RFR: 8302459: Missing late inline cleanup causes compiler/vectorapi/VectorLogicalOpIdentityTest.java IR failure [v5]

Damon Fenacci dfenacci at openjdk.org
Mon Mar 24 10:02:16 UTC 2025


On Thu, 20 Mar 2025 12:49:51 GMT, Tobias Hartmann <thartmann at openjdk.org> wrote:

>> Damon Fenacci has updated the pull request incrementally with two additional commits since the last revision:
>> 
>>  - JDK-8302459: refactor helper method
>>  - JDK-8302459: reshape infinite loop check
>
> src/hotspot/share/opto/callnode.cpp line 1117:
> 
>> 1115:       if (phase->C->print_inlining()) {
>> 1116:         phase->C->inline_printer()->record(cg->method(), cg->call_node()->jvms(), InliningResult::FAILURE,
>> 1117:           "static call node changed: trying again");
> 
> FTR, could you share how the PrintInlining output looks now when this code is triggered?

It looks like this:

@ 192   jdk.internal.vm.vector.VectorSupport::binaryOp (38 bytes)   failed to inline: failed to inline (intrinsic)   failed to inline: static call node changed: trying again   (intrinsic)   late inline succeeded


It seems a bit redundant: the first `failed to inline: failed to inline (intrinsic)` doesn't seem to be needed.

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

PR Review Comment: https://git.openjdk.org/jdk/pull/21682#discussion_r2009841029


More information about the hotspot-compiler-dev mailing list