Details
Diff Detail
- Repository
- rG LLVM Github Monorepo
Event Timeline
compiler-rt/test/dfsan/reaches_function.c | ||
---|---|---|
5 | Can you make this test build and run a second time with origin tracking enabled? When origin tracking is enabled (use #ifdef), you could also call dfsan_print_origin_id_trace and add filecheck expectations for that output. | |
21–22 | It may be a better approach to:
The advantages of this would be:
|
compiler-rt/test/dfsan/reaches_function.c | ||
---|---|---|
53 | To clarify, this is triggered by the tainted value is returned from add? Should we expect the location to be here-1 because it should be the line number of the containing function, or should the line number be down at 60? | |
llvm/lib/Transforms/Instrumentation/DataFlowSanitizer.cpp | ||
1061–1065 | Should this attempt to use debug information from Instruction &I (the instruction where this occurs), rather than just using the debug loc for the containing function? |
LGTM.
Test failures look unrelated, but rebasing might make it green again.
Let me know if you'd like me to commit it for you.
Can you make this test build and run a second time with origin tracking enabled?
e.g. https://github.com/llvm/llvm-project/blob/91b38c6aaddefabad2a4c959ea3865e356761ed5/compiler-rt/test/dfsan/flush.c#L3
When origin tracking is enabled (use #ifdef), you could also call dfsan_print_origin_id_trace and add filecheck expectations for that output.