC] alignment required: %u \n", alignments);
+            printf("[ALIGNED_ALLOC] address of allocated memory: %p \n\n", (void *)buf);
+            //Address of obtained block must be aligned with selected value
+            TEST_ASSERT(((intptr_t)buf & (alignments - 1)) == 0);
+
+            //Write some data, if it corrupts memory probably the heap
+            //canary verification will fail:
+            memset(buf, 0xA5, (alignments + 137));
+
+            heap_caps_free(buf);
+        }
+    }
+
+    //Check if memory is initialized with zero:
+    uint8_t byte_array[1024];
+    memset(&byte_array, 0, sizeof(byte_array));
+    uint8_t *buf = (uint8_t *)heap_caps_aligned_calloc(1024, 1, 1024, MALLOC_CAP_DEFAULT);
+    TEST_ASSERT(memcmp(byte_array, buf, sizeof(byte_array)) == 0);
+    heap_caps_free(buf);
+
+    //Same size, but different chunk:
+    buf = (uint8_t *)heap_caps_aligned_calloc(1024, 1024, 1, MALLOC_CAP_DEFAULT);
+    TEST_ASSERT(memcmp(byte_array, buf, sizeof(byte_array)) == 0);
+    heap_caps_free(buf);
+
+    //Alloc from a non permitted area:
+    uint32_t *not_permitted_buf = (uint32_t *)heap_caps_aligned_calloc(alignments, 1, (alignments + 137), MALLOC_CAP_32BIT);
+    TEST_ASSERT( not_permitted_buf == NULL );
+
+#if CONFIG_ESP32_SPIRAM_SUPPORT || CONFIG_ESP32S2_SPIRAM_SUPPORT
+    alignments = 0;
+    printf("[ALIGNED_ALLOC] Allocating from external memory: \n");
+
+    for(;alignments <= 1024 * 1024; alignments++) {
+        //Now try to take aligned memory from IRAM:
+        uint8_t *buf = (uint8_t *)(uint8_t *)heap_caps_aligned_calloc(alignments, 1, 10*1024, MALLOC_CAP_SPIRAM);
+        if(((alignments & (alignments - 1)) != 0) || (!alignments)) {
+            TEST_ASSERT( buf == NULL );
+            //printf("[ALIGNED_ALLOC] alignment: %u is not a power of two, don't allow allocation \n", aligments);
+        } else {
+            TEST_ASSERT( buf != NULL );
+            printf("[ALIGNED_ALLOC] alignment required: %u \n", alignments);
+            printf("[ALIGNED_ALLOC] address of allocated memory: %p \n\n", (void *)buf);
+            //Address of obtained block must be aligned with selected value
+            TEST_ASSERT(((intptr_t)buf & (alignments - 1)) == 0);
+
+            //Write some data, if it corrupts memory probably the heap
+            //canary verification will fail:
+            memset(buf, 0xA5, (10*1024));
+            heap_caps_free(buf);
+        }
+    }
+#endif
+
+}
+
+TEST_CASE("aligned_alloc(0) should return a NULL pointer", "[heap]")
+{
+    void *p;
+    p = heap_caps_aligned_alloc(32, 0, MALLOC_CAP_DEFAULT);
+    TEST_ASSERT(p == NULL);
+}

