C2: loop optimization performs split_thru_phi infinitely many times

Compile::optimize_loops performs always the same optimization until it reaches the `LoopOptsCount` limit. `Major progress` is set in `PhaseIdealLoop::split_if_with_blocks` -> `PhaseIdealLoop::split_if_with_blocks_pre` after performing `PhaseIdealLoop::split_thru_phi`. This creates a chain of `phis` in the graph without any benefit for performance. Without the limit of `LoopOptsCount` this would run infinitely many times.

HOW TO REPRODUCE

Run the following command on a debug JDK build:

$ java -Xbatch -XX:-PartialPeelLoop -XX:CompileCommand=compileonly,Test::run Test.java

# A fatal error has been detected by the Java Runtime Environment:
#
# Internal Error (src/hotspot/share/opto/compile.cpp:2065), pid=7080, tid=7093
# assert(cnt++ < 40) failed: infinite cycle in loop optimization
#
# JRE version: Java(TM) SE Runtime Environment (19.0+23) (fastdebug build 19-ea+23-1706)
# Java VM: Java HotSpot(TM) 64-Bit Server VM (fastdebug 19-ea+23-1706, mixed mode, sharing, tiered, compressed oops, compressed class ptrs, g1 gc, linux-amd64)
# Problematic frame:
# V [libjvm.so+0xaba94c] Compile::Optimize()+0x180c