Skip to content

Commit 6bc1c1d

Browse files
eastmadcclaude
eastmadc
and
claude
committed
feat(hw-fw): NVIDIA Tegra/L4T CVE pins in known_firmware.yaml
Closes Reviewer A C2 + Reviewer B F-FORENSIC-03 deferred from postmortem hw-firmware-adaptive-session-2026-05-18 Rec #3. 6 new curated CVE family entries covering the NVIDIA Tegra/L4T BSP cluster that DS1 (project d360f8f5) and other Jetson-based firmware images depend on. Per-CVE NVD CPE list verified via direct WebFetch on services.nvd.nist.gov/rest/json/cves/2.0?cveId=<CVE> recursively for EACH of the 6 CVEs — Scout report's NVD claims independently re-fetched before commit. Recursive-verification discipline (Reviewer B 2026-05-15..18 caught CVE-attribution failure modes 4 sessions in a row): every CVE- attribution claim — whether from user prompt, scout report, or reviewer finding — gets the direct NVD URL fetch before pinning. The 6 fetches surfaced 3 user-prompt-discrepancies that NVD overruled: * CVE-2021-1111 — user said "ALL Jetsons"; NVD CPE EXCLUDES TX1/Nano (only AGX-Xavier+TX2+TX2-NX+Xavier-NX listed). chipset_regex narrowed accordingly. * CVE-2021-34372 — user said product=tegra_tos_trusty; NVD CPE uses product=jetson_linux. wairz curated families: schema has no `product` field (only banner_cve_pin schema does); we narrow via vendor=nvidia + category=tee, which is the schematic equivalent for the Trusty OTE attack surface. * CVE-2022-42269 — user said "AGX-Xavier/TX2/Xavier-family"; NVD CPE ALSO includes jetson_tx1. chipset_regex expanded to include T210/TX1 (Nano stays excluded — different hardware anchor). NVD wins per Rule #19 evidence-first in all 3 cases. All 6 pins satisfy F-FORENSIC-10 narrowing (chipset_regex AND/OR version_regex present — no family-only attribution). Per-CVE summary =============== | CVE | Fix at | Scope | Discrepancy? | |-----|--------|-------|--------------| | CVE-2019-5680 | R32.2 | TX1 only | None (R32.2 IS fix) | | CVE-2021-1111 | R32.6.1 | AGX-Xav+TX2+TX2-NX+Xav-NX | NVD excludes TX1+Nano | | CVE-2021-34372 | R32.5.1 | ALL Jetsons (TEE) | product=jetson_linux not tegra_tos_trusty | | CVE-2021-34397 | R32.5.1 | TX2+Xavier family | None | | CVE-2022-42269 | R32.7.2 | AGX-Xav+TX2+Xav-NX+TX1 | NVD adds TX1 | | CVE-2022-42270 | R32.7.2 | Xavier only (NVDLA) | None | Forward-prepared note ===================== version_regex requires the L4T BSP release string ("R32.x.y" style) to appear in blob.version OR any blob metadata value. The current Tegra parser (commit 8054d22) does NOT yet extract L4T release from blob content — these pins will fire when a future enhancement adds L4T release extraction (e.g. parse /etc/nv_tegra_release on the firmware tree → populate firmware.device_metadata["l4t_release"], OR scan Tegra-blob head for "R<N> (release), REVISION: <x.y>" banner string and store in blob.metadata). Strict version_regex is the right discipline NOW per Reviewer B discipline — better forward-prepared with zero false-positives than firing on every Tegra blob regardless of L4T version. Out-of-scope (explicit per user direction) ========================================== CVE-2021-34373..34396 from the same NVIDIA security bulletin disclosure batch — DO NOT batch-extrapolate. Each future CVE needs its own NVD WebFetch verification before pinning. The disclosure- batch antipattern has bitten wairz 4 sessions in a row (BTFM 2026-05-15, CVE-2021-28139 2026-05-16, CVE-2021-34147/31609/31612 2026-05-17, CVE-2019-5680 Selfblow 2026-05-18); the per-NVD-CPE recipe applies recursively for any future Tegra coverage extension. Tests (10 new in test_hardware_firmware_cve_matcher.py) ======================================================= - test_tegra_cve_pins_all_six_loaded — all 6 entries present - test_tegra_cve_pins_satisfy_f_forensic_10_narrowing — chipset_regex or version_regex on every pin (no family-only) - test_cve_2019_5680_chipset_regex_tx1_only — accepts TX1, rejects TX2/Xavier/Nano (canonical Selfblow scope) - test_cve_2021_1111_chipset_regex_excludes_tx1_and_nano — Rule #46 paired canary for the user-prompt discrepancy - test_cve_2022_42269_chipset_regex_includes_tx1 — Rule #46 paired canary for the TX1-inclusion discrepancy - test_cve_2022_42270_xavier_only_excludes_tx2 — NVDLA Xavier-only - test_cve_2021_34397_excludes_tx1_and_nano — TX2/Xavier family - test_cve_2021_34372_has_no_chipset_regex_per_nvd_all_jetsons — confirms intentional chipset_regex absence (ALL-Jetsons scope) - test_tegra_version_regex_matches_pre_fix_l4t_releases — forward-prepared canary that pre-fix L4T release strings WILL match when extraction lands - test_tegra_cve_pins_carry_nvd_url_reference — verifiability discipline: each pin's notes cites the NVD URL 66/66 cve_matcher tests pass. No regressions. Refs: - Postmortem hw-firmware-adaptive-session-2026-05-18 Rec #3 - NVD CPE WebFetch for each CVE (see notes field URLs) - CLAUDE.md Rule #19 (evidence-first — NVD CPE list IS truth) - CLAUDE.md Rule #25 (per-piece commits) - CLAUDE.md Rule #46 (paired-canary discipline) Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
1 parent 8054d22 commit 6bc1c1d