components/heap/test/test_allocator_timings.c

@@ -0,0 +1,108 @@
+#include "freertos/FreeRTOS.h"
+#include <esp_types.h>
+#include <stdio.h>
+#include "unity.h"
+#include "esp_attr.h"
+#include "esp_heap_caps.h"
+#include <stdlib.h>
+#include <sys/param.h>
+#include <string.h>
+#include <test_utils.h>
+
+//This test only makes sense with poisoning disabled (light or comprehensive)
+#if !defined(CONFIG_HEAP_POISONING_COMPREHENSIVE) && !defined(CONFIG_HEAP_POISONING_LIGHT)
+
+#define NUM_POINTERS 128
+#define ITERATIONS 10000
+
+TEST_CASE("Heap many random allocations timings", "[heap]")
+{
+    void *p[NUM_POINTERS] = { 0 };
+    size_t s[NUM_POINTERS] = { 0 };
+
+    uint32_t cycles_before;
+    uint64_t alloc_time_average = 0;
+    uint64_t free_time_average = 0;
+    uint64_t realloc_time_average = 0;
+
+    for (int i = 0; i < ITERATIONS; i++) {
+        uint8_t n = esp_random() % NUM_POINTERS;
+
+        if (esp_random() % 4 == 0) {
+            /* 1 in 4 iterations, try to realloc the buffer instead
+               of using malloc/free
+            */
+            size_t new_size = esp_random() % 1024;
+
+            cycles_before = portGET_RUN_TIME_COUNTER_VALUE();
+            void *new_p = heap_caps_realloc(p[n], new_size, MALLOC_CAP_DEFAULT);
+            realloc_time_average = portGET_RUN_TIME_COUNTER_VALUE() - cycles_before;
+
+            printf("realloc %p -> %p (%zu -> %zu) time spent cycles: %lld \n", p[n], new_p, s[n], new_size, realloc_time_average);
+            heap_caps_check_integrity(MALLOC_CAP_DEFAULT, true);
+            if (new_size == 0 || new_p != NULL) {
+                p[n] = new_p;
+                s[n] = new_size;
+                if (new_size > 0) {
+                    memset(p[n], n, new_size);
+                }
+            }
+            continue;
+        }
+
+        if (p[n] != NULL) {
+            if (s[n] > 0) {
+                /* Verify pre-existing contents of p[n] */
+                uint8_t compare[s[n]];
+                memset(compare, n, s[n]);
+                TEST_ASSERT(( memcmp(compare, p[n], s[n]) == 0 ));
+            }
+            TEST_ASSERT(heap_caps_check_integrity(MALLOC_CAP_DEFAULT, true));
+
+            cycles_before = portGET_RUN_TIME_COUNTER_VALUE();
+            heap_caps_free(p[n]);
+            free_time_average = portGET_RUN_TIME_COUNTER_VALUE() - cycles_before;
+
+            printf("freed %p (%zu) time spent cycles: %lld\n", p[n], s[n], free_time_average);
+
+            if (!heap_caps_check_integrity(MALLOC_CAP_DEFAULT, true)) {
+                printf("FAILED iteration %d after freeing %p\n", i, p[n]);
+                heap_caps_dump(MALLOC_CAP_DEFAULT);
+                TEST_ASSERT(0);
+            }
+        }
+
+        s[n] = rand() % 1024;
+        heap_caps_check_integrity(MALLOC_CAP_DEFAULT, true);
+        cycles_before = portGET_RUN_TIME_COUNTER_VALUE();
+        p[n] = heap_caps_malloc(s[n], MALLOC_CAP_DEFAULT);
+        alloc_time_average = portGET_RUN_TIME_COUNTER_VALUE() - cycles_before;
+
+        printf("malloc %p (%zu) time spent cycles: %lld \n", p[n], s[n], alloc_time_average);
+
+        if (!heap_caps_check_integrity(MALLOC_CAP_DEFAULT, true)) {
+            printf("FAILED iteration %d after mallocing %p (%zu bytes)\n", i, p[n], s[n]);
+            heap_caps_dump(MALLOC_CAP_DEFAULT);
+            TEST_ASSERT(0);
+        }
+
+        if (p[n] != NULL) {
+            memset(p[n], n, s[n]);
+        }
+    }
+
+    for (int i = 0; i < NUM_POINTERS; i++) {
+        cycles_before = portGET_RUN_TIME_COUNTER_VALUE();
+        heap_caps_free( p[i]);
+        free_time_average = portGET_RUN_TIME_COUNTER_VALUE() - cycles_before;
+
+        if (!heap_caps_check_integrity(MALLOC_CAP_DEFAULT, true)) {
+            printf("FAILED during cleanup after freeing %p\n", p[i]);
+            heap_caps_dump(MALLOC_CAP_DEFAULT);
+            TEST_ASSERT(0);
+        }
+    }
+
+    TEST_ASSERT(heap_caps_check_integrity(MALLOC_CAP_DEFAULT, true));
+}
+#endif

components/heap/test/test_diram.c

@@ -0,0 +1,74 @@
+/*
+ Tests for D/IRAM support in heap capability allocator
+*/
+
+#include <esp_types.h>
+#include <stdio.h>
+#include "unity.h"
+#include "esp_heap_caps.h"
+#include "soc/soc_memory_layout.h"
+
+#define ALLOC_SZ 1024
+
+static void *malloc_block_diram(uint32_t caps)
+{
+    void *attempts[256] = { 0 }; // Allocate up to 256 ALLOC_SZ blocks to exhaust all non-D/IRAM memory temporarily
+    int count = 0;
+    void *result;
+
+    while(count < sizeof(attempts)/sizeof(void *)) {
+        result = heap_caps_malloc(ALLOC_SZ, caps);
+        TEST_ASSERT_NOT_NULL_MESSAGE(result, "not enough free heap to perform test");
+
+        if (esp_ptr_in_diram_dram(result) || esp_ptr_in_diram_iram(result)) {
+            break;
+        }
+
+        attempts[count] = result;
+        result = NULL;
+        count++;
+    }
+
+    for (int i = 0; i < count; i++) {
+        free(attempts[i]);
+    }
+
+    TEST_ASSERT_NOT_NULL_MESSAGE(result, "not enough D/IRAM memory is free");
+    return result;
+}
+
+TEST_CASE("Allocate D/IRAM as DRAM", "[heap]")
+{
+    uint32_t *dram = malloc_block_diram(MALLOC_CAP_8BIT | MALLOC_CAP_INTERNAL);
+
+    for (int i = 0; i < ALLOC_SZ / sizeof(uint32_t); i++) {
+        uint32_t v = i + 0xAAAA;
+        dram[i] = v;
+        volatile uint32_t *iram = esp_ptr_diram_dram_to_iram(dram + i);
+        TEST_ASSERT_EQUAL(v, dram[i]);
+        TEST_ASSERT_EQUAL(v, *iram);
+        *iram = UINT32_MAX;
+        TEST_ASSERT_EQUAL(UINT32_MAX, *iram);
+        TEST_ASSERT_EQUAL(UINT32_MAX, dram[i]);
+    }
+
+    free(dram);
+}
+
+TEST_CASE("Allocate D/IRAM as IRAM", "[heap]")
+{
+    uint32_t *iram = malloc_block_diram(MALLOC_CAP_EXEC);
+
+    for (int i = 0; i < ALLOC_SZ / sizeof(uint32_t); i++) {
+        uint32_t v = i + 0xEEE;
+        iram[i] = v;
+        volatile uint32_t *dram = esp_ptr_diram_iram_to_dram(iram + i);
+        TEST_ASSERT_EQUAL_HEX32(v, iram[i]);
+        TEST_ASSERT_EQUAL_HEX32(v, *dram);
+        *dram = UINT32_MAX;
+        TEST_ASSERT_EQUAL_HEX32(UINT32_MAX, *dram);
+        TEST_ASSERT_EQUAL_HEX32(UINT32_MAX, iram[i]);
+    }
+
+    free(iram);
+}

