Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <20261119075006.GA2914@openwall.com>
Date: Thu, 19 Nov 2026 08:50:06 +0100
From: Solar Designer <solar@...nwall.com>
To: oss-security@...ts.openwall.com
Subject: Re: CVE-2026-114938: OpenJDK pre-auth RCE via memory-corruption in Base64 decoder

On Tue, Nov 17, 2026 at 04:53:52PM -0500, Jan Schaumann wrote:
> Today's internet hyped vulnerability appears to be
> CVE-2026-114938, obligatorily named
> "<something>[2|4]shell".
> 
> In this case, it's Java, specifically a
> memory-corruption vulnerability in OpenJDK's Base64
> decoder on x86-64 processors with AVX-512 VBMI
> support leading to memory disclosure and RCE.
> 
> 
> Calif.io set their priorities on publicity-maxxing, it
> seems:
> 
> "After we disclosed PHP2Shell, we realized many people
> didn't know what PHP was. So we decided to go after
> something a little more familiar: Java."
> 
> Anyway, full details here:
> https://calif.io/java4shell/
> 
> OpenJDK advisory (without any technical details) here:
> https://openjdk.org/groups/vulnerability/advisories/2026-11-17
> 
> Oracle JDK release notes here:
> https://www.oracle.com/security-alerts/cspunov2026.html
> 
> I haven't yet seen a link to the actual source code
> fixes.

The fix merged just in time for this disclosure (but prepared earlier):

https://github.com/openjdk/jdk/commit/5177a70b385c16f26ede191fdac7aaba0bed0f8b

>From 5177a70b385c16f26ede191fdac7aaba0bed0f8b Mon Sep 17 00:00:00 2001
From: Tobias Hartmann <thartmann@...njdk.org>
Date: Tue, 13 Oct 2026 09:27:30 +0000
Subject: [PATCH] 8393212: Base64 optimization improvements

Co-authored-by: Albert Mingkun Yang <ayang@...njdk.org>
Reviewed-by: rhalade, mschoene, chagedorn, rcastanedalo
---
 src/hotspot/cpu/x86/stubGenerator_x86_64.cpp | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/src/hotspot/cpu/x86/stubGenerator_x86_64.cpp b/src/hotspot/cpu/x86/stubGenerator_x86_64.cpp
index d6c5ca188b34..d14929b77735 100644
--- a/src/hotspot/cpu/x86/stubGenerator_x86_64.cpp
+++ b/src/hotspot/cpu/x86/stubGenerator_x86_64.cpp
@@ -3395,6 +3395,10 @@ address StubGenerator::generate_base64_decodeBlock() {
     __ BIND(L_finalBit);
     // Now have 1 to 63 bytes left to decode
 
+    // A padded Base64 tail shorter than one complete block is handled by Java.
+    __ cmpl(length, 4);
+    __ jcc(Assembler::less, L_bruteForce);
+
     // I was going to let Java take care of the final fragment
     // however it will repeatedly call this routine for every 4 bytes
     // of input data, so handle the rest here.

PR with the same fix submitted and closed without merging 3 months earlier:

https://github.com/openjdk/jdk/pull/32781

Third-party "CVE-2026-114938 Runtime Patch (Temporary Patch Provided by Xint)":

https://github.com/theori-io/java4shell-runtime-patch-CVE-2026-114938

Writeup on independent rediscovery of the same issue:

https://www.chainguard.dev/unchained/java4shell-how-a-malformed-base64-string-breaks-the-jvm-on-avx-512

about this rediscovery Dan Lorenc writes on Twitter: "I double checked
mine after and it was found through fuzzing. I only saw that issue when
cleaning things up after."

Alexander

Powered by blists - more mailing lists

Please check out the Open Source Software Security Wiki, which is counterpart to this mailing list.

Confused about mailing lists and their use? Read about mailing lists on Wikipedia and check out these guidelines on proper formatting of your messages.