Skip to main content

Process a built Yocto project to create a Black Duck project version

Project description

Synopsys Import Yocto Build Manifest - import_yocto_bm.py - v2.8

INTRODUCTION

This script is provided under an OSS license as an example of how to use the Black Duck APIs to import components from a Yocto project manifest.

It does not represent any extension of licensed functionality of Synopsys software itself and is provided as-is, without warranty or liability.

If you have comments or issues, please raise a GitHub issue here. Synopsys support is not able to respond to support tickets for this OSS utility.

OVERVIEW

The import_yocto_bm.py script is designed to import a Yocto project build manifest created by Bitbake. It replaces all previous scripts (including https://github.com/matthewb66/import_yocto_build_manifest).

It can be used as an alternative to the standard Yocto scan process for Black Duck provided within Synopsys Detect, and provides additional capabilities including checking recipes against the Black Duck KB, replacing recipe specifications where required and propagating locally patched CVEs to the Black Duck project.

Version 2.1 provides an optional wizard mode to request additional information where it is not provided (use --nowizard to disable this mode for batch operation).

YOCTO SCANNING IN BLACK DUCK USING SYNOPSYS DETECT

The default Yocto scan process for Black Duck is to determine Bitbake dependencies using Synopsys Detect (see Synopsys Detect - scanning Yocto).

You should NOT use Black Duck Signature scanning for whole Yocto projects (although Signature scanning is useful for custom recipes which may contain OSS which you want to identify). Signature scanning will not function accurately on Yocto projects because the code is modified locally by change/diff files and no reference copy of the changed code is available, meaning that Signature matching will not be able to determine the correct Yocto modified OSS components. Yocto projects are also very large and Signature scanning can consume large volumes of server resources without providing accurate matching.

When used to scan a Yocto project, Synopsys Detect calls Bitbake to report the project layers and recipes, checking against the recipes reported at OpenEmbedded.org (refer to layers.openembedded.org to identify layers and recipes which will be mapped) and creating a Black Duck project containing the mapped components. Recipes in layers not reported at OpenEmbedded.org will not be imported into Black Duck, and you should consider using other Black Duck scan techniques to identify OSS within specific custom recipes referenced in the build. Note also that moving original OSS recipes to new layers will also stop them being reported.

The resulting project will contain mapped recipes/layers including those used to create the image but which may not delivered in the built image. Recipe matching also requires that OSS components are used within standard recipes/revisions and within the original layer. If you copy a recipe to a new or different layer, it will not be matched in the Black Duck project by Synopsys Detect.

To perform a Yocto scan using Synopsys Detect:

  • Change to the poky folder of a Yocto project
  • Run Synopsys Detect adding the options --detect.tools=DETECTOR --detect.bitbake.package.names=core-image-sato (where core-image-sato is the package name).
  • New bitbake options have been added in Synopsys Detect 7.9 including the ability to locate the manifest files and remove build dependencies ( for example --detect.bitbake.dependency.types.excluded=BUILD - see (here)[https://community.synopsys.com/s/document-item?bundleId=integrations-detect&topicId=properties%2Fdetectors%2Fbitbake.html] ).

WHY THIS SCRIPT

This script provides an alternative to scanning Yocto projects in Synopsys Detect.

It should be considered in the following scenarios:

  • Where many standard OpenEmbedded recipes have been moved to new layers (meaning they will not be matched by Synopsys Detect)
  • Where OpenEmbedded recipe versions or revisions have been modified from the original
  • Where a report is required for custom recipes so they can be scanned independently
  • The identification of locally patched CVEs is desired

The script operates on a built Yocto project, by identifying the build manifest containing only the recipes which are within the built image as opposed to the superset of all recipes used to build the distribution (including build tools etc.) produced by Synopsys Detect by default. This script also optionally supports extracting a list of locally patched CVEs from Bitbake and marking them as patched in the Black Duck project.

Recipes/layers are checked against the recipes maintained in the Black Duck KB which have been extracted from OpenEmbedded.org and those which match are added to the created Black Duck project.

The script should be executed on a Linux workstation where Yocto has been installed and after a successful Bitbake build.

