Description
In the Linux kernel, the following vulnerability has been resolved:

ice: Restore Ordered MMIO Writes for Tx Doorbells

The DQL accounting state which tracks the number of bytes queued for the
NIC to transmit, namely dql->num_queued, is updated by the ICE driver
when it invokes __netdev_tx_sent_queue(). Subsequently the driver
updates the hardware queue tail allowing the NIC to begin processing the
work queue. The queue accounting update should be ordered before the NIC
begins transmitting descriptors.

The transmit path currently updates the hardware doorbell using
writel_relaxed(), which (on arm64) does not provide the same ordering
guarantees between writes to normal memory and writes to MMIO registers
that writel() does. This introduces a potential race where the NIC
begins transmitting descriptors before the dql->num_queued update is
globally visible. If the NIC finishes before the update is observed by
dql_completed(), it detects an invalid state where more bytes have been
completed than have been queued. When this happens the following BUG_ON
is triggered.

BUG_ON(count > num_queued - dql->num_completed);

This has been observed and manifests as the following crash (note that
the trace has been trimmed) in an environment with sustained network
load that uses an Intel Corporation Ethernet Controller E810-XXV for SFP
(rev 02) on an arm64 machine. Replacing writel_relaxed() with writel()
in a test kernel eliminated the crash in the user's workload, which
previously reproduced the issue reliably.

kernel BUG at lib/dynamic_queue_limits.c:99

Internal error: Oops - BUG: 00000000f2000800 [#1] SMP
pc : dql_completed+0x268/0x2a0
lr : ice_clean_tx_irq+0x1d4/0x620 [ice]
Call trace:
dql_completed+0x268/0x2a0 (P)
ice_napi_poll+0x94/0x520 [ice]
__napi_poll+0x48/0x3f0
net_rx_action+0x194/0x420

This restores the behaviour prior to commit ccde82e90946 ("ice: add E830
Earliest TxTime First Offload support"), which changed the notification
mechanism from writel() to writel_relaxed().
Published: 2026-10-23
Score: 7.5 High
EPSS: < 1% Very Low
KEV: No
Impact: Kernel crash and denial of service via a race condition in the ICE driver
Action: Apply Patch
AI Analysis

Impact

The ICE driver on arm64 performs an MMIO write to the NIC doorbell using writel_relaxed, which does not enforce ordering with normal memory writes. A race condition can occur when the NIC starts transmitting descriptors before the driver’s queue‑accounting variable dql->num_queued is globally visible. This triggers the internal BUG_ON and causes the kernel to panic, as observed under sustained network load on an Intel Ethernet E810‑XXV. Based on the description, it is inferred that sustained traffic is required to reliably trigger the crash.

Affected Systems

Linux kernel machines on arm64 that load the ICE driver for Intel Corporation Ethernet Controllers E810‑XXV. The flaw exists in kernel versions preceding the patch that restores writel() ordering; no specific version numbers are listed, so any kernel before that commit is potentially vulnerable.

Risk and Exploitability

The flaw is a race condition that requires the attacker to generate continuous NIC traffic; the likely attack vector is network traffic from a remote or local user capable of sending large amounts of data to the NIC. The EPSS score is below 1%, indicating a low probability of widespread exploitation, while the CVSS score of 7.5 denotes high severity. The bug can lead to a kernel‑panic denial of service, but the vulnerability is not currently listed in CISA KEV and no public exploit is documented.

Generated by OpenCVE AI on October 29, 2026 at 20:01 UTC.

Remediation

No solution or workaround provided in the CVE record.

OpenCVE Recommended Actions

  • Upgrade to a kernel version that includes the ICE driver fix which restores writel() ordering for doorbell writes.
  • If an immediate kernel upgrade is not possible, reduce NIC traffic by restarting the networking service or applying traffic shaping to avoid continuous load while the race remains present.
  • Monitor system logs such as dmesg or journalctl for BUG_ON messages related to dynamic queue limits; set alerts to detect repeated crashes promptly.

Generated by OpenCVE AI on October 29, 2026 at 20:01 UTC.

