CVE-2026-98277
NoneCVSS
—
EPSS
0.2%
Published
Oct 6, 2026
CISA due date
—
Cloudanix analysis
A networking bug in Linux can cause traffic to stop flowing on certain network cards when packets are dropped during bursts of activity.
- Exploitation
- Other
- Attacker needs
- Local User
- Complexity
- High
- Impact
- Denial Of Service
- Component
- OS kernel
- Class
- Race Condition
- Remediation
- Upgrade
- Effort
- Medium
- Patch risk
- Low
What it actually does
In the fbnic transmit path, a burst of packets with xmit_more set skips doorbell writes and completion requests on all but the final packet; when that final packet is dropped by skb_put_padto, skb_cow_head, or DMA mapping failure, the preceding descriptors remain invisible to hardware, leaving the ring stalled until the next transmit.
Why it matters to the business
Network transmit queues on affected systems may stall under burst-then-idle workloads, causing packet loss, reduced throughput, or temporary denial of service for network-dependent services.
How an attack unfolds
An attacker generates a burst of packets on an fbnic interface that ends with a drop condition, preventing the hardware from seeing prior descriptors and stalling the queue; the ring remains unusable until subsequent traffic or manual intervention occurs.
How to fix it
Update the Linux kernel to a version containing the fbnic fix that remembers the last undelivered meta descriptor and flushes it from error paths with the completion request set.
How to confirm the fix worked
Confirm the kernel source contains the fbnic_tx_map change that stores the last meta descriptor and forces a doorbell plus completion request from drop paths; re-test under QEMU with injected drop conditions to verify the ring no longer stalls.
Typical attack path
- Identify systems using the fbnic driver
- Send a burst of packets configured to trigger skb_put_padto, skb_cow_head, or DMA mapping failure on the final packet
- Observe that preceding descriptors are never published to hardware
- Confirm the transmit ring stalls and remains stalled under low subsequent traffic
How to detect it
Review kernel logs for fbnic transmit ring warnings, stalled queue messages, or DMA mapping failures followed by prolonged lack of completions; correlate with network interface statistics showing transmit descriptors not advancing.
What your security team should do
- Prioritize kernel updates on any hosts exposing fbnic interfaces
- Monitor network performance counters for unexpected transmit ring stalls
- Apply the fix even on non-production systems given the low but non-zero EPSS and similarity to prior bnxt issues
The record
In the Linux kernel, the following vulnerability has been resolved: eth: fbnic: ring the doorbell if a burst ends in a drop fbnic_tx_map() skips the doorbell write, and the completion request, for every packet handed to it with xmit_more set, counting on the packet which ends the burst to publish them all. When that packet is dropped instead - skb_put_padto(), skb_cow_head() or a DMA mapping failure - nothing rings. The descriptors of the preceding packets stay invisible to the HW until the next transmit on that queue, which for a burst-then-idle workload may never come. Remember the meta descriptor of the last packet left without a doorbell and flush it from the error paths. The completion request has to be set on that descriptor rather than simply writing the tail, otherwise the HW would transmit the packets but never report a head, and the ring would fill up and stall for good. This is very similar to Joe's recent series of fixes for bnxt. Not seen in real life, reproduced under QEMU with failure injection.
Affected packages
| Ecosystem | Package | Introduced | Fixed |
|---|