Repository navigation
Conversation
The non-MMX 8-bit loop uses bmax without initializing it, which can iterate beyond the pixel buffer and crash. Move the existing assignment before the MMX/non-MMX branch so both paths use the same initialized bound.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
The scalar 8-bit blur loop uses an uninitialized upper bound: render_blur()
assigns bmax only inside the MMX branch, although the non-MMX branch also
uses it to stop iterating over the pixel buffer. This can run past the
buffer and crash while rendering an 8-bit surface.
Move the existing assignment before the MMX/non-MMX branch so both paths
use the same initialized bound. The loop and bound calculation are
otherwise unchanged.
Validation with a headless libvisual client rendering 16 frames into a
320x240 8-bit surface, fed with synthetic stereo audio:
QEMU. GDB identifies the scalar loop at renderer.c:421.
native amd64 and sh4/QEMU.
the failing case to the 8-bit path.
This targets 0.4.x and changes only the JESS renderer.
CI status: all four Linux/macOS build-and-test jobs and the help-output
check pass. The translation-sync job stops during the core configure
step with "Invalid configuration '\'", before catalog comparison. Its
translations.yml has a trailing backslash on the final --prefix line;
that workflow is unchanged between the fork base and current 0.4.x.
No workflow or catalog files are changed by this PR.