ZJIT: Create delta debugging script to narrow JIT failures (ruby#14041)

Add support for `--zjit-allowed-iseqs=SomeFile` and
`--zjit-log-compiled-iseqs=SomeFile` so we can restrict and inspect
which ISEQs get compiled.

Then add `jit_bisect.rb` which we can run to try and narrow a failing
script. For example:

    plum% ../tool/zjit_bisect.rb ../build-dev/miniruby "test.rb"
    I, [2025-07-29T12:41:18.657177 #96899]  INFO -- : Starting with JIT list of 4 items.
    I, [2025-07-29T12:41:18.657229 #96899]  INFO -- : Verifying items
    I, [2025-07-29T12:41:