Rockwell 1718/1719-AENTR DoS: CVE-2026-9140 and the 3.012 Fix

Industrial automation cell with an engineer reviewing a control dashboard on a tablet

In short: Rockwell Automation advisory SD1778 (published 14 July 2026) describes CVE-2026-9140, a denial-of-service flaw in the 1718-AENTR and 1719-AENTR EtherNet/IP adapters for the 1718/1719 Ex I/O system. A sustained UDP unicast network storm is not handled correctly, the adapter becomes overloaded and loses communication with the controller, and only a power cycle brings it back. Firmware 3.011 is the affected release; Rockwell corrected it in 3.012. The CNA scores it 7.5 on CVSS 3.1 and 8.7 on CVSS 4.0 — both High. Rockwell lists no known exploitation and no workaround.

For maintenance and controls teams the practical problem is not the CVSS number. It is that recovery from this bug is a physical act on a device that often sits in a hazardous-area enclosure, and that switch-level “storm control” — the usual first reflex — does not catch a unicast flood. This guide covers what the advisory says, which versions are affected, how the two scores differ, and what to do if you cannot flash 3.012 this week.

What exactly did Rockwell disclose in SD1778?

SD1778 is a High-severity advisory from Rockwell’s Product Security Incident Response Team. It was released on 14 July 2026 at revision 1.0 and has not been revised since. Rockwell states the defect was found internally during routine testing, not reported by an external researcher, and that no workaround is published — the only remedy listed is the corrected firmware.

Advisory fieldValue
Advisory IDSD1778
Published / revision14 July 2026, revision 1.0
SeverityHigh
CVECVE-2026-9140 (CNA: Rockwell Automation)
WeaknessCWE-770 — Allocation of Resources Without Limits or Throttling
Affected products1718-AENTR, 1719-AENTR (1718/1719 Ex I/O EtherNet/IP Adapter)
Affected firmware3.011
Corrected in3.012
Known exploited (KEV)No
Workaround publishedNo
SourceRockwell SD1778

The CVE record was reserved on 20 May 2026 and published on 14 July 2026 at 14:52 UTC, the same day as the advisory. Its description is short and matches Rockwell’s wording: improper handling of a UDP unicast network storm causes the device to become overloaded and lose communication, and a power cycle is required to recover.

Which adapters and firmware versions are affected?

Two catalog numbers are in scope: 1718-AENTR and 1719-AENTR. Both are the EtherNet/IP adapter for the 1718/1719 Ex I/O family, the network interface that connects Ex I/O modules in hazardous-location installations to the control system. Rockwell publishes a single firmware package for both catalog numbers.

Catalog numberRoleAffected firmwareFirst fixed releaseFirmware posted now
1718-AENTREx I/O EtherNet/IP adapter3.0113.0123.013
1719-AENTREx I/O EtherNet/IP adapter3.0113.0123.013

Two details are worth reading carefully. First, the CVE record lists version 3.011 as affected and sets the default status to “unaffected”, which means Rockwell did not enumerate earlier releases in the record. If your adapters are older than 3.011, do not assume they are safe — check with Rockwell or simply move to the current release. Second, 3.012 is the minimum fixed version, not the newest one: Rockwell’s product pages currently offer firmware 3.013 for 1718_1719_AENTR (a 237 MB package). Upgrading to 3.013 satisfies the advisory.

How does a UDP unicast storm take the adapter offline?

The mechanism is resource exhaustion, not memory corruption. CWE-770 is the classification for allocating resources without limits or throttling; in practice the adapter keeps accepting and processing UDP traffic until its receive path is saturated, and EtherNet/IP communication stops. The advisory does not name the destination port, so treat any UDP service on the adapter as a candidate path.

  1. Reachability. The attacker needs a UDP path to the adapter. Base EtherNet/IP has no authentication at the transport layer, so being on the same VLAN or on a routed path with UDP permitted is enough.
  2. Flood. A sustained stream of UDP unicast frames is sent to the adapter’s address. Because it is unicast, the frames are forwarded to that one port rather than flooded to the whole segment.
  3. Overload. The adapter does not throttle or shed the load (CWE-770). Processing capacity is consumed by the flood.
  4. Loss of communication. CIP connections and I/O updates time out; the controller sees the adapter as unresponsive and the I/O modules behind it drop out of the scan.
  5. Power cycle. The device does not recover on its own. Rockwell states a power cycle is required, which means a person at the panel — not a remote reboot.

