While the preemptirq patch set is built into the kernel with PREEMPT_TRACER and IRQ_TRACER enabled, we need to verify that when the tracers are inactive (not tracing), there is no instruction cache (iCache) performance degradation.

Even with similar code paths, changes in code layout, alignment, function sizes, or branch patterns could negatively affect iCache behavior and introduce performance overhead.

Task

The goal of this ticket is to create a test case that validates iCache performance when preemptirq tracers are built but inactive. The test should:

1. Build the kernel without the preemptirq patch set applied.
2. Build the kernel with the preemptirq patch set applied and PREEMPT_TRACER and IRQ_TRACER Kconfig options enabled, but ensure tracers are not actively tracing.
3. Run iCache performance benchmarks on both kernels to measure:
   • Instruction cache miss rates
   • Performance counters related to instruction fetch stalls
   • Microbenchmarks targeting hot paths in the scheduler and interrupt handling
4. Compare results to ensure no measurable iCache performance degradation when tracers are inactive.

Related Issue

This task is split from RHEL-121551 and focuses on validating runtime iCache performance characteristics.