xref: /openbmc/openbmc/poky/meta/classes/create-spdx-2.2.bbclass (revision c9537f57ab488bf5d90132917b0184e2527970a5)
1#
2# Copyright OpenEmbedded Contributors
3#
4# SPDX-License-Identifier: GPL-2.0-only
5#
6
7inherit spdx-common
8
9SPDX_VERSION = "2.2"
10
11SPDX_ORG ??= "OpenEmbedded ()"
12SPDX_SUPPLIER ??= "Organization: ${SPDX_ORG}"
13SPDX_SUPPLIER[doc] = "The SPDX PackageSupplier field for SPDX packages created from \
14    this recipe. For SPDX documents create using this class during the build, this \
15    is the contact information for the person or organization who is doing the \
16    build."
17
18SPDX_ARCHIVE_SOURCES ??= "0"
19SPDX_ARCHIVE_PACKAGED ??= "0"
20
21def get_namespace(d, name):
22    import uuid
23    namespace_uuid = uuid.uuid5(uuid.NAMESPACE_DNS, d.getVar("SPDX_UUID_NAMESPACE"))
24    return "%s/%s-%s" % (d.getVar("SPDX_NAMESPACE_PREFIX"), name, str(uuid.uuid5(namespace_uuid, name)))
25
26
27def create_annotation(d, comment):
28    from datetime import datetime, timezone
29
30    creation_time = datetime.now(tz=timezone.utc).strftime("%Y-%m-%dT%H:%M:%SZ")
31    annotation = oe.spdx.SPDXAnnotation()
32    annotation.annotationDate = creation_time
33    annotation.annotationType = "OTHER"
34    annotation.annotator = "Tool: %s - %s" % (d.getVar("SPDX_TOOL_NAME"), d.getVar("SPDX_TOOL_VERSION"))
35    annotation.comment = comment
36    return annotation
37
38def recipe_spdx_is_native(d, recipe):
39    return any(a.annotationType == "OTHER" and
40      a.annotator == "Tool: %s - %s" % (d.getVar("SPDX_TOOL_NAME"), d.getVar("SPDX_TOOL_VERSION")) and
41      a.comment == "isNative" for a in recipe.annotations)
42
43def get_json_indent(d):
44    if d.getVar("SPDX_PRETTY") == "1":
45        return 2
46    return None
47
48
49def convert_license_to_spdx(lic, license_data, document, d, existing={}):
50    from pathlib import Path
51    import oe.spdx
52
53    extracted = {}
54
55    def add_extracted_license(ident, name):
56        nonlocal document
57
58        if name in extracted:
59            return
60
61        extracted_info = oe.spdx.SPDXExtractedLicensingInfo()
62        extracted_info.name = name
63        extracted_info.licenseId = ident
64        extracted_info.extractedText = None
65
66        if name == "PD":
67            # Special-case this.
68            extracted_info.extractedText = "Software released to the public domain"
69        else:
70            # Seach for the license in COMMON_LICENSE_DIR and LICENSE_PATH
71            for directory in [d.getVar('COMMON_LICENSE_DIR')] + (d.getVar('LICENSE_PATH') or '').split():
72                try:
73                    with (Path(directory) / name).open(errors="replace") as f:
74                        extracted_info.extractedText = f.read()
75                        break
76                except FileNotFoundError:
77                    pass
78            if extracted_info.extractedText is None:
79                # If it's not SPDX or PD, then NO_GENERIC_LICENSE must be set
80                entry = d.getVarFlag('NO_GENERIC_LICENSE', name).split(';')
81                filename = entry[0]
82                params = {i.split('=')[0]: i.split('=')[1] for i in entry[1:] if '=' in i}
83                beginline = int(params.get('beginline', 1))
84                endline = params.get('endline', None)
85                if endline:
86                    endline = int(endline)
87                if filename:
88                    filename = d.expand("${S}/" + filename)
89                    with open(filename, errors="replace") as f:
90                        extracted_info.extractedText = "".join(line for idx, line in enumerate(f, 1) if beginline <= idx and idx <= (endline or idx))
91                else:
92                    bb.fatal("Cannot find any text for license %s" % name)
93
94        extracted[name] = extracted_info
95        document.hasExtractedLicensingInfos.append(extracted_info)
96
97    def convert(l):
98        if l == "(" or l == ")":
99            return l
100
101        if l == "&":
102            return "AND"
103
104        if l == "|":
105            return "OR"
106
107        if l == "CLOSED":
108            return "NONE"
109
110        spdx_license = d.getVarFlag("SPDXLICENSEMAP", l) or l
111        if spdx_license in license_data["licenses"]:
112            return spdx_license
113
114        try:
115            spdx_license = existing[l]
116        except KeyError:
117            spdx_license = "LicenseRef-" + l
118            add_extracted_license(spdx_license, l)
119
120        return spdx_license
121
122    lic_split = lic.replace("(", " ( ").replace(")", " ) ").replace("|", " | ").replace("&", " & ").split()
123
124    return ' '.join(convert(l) for l in lic_split)
125
126def add_package_files(d, doc, spdx_pkg, topdir, get_spdxid, get_types, *, archive=None, ignore_dirs=[], ignore_top_level_dirs=[]):
127    from pathlib import Path
128    import oe.spdx
129    import oe.spdx_common
130    import hashlib
131
132    source_date_epoch = d.getVar("SOURCE_DATE_EPOCH")
133    if source_date_epoch:
134        source_date_epoch = int(source_date_epoch)
135
136    sha1s = []
137    spdx_files = []
138
139    file_counter = 1
140    for subdir, dirs, files in os.walk(topdir):
141        dirs[:] = [d for d in dirs if d not in ignore_dirs]
142        if subdir == str(topdir):
143            dirs[:] = [d for d in dirs if d not in ignore_top_level_dirs]
144
145        for file in files:
146            filepath = Path(subdir) / file
147            filename = str(filepath.relative_to(topdir))
148
149            if not filepath.is_symlink() and filepath.is_file():