components/heap/test/test_heap_trace.c

@@ -0,0 +1,164 @@
+/*
+ Generic test for heap tracing support
+
+ Only compiled in if CONFIG_HEAP_TRACING is set
+*/
+
+#include <esp_types.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include "sdkconfig.h"
+#include "unity.h"
+
+#include "freertos/FreeRTOS.h"
+#include "freertos/task.h"
+
+#ifdef CONFIG_HEAP_TRACING
+// only compile in heap tracing tests if tracing is enabled
+
+#include "esp_heap_trace.h"
+
+TEST_CASE("heap trace leak check", "[heap]")
+{
+    heap_trace_record_t recs[8];
+    heap_trace_init_standalone(recs, 8);
+
+    printf("Leak check test\n"); // Print something before trace starts, or stdout allocations skew total counts
+    fflush(stdout);
+
+    heap_trace_start(HEAP_TRACE_LEAKS);
+
+    void *a = malloc(64);
+    memset(a, '3', 64);
+
+    void *b = malloc(96);
+    memset(b, '4', 11);
+
+    printf("a.address %p vs %p b.address %p vs %p\n", a, recs[0].address, b, recs[1].address);
+
+    heap_trace_dump();
+    TEST_ASSERT_EQUAL(2, heap_trace_get_count());
+
+    heap_trace_record_t trace_a, trace_b;
+    heap_trace_get(0, &trace_a);
+    heap_trace_get(1, &trace_b);
+
+    printf("trace_a.address %p trace_bb.address %p\n", trace_a.address, trace_b.address);
+
+    TEST_ASSERT_EQUAL_PTR(a, trace_a.address);
+    TEST_ASSERT_EQUAL_PTR(b, trace_b.address);
+
+    TEST_ASSERT_EQUAL_PTR(recs[0].address, trace_a.address);
+    TEST_ASSERT_EQUAL_PTR(recs[1].address, trace_b.address);
+
+    free(a);
+
+    TEST_ASSERT_EQUAL(1, heap_trace_get_count());
+
+    heap_trace_get(0, &trace_b);
+    TEST_ASSERT_EQUAL_PTR(b, trace_b.address);
+
+    /* buffer deletes trace_a when freed,
+       so trace_b at head of buffer */
+    TEST_ASSERT_EQUAL_PTR(recs[0].address, trace_b.address);
+
+    heap_trace_stop();
+}
+
+TEST_CASE("heap trace wrapped buffer check", "[heap]")
+{
+    const size_t N = 8;
+    heap_trace_record_t recs[N];
+    heap_trace_init_standalone(recs, N);
+
+    heap_trace_start(HEAP_TRACE_LEAKS);
+
+    void *ptrs[N+1];
+    for (int i = 0; i < N+1; i++) {
+        ptrs[i] = malloc(i*3);
+    }
+
+    // becuase other mallocs happen as part of this control flow,
+    // we can't guarantee N entries of ptrs[] are in the heap check buffer.
+    // but we should guarantee at least the last one is
+    bool saw_last_ptr = false;
+    for (int i = 0; i < N; i++) {
+        heap_trace_record_t rec;
+        heap_trace_get(i, &rec);
+        if (rec.address == ptrs[N-1]) {
+            saw_last_ptr = true;
+        }
+    }
+    TEST_ASSERT(saw_last_ptr);
+
+    void *other = malloc(6);
+
+    heap_trace_dump();
+
+    for (int i = 0; i < N+1; i++) {
+        free(ptrs[i]);
+    }
+
+    heap_trace_dump();
+
+    bool saw_other = false;
+
+    for (int i = 0; i < heap_trace_get_count(); i++) {
+        heap_trace_record_t rec;
+        heap_trace_get(i, &rec);
+
+        // none of ptr[]s should be in the heap trace any more
+        for (int j = 0; j < N+1; j++) {
+            TEST_ASSERT_NOT_EQUAL(ptrs[j], rec.address);
+        }
+        if (rec.address == other) {
+            saw_other = true;
+        }
+    }
+
+    // 'other' pointer should be somewhere in the leak dump
+    TEST_ASSERT(saw_other);
+
+    heap_trace_stop();
+}
+
+static void print_floats_task(void *ignore)
+{
+    heap_trace_start(HEAP_TRACE_ALL);
+    char buf[16] = { };
+    volatile float f = 12.3456;
+    sprintf(buf, "%.4f", f);
+    TEST_ASSERT_EQUAL_STRING("12.3456", buf);
+    heap_trace_stop();
+
+    vTaskDelete(NULL);
+}
+
+TEST_CASE("can trace allocations made by newlib", "[heap]")
+{
+    const size_t N = 8;
+    heap_trace_record_t recs[N];
+    heap_trace_init_standalone(recs, N);
+
+    /* Verifying that newlib code performs an allocation is very fiddly:
+
+       - Printing a float allocates data associated with the task, but only the
+       first time a task prints a float of this length. So we do it in a one-shot task
+       to avoid possibility it already happened.
+
+       - If newlib is updated this test may start failing if the printf() implementation
+       changes. (This version passes for both nano & regular formatting in newlib 2.2.0)
+
+       - We also do the tracing in the task so we only capture things directly related to it.
+    */
+
+    xTaskCreate(print_floats_task, "print_float", 4096, NULL, 5, NULL);
+    vTaskDelay(10);
+
+    /* has to be at least a few as newlib allocates via multiple different function calls */
+    TEST_ASSERT(heap_trace_get_count() > 3);
+}
+
+
+#endif

