Skip to contentSkip to conversation

improve_kernel_cve_report: use numeric versions instead of cpeApplicability

PATCH In review [email protected] 3 participants 14 messages Apr 17 – Apr 30

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 improve_kernel_cve_report: use numeric versions instead of cpeApplicability

From: Daniel Turull <[email protected]>

git shas or versions should be use instead of cpeApplicability. Reuse the same logic as generate-cve-exclusions, so outputs are consistent.

cpeApplicability does not provide accurate version information and for some
CVEs the information is not the same. This came from a discussion that
we had with Greg Kroah-Hartma, member of the Linux security team.

---
 scripts/contrib/improve_kernel_cve_report.py | 247 ++++++++-----------
 1 file changed, 104 insertions(+), 143 deletions(-)
scripts/contrib/improve_kernel_cve_report.py +104 -143
index b386c9383a..a5458f8dd3 100755
@@ -82,16 +82,59 @@ def get_kernel_cves(datadir, compiled_files, version):
                 "description": f"Rejected by CNA"
             }
             continue
-        if any(elem in cve_file for elem in ["review", "reverved", "testing"]):
+        if any(elem in cve_file for elem in ["review", "reverved", "testing", "tmp"]):
             continue
 
-        is_vulnerable, first_affected, last_affected, better_match_first, better_match_last, affected_versions = get_cpe_applicability(cve_info, version)
+        first_affected, fixed, backport_ver = get_fixed_versions(cve_info, base_version)
 
-        logging.debug("%s: %s (%s - %s) (%s - %s)", cve_id, is_vulnerable, better_match_first, better_match_last, first_affected, last_affected)
+        logging.debug("%s: first_affected=%s fixed=%s backport=%s", cve_id, first_affected, fixed, backport_ver)
 
-        if is_vulnerable is None:
-            logging.warning("%s doesn't have good metadata", cve_id)
-        if is_vulnerable:
+        if not fixed:
+            logging.warning("%s has no known resolution", cve_id)
+            cves[cve_id] = {
+                "id": cve_id,
+                "status": "Unpatched",
+                "detail": "known-affected",
+                "summary": description,
+                "description": "No known resolution"
+            }
+            vulnerable += 1
+            continue
+        elif first_affected and version < first_affected:
+            logging.debug('%s - fixed-version: only affects %s onwards',
+                          cve_id, first_affected)
+            cves[cve_id] = {
+                "id": cve_id,
+                "status": "Patched",
+                "detail": "fixed-version",
+                "summary": description,
+                "description": f"only affects {first_affected} onwards"
+            }
+            not_vulnerable += 1
+        elif fixed <= version:
+            logging.debug("%s - fixed-version: Fixed from version %s",
+                          cve_id, fixed)
+            cves[cve_id] = {
+                "id": cve_id,
+                "status": "Patched",
+                "detail": "fixed-version",
+                "summary": description,
+                "description": f"Fixed from version {fixed}"
+            }
+            not_vulnerable += 1
+        elif backport_ver and backport_ver <= version:
+            logging.debug("%s - cpe-stable-backport: Backported in %s",
+                          cve_id, backport_ver)
+            cves[cve_id] = {
+                "id": cve_id,
+                "status": "Patched",
+                "detail": "cpe-stable-backport",
+                "summary": description,
+                "description": f"Backported in {backport_ver}"
+            }
+            not_vulnerable += 1
+        else:
+            # Vulnerable - may need backporting
             is_affected = True
             affected_files = []
             if check_config:
@@ -108,87 +151,21 @@ def get_kernel_cves(datadir, compiled_files, version):
                     "summary": description,
                     "description": f"Source code not compiled by config. {sorted(affected_files)}"
                 }
-                not_applicable_config +=1
-            # Check if we have backport
+                not_applicable_config += 1
             else:
-                if not better_match_last:
-                    fixed_in = last_affected
-                else:
-                    fixed_in = better_match_last
+                fixed_in = backport_ver if backport_ver else fixed
                 logging.debug("%s needs backporting (fixed from %s)", cve_id, fixed_in)
-                cves[cve_id] = {
-                        "id": cve_id,
-                        "status": "Unpatched",
-                        "detail": "version-in-range",
-                        "summary": description,
-                        "description": f"Needs backporting (fixed from {fixed_in})"
-                }
-                vulnerable += 1
-                if (better_match_last and
-                    Version(f"{better_match_last.major}.{better_match_last.minor}") == base_version):
-                    fixed_as_later_backport += 1
-        # Not vulnerable
-        else:
-            if not first_affected:
-                logging.debug("%s - not known affected %s",
-                              cve_id,
-                              better_match_last)
                 cves[cve_id] = {
                     "id": cve_id,
-                    "status": "Patched",
-                    "detail": "version-not-in-range",
+                    "status": "Unpatched",
+                    "detail": "version-in-range",
                     "summary": description,
-                    "description": "No CPE match"
+                    "description": f"Needs backporting (fixed from {fixed_in})"
                 }
-                not_vulnerable += 1
-                continue
-            backport_base = Version(f"{better_match_last.major}.{better_match_last.minor}")
-            if version < first_affected:
-                logging.debug('%s - fixed-version: only affects %s onwards',
-                              cve_id,
-                              first_affected)
-                cves[cve_id] = {
-                    "id": cve_id,
-                    "status": "Patched",
-                    "detail": "fixed-version",
-                    "summary": description,
-                    "description": f"only affects {first_affected} onwards"
-                }
-                not_vulnerable += 1
-            elif last_affected <= version:
-                logging.debug("%s - fixed-version: Fixed from version %s",
-                              cve_id,
-                              last_affected)
-                cves[cve_id] = {
-                    "id": cve_id,
-                    "status": "Patched",
-                    "detail": "fixed-version",
-                    "summary": description,
-                    "description": f"Fixed from version {last_affected}"
-                }
-                not_vulnerable += 1
-            elif backport_base == base_version:
-                logging.debug("%s - cpe-stable-backport: Backported in %s",
-                              cve_id,
-                              better_match_last)
-                cves[cve_id] = {
-                    "id": cve_id,
-                    "status": "Patched",
-                    "detail": "cpe-stable-backport",
-                    "summary": description,
-                    "description": f"Backported in {better_match_last}"
-                }
-                not_vulnerable += 1
-            else:
-                logging.debug("%s - version not affected %s", cve_id, str(affected_versions))
-                cves[cve_id] = {
-                    "id": cve_id,
-                    "status": "Patched",
-                    "detail": "version-not-in-range",
-                    "summary": description,
-                    "description": f"Range {affected_versions}"
-                }
-                not_vulnerable += 1
+                vulnerable += 1
+                if (backport_ver and
+                    Version(f"{backport_ver.major}.{backport_ver.minor}") == base_version):
+                    fixed_as_later_backport += 1
 
     logging.info("Total CVEs ignored due to not applicable config: %d", not_applicable_config)
     logging.info("Total CVEs not vulnerable due version-not-in-range: %d", not_vulnerable)
@@ -276,70 +253,54 @@ def check_kernel_compiled_files(compiled_files, cve_info):
                 is_affected = True
     return is_affected, files_affected
 
-def get_cpe_applicability(cve_info, v):
+def get_fixed_versions(cve_info, base_version):
     '''
-    Check if version is affected and return affected versions
+    Get fixed versionss
     '''
-    base_branch = Version(f"{v.major}.{v.minor}")
-    affected = []
-    if not 'cpeApplicability' in cve_info["containers"]["cna"]:
-        return None, None, None, None, None, None
-
-    for nodes in cve_info["containers"]["cna"]["cpeApplicability"]:
-        for node in nodes.values():
-            vulnerable = False
-            matched_branch = False
-            first_affected = Version("5000")
-            last_affected = Version("0")
-            better_match_first = Version("0")
-            better_match_last = Version("5000")
-
-            if len(node[0]['cpeMatch']) == 0:
-                first_affected = None
-                last_affected = None
-                better_match_first = None
-                better_match_last = None
-
-            for cpe_match in node[0]['cpeMatch']:
-                version_start_including = Version("0")
-                version_end_excluding = Version("0")
-                if 'versionStartIncluding' in cpe_match:
-                    version_start_including = Version(cpe_match['versionStartIncluding'])
-                else:
-                    version_start_including = Version("0")
-                # if versionEndExcluding is missing we are in a branch, which is not fixed.
-                if "versionEndExcluding" in cpe_match:
-                    version_end_excluding = Version(cpe_match["versionEndExcluding"])
-                else:
-                    # if versionEndExcluding is missing we are in a branch, which is not fixed.
-                    version_end_excluding = Version(
-                        f"{version_start_including.major}.{version_start_including.minor}.5000"
-                    )
-                affected.append(f" {version_start_including}-{version_end_excluding}")
-                # Detect if versionEnd is in fixed in base branch. It has precedence over the rest
-                branch_end = Version(f"{version_end_excluding.major}.{version_end_excluding.minor}")
-                if branch_end == base_branch:
-                    if version_start_including <= v < version_end_excluding:
-                        vulnerable = cpe_match['vulnerable']
-                    # If we don't match in our branch, we are not vulnerable,
-                    # since we have a backport
-                    matched_branch = True
-                    better_match_first = version_start_including
-                    better_match_last = version_end_excluding
-                if version_start_including <= v < version_end_excluding and not matched_branch:
-                    if version_end_excluding < better_match_last:
-                        better_match_first = max(version_start_including, better_match_first)
-                        better_match_last = min(better_match_last, version_end_excluding)
-                        vulnerable = cpe_match['vulnerable']
-                        matched_branch = True
-
-                first_affected = min(version_start_including, first_affected)
-                last_affected = max(version_end_excluding, last_affected)
-            # Not a better match, we use the first and last affected instead of the fake .5000
-            if vulnerable and better_match_last == Version(f"{base_branch}.5000"):
-                better_match_last = last_affected
-                better_match_first = first_affected
-    return vulnerable, first_affected, last_affected, better_match_first, better_match_last, affected
+    first_affected = None
+    fixed = None
+    fixed_backport = None
+    next_version = Version(str(base_version) + ".5000")
+    for affected in cve_info["containers"]["cna"]["affected"]:
+        # In case the CVE info is not complete, it might not have default status and therefore
+        # we don't know the status of this CVE.
+        if not "defaultStatus" in affected:
+            return first_affected, fixed, fixed_backport
+        if affected["defaultStatus"] == "affected":
+            for version in affected["versions"]:
+                v = Version(version["version"])
+                if v == Version('0'):
+                    #Skiping non-affected
+                    continue
+                if version["status"] == "unaffected" and first_affected and v < first_affected:
+                    first_affected = Version(f"{v.major}.{v.minor}")
+                if version["status"] == "affected" and not first_affected:
+                    first_affected = v
+                elif (version["status"] == "unaffected" and
+                    version['versionType'] == "original_commit_for_fix"):
+                    fixed = v
+                elif base_version < v and v < next_version:
+                    fixed_backport = v
+        elif affected["defaultStatus"] == "unaffected":
+            # Only specific versions are affected. We care only about our base version
+            if "versions" not in affected:
+                continue
+            for version in affected["versions"]:
+                if "versionType" not in version:
+                    continue
+                if version["versionType"] == "git":
+                    continue
+                v = Version(version["version"])
+                # in case it is not in our base version
+                less_than = Version(version["lessThan"])
+
+                if not first_affected:
+                    first_affected = v
+                fixed = less_than
+                if base_version < v and v < next_version:
+                    fixed_backport = less_than
+
+    return first_affected, fixed, fixed_backport
 