150                spdx_file = oe.spdx.SPDXFile()
151                spdx_file.SPDXID = get_spdxid(file_counter)
152                for t in get_types(filepath):
153                    spdx_file.fileTypes.append(t)
154                spdx_file.fileName = filename
155
156                if archive is not None:
157                    with filepath.open("rb") as f:
158                        info = archive.gettarinfo(fileobj=f)
159                        info.name = filename
160                        info.uid = 0
161                        info.gid = 0
162                        info.uname = "root"
163                        info.gname = "root"
164
165                        if source_date_epoch is not None and info.mtime > source_date_epoch:
166                            info.mtime = source_date_epoch
167
168                        archive.addfile(info, f)
169
170                sha1 = bb.utils.sha1_file(filepath)
171                sha1s.append(sha1)
172                spdx_file.checksums.append(oe.spdx.SPDXChecksum(
173                        algorithm="SHA1",
174                        checksumValue=sha1,
175                    ))
176                spdx_file.checksums.append(oe.spdx.SPDXChecksum(
177                        algorithm="SHA256",
178                        checksumValue=bb.utils.sha256_file(filepath),
179                    ))
180
181                if "SOURCE" in spdx_file.fileTypes:
182                    extracted_lics = oe.spdx_common.extract_licenses(filepath)
183                    if extracted_lics:
184                        spdx_file.licenseInfoInFiles = extracted_lics
185
186                doc.files.append(spdx_file)
187                doc.add_relationship(spdx_pkg, "CONTAINS", spdx_file)
188                spdx_pkg.hasFiles.append(spdx_file.SPDXID)
189
190                spdx_files.append(spdx_file)
191
192                file_counter += 1
193
194    sha1s.sort()
195    verifier = hashlib.sha1()
196    for v in sha1s:
197        verifier.update(v.encode("utf-8"))
198    spdx_pkg.packageVerificationCode.packageVerificationCodeValue = verifier.hexdigest()
199
200    return spdx_files
201
202
203def add_package_sources_from_debug(d, package_doc, spdx_package, package, package_files, sources):
204    from pathlib import Path
205    import hashlib
206    import oe.packagedata
207    import oe.spdx
208
209    debug_search_paths = [
210        Path(d.getVar('PKGD')),
211        Path(d.getVar('STAGING_DIR_TARGET')),
212        Path(d.getVar('STAGING_DIR_NATIVE')),
213        Path(d.getVar('STAGING_KERNEL_DIR')),
214    ]
215
216    pkg_data = oe.packagedata.read_subpkgdata_extended(package, d)
217
218    if pkg_data is None:
219        return
220
221    for file_path, file_data in pkg_data["files_info"].items():
222        if not "debugsrc" in file_data:
223            continue
224
225        for pkg_file in package_files:
226            if file_path.lstrip("/") == pkg_file.fileName.lstrip("/"):
227                break
228        else:
229            bb.fatal("No package file found for %s in %s; SPDX found: %s" % (str(file_path), package,
230                " ".join(p.fileName for p in package_files)))
231            continue
232
233        for debugsrc in file_data["debugsrc"]:
234            ref_id = "NOASSERTION"
235            for search in debug_search_paths:
236                if debugsrc.startswith("/usr/src/kernel"):
237                    debugsrc_path = search / debugsrc.replace('/usr/src/kernel/', '')
238                else:
239                    debugsrc_path = search / debugsrc.lstrip("/")
240                # We can only hash files below, skip directories, links, etc.
241                if not os.path.isfile(debugsrc_path):
242                    continue
243
244                file_sha256 = bb.utils.sha256_file(debugsrc_path)
245
246                if file_sha256 in sources:
247                    source_file = sources[file_sha256]
248
249                    doc_ref = package_doc.find_external_document_ref(source_file.doc.documentNamespace)
250                    if doc_ref is None:
251                        doc_ref = oe.spdx.SPDXExternalDocumentRef()
252                        doc_ref.externalDocumentId = "DocumentRef-dependency-" + source_file.doc.name
253                        doc_ref.spdxDocument = source_file.doc.documentNamespace
254                        doc_ref.checksum.algorithm = "SHA1"
255                        doc_ref.checksum.checksumValue = source_file.doc_sha1
256                        package_doc.externalDocumentRefs.append(doc_ref)
257
258                    ref_id = "%s:%s" % (doc_ref.externalDocumentId, source_file.file.SPDXID)
259                else:
260                    bb.debug(1, "Debug source %s with SHA256 %s not found in any dependency" % (str(debugsrc_path), file_sha256))
261                break
262            else:
263                bb.debug(1, "Debug source %s not found" % debugsrc)
264
265            package_doc.add_relationship(pkg_file, "GENERATED_FROM", ref_id, comment=debugsrc)
266
267add_package_sources_from_debug[vardepsexclude] += "STAGING_KERNEL_DIR"
268
269def collect_dep_recipes(d, doc, spdx_recipe):
270    import json
271    from pathlib import Path
272    import oe.sbom
273    import oe.spdx
274    import oe.spdx_common
275
276    deploy_dir_spdx = Path(d.getVar("DEPLOY_DIR_SPDX"))
277    package_archs = d.getVar("SPDX_MULTILIB_SSTATE_ARCHS").split()
278    package_archs.reverse()
279
280    dep_recipes = []
281
282    deps = oe.spdx_common.get_spdx_deps(d)
283
284    for dep in deps:
285        # If this dependency is not calculated in the taskhash skip it.