components/heap/test/test_leak.c

@@ -0,0 +1,60 @@
+/*
+ Tests for a leak tag
+*/
+
+#include <stdio.h>
+#include "unity.h"
+#include "esp_heap_caps_init.h"
+#include "esp_system.h"
+#include <stdlib.h>
+
+
+static char* check_calloc(int size)
+{
+    char *arr = calloc(size, sizeof(char));
+    TEST_ASSERT_NOT_NULL(arr);
+    return arr;
+}
+
+TEST_CASE("Check for leaks (no leak)", "[heap]")
+{
+    char *arr = check_calloc(1000);
+    free(arr);
+}
+
+TEST_CASE("Check for leaks (leak)", "[heap][ignore]")
+{
+    check_calloc(1000);
+}
+
+TEST_CASE("Not check for leaks", "[heap][leaks]")
+{
+    check_calloc(1000);
+}
+
+TEST_CASE("Set a leak level = 7016", "[heap][leaks=7016]")
+{
+    check_calloc(7000);
+}
+
+static void test_fn(void)
+{
+    check_calloc(1000);
+}
+
+TEST_CASE_MULTIPLE_STAGES("Not check for leaks in MULTIPLE_STAGES mode", "[heap][leaks]", test_fn, test_fn, test_fn);
+
+TEST_CASE_MULTIPLE_STAGES("Check for leaks in MULTIPLE_STAGES mode (leak)", "[heap][ignore]", test_fn, test_fn, test_fn);
+
+static void test_fn2(void)
+{
+    check_calloc(1000);
+    esp_restart();
+}
+
+static void test_fn3(void)
+{
+    check_calloc(1000);
+}
+
+TEST_CASE_MULTIPLE_STAGES("Check for leaks in MULTIPLE_STAGES mode (manual reset)", "[heap][leaks][reset=SW_CPU_RESET, SW_CPU_RESET]", test_fn2, test_fn2, test_fn3);

components/heap/test/test_malloc.c

@@ -0,0 +1,134 @@
+/*
+ Generic test for malloc/free
+*/
+
+#include <esp_types.h>
+#include <stdio.h>
+
+#include "freertos/FreeRTOS.h"
+#include "freertos/task.h"
+#include "freertos/semphr.h"
+#include "freertos/queue.h"
+#include "unity.h"
+#include "esp_heap_caps.h"
+
+#include "sdkconfig.h"
+
+
+static int **allocatedMem;
+static int noAllocated;
+
+
+static int tryAllocMem(void) {
+    int i, j;
+    const int allocateMaxK=1024*5; //try to allocate a max of 5MiB
+
+    allocatedMem=malloc(sizeof(int *)*allocateMaxK);
+    if (!allocatedMem) return 0;
+
+    for (i=0; i<allocateMaxK; i++) {
+        allocatedMem[i]=malloc(1024);
+        if (allocatedMem[i]==NULL) break;
+        for (j=0; j<1024/4; j++) allocatedMem[i][j]=(0xdeadbeef);
+    }
+    noAllocated=i;
+    return i;
+}
+
+
+static void tryAllocMemFree(void) {
+    int i, j;
+    for (i=0; i<noAllocated; i++) {
+        for (j=0; j<1024/4; j++) {
+            TEST_ASSERT(allocatedMem[i][j]==(0xdeadbeef));
+        }
+        free(allocatedMem[i]);
+    }
+    free(allocatedMem);
+}
+
+
+TEST_CASE("Malloc/overwrite, then free all available DRAM", "[heap]")
+{
+    int m1=0, m2=0;
+    m1=tryAllocMem();
+    tryAllocMemFree();
+    m2=tryAllocMem();
+    tryAllocMemFree();
+    printf("Could allocate %dK on first try, %dK on 2nd try.\n", m1, m2);
+    TEST_ASSERT(m1==m2);
+}
+
+#if CONFIG_SPIRAM_USE_MALLOC
+
+#if (CONFIG_SPIRAM_MALLOC_RESERVE_INTERNAL > 1024)
+TEST_CASE("Check if reserved DMA pool still can allocate even when malloc()'ed memory is exhausted", "[heap]")
+{
+    char** dmaMem=malloc(sizeof(char*)*512);
+    assert(dmaMem);
+    int m=tryAllocMem();
+    int i=0;
+    for (i=0; i<512; i++) {
+        dmaMem[i]=heap_caps_malloc(1024, MALLOC_CAP_DMA);
+        if (dmaMem[i]==NULL) break;
+    }
+    for (int j=0; j<i; j++) free(dmaMem[j]);
+    free(dmaMem);
+    tryAllocMemFree();
+    printf("Could allocate %dK of DMA memory after allocating all of %dK of normal memory.\n", i, m);
+    TEST_ASSERT(i);
+}
+#endif
+
+#endif
+
+
+/* As you see, we are desperately trying to outsmart the compiler, so that it
+ * doesn't warn about oversized allocations in the next two unit tests.
+ * To be removed when we switch to GCC 8.2 and add
+ * -Wno-alloc-size-larger-than=PTRDIFF_MAX to CFLAGS for this file.
+ */
+void* (*g_test_malloc_ptr)(size_t) = &malloc;
+void* (*g_test_calloc_ptr)(size_t, size_t) = &calloc;
+
+void* test_malloc_wrapper(size_t size)
+{
+    return (*g_test_malloc_ptr)(size);
+}
+
+void* test_calloc_wrapper(size_t count, size_t size)
+{
+    return (*g_test_calloc_ptr)(count, size);
+}
+
+TEST_CASE("alloc overflows should all fail", "[heap]")
+{
+    /* allocates 8 bytes if size_t overflows */
+    TEST_ASSERT_NULL(test_calloc_wrapper(SIZE_MAX / 2 + 4, 2));
+
+    /* will overflow if any poisoning is enabled
+       (should fail for sensible OOM reasons, otherwise) */
+    TEST_ASSERT_NULL(test_malloc_wrapper(SIZE_MAX - 1));
+    TEST_ASSERT_NULL(test_calloc_wrapper(SIZE_MAX - 1, 1));
+
+    /* will overflow when the size is rounded up to word align it */
+    TEST_ASSERT_NULL(heap_caps_malloc(SIZE_MAX-1, MALLOC_CAP_32BIT));
+
+    TEST_ASSERT_NULL(heap_caps_malloc(SIZE_MAX-1, MALLOC_CAP_EXEC));
+}
+
+TEST_CASE("unreasonable allocs should all fail", "[heap]")
+{
+    TEST_ASSERT_NULL(test_calloc_wrapper(16, 1024*1024));
+    TEST_ASSERT_NULL(test_malloc_wrapper(16*1024*1024));
+    TEST_ASSERT_NULL(test_malloc_wrapper(SIZE_MAX / 2));
+    TEST_ASSERT_NULL(test_malloc_wrapper(SIZE_MAX - 256));
+    TEST_ASSERT_NULL(test_malloc_wrapper(xPortGetFreeHeapSize() - 1));
+}
+
+TEST_CASE("malloc(0) should return a NULL pointer", "[heap]")
+{
+    void *p;
+    p = malloc(0);
+    TEST_ASSERT(p == NULL);
+}

