PIX validation fallback should inspect canonical DXIL

Open

#8,873 opened on Sep 2, 2026 0 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

Assessment

Difficulty: 3/5
Estimated time: 1-2 days
Newbie friendliness: 76/100

Issue type: Bug
Clarity: Clearly specified
Activity status: Active
Tech stack: cpp
Domain: testing-qa

Research direction

Start at the PIX test validation helper's direct validation and metadata-stripping fallback, comparing how each selects the DFCC_DXIL and DFCC_ShaderDebugInfoDXIL parts. Add the divergent-container regression test and canonical known-PIX positive control described in the issue; done means the fallback parses, clones, strips, and revalidates the canonical DFCC_DXIL module.

Written by the indexing model from the issue text.

Description

bug needs-triage

The PIX test validation helper directly validates the canonical DFCC_DXIL container part, but its metadata-stripping fallback parses DFCC_ShaderDebugInfoDXIL (ILDB). Those parts can diverge in content that survives debug-info stripping, so the fallback can classify validation using a different module from the one the validator rejected.

Reproduction

Construct a DXIL container whose canonical DFCC_DXIL and DFCC_ShaderDebugInfoDXIL parts contain different unused metadata.
Put a real non-PIX validation defect in the canonical part while the debug-info part contains only metadata permitted by the PIX test helper.
Run the helper's direct validation and metadata-stripping fallback.

Direct validation rejects the canonical module, but the fallback inspects the ILDB module. In a divergent container, this can excuse or misclassify a genuine canonical validation failure.

Proposed fix

Parse, clone, strip, and revalidate the same canonical DFCC_DXIL part targeted by direct validation. Add a focused divergent-container regression test plus a positive control for canonical known-PIX metadata.

This is a test-harness correctness issue, not a compiler behavior change. It was discovered while reviewing PR #8841 and should be handled in a separate follow-up commit rather than expanding that review-feedback amendment.

Dominant language: C++
Stars: 3.7k
Forks: 900
Avg merge: 2d 14h
Merged PRs (30d): 39

Contributor guide

Open the contributing guide

First steps

Read the whole issue, then the project's contributing guide.
Comment on the issue to say you are picking it up — it saves two people doing the same work.
Fork the repository and make your change on a branch.
Open a pull request that references the issue number.

Similar issues

[Bug] NULL pointer dereference in llm_running_callback (NGZA-AX-LLM #3) Open

Difficulty 1/5 Under an hour Newbie friendliness 90/100

AXERA-TECH/ax-llm#77 · Opened Sep 22, 2026
resume() sends a second HTTP response (400) after a successful resume Open

Difficulty 1/5 Under an hour Newbie friendliness 90/100

games-on-whales/wolf#509 · Opened Sep 21, 2026
pn7160: reader goes permanently deaf — post-read IRQ-low check fails reads that succeeded Open

Difficulty 2/5 1-3 hours Newbie friendliness 74/100

esphome/esphome#19452 · Opened Sep 21, 2026
Eval bug: Ministral-3 tool-calls during reasoning are not processed Open

bug-unconfirmed

Difficulty 2/5 1-3 hours Newbie friendliness 76/100

ggml-org/llama.cpp#29248 · Opened Sep 21, 2026
run_tests.py fails on windows, but running tests individually works Open

Difficulty 2/5 1-3 hours Newbie friendliness 74/100

NVIDIA/cuda-samples#453 · Opened Sep 21, 2026

More C++ issues