 def copy_data(old, new):
     '''Update dictionary with new entries, while keeping the old ones'''
Signature · 1 line
2.43.0
#1 lore raw

Hello Daniel,

Quoted 8 lines · On Friday, April 17, 2026 at 3:24 PM, [email protected] wrote:

From: Daniel Turull <[email protected]>

git shas or versions should be use instead of cpeApplicability. Reuse the same logic as generate-cve-exclusions, so outputs are consistent.

cpeApplicability does not provide accurate version information and for some
CVEs the information is not the same. This came from a discussion that
we had with Greg Kroah-Hartma, member of the Linux security team.

Indeed "cpeApplicability" does not provide the same kind of information that the "versions" node.
In sbom-cve-check (the latest version in main branch) we are using both sources of information.
But you are saying that "cpeApplicability" does not provide accurate version information. Could you elaborate and give various examples? I never saw something invalid in "cpeApplicability".

Quoted 4 lines
---
 scripts/contrib/improve_kernel_cve_report.py | 247 ++++++++-----------
 1 file changed, 104 insertions(+), 143 deletions(-)
Signature · 3 lines
Benjamin Robin, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com
#2 lore raw

Hi,
We had Greg visiting us and I asked him what is better to use and he said git or versions, not cvepAplicability that has issues defining trees.

I have done some comparison with 6.6.100 and 6.18.22

=== 6.6.100 ===
Old: 10900 New: 10898 Match: 10418 Only old: 2 Only new: 0 Diff: 480

    327  Unpatched/version-in-range -> Unpatched/version-in-range
    106  Patched/fixed-version -> Patched/fixed-version
     16  Patched/cpe-stable-backport -> Unpatched/version-in-range
     15  Unpatched/version-in-range -> Patched/fixed-version
      5  Patched/fixed-version -> Unpatched/known-affected
      5  Patched/fixed-version -> Unpatched/version-in-range
      2  Patched/cpe-stable-backport -> Patched/fixed-version
      1  Patched/version-not-in-range -> Unpatched/version-in-range
      1  Patched/cpe-stable-backport -> Unpatched/known-affected
      1  Patched/version-not-in-range -> Patched/fixed-version
      1  Unpatched/version-in-range -> Unpatched/known-affected
  Only in old:
    CVE-2024-0000: Patched/version-not-in-range
    CVE-2024-0053: Patched/version-not-in-range

=== 6.18.22 ===
Old: 10900 New: 10898 Match: 10877 Only old: 2 Only new: 0 Diff: 21

      7  Unpatched/version-in-range -> Unpatched/version-in-range
      6  Patched/fixed-version -> Patched/fixed-version
      6  Patched/fixed-version -> Unpatched/known-affected
      1  Unpatched/version-in-range -> Unpatched/known-affected
      1  Unpatched/version-in-range -> Patched/cpe-stable-backport
  Only in old:
    CVE-2024-0000: Patched/version-not-in-range
    CVE-2024-0053: Patched/version-not-in-range

old vs new outputs for kernel 6.18.22

Old total: 10900
New total: 10898
Matching: 10877
Only in old: 2
Only in new: 0
Different: 21

Difference categories:

      7  Unpatched/version-in-range -> Unpatched/version-in-range
      6  Patched/fixed-version -> Patched/fixed-version
      6  Patched/fixed-version -> Unpatched/known-affected
      1  Unpatched/version-in-range -> Unpatched/known-affected
      1  Unpatched/version-in-range -> Patched/cpe-stable-backport

Only in old (2):
CVE-2024-0000: Patched/version-not-in-range: No CPE match
CVE-2024-0053: Patched/version-not-in-range: No CPE match

Different (all 21):
CVE-2021-47295:
old: Patched/fixed-version: Fixed from version 6.2.5
new: Patched/fixed-version: Fixed from version 5.14
CVE-2021-47342:
old: Patched/fixed-version: Fixed from version 5.12.5000
new: Patched/fixed-version: Fixed from version 5.10.77
CVE-2022-50396:
old: Patched/fixed-version: Fixed from version 6.2.5
new: Patched/fixed-version: Fixed from version 6.2
CVE-2023-53012:
old: Patched/fixed-version: Fixed from version 6.1.5000
new: Unpatched/known-affected: No known resolution
CVE-2023-53187:
old: Patched/fixed-version: Fixed from version 5.15.5000
new: Unpatched/known-affected: No known resolution
CVE-2024-49854:
old: Patched/fixed-version: Fixed from version 5.10.5000
new: Unpatched/known-affected: No known resolution
CVE-2025-38656:
old: Patched/fixed-version: Fixed from version 6.12.5000
new: Unpatched/known-affected: No known resolution
CVE-2025-68195:
old: Patched/fixed-version: Fixed from version 6.12.5000
new: Unpatched/known-affected: No known resolution
CVE-2025-68357:
old: Patched/fixed-version: Fixed from version 6.17.5000
new: Patched/fixed-version: Fixed from version 6.12.64
CVE-2025-71145:
old: Patched/fixed-version: Fixed from version 5.10.5000
new: Unpatched/known-affected: No known resolution
CVE-2026-23288:
old: Patched/fixed-version: only affects 6.19.4 onwards
new: Patched/fixed-version: only affects 6.19 onwards
CVE-2026-23327:
old: Unpatched/version-in-range: Needs backporting (fixed from 6.19.7)
new: Unpatched/version-in-range: Needs backporting (fixed from 7.0rc2)
CVE-2026-23328:
old: Unpatched/version-in-range: Needs backporting (fixed from 6.19.7)
new: Unpatched/version-in-range: Needs backporting (fixed from 7.0rc3)
CVE-2026-23333:
old: Unpatched/version-in-range: Needs backporting (fixed from 6.19.4)
new: Unpatched/known-affected: No known resolution
CVE-2026-23341:
old: Patched/fixed-version: only affects 6.19.4 onwards
new: Patched/fixed-version: only affects 6.19 onwards
CVE-2026-23355:
old: Unpatched/version-in-range: Needs backporting (fixed from 7.0rc3)
new: Patched/cpe-stable-backport: Backported in 6.18.18
CVE-2026-23371:
old: Unpatched/version-in-range: Needs backporting (fixed from 6.19.7)
new: Unpatched/version-in-range: Needs backporting (fixed from 7.0rc3)
CVE-2026-23374:
old: Unpatched/version-in-range: Needs backporting (fixed from 6.19.7)
new: Unpatched/version-in-range: Needs backporting (fixed from 7.0rc3)
CVE-2026-23377:
old: Unpatched/version-in-range: Needs backporting (fixed from 6.19.7)
new: Unpatched/version-in-range: Needs backporting (fixed from 7.0rc3)
CVE-2026-23389:
old: Unpatched/version-in-range: Needs backporting (fixed from 6.19.7)
new: Unpatched/version-in-range: Needs backporting (fixed from 7.0rc3)
CVE-2026-23394:
old: Unpatched/version-in-range: Needs backporting (fixed from 6.19.10)
new: Unpatched/version-in-range: Needs backporting (fixed from 7.0rc5)

Best regards,
Daniel

Quoted 51 lines

-----Original Message-----
From: Benjamin Robin <[email protected]>
Sent: Friday, 17 April 2026 15:32
To: [email protected]; Daniel Turull <[email protected]>
Subject: Re: [PATCH] improve_kernel_cve_report: use numeric versions instead of cpeApplicability

[You don't often get email from [email protected]. Learn why this is important at https://aka.ms/LearnAboutSenderIdentification ]

Hello Daniel,

Quoted 8 lines · On Friday, April 17, 2026 at 3:24 PM, [email protected] wrote:

From: Daniel Turull <[email protected]>

git shas or versions should be use instead of cpeApplicability. Reuse the same logic as generate-cve-exclusions, so outputs are consistent.

cpeApplicability does not provide accurate version information and for some CVEs the information is not the same. This came from a discussion that we had with Greg Kroah-Hartma, member of the Linux

security team.

