On Fri, 8 May 2026 00:29:14 GMT, Yasumasa Suenaga <[email protected]> wrote:
> serviceability/sa/TestJhsdbJstackMixedWithXComp.java failure on CentOS 7 was > reported. > > That error is caused that glibc on CentOS 7 has Personality Routine and LSDA > (language-specific data area). SA does not support them, and skip them so > far. Thus DWARF parser could not continue to unwind call stacks. > > Personality Routine and LSDA are for exception handling for the language > which supports exception like C++ (See [LLVM > document](https://llvm.org/docs/ExceptionHandling.html) for details). They > are not needed for stack unwinding in debugger like SA, but we need to skip > DWARF augmentation correctly. > > This change has passed serviceability/sa tests on both CentOS 7 and Fedora 44. > > --------- > - [x] I confirm that I make this contribution in accordance with the [OpenJDK > Interim AI Policy](https://openjdk.org/legal/ai). Overall this change looks good. I still need to do some testing. src/jdk.hotspot.agent/linux/native/libsaproc/dwarf.cpp line 116: > 114: while (*augmentation_string != '\0') { > 115: if (*augmentation_string == 'R') { > 116: _encoding = *_buf++; This is the FDE pointer encoding, right? If so can you name it something like fde_ptr_encoding? ------------- PR Review: https://git.openjdk.org/jdk/pull/31085#pullrequestreview-4286293454 PR Review Comment: https://git.openjdk.org/jdk/pull/31085#discussion_r3238234644