2 files changed

Lines changed: 443 additions & 0 deletions

File tree

‎backend/app/services/hardware_firmware/known_firmware.yaml‎

Lines changed: 167 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -1144,3 +1144,170 @@ families:
11441144
cache side-channel), -9496 (invalid-curve ECC), -13377 (SAE
11451145
Brainpool timing). All four fixed by wpa_supplicant 2.9.
11461146
Affects any pre-2.9 build.
1147+
1148+
# ===================================================================
1149+
# NVIDIA Tegra / L4T BSP CVE cluster — added 2026-05-15 per
1150+
# postmortem hw-firmware-adaptive-session-2026-05-18 Rec #3
1151+
# (Reviewer A C2 + Reviewer B F-FORENSIC-03 deferred). Each pin
1152+
# verified against NVD CPE list via direct WebFetch on
1153+
# services.nvd.nist.gov/rest/json/cves/2.0?cveId=<CVE>. Recursive-
1154+
# verification discipline (Reviewer B 2026-05-15..18) applied:
1155+
# Scout report's per-CVE CPE claims independently re-fetched
1156+
# before commit.
1157+
#
1158+
# 3 of 6 had user-prompt-discrepancies that NVD overruled:
1159+
# * CVE-2021-1111 — user said "ALL Jetsons"; NVD CPE EXCLUDES
1160+
# TX1/Nano (only AGX-Xavier+TX2+TX2-NX+Xavier-NX).
1161+
# * CVE-2021-34372 — user said product=tegra_tos_trusty; NVD CPE
1162+
# uses product=jetson_linux. wairz curated
1163+
# schema has no `product` field; we narrow via
1164+
# vendor=nvidia + category=tee instead.
1165+
# * CVE-2022-42269 — user said "AGX-Xavier/TX2/Xavier-family";
1166+
# NVD CPE ALSO includes jetson_tx1 (T210).
1167+
#
1168+
# NVD wins per Rule #19 evidence-first in all 3 cases.
1169+
#
1170+
# FORWARD-PREPARED NOTE: version_regex requires L4T BSP release
1171+
# string ("R32.3.1" etc.) to appear in blob.version OR any blob
1172+
# metadata value. The current Tegra parser (parsers/tegra_blob.py,
1173+
# shipped commit 8054d22 2026-05-15) does NOT yet extract L4T
1174+
# release from blob content. These pins will fire when a future
1175+
# enhancement adds L4T release extraction (e.g. parse
1176+
# /etc/nv_tegra_release on the firmware tree → populate
1177+
# firmware.device_metadata["l4t_release"], OR scan Tegra-blob
1178+
# head for "R<N> (release), REVISION: <x.y>" banner string and
1179+
# store in blob.metadata). Until then the pins are documented
1180+
# forward-prepared entries — strict version_regex prevents
1181+
# over-attribution (Reviewer B 2026-05-18 discipline).
1182+
#
1183+
# OUT OF SCOPE this session per user direction: CVE-2021-34373..
1184+
# 34396 from the same NVIDIA security bulletin disclosure batch.
1185+
# The per-NVD-CPE recipe applies recursively for any future session
1186+
# extending Tegra coverage — DO NOT batch-extrapolate. Each CVE
1187+
# gets its own NVD WebFetch verification before pinning.
1188+
#
1189+
# Tegra chipset codes: T210 = TX1/Nano (Maxwell), T186 = TX2
1190+
# (Pascal), T194 = Xavier (Volta), T234 = Orin (Ampere; not
1191+
# affected by any of these 6 CVEs per NVD CPE).
1192+
