{"api_version":"1","generated_at":"2026-09-26T03:03:19+00:00","cve":"CVE-2026-98104","urls":{"html":"https://cve.report/CVE-2026-98104","api":"https://cve.report/api/cve/CVE-2026-98104.json","docs":"https://cve.report/api","cve_org":"https://www.cve.org/CVERecord?id=CVE-2026-98104","nvd":"https://nvd.nist.gov/vuln/detail/CVE-2026-98104"},"summary":{"title":"net/sched: cls_u32: fix duplicate handle when node ID pool is exhausted","description":"In the Linux kernel, the following vulnerability has been resolved:\n\nnet/sched: cls_u32: fix duplicate handle when node ID pool is exhausted\n\ngen_new_kid() falls back to returning max (htid | 0xFFF) when both\nidr_alloc_u32() ranges are full, instead of reporting an error.\nu32_change() trusts that value and inserts a new knode with a handle\nthat is already live in the hash table, breaking handle uniqueness\nwithin the table's node ID space.\n\nThe handle was never reserved in ht->handle_idr, so every later error\npath that does idr_remove(&ht->handle_idr, handle) removes the\nreservation of a different, live knode, which is then reused — one\nfailed add compounds into further duplicates.\n\nThe 4095 limit is per (table, bucket) — ht->handle_idr is per hash\ntable and the range is derived from htid (bucketid), so a table with\ndivisor 256 can legitimately hold 256*4095 knodes.\n\nThe sibling helper gen_new_htid() has the same silent in-band failure:\nit returns 0 when the tp_c handle pool (1..0x7FF) is full, and\nu32_init() publishes the root hash table with handle 0 without\nchecking.  Two root tables with handle 0 alias in u32_lookup_ht(),\nallowing cross-tcf_proto knode add/lookup/delete.  Add the same\nexhaustion check that the divisor path already has.\n\nReturn an error so u32_change() fails with ENOSPC/ENOMEM when the\nnode ID space is exhausted, and so u32_init() fails with -ENOMEM\nwhen the hash table ID space is exhausted.  The extack message\ndistinguishes pool exhaustion (-ENOSPC) from a transient allocation\nfailure (-ENOMEM).\n\nConditions to recreate the bug:\n- CONFIG_NET_SCHED=y, CONFIG_CLS_U32=y (or =m with module loaded)\n- Create a clsact qdisc on a device, then add 4095 u32 filters with\n  auto-generated handles to fill the node ID space for the root hash\n  table (single bucket). The 4096th auto-handle filter add triggers\n  the duplicate handle (fh 800::fff reused). Reachable at Level 2\n  (unshare -Urn, namespace-local CAP_NET_ADMIN).\n- For gen_new_htid: create 2047 u32 proto entries on the same block\n  to fill the tp_c handle pool, then create one more. The root table\n  gets handle 0 and aliases with other handle-0 root tables.","state":"PUBLISHED","assigner":"Linux","published_at":"2026-09-25 11:17:40","updated_at":"2026-09-25 11:17:40"},"problem_types":[],"metrics":[],"references":[{"url":"https://git.kernel.org/stable/c/feab9261b537df4ebb8350e4779bc185373059fd","name":"https://git.kernel.org/stable/c/feab9261b537df4ebb8350e4779bc185373059fd","refsource":"416baaa9-dc9f-4396-8d5f-8c081fb06d67","tags":[],"title":"","mime":"","httpstatus":"","archivestatus":"0"},{"url":"https://git.kernel.org/stable/c/f594f04268d01c5fdc975f3f51fc219ea2159ac6","name":"https://git.kernel.org/stable/c/f594f04268d01c5fdc975f3f51fc219ea2159ac6","refsource":"416baaa9-dc9f-4396-8d5f-8c081fb06d67","tags":[],"title":"","mime":"","httpstatus":"","archivestatus":"0"},{"url":"https://git.kernel.org/stable/c/d7e7e98d23f42a92d9ab7e36302bd96bd9b33b5f","