Description
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.
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.
No analysis available yet.
Remediation
No remediation available yet.
Tracking
Sign in to view the affected projects.
Advisories
| Source | ID | Title |
|---|---|---|
Debian DLA |
DLA-4327-1 | linux security update |
Debian DLA |
DLA-4328-1 | linux-6.1 security update |
Debian DSA |
DSA-6009-1 | linux security update |
EUVD |
EUVD-2025-26748 | 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. |
Ubuntu USN |
USN-7909-1 | Linux kernel vulnerabilities |
Ubuntu USN |
USN-7909-2 | Linux kernel (Real-time) vulnerabilities |
Ubuntu USN |
USN-7909-3 | Linux kernel (FIPS) vulnerabilities |
Ubuntu USN |
USN-7910-1 | Linux kernel (Azure FIPS) vulnerabilities |
Ubuntu USN |
USN-7909-4 | Linux kernel vulnerabilities |
Ubuntu USN |
USN-7910-2 | Linux kernel (Azure) vulnerabilities |
Ubuntu USN |
USN-7909-5 | Linux kernel (Raspberry Pi) vulnerabilities |
Ubuntu USN |
USN-7933-1 | Linux kernel (KVM) vulnerabilities |
Ubuntu USN |
USN-7938-1 | Linux kernel (Azure) vulnerabilities |
Ubuntu USN |
USN-8028-1 | Linux kernel vulnerabilities |
Ubuntu USN |
USN-8028-2 | Linux kernel (Real-time) vulnerabilities |
Ubuntu USN |
USN-8031-1 | Linux kernel (GCP) vulnerabilities |
Ubuntu USN |
USN-8028-3 | Linux kernel (Real-time) vulnerabilities |
Ubuntu USN |
USN-8028-4 | Linux kernel (FIPS) vulnerabilities |
Ubuntu USN |
USN-8028-5 | Linux kernel vulnerabilities |
Ubuntu USN |
USN-8031-2 | Linux kernel (GCP FIPS) vulnerabilities |
Ubuntu USN |
USN-8028-6 | Linux kernel (HWE) vulnerabilities |
Ubuntu USN |
USN-8031-3 | Linux kernel vulnerabilities |
Ubuntu USN |
USN-8052-1 | Linux kernel (Low Latency) vulnerabilities |
Ubuntu USN |
USN-8028-7 | Linux kernel (Low Latency NVIDIA) vulnerabilities |
Ubuntu USN |
USN-8028-8 | Linux kernel (IBM) vulnerabilities |
Ubuntu USN |
USN-8052-2 | Linux kernel (Xilinx) vulnerabilities |
Ubuntu USN |
USN-8074-1 | Linux kernel (Azure) vulnerabilities |
Ubuntu USN |
USN-8074-2 | Linux kernel (Azure FIPS) vulnerabilities |
Ubuntu USN |
USN-8126-1 | Linux kernel (Azure) vulnerabilities |
References
History
Fri, 09 Jan 2026 16:00:00 +0000
| Type | Values Removed | Values Added |
|---|---|---|
| First Time appeared |
Debian
Debian debian Linux |
|
| Weaknesses | NVD-CWE-Other | |
| CPEs | cpe:2.3:o:debian:debian_linux:11.0:*:*:*:*:*:*:* cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:* cpe:2.3:o:linux:linux_kernel:2.6.18:-:*:*:*:*:*:* cpe:2.3:o:linux:linux_kernel:2.6.18:rc5:*:*:*:*:*:* cpe:2.3:o:linux:linux_kernel:2.6.18:rc6:*:*:*:*:*:* cpe:2.3:o:linux:linux_kernel:2.6.18:rc7:*:*:*:*:*:* cpe:2.3:o:linux:linux_kernel:6.17:rc1:*:*:*:*:*:* |
|
| Vendors & Products |
Debian
Debian debian Linux |
Mon, 03 Nov 2025 18:30:00 +0000
| Type | Values Removed | Values Added |
|---|---|---|
| References |
|
Fri, 05 Sep 2025 14:15:00 +0000
| Type | Values Removed | Values Added |
|---|---|---|
| First Time appeared |
Linux
Linux linux Kernel |
|
| Vendors & Products |
Linux
Linux linux Kernel |
Fri, 05 Sep 2025 00:15:00 +0000
| Type | Values Removed | Values Added |
|---|---|---|
| References |
| |
| Metrics |
threat_severity
|
cvssV3_1
|
Thu, 04 Sep 2025 15:45:00 +0000
| Type | Values Removed | Values Added |
|---|---|---|
| Description | 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. | |
| Title | netfilter: ctnetlink: fix refcount leak on table dump | |
| References |
|
|
Status: PUBLISHED
Assigner: Linux
Published:
Updated: 2025-11-03T17:41:50.589Z
Reserved: 2025-04-16T04:51:24.033Z
Link: CVE-2025-38721
No data.
Status : Analyzed
Published: 2025-09-04T16:15:41.843
Modified: 2026-01-09T15:57:13.957
Link: CVE-2025-38721
OpenCVE Enrichment
Updated: 2025-09-05T14:02:34Z
Weaknesses
Debian DLA
Debian DSA
EUVD
Ubuntu USN