Issue

The test_inline_stack_discipline_idempotence test is failing because running inline_functions twice produces different results (89 bytes → 94 bytes).

Expected Behavior

Inlining should be idempotent - running it multiple times should produce the same result after the first pass.

Current Behavior

First inline: input → 89 bytes
Second inline: 89 bytes → 94 bytes

Root Cause

The inlining pass may be creating new inlining opportunities after the first pass completes.

Fix Required

Ensure inline_functions() is truly idempotent by either:

1. Running to fixed point within the function, or
2. Marking inlined functions to prevent re-inlining

Location

loom-core/tests/optimization_tests.rs:1829
loom-core/src/lib.rs:6804 (inline_functions implementation)