On Wed, 28 Oct 2026 11:38:04 GMT, Matthias Baesken <[email protected]> wrote:
>> src/java.base/share/native/libverify/check_code.c line 452: >> >>> 450: ((type *)CCalloc(context, (count)*(sizeof(type)), JNI_FALSE)) >>> 451: #define ZNEW(type, count) \ >>> 452: ((type *)CCalloc(context, (count)*(sizeof(type)), JNI_TRUE)) >> >> Seems like `jboolean zero` parameter for `CCalloc` is no longer needed after >> this removal. > > Yeah seems so. But the question now is, should we now completely remove the > zeroing (memset 0) in CCalloc with the boolean parameter ? Yes, I'd prefer to clean it up. >> src/jdk.hotspot.agent/linux/native/libsaproc/ps_proc.c line 157: >> >>> 155: } >>> 156: return true; >>> 157: #elif defined(PTRACE_GETREGS_REQ) >> >> Do we still need this old code? `PTRACE_GETREGSET` exists since Linux >> 2.6.34. Shouldn't it always be defined? Even on Alpine? > > Good question, probably we can remove it. But I am not 100% sure when it > became available in user-space headers and if there are differences between > CPU-architectures. > Should I remove the fallback and always count on PTRACE_GETREGSET ? I suggest to try and run through Oracle's and our testing. ------------- PR Review Comment: https://git.openjdk.org/jdk/pull/33468#discussion_r4400938909 PR Review Comment: https://git.openjdk.org/jdk/pull/33468#discussion_r4400945965