286        # Otherwise, it can result in broken links since this task won't
287        # rebuild and see the new SPDX ID if the dependency changes
288        if not dep.in_taskhash:
289            continue
290
291        dep_recipe_path = oe.sbom.doc_find_by_hashfn(deploy_dir_spdx, package_archs, "recipe-" + dep.pn, dep.hashfn)
292        if not dep_recipe_path:
293            bb.fatal("Cannot find any SPDX file for recipe %s, %s" % (dep.pn, dep.hashfn))
294
295        spdx_dep_doc, spdx_dep_sha1 = oe.sbom.read_doc(dep_recipe_path)
296
297        for pkg in spdx_dep_doc.packages:
298            if pkg.name == dep.pn:
299                spdx_dep_recipe = pkg
300                break
301        else:
302            continue
303
304        dep_recipes.append(oe.sbom.DepRecipe(spdx_dep_doc, spdx_dep_sha1, spdx_dep_recipe))
305
306        dep_recipe_ref = oe.spdx.SPDXExternalDocumentRef()
307        dep_recipe_ref.externalDocumentId = "DocumentRef-dependency-" + spdx_dep_doc.name
308        dep_recipe_ref.spdxDocument = spdx_dep_doc.documentNamespace
309        dep_recipe_ref.checksum.algorithm = "SHA1"
310        dep_recipe_ref.checksum.checksumValue = spdx_dep_sha1
311
312        doc.externalDocumentRefs.append(dep_recipe_ref)
313
314        doc.add_relationship(
315            "%s:%s" % (dep_recipe_ref.externalDocumentId, spdx_dep_recipe.SPDXID),
316            "BUILD_DEPENDENCY_OF",
317            spdx_recipe
318        )
319
320    return dep_recipes
321
322collect_dep_recipes[vardepsexclude] = "SPDX_MULTILIB_SSTATE_ARCHS"
323
324def collect_dep_sources(d, dep_recipes):
325    import oe.sbom
326
327    sources = {}
328    for dep in dep_recipes:
329        # Don't collect sources from native recipes as they
330        # match non-native sources also.
331        if recipe_spdx_is_native(d, dep.recipe):
332            continue
333        recipe_files = set(dep.recipe.hasFiles)
334
335        for spdx_file in dep.doc.files:
336            if spdx_file.SPDXID not in recipe_files:
337                continue
338
339            if "SOURCE" in spdx_file.fileTypes:
340                for checksum in spdx_file.checksums:
341                    if checksum.algorithm == "SHA256":
342                        sources[checksum.checksumValue] = oe.sbom.DepSource(dep.doc, dep.doc_sha1, dep.recipe, spdx_file)
343                        break
344
345    return sources
346
347def add_download_packages(d, doc, recipe):
348    import os.path
349    from bb.fetch2 import decodeurl, CHECKSUM_LIST
350    import bb.process
351    import oe.spdx
352    import oe.sbom
353
354    for download_idx, src_uri in enumerate(d.getVar('SRC_URI').split()):
355        f = bb.fetch2.FetchData(src_uri, d)
356
357        package = oe.spdx.SPDXPackage()
358        package.name = "%s-source-%d" % (d.getVar("PN"), download_idx + 1)
359        package.SPDXID = oe.sbom.get_download_spdxid(d, download_idx + 1)
360
361        if f.type == "file":
362            continue
363
364        if f.method.supports_checksum(f):
365            for checksum_id in CHECKSUM_LIST:
366                if checksum_id.upper() not in oe.spdx.SPDXPackage.ALLOWED_CHECKSUMS:
367                    continue
368
369                expected_checksum = getattr(f, "%s_expected" % checksum_id)
370                if expected_checksum is None:
371                    continue
372
373                c = oe.spdx.SPDXChecksum()
374                c.algorithm = checksum_id.upper()
375                c.checksumValue = expected_checksum
376                package.checksums.append(c)
377
378        package.downloadLocation = oe.spdx_common.fetch_data_to_uri(f, f.name)
379        doc.packages.append(package)
380        doc.add_relationship(doc, "DESCRIBES", package)
381        # In the future, we might be able to do more fancy dependencies,
382        # but this should be sufficient for now
383        doc.add_relationship(package, "BUILD_DEPENDENCY_OF", recipe)
384
385def get_license_list_version(license_data, d):
386    # Newer versions of the SPDX license list are SemVer ("MAJOR.MINOR.MICRO"),
387    # but SPDX 2 only uses "MAJOR.MINOR".
