RFR: JDK-8326376: java -version failed with CONF=fastdebug -XX:InitialCodeCacheSize=1024K -XX:ReservedCodeCacheSize=1200k
SUN Guoyun
duke at openjdk.org
Thu Feb 22 01:57:56 UTC 2024
On Wed, 21 Feb 2024 08:16:27 GMT, Thomas Stuefe <stuefe at openjdk.org> wrote:
> Please review this fix of a regression brought by https://bugs.openjdk.org/browse/JDK-8318444. Thanks to @sunny868 for reporting this.
>
> We bail out during compiler initialization of C2. The thread's ciEnv does not have a CompileTask associated in that case, and that's why we are faulting.
>
> I tested the patch on Linux x64 and confirmed the fix. I also tested with a compiler memory ceiling and the crash option (see https://bugs.openjdk.org/browse/JDK-8318016) and confirmed that the hs-err file contains valid compile ids when we have a valid compile task.
loongarch64 used a larger CodeCache, so `compiler/codecache/CodeCacheFullCountTest.java` exposed this issue under default parameters..
-------------
PR Comment: https://git.openjdk.org/jdk/pull/17941#issuecomment-1958519885
More information about the hotspot-compiler-dev
mailing list