[jdk11u-dev] RFR: 8357105: C2: compilation fails with "assert(false) failed: empty program detected during loop optimization"

Antonio Vieiro avieiro at openjdk.org
Fri Sep 19 07:31:06 UTC 2025


On Mon, 18 Aug 2025 14:54:23 GMT, Antonio Vieiro <avieiro at openjdk.org> wrote:

> Mostly clean backport of [JDK-8357105](https://bugs.openjdk.org/browse/JDK-8357105). Backport is not clean because JDK11 C++ uses `NULL` instead of `nulllptr`.
> 
> Tested on Linux with:
> 
> tier1 tests:
> 
> ==============================
> Test summary
> ==============================
>    TEST                                              TOTAL  PASS  FAIL ERROR   
>    jtreg:test/hotspot/jtreg:tier1                     1497  1497     0     0   
>    jtreg:test/jdk:tier1                               1899  1899     0     0   
>    jtreg:test/langtools:tier1                         3941  3941     0     0   
>    jtreg:test/nashorn:tier1                              0     0     0     0   
>    jtreg:test/jaxp:tier1                                 0     0     0     0   
> ==============================
> TEST SUCCESS
> 
> 
> The test itself
> 
> ==============================
> Test summary
> ==============================
>    TEST                                              TOTAL  PASS  FAIL ERROR   
>    jtreg:test/hotspot/jtreg/compiler/stringopts/TestStackedConcatsAppendUncommonTrap.java
>                                                          1     1     0     0   
> ==============================
> TEST SUCCESS
> 
> 
> and gtests
> 
> 
> ==============================
> Test summary
> ==============================
>    TEST                                              TOTAL  PASS  FAIL ERROR   
>    gtest:all/server                                    504   504     0     0   
> ==============================
> TEST SUCCESS
> 
> 
> Reproducing this on 11 requires the additional flags `-XX:+UnlockDiagnosticVMOptions -XX:+AbortVMOnCompilationFailure `. I'm wondering if we want to add these additional flags to the new test too for 11...

This is `jdk11u-defer-next`, so let's wait.

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

PR Comment: https://git.openjdk.org/jdk11u-dev/pull/3071#issuecomment-3310982009


More information about the jdk-updates-dev mailing list