388    return ".".join(license_data["licenseListVersion"].split(".")[:2])
389
390
391python do_create_spdx() {
392    from datetime import datetime, timezone
393    import oe.sbom
394    import oe.spdx
395    import oe.spdx_common
396    import uuid
397    from pathlib import Path
398    from contextlib import contextmanager
399    import oe.cve_check
400
401    license_data = oe.spdx_common.load_spdx_license_data(d)
402
403    @contextmanager
404    def optional_tarfile(name, guard, mode="w"):
405        import tarfile
406        import bb.compress.zstd
407
408        num_threads = int(d.getVar("BB_NUMBER_THREADS"))
409
410        if guard:
411            name.parent.mkdir(parents=True, exist_ok=True)
412            with bb.compress.zstd.open(name, mode=mode + "b", num_threads=num_threads) as f:
413                with tarfile.open(fileobj=f, mode=mode + "|") as tf:
414                    yield tf
415        else:
416            yield None
417
418
419    deploy_dir_spdx = Path(d.getVar("DEPLOY_DIR_SPDX"))
420    spdx_workdir = Path(d.getVar("SPDXWORK"))
421    include_sources = d.getVar("SPDX_INCLUDE_SOURCES") == "1"
422    archive_sources = d.getVar("SPDX_ARCHIVE_SOURCES") == "1"
423    archive_packaged = d.getVar("SPDX_ARCHIVE_PACKAGED") == "1"
424    pkg_arch = d.getVar("SSTATE_PKGARCH")
425
426    creation_time = datetime.now(tz=timezone.utc).strftime("%Y-%m-%dT%H:%M:%SZ")
427
428    doc = oe.spdx.SPDXDocument()
429
430    doc.name = "recipe-" + d.getVar("PN")
431    doc.documentNamespace = get_namespace(d, doc.name)
432    doc.creationInfo.created = creation_time
433    doc.creationInfo.comment = "This document was created by analyzing recipe files during the build."
434    doc.creationInfo.licenseListVersion = get_license_list_version(license_data, d)
435    doc.creationInfo.creators.append("Tool: OpenEmbedded Core create-spdx.bbclass")
436    doc.creationInfo.creators.append("Organization: %s" % d.getVar("SPDX_ORG"))
437    doc.creationInfo.creators.append("Person: N/A ()")
438
439    recipe = oe.spdx.SPDXPackage()
440    recipe.name = d.getVar("PN")
441    recipe.versionInfo = d.getVar("PV")
442    recipe.SPDXID = oe.sbom.get_recipe_spdxid(d)
443    recipe.supplier = d.getVar("SPDX_SUPPLIER")
444    if bb.data.inherits_class("native", d) or bb.data.inherits_class("cross", d):
445        recipe.annotations.append(create_annotation(d, "isNative"))
446
447    homepage = d.getVar("HOMEPAGE")
448    if homepage:
449        recipe.homepage = homepage
450
451    license = d.getVar("LICENSE")
452    if license:
453        recipe.licenseDeclared = convert_license_to_spdx(license, license_data, doc, d)
454
455    summary = d.getVar("SUMMARY")
456    if summary:
457        recipe.summary = summary
458
459    description = d.getVar("DESCRIPTION")
460    if description:
461        recipe.description = description
462
463    if d.getVar("SPDX_CUSTOM_ANNOTATION_VARS"):
464        for var in d.getVar('SPDX_CUSTOM_ANNOTATION_VARS').split():
465            recipe.annotations.append(create_annotation(d, var + "=" + d.getVar(var)))
466
467    # Some CVEs may be patched during the build process without incrementing the version number,
468    # so querying for CVEs based on the CPE id can lead to false positives. To account for this,
469    # save the CVEs fixed by patches to source information field in the SPDX.
470    patched_cves = oe.cve_check.get_patched_cves(d)
471    patched_cves = list(patched_cves)
472    patched_cves = ' '.join(patched_cves)
473    if patched_cves:
474        recipe.sourceInfo = "CVEs fixed: " + patched_cves
475
476    cpe_ids = oe.cve_check.get_cpe_ids(d.getVar("CVE_PRODUCT"), d.getVar("CVE_VERSION"))
477    if cpe_ids:
478        for cpe_id in cpe_ids:
479            cpe = oe.spdx.SPDXExternalReference()
480            cpe.referenceCategory = "SECURITY"
481            cpe.referenceType = "http://spdx.org/rdf/references/cpe23Type"
482            cpe.referenceLocator = cpe_id
483            recipe.externalRefs.append(cpe)
484
485    doc.packages.append(recipe)
486    doc.add_relationship(doc, "DESCRIBES", recipe)
487
488    add_download_packages(d, doc, recipe)
489
490    if oe.spdx_common.process_sources(d) and include_sources:
491        recipe_archive = deploy_dir_spdx / "recipes" / (doc.name + ".tar.zst")
492        with optional_tarfile(recipe_archive, archive_sources) as archive:
493            oe.spdx_common.get_patched_src(d)
494
495            add_package_files(
496                d,
497                doc,
498                recipe,
499                spdx_workdir,
500                lambda file_counter: "SPDXRef-SourceFile-%s-%d" % (d.getVar("PN"), file_counter),
501                lambda filepath: ["SOURCE"],
502                ignore_dirs=[".git"],
503                ignore_top_level_dirs=["temp"],
504                archive=archive,
505            )
506
507            if archive is not None:
508                recipe.packageFileName = str(recipe_archive.name)
509
510    dep_recipes = collect_dep_recipes(d, doc, recipe)
511
512    doc_sha1 = oe.sbom.write_doc(d, doc, pkg_arch, "recipes", indent=get_json_indent(d))
513    dep_recipes.append(oe.sbom.DepRecipe(doc, doc_sha1, recipe))
514
515    recipe_ref = oe.spdx.SPDXExternalDocumentRef()
516    recipe_ref.externalDocumentId = "DocumentRef-recipe-" + recipe.name
517    recipe_ref.spdxDocument = doc.documentNamespace
518    recipe_ref.checksum.algorithm = "SHA1"
519    recipe_ref.checksum.checksumValue = doc_sha1
520
521    sources = collect_dep_sources(d, dep_recipes)
522    found_licenses = {license.name:recipe_ref.externalDocumentId + ":" + license.licenseId for license in doc.hasExtractedLicensingInfos}
523
524    if not recipe_spdx_is_native(d, recipe):
525        bb.build.exec_func("read_subpackage_metadata", d)
526
527        pkgdest = Path(d.getVar("PKGDEST"))
528        for package in d.getVar("PACKAGES").split():
529            if not oe.packagedata.packaged(package, d):
530                continue
531
532            package_doc = oe.spdx.SPDXDocument()
533            pkg_name = d.getVar("PKG:%s" % package) or package
534            package_doc.name = pkg_name
535            package_doc.documentNamespace = get_namespace(d, package_doc.name)
536            package_doc.creationInfo.created = creation_time
537            package_doc.creationInfo.comment = "This document was created by analyzing packages created during the build."
