It was first found during investigation of the bug 6357559.
(Mustang JDK build 61 was used with different product VM).

Product and fastdebug version of C1 hangs on P4 Win2K
after the next putback 20051022115706.never.merge:
Fixed 6320351: new register allocator for c1
Fixed 4627997: JDI testcase crexreq001 failing with -client -Xcomp only, all platforms
Fixed 6311902: VM with a new c1 crashes in infinite recursion
Fixed 6180872: Add methodDataOop profiling to client compiler

C:\6357559>\jdk1.6.0_b61\bin\java -client -version
java version "1.6.0-rc"
Java(TM) 2 Runtime Environment, Standard Edition (build 1.6.0-rc-b61)
Java HotSpot(TM) Client VM (build 20051022115706.never.merge, mixed mode)

C:\6357559>\jdk1.6.0_b61\bin\java -client Tan

Previous putback is fine:

C:\6357559>\jdk1.6.0_b61\bin\java -client -version
java version "1.6.0-rc"
Java(TM) 2 Runtime Environment, Standard Edition (build 1.6.0-rc-b61)
Java HotSpot(TM) Client VM (build 20051021233148.jrose.mustang-dependencies, mix
ed mode)

C:\6357559>\jdk1.6.0_b61\bin\java -client Tan
Loose: 223ns/call -1845293.281085382
...