That last point has an operational consequence most DoS advisories do not: in an Ex installation, the panel may be inside a restricted area, and re-energising it may need a permit or a process shutdown. The recovery cost can be much larger than the seconds it takes to cycle power.

Why does the same bug score 7.5 and 8.7?

Because they are two different scoring systems applied to the same technical facts, not two different assessments. Rockwell is the CNA for this CVE and published both vectors. The 7.5 figure is CVSS 3.1; the 8.7 figure is CVSS 4.0. Both are “High”, and neither supersedes the other — a vulnerability manager can legitimately report either, as long as the version is stated.

ComponentCVSS 3.1CVSS 4.0What it means here
Base score7.5 High8.7 HighSame defect, different equations
Attack vectorAV:NAV:NNetwork reachable
Attack complexityAC:LAC:LNo special conditions to exploit
Attack requirements— (not in v3.1)AT:NNo additional preconditions in v4.0
Privileges requiredPR:NPR:NNo credentials
User interactionUI:NUI:NNo operator action
ConfidentialityC:NVC:N / SC:NNo data exposure
IntegrityI:NVI:N / SI:NNo data modification
AvailabilityA:HVA:H / SA:NVulnerable device fully loses availability; no downstream-system impact scored
Scope (v3.1 only)S:U— (split into VC/VI/VA and SC/SI/SA)Impact stays inside the adapter

In plain terms: CVSS 4.0 removes the single “Scope” switch and scores impact on the vulnerable system and on subsequent systems separately, and it adds an explicit “attack requirements” metric. With no privileges, no user interaction and a total loss of availability confined to the adapter, the v4.0 formula lands at 8.7 while v3.1 lands at 7.5. Use the version your vulnerability policy specifies; do not average them.

One more sourcing note: NVD has published the CVE record but marks it “Awaiting Enrichment” — NVD has not issued its own score, so the only base metrics in the record are Rockwell’s. The NVD entry also carries the CWE-770 classification and a reference to SD1778.

Is CVE-2026-9140 being exploited in the wild?

As of publication, there is no evidence of exploitation. Rockwell’s advisory lists “Known Exploited Vulnerability: No”. CISA’s ADP enrichment in the CVE record adds an SSVC decision of Exploitation: none, Automatable: yes, Technical Impact: partial, timestamped 14 July 2026. The “automatable: yes” flag is the one to respect: the attack is a traffic pattern, not a bespoke exploit, so a script or a misconfigured device can reproduce it without an exploit kit.

Absence from the KEV catalog is not a reason to defer patching, but it does change sequencing. If your Ex I/O networks are flat and reachable from IT or from the internet, treat this as urgent; if they are already segmented behind an industrial firewall with strict UDP rules, the interim exposure is much smaller.

How do you fix it? Upgrade to firmware 3.012 or later

The fix is a firmware upgrade, and Rockwell lists no configuration workaround. The sequence below is the one we recommend for a live plant.

  1. Record the current revision. Check the adapter’s module properties in Studio 5000 or the device’s web/configuration interface and write down the firmware revision and the exact catalog number (1718-AENTR or 1719-AENTR).
  2. Download the package. Rockwell’s product pages currently list “Firmware for 1718_1719_AENTR v3.013”. 3.012 or later satisfies SD1778; 3.013 is the current drop.
  3. Plan the window. Flashing an adapter interrupts I/O. In hazardous-area installations, coordinate the power cycle and any enclosure access with site permits before you start.
  4. Flash and verify. Use ControlFLASH or the Studio 5000 firmware workflow, then confirm the revision reads 3.012 or higher after the restart.
  5. Re-check the I/O. Confirm every CIP connection re-establishes, the module status is healthy, and no module behind the adapter is still faulted.
  6. Update your asset register. Log the new revision against the asset so the next audit does not flag it again.

If you cannot patch this week: interim mitigations that actually help

All of these reduce exposure. None of them removes the defect, and Rockwell does not publish a supported workaround — so treat them as a bridge to 3.012, not a substitute.

  • Segment the Ex I/O network. Keep adapters off flat IT VLANs and never expose them to the internet. A cell/area zone behind an industrial firewall is the baseline.
  • Filter and rate-limit UDP, not just broadcast. Standard “storm control” watches broadcast and multicast; a unicast flood passes straight through it. Use per-port rate limiting and ACLs that permit UDP only from the controller and engineering hosts that need it.
  • Deny unnecessary UDP services. If nothing on a segment needs discovery or a given UDP service, block it at the boundary rather than relying on the adapter to cope.
  • Watch the right signals. CIP connection timeouts, adapter status faults and switch interface counters (drops, unknown-unicast) are the observable evidence of this attack; alert on them rather than waiting for a line stop.
  • Write the recovery step down. Because recovery needs a power cycle, make sure the on-call procedure names the panel, the breaker and the permit requirement.

