netfilter: ctnetlink: fix refcount leak on table dump
In the Linux kernel, the following vulnerability has been resolved: netfilter: ctnetlink: fix refcount leak on table dump There is a reference count leak in ctnetlink_dump_table(): if (res < 0) { nf_conntrack_get(&ct->ct_general); // HERE cb->args[1] = (unsigned long)ct; ... While its very unlikely, its possible that ct == last. If this happens, then the refcount of ct was already incremented. This 2nd increment is never undone. This prevents the conntrack object from being released, which in turn keeps prevents cnet->count from dropping back to 0. This will then block the netns dismantle (or conntrack rmmod) as nf_conntrack_cleanup_net_list() will wait forever. This can be reproduced by running conntrack_resize.sh selftest in a loop. It takes ~20 minutes for me on a preemptible kernel on average before I see a runaway kworker spinning in nf_conntrack_cleanup_net_list. One fix would to change this to: if (res < 0) { if (ct != last) nf_conntrack_get(&ct->ct_general); But this reference counting isn't needed in the first place. We can just store a cookie value instead. A followup patch will do the same for ctnetlink_exp_dump_table, it looks to me as if this has the same problem and like ctnetlink_dump_table, we only need a 'skip hint', not the actual object so we can apply the same cookie strategy there as well.
- Products
- Kernel
- Affected versions
- Not specified
- Fixed versions
- 5.10.241, 5.15.190, 5.4.297, 6.1.149, 6.12.43, 6.15.11, 6.16.2, 6.6.103
- CVEs
- CVE-2025-38721
- Confirmed fixed CVEs
- Not specified
- Updated in this feed
- 2 Nov 2026, 22:43 UTC · Revision 1
Source evidence
netfilter: ctnetlink: fix refcount leak on table dump
core-osv · Original sourcePublished 4 Sept 2025, 15:33 UTC
In the Linux kernel, the following vulnerability has been resolved: netfilter: ctnetlink: fix refcount leak on table dump There is a reference count leak in ctnetlink_dump_table(): if (res < 0) { nf_conntrack_get(&ct->ct_general); // HERE cb->args[1] = (unsigned long)ct; ... While its very unlikely, its possible that ct == last. If this happens, then the refcount of ct was already incremented. This 2nd increment is never undone. This prevents the conntrack object from being released, which in turn keeps prevents cnet->count from dropping back to 0. This will then block the netns dismantle (or conntrack rmmod) as nf_conntrack_cleanup_net_list() will wait forever. This can be reproduced by running conntrack_resize.sh selftest in a loop. It takes ~20 minutes for me on a preemptible kernel on average before I see a runaway kworker spinning in nf_conntrack_cleanup_net_list. One fix would to change this to: if (res < 0) { if (ct != last) nf_conntrack_get(&ct->ct_general); But this reference counting isn't needed in the first place. We can just store a cookie value instead. A followup patch will do the same for ctnetlink_exp_dump_table, it looks to me as if this has the same problem and like ctnetlink_dump_table, we only need a 'skip hint', not the actual object so we can apply the same cookie strategy there as well.
osv:CVE-2025-38721
Kernel
Fixed versions: 5.4.297, 5.10.241, 5.15.190, 6.1.149, 6.6.103, 6.12.43, 6.15.11, 6.16.2
Complete source record
{
"id": "7f230c4b61a1e356efb3bc80b82fa46b333b8ccfc0608d09fcc973ef3f35f6a4",
"url": "https://osv.dev/vulnerability/CVE-2025-38721",
"method": "structured",
"payload": {
"kind": "advisory",
"title": "netfilter: ctnetlink: fix refcount leak on table dump",
"summary": "In the Linux kernel, the following vulnerability has been resolved: netfilter: ctnetlink: fix refcount leak on table dump There is a reference count leak in ctnetlink_dump_table(): if (res < 0) { nf_conntrack_get(&ct->ct_general); // HERE cb->args[1] = (unsigned long)ct; ... While its very unlikely, its possible that ct == last. If this happens, then the refcount of ct was already incremented. This 2nd increment is never undone. This prevents the conntrack object from being released, which in turn keeps prevents cnet->count from dropping back to 0. This will then block the netns dismantle (or conntrack rmmod) as nf_conntrack_cleanup_net_list() will wait forever. This can be reproduced by running conntrack_resize.sh selftest in a loop. It takes ~20 minutes for me on a preemptible kernel on average before I see a runaway kworker spinning in nf_conntrack_cleanup_net_list. One fix would to change this to: if (res < 0) { if (ct != last) nf_conntrack_get(&ct->ct_general); But this reference counting isn't needed in the first place. We can just store a cookie value instead. A followup patch will do the same for ctnetlink_exp_dump_table, it looks to me as if this has the same problem and like ctnetlink_dump_table, we only need a 'skip hint', not the actual object so we can apply the same cookie strategy there as well.",
