On Tue, 13 Jun 2023 09:15:13 GMT, Chen Liang <li...@openjdk.org> wrote:
>> Adam Sotona has updated the pull request incrementally with one additional >> commit since the last revision: >> >> ClassHierarchyResolver.DEFAULT_RESOLVER gets info from system class loader >> and does not cache >> method-level caching is enabled if the provided resolver is not already >> cached > > src/java.base/share/classes/jdk/internal/classfile/impl/DirectCodeBuilder.java > line 334: > >> 332: maxLocals = original.maxLocals(); >> 333: maxStack = original.maxStack(); >> 334: stackMapAttr = >> original.findAttribute(Attributes.STACK_MAP_TABLE).orElse(null); > > On the baseline, if a stack map table attribute can be reused but this > findAttribute doesn't find an attribute (implying an implicit stack map, > potentially for simple code?), we will not perform a stack map generation and > simply reuse a trivial stack map (null), which should be a correct algorithm. Yes, I caused this regression by "Revert of DirectCodeBuilder.needsStackMap pre-calculation" commit above. The needStackMap pre-calculation on the other side caused regression in lambda performance, so one regression has been replaced with another. ------------- PR Review Comment: https://git.openjdk.org/jdk/pull/14180#discussion_r1228001362
