gblinear to predict best feature values

Hi, how do we derive the best feature variable values to get the highest target variable value using gblinear?

The issue mentions gblinear but names no files, tests, or entry points. First clarify whether this is a request for an optimization capability or guidance on using an existing model, then define the expected inputs, outputs, and completion criteria before locating the relevant XGBoost implementation or documentation.

Written by the indexing model from the issue text.