components/heap/test/test_malloc_caps.c

@@ -0,0 +1,247 @@
+/*
+ Tests for the capabilities-based memory allocator.
+*/
+
+#include <esp_types.h>
+#include <stdio.h>
+#include "unity.h"
+#include "esp_attr.h"
+#include "esp_heap_caps.h"
+#include "esp_spi_flash.h"
+#include <stdlib.h>
+#include <sys/param.h>
+
+#ifndef CONFIG_ESP_SYSTEM_MEMPROT_FEATURE
+TEST_CASE("Capabilities allocator test", "[heap]")
+{
+    char *m1, *m2[10];
+    int x;
+    size_t free8start, free32start, free8, free32;
+
+    /* It's important we printf() something before we take the empty heap sizes,
+       as the first printf() in a task allocates heap resources... */
+    printf("Testing capabilities allocator...\n");
+
+    free8start = heap_caps_get_free_size(MALLOC_CAP_8BIT);
+    free32start = heap_caps_get_free_size(MALLOC_CAP_32BIT);
+    printf("Free 8bit-capable memory (start): %dK, 32-bit capable memory %dK\n", free8start, free32start);
+    TEST_ASSERT(free32start >= free8start);
+
+    printf("Allocating 10K of 8-bit capable RAM\n");
+    m1= heap_caps_malloc(10*1024, MALLOC_CAP_8BIT);
+    printf("--> %p\n", m1);
+    free8 = heap_caps_get_free_size(MALLOC_CAP_8BIT);
+    free32 = heap_caps_get_free_size(MALLOC_CAP_32BIT);
+    printf("Free 8bit-capable memory (both reduced): %dK, 32-bit capable memory %dK\n", free8, free32);
+    //Both should have gone down by 10K; 8bit capable ram is also 32-bit capable
+    TEST_ASSERT(free8<=(free8start-10*1024));
+    TEST_ASSERT(free32<=(free32start-10*1024));
+    //Assume we got DRAM back
+    TEST_ASSERT((((int)m1)&0xFF000000)==0x3F000000);
+    free(m1);
+
+    //The goal here is to allocate from IRAM. Since there is no external IRAM (yet)
+    //the following gives size of IRAM-only (not D/IRAM) memory.
+    size_t free_iram = heap_caps_get_free_size(MALLOC_CAP_INTERNAL) -
+                           heap_caps_get_free_size(MALLOC_CAP_8BIT | MALLOC_CAP_INTERNAL);
+    size_t alloc32 = MIN(free_iram / 2, 10*1024) & (~3);
+    if(free_iram) {
+        printf("Freeing; allocating %u bytes of 32K-capable RAM\n", alloc32);
+        m1 = heap_caps_malloc(alloc32, MALLOC_CAP_32BIT);
+        printf("--> %p\n", m1);
+        //Check that we got IRAM back
+        TEST_ASSERT((((int)m1)&0xFF000000)==0x40000000);
+        free8 = heap_caps_get_free_size(MALLOC_CAP_8BIT);
+        free32 = heap_caps_get_free_size(MALLOC_CAP_32BIT);
+        printf("Free 8bit-capable memory (after 32-bit): %dK, 32-bit capable memory %dK\n", free8, free32);
+        //Only 32-bit should have gone down by alloc32: 32-bit isn't necessarily 8bit capable
+        TEST_ASSERT(free32<=(free32start-alloc32));
+        TEST_ASSERT(free8==free8start);
+        free(m1);
+    } else {
+        printf("This platform has no 32-bit only capable RAM, jumping to next test \n");
+    }
+
+    printf("Allocating impossible caps\n");
+    m1= heap_caps_malloc(10*1024, MALLOC_CAP_8BIT|MALLOC_CAP_EXEC);
+    printf("--> %p\n", m1);
+    TEST_ASSERT(m1==NULL);
+
+    if(free_iram) {
+        printf("Testing changeover iram -> dram");
+        // priorities will exhaust IRAM first, then start allocating from DRAM
+        for (x=0; x<10; x++) {
+            m2[x]= heap_caps_malloc(alloc32, MALLOC_CAP_32BIT);
+            printf("--> %p\n", m2[x]);
+        }
+        TEST_ASSERT((((int)m2[0])&0xFF000000)==0x40000000);
+        TEST_ASSERT((((int)m2[9])&0xFF000000)==0x3F000000);
+
+    } else {
+        printf("This platform has no IRAM-only so changeover will never occur, jumping to next test\n");
+    }
+
+    printf("Test if allocating executable code still gives IRAM, even with dedicated IRAM region depleted\n");
+    if(free_iram) {
+        // (the allocation should come from D/IRAM)
+        free_iram = heap_caps_get_free_size(MALLOC_CAP_EXEC);
+        m1= heap_caps_malloc(MIN(free_iram / 2, 10*1024), MALLOC_CAP_EXEC);
+        printf("--> %p\n", m1);
+        TEST_ASSERT((((int)m1)&0xFF000000)==0x40000000);
+        for (x=0; x<10; x++) free(m2[x]);
+
+    } else {
+        // (the allocation should come from D/IRAM)
+        free_iram = heap_caps_get_free_size(MALLOC_CAP_EXEC);
+        m1= heap_caps_malloc(MIN(free_iram / 2, 10*1024), MALLOC_CAP_EXEC);
+        printf("--> %p\n", m1);
+        TEST_ASSERT((((int)m1)&0xFF000000)==0x40000000);
+    }
+
+    free(m1);
+    printf("Done.\n");
+}
+#endif
+
+#ifdef CONFIG_ESP32_IRAM_AS_8BIT_ACCESSIBLE_MEMORY
+TEST_CASE("IRAM_8BIT capability test", "[heap]")
+{
+    uint8_t *ptr;
+    size_t free_size, free_size32, largest_free_size;
+
+    /* need to print something as first printf allocates some heap */
+    printf("IRAM_8BIT capability test\n");
+
+    free_size = heap_caps_get_free_size(MALLOC_CAP_IRAM_8BIT);
+    free_size32 = heap_caps_get_free_size(MALLOC_CAP_32BIT);