If invoked in the Yocto build folder (or the build folder is manually specified using the -y option), then it will locate the license-manifest file automatically in the tmp/deploy hierarchy. Alternatively you can specify the license.manifest file as a script option or the default wizard mode will ask you to specify the file.

If the Bitbake build was performed with the Yocto cve_check class configured, then the script will also locate the CVE log exported by CVE check, extract patched CVEs and set the remediation status of matching CVEs in the Black Duck project.

The script requires access to the Black Duck server via the API (see Prerequisites below) unless the -o option is used to create the output scan file for manual upload.

OVERALL SCAN PROCESS USING IMPORT_YOCTO_BM

Using this script will allow standard OSS recipes to be imported to the Black Duck project. It will also identify where standard OSS recipes have been moved to new or different layers, and match close revisions or versions which exist in the Black Duck KB.

The proposed process to scan a Yocto project using this script is:

  1. Use the import_yocto_bm to identify original OSS recipes within the project and create a Black Duck project version.

  2. Add the --report rep.txt option to export a report of the matched recipes. This will include reports categorised as follows:

  • OK = recipe matched the Black Duck KB and will be included in the Black Duck project
  • REPLACED = recipe has been moved to a new layer and the script has referenced the original layer in the KB or the recipe was changed using a replacefile entry - will be included in the Black Duck project
  • REPLACED_NOREVISION = the script has replaced the revision to match the KB - will be included in project
  • REPLACED_NOLAYER+REVISION = recipe has been moved to a new layer with a new revision and the script has referenced the original layer and revision in the KB - will be included in the Black Duck project
  • NOTREPLACED_NOVERSION = the layer and recipe exist in the KB but the version does not - will not be included in the Black Duck project
  • NOTREPLACED_NOLAYER+VERSION = the recipe exists in the KB but the layer and version do not - will not be included in the Black Duck project
  • MISSING = recipe does not exist in the KB and will not be included in the Black Duck project
  • SKIPPED = recipe is not mapped in bblayers data (corrupt yocto project?) - skipped
  1. For the NOTREPLACED_NOVERSION and NOTREPLACED_NOLAYER+VERSION recipes, you could consider using the --replacefile repfile option to map to layers/recipes and version/revisions which exist in the KB, rerunning the script to import them. See the section REPLACING RECIPE NAMES below.

  2. For the MISSING recipes, these are almost certainly custom recipes containing either a mix of custom application code (potentially with OSS sources) or new OSS not obtained from Openembedded.org as standard OSS recipes. Identify the layers which contain mainly custom recipes, and use standard Black Duck signature scanning (and optionally snippet scanning) to search for modified OSS within these sub-folders only. Do not run a signature or snippet scan on the entire Yocto project - only use signature scanning for custom layers and recipes not mapped by this script. You can optionally combine the dependency and signature scans in the same Black Duck project.

SUPPORTED YOCTO PROJECTS

This script is designed to support Yocto versions 2.0 up to 3.3 for building projects.

OSS components from OpenEmbedded recipes maintained at layers.openbedded.org should be detected by the scan. Additional OSS components managed by custom recipes will not be detected.

PREREQUISITES

  1. Must be run on Linux

  2. Python 3 must be installed.

  3. A supported Yocto environment (version 2.0 to 3.1) should be loaded to the current shell (see Preconfiguration section below). Alternatively use the --bblayer_out option to specify the output of the bitbake-layers show-recipes command run separately.

  4. The Yocto project must have been pre-built with a license.manifest files available.

  5. OPTIONAL: For patched CVE remediation in the Black Duck project, you will need to add the cve_check bbclass to the Yocto build configuration to generate the CVE check log output. Add the following line to the build/conf/local.conf file:

    INHERIT += "cve-check"
    

Then use the Yocto build command (e.g. bitbake core-image-sato which will incrementally build without needing to rerun the full build, but will add the CVE check action to generate the log files.

INSTALLATION

Install the utility using pip - for example:

pip3 install import_yocto_bm

Alternatively, clone the repository and run directly using:

python3 import_yocto_bm/main.py

STANDARD USAGE

Run the command import_yocto_bm without arguments to invoke the wizard to guide you through the required information and options.

