On Sat, 27 Jun 2026 17:55:03 GMT, Chris Plummer <[email protected]> wrote:
>>> What happens if the core was generated on a machine with PAC support but >>> debugged on a machine without it. Is _pac_mask going to be setup in that >>> case? >> >> `_pac_mask` would be exposed via VMStructs, so we can use it in SA when we >> attach the core. Then we can unwind call stacks on PAC-disabled machine. >> >> For unwinding, we need to restore caller PC by AND operation with >> `_pac_mask`. So it is ok to know just `_pac_mask` as a bitmask. > > But if PAC is disabled on the machine that SA is running on then _pac_mask > won't be setup. This means you won't be able to do the AND operation to > restore the caller PC. _pac_mask would be exposed via VMStructs, we can retrieve the value from CIntgerField. IIUC it means _pac_mask comes from debuggee, in the coredump when we attach it. Thus we can refer appropriate mask value, no? ------------- PR Review Comment: https://git.openjdk.org/jdk/pull/31669#discussion_r3487122306