538            package_doc.creationInfo.licenseListVersion = get_license_list_version(license_data, d)
539            package_doc.creationInfo.creators.append("Tool: OpenEmbedded Core create-spdx.bbclass")
540            package_doc.creationInfo.creators.append("Organization: %s" % d.getVar("SPDX_ORG"))
541            package_doc.creationInfo.creators.append("Person: N/A ()")
542            package_doc.externalDocumentRefs.append(recipe_ref)
543
544            package_license = d.getVar("LICENSE:%s" % package) or d.getVar("LICENSE")
545
546            spdx_package = oe.spdx.SPDXPackage()
547
548            spdx_package.SPDXID = oe.sbom.get_package_spdxid(pkg_name)
549            spdx_package.name = pkg_name
550            spdx_package.versionInfo = d.getVar("PV")
551            spdx_package.licenseDeclared = convert_license_to_spdx(package_license, license_data, package_doc, d, found_licenses)
552            spdx_package.supplier = d.getVar("SPDX_SUPPLIER")
553
554            package_doc.packages.append(spdx_package)
555
556            package_doc.add_relationship(spdx_package, "GENERATED_FROM", "%s:%s" % (recipe_ref.externalDocumentId, recipe.SPDXID))
557            package_doc.add_relationship(package_doc, "DESCRIBES", spdx_package)
558
559            package_archive = deploy_dir_spdx / "packages" / (package_doc.name + ".tar.zst")
560            with optional_tarfile(package_archive, archive_packaged) as archive:
561                package_files = add_package_files(
562                    d,
563                    package_doc,
564                    spdx_package,
565                    pkgdest / package,
566                    lambda file_counter: oe.sbom.get_packaged_file_spdxid(pkg_name, file_counter),
567                    lambda filepath: ["BINARY"],
568                    ignore_top_level_dirs=['CONTROL', 'DEBIAN'],
569                    archive=archive,
570                )
571
572                if archive is not None:
573                    spdx_package.packageFileName = str(package_archive.name)
574
575            add_package_sources_from_debug(d, package_doc, spdx_package, package, package_files, sources)
576
577            oe.sbom.write_doc(d, package_doc, pkg_arch, "packages", indent=get_json_indent(d))
578}
579do_create_spdx[vardepsexclude] += "BB_NUMBER_THREADS"
580# NOTE: depending on do_unpack is a hack that is necessary to get it's dependencies for archive the source
581addtask do_create_spdx after do_package do_packagedata do_unpack do_collect_spdx_deps before do_populate_sdk do_build do_rm_work
582
583SSTATETASKS += "do_create_spdx"
584do_create_spdx[sstate-inputdirs] = "${SPDXDEPLOY}"
585do_create_spdx[sstate-outputdirs] = "${DEPLOY_DIR_SPDX}"
586
587python do_create_spdx_setscene () {
588    sstate_setscene(d)
589}
590addtask do_create_spdx_setscene
591
592do_create_spdx[dirs] = "${SPDXWORK}"
593do_create_spdx[cleandirs] = "${SPDXDEPLOY} ${SPDXWORK}"
594do_create_spdx[depends] += " \
595    ${PATCHDEPENDENCY} \
596    ${@create_spdx_source_deps(d)} \
597"
598
599python do_create_runtime_spdx() {
600    from datetime import datetime, timezone
601    import oe.sbom
602    import oe.spdx
603    import oe.spdx_common
604    import oe.packagedata
605    from pathlib import Path
606
607    deploy_dir_spdx = Path(d.getVar("DEPLOY_DIR_SPDX"))
608    spdx_deploy = Path(d.getVar("SPDXRUNTIMEDEPLOY"))
609    is_native = bb.data.inherits_class("native", d) or bb.data.inherits_class("cross", d)
610
611    creation_time = datetime.now(tz=timezone.utc).strftime("%Y-%m-%dT%H:%M:%SZ")
612
613    license_data = oe.spdx_common.load_spdx_license_data(d)
614
615    providers = oe.spdx_common.collect_package_providers(d)
616    pkg_arch = d.getVar("SSTATE_PKGARCH")
617    package_archs = d.getVar("SPDX_MULTILIB_SSTATE_ARCHS").split()
618    package_archs.reverse()
619
620    if not is_native:
621        bb.build.exec_func("read_subpackage_metadata", d)
622
623        dep_package_cache = {}
624
625        pkgdest = Path(d.getVar("PKGDEST"))
626        for package in d.getVar("PACKAGES").split():
627            localdata = bb.data.createCopy(d)
628            pkg_name = d.getVar("PKG:%s" % package) or package
629            localdata.setVar("PKG", pkg_name)
630            localdata.setVar('OVERRIDES', d.getVar("OVERRIDES", False) + ":" + package)
631
632            if not oe.packagedata.packaged(package, localdata):
633                continue
634
635            pkg_spdx_path = oe.sbom.doc_path(deploy_dir_spdx, pkg_name, pkg_arch, "packages")
636
637            package_doc, package_doc_sha1 = oe.sbom.read_doc(pkg_spdx_path)
638
639            for p in package_doc.packages:
640                if p.name == pkg_name:
641                    spdx_package = p
642                    break
643            else:
644                bb.fatal("Package '%s' not found in %s" % (pkg_name, pkg_spdx_path))
645
646            runtime_doc = oe.spdx.SPDXDocument()
647            runtime_doc.name = "runtime-" + pkg_name
648            runtime_doc.documentNamespace = get_namespace(localdata, runtime_doc.name)
649            runtime_doc.creationInfo.created = creation_time
650            runtime_doc.creationInfo.comment = "This document was created by analyzing package runtime dependencies."