Indeed "cpeApplicability" does not provide the same kind of information that the "versions" node.
In sbom-cve-check (the latest version in main branch) we are using both sources of information.
But you are saying that "cpeApplicability" does not provide accurate version information. Could you elaborate and give various examples? I never saw something invalid in "cpeApplicability".

Quoted 5 lines
---
 scripts/contrib/improve_kernel_cve_report.py | 247
++++++++-----------
 1 file changed, 104 insertions(+), 143 deletions(-)

--
Benjamin Robin, Bootlin
Embedded Linux and Kernel engineering
https://eur02.safelinks.protection.outlook.com/?url=https%3A%2F%2Fbootlin
.com%2F&data=05%7C02%7Cdaniel.turull%40ericsson.com%7C3ca7f33551f2 46de878808de9c85c084%7C92e84cebfbfd47abbe52080c6b87953f%7C0%7C0 %7C639120295445918854%7CUnknown%7CTWFpbGZsb3d8eyJFbXB0eU1hcG kiOnRydWUsIlYiOiIwLjAuMDAwMCIsIlAiOiJXaW4zMiIsIkFOIjoiTWFpbCIsIldUIj oyfQ%3D%3D%7C0%7C%7C%7C&sdata=F%2F5gD%2FMGt0tXSMT1fZ9yh6%2 BJJN3tn%2BqOCJ%2F2TchpDO4%3D&reserved=0

#3 lore raw
Quoted 2 lines · Daniel Turull

Hi,
We had Greg visiting us and I asked him what is better to use and he said git or versions, not cvepAplicability that has issues defining trees.

You reply is technically not responding to my question :) Could you provide at least one example with an entry that is not correct in the cpeApplicability node?

The script had previously various issue (or at least it looked like it). I preferred to use a completely different algorithm (and using all sources of information)

But since I am also using NVD entries this degrade a bit the quality of the generated assessment message

Quoted 151 lines

I have done some comparison with 6.6.100 and 6.18.22

=== 6.6.100 ===
Old: 10900 New: 10898 Match: 10418 Only old: 2 Only new: 0 Diff: 480

    327  Unpatched/version-in-range -> Unpatched/version-in-range
    106  Patched/fixed-version -> Patched/fixed-version
     16  Patched/cpe-stable-backport -> Unpatched/version-in-range
     15  Unpatched/version-in-range -> Patched/fixed-version
      5  Patched/fixed-version -> Unpatched/known-affected
      5  Patched/fixed-version -> Unpatched/version-in-range
      2  Patched/cpe-stable-backport -> Patched/fixed-version
      1  Patched/version-not-in-range -> Unpatched/version-in-range
      1  Patched/cpe-stable-backport -> Unpatched/known-affected
      1  Patched/version-not-in-range -> Patched/fixed-version
      1  Unpatched/version-in-range -> Unpatched/known-affected
  Only in old:
    CVE-2024-0000: Patched/version-not-in-range
    CVE-2024-0053: Patched/version-not-in-range

=== 6.18.22 ===
Old: 10900 New: 10898 Match: 10877 Only old: 2 Only new: 0 Diff: 21

      7  Unpatched/version-in-range -> Unpatched/version-in-range
      6  Patched/fixed-version -> Patched/fixed-version
      6  Patched/fixed-version -> Unpatched/known-affected
      1  Unpatched/version-in-range -> Unpatched/known-affected
      1  Unpatched/version-in-range -> Patched/cpe-stable-backport
  Only in old:
    CVE-2024-0000: Patched/version-not-in-range
    CVE-2024-0053: Patched/version-not-in-range

old vs new outputs for kernel 6.18.22

Old total: 10900
New total: 10898
Matching: 10877
Only in old: 2
Only in new: 0
Different: 21

Difference categories:

      7  Unpatched/version-in-range -> Unpatched/version-in-range
      6  Patched/fixed-version -> Patched/fixed-version
      6  Patched/fixed-version -> Unpatched/known-affected
      1  Unpatched/version-in-range -> Unpatched/known-affected
      1  Unpatched/version-in-range -> Patched/cpe-stable-backport

Only in old (2):
CVE-2024-0000: Patched/version-not-in-range: No CPE match
CVE-2024-0053: Patched/version-not-in-range: No CPE match

Different (all 21):
CVE-2021-47295:
old: Patched/fixed-version: Fixed from version 6.2.5
new: Patched/fixed-version: Fixed from version 5.14
CVE-2021-47342:
old: Patched/fixed-version: Fixed from version 5.12.5000
new: Patched/fixed-version: Fixed from version 5.10.77
CVE-2022-50396:
old: Patched/fixed-version: Fixed from version 6.2.5
new: Patched/fixed-version: Fixed from version 6.2
CVE-2023-53012:
old: Patched/fixed-version: Fixed from version 6.1.5000
new: Unpatched/known-affected: No known resolution
CVE-2023-53187:
old: Patched/fixed-version: Fixed from version 5.15.5000
new: Unpatched/known-affected: No known resolution
CVE-2024-49854:
old: Patched/fixed-version: Fixed from version 5.10.5000
new: Unpatched/known-affected: No known resolution
CVE-2025-38656:
old: Patched/fixed-version: Fixed from version 6.12.5000
new: Unpatched/known-affected: No known resolution
CVE-2025-68195:
old: Patched/fixed-version: Fixed from version 6.12.5000
new: Unpatched/known-affected: No known resolution
CVE-2025-68357:
old: Patched/fixed-version: Fixed from version 6.17.5000
new: Patched/fixed-version: Fixed from version 6.12.64
CVE-2025-71145:
old: Patched/fixed-version: Fixed from version 5.10.5000
new: Unpatched/known-affected: No known resolution
CVE-2026-23288:
old: Patched/fixed-version: only affects 6.19.4 onwards
new: Patched/fixed-version: only affects 6.19 onwards
CVE-2026-23327:
old: Unpatched/version-in-range: Needs backporting (fixed from 6.19.7)
new: Unpatched/version-in-range: Needs backporting (fixed from 7.0rc2)
CVE-2026-23328:
old: Unpatched/version-in-range: Needs backporting (fixed from 6.19.7)
new: Unpatched/version-in-range: Needs backporting (fixed from 7.0rc3)
CVE-2026-23333:
old: Unpatched/version-in-range: Needs backporting (fixed from 6.19.4)
new: Unpatched/known-affected: No known resolution
CVE-2026-23341:
old: Patched/fixed-version: only affects 6.19.4 onwards
new: Patched/fixed-version: only affects 6.19 onwards
CVE-2026-23355:
old: Unpatched/version-in-range: Needs backporting (fixed from 7.0rc3)
new: Patched/cpe-stable-backport: Backported in 6.18.18
CVE-2026-23371:
old: Unpatched/version-in-range: Needs backporting (fixed from 6.19.7)
new: Unpatched/version-in-range: Needs backporting (fixed from 7.0rc3)
CVE-2026-23374:
old: Unpatched/version-in-range: Needs backporting (fixed from 6.19.7)
new: Unpatched/version-in-range: Needs backporting (fixed from 7.0rc3)
CVE-2026-23377:
old: Unpatched/version-in-range: Needs backporting (fixed from 6.19.7)
new: Unpatched/version-in-range: Needs backporting (fixed from 7.0rc3)
CVE-2026-23389:
old: Unpatched/version-in-range: Needs backporting (fixed from 6.19.7)
new: Unpatched/version-in-range: Needs backporting (fixed from 7.0rc3)
CVE-2026-23394:
old: Unpatched/version-in-range: Needs backporting (fixed from 6.19.10)
new: Unpatched/version-in-range: Needs backporting (fixed from 7.0rc5)

Best regards,
Daniel

Quoted 32 lines

-----Original Message-----
From: Benjamin Robin <[email protected]>
Sent: Friday, 17 April 2026 15:32
To: [email protected]; Daniel Turull <[email protected]>
Subject: Re: [PATCH] improve_kernel_cve_report: use numeric versions instead of cpeApplicability

[You don't often get email from [email protected]. Learn why this is important at https://aka.ms/LearnAboutSenderIdentification ]

Hello Daniel,

Quoted 8 lines · On Friday, April 17, 2026 at 3:24 PM, [email protected] wrote:

From: Daniel Turull <[email protected]>

git shas or versions should be use instead of cpeApplicability. Reuse the same logic as generate-cve-exclusions, so outputs are consistent.

cpeApplicability does not provide accurate version information and for some CVEs the information is not the same. This came from a discussion that we had with Greg Kroah-Hartma, member of the Linux

security team.

Indeed "cpeApplicability" does not provide the same kind of information that the "versions" node.
In sbom-cve-check (the latest version in main branch) we are using both sources of information.
But you are saying that "cpeApplicability" does not provide accurate version information. Could you elaborate and give various examples? I never saw something invalid in "cpeApplicability".

Signature · 3 lines
Benjamin Robin, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com
#4 lore raw

True,
I sent CVEs that had different responses. I used an old checkout and rerun with the old and the new script.

I'm starting to find some issues with the data that we should clarify before merging this patch. Let's pause it and have it correct.

For example,
https://git.kernel.org/pub/scm/linux/security/vulns.git/tree/cve/published/2025/CVE-2025-40067.json

"defaultStatus": "affected",

{
                     "version": "6.6.112",
                     "lessThanOrEqual": "6.6.*",
                     "status": "unaffected",
                     "versionType": "semver"
                  },

"negate": false,

{
                           "vulnerable": true,
                           "criteria": "cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*",
                           "versionStartIncluding": "6.6.102",
                           "versionEndExcluding": "6.6.112"
                        },

In this case, the information in one of the entries is not correct. For example, in 6.6.100 is vulnerable in the version not in the cpeApplicability, but git versions and cpeApplicability match if we do a git describe on them.

