On Mon, 13 Jan 2025 10:11:27 GMT, Per Minborg <pminb...@openjdk.org> wrote:
>> Going forward, converting older JDK code to use the relatively new FFM API
>> requires system calls that can provide `errno` and the likes to explicitly
>> allocate a MemorySegment to capture potential error states. This can lead to
>> negative performance implications if not designed carefully and also
>> introduces unnecessary code complexity.
>>
>> Hence, this PR proposes to add a _JDK internal_ method handle adapter that
>> can be used to handle system calls with `errno`, `GetLastError`, and
>> `WSAGetLastError`.
>>
>> It currently relies on a thread-local cache of MemorySegments to allide
>> allocations. If, in the future, a more efficient thread-associated
>> allocation scheme becomes available, we could easily migrate to that one.
>>
>> Here are some benchmarks:
>>
>>
>> Benchmark Mode Cnt Score Error Units
>> CaptureStateUtilBench.explicitFail avgt 30 42.061 ? 1.261 ns/op
>> CaptureStateUtilBench.explicitSuccess avgt 30 23.142 ? 0.739 ns/op
>> CaptureStateUtilBench.tlFail avgt 30 23.580 ? 0.267 ns/op
>> CaptureStateUtilBench.tlSuccess avgt 30 1.753 ? 0.016 ns/op
>>
>>
>> Explicit allocation:
>>
>> try (var arena = Arena.ofConfined()) {
>> return (int) HANDLE.invoke(arena.allocate(4), 0, 0);
>> }
>>
>>
>> Thread Local (tl):
>>
>> return (int) ADAPTED_HANDLE.invoke(arena.allocate(4), 0, 0);
>>
>>
>> The graph below shows the difference in latency for a successful call:
>>
>> 
>>
>> This is a >10x improvement on the happy path.
>>
>>
>> Tested and passed tiers 1-3.
>
> Per Minborg has updated the pull request incrementally with two additional
> commits since the last revision:
>
> - Clean up benchmark
> - Fix allocation problem in benchmark
src/java.base/share/classes/jdk/internal/foreign/CaptureStateUtil.java line 214:
> 212: // Used reflectively by `getAsIntHandle(MemoryLayout layout)`
> 213: private static int getStateAsInt(VarHandle handle) {
> 214: if (Thread.currentThread().isVirtual() &&
> ContinuationSupport.isSupported()) {
Shouldn't we pin for the entire duration of the native call + errno retrieval?
Otherwise, can't we end up in a situation where we have a back to back native
call where the second native call overwrites the errno state of the first?
-------------
PR Review Comment: https://git.openjdk.org/jdk/pull/22391#discussion_r1913019319