651            runtime_doc.creationInfo.licenseListVersion = get_license_list_version(license_data, d)
652            runtime_doc.creationInfo.creators.append("Tool: OpenEmbedded Core create-spdx.bbclass")
653            runtime_doc.creationInfo.creators.append("Organization: %s" % d.getVar("SPDX_ORG"))
654            runtime_doc.creationInfo.creators.append("Person: N/A ()")
655
656            package_ref = oe.spdx.SPDXExternalDocumentRef()
657            package_ref.externalDocumentId = "DocumentRef-package-" + package
658            package_ref.spdxDocument = package_doc.documentNamespace
659            package_ref.checksum.algorithm = "SHA1"
660            package_ref.checksum.checksumValue = package_doc_sha1
661
662            runtime_doc.externalDocumentRefs.append(package_ref)
663
664            runtime_doc.add_relationship(
665                runtime_doc.SPDXID,
666                "AMENDS",
667                "%s:%s" % (package_ref.externalDocumentId, package_doc.SPDXID)
668            )
669
670            deps = bb.utils.explode_dep_versions2(localdata.getVar("RDEPENDS") or "")
671            seen_deps = set()
672            for dep, _ in deps.items():
673                if dep in seen_deps:
674                    continue
675
676                if dep not in providers:
677                    continue
678
679                (dep, dep_hashfn) = providers[dep]
680
681                if not oe.packagedata.packaged(dep, localdata):
682                    continue
683
684                dep_pkg_data = oe.packagedata.read_subpkgdata_dict(dep, d)
685                dep_pkg = dep_pkg_data["PKG"]
686
687                if dep in dep_package_cache:
688                    (dep_spdx_package, dep_package_ref) = dep_package_cache[dep]
689                else:
690                    dep_path = oe.sbom.doc_find_by_hashfn(deploy_dir_spdx, package_archs, dep_pkg, dep_hashfn)
691                    if not dep_path:
692                        bb.fatal("No SPDX file found for package %s, %s" % (dep_pkg, dep_hashfn))
693
694                    spdx_dep_doc, spdx_dep_sha1 = oe.sbom.read_doc(dep_path)
695
696                    for pkg in spdx_dep_doc.packages:
697                        if pkg.name == dep_pkg:
698                            dep_spdx_package = pkg
699                            break
700                    else:
701                        bb.fatal("Package '%s' not found in %s" % (dep_pkg, dep_path))
702
703                    dep_package_ref = oe.spdx.SPDXExternalDocumentRef()
704                    dep_package_ref.externalDocumentId = "DocumentRef-runtime-dependency-" + spdx_dep_doc.name
705                    dep_package_ref.spdxDocument = spdx_dep_doc.documentNamespace
706                    dep_package_ref.checksum.algorithm = "SHA1"
707                    dep_package_ref.checksum.checksumValue = spdx_dep_sha1
708
709                    dep_package_cache[dep] = (dep_spdx_package, dep_package_ref)
710
711                runtime_doc.externalDocumentRefs.append(dep_package_ref)
712
713                runtime_doc.add_relationship(
714                    "%s:%s" % (dep_package_ref.externalDocumentId, dep_spdx_package.SPDXID),
715                    "RUNTIME_DEPENDENCY_OF",
716                    "%s:%s" % (package_ref.externalDocumentId, spdx_package.SPDXID)
717                )
718                seen_deps.add(dep)
719
720            oe.sbom.write_doc(d, runtime_doc, pkg_arch, "runtime", spdx_deploy, indent=get_json_indent(d))
721}
722
723do_create_runtime_spdx[vardepsexclude] += "OVERRIDES SPDX_MULTILIB_SSTATE_ARCHS"
724
725addtask do_create_runtime_spdx after do_create_spdx before do_build do_rm_work
