CVE-2024-26921: inet: inet_defrag: prevent sk release while still in use
In the Linux kernel, the following vulnerability has been resolved:
inet: inet_defrag: prevent sk release while still in use
ip_local_out() and other functions can pass skb->sk as function argument.
If the skb is a fragment and reassembly happens before such function call
returns, the sk must not be released.
This affects skb fragments reassembled via netfilter or similar
modules, e.g. openvswitch or ct_act.c, when run as part of tx pipeline.
Eric Dumazet made an initial analysis of this bug. Quoting Eric:
Calling ip_defrag() in output path is also implying skb_orphan(),
which is buggy because output path relies on sk not disappearing.
A relevant old patch about the issue was :
8282f27449bf ("inet: frag: Always orphan skbs inside ip_defrag()")
[..]
net/ipv4/ip_output.c depends on skb->sk being set, and probably to an
inet socket, not an arbitrary one.
If we orphan the packet in ipvlan, then downstream things like FQ
packet scheduler will not work properly.
We need to change ip_defrag() to only use skb_orphan() when really
needed, ie whenever frag_list is going to be used.
Eric suggested to stash sk in fragment queue and made an initial patch.
However there is a problem with this:
If skb is refragmented again right after, ip_do_fragment() will copy
head->sk to the new fragments, and sets up destructor to sock_wfree.
IOW, we have no choice but to fix up sk_wmem accouting to reflect the
fully reassembled skb, else wmem will underflow.
This change moves the orphan down into the core, to last possible moment.
As ip_defrag_offset is aliased with sk_buff->sk member, we must move the
offset into the FRAG_CB, else skb->sk gets clobbered.
This allows to delay the orphaning long enough to learn if the skb has
to be queued or if the skb is completing the reasm queue.
In the former case, things work as before, skb is orphaned. This is
safe because skb gets queued/stolen and won't continue past reasm engine.
In the latter case, we will steal the skb->sk reference, reattach it to
the head skb, and fix up wmem accouting when inet_frag inflates truesize.
Security readout for executives and security teams
Plain-English summary
A Linux kernel networking flaw can release a socket while transmitted IPv4 fragments still depend on it. On affected systems, a local low-privileged user may be able to trigger serious confidentiality, integrity, or availability impact. Exposure depends on kernel version and transmit-path fragment reassembly.
Executive priority
Treat this as a high-priority kernel maintenance issue, especially on multi-user, container-hosting, virtual-switching, or network-processing systems. Accelerate version verification and vendor-supported updates, but do not characterize it as an active exploitation emergency based on the supplied evidence.
Technical view
IPv4 defragmentation could orphan a fragmented packet's socket reference before output-path processing finished. Reassembly through netfilter or similar transmit-pipeline modules, including Open vSwitch and ct_act, could therefore leave downstream code using an invalid socket reference. The fix delays orphaning, preserves the socket when reassembly completes, and corrects memory accounting.
Likely exposure
Potentially exposed assets are Linux systems running affected kernels where IPv4 fragments are reassembled within transmit processing through netfilter, Open vSwitch, ct_act, or similar modules. The supplied affected-version data is unusually formatted, so confirm exact exposure against distribution advisories and applicable stable commits.
Exploitation context
The CVSS vector describes local access, low complexity, low privileges, no user interaction, and potentially high confidentiality, integrity, and availability impact. The supplied record is not in KEV, and the sources provide no evidence of active exploitation or public weaponization.
Researcher notes
The defect concerns socket lifetime and write-memory accounting during output-path fragment reassembly. Relevant code paths include ip_defrag(), skb_orphan(), ip_local_out(), and refragmentation handling. The source bundle lists multiple stable-kernel commits, but its flattened version list does not reliably express branch-specific affected and fixed boundaries.
Mitigation direction
Install the distribution-supported kernel update containing the applicable upstream stable fix.
Prioritize systems using transmit-path IPv4 reassembly through netfilter, Open vSwitch, ct_act, or similar modules.
If immediate updating is impossible, consult vendor guidance; the supplied sources name no supported temporary mitigation.
Reboot into the updated kernel where required by the operating system or vendor.
Validation and detection
Inventory running kernel versions and distribution package revisions across hosts, appliances, and network nodes.
Identify transmit paths using netfilter, Open vSwitch, ct_act, or comparable IPv4 fragment reassembly.
Match installed packages against vendor advisories or confirm inclusion of the applicable upstream stable commit.
After updating, verify the running kernel changed following reboot and repeat exposure checks.
Generated from the cited source records. This long-tail analysis has not been individually reviewed by a named human.
Potential ATT&CK relevance
Conservative CVE-to-ATT&CK context
These mappings and lookup hints may be relevant to the vulnerability behavior, CWE, affected product, or exposure path. Glexia-inferred context is not an official MITRE, ATT&CK, CWE, or CVE Program mapping.
ATT&CK lookup starting points
Use these exact CWE pages and searches to review the Glexia ATT&CK library from this CVE's weakness and description context.
cve · low confidence lookup
CVE-2024-26921 mapping review
Open the CVE-to-ATT&CK bridge for reviewed, inferred, or future official mappings tied to this CVE.
These fields come from the CVE record and ADP containers, not from Glexia's Take. They preserve time-varying source decisions such as CISA SSVC, KEV status, CVSS metrics, and provider references.
We collect every scored CVSS vector available in the official CNA and ADP containers. When more than one version is present, the table keeps the source vectors side by side instead of collapsing them into the highest score.