Attention is currently required from: fixeria, laforge.
1 comment:
Patchset:
Alternatively, we may clarify the API docs and state that it's responsibility of the caller to [zero-]initialize the output buffer.
I really don't like this idea. It is nothing but a weird quirk of the particular code implementation used for FRv1 and EFR (but not for HRv1 or CSD) in that the code only sets bits to 1 but doesn't directly set them to 0. Clean interfaces between components (APIs) should be based on the good abstract design, not retro-defined around oddball quirks of a particular legacy code implementation. I struggle to think of any widely known API where the result of a function is wrong unless the caller presets the output buffer to some particular value or byte pattern.
You should also note that the unit test in the following patch fails unless the present patch is applied - that's how I caught this bug in the first place.
To view, visit change 38138. To unsubscribe, or for help writing mail filters, visit settings.