726SSTATETASKS += "do_create_runtime_spdx"
727do_create_runtime_spdx[sstate-inputdirs] = "${SPDXRUNTIMEDEPLOY}"
728do_create_runtime_spdx[sstate-outputdirs] = "${DEPLOY_DIR_SPDX}"
729
730python do_create_runtime_spdx_setscene () {
731    sstate_setscene(d)
732}
733addtask do_create_runtime_spdx_setscene
734
735do_create_runtime_spdx[dirs] = "${SPDXRUNTIMEDEPLOY}"
736do_create_runtime_spdx[cleandirs] = "${SPDXRUNTIMEDEPLOY}"
737do_create_runtime_spdx[rdeptask] = "do_create_spdx"
738
739do_rootfs[recrdeptask] += "do_create_spdx do_create_runtime_spdx"
740do_rootfs[cleandirs] += "${SPDXIMAGEWORK}"
741
742ROOTFS_POSTUNINSTALL_COMMAND =+ "image_combine_spdx"
743
744do_populate_sdk[recrdeptask] += "do_create_spdx do_create_runtime_spdx"
745do_populate_sdk[cleandirs] += "${SPDXSDKWORK}"
746POPULATE_SDK_POST_HOST_COMMAND:append:task-populate-sdk = " sdk_host_combine_spdx"
747POPULATE_SDK_POST_TARGET_COMMAND:append:task-populate-sdk = " sdk_target_combine_spdx"
748
749python image_combine_spdx() {
750    import os
751    import oe.sbom
752    from pathlib import Path
753    from oe.rootfs import image_list_installed_packages
754
755    image_name = d.getVar("IMAGE_NAME")
756    image_link_name = d.getVar("IMAGE_LINK_NAME")
757    imgdeploydir = Path(d.getVar("IMGDEPLOYDIR"))
758    img_spdxid = oe.sbom.get_image_spdxid(image_name)
759    packages = image_list_installed_packages(d)
760
761    combine_spdx(d, image_name, imgdeploydir, img_spdxid, packages, Path(d.getVar("SPDXIMAGEWORK")))
762
763    def make_image_link(target_path, suffix):
764        if image_link_name:
765            link = imgdeploydir / (image_link_name + suffix)
766            if link != target_path:
767                link.symlink_to(os.path.relpath(target_path, link.parent))
768
769    spdx_tar_path = imgdeploydir / (image_name + ".spdx.tar.zst")
770    make_image_link(spdx_tar_path, ".spdx.tar.zst")
771}
772
773python sdk_host_combine_spdx() {
774    sdk_combine_spdx(d, "host")
775}
776
777python sdk_target_combine_spdx() {
778    sdk_combine_spdx(d, "target")
779}
780
781def sdk_combine_spdx(d, sdk_type):
782    import oe.sbom
783    from pathlib import Path
784    from oe.sdk import sdk_list_installed_packages
785
786    sdk_name = d.getVar("TOOLCHAIN_OUTPUTNAME") + "-" + sdk_type
787    sdk_deploydir = Path(d.getVar("SDKDEPLOYDIR"))
788    sdk_spdxid = oe.sbom.get_sdk_spdxid(sdk_name)
789    sdk_packages = sdk_list_installed_packages(d, sdk_type == "target")
790    combine_spdx(d, sdk_name, sdk_deploydir, sdk_spdxid, sdk_packages, Path(d.getVar('SPDXSDKWORK')))
791
792def combine_spdx(d, rootfs_name, rootfs_deploydir, rootfs_spdxid, packages, spdx_workdir):
793    import os
794    import oe.spdx
795    import oe.sbom
796    import oe.spdx_common
797    import io
798    import json
799    from datetime import timezone, datetime
800    from pathlib import Path
801    import tarfile
802    import bb.compress.zstd
803
804    license_data = oe.spdx_common.load_spdx_license_data(d)
805
806    providers = oe.spdx_common.collect_package_providers(d)
807    package_archs = d.getVar("SPDX_MULTILIB_SSTATE_ARCHS").split()
808    package_archs.reverse()
809
810    creation_time = datetime.now(tz=timezone.utc).strftime("%Y-%m-%dT%H:%M:%SZ")
811    deploy_dir_spdx = Path(d.getVar("DEPLOY_DIR_SPDX"))
812    source_date_epoch = d.getVar("SOURCE_DATE_EPOCH")
813
814    doc = oe.spdx.SPDXDocument()
815    doc.name = rootfs_name
816    doc.documentNamespace = get_namespace(d, doc.name)
817    doc.creationInfo.created = creation_time
818    doc.creationInfo.comment = "This document was created by analyzing the source of the Yocto recipe during the build."