name":"https://git.kernel.org/stable/c/d7e7e98d23f42a92d9ab7e36302bd96bd9b33b5f","refsource":"416baaa9-dc9f-4396-8d5f-8c081fb06d67","tags":[],"title":"","mime":"","httpstatus":"","archivestatus":"0"},{"url":"https://git.kernel.org/stable/c/6890e28840bae4f6805e8de981c4ec8e12a4e064","name":"https://git.kernel.org/stable/c/6890e28840bae4f6805e8de981c4ec8e12a4e064","refsource":"416baaa9-dc9f-4396-8d5f-8c081fb06d67","tags":[],"title":"","mime":"","httpstatus":"","archivestatus":"0"},{"url":"https://www.cve.org/CVERecord?id=CVE-2026-98104","name":"CVE Program record","refsource":"CVE.ORG","tags":["canonical"]},{"url":"https://nvd.nist.gov/vuln/detail/CVE-2026-98104","name":"NVD vulnerability detail","refsource":"NVD","tags":["canonical","analysis"]}],"affected":[{"source":"CNA","vendor":"Linux","product":"Linux","version":"affected 7801db8aec957fa6610efe0ee26a6c8bc0f1d73b 6890e28840bae4f6805e8de981c4ec8e12a4e064 git","platforms":[]},{"source":"CNA","vendor":"Linux","product":"Linux","version":"affected 7801db8aec957fa6610efe0ee26a6c8bc0f1d73b feab9261b537df4ebb8350e4779bc185373059fd git","platforms":[]},{"source":"CNA","vendor":"Linux","product":"Linux","version":"affected 7801db8aec957fa6610efe0ee26a6c8bc0f1d73b f594f04268d01c5fdc975f3f51fc219ea2159ac6 git","platforms":[]},{"source":"CNA","vendor":"Linux","product":"Linux","version":"affected 7801db8aec957fa6610efe0ee26a6c8bc0f1d73b d7e7e98d23f42a92d9ab7e36302bd96bd9b33b5f git","platforms":[]},{"source":"CNA","vendor":"Linux","product":"Linux","version":"affected 3.16","platforms":[]},{"source":"CNA","vendor":"Linux","product":"Linux","version":"unaffected 3.16 semver","platforms":[]},{"source":"CNA","vendor":"Linux","product":"Linux","version":"unaffected 6.12.111 6.12.* semver","platforms":[]},{"source":"CNA","vendor":"Linux","product":"Linux","version":"unaffected 6.18.53 6.18.* semver","platforms":[]},{"source":"CNA","vendor":"Linux","product":"Linux","version":"unaffected 7.2.7 7.2.* semver","platforms":[]},{"source":"CNA","vendor":"Linux","product":"Linux","version":"unaffected 7.3-rc2 * original_commit_for_fix","platforms":[]}],"timeline":[],"solutions":[],"workarounds":[],"exploits":[],"credits":[],"nvd_cpes":[],"vendor_comments":[],"enrichments":{"kev":null,"epss":null,"legacy_qids":[]},"source_records":{"cve_program":{"containers":{"cna":{"affected":[{"defaultStatus":"unaffected","product":"Linux","programFiles":["net/sched/cls_u32.c"],"repo":"https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git","vendor":"Linux","versions":[{"lessThan":"6890e28840bae4f6805e8de981c4ec8e12a4e064","status":"affected","version":"7801db8aec957fa6610efe0ee26a6c8bc0f1d73b","versionType":"git"},{"lessThan":"feab9261b537df4ebb8350e4779bc185373059fd","status":"affected","version":"7801db8aec957fa6610efe0ee26a6c8bc0f1d73b","versionType":"git"},{"lessThan":"f594f04268d01c5fdc975f3f51fc219ea2159ac6","status":"affected","version":"7801db8aec957fa6610efe0ee26a6c8bc0f1d73b","versionType":"git"},{"lessThan":"d7e7e98d23f42a92d9ab7e36302bd96bd9b33b5f","status":"affected","version":"7801db8aec957fa6610efe0ee26a6c8bc0f1d73b","versionType":"git"}]},{"defaultStatus":"affected","product":"Linux","programFiles":["net/sched/cls_u32.c"],"repo":"https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git","vendor":"Linux","versions":[{"status":"affected","version":"3.16"},{"lessThan":"3.16","status":"unaffected","version":"0","versionType":"semver"},{"lessThanOrEqual":"6.12.