I'll send an email to ask for clarification to the kernel security team and try to see other similar cases. I must say that this is a minority of all CVEs.

Also I need to look if this could be integrated in the sbom-cve-check, so we have it only one place. I want to be able to run the script as well with older releases or just telling the kernel to use without SBOM.

Best regards,
Daniel

Quoted 191 lines

-----Original Message-----
From: Benjamin Robin <[email protected]>
Sent: Friday, 17 April 2026 15:55
To: [email protected]; Daniel Turull <[email protected]>
Subject: Re: [PATCH] improve_kernel_cve_report: use numeric versions instead of cpeApplicability

Quoted 2 lines · Daniel Turull

Hi,
We had Greg visiting us and I asked him what is better to use and he said git

or versions, not cvepAplicability that has issues defining trees.

You reply is technically not responding to my question :) Could you provide at least one example with an entry that is not correct in the cpeApplicability node?

The script had previously various issue (or at least it looked like it). I preferred to use a completely different algorithm (and using all sources of information)

But since I am also using NVD entries this degrade a bit the quality of the generated assessment message

Quoted 117 lines

I have done some comparison with 6.6.100 and 6.18.22

=== 6.6.100 ===
Old: 10900 New: 10898 Match: 10418 Only old: 2 Only new: 0 Diff: 480

    327  Unpatched/version-in-range -> Unpatched/version-in-range
    106  Patched/fixed-version -> Patched/fixed-version
     16  Patched/cpe-stable-backport -> Unpatched/version-in-range
     15  Unpatched/version-in-range -> Patched/fixed-version
      5  Patched/fixed-version -> Unpatched/known-affected
      5  Patched/fixed-version -> Unpatched/version-in-range
      2  Patched/cpe-stable-backport -> Patched/fixed-version
      1  Patched/version-not-in-range -> Unpatched/version-in-range
      1  Patched/cpe-stable-backport -> Unpatched/known-affected
      1  Patched/version-not-in-range -> Patched/fixed-version
      1  Unpatched/version-in-range -> Unpatched/known-affected
  Only in old:
    CVE-2024-0000: Patched/version-not-in-range
    CVE-2024-0053: Patched/version-not-in-range

=== 6.18.22 ===
Old: 10900 New: 10898 Match: 10877 Only old: 2 Only new: 0 Diff: 21

      7  Unpatched/version-in-range -> Unpatched/version-in-range
      6  Patched/fixed-version -> Patched/fixed-version
      6  Patched/fixed-version -> Unpatched/known-affected
      1  Unpatched/version-in-range -> Unpatched/known-affected
      1  Unpatched/version-in-range -> Patched/cpe-stable-backport
  Only in old:
    CVE-2024-0000: Patched/version-not-in-range
    CVE-2024-0053: Patched/version-not-in-range

old vs new outputs for kernel 6.18.22

Old total: 10900
New total: 10898
Matching: 10877
Only in old: 2
Only in new: 0
Different: 21

Difference categories:

      7  Unpatched/version-in-range -> Unpatched/version-in-range
      6  Patched/fixed-version -> Patched/fixed-version
      6  Patched/fixed-version -> Unpatched/known-affected
      1  Unpatched/version-in-range -> Unpatched/known-affected
      1  Unpatched/version-in-range -> Patched/cpe-stable-backport

Only in old (2):
CVE-2024-0000: Patched/version-not-in-range: No CPE match
CVE-2024-0053: Patched/version-not-in-range: No CPE match

Different (all 21):
CVE-2021-47295:
old: Patched/fixed-version: Fixed from version 6.2.5
new: Patched/fixed-version: Fixed from version 5.14
CVE-2021-47342:
old: Patched/fixed-version: Fixed from version 5.12.5000
new: Patched/fixed-version: Fixed from version 5.10.77
CVE-2022-50396:
old: Patched/fixed-version: Fixed from version 6.2.5
new: Patched/fixed-version: Fixed from version 6.2
CVE-2023-53012:
old: Patched/fixed-version: Fixed from version 6.1.5000
new: Unpatched/known-affected: No known resolution
CVE-2023-53187:
old: Patched/fixed-version: Fixed from version 5.15.5000
new: Unpatched/known-affected: No known resolution
CVE-2024-49854:
old: Patched/fixed-version: Fixed from version 5.10.5000
new: Unpatched/known-affected: No known resolution
CVE-2025-38656:
old: Patched/fixed-version: Fixed from version 6.12.5000
new: Unpatched/known-affected: No known resolution
CVE-2025-68195:
old: Patched/fixed-version: Fixed from version 6.12.5000
new: Unpatched/known-affected: No known resolution
CVE-2025-68357:
old: Patched/fixed-version: Fixed from version 6.17.5000
new: Patched/fixed-version: Fixed from version 6.12.64
CVE-2025-71145:
old: Patched/fixed-version: Fixed from version 5.10.5000
new: Unpatched/known-affected: No known resolution
CVE-2026-23288:
old: Patched/fixed-version: only affects 6.19.4 onwards
new: Patched/fixed-version: only affects 6.19 onwards
CVE-2026-23327:
old: Unpatched/version-in-range: Needs backporting (fixed from 6.19.7)
new: Unpatched/version-in-range: Needs backporting (fixed from 7.0rc2)
CVE-2026-23328:
old: Unpatched/version-in-range: Needs backporting (fixed from 6.19.7)
new: Unpatched/version-in-range: Needs backporting (fixed from 7.0rc3)
CVE-2026-23333:
old: Unpatched/version-in-range: Needs backporting (fixed from 6.19.4)
new: Unpatched/known-affected: No known resolution
CVE-2026-23341:
old: Patched/fixed-version: only affects 6.19.4 onwards
new: Patched/fixed-version: only affects 6.19 onwards
CVE-2026-23355:
old: Unpatched/version-in-range: Needs backporting (fixed from 7.0rc3)
new: Patched/cpe-stable-backport: Backported in 6.18.18
CVE-2026-23371:
old: Unpatched/version-in-range: Needs backporting (fixed from 6.19.7)
new: Unpatched/version-in-range: Needs backporting (fixed from 7.0rc3)
CVE-2026-23374:
old: Unpatched/version-in-range: Needs backporting (fixed from 6.19.7)
new: Unpatched/version-in-range: Needs backporting (fixed from 7.0rc3)
CVE-2026-23377:
old: Unpatched/version-in-range: Needs backporting (fixed from 6.19.7)
new: Unpatched/version-in-range: Needs backporting (fixed from 7.0rc3)
CVE-2026-23389:
old: Unpatched/version-in-range: Needs backporting (fixed from 6.19.7)
new: Unpatched/version-in-range: Needs backporting (fixed from 7.0rc3)
CVE-2026-23394:
old: Unpatched/version-in-range: Needs backporting (fixed from 6.19.10)
new: Unpatched/version-in-range: Needs backporting (fixed from
7.0rc5)

Quoted 22 lines

Best regards,
Daniel

Quoted 19 lines

-----Original Message-----
From: Benjamin Robin <[email protected]>
Sent: Friday, 17 April 2026 15:32
To: [email protected]; Daniel Turull <[email protected]>
Subject: Re: [PATCH] improve_kernel_cve_report: use numeric versions instead of cpeApplicability

[You don't often get email from [email protected]. Learn why this is important at
https://aka.ms/LearnAboutSenderIdentification ]

Hello Daniel,

Quoted 4 lines · On Friday, April 17, 2026 at 3:24 PM, [email protected] wrote:

From: Daniel Turull <[email protected]>

git shas or versions should be use instead of cpeApplicability. Reuse the same logic as generate-cve-exclusions, so outputs are

consistent.

Quoted 6 lines
Quoted 6 lines
Quoted 4 lines

cpeApplicability does not provide accurate version information and for some CVEs the information is not the same. This came from a discussion that we had with Greg Kroah-Hartma, member of the Linux

security team.

Quoted 8 lines
Quoted 8 lines

Indeed "cpeApplicability" does not provide the same kind of information that the "versions" node.
In sbom-cve-check (the latest version in main branch) we are using both sources of information.
But you are saying that "cpeApplicability" does not provide accurate version information. Could you elaborate and give various examples? I never saw something invalid in "cpeApplicability".

--
Benjamin Robin, Bootlin
Embedded Linux and Kernel engineering
https://eur02.safelinks.protection.outlook.com/?url=https%3A%2F%2Fbootlin
.com%2F&data=05%7C02%7Cdaniel.turull%40ericsson.com%7Cdbbb990e606 245d2f01c08de9c88ecab%7C92e84cebfbfd47abbe52080c6b87953f%7C0%7C0 %7C639120309071164067%7CUnknown%7CTWFpbGZsb3d8eyJFbXB0eU1hcG kiOnRydWUsIlYiOiIwLjAuMDAwMCIsIlAiOiJXaW4zMiIsIkFOIjoiTWFpbCIsIldUIj oyfQ%3D%3D%7C0%7C%7C%7C&sdata=mzgiHIfwb7lLWnnQ%2Fo8KpUkKlUN 5KrZb7eij1g%2BTaIg%3D&reserved=0

#5 lore raw
Quoted 16 lines · Daniel Turull

True,
I sent CVEs that had different responses. I used an old checkout and rerun with the old and the new script.

I'm starting to find some issues with the data that we should clarify before merging this patch. Let's pause it and have it correct.