The import_yocto_bm parameters for command line usage are shown below:

usage: import_yocto_bm [-h] [-p PROJECT] [-v VERSION] [-y YOCTO_FOLDER]
			[-t TARGET] [-o OUTPUT_JSON] [-m MANIFEST]
			[-b BUILDCONF] [-l LOCALCONF] [--arch ARCH]
			[-r repfile] [--cve_check_only] [--no_cve_check]
			[--cve_check_file CVE_CHECK_FILE] [--report rep.txt] [--nowizard|-wizard]

Import Yocto build manifest to BD project version

optional arguments:
  -h, --help            show this help message and exit
  --blackduck_url BLACKDUCK_URL
                        Black Duck server URL (also uses $BLACKDUCK_URL environment variable)
  --blackduck_api_token BLACKDUCK_API_TOKEN
                        Black Duck API token (also uses $BLACKDUCK_API_TOKEN environment variable)
  --blackduck_trust_cert
                        Black Duck trust server cert
  -p PROJECT, --project PROJECT
			Black Duck project to create (REQUIRED)
  -v VERSION, --version VERSION
			Black Duck project version to create (REQUIRED)
  -y YOCTO_FOLDER, --yocto_build_folder YOCTO_FOLDER
  			Yocto build folder (required if CVE check required or
			manifest file not specified)
  -o OUTPUT_JSON, --output_json OUTPUT_JSON
			Output JSON bom file for manual import to Black Duck
			(instead of uploading the scan automatically)
  -t TARGET, --target TARGET
			Yocto target (default core-poky-sato)
  -m MANIFEST, --manifest MANIFEST
			Input build manifest file (if not specified will be
			determined from conf files) - must be the 
			license.manifest file (not build.manifest)
  -b BUILDCONF, --buildconf BUILDCONF
			Build config file (if not specified 
			poky/meta/conf/bitbake.conf will be used)
  -l LOCALCONF, --localconf LOCALCONF
			Local config file (if not specified 
			poky/build/conf/local.conf will be used)
  -r REPFILE, --replacefile REPFILE
  			Replace file used to replace layer and recipe names
  --arch ARCH           Architecture (if not specified then will be determined
			from conf files)
  --cve_check_only      Only check for patched CVEs from cve_check and update
			existing project
  --no_cve_check        Skip check for and update of patched CVEs
  --cve_check_file CVE_CHECK_FILE
			CVE check output file (if not specified will be
			determined from conf files)
  --no_kb_check         Do not check recipes against KB
  --kb_recipe_dir KB_RECIPE_DIR
                        KB recipe data directory local copy
  --report rep.txt	If KB check is performed, produce a list of matched. modified and unmatched recipes
  --bblayers_out bbfile
  			Can be used to scan a build without access to the build environment - a file 
			containing the output of the command `bitbake-layers show-recipes '*'` instead of 
			calling the command directly as well as the license.manifest file to be specified 
			with -m. Also skips identification of recipe revisions (assumes all revisions are -r0)
  --wizard
            Start command line wizard (Wizard will run by default if config incomplete)
  --nowizard
            Do not start command line wizard even if config incomplete (batch mode)

The script will use the invocation folder as the Yocto build folder (e.g. yocto_zeus/poky/build) by default (if there is a build sub-folder then it will be used instead). The --yocto_folder option can be used to specify the Yocto build folder as opposed to the invocation folder.

The --project and --version options are required to define the Black Duck project and version names for inclusion in the json output file (and update CVE patch status).

The --output_json option can be used to specify an output file for the project scan. If specified, then the scan will not be uploaded automatically and CVE patch checking will be skipped.

The --replacefile option can be used to specify a layer/recipe/version replacement file (see REPLACING LAYER AND RECIPE NAMES below).

The Yocto target and architecture values are required to locate the manifest and cve_check log files and will be extracted from the Bitbake config files automatically, but the --target and --arch options can be used to specify these manually.

The most recent Bitbake output manifest file (located in the build/tmp/deploy/licenses/<image>-<target>-<datetime>/license.manifest file) will be located automatically. Use the --manifest option to specify the manifest file manually.

The most recent cve_check log file build/tmp/deploy/images/<arch>/<image>-<target>-<datetime>.rootfs.cve will be located automatically if it exists. Use the --cve_check_file option to specify the cve_check log file location manually (for example to use an older copy).

