In StubGenerator::generate_base64_decodeBlock() (x86_64 AVX-512 VBMI path), L_finalBit handles 1-63 leftover bytes with a masked vector sequence that computes output_size = (len/4)*3. Valid inputs always have a length multiple of 4, so this bug triggers only on invalid inputs: with length < 4 (e.g. "A".repeat(64)+"=" leaves 1 byte after L_process64Loop), output_size starts at 0 and the padding path decrements it below zero, producing a wrong output_mask. The stub stores past the logical output length before returning to Java, corrupting the heap.
The suggested fix is to early-return to Java if length < 4.
- I confirm that I make this contribution in accordance with the OpenJDK Interim AI Policy.
Progress
- Change must not contain extraneous whitespace
- Commit message must refer to an issue
- Change must be properly reviewed (2 reviews required, with at least 1 Reviewer, 1 Author)
Integration blocker
8392044. Please make sure it exists and is accessible.
Issue
⚠️ Failed to retrieve information on issue8392044.
Reviewing
Using git
Checkout this PR locally:
$ git fetch https://git.openjdk.org/jdk.git pull/32781/head:pull/32781
$ git checkout pull/32781
Update a local copy of the PR:
$ git checkout pull/32781
$ git pull https://git.openjdk.org/jdk.git pull/32781/head
Using Skara CLI tools
Checkout this PR locally:
$ git pr checkout 32781
View PR using the GUI difftool:
$ git pr show -t 32781
Using diff file
Download this PR as a diff file:
https://git.openjdk.org/jdk/pull/32781.diff