# ===================================================================
1193+
1194+
- name: Tegra TX1 nvtboot Selfblow (CVE-2019-5680)
1195+
vendor: nvidia
1196+
category: bootloader
1197+
chipset_regex: "(?i)^(t210|tegra210|tx1)$"
1198+
version_regex: "(?i)(r3[01]\\b|r32\\.[01]\\b|r32\\.0\\.|r32\\.1\\.)"
1199+
strict_chipset: false
1200+
cves:
1201+
- CVE-2019-5680
1202+
severity: high
1203+
cvss_score: 6.7
1204+
notes: |
1205+
nvtboot load-address validation bypass on Tegra TX1 (T210).
1206+
NVD CPE: jetson_tx1_firmware versionEndExcluding=r32.2.
1207+
TX1-ONLY — explicitly EXCLUDES TX2 / Xavier / Nano.
1208+
Fix landed in L4T R32.2 (June 2019); R32.2 IS the fix cohort
1209+
(Reviewer B 2026-05-18 caught earlier R32.3.1 mis-attribution).
1210+
Reference: https://nvd.nist.gov/vuln/detail/CVE-2019-5680
1211+
1212+
- name: Tegra NV3P USB recovery bounds-check (CVE-2021-1111)
1213+
vendor: nvidia
1214+
category: bootloader
1215+
chipset_regex: "(?i)^(t186|t194|tegra186|tegra194|tx2|tx2-nx|tx2nx|xavier-nx|xaviernx|agx-xavier|agxxavier)$"
1216+
version_regex: "(?i)(r32\\.[1-5]\\b|r32\\.6\\.0\\b|r32\\.[1-5]\\.|r32\\.6\\.0\\.)"
1217+
strict_chipset: false
1218+
cves:
1219+
- CVE-2021-1111
1220+
severity: medium
1221+
cvss_score: 6.0
1222+
notes: |
1223+
NV3P recovery-mode bounds-check failure (USB physical access).
1224+
NVD CPE: jetson_linux versionStartIncluding=32.1 AND
1225+
versionEndExcluding=32.6.1. Hardware = AGX-Xavier + TX2 +
1226+
TX2-NX + Xavier-NX ONLY. TX1 + Nano + Nano-2GB EXPLICITLY
1227+
EXCLUDED per NVD (contradicts session-prompt's "ALL Jetsons"
1228+
claim — NVD wins per Rule #19 evidence-first).
1229+
Fix in L4T R32.6.1 (Aug 2021).
1230+
Reference: https://nvd.nist.gov/vuln/detail/CVE-2021-1111
1231+
1232+
- name: Tegra Trusty OTE integer overflow (CVE-2021-34372)
1233+
vendor: nvidia
1234+
category: tee
1235+
version_regex: "(?i)(r3[01]\\b|r32\\.[0-4]\\b|r32\\.5\\.0\\b|r3[01]\\.|r32\\.[0-4]\\.|r32\\.5\\.0\\.)"
1236+
strict_chipset: false
1237+
cves:
1238+
- CVE-2021-34372
1239+
severity: high
1240+
cvss_score: 7.8
1241+
notes: |
1242+
Integer overflow in malloc() size calc in Trusty OS OTE
1243+
protocol parser. NVD CPE: jetson_linux versionEndExcluding=
1244+
32.5.1. Hardware list = ALL Jetsons (TX1 + TX2 family + AGX-
1245+
Xavier + Xavier-NX + Nano + Nano-2GB) — the broadest scope of
1246+
the 6 Tegra CVEs.
1247+
Session prompt claimed product=tegra_tos_trusty; NVD CPE uses
1248+
product=jetson_linux (NVD wins per Rule #19). wairz curated
1249+
schema has no `product` field; we narrow via vendor=nvidia +
1250+
category=tee instead. Fix in L4T R32.5.1.
1251+
Reference: https://nvd.nist.gov/vuln/detail/CVE-2021-34372
1252+
1253+
- name: Tegra MB2 free-the-wrong-heap (CVE-2021-34397)
1254+
vendor: nvidia
1255+
category: bootloader
1256+
chipset_regex: "(?i)^(t186|t194|tegra186|tegra194|tx2|tx2-4gb|tx2-nx|tx2nx|tx2i|xavier-nx|xaviernx|agx-xavier|agxxavier)$"