For example,
https://git.kernel.org/pub/scm/linux/security/vulns.git/tree/cve/published/2025/CVE-2025-40067.json

"defaultStatus": "affected",

{
                     "version": "6.6.112",
                     "lessThanOrEqual": "6.6.*",
                     "status": "unaffected",
                     "versionType": "semver"
                  },
Quoted 7 lines

"negate": false,

{
                           "vulnerable": true,
                           "criteria": "cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*",
                           "versionStartIncluding": "6.6.102",
                           "versionEndExcluding": "6.6.112"
                        },

I see nothing wrong here.
In [>6.6.102, <6.6.112] => vulnerable, and [>=6.6.112, <=6.6.*] => not vulnerable.

Quoted 1 line

In this case, the information in one of the entries is not correct. For example, in 6.6.100 is vulnerable in the version not in the cpeApplicability, but git versions and cpeApplicability match if we do a git describe on them.

I really did not understood this sentence.
Be aware that if you do a git describe, this is the tag that can be reached that is displayed, not the tag that include the commit!

Could you elaborate in more details, I may be wrong here...

Quoted 3 lines

I'll send an email to ask for clarification to the kernel security team and try to see other similar cases. I must say that this is a minority of all CVEs.

Also I need to look if this could be integrated in the sbom-cve-check, so we have it only one place. I want to be able to run the script as well with older releases or just telling the kernel to use without SBOM.

This is already integrated to sbom-cve-check (I pushed today in main), but the algorithm is completely different.
And again since NVD database is enabled (by default) the result is not as good as if you are only using the CVEList entries.

I was going to propose to drop this Python script that you are working on.

Quoted 30 lines

Best regards,
Daniel

Quoted 25 lines

-----Original Message-----
From: Benjamin Robin <[email protected]>
Sent: Friday, 17 April 2026 15:55
To: [email protected]; Daniel Turull <[email protected]>
Subject: Re: [PATCH] improve_kernel_cve_report: use numeric versions instead of cpeApplicability

Quoted 2 lines · Daniel Turull

Hi,
We had Greg visiting us and I asked him what is better to use and he said git

or versions, not cvepAplicability that has issues defining trees.

You reply is technically not responding to my question :) Could you provide at least one example with an entry that is not correct in the cpeApplicability node?

The script had previously various issue (or at least it looked like it). I preferred to use a completely different algorithm (and using all sources of information)

But since I am also using NVD entries this degrade a bit the quality of the generated assessment message

Quoted 1 line

I have done some comparison with 6.6.100 and 6.18.22

Signature · 3 lines
Benjamin Robin, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com
#6 lore raw

What I tried to say was that in this case using versions if we have 6.6.100 we think we are vulnerable, but the cve was introduced in a backport in 6.6.102. Only the range 6.6.102 to 6.6.112 should be vulnerable. I'll see if I can report the issues so the source data is correct.

If we plan to drop the script, I should probably not spend too much time on the script itself but on the data quality.

Have a nice weekend
Daniel

________________________________
From: Benjamin Robin <[email protected]>
Sent: Friday, April 17, 2026 4:47 PM
To: [email protected] <[email protected]>; Daniel Turull <[email protected]>
Subject: Re: [PATCH] improve_kernel_cve_report: use numeric versions instead of cpeApplicability

Quoted 16 lines · Daniel Turull

True,
I sent CVEs that had different responses. I used an old checkout and rerun with the old and the new script.

I'm starting to find some issues with the data that we should clarify before merging this patch. Let's pause it and have it correct.

For example,
https://eur02.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgit.kernel.org%2Fpub%2Fscm%2Flinux%2Fsecurity%2Fvulns.git%2Ftree%2Fcve%2Fpublished%2F2025%2FCVE-2025-40067.json&data=05%7C02%7Cdaniel.turull%40ericsson.com%7C51a3d6a1ffb84372d85e08de9c903527%7C92e84cebfbfd47abbe52080c6b87953f%7C0%7C0%7C639120340343085371%7CUnknown%7CTWFpbGZsb3d8eyJFbXB0eU1hcGkiOnRydWUsIlYiOiIwLjAuMDAwMCIsIlAiOiJXaW4zMiIsIkFOIjoiTWFpbCIsIldUIjoyfQ%3D%3D%7C0%7C%7C%7C&sdata=o0yl%2FWDeKZqspA3cFaha2hCBt9hvxVkAygWN6UqXlkY%3D&reserved=0<https://git.kernel.org/pub/scm/linux/security/vulns.git/tree/cve/published/2025/CVE-2025-40067.json>

"defaultStatus": "affected",

{
                     "version": "6.6.112",
                     "lessThanOrEqual": "6.6.*",
                     "status": "unaffected",
                     "versionType": "semver"
                  },
Quoted 7 lines

"negate": false,

{
                           "vulnerable": true,
                           "criteria": "cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*",
                           "versionStartIncluding": "6.6.102",
                           "versionEndExcluding": "6.6.112"
                        },

I see nothing wrong here.
In [>6.6.102, <6.6.112] => vulnerable, and [>=6.6.112, <=6.6.*] => not vulnerable.

Quoted 1 line

In this case, the information in one of the entries is not correct. For example, in 6.6.100 is vulnerable in the version not in the cpeApplicability, but git versions and cpeApplicability match if we do a git describe on them.

I really did not understood this sentence.
Be aware that if you do a git describe, this is the tag that can be reached that is displayed, not the tag that include the commit!

Could you elaborate in more details, I may be wrong here...

Quoted 3 lines

I'll send an email to ask for clarification to the kernel security team and try to see other similar cases. I must say that this is a minority of all CVEs.

Also I need to look if this could be integrated in the sbom-cve-check, so we have it only one place. I want to be able to run the script as well with older releases or just telling the kernel to use without SBOM.

This is already integrated to sbom-cve-check (I pushed today in main), but the algorithm is completely different.
And again since NVD database is enabled (by default) the result is not as good as if you are only using the CVEList entries.

I was going to propose to drop this Python script that you are working on.

Quoted 30 lines

Best regards,
Daniel

Quoted 25 lines

-----Original Message-----
From: Benjamin Robin <[email protected]>
Sent: Friday, 17 April 2026 15:55
To: [email protected]; Daniel Turull <[email protected]>
Subject: Re: [PATCH] improve_kernel_cve_report: use numeric versions instead of cpeApplicability

Quoted 2 lines · Daniel Turull

Hi,
We had Greg visiting us and I asked him what is better to use and he said git

or versions, not cvepAplicability that has issues defining trees.

You reply is technically not responding to my question :) Could you provide at least one example with an entry that is not correct in the cpeApplicability node?

The script had previously various issue (or at least it looked like it). I preferred to use a completely different algorithm (and using all sources of information)

But since I am also using NVD entries this degrade a bit the quality of the generated assessment message

Quoted 1 line

I have done some comparison with 6.6.100 and 6.18.22

--
Benjamin Robin, Bootlin
Embedded Linux and Kernel engineering
https://eur02.safelinks.protection.outlook.com/?url=https%3A%2F%2Fbootlin.com%2F&data=05%7C02%7Cdaniel.turull%40ericsson.com%7C51a3d6a1ffb84372d85e08de9c903527%7C92e84cebfbfd47abbe52080c6b87953f%7C0%7C0%7C639120340343100584%7CUnknown%7CTWFpbGZsb3d8eyJFbXB0eU1hcGkiOnRydWUsIlYiOiIwLjAuMDAwMCIsIlAiOiJXaW4zMiIsIkFOIjoiTWFpbCIsIldUIjoyfQ%3D%3D%7C0%7C%7C%7C&sdata=xAn%2BE3rQ%2BUyoTyyLFkMBoTgm2ibB328y1PRu0FY9afY%3D&reserved=0<https://bootlin.com/>

#7 lore raw

Hello Daniel,

Quoted 1 line · Daniel Turull

What I tried to say was that in this case using versions if we have 6.6.100 we think we are vulnerable, but the cve was introduced in a backport in 6.6.102. Only the range 6.6.102 to 6.6.112 should be vulnerable. I'll see if I can report the issues so the source data is correct.

From the version ranges:
- https://cveawg.mitre.org/api/cve/CVE-2025-40067
- https://git.kernel.org/pub/scm/linux/security/vulns.git/tree/cve/published/2025/CVE-2025-40067.json

The "cpeApplicability" node says:

{
    "vulnerable": true,
    "criteria": "cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*",
    "versionStartIncluding": "6.6.102",
    "versionEndExcluding": "6.6.112"
},

So it is very clear that the first vulnerable version starts from 6.6.102

And the "versions" node says:

{
    "version": "6.6.112",
    "lessThanOrEqual": "6.6.*",
    "status": "unaffected",
    "versionType": "semver"
},

Which is coherent, version 6.6.112 is not vulnerable.

So in summary we are having:
- [>=6.6.0, <6.6.102] => not vulnerable,
- [>=6.6.102, <6.6.112] => vulnerable,
- [>=6.6.112, <=6.6.*] => not vulnerable.

To have a full picture, we must use all the information ("cpeApplicability" and "versions" nodes).
But now I am starting to understand what you are saying :)

Quoted 1 line

in this case using versions if we have 6.6.100 we think we are vulnerable

If we are only looking at the "versions" node, we may think that any version below 6.6.102 is vulnerable (e.g., the 6.6.100 version).

This is not the only CVE that is like that; a *lot* of kernel CVEs are in this case. So I don't know why Greg Kroah-Hartma said that cpeApplicability nodes should not be used. Without it, the full picture of which version is vulnerable cannot be obtained.

Quoted 1 line

If we plan to drop the script, I should probably not spend too much time on the script itself but on the data quality.

I don't know if we are going to drop the script. Previously the script had some flows (and it still has), and since sbom-cve-check (in the next release) should do better, I was going to propose its suppression (In a few weeks).