"evidence": "osv:CVE-2025-38721",
"products": [
{
"name": {
"value": "Kernel",
"evidence": "/affected/1/package/name"
},
"prerequisites": [],
"fixed_versions": [
{
"value": "5.4.297",
"evidence": "/affected/1/ranges"
},
{
"value": "5.10.241",
"evidence": "/affected/1/ranges"
},
{
"value": "5.15.190",
"evidence": "/affected/1/ranges"
},
{
"value": "6.1.149",
"evidence": "/affected/1/ranges"
},
{
"value": "6.6.103",
"evidence": "/affected/1/ranges"
},
{
"value": "6.12.43",
"evidence": "/affected/1/ranges"
},
{
"value": "6.15.11",
"evidence": "/affected/1/ranges"
},
{
"value": "6.16.2",
"evidence": "/affected/1/ranges"
}
],
"affected_versions": []
}
],
"severity": "unknown",
"structured": {
"id": "CVE-2025-38721",
"aliases": [
"CVE-2025-38721"
],
"affected": [
{
"ranges": [
{
"repo": "https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git",
"type": "GIT",
"events": [
{
"introduced": "d205dc40798d97d63ad348bfaf7394f445d152d4"
},
{
"fixed": "586892e341fbf698e7cbaca293e1353957db725a"
},
{
"fixed": "962518c6ca9f9a13df099cafa429f72f68ad61f0"
},
{
"fixed": "19b909a4b1452fb97e477d2f08b97f8d04095619"
},
{
"fixed": "41462f4cfc583513833f87f9ee55d12da651a7e3"
},
{
"fixed": "30cf811058552b8cd0e98dff677ef3f89d6d34ce"
},
{
"fixed": "a2cb4df7872de069f809de2f076ec8e54d649fe3"
},
{
"fixed": "e14f72aa66c029db106921d621edcedef68e065b"
},
{
"fixed": "a62d6aa3f31f216b637a4c71b7a8bfc7c57f049b"
},
{
"fixed": "de788b2e6227462b6dcd0e07474e72c089008f74"
}
]
}
],
"database_specific": {
"source": "https://storage.googleapis.com/cve-osv-conversion/osv-output/CVE-2025-38721.json"
}
},
{
"ranges": [
{
"type": "ECOSYSTEM",
"events": [
{
"introduced": "2.6.18"
},
{
"fixed": "5.4.297"
}
]
},
{
"type": "ECOSYSTEM",
"events": [
{
"introduced": "5.5.0"
},
{
"fixed": "5.10.241"
}
]
},
{
"type": "ECOSYSTEM",
"events": [
{
"introduced": "5.11.0"
},
{
"fixed": "5.15.190"
}
]
},
{
"type": "ECOSYSTEM",
"events": [
{
"introduced": "5.16.0"
},
{
"fixed": "6.1.149"
}
]
},
{
"type": "ECOSYSTEM",
"events": [
{
"introduced": "6.2.0"
},
{
"fixed": "6.6.103"
}
]
},
{
"type": "ECOSYSTEM",
"events": [
{
"introduced": "6.7.0"
},
{
"fixed": "6.12.43"
}
]
},
{
"type": "ECOSYSTEM",
"events": [
{
"introduced": "6.13.0"
},
{
"fixed": "6.15.11"
}
]
},
{
"type": "ECOSYSTEM",
"events": [
{
"introduced": "6.16.0"
},
{
"fixed": "6.16.2"
}
]
}
],
"package": {
"name": "Kernel",
"ecosystem": "Linux"
},
"database_specific": {
"source": "https://storage.googleapis.com/cve-osv-conversion/osv-output/CVE-2025-38721.json"
}
}
],
"modified": "2026-10-28T02:30:23.961670195Z",
"provider": "osv",
"references": [
"https://cert-portal.siemens.com/productcert/html/ssa-019113.html",
"https://cert-portal.siemens.com/productcert/html/ssa-032379.html",
"https://cert-portal.siemens.com/productcert/html/ssa-082556.html",
"https://cert-portal.siemens.com/productcert/html/ssa-457936.html",
"https://git.kernel.org/stable/c/19b909a4b1452fb97e477d2f08b97f8d04095619",
"https://git.kernel.org/stable/c/30cf811058552b8cd0e98dff677ef3f89d6d34ce",
"https://git.kernel.org/stable/c/41462f4cfc583513833f87f9ee55d12da651a7e3",
"https://git.kernel.org/stable/c/586892e341fbf698e7cbaca293e1353957db725a",
"https://git.kernel.org/stable/c/962518c6ca9f9a13df099cafa429f72f68ad61f0",
"https://git.kernel.org/stable/c/a2cb4df7872de069f809de2f076ec8e54d649fe3",
"https://git.kernel.org/stable/c/a62d6aa3f31f216b637a4c71b7a8bfc7c57f049b",
"https://git.kernel.org/stable/c/de788b2e6227462b6dcd0e07474e72c089008f74",
"https://git.kernel.org/stable/c/e14f72aa66c029db106921d621edcedef68e065b",
"https://lists.debian.org/debian-lts-announce/2025/10/msg00007.html",
"https://lists.debian.org/debian-lts-announce/2025/10/msg00008.html",
"https://github.com/CVEProject/cvelistV5/tree/main/cves/2025/38xxx/CVE-2025-38721.json",
"https://nvd.nist.gov/vuln/detail/CVE-2025-38721",
"https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git"
],
"provider_metadata": {
"cna_assigner": "Linux",
"osv_generated_from": "https://github.com/CVEProject/cvelistV5/tree/main/cves/2025/38xxx/CVE-2025-38721.json"
}
},
"related_cves": [
{
"value": "CVE-2025-38721",
"evidence": "/id"
}
],
"security_relevant": true,
"severity_evidence": null
},
"source_id": "core-osv",
"document_id": 490637,
"published_at": "2025-09-04T15:33:14.891+00:00"
}