1257+
version_regex: "(?i)(r3[01]\\b|r32\\.[0-4]\\b|r32\\.5\\.0\\b|r3[01]\\.|r32\\.[0-4]\\.|r32\\.5\\.0\\.)"
1258+
strict_chipset: false
1259+
cves:
1260+
- CVE-2021-34397
1261+
severity: medium
1262+
cvss_score: 5.5
1263+
notes: |
1264+
MB2 bootloader free-the-wrong-heap → limited DoS. NVD CPE:
1265+
jetson_linux versionEndExcluding=32.5.1. Hardware anchors =
1266+
AGX-Xavier (8/16/32GB) + TX2 (incl. 4GB/i/NX) + Xavier-NX
1267+
(dev/prod). TX1 + Nano EXPLICITLY EXCLUDED per NVD.
1268+
Fix in L4T R32.5.1.
1269+
Reference: https://nvd.nist.gov/vuln/detail/CVE-2021-34397
1270+
1271+
- name: Tegra Trusty SMC input validation (CVE-2022-42269)
1272+
vendor: nvidia
1273+
category: tee
1274+
chipset_regex: "(?i)^(t210|t186|t194|tegra210|tegra186|tegra194|tx1|tx2|tx2-4gb|tx2-nx|tx2nx|tx2i|xavier-nx|xaviernx|xavier-nx-16gb|agx-xavier|agxxavier|agx-xavier-8gb|agx-xavier-16gb|agx-xavier-32gb|agx-xavier-64gb|agx-xavier-industrial)$"
1275+
version_regex: "(?i)(r3[01]\\b|r32\\.[0-6]\\b|r32\\.7\\.[01]\\b|r3[01]\\.|r32\\.[0-6]\\.|r32\\.7\\.[01]\\.)"
1276+
strict_chipset: false
1277+
cves:
1278+
- CVE-2022-42269
1279+
severity: high
1280+
cvss_score: 7.9
1281+
notes: |
1282+
Trusty OS SMC handler missing input validation → info disclosure
1283+
+ integrity compromise. NVD CPE: jetson_linux versionEnd
1284+
Excluding=32.7.2. Hardware anchors INCLUDE jetson_tx1 in
1285+
addition to AGX-Xavier (6 SKUs) + TX2 (4 SKUs incl. TX2-NX) +
1286+
Xavier-NX (2 SKUs) — contradicts session-prompt's
1287+
"AGX-Xavier/TX2/Xavier-family" scope which omitted TX1.
1288+
NVD wins per Rule #19 evidence-first; TX1 added to
1289+
chipset_regex. Jetson Nano (also T210 but a different
1290+
hardware anchor) is NOT in the NVD CPE list and is excluded.
1291+
Fix in L4T R32.7.2 (Aug 2022).
1292+
Reference: https://nvd.nist.gov/vuln/detail/CVE-2022-42269
1293+
1294+
- name: NVDLA emu_task_submit Xavier-only (CVE-2022-42270)
1295+
vendor: nvidia
1296+
category: kernel_module
1297+
chipset_regex: "(?i)^(t194|tegra194|xavier-nx|xaviernx|xavier-nx-16gb|agx-xavier|agxxavier|agx-xavier-8gb|agx-xavier-16gb|agx-xavier-32gb|agx-xavier-64gb|agx-xavier-industrial)$"
1298+
version_regex: "(?i)(r3[01]\\b|r32\\.[0-6]\\b|r32\\.7\\.[01]\\b|r3[01]\\.|r32\\.[0-6]\\.|r32\\.7\\.[01]\\.)"
1299+
strict_chipset: false
1300+
cves:
1301+
- CVE-2022-42270
1302+
severity: high
1303+
cvss_score: 7.8
1304+
notes: |
1305+
NVDLA emu_task_submit stack-based buffer overflow. NVD CPE:
1306+
jetson_linux versionEndExcluding=32.7.2. Xavier-ONLY: NVDLA
1307+
Deep Learning Accelerator is only on T194/Xavier silicon.
1308+
Hardware anchors = AGX-Xavier (6 SKUs) + Xavier-NX (2 SKUs).
1309+
TX1 / TX2 / Nano EXPLICITLY EXCLUDED per NVD.
1310+
category=kernel_module because NVDLA driver lives in
1311+
nvidia.ko / nvgpu kernel module.
1312+
Fix in L4T R32.7.2.
1313+
Reference: https://nvd.nist.gov/vuln/detail/CVE-2022-42270

0 commit comments

Comments
 (0)