RFR: 8342902: Deduplication of acquire calls in BindingSpecializer causes escape-analyisis failure

[Maurizio Cimadamore]

On Fri, 25 Oct 2024 08:37:30 GMT, Maurizio Cimadamore <mcimadamore at openjdk.org> wrote:

> This PR fixes an issue where passing many by-reference parameters to downcall results in escape analysis failures.
> The problem is that, as the parameters grow, the generated code in the trampoline stub we generate also grows.
> When it reaches a certain threshold, it becomes too big, and it is no longer inlined in the caller.
> When this happens, allocations for by-reference parameters (e.g. a segment constructed from `MemorySegment::ofAddress`) can no longer be eliminated.
> 
> The solution is two-fold. First, we annotate the generated trampoline with `@ForceInline`. After all, it is rather critical, to guarantee optimal performance, that this code can be always inlined.
> Second, to keep the size of the generated code under control, we also put a limit on the max number of comparisons that are generated in order to "deduplicate" scope acquire/release calls.
> The deduplication logic is a bit finicky -- it was put in place because, when confined and shared are passed by-ref, we need to prevent them from being closed in the middle of a native call.
> So, we save all the seen scopes in a bunch of locals, and then we compare each new scope with _all_ the previous cached locals, and skip acquire if we can.
> 
> While this strategy work it does not scale when there are many by-ref parameters - as a by-ref parameter N will need N - 1 comparisons - which means a quadratic number of comparisons is generated.
> This is fixed in this PR by putting a lid on the maximum number of comparisons that are generated. We also make the comparisons a bit smarter, by always skipping the very first by-ref argument -- the downcall address.
> It is in fact very common for the downcall address to be in a different scope than that of the other by-ref arguments anyway.
> 
> A nice property of the new logic is that by configuring the max number of comparisons we can effectively select between different strategies:
> * max = 0, means no dedup
> * max = 1, means one-element cache
> * max = N, means full dedup (like before)
> 
> Thanks to Ioannis (@spasi) for the report and the benchmark. I've beefed the benchmark up by adding a case for 10 arguments, and also adding support for critical downcalls, so we could also test passing by-ref heap segments. Benchmark result will be provided in a separate comment.

src/java.base/share/classes/jdk/internal/foreign/abi/BindingSpecializer.java line 85:

> 83:         = GetBooleanAction.privilegedGetProperty("jdk.internal.foreign.abi.Specializer.PERFORM_VERIFICATION");
> 84:     private static final int SCOPE_DEDUP_DEPTH
> 85:             = GetIntegerAction.privilegedGetProperty("jdk.internal.foreign.abi.Specializer.SCOPE_DEDUP_DEPTH", 2);

A two-comparison logic seems like a good compromise. Note this doesn't really mean a two-element cache - it is still possible to have e.g. the first two arguments with the same scope and then a third argumemt with another scope - in which case the third argument won't be deduped. This is a trade-off - to keep the generated code simple we need to have a 1-1 mapping between by-ref parameters and locals.

-------------

PR Review Comment: https://git.openjdk.org/jdk/pull/21706#discussion_r1816268042