Quoted 43 lines

Have a nice weekend
Daniel

________________________________
From: Benjamin Robin <[email protected]>
Sent: Friday, April 17, 2026 4:47 PM
To: [email protected] <[email protected]>; Daniel Turull <[email protected]>
Subject: Re: [PATCH] improve_kernel_cve_report: use numeric versions instead of cpeApplicability

Quoted 15 lines · Daniel Turull

True,
I sent CVEs that had different responses. I used an old checkout and rerun with the old and the new script.

I'm starting to find some issues with the data that we should clarify before merging this patch. Let's pause it and have it correct.

For example,

"defaultStatus": "affected",

{
                     "version": "6.6.112",
                     "lessThanOrEqual": "6.6.*",
                     "status": "unaffected",
                     "versionType": "semver"
                  },
Quoted 7 lines

"negate": false,

{
                           "vulnerable": true,
                           "criteria": "cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*",
                           "versionStartIncluding": "6.6.102",
                           "versionEndExcluding": "6.6.112"
                        },

I see nothing wrong here.
In [>6.6.102, <6.6.112] => vulnerable, and [>=6.6.112, <=6.6.*] => not vulnerable.

Quoted 1 line

In this case, the information in one of the entries is not correct. For example, in 6.6.100 is vulnerable in the version not in the cpeApplicability, but git versions and cpeApplicability match if we do a git describe on them.

I really did not understood this sentence.
Be aware that if you do a git describe, this is the tag that can be reached that is displayed, not the tag that include the commit!

Could you elaborate in more details, I may be wrong here...

Signature · 3 lines
Benjamin Robin, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com
#8 lore raw
Quoted 54 lines

-----Original Message-----
From: [email protected] <openembedded- [email protected]> On Behalf Of Benjamin Robin via lists.openembedded.org
Sent: Sunday, 19 April 2026 11:07
To: [email protected]; Daniel Turull <[email protected]>
Subject: Re: [OE-core] [PATCH] improve_kernel_cve_report: use numeric versions instead of cpeApplicability

Hello Daniel,

Quoted 1 line · Daniel Turull

What I tried to say was that in this case using versions if we have 6.6.100 we

think we are vulnerable, but the cve was introduced in a backport in 6.6.102. Only the range 6.6.102 to 6.6.112 should be vulnerable. I'll see if I can report the issues so the source data is correct.

From the version ranges:
-

The "cpeApplicability" node says:

{
    "vulnerable": true,
    "criteria": "cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*",
    "versionStartIncluding": "6.6.102",
    "versionEndExcluding": "6.6.112"
},

So it is very clear that the first vulnerable version starts from 6.6.102

And the "versions" node says:

{
    "version": "6.6.112",
    "lessThanOrEqual": "6.6.*",
    "status": "unaffected",
    "versionType": "semver"
},

Which is coherent, version 6.6.112 is not vulnerable.

So in summary we are having:
- [>=6.6.0, <6.6.102] => not vulnerable,
- [>=6.6.102, <6.6.112] => vulnerable,
- [>=6.6.112, <=6.6.*] => not vulnerable.

To have a full picture, we must use all the information ("cpeApplicability" and "versions" nodes).
But now I am starting to understand what you are saying :)

Quoted 5 lines
Quoted 2 lines

in this case using versions if we have 6.6.100 we think we are vulnerable

If we are only looking at the "versions" node, we may think that any version below 6.6.102 is vulnerable (e.g., the 6.6.100 version).

Exactly.

Quoted 5 lines

This is not the only CVE that is like that; a *lot* of kernel CVEs are in this case. So I don't know why Greg Kroah-Hartma said that cpeApplicability nodes should not be used. Without it, the full picture of which version is vulnerable cannot be obtained.

I think is because if you look literally the versions without looking like a tree (which we do) the information is not correct. For example, imagine that the correction is 6.6.100, then any 6.7.X is not affected, but that 6.6.100 was a backport and it was never backported to 6.7.X since it is not supported anymore.

Quoted 8 lines
Quoted 2 lines

If we plan to drop the script, I should probably not spend too much time on

the script itself but on the data quality.

I don't know if we are going to drop the script.
Previously the script had some flows (and it still has), and since sbom-cve- check (in the next release) should do better, I was going to propose its suppression (In a few weeks).

Can you point to the flaws, so I'm aware?

Thanks
Daniel

Quoted 57 lines
Quoted 12 lines

Have a nice weekend
Daniel

________________________________
From: Benjamin Robin <[email protected]>
Sent: Friday, April 17, 2026 4:47 PM
To: [email protected] <[email protected]>; Daniel Turull <[email protected]>
Subject: Re: [PATCH] improve_kernel_cve_report: use numeric versions instead of cpeApplicability

Quoted 3 lines
Quoted 2 lines · Daniel Turull

True,
I sent CVEs that had different responses. I used an old checkout and rerun

with the old and the new script.

Quoted 2 lines
Quoted 2 lines

I'm starting to find some issues with the data that we should clarify before

merging this patch. Let's pause it and have it correct.

Quoted 19 lines
Quoted 11 lines

For example,

"defaultStatus": "affected",

{
                     "version": "6.6.112",
                     "lessThanOrEqual": "6.6.*",
                     "status": "unaffected",
                     "versionType": "semver"
                  },
Quoted 7 lines

"negate": false,

{
                           "vulnerable": true,
                           "criteria": "cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*",
                           "versionStartIncluding": "6.6.102",
                           "versionEndExcluding": "6.6.112"
                        },
Quoted 2 lines

I see nothing wrong here.
In [>6.6.102, <6.6.112] => vulnerable, and [>=6.6.112, <=6.6.*] => not

vulnerable.

Quoted 2 lines
Quoted 2 lines

In this case, the information in one of the entries is not correct. For

example, in 6.6.100 is vulnerable in the version not in the cpeApplicability, but git versions and cpeApplicability match if we do a git describe on them.

Quoted 6 lines

I really did not understood this sentence.
Be aware that if you do a git describe, this is the tag that can be reached that is displayed, not the tag that include the commit!

Could you elaborate in more details, I may be wrong here...

--
Benjamin Robin, Bootlin
Embedded Linux and Kernel engineering

#9 lore raw
Quoted 62 lines · Daniel Turull
Quoted 53 lines

-----Original Message-----
From: [email protected] <openembedded- [email protected]> On Behalf Of Benjamin Robin via lists.openembedded.org
Sent: Sunday, 19 April 2026 11:07
To: [email protected]; Daniel Turull <[email protected]>
Subject: Re: [OE-core] [PATCH] improve_kernel_cve_report: use numeric versions instead of cpeApplicability

Hello Daniel,

Quoted 1 line · Daniel Turull

What I tried to say was that in this case using versions if we have 6.6.100 we

think we are vulnerable, but the cve was introduced in a backport in 6.6.102. Only the range 6.6.102 to 6.6.112 should be vulnerable. I'll see if I can report the issues so the source data is correct.

From the version ranges:
-

The "cpeApplicability" node says:

{
    "vulnerable": true,
    "criteria": "cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*",
    "versionStartIncluding": "6.6.102",
    "versionEndExcluding": "6.6.112"
},

So it is very clear that the first vulnerable version starts from 6.6.102

And the "versions" node says:

{
    "version": "6.6.112",
    "lessThanOrEqual": "6.6.*",
    "status": "unaffected",
    "versionType": "semver"
},

Which is coherent, version 6.6.112 is not vulnerable.

So in summary we are having:
- [>=6.6.0, <6.6.102] => not vulnerable,
- [>=6.6.102, <6.6.112] => vulnerable,
- [>=6.6.112, <=6.6.*] => not vulnerable.

To have a full picture, we must use all the information ("cpeApplicability" and "versions" nodes).
But now I am starting to understand what you are saying :)

Quoted 5 lines
Quoted 2 lines

in this case using versions if we have 6.6.100 we think we are vulnerable

If we are only looking at the "versions" node, we may think that any version below 6.6.102 is vulnerable (e.g., the 6.6.100 version).

Exactly.

Quoted 1 line
Quoted 8 lines
Quoted 4 lines

This is not the only CVE that is like that; a *lot* of kernel CVEs are in this case. So I don't know why Greg Kroah-Hartma said that cpeApplicability nodes should not be used. Without it, the full picture of which version is vulnerable cannot be obtained.

I think is because if you look literally the versions without looking like a tree (which we do) the information is not correct. For example, imagine that the correction is 6.6.100, then any 6.7.X is not affected, but that 6.6.100 was a backport and it was never backported to 6.7.X since it is not supported anymore.

Yes I understand that, but why we should not look it like a tree? sbom-cve-check in main branch has a special case for the kernel which do a specific processing for all these version ranges provided by the kernel CNA.

Quoted 9 lines
Quoted 7 lines
Quoted 1 line

If we plan to drop the script, I should probably not spend too much time on

the script itself but on the data quality.

I don't know if we are going to drop the script.
Previously the script had some flows (and it still has), and since sbom-cve- check (in the next release) should do better, I was going to propose its suppression (In a few weeks).

Can you point to the flaws, so I'm aware?

The current version of improve_kernel_cve_report.py does not process the "versions" nodes. We need both sources of information.

I did not double check if these findings are right (but at first glance it looked wrong):
- If at least one entry does not have versionStartIncluding defined, first_affected is going to be 0.
- It does not handle versionStartExcluding
- It does not handle the fact that we could have multiples ranges within the same branch that is vulnerable.
- Maybe there are others issues, I did not spent to much time analyzing it since I choose to use a different way to do that.

