The flexible spin-lock model (FSLM) unifies suspension-based and spin-based resource sharing protocols for partitioned fixed-priority preemptive scheduling based real-time multiprocessor platforms. Recent work has been done in defining the protocol for FSLM and providing a schedulability analysis without accounting for the implementation overheads. In this paper, we extend the analysis for FSLM with implementation overheads. Utilizing an initial implementation of FSLM in the OSEK/VDX-compliant Erika Enterprise RTOS on an Altera Nios II platform using 4 soft-core processors, we present an improved implementation. Given the design of the implementation, the overheads are characterized and incorporated in specific terms of the existing analysis. The paper also supplements the analysis with measurement results, enabling an analytical comparison of FSLM with the natively provided multiprocessor stack resource policy (MSRP), which may serve as a guideline for the choice of FSLM or MSRP for a specific application.