+
+    largest_free_size = heap_caps_get_largest_free_block(MALLOC_CAP_IRAM_8BIT);
+
+    ptr = heap_caps_malloc(largest_free_size, MALLOC_CAP_IRAM_8BIT);
+
+    TEST_ASSERT((((int)ptr)&0xFF000000)==0x40000000);
+
+    TEST_ASSERT(heap_caps_get_free_size(MALLOC_CAP_IRAM_8BIT) == (free_size - heap_caps_get_allocated_size(ptr)));
+    TEST_ASSERT(heap_caps_get_free_size(MALLOC_CAP_32BIT) == (free_size32 - heap_caps_get_allocated_size(ptr)));
+
+    free(ptr);
+}
+#endif
+
+TEST_CASE("heap_caps metadata test", "[heap]")
+{
+    /* need to print something as first printf allocates some heap */
+    printf("heap_caps metadata test\n");
+    heap_caps_print_heap_info(MALLOC_CAP_8BIT);
+
+    multi_heap_info_t original;
+    heap_caps_get_info(&original, MALLOC_CAP_8BIT);
+
+    void *b = heap_caps_malloc(original.largest_free_block, MALLOC_CAP_8BIT);
+    TEST_ASSERT_NOT_NULL(b);
+
+    printf("After allocating %d bytes:\n", original.largest_free_block);
+    heap_caps_print_heap_info(MALLOC_CAP_8BIT);
+
+    multi_heap_info_t after;
+    heap_caps_get_info(&after, MALLOC_CAP_8BIT);
+    TEST_ASSERT(after.largest_free_block <= original.largest_free_block);
+    TEST_ASSERT(after.total_free_bytes <= original.total_free_bytes);
+
+    free(b);
+    heap_caps_get_info(&after, MALLOC_CAP_8BIT);
+
+    printf("\n\n After test, heap status:\n");
+    heap_caps_print_heap_info(MALLOC_CAP_8BIT);
+
+    /* Allow some leeway here, because LWIP sometimes allocates up to 144 bytes in the background
+       as part of timer management.
+    */
+    TEST_ASSERT_INT32_WITHIN(200, after.total_free_bytes, original.total_free_bytes);
+    TEST_ASSERT_INT32_WITHIN(200, after.largest_free_block, original.largest_free_block);
+    TEST_ASSERT(after.minimum_free_bytes < original.total_free_bytes);
+}
+
+/* Small function runs from IRAM to check that malloc/free/realloc
+   all work OK when cache is disabled...
+*/
+static IRAM_ATTR __attribute__((noinline)) bool iram_malloc_test(void)
+{
+    spi_flash_guard_get()->start(); // Disables flash cache
+
+    bool result = true;
+    void *x = heap_caps_malloc(64, MALLOC_CAP_EXEC);
+    result = result && (x != NULL);
+    void *y = heap_caps_realloc(x, 32, MALLOC_CAP_EXEC);
+    result = result && (y != NULL);
+    heap_caps_free(y);
+
+    spi_flash_guard_get()->end(); // Re-enables flash cache
+
+    return result;
+}
+
+
+TEST_CASE("heap_caps_xxx functions work with flash cache disabled", "[heap]")
+{
+    TEST_ASSERT( iram_malloc_test() );
+}
+
+#ifdef CONFIG_HEAP_ABORT_WHEN_ALLOCATION_FAILS
+TEST_CASE("When enabled, allocation operation failure generates an abort", "[heap][reset=abort,SW_CPU_RESET]")
+{
+    const size_t stupid_allocation_size = (128 * 1024 * 1024);
+    void *ptr = heap_caps_malloc(stupid_allocation_size, MALLOC_CAP_DEFAULT);
+    (void)ptr;
+    TEST_FAIL_MESSAGE("should not be reached");
+}
+#endif
+
+static bool called_user_failed_hook = false;
+
+void heap_caps_alloc_failed_hook(size_t requested_size, uint32_t caps, const char *function_name)
+{
+    printf("%s was called but failed to allocate %d bytes with 0x%X capabilities. \n",function_name, requested_size, caps);
+    called_user_failed_hook = true;
+}
+
+TEST_CASE("user provided alloc failed hook must be called when allocation fails", "[heap]")
+{
+    TEST_ASSERT(heap_caps_register_failed_alloc_callback(heap_caps_alloc_failed_hook) == ESP_OK);
+
+    const size_t stupid_allocation_size = (128 * 1024 * 1024);
+    void *ptr = heap_caps_malloc(stupid_allocation_size, MALLOC_CAP_DEFAULT);
+    TEST_ASSERT(called_user_failed_hook != false);
+
+    called_user_failed_hook = false;
+    ptr = heap_caps_realloc(ptr, stupid_allocation_size, MALLOC_CAP_DEFAULT);
+    TEST_ASSERT(called_user_failed_hook != false);
+
+    called_user_failed_hook = false;
+    ptr = heap_caps_aligned_alloc(0x200, stupid_allocation_size, MALLOC_CAP_DEFAULT);
+    TEST_ASSERT(called_user_failed_hook != false);
+
+    (void)ptr;
+}
+
+TEST_CASE("allocation with invalid capability should also trigger the alloc failed hook", "[heap]")
+{
+    const size_t allocation_size = 64;
+    const uint32_t invalid_cap = MALLOC_CAP_INVALID;
+
+    TEST_ASSERT(heap_caps_register_failed_alloc_callback(heap_caps_alloc_failed_hook) == ESP_OK);
+
+    called_user_failed_hook = false;
+    void *ptr = heap_caps_malloc(allocation_size, invalid_cap);
+    TEST_ASSERT(called_user_failed_hook != false);
+
+    called_user_failed_hook = false;
+    ptr = heap_caps_realloc(ptr, allocation_size, invalid_cap);
+    TEST_ASSERT(called_user_failed_hook != false);
+
+    called_user_failed_hook = false;
+    ptr = heap_caps_aligned_alloc(0x200, allocation_size, invalid_cap);
+    TEST_ASSERT(called_user_failed_hook != false);
+
+    (void)ptr;
+}