Tracking

Sign in to view the affected projects.

Advisories
Source ID Title
Ubuntu USN Ubuntu USN USN-9018-1 Linux kernel vulnerabilities
Ubuntu USN Ubuntu USN USN-9018-2 Linux kernel (HWE) vulnerabilities
Ubuntu USN Ubuntu USN USN-9019-1 Linux kernel (Azure) vulnerabilities
Ubuntu USN Ubuntu USN USN-9020-1 Linux kernel vulnerabilities
History

Thu, 29 Oct 2026 12:15:00 +0000


Sat, 24 Oct 2026 17:00:00 +0000

Type Values Removed Values Added
Weaknesses CWE-362

Sat, 24 Oct 2026 14:15:00 +0000

Type Values Removed Values Added
Metrics cvssV3_1

{'score': 7.5, 'vector': 'CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H'}


Fri, 23 Oct 2026 18:00:00 +0000

Type Values Removed Values Added
First Time appeared Linux linux
Weaknesses CWE-362
Vendors & Products Linux linux

Fri, 23 Oct 2026 12:30:00 +0000

Type Values Removed Values Added
Description In the Linux kernel, the following vulnerability has been resolved: ice: Restore Ordered MMIO Writes for Tx Doorbells The DQL accounting state which tracks the number of bytes queued for the NIC to transmit, namely dql->num_queued, is updated by the ICE driver when it invokes __netdev_tx_sent_queue(). Subsequently the driver updates the hardware queue tail allowing the NIC to begin processing the work queue. The queue accounting update should be ordered before the NIC begins transmitting descriptors. The transmit path currently updates the hardware doorbell using writel_relaxed(), which (on arm64) does not provide the same ordering guarantees between writes to normal memory and writes to MMIO registers that writel() does. This introduces a potential race where the NIC begins transmitting descriptors before the dql->num_queued update is globally visible. If the NIC finishes before the update is observed by dql_completed(), it detects an invalid state where more bytes have been completed than have been queued. When this happens the following BUG_ON is triggered. BUG_ON(count > num_queued - dql->num_completed); This has been observed and manifests as the following crash (note that the trace has been trimmed) in an environment with sustained network load that uses an Intel Corporation Ethernet Controller E810-XXV for SFP (rev 02) on an arm64 machine. Replacing writel_relaxed() with writel() in a test kernel eliminated the crash in the user's workload, which previously reproduced the issue reliably. kernel BUG at lib/dynamic_queue_limits.c:99 Internal error: Oops - BUG: 00000000f2000800 [#1] SMP pc : dql_completed+0x268/0x2a0 lr : ice_clean_tx_irq+0x1d4/0x620 [ice] Call trace: dql_completed+0x268/0x2a0 (P) ice_napi_poll+0x94/0x520 [ice] __napi_poll+0x48/0x3f0 net_rx_action+0x194/0x420 This restores the behaviour prior to commit ccde82e90946 ("ice: add E830 Earliest TxTime First Offload support"), which changed the notification mechanism from writel() to writel_relaxed().
Title ice: Restore Ordered MMIO Writes for Tx Doorbells
First Time appeared Linux
Linux linux Kernel
CPEs cpe:2.3:o:linux:linux_kernel:*:*:*:*:*:*:*:*
Vendors & Products Linux
Linux linux Kernel
References

Subscriptions

Linux Linux Linux Kernel
cve-icon MITRE

Status: PUBLISHED

Assigner: Linux

Published:

Updated: 2026-10-24T14:09:41.904Z

Reserved: 2026-09-25T10:25:14.383Z

Link: CVE-2026-98965

cve-icon Vulnrichment

No data.

cve-icon NVD

Status : Received

Published: 2026-10-23T13:17:17.950

Modified: 2026-10-24T14:16:49.640

Link: CVE-2026-98965

cve-icon Redhat

Severity : Moderate

Publid Date: 2026-10-23T00:00:00Z

Links: CVE-2026-98965 - Bugzilla

cve-icon OpenCVE Enrichment

Updated: 2026-10-29T20:15:14Z

Weaknesses