*","status":"unaffected","version":"6.12.111","versionType":"semver"},{"lessThanOrEqual":"6.18.*","status":"unaffected","version":"6.18.53","versionType":"semver"},{"lessThanOrEqual":"7.2.*","status":"unaffected","version":"7.2.7","versionType":"semver"},{"lessThanOrEqual":"*","status":"unaffected","version":"7.3-rc2","versionType":"original_commit_for_fix"}]}],"cpeApplicability":[{"nodes":[{"cpeMatch":[{"criteria":"cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*","versionEndExcluding":"6.12.111","versionStartIncluding":"3.16","vulnerable":true},{"criteria":"cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*","versionEndExcluding":"6.18.53","versionStartIncluding":"3.16","vulnerable":true},{"criteria":"cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*","versionEndExcluding":"7.2.7","versionStartIncluding":"3.16","vulnerable":true},{"criteria":"cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*","versionEndExcluding":"7.3-rc2","versionStartIncluding":"3.16","vulnerable":true}],"negate":false,"operator":"OR"}]}],"descriptions":[{"lang":"en","value":"In the Linux kernel, the following vulnerability has been resolved:\n\nnet/sched: cls_u32: fix duplicate handle when node ID pool is exhausted\n\ngen_new_kid() falls back to returning max (htid | 0xFFF) when both\nidr_alloc_u32() ranges are full, instead of reporting an error.\nu32_change() trusts that value and inserts a new knode with a handle\nthat is already live in the hash table, breaking handle uniqueness\nwithin the table's node ID space.\n\nThe handle was never reserved in ht->handle_idr, so every later error\npath that does idr_remove(&ht->handle_idr, handle) removes the\nreservation of a different, live knode, which is then reused — one\nfailed add compounds into further duplicates.\n\nThe 4095 limit is per (table, bucket) — ht->handle_idr is per hash\ntable and the range is derived from htid (bucketid), so a table with\ndivisor 256 can legitimately hold 256*4095 knodes.\n\nThe sibling helper gen_new_htid() has the same silent in-band failure:\nit returns 0 when the tp_c handle pool (1..0x7FF) is full, and\nu32_init() publishes the root hash table with handle 0 without\nchecking.  Two root tables with handle 0 alias in u32_lookup_ht(),\nallowing cross-tcf_proto knode add/lookup/delete.  Add the same\nexhaustion check that the divisor path already has.\n\nReturn an error so u32_change() fails with ENOSPC/ENOMEM when the\nnode ID space is exhausted, and so u32_init() fails with -ENOMEM\nwhen the hash table ID space is exhausted.  The extack message\ndistinguishes pool exhaustion (-ENOSPC) from a transient allocation\nfailure (-ENOMEM).\n\nConditions to recreate the bug:\n- CONFIG_NET_SCHED=y, CONFIG_CLS_U32=y (or =m with module loaded)\n- Create a clsact qdisc on a device, then add 4095 u32 filters with\n  auto-generated handles to fill the node ID space for the root hash\n  table (single bucket). The 4096th auto-handle filter add triggers\n  the duplicate handle (fh 800::fff reused). Reachable at Level 2\n  (unshare -Urn, namespace-local CAP_NET_ADMIN).\n- For gen_new_htid: create 2047 u32 proto entries on the same block\n  to fill the tp_c handle pool, then create one more. The root table\n  gets handle 0 and aliases with other handle-0 root tables."