Use the --cve_check_only option to skip the scanning of the project and creation of a project, only looking for a CVE check output log file to identify and patching matched CVEs within an existing Black Duck project (which must have been created previously).

Use the --no_cve_check option to skip the patched CVE identification and update of CVE status in the Black Duck project.

PRECONFIGURATION

The import_yocto_bm script calls Bitbake commands by default, so you will need to run the Yocto project config (change the location as required) before using the script - for example:

cd /home/users/myuser/yocto_zeus/poky
source oe-init-build-env

This will set the Yocto environment and change directory into the Yocto build sub-folder.

Alternatively, you can run the bitbake-layer show-recipes command separately to create an output file, then use the --bblayers_out file option (change file to the output file name) or specify the file in the wizard mode.

If you want to upload the resulting project direct to Black Duck (and are not using the --output_json option), you specify the Black Duck server with URL & API_TOKEN either using command line options:

  --blackduck_url SERVER_URL
  --blackduck_api_token TOKEN
  --blackduck_trust_cert (specify if untrusted CA certificate used for BD server)

OR by setting environment variables:

  BLACKDUCK_URL=https://SERVER_URL
  BLACKDUCK_API_TOKEN=TOKEN

Where SERVER_URL is the Black Duck server URL and TOKEN is the Black Duck API token.

REPLACING RECIPE NAMES

Layers and recipes extracted from the project are combined by Black Duck and used to lookup original OSS components at https://layers.openembedded.org. If OSS components are moved from original layers to a new (custom) or different layer which is not shown at https://layers.openembedded.org and which cannot be mapped by a close neighbour in the script, then they will not be mapped in the resulting Black Duck project.

To reference a specific original component, you can use the --replacefile REPFILE option to map OSS components back to original recipes.

The replacefile option can also be used to remap new OSS component versions (not listed at https://layers.openembedded.org) to previous versions which are listed.

Example REPFILE content is shown below:

RECIPE meta-customlayer/alsa-lib/1.2.1.2-r5 meta/alsa-lib/1.2.1.2-r0

which will remap recipe and version alsa-lib/1.2.1.2-r5 in the meta-customlayer to meta/alsa-lib/1.2.1.2-r0 in teh Black Duck KB.

EXAMPLE USAGE

Check the Preconfiguration section above before running the script.

To run the utility in wizard mode, simply use the command import_yocto_bm and it will ask questions to determine the scan parameters.

Use the option --nowizard to run in batch mode and bypass the wizard mode, noting that you will need to specify all required options on the command line correctly.

Use the following command to scan a Yocto build, create Black Duck project myproject and version v1.0, then update CVE patch status for identified CVEs (will require the OE environment to have been loaded previously):

import_yocto_bm -p myproject -v v1.0

To scan a Yocto project specifying a different build manifest as opposed to the most recent one (will require the OE environment to have been loaded previously):

import_yocto_bm -p myproject -v v1.0 -m tmp/deploy/licenses/core-image-sato-qemux86-64-20200728105751/package.manifest

To scan the most recent Yocto build in a different build folder location (not the current folder - will require the OE environment to have been loaded previously):

import_yocto_bm -p myproject -v v1.0 --y $HOME/newyocto/poky/build

To perform a CVE check patch analysis only (to update an existing Black Duck project created previously by the script with patched vulnerabilities) use the command (will require the OE environment to have been loaded previously):

import_yocto_bm -p myproject -v v1.0 --cve_check_only

To create a JSON output scan without uploading (and no CVE patch update) use (will require the OE environment to have been loaded previously):

import_yocto_bm -p myproject -v v1.0 -o my.jsonld

To create a JSON output scan without uploading (and no CVE patch update) without checking recipes against the KB recipe list downloaded from Github (will require the OE environment to have been loaded previously):

import_yocto_bm -p myproject -v v1.0 -o my.jsonld --no_kb_check

If you are unable to load the OE environment prior to running the utility, ensure you have the output of the command bitbake-layers show-recipes in a file, and use the option --bblayers_out OUTPUT_FILE to specify the file OUTPUT_FILE containing this output, for example:

import_yocto_bm -p myproject -v v1.0 -m poky/build/tmp/output/license.manifest --bblayers_out bblayers.txt

CVEs from cve_check Versus Black Duck

The Yocto cve_check class works on the Bitbake dependencies within the dev environment, and produces a list of CVEs identified from the NVD for ALL packages in the development environment.

