HIGH
net OFOQueue Exhaustion
CVE-2018-5390
CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H
KernelScan AI7.5HIGH
01Description
Linux kernel versions 4.9+ can be forced to make very expensive calls to tcp_collapse_ofo_queue() and tcp_prune_ofo_queue() for every incoming packet which can lead to a denial of service.
02KernelScan AI Analysis
Risk summary
Hosts running affected kernels with reachable TCP endpoints are at risk from remote, unauthenticated attackers sending crafted traffic. Streams of tiny, highly out-of-order packets can consume excessive CPU and kernel networking time, severely degrading or denying network service without providing data disclosure or privilege escalation.
Vulnerability analysis
The original TCP input processing repeatedly scanned and copied large numbers of tiny, completely out-of-order queued packets even though this work did not reduce memory consumption, allowing incoming packet volume to cause disproportionate processing cost. The fix stops attempting to coalesce tiny segments and abandons a pathological range once its accumulated size becomes excessive, allowing the less expensive queue-pruning behavior to handle it. A remote TCP peer with network access to an affected endpoint can trigger the condition without authentication or local privileges, causing major availability impact; no special hardware or nondefault configuration is indicated.
Lifecycle
03Fix Versions
| Branch | Introduced | Fixed in | Patch commit |
|---|---|---|---|
| mainline | — | 4.18 | 3d4bf93ac120 |
04CIP Kernels
State of the newest release of each Civil Infrastructure Platform SLTS tree, from CIP's own fix records.
| CIP tree | State |
|---|---|
| cip/4.4 | Affected — no CIP fix yet (latest 4.4.302-cip115) |
| cip/4.4-rt | Affected — no CIP fix yet (latest 4.4.302-cip115-rt63) |