}],"providerMetadata":{"dateUpdated":"2026-09-25T10:35:54.544Z","orgId":"416baaa9-dc9f-4396-8d5f-8c081fb06d67","shortName":"Linux"},"references":[{"url":"https://git.kernel.org/stable/c/6890e28840bae4f6805e8de981c4ec8e12a4e064"},{"url":"https://git.kernel.org/stable/c/feab9261b537df4ebb8350e4779bc185373059fd"},{"url":"https://git.kernel.org/stable/c/f594f04268d01c5fdc975f3f51fc219ea2159ac6"},{"url":"https://git.kernel.org/stable/c/d7e7e98d23f42a92d9ab7e36302bd96bd9b33b5f"}],"title":"net/sched: cls_u32: fix duplicate handle when node ID pool is exhausted","x_generator":{"engine":"bippy-1.2.0"}}},"cveMetadata":{"assignerOrgId":"416baaa9-dc9f-4396-8d5f-8c081fb06d67","assignerShortName":"Linux","cveId":"CVE-2026-98104","datePublished":"2026-09-25T10:35:54.544Z","dateReserved":"2026-09-25T10:25:14.315Z","dateUpdated":"2026-09-25T10:35:54.544Z","state":"PUBLISHED"},"dataType":"CVE_RECORD","dataVersion":"5.2"},"nvd":{"publishedDate":"2026-09-25 11:17:40","lastModifiedDate":"2026-09-25 11:17:40","problem_types":[],"metrics":[],"configurations":[]},"legacy_mitre":{"record":{"CveYear":"2026","CveId":"98104","Ordinal":"1","Title":"net/sched: cls_u32: fix duplicate handle when node ID pool is ex","CVE":"CVE-2026-98104","Year":"2026"},"notes":[{"CveYear":"2026","CveId":"98104","Ordinal":"1","NoteData":"In the Linux kernel, the following vulnerability has been resolved:\n\nnet/sched: cls_u32: fix duplicate handle when node ID pool is exhausted\n\ngen_new_kid() falls back to returning max (htid | 0xFFF) when both\nidr_alloc_u32() ranges are full, instead of reporting an error.\nu32_change() trusts that value and inserts a new knode with a handle\nthat is already live in the hash table, breaking handle uniqueness\nwithin the table's node ID space.\n\nThe handle was never reserved in ht->handle_idr, so every later error\npath that does idr_remove(&ht->handle_idr, handle) removes the\nreservation of a different, live knode, which is then reused — one\nfailed add compounds into further duplicates.\n\nThe 4095 limit is per (table, bucket) — ht->handle_idr is per hash\ntable and the range is derived from htid (bucketid), so a table with\ndivisor 256 can legitimately hold 256*4095 knodes.\n\nThe sibling helper gen_new_htid() has the same silent in-band failure:\nit returns 0 when the tp_c handle pool (1..0x7FF) is full, and\nu32_init() publishes the root hash table with handle 0 without\nchecking.  Two root tables with handle 0 alias in u32_lookup_ht(),\nallowing cross-tcf_proto knode add/lookup/delete.  Add the same\nexhaustion check that the divisor path already has.\n\nReturn an error so u32_change() fails with ENOSPC/ENOMEM when the\nnode ID space is exhausted, and so u32_init() fails with -ENOMEM\nwhen the hash table ID space is exhausted.  The extack message\ndistinguishes pool exhaustion (-ENOSPC) from a transient allocation\nfailure (-ENOMEM).\n\nConditions to recreate the bug:\n- CONFIG_NET_SCHED=y, CONFIG_CLS_U32=y (or =m with module loaded)\n- Create a clsact qdisc on a device, then add 4095 u32 filters with\n  auto-generated handles to fill the node ID space for the root hash\n  table (single bucket). The 4096th auto-handle filter add triggers\n  the duplicate handle (fh 800::fff reused). Reachable at Level 2\n  (unshare -Urn, namespace-local CAP_NET_ADMIN).\n- For gen_new_htid: create 2047 u32 proto entries on the same block\n  to fill the tp_c handle pool, then create one more. The root table\n  gets handle 0 and aliases with other handle-0 root tables.","Type":"Description","Title":"net/sched: cls_u32: fix duplicate handle when node ID pool is ex"}]}}}