Incident sheet · /security

Filed
Read
5 min · 1,007 words
Severity
Routine

Java4Shell Flaw Scores 9.2, Exposes Heap Memory Across Enterprise Estates

Java4Shell (CVE-2026-114938) scores 9.2 in CVSS 4.0 and lets unauthenticated attackers read heap memory and run code. Patching depends on CPU silicon and vendor distribution.

A new critical Java hole illustrates the enterprise IT visibility problem
FeedA new critical Java hole illustrates the enterprise IT visibility problem — AI-generated

Event log

  1. Java4Shell, CVE-2026-114938, scores 9.2 (Critical) under CVSS 4.0; Oracle rated it 8.1 under CVSS 3.1.
  2. The flaw in the OpenJDK x86-64 Base64 decoder can be triggered by 64 'A' characters plus an equals sign (128 on older OpenJDK 17).
  3. Exploitation requires x86-64 processors supporting AVX-512 VBMI instructions.
  4. Patching requires fixes from multiple vendors including Oracle, Amazon, Eclipse, Microsoft, Red Hat, and Azul.
  5. Client-side compromise is possible via connection to a malicious TLS server with no exposed service.

By Rebecca Stone

A critical Java vulnerability scored 9.2 under CVSS 4.0 allows unauthenticated attackers to read heap memory, steal private keys and tokens, and reach remote code execution. Research firm Calif published details of the flaw, dubbed Java4Shell and tracked as CVE-2026-114938, on Tuesday. Oracle rated the same bug 8.1 (High) under CVSS 3.1.

The bug sits in the OpenJDK x86-64 Base64 decoder. Base64 is a common method of representing binary data as text. Calif said it is used in authentication protocols, file upload encoding, and network message delivery. "To decode Base64 faster, OpenJDK uses specialized x86-64 CPU instructions to process many bytes at once," Calif wrote in its advisory. "We discovered a bug in optimized machine-level code that lets malformed input overwrite neighboring memory. With careful exploitation, the corruption can reveal sensitive memory, as Heartbleed did, or lead to remote code execution."

Triggering the corruption requires minimal input. Calif said a string of 64 "A" characters followed by a single equals sign suffices. Older OpenJDK 17 versions require 128 "A" characters before the equals sign.

Why does a memory bug threaten Java specifically?

Java is known for memory safety. The runtime manages memory on developers' behalf and checks memory accesses. That reputation makes this flaw unusual and dangerous.

"Memory corruption is the most common class of vulnerability on the internet," Calif said. "But one of Java's biggest selling points is memory safety. But like PHP and many other technologies, much of Java itself is written in low-level languages like C, C++, and, in this case, assembly, where memory corruption is still a part of life."

Exploitability is conditional. The flaw requires a hardware-accelerated decoding path present only on x86-64 processors supporting AVX-512 VBMI instructions. Patching requires fixes from a lengthy list of vendors, not just Oracle's Tuesday patch for its own Java builds.

Noah Kenney, principal consultant at Digital 520, called the potential damage from an unpatched, exposed Java instance extreme. "The researchers were able to read Java heap memory before they achieved code execution, so an attacker may be able to steal tokens, credentials, or keys from those systems," he said. "That access can remain useful even after the vulnerable server has been patched. Any unexplained crashes on an exposed system should lead to a review of what secrets were in memory and whether they need to be rotated."

How does CPU hardware change exposure?

Not every Java instance is vulnerable. Hardware determines whether the faulty decoder runs at all.

"The same Java application and version can be vulnerable on a newer x86 server and unaffected on older hardware because the faulty decoder only runs when AVX-512 is available," Kenney said. "That makes a normal software inventory less useful than it usually is." He added that installing a newer processor can expose an application that was safe on the server it replaced, with no application change.

Enterprises also run Java from many distributions. "Oracle released a fix for its Java builds, but enterprises also run Java from Amazon, Eclipse, Microsoft, Red Hat, and Azul, along with runtimes buried inside commercial products and containers," Kenney said. He advised CISOs to start with identity systems and software-delivery infrastructure, identify the runtime each uses, apply the appropriate vendor fix, and restart or rebuild the workload.

Outbound connections also demand evaluation. Kenney noted that researchers demonstrated a Java program could be compromised by connecting to a malicious TLS server even with no service exposed to the internet. Aman Mahapatra, chief strategy officer at New York-based Tribeca Softech, said the client-side exploit puts any Java process that connects to a malicious TLS server at risk, including integrations, connectors, crawlers, and batch jobs.

Can SBOMs detect the real exposure?

Software bills of materials fall short, Mahapatra said. "Your SBOM can't see the real exposure. Whether a service is vulnerable depends on the silicon underneath as much as the JDK version," he said. "The same container image is exploitable on an Ice Lake or Zen 4 node and safe on ARM or older hardware, and autoscaling can move it between the two. For once, the organizations that refreshed their hardware are the exposed ones."

Raj Annaladasu, staff customer success engineer at Palo Alto Networks, said the flaw exposes a blind spot in how enterprises identify, prioritize, and validate vulnerabilities. "Java4Shell is a reminder that vulnerabilities can originate below the application layer, inside the managed runtime itself," he said. He explained that exploitability depends on an affected runtime, the relevant CPU capabilities, the vulnerable execution path being enabled, and attacker-controlled input reaching the decoder.

Runtimes hide in third-party products, container images, appliances, custom distributions, and application-specific directories, Annaladasu said. A clean dependency scan or CMDB does not establish the environment is free of vulnerable Java runtimes.

What happens if one server is missed?

Frank Dickson, principal analyst at Dickson Research, said the odds favor attackers. "Java4Shell is a big deal because a single missed server with Java is all an attacker needs, allowing someone who never logs in to read its memory, including personal data, private keys, and tokens," he said. "If that server is an identity system like Keycloak or PingFederate, the attacker now owns you." He added: "Will a CISO miss one? The odds say yes. Java hides everywhere. Each Java instance needs its own fix."

Brian Levine, a partner at Control Risks, said Java's popularity is its greatest weakness. "Java runs the back office of corporate America: logins, code pipelines, ticketing, internal wikis," he said. "Java4Shell lets an attacker who hasn't even logged in take over some of the most sensitive of those systems by sending a short string of text." He cited Keycloak, PingFederate, Jenkins, Artifactory, Jira, and Confluence as high-value targets. He noted one mitigation: the bug only fires on certain newer Intel and AMD chips. The catch, he said, is that those chips fill modern data centers and cloud fleets.

via infoworld.com/feed/ (Source)

Filed under

  • java4shell
  • cve-114938
  • openjdk
  • memory-corruption
  • vulnerability-management
Share this article:

More from Rebecca Stone

Rebecca Stone

Show full bio

Senior reporter covering money, funding and market shifts around enterprise information technology and the leaders who run at CIO Monitor.

199 articles

The newsroom digest

One email a day with the stories that matter. No noise.

By subscribing you agree to the privacy policy.

Cross-references

‹ Previous articleNext article ›