Which EtherNet/IP ports should you review?

SD1778 does not name the destination port used in the storm, so the safe approach is to inventory every UDP service the adapter exposes and permit only what the application needs. The three ports below cover normal EtherNet/IP operation.

PortProtocol / rolePractical rule
44818/TCPEtherNet/IP explicit messaging — configuration, HMI and engineering accessPermit only from engineering workstations and HMIs that need it; never from IT subnets by default
44818/UDPEtherNet/IP discovery (ListIdentity) and some explicit servicesUsually unnecessary across zone boundaries; block at the firewall unless a tool depends on it
2222/UDPEtherNet/IP implicit (cyclic I/O) messagingPermit only between the adapter and its scanning controller; deny everything else

If your architecture already uses CIP Security or a deny-by-default conduit model, the change is small: confirm the adapter firmware is at 3.012+ and that no “allow any UDP” rule exists for the Ex I/O zone. If it does not, this CVE is a good trigger to build one — the same rule set also blunts the next traffic-pattern bug in this family.

What this means for Ex I/O spares and maintenance planning

Three practical points for maintenance and procurement teams. First, an adapter that has been hit by this bug is not broken — it recovers on a power cycle — so do not replace hardware on a hunch; confirm the firmware revision first. Second, replacement lead times matter: Rockwell’s product page for the 1718-AENTR lists an estimated lead time of about 101 days at the time of writing, so a spare on the shelf is cheaper than an air-freight emergency. Third, any adapter you receive as a spare should be checked and, if needed, flashed to 3.012+ before it goes into a panel.

If you are refreshing an Ex I/O station, the same rules apply to the modules behind the adapter. Our I/O and function modules index covers the digital and analog modules commonly paired with these adapters, and the communication modules index covers the network side. For the wider Allen-Bradley catalogue, start at the Allen-Bradley brand page.

FAQ: CVE-2026-9140 and the 1718/1719 Ex I/O adapters

Does CVE-2026-9140 affect 1734-AENTR or 1794-AENTR?

No. SD1778 lists only 1718-AENTR and 1719-AENTR. The 1734 POINT I/O, 1794 FLEX I/O and 1756 ControlLogix families are separate product lines with their own advisories — for example, our notes on 1734-8CFG and 1794-OB8 are unrelated to this CVE. Check Rockwell’s advisory index for the specific catalog number before assuming coverage.

Can a firewall rule fix it without upgrading?

No. Filtering and rate limiting reduce the chance of the adapter being flooded, but the defect is in how the device handles the traffic that does arrive. Rockwell publishes no workaround; 3.012 is the fix.

Will the adapter recover by itself after the storm stops?

No. The advisory states that a power cycle is required to recover. Removing the traffic source is necessary but not sufficient — the adapter stays unresponsive until it is restarted.

Which firmware version is safe?

3.012 or later. Rockwell corrected the issue in 3.012 and currently posts 3.013 for the combined 1718_1719_AENTR package. Anything reporting 3.011 — or an older revision — should be upgraded.

Is authentication needed to trigger it?

No. Both CVSS vectors use PR:N and UI:N — no privileges and no user interaction. Network reachability to a UDP service on the adapter is the only precondition, which is why segmentation matters as much as patching.

Sources

Need a spare adapter while you plan the upgrade?

Autronexa supplies Allen-Bradley Ex I/O and EtherNet/IP hardware, including new-surplus and refurbished adapters and I/O modules. If you are scheduling a firmware window and want a known-good spare on the shelf first, send us the catalog number and the firmware revision you need — we will confirm condition, revision and lead time. Start with a quote request, or read how to verify an automation part number before you order if the label on the module is unclear.

Related reading: Rockwell CompactLogix 5390: 43% smaller, 128 axes, built-in security, and our series guides for the 1734 / 1746 / 1762 / 1794 families and the 1756 ControlLogix family.

Disclaimer: this article summarises a public vendor advisory and is provided for information only. It is not a substitute for Rockwell’s official documentation or your site’s change-control process. Verify firmware compatibility against your own project before flashing any device.

Leave a comment

Your email address will not be published. Required fields are marked *

Need a part? Ask us.

We reply within one business day.