Implement splat for cfuncs. Split exit exit cases to better capture where we are exiting (#6929)

YJIT: Implement splat for cfuncs. Split exit cases

This also implements a new check for ruby2keywords as the last
argument of a splat. This does mean that we generate more code, but in
actual benchmarks where we gained speed from this (binarytrees) I
don't see any significant slow down. I did have to struggle here with
the register allocator to find code that didn't allocate too many
registers. It's a bit hard when everything is implicit. But I think I
got to the minimal amount of copying and stuff given our current
allocation strategy.

1 files changed, 3 insertions, 0 deletions

diff --git a/yjit/bindgen/src/main.rs b/yjit/bindgen/src/main.rs
index 98d9161866..229b02bdcd 100644
--- a/yjit/bindgen/src/main.rs
+++ b/yjit/bindgen/src/main.rs
@@ -227,6 +227,9 @@ fn main() {
         // From include/ruby/internal/value_type.h
         .allowlist_type("ruby_value_type") // really old C extension API
 
+        // From include/ruby/internal/hash.h
+        .allowlist_type("ruby_rhash_flags") // really old C extension API
+
         // Autogenerated into id.h
         .allowlist_type("ruby_method_ids")