RFR: 8319764: C2 compilation asserts during incremental inlining because Phi input is out of bounds [v2]

[Roland Westrelin]

On Wed, 15 Nov 2023 14:58:41 GMT, Christian Hagedorn <chagedorn at openjdk.org> wrote:

> That looks reasonable to me, too!

Thanks for reviewing and making suggestions.

> src/hotspot/share/opto/replacednodes.cpp line 155:
> 
>> 153:         if (n->is_Phi()) {
>> 154:           Node* region = n->in(0);
>> 155:           if (n->req() == region->req()) { // dead phi?
> 
> Maybe you can change this comment to "ignore dead phis".

Done in new commit.

> test/hotspot/jtreg/compiler/inlining/TestLateInlineReplacedNodesExceptionPath.java line 29:
> 
>> 27:  * @summary C2 compilation asserts during incremental inlining because Phi input is out of bounds
>> 28:  * @run main/othervm -XX:-TieredCompilation -XX:-BackgroundCompilation -XX:CompileCommand=dontinline,TestLateInlineReplacedNodesExceptionPath::notInlined
>> 29:  *                   -XX:+UnlockDiagnosticVMOptions -XX:+StressIGVN -XX:StressSeed=1246687813 TestLateInlineReplacedNodesExceptionPath
> 
> Since `StressIGVN` is C2 specific, you should add a `@requires vm.compiler2.enabled`.

Done in new commit.

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

PR Comment: https://git.openjdk.org/jdk/pull/16648#issuecomment-1812834277
PR Review Comment: https://git.openjdk.org/jdk/pull/16648#discussion_r1394434168
PR Review Comment: https://git.openjdk.org/jdk/pull/16648#discussion_r1394433731