This script extracts the packages from the build manifest (which will be a subset of those in the full Bitbake dependencies for build environment) and creates a Black Duck project.

The list of CVEs reported by cve_check will therefore be considerably larger than seen in the Black Duck project (which is the expected situation).

OUTSTANDING ISSUES

The identification of the Linux Kernel version from the Bitbake recipes and association with the upstream component in the KB has not been completed yet. Until an automatic identification is possible, the required Linux Kernel component can be added manually to the Black Duck project.

UPDATE HISTORY

V2.8

  • Updated kb data & fixed bug refering to old file

V2.7

  • Fixed issue with trustcert option usage

V2.6

  • Some wizard fixes

V2.5

  • Addressed API limits introduced in 2022.2

V2.4

  • Fixed another issue with replacefile

V2.3

  • Fixed issue with replacefile usage

V2.2

  • Various minor fixes

V2.1

  • Added extra information to report output file
  • Added defaults to wizard mode & check for bd_trustcert

V2.0

  • Added new Wizard mode
  • Other fixes to manage component version epochs
  • Updated KB data

V1.17 Beta

  • Update to new KB data

V1.16 Beta

  • Fixed CVE patch process to use new Client API class

V1.15 Beta

  • Changed KB recipe download to use json formatted files
  • Changed --kb_recipe_file to --kb_recipe_dir to support json files

V1.14 Beta

  • Reworked replacefile logic and improved console reporting

V1.13 Beta

  • replaced the --debug option with --bblayers_out with the ability to define the input file.
  • Fixed the --report option to use the specified report file (previously always wrote to report.txt)
  • Changed the recipe matching logic to also look for KB matches without revisions (-rX) to support older Yocto versions

V1.12

  • Minor bug fixes

V1.11

  • Fixed a minor issue with AUTOINC component versions.

Previous Versions

Versions 1.10 and 1.9 added fixes for the KB recipe lookup, and a new report output showing the list of recipes matched, modified or not matchable (use --report reort.txt option) as well as updating the list of KB recipes located in the data file.

Version 1.8 will automatically download a list of all Yocto projects stored in the Black Duck KB and checks the recipes in the local project, remapping recipes to origin layers or revisions to reduce the number of non matched components in the resulting Black Duck project. This functionality can be skipped using the --no_kb_check option, or if the download of the KB data from Github is not possible within the script, the --kb_recipe_file option allows a local copy to be used.

The DEPLOY_DIR and TMP_DIR values extracted from conf files now support the use of environment variables. Other fixes have been applied especially to the processing of the recipeinfo files and extraction of recipe revisions.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

import_yocto_bm-2.8.tar.gz (32.9 kB view details)

Uploaded Source

Built Distribution

import_yocto_bm-2.8-py3-none-any.whl (28.6 kB view details)

Uploaded Python 3

File details

Details for the file import_yocto_bm-2.8.tar.gz.

File metadata

  • Download URL: import_yocto_bm-2.8.tar.gz
  • Upload date:
  • Size: 32.9 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/4.0.2 CPython/3.11.3

File hashes

Hashes for import_yocto_bm-2.8.tar.gz
Algorithm Hash digest
SHA256 66d83e4a08d466e646a1a746eeec3ada74162fa973a7a82140497504ee57fd7f
MD5 69275cba3204683ce6182d57e7e14f34
BLAKE2b-256 3d2a54aba83599bef8ca2b6a0b5b04162508a0619c0236799e4d703190f47842

See more details on using hashes here.

File details

Details for the file import_yocto_bm-2.8-py3-none-any.whl.

File metadata

File hashes

Hashes for import_yocto_bm-2.8-py3-none-any.whl
Algorithm Hash digest
SHA256 10e188bc800a225bd98877066ee11a96216ab0eb36c4ef1806515487b51c1f7d
MD5 f8841ee203837514db7c17b8f6b1bb64
BLAKE2b-256 13c1faffade957f464fbecb3e6f742e7c60f3590e98336a832e74b52ce8f6162

See more details on using hashes here.

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page