components/heap/test/test_realloc.c

@@ -0,0 +1,67 @@
+/*
+ Generic test for realloc
+*/
+
+#include <stdlib.h>
+#include <string.h>
+#include "unity.h"
+#include "sdkconfig.h"
+#include "esp_heap_caps.h"
+#include "soc/soc_memory_layout.h"
+
+
+#ifndef CONFIG_HEAP_POISONING_COMPREHENSIVE
+/* (can't realloc in place if comprehensive is enabled) */
+
+TEST_CASE("realloc shrink buffer in place", "[heap]")
+{
+    void *x = malloc(64);
+    TEST_ASSERT(x);
+    void *y = realloc(x, 48);
+    TEST_ASSERT_EQUAL_PTR(x, y);
+}
+
+#endif
+
+#ifndef CONFIG_ESP_SYSTEM_MEMPROT_FEATURE
+TEST_CASE("realloc shrink buffer with EXEC CAPS", "[heap]")
+{
+    const size_t buffer_size = 64;
+
+    void *x = heap_caps_malloc(buffer_size, MALLOC_CAP_EXEC);
+    TEST_ASSERT(x);
+    void *y = heap_caps_realloc(x, buffer_size - 16, MALLOC_CAP_EXEC);
+    TEST_ASSERT(y);
+
+    //y needs to fall in a compatible memory area of IRAM:
+    TEST_ASSERT(esp_ptr_executable(y)|| esp_ptr_in_iram(y) || esp_ptr_in_diram_iram(y));
+
+    free(y);
+}
+
+TEST_CASE("realloc move data to a new heap type", "[heap]")
+{
+    const char *test = "I am some test content to put in the heap";
+    char buf[64];
+    memset(buf, 0xEE, 64);
+    strlcpy(buf, test, 64);
+
+    char *a = malloc(64);
+    memcpy(a, buf, 64);
+    // move data from 'a' to IRAM
+    char *b = heap_caps_realloc(a, 64, MALLOC_CAP_EXEC);
+    TEST_ASSERT_NOT_NULL(b);
+    TEST_ASSERT_NOT_EQUAL(a, b);
+    TEST_ASSERT(heap_caps_check_integrity(MALLOC_CAP_INVALID, true));
+    TEST_ASSERT_EQUAL_HEX32_ARRAY(buf, b, 64 / sizeof(uint32_t));
+
+    // Move data back to DRAM
+    char *c = heap_caps_realloc(b, 48, MALLOC_CAP_8BIT);
+    TEST_ASSERT_NOT_NULL(c);
+    TEST_ASSERT_NOT_EQUAL(b, c);
+    TEST_ASSERT(heap_caps_check_integrity(MALLOC_CAP_INVALID, true));
+    TEST_ASSERT_EQUAL_HEX8_ARRAY(buf, c, 48);
+
+    free(c);
+}
+#endif

