On Wed, 8 May 2024 19:22:31 GMT, Chris Plummer <cjplum...@openjdk.org> wrote:
> The special popFrame check needs to go in the first loop only, so it > shouldn't be a problem or add any complexity that we don't already have. Sounds good. > One things to resolve is if we enter a regular monitor while holding a leaf > monitor, is this a "rank" failure or "leaf" failure. Currently in the code it > is a "rank" failure. Your change would make it a "leaf" failure. A "leaf" failure is more specific than a "rank order" failure, so it is better to report it first. Each "leaf" failure is also a "rank order" failure (AFAICS). > I'm not sure why you added the "i != rank" check with the leaf check. We > should never be re-entering a leaf monitor. The same > "dbg_monitor->ownerThread != NULL" check as in the first loop should be used. You are right. This check is not needed and has to be removed. I was thinking a leaf monitor can be grabbed recursively. ------------- PR Review Comment: https://git.openjdk.org/jdk/pull/19044#discussion_r1594583974