819    doc.creationInfo.licenseListVersion = get_license_list_version(license_data, d)
820    doc.creationInfo.creators.append("Tool: OpenEmbedded Core create-spdx.bbclass")
821    doc.creationInfo.creators.append("Organization: %s" % d.getVar("SPDX_ORG"))
822    doc.creationInfo.creators.append("Person: N/A ()")
823
824    image = oe.spdx.SPDXPackage()
825    image.name = d.getVar("PN")
826    image.versionInfo = d.getVar("PV")
827    image.SPDXID = rootfs_spdxid
828    image.supplier = d.getVar("SPDX_SUPPLIER")
829
830    doc.packages.append(image)
831
832    if packages:
833        for name in sorted(packages.keys()):
834            if name not in providers:
835                bb.fatal("Unable to find SPDX provider for '%s'" % name)
836
837            pkg_name, pkg_hashfn = providers[name]
838
839            pkg_spdx_path = oe.sbom.doc_find_by_hashfn(deploy_dir_spdx, package_archs, pkg_name, pkg_hashfn)
840            if not pkg_spdx_path:
841                bb.fatal("No SPDX file found for package %s, %s" % (pkg_name, pkg_hashfn))
842
843            pkg_doc, pkg_doc_sha1 = oe.sbom.read_doc(pkg_spdx_path)
844
845            for p in pkg_doc.packages:
846                if p.name == name:
847                    pkg_ref = oe.spdx.SPDXExternalDocumentRef()
848                    pkg_ref.externalDocumentId = "DocumentRef-%s" % pkg_doc.name
849                    pkg_ref.spdxDocument = pkg_doc.documentNamespace
850                    pkg_ref.checksum.algorithm = "SHA1"
851                    pkg_ref.checksum.checksumValue = pkg_doc_sha1
852
853                    doc.externalDocumentRefs.append(pkg_ref)
854                    doc.add_relationship(image, "CONTAINS", "%s:%s" % (pkg_ref.externalDocumentId, p.SPDXID))
855                    break
856            else:
857                bb.fatal("Unable to find package with name '%s' in SPDX file %s" % (name, pkg_spdx_path))
858
859            runtime_spdx_path = oe.sbom.doc_find_by_hashfn(deploy_dir_spdx, package_archs, "runtime-" + name, pkg_hashfn)
860            if not runtime_spdx_path:
861                bb.fatal("No runtime SPDX document found for %s, %s" % (name, pkg_hashfn))
862
863            runtime_doc, runtime_doc_sha1 = oe.sbom.read_doc(runtime_spdx_path)
864
865            runtime_ref = oe.spdx.SPDXExternalDocumentRef()
866            runtime_ref.externalDocumentId = "DocumentRef-%s" % runtime_doc.name
867            runtime_ref.spdxDocument = runtime_doc.documentNamespace
868            runtime_ref.checksum.algorithm = "SHA1"
869            runtime_ref.checksum.checksumValue = runtime_doc_sha1
870
871            # "OTHER" isn't ideal here, but I can't find a relationship that makes sense
872            doc.externalDocumentRefs.append(runtime_ref)
873            doc.add_relationship(
874                image,
875                "OTHER",
876                "%s:%s" % (runtime_ref.externalDocumentId, runtime_doc.SPDXID),
877                comment="Runtime dependencies for %s" % name
878            )
879    bb.utils.mkdirhier(spdx_workdir)
880    image_spdx_path = spdx_workdir / (rootfs_name + ".spdx.json")
881
882    with image_spdx_path.open("wb") as f:
883        doc.to_json(f, sort_keys=True, indent=get_json_indent(d))
884
885    num_threads = int(d.getVar("BB_NUMBER_THREADS"))
886
887    visited_docs = set()
888
889    index = {"documents": []}
890
891    spdx_tar_path = rootfs_deploydir / (rootfs_name + ".spdx.tar.zst")
892    with bb.compress.zstd.open(spdx_tar_path, "w", num_threads=num_threads) as f:
893        with tarfile.open(fileobj=f, mode="w|") as tar:
894            def collect_spdx_document(path):
895                nonlocal tar
896                nonlocal deploy_dir_spdx
897                nonlocal source_date_epoch
898                nonlocal index
899
900                if path in visited_docs:
901                    return
902
903                visited_docs.add(path)
904
905                with path.open("rb") as f:
906                    doc, sha1 = oe.sbom.read_doc(f)
907                    f.seek(0)
908
909                    if doc.documentNamespace in visited_docs:
910                        return
911
912                    bb.note("Adding SPDX document %s" % path)
913                    visited_docs.add(doc.documentNamespace)
914                    info = tar.gettarinfo(fileobj=f)
915
916                    info.name = doc.name + ".spdx.json"
917                    info.uid = 0
918                    info.gid = 0
919                    info.uname = "root"
920                    info.gname = "root"
921
922                    if source_date_epoch is not None and info.mtime > int(source_date_epoch):
923                        info.mtime = int(source_date_epoch)
924
925                    tar.addfile(info, f)
926
927                    index["documents"].append({
928                        "filename": info.name,
929                        "documentNamespace": doc.documentNamespace,
930                        "sha1": sha1,
931                    })
932
933                for ref in doc.externalDocumentRefs:
934                    ref_path = oe.sbom.doc_find_by_namespace(deploy_dir_spdx, package_archs, ref.spdxDocument)
935                    if not ref_path:
936                        bb.fatal("Cannot find any SPDX file for document %s" % ref.spdxDocument)
937                    collect_spdx_document(ref_path)
938
939            collect_spdx_document(image_spdx_path)
940
941            index["documents"].sort(key=lambda x: x["filename"])
942
943            index_str = io.BytesIO(json.dumps(
944                index,
945                sort_keys=True,
946                indent=get_json_indent(d),
947            ).encode("utf-8"))
948
949            info = tarfile.TarInfo()
950            info.name = "index.json"
951            info.size = len(index_str.getvalue())
952            info.uid = 0
953            info.gid = 0
954            info.uname = "root"
955            info.gname = "root"
956
957            tar.addfile(info, fileobj=index_str)
958
959combine_spdx[vardepsexclude] += "BB_NUMBER_THREADS SPDX_MULTILIB_SSTATE_ARCHS"
960