Update Oracle has shipped fixes. Only builds ending in .1 are patched (17.0.21.1 · 21.0.13.1 · 25.0.5.1 · 27.0.1.1). Check your version →
Calif claims pre-auth RCE ● Patch available

Java4Shell.
Patch is out. Are you?

A heap overflow in Java's AVX-512 Base64 decoder, reachable before login. Calif reports remote code execution on Keycloak, Jira, Confluence, Jenkins and more. Upwind reproduced the memory corruption in our lab.

9.2CVSS 4.0 (Calif)
62 BHeap write, lab-confirmed
x86 onlyAVX-512 VBMI. ARM safe
CVE-2026-114938 · Last updated Nov 17, 2026, 22:00 UTC · Upwind Research & MDR

Am I vulnerable?

Paste your Java version and pick your CPU.
Run java -version on the host or inside the container.
Fixed versions

Upgrade to a build ending in .1

Oracle's November 17 Critical Patch Update fixes the bug. The same version number without the .1 suffix is still vulnerable. Downstream builds (Temurin, Corretto, Zulu, and others) are rolling out; check your vendor.

Java lineVulnerableFixed inStatus
1717.0.6 – 17.0.2117.0.21.1Affected
21all before 21.0.13.121.0.13.1Affected
25all before 25.0.5.125.0.5.1Affected
27all before 27.0.1.127.0.1.1Affected
11all—Not affected
8OpenJDK 8See Oracle advisoryCheck Oracle
⚠️
Java 8: standard OpenJDK 8 has no AVX-512 Base64 decoder and was clean in our tests, but Oracle's advisory lists a Java SE 8 -perf build as affected. If you run Oracle Java SE 8, confirm against Oracle's page.

Sources: Oracle CPU Nov 2026 · OpenJDK advisory · Calif write-up

Proven vs. claimed

What's confirmed, and by whom

We keep our own lab results separate from what researchers report, so you know how much weight each claim carries.

🧪 Upwind lab

Reproduced on AWS EC2, Intel Ice Lake (AVX-512 VBMI)
PROVEN62-byte out-of-bounds heap write from a malformed Base64 string
PROVENAdjacent heap object corrupted, then JVM crash during garbage collection (remote DoS)
PROVENCorrupted array length turned into heap read/write primitives
PROVEN-XX:-UseBASE64Intrinsics blocks the bug on every affected version

📣 Calif (discoverer)

Reported, not yet reproduced by Upwind. Exploit chains not published.
CLAIMEDRemote code execution in default configs.
KeycloakJiraConfluenceArtifactoryJenkinsPingFederate
KafkaSolrElasticsearchGerritSpinnaker
■ pre-auth   ■ post-auth
CLAIMEDClient-side path: a Java app connecting to a malicious TLS server
Do this now

Four steps for security teams

01

Inventory

Every Java 17/21/25/27 runtime, including JREs bundled in containers and vendor products like Keycloak, Jira or Jenkins.

02

Prioritize

Internet-facing login and API services on AVX-512 x86 hosts first. They decode Base64 from strangers by design.

03

Patch

Move to the .1 builds. If you can't yet, apply the stopgap:

-XX:-UseBASE64Intrinsics
04

Watch

Hunt for malformed Base64, decoder exceptions and JVM crash loops. See detection below.

⚠️ Broken advice is going around Some posts recommend this flag on its own. It stops the JVM from starting. Pushed fleet-wide via JAVA_TOOL_OPTIONS, it causes an outage:
-XX:DisableIntrinsic=_base64_decodeBlock
Error: VM option 'DisableIntrinsic' is diagnostic and must be enabled via -XX:+UnlockDiagnosticVMOptions.
Use -XX:-UseBASE64Intrinsics instead. -XX:UseAVX=2 also works, but only on x86-64: on ARM the JVM won't start with it.
Detect attempts

What exploitation looks like

Three signals from our lab, strongest first. Legitimate padded Base64 always has a length that is a multiple of 4, so the first rule is low-noise.

High signal

Malformed Base64 in requests

A token in headers or body that contains = but whose length isn't a multiple of 4. Strip the auth scheme first (Basic, Bearer…). Flag anything that doesn't match:

^(?:[A-Za-z0-9+/_-]{4})*(?:[A-Za-z0-9+/_-]{2}==|[A-Za-z0-9+/_-]{3}=)?$
High signal

Decoder exceptions in app logs

Fired on every trigger attempt in our tests. It's thrown after the overflow, so catching it doesn't prevent the corruption.

IllegalArgumentException: Input byte array has wrong 4-byte ending unit
Supporting

JVM crashes & respawn bursts

SIGSEGV in the garbage collector, new hs_err_pid*.log files, and services restarting in a loop. Don't rely on this alone: crashes depend on heap layout.

SIGSEGV … G1ParScanThreadState::copy_to_survivor_space
Timeline

How it unfolded

Details leak ahead of Oracle's patch. CVE-2026-114938 reserved.

Upwind reproduces the 62-byte heap overflow on Ice Lake hardware; confirms 17.0.6+, 21, 25 affected; 11 and ARM clean.

Calif publishes its write-up and RCE claims across 11 products. Exploit chains held back.

Patch available: 17.0.21.1, 21.0.13.1, 25.0.5.1, 27.0.1.1. Downstream builds rolling out.

We'll update this page as vendors ship builds and as RCE claims are confirmed or refuted.

Upwind customers

Find exposed Java in minutes

Upwind shows which runtimes run an affected JDK, on which CPU, and whether they're reachable from the internet, with runtime detection for post-exploitation behavior.