components/heap/test/test_runtime_heap_reg.c

@@ -0,0 +1,72 @@
+/*
+ Tests for registering new heap memory at runtime
+*/
+
+#include <stdio.h>
+#include "unity.h"
+#include "esp_heap_caps_init.h"
+#include "esp_system.h"
+#include <stdlib.h>
+
+
+/* NOTE: This is not a well-formed unit test, it leaks memory */
+TEST_CASE("Allocate new heap at runtime", "[heap][ignore]")
+{
+    const size_t BUF_SZ = 1000;
+    const size_t HEAP_OVERHEAD_MAX = 200;
+    void *buffer = malloc(BUF_SZ);
+    TEST_ASSERT_NOT_NULL(buffer);
+    uint32_t before_free = esp_get_free_heap_size();
+    TEST_ESP_OK( heap_caps_add_region((intptr_t)buffer, (intptr_t)buffer + BUF_SZ) );
+    uint32_t after_free = esp_get_free_heap_size();
+    printf("Before %u after %u\n", before_free, after_free);
+    /* allow for some 'heap overhead' from accounting structures */
+    TEST_ASSERT(after_free >= before_free + BUF_SZ - HEAP_OVERHEAD_MAX);
+}
+
+/* NOTE: This is not a well-formed unit test, it leaks memory and
+   may fail if run twice in a row without a reset.
+*/
+TEST_CASE("Allocate new heap with new capability", "[heap][ignore]")
+{
+    const size_t BUF_SZ = 100;
+#ifdef CONFIG_ESP_SYSTEM_MEMPROT_FEATURE
+    const size_t ALLOC_SZ = 32;
+#else
+    const size_t ALLOC_SZ = 64; // More than half of BUF_SZ
+#endif
+    const uint32_t MALLOC_CAP_INVENTED = (1 << 30); /* this must be unused in esp_heap_caps.h */
+
+    /* no memory exists to provide this capability */
+    TEST_ASSERT_NULL( heap_caps_malloc(ALLOC_SZ, MALLOC_CAP_INVENTED) );
+
+    void *buffer = malloc(BUF_SZ);
+    TEST_ASSERT_NOT_NULL(buffer);
+    uint32_t caps[SOC_MEMORY_TYPE_NO_PRIOS] = { MALLOC_CAP_INVENTED };
+    TEST_ESP_OK( heap_caps_add_region_with_caps(caps, (intptr_t)buffer, (intptr_t)buffer + BUF_SZ) );
+
+    /* ta-da, it's now possible! */
+    TEST_ASSERT_NOT_NULL( heap_caps_malloc(ALLOC_SZ, MALLOC_CAP_INVENTED) );
+}
+
+/* NOTE: This is not a well-formed unit test.
+ * If run twice without a reset, it will failed.
+ */
+
+TEST_CASE("Add .bss memory to heap region runtime", "[heap][ignore]")
+{
+#define BUF_SZ 1000
+#define HEAP_OVERHEAD_MAX 200
+    static uint8_t s_buffer[BUF_SZ];
+
+    printf("s_buffer start %08x end %08x\n", (intptr_t)s_buffer, (intptr_t)s_buffer + BUF_SZ);
+    uint32_t before_free = esp_get_free_heap_size();
+    TEST_ESP_OK( heap_caps_add_region((intptr_t)s_buffer, (intptr_t)s_buffer + BUF_SZ) );
+    uint32_t after_free = esp_get_free_heap_size();
+    printf("Before %u after %u\n", before_free, after_free);
+    /* allow for some 'heap overhead' from accounting structures */
+    TEST_ASSERT(after_free >= before_free + BUF_SZ - HEAP_OVERHEAD_MAX);
+
+    /* Twice add must be failed */
+    TEST_ASSERT( (heap_caps_add_region((intptr_t)s_buffer, (intptr_t)s_buffer + BUF_SZ) != ESP_OK) );
+}

components/heap/test_multi_heap_host/Makefile

@@ -0,0 +1,54 @@
+TEST_PROGRAM=test_multi_heap
+all: $(TEST_PROGRAM)
+
+ifneq ($(filter clean,$(MAKECMDGOALS)),)
+.NOTPARALLEL:  # prevent make clean racing the other targets
+endif
+
+SOURCE_FILES = $(abspath \
+    ../multi_heap.c \
+    ../heap_tlsf.c \
+	../multi_heap_poisoning.c \
+	test_multi_heap.cpp \
+	main.cpp \
+    )
+
+INCLUDE_FLAGS = -I../include -I../../../tools/catch
+
+GCOV ?= gcov
+
+CPPFLAGS += $(INCLUDE_FLAGS) -D CONFIG_LOG_DEFAULT_LEVEL -g -fstack-protector-all -m32  -DCONFIG_HEAP_POISONING_COMPREHENSIVE
+CFLAGS += -Wall -Werror -fprofile-arcs -ftest-coverage
+CXXFLAGS += -std=c++11 -Wall -Werror  -fprofile-arcs -ftest-coverage
+LDFLAGS += -lstdc++ -fprofile-arcs -ftest-coverage -m32
+
+OBJ_FILES = $(filter %.o, $(SOURCE_FILES:.cpp=.o) $(SOURCE_FILES:.c=.o))
+
+COVERAGE_FILES = $(OBJ_FILES:.o=.gc*)
+
+$(TEST_PROGRAM): $(OBJ_FILES)
+	g++ $(LDFLAGS) -o $(TEST_PROGRAM) $(OBJ_FILES)
+
+$(OUTPUT_DIR):
+	mkdir -p $(OUTPUT_DIR)
+
+test: $(TEST_PROGRAM)
+	./$(TEST_PROGRAM)
+
+$(COVERAGE_FILES): $(TEST_PROGRAM) test
+
+coverage.info: $(COVERAGE_FILES)
+	find ../ -name "*.gcno" -exec $(GCOV) -r -pb {} +
+	lcov --capture --directory $(abspath ../) --no-external --output-file coverage.info --gcov-tool $(GCOV)
+
+coverage_report: coverage.info
+	genhtml coverage.info --output-directory coverage_report
+	@echo "Coverage report is in coverage_report/index.html"
+
+clean:
+	rm -f $(OBJ_FILES) $(TEST_PROGRAM)
+	rm -f $(COVERAGE_FILES) *.gcov
+	rm -rf coverage_report/
+	rm -f coverage.info
+
+.PHONY: clean all test

components/heap/test_multi_heap_host/main.cpp

@@ -0,0 +1,2 @@
+#define CATCH_CONFIG_MAIN
+#include "catch.hpp"