[pull] master from ruby:master#892
Merged
pull[bot] merged 23 commits intoturkdevops:masterfrom Mar 28, 2026
Merged
Conversation
When JIT-to-JIT calls hit a function stub for lazy compilation, check if the native stack is nearly exhausted before proceeding. If so, bail out to the interpreter instead of compiling and entering more JIT code. This complements the check in rb_zjit_iseq_gen_entry_point (which guards interpreter-to-JIT entry) by also guarding JIT-to-JIT entry via function stubs, matching what YJIT does in branch_stub_hit.
Add specific SendFallbackReason variants for InvokeBlock, SendForward, and InvokeSuperForward instructions that previously stayed as Uncategorized in type_specialize. Also add SingleRactorModeRequired for paths in Send handling that failed due to multi-ractor mode without setting a reason. This reduces the "uncategorized" percentage from 53.6% to 0.2% on the lobsters benchmark, making the stats output actionable.
…lize Move the InvokeBlockNotSpecialized, SendForwardNotSpecialized, and InvokeSuperForwardNotSpecialized reasons from type_specialize match arms to HIR construction, since type_specialize doesn't attempt to specialize these instructions anyway.
It appears in documentation on https://docs.ruby-lang.org/ under `Object::Ripper` ruby/prism@6c1000661c
When profiling shows a monomorphic IFUNC block handler, emit InvokeBlockIfunc which calls rb_vm_yield_with_cfunc directly instead of falling back to rb_vm_invokeblock. This avoids the interpreter dispatch overhead for the most common invokeblock case on lobsters (77% of all invokeblock calls). The implementation: - Loads the block handler from the local EP - Guards tag bits (& 0x3 == 0x3) to confirm IFUNC - Untags the pointer to get the captured block - Calls rb_vm_yield_with_cfunc(ec, captured, argc, argv)
Test Enumerable#map on a class with a Ruby `each` that uses `yield`. This exercises the IFUNC block handler path in invokeblock, where Enumerable C code creates an IFUNC wrapping the user's block.
Use IntAnd + GuardBitEquals to check (block_handler & 0x3) == 0x3, matching VM_BH_IFUNC_P() in the interpreter. Add the IntAnd HIR instruction (placed next to IntOr) for bitwise AND on raw C integers.
When rb_bug() is called (e.g. from a ZJIT Rust panic), the crash report file could be truncated and left nearly empty due to two issues: 1. The crash report file was opened with default (buffered) I/O. If rb_vm_bugreport() triggered a secondary crash, buffered output was lost. Fix by setting the file to unbuffered with setvbuf(). 2. After rb_bug() wrote the crash report, die() called abort() which triggered our custom SIGABRT handler. The handler re-opened the same crash report file with "w" mode, truncating the report that was already written. Fix by resetting SIGABRT to SIG_DFL before abort().
Co-authored-by: Alan Wu <alanwu@ruby-lang.org>
Remove redundant rb_zjit_enabled_p checks from rb_zjit_cfp_pc and rb_zjit_cfp_iseq since CFP_HAS_JIT_RETURN already checks it. Rewrite rb_zjit_cfp_has_iseq and rb_zjit_cfp_has_pc as thin wrappers around rb_zjit_cfp_iseq and rb_zjit_cfp_pc.
Add documentation to JITFrame fields explaining their purpose and nullability. Split JITFrame::new into new_iseq and new_cfunc constructors to make the two distinct use cases explicit.
Add mark() and update_references() methods to JITFrame so GC logic is encapsulated. Store *mut JITFrame in ZJITState to avoid unsafe const-to-mut casts in gc.rs.
Add rb_iseq_bare_opcode_at_pc which uses rb_vm_insn_addr2insn to return the base instruction, stripping trace/zjit variants. Use it in iseq_may_write_block_code to simplify the match arms.
The jit_return = 0 write in rb_set_cfp_pc was only needed by ZJIT. Move it to the codegen call site next to where iseq is also set, making the intent clearer.
Make zjit.h the single source of truth for the JITFrame struct layout. Remove the manual #[repr(C)] definition from jit_frame.rs and replace it with a type alias to the bindgen-generated zjit_jit_frame.
Expose rb_zjit_jit_frame_update_references so vm.c can call into Rust to relocate the iseq pointer during GC compaction, instead of manually manipulating JITFrame fields in C. This keeps JITFrame field access consolidated in Rust.
Per tenderlove's review, every cfp->iseq read should go through an accessor that accounts for JITFrame. Rename the field to _iseq so direct access causes a compile error, and rename the accessors to drop the zjit prefix (rb_cfp_iseq, rb_cfp_has_iseq, rb_cfp_pc, rb_cfp_has_pc).
The interpreter always has a valid _iseq field (it's written on exit from JIT code), so GET_ISEQ() doesn't need to go through rb_cfp_iseq(). Code in vm_insnhelper.c that may be called as a ZJIT fallback should use rb_cfp_iseq() instead.
Rename for consistency with ISEQ_BODY naming convention. Remove the separate rb_cfp_has_iseq/rb_cfp_has_pc wrappers — CFP_ISEQ and CFP_PC return NULL for missing values, which works directly in if conditions.
Return the pointer directly instead of a bool, for consistency with CFP_ISEQ and CFP_PC. The pointer is truthy in if conditions so all callers work unchanged.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
See Commits and Changes for more details.
Created by
pull[bot] (v2.0.0-alpha.4)
Can you help keep this open source service alive? 💖 Please sponsor : )