RFR: 8325681: C2 inliner rejects to inline a deeper callee because the methoddata of caller is immature. [v2]

[Vladimir Ivanov]

On Tue, 27 Feb 2024 07:04:45 GMT, Xin Liu <xliu at openjdk.org> wrote:

>> test/hotspot/jtreg/compiler/c2/irTests/TestUnderProfiledSubprocedure.java line 61:
>> 
>>> 59:     Object[] setupCondition() {
>>> 60:         // return true with ODD% possibility.
>>> 61:         return new Object[]{Boolean.valueOf(RANDOM.nextInt(100) < ODD)};
>> 
>> Why don't you produce a deterministic sequence with desired distribution instead?
>> Random input data generation may be a source of intermittent test failures. 
>> E.g., what are the chances that condition will never be true? 
>> (It's the first time I see `@Setup` usage, so I have a hard time reasoning about its behavior.)
>
> you mean I create a pattern such as 10-length array '10000 00000' and repeat it?

Yes, sort of. Take a look at [`SetupExample.setupDeterministic()`  ](https://github.com/openjdk/jdk/blob/master/test/hotspot/jtreg/testlibrary_tests/ir_framework/examples/SetupExample.java#L139).

-------------

PR Review Comment: https://git.openjdk.org/jdk/pull/17957#discussion_r1504754721