Quoted 3 lines

Thanks
Daniel

Signature · 3 lines
Benjamin Robin, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com
#10 lore raw
Quoted 85 lines

-----Original Message-----
From: Benjamin Robin <[email protected]>
Sent: Monday, 20 April 2026 09:30
To: [email protected]; Daniel Turull <[email protected]>
Subject: Re: [OE-core] [PATCH] improve_kernel_cve_report: use numeric versions instead of cpeApplicability

Quoted 11 lines · Daniel Turull
Quoted 11 lines

-----Original Message-----
From: [email protected] <openembedded- [email protected]> On Behalf Of Benjamin Robin via lists.openembedded.org
Sent: Sunday, 19 April 2026 11:07
To: [email protected]; Daniel Turull <[email protected]>
Subject: Re: [OE-core] [PATCH] improve_kernel_cve_report: use numeric versions instead of cpeApplicability

Quoted 6 lines
Quoted 6 lines

Hello Daniel,

Quoted 2 lines · Daniel Turull

What I tried to say was that in this case using versions if we have 6.6.100 we

think we are vulnerable, but the cve was introduced in a backport in

6.6.102.

Quoted 38 lines
Quoted 38 lines

Only the range 6.6.102 to 6.6.112 should be vulnerable. I'll see if I can report the issues so the source data is correct.

From the version ranges:
-

The "cpeApplicability" node says:

{
    "vulnerable": true,
    "criteria": "cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*",
    "versionStartIncluding": "6.6.102",
    "versionEndExcluding": "6.6.112"
},

So it is very clear that the first vulnerable version starts from 6.6.102

And the "versions" node says:

{
    "version": "6.6.112",
    "lessThanOrEqual": "6.6.*",
    "status": "unaffected",
    "versionType": "semver"
},

Which is coherent, version 6.6.112 is not vulnerable.

So in summary we are having:
- [>=6.6.0, <6.6.102] => not vulnerable,
- [>=6.6.102, <6.6.112] => vulnerable,
- [>=6.6.112, <=6.6.*] => not vulnerable.

To have a full picture, we must use all the information ("cpeApplicability" and "versions" nodes).
But now I am starting to understand what you are saying :)

Quoted 10 lines
Quoted 5 lines
Quoted 2 lines

in this case using versions if we have 6.6.100 we think we are vulnerable

If we are only looking at the "versions" node, we may think that any version below 6.6.102 is vulnerable (e.g., the 6.6.100 version).

Exactly.

Quoted 2 lines

This is not the only CVE that is like that; a *lot* of kernel CVEs are in this

case.

Quoted 7 lines
Quoted 3 lines

So I don't know why Greg Kroah-Hartma said that cpeApplicability nodes should not be used. Without it, the full picture of which version is vulnerable cannot be obtained.

I think is because if you look literally the versions without looking like a tree (which we do) the information is not correct. For example, imagine that the correction is 6.6.100, then any 6.7.X is not affected, but that

Quoted 6 lines

6.6.100 was a backport and it was never backported to 6.7.X since it is not supported anymore.

Yes I understand that, but why we should not look it like a tree? sbom-cve- check in main branch has a special case for the kernel which do a specific processing for all these version ranges provided by the kernel CNA.

I'm not saying that we should not look that a tree. We are doing the correct thing. Only if you take the naive approach and compare versions 6.6.100 < 6.7.1, even 6.6.100 has a fix and 6.7.1 not.

Quoted 14 lines
Quoted 10 lines
Quoted 8 lines
Quoted 2 lines

If we plan to drop the script, I should probably not spend too much time on

the script itself but on the data quality.

I don't know if we are going to drop the script. Previously the script had some flows (and it still has), and since sbom-cve- check (in the next release) should do better, I was going to propose its suppression (In a few weeks).

Can you point to the flaws, so I'm aware?

The current version of improve_kernel_cve_report.py does not process the "versions" nodes. We need both sources of information.

Quoted 9 lines

I did not double check if these findings are right (but at first glance it looked wrong):
- If at least one entry does not have versionStartIncluding defined, first_affected is going to be 0.
- It does not handle versionStartExcluding
- It does not handle the fact that we could have multiples ranges within the same branch that is vulnerable.
- Maybe there are others issues, I did not spent to much time analyzing it since I choose to use a different way to do that.

Thanks, I need to look at start using sbom-cve-check. It is true, there are a few corner cases.

Quoted 16 lines
Quoted 3 lines

Thanks
Daniel

--
Benjamin Robin, Bootlin
Embedded Linux and Kernel engineering
https://eur02.safelinks.protection.outlook.com/?url=https%3A%2F%2Fbootlin
.com%2F&data=05%7C02%7Cdaniel.turull%40ericsson.com%7C02c979105b9 a41cfe17908de9eaea603%7C92e84cebfbfd47abbe52080c6b87953f%7C0%7C0 %7C639122670115586461%7CUnknown%7CTWFpbGZsb3d8eyJFbXB0eU1hcG kiOnRydWUsIlYiOiIwLjAuMDAwMCIsIlAiOiJXaW4zMiIsIkFOIjoiTWFpbCIsIldUIj oyfQ%3D%3D%7C0%7C%7C%7C&sdata=xJjwN9UNO5bFS5i8MciYNCwue%2FD yWf5BFd7cwFb%2BrdQ%3D&reserved=0

#11 lore raw
Quoted 119 lines · Daniel Turull via lists.openembedded.org
Quoted 85 lines

-----Original Message-----
From: Benjamin Robin <[email protected]>
Sent: Monday, 20 April 2026 09:30
To: [email protected]; Daniel Turull <[email protected]>
Subject: Re: [OE-core] [PATCH] improve_kernel_cve_report: use numeric versions instead of cpeApplicability

Quoted 11 lines · Daniel Turull
Quoted 11 lines

-----Original Message-----
From: [email protected] <openembedded- [email protected]> On Behalf Of Benjamin Robin via lists.openembedded.org
Sent: Sunday, 19 April 2026 11:07
To: [email protected]; Daniel Turull <[email protected]>
Subject: Re: [OE-core] [PATCH] improve_kernel_cve_report: use numeric versions instead of cpeApplicability

Quoted 6 lines
Quoted 6 lines

Hello Daniel,

Quoted 2 lines · Daniel Turull

What I tried to say was that in this case using versions if we have 6.6.100 we

think we are vulnerable, but the cve was introduced in a backport in

6.6.102.

Quoted 38 lines
Quoted 38 lines

Only the range 6.6.102 to 6.6.112 should be vulnerable. I'll see if I can report the issues so the source data is correct.

From the version ranges:
 -

The "cpeApplicability" node says:

{
    "vulnerable": true,
    "criteria": "cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*", "versionStartIncluding": "6.6.102", "versionEndExcluding": "6.6.112"
},

So it is very clear that the first vulnerable version starts from 6.6.102

And the "versions" node says:

{
    "version": "6.6.112",
    "lessThanOrEqual": "6.6.*",
    "status": "unaffected",
    "versionType": "semver"
},

Which is coherent, version 6.6.112 is not vulnerable.

So in summary we are having:
 - [>=6.6.0, <6.6.102] => not vulnerable,
 - [>=6.6.102, <6.6.112] => vulnerable,
 - [>=6.6.112, <=6.6.*] => not vulnerable.

To have a full picture, we must use all the information ("cpeApplicability" and "versions" nodes).
But now I am starting to understand what you are saying :)

Quoted 10 lines
Quoted 5 lines
Quoted 2 lines

in this case using versions if we have 6.6.100 we think we are vulnerable

If we are only looking at the "versions" node, we may think that any version below 6.6.102 is vulnerable (e.g., the 6.6.100 version).

Exactly.

Quoted 2 lines

This is not the only CVE that is like that; a *lot* of kernel CVEs are in this

case.

Quoted 7 lines
Quoted 3 lines

So I don't know why Greg Kroah-Hartma said that cpeApplicability nodes should not be used. Without it, the full picture of which version is vulnerable cannot be obtained.

I think is because if you look literally the versions without looking like a tree (which we do) the information is not correct. For example, imagine that the correction is 6.6.100, then any 6.7.X is not affected, but that

Quoted 6 lines

6.6.100 was a backport and it was never backported to 6.7.X since it is not supported anymore.

Yes I understand that, but why we should not look it like a tree? sbom-cve- check in main branch has a special case for the kernel which do a specific processing for all these version ranges provided by the kernel CNA.

I'm not saying that we should not look that a tree. We are doing the correct thing. Only if you take the naive approach and compare versions 6.6.100 < 6.7.1, even 6.6.100 has a fix and 6.7.1 not.

Quoted 14 lines
Quoted 10 lines
Quoted 8 lines
Quoted 2 lines

If we plan to drop the script, I should probably not spend too much time on

the script itself but on the data quality.

I don't know if we are going to drop the script. Previously the script had some flows (and it still has), and since sbom-cve- check (in the next release) should do better, I was going to propose its suppression (In a few weeks).

Can you point to the flaws, so I'm aware?

The current version of improve_kernel_cve_report.py does not process the "versions" nodes. We need both sources of information.

Quoted 9 lines

I did not double check if these findings are right (but at first glance it looked wrong):
 - If at least one entry does not have versionStartIncluding defined, first_affected is going to be 0.
 - It does not handle versionStartExcluding
 - It does not handle the fact that we could have multiples ranges within the same branch that is vulnerable.
 - Maybe there are others issues, I did not spent to much time analyzing it since I choose to use a different way to do that.

Quoted 1 line

Thanks, I need to look at start using sbom-cve-check. It is true, there are a few corner cases.

Did we reach a consensus on what we should do with this patch?

We (as in patch review/triage) couldn't quite understand what the end result was!

Cheers,

Richard

#12 lore raw
Quoted 120 lines · Richard Purdie
Quoted 119 lines · Daniel Turull via lists.openembedded.org
Quoted 85 lines

-----Original Message-----
From: Benjamin Robin <[email protected]>
Sent: Monday, 20 April 2026 09:30
To: [email protected]; Daniel Turull <[email protected]>
Subject: Re: [OE-core] [PATCH] improve_kernel_cve_report: use numeric versions instead of cpeApplicability

Quoted 11 lines · Daniel Turull
Quoted 11 lines

-----Original Message-----
From: [email protected] <openembedded- [email protected]> On Behalf Of Benjamin Robin via lists.openembedded.org
Sent: Sunday, 19 April 2026 11:07
To: [email protected]; Daniel Turull <[email protected]>
Subject: Re: [OE-core] [PATCH] improve_kernel_cve_report: use numeric versions instead of cpeApplicability

Quoted 6 lines
Quoted 6 lines

Hello Daniel,

Quoted 2 lines · Daniel Turull

What I tried to say was that in this case using versions if we have 6.6.100 we

think we are vulnerable, but the cve was introduced in a backport in

6.6.102.

Quoted 38 lines
Quoted 38 lines

Only the range 6.6.102 to 6.6.112 should be vulnerable. I'll see if I can report the issues so the source data is correct.

From the version ranges:
-

The "cpeApplicability" node says:

{
    "vulnerable": true,
    "criteria": "cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*",
    "versionStartIncluding": "6.6.102",
    "versionEndExcluding": "6.6.112"
},

So it is very clear that the first vulnerable version starts from 6.6.102

And the "versions" node says:

{
    "version": "6.6.112",
    "lessThanOrEqual": "6.6.*",
    "status": "unaffected",
    "versionType": "semver"
},

Which is coherent, version 6.6.112 is not vulnerable.

So in summary we are having:
- [>=6.6.0, <6.6.102] => not vulnerable,
- [>=6.6.102, <6.6.112] => vulnerable,
- [>=6.6.112, <=6.6.*] => not vulnerable.

To have a full picture, we must use all the information ("cpeApplicability" and "versions" nodes).
But now I am starting to understand what you are saying :)

Quoted 10 lines
Quoted 5 lines
Quoted 2 lines

in this case using versions if we have 6.6.100 we think we are vulnerable

If we are only looking at the "versions" node, we may think that any version below 6.6.102 is vulnerable (e.g., the 6.6.100 version).

Exactly.

Quoted 2 lines

This is not the only CVE that is like that; a *lot* of kernel CVEs are in this

case.

Quoted 7 lines
Quoted 3 lines

So I don't know why Greg Kroah-Hartma said that cpeApplicability nodes should not be used. Without it, the full picture of which version is vulnerable cannot be obtained.

I think is because if you look literally the versions without looking like a tree (which we do) the information is not correct. For example, imagine that the correction is 6.6.100, then any 6.7.X is not affected, but that

Quoted 6 lines

6.6.100 was a backport and it was never backported to 6.7.X since it is not supported anymore.

Yes I understand that, but why we should not look it like a tree? sbom-cve- check in main branch has a special case for the kernel which do a specific processing for all these version ranges provided by the kernel CNA.

I'm not saying that we should not look that a tree. We are doing the correct thing. Only if you take the naive approach and compare versions 6.6.100 < 6.7.1, even 6.6.100 has a fix and 6.7.1 not.

Quoted 23 lines
Quoted 10 lines
Quoted 8 lines
Quoted 2 lines

If we plan to drop the script, I should probably not spend too much time on

the script itself but on the data quality.

I don't know if we are going to drop the script. Previously the script had some flows (and it still has), and since sbom-cve- check (in the next release) should do better, I was going to propose its suppression (In a few weeks).

Can you point to the flaws, so I'm aware?

The current version of improve_kernel_cve_report.py does not process the "versions" nodes. We need both sources of information.

I did not double check if these findings are right (but at first glance it looked wrong):
- If at least one entry does not have versionStartIncluding defined, first_affected is going to be 0.
- It does not handle versionStartExcluding
- It does not handle the fact that we could have multiples ranges within the same branch that is vulnerable.
- Maybe there are others issues, I did not spent to much time analyzing it since I choose to use a different way to do that.

Quoted 3 lines
Quoted 1 line

Thanks, I need to look at start using sbom-cve-check. It is true, there are a few corner cases.

Did we reach a consensus on what we should do with this patch?

I am not sure.
Currently this patch is "not good enough" (see the whole discussion for the reasons).

But, sbom-cve-check should have a "better" algorithm, and this script should no longer be necessary. And since cve-check class was removed, I don't know if this script is still useful.

Quoted 5 lines

We (as in patch review/triage) couldn't quite understand what the end result was!

Cheers,

Richard

Signature · 3 lines
Benjamin Robin, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com
#13 lore raw
In reply to Richard Purdie

Let’s drop it for now. I’m working with Greg and Sasha from the cve kernel team to fix the scripts that generate the cve metadata. Then I’ll update both improve_kernel_cve_report and gen_linux_exlcusions.sh to report the same and be consistent.

I still believe that since this script is in contrib and independent, we should keep regarless what sbom-cve-check does. I'll maintain it and make sure we have the most accurate information

Best regards
Daniel
________________________________
From: Richard Purdie <[email protected]>
Sent: Thursday, April 30, 2026 11:34:13 AM
To: Daniel Turull <[email protected]>; Benjamin Robin <[email protected]>; [email protected] <[email protected]>
Subject: Re: [OE-core] [PATCH] improve_kernel_cve_report: use numeric versions instead of cpeApplicability

Quoted 119 lines · Daniel Turull via lists.openembedded.org
Quoted 85 lines

-----Original Message-----
From: Benjamin Robin <[email protected]>
Sent: Monday, 20 April 2026 09:30
To: [email protected]; Daniel Turull <[email protected]>
Subject: Re: [OE-core] [PATCH] improve_kernel_cve_report: use numeric versions instead of cpeApplicability

Quoted 11 lines · Daniel Turull
Quoted 11 lines

-----Original Message-----
From: [email protected] <openembedded- [email protected]> On Behalf Of Benjamin Robin via lists.openembedded.org
Sent: Sunday, 19 April 2026 11:07
To: [email protected]; Daniel Turull <[email protected]>
Subject: Re: [OE-core] [PATCH] improve_kernel_cve_report: use numeric versions instead of cpeApplicability

Quoted 6 lines
Quoted 6 lines

Hello Daniel,

Quoted 2 lines · Daniel Turull

What I tried to say was that in this case using versions if we have 6.6.100 we

think we are vulnerable, but the cve was introduced in a backport in

6.6.102.

Quoted 38 lines
Quoted 38 lines

Only the range 6.6.102 to 6.6.112 should be vulnerable. I'll see if I can report the issues so the source data is correct.

From the version ranges:
-

The "cpeApplicability" node says:

{
    "vulnerable": true,
    "criteria": "cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*",
    "versionStartIncluding": "6.6.102",
    "versionEndExcluding": "6.6.112"
},

So it is very clear that the first vulnerable version starts from 6.6.102

And the "versions" node says:

{
    "version": "6.6.112",
    "lessThanOrEqual": "6.6.*",
    "status": "unaffected",
    "versionType": "semver"
},

Which is coherent, version 6.6.112 is not vulnerable.

So in summary we are having:
- [>=6.6.0, <6.6.102] => not vulnerable,
- [>=6.6.102, <6.6.112] => vulnerable,
- [>=6.6.112, <=6.6.*] => not vulnerable.

To have a full picture, we must use all the information ("cpeApplicability" and "versions" nodes).
But now I am starting to understand what you are saying :)

Quoted 10 lines
Quoted 5 lines
Quoted 2 lines

in this case using versions if we have 6.6.100 we think we are vulnerable

If we are only looking at the "versions" node, we may think that any version below 6.6.102 is vulnerable (e.g., the 6.6.100 version).

Exactly.

Quoted 2 lines

This is not the only CVE that is like that; a *lot* of kernel CVEs are in this

case.

Quoted 7 lines
Quoted 3 lines

So I don't know why Greg Kroah-Hartma said that cpeApplicability nodes should not be used. Without it, the full picture of which version is vulnerable cannot be obtained.

I think is because if you look literally the versions without looking like a tree (which we do) the information is not correct. For example, imagine that the correction is 6.6.100, then any 6.7.X is not affected, but that

Quoted 6 lines

6.6.100 was a backport and it was never backported to 6.7.X since it is not supported anymore.

Yes I understand that, but why we should not look it like a tree? sbom-cve- check in main branch has a special case for the kernel which do a specific processing for all these version ranges provided by the kernel CNA.

I'm not saying that we should not look that a tree. We are doing the correct thing. Only if you take the naive approach and compare versions 6.6.100 < 6.7.1, even 6.6.100 has a fix and 6.7.1 not.

Quoted 14 lines
Quoted 10 lines
Quoted 8 lines
Quoted 2 lines

If we plan to drop the script, I should probably not spend too much time on

the script itself but on the data quality.

I don't know if we are going to drop the script. Previously the script had some flows (and it still has), and since sbom-cve- check (in the next release) should do better, I was going to propose its suppression (In a few weeks).

Can you point to the flaws, so I'm aware?

The current version of improve_kernel_cve_report.py does not process the "versions" nodes. We need both sources of information.

Quoted 9 lines

I did not double check if these findings are right (but at first glance it looked wrong):
- If at least one entry does not have versionStartIncluding defined, first_affected is going to be 0.
- It does not handle versionStartExcluding
- It does not handle the fact that we could have multiples ranges within the same branch that is vulnerable.
- Maybe there are others issues, I did not spent to much time analyzing it since I choose to use a different way to do that.

Quoted 1 line

Thanks, I need to look at start using sbom-cve-check. It is true, there are a few corner cases.

Did we reach a consensus on what we should do with this patch?

We (as in patch review/triage) couldn't quite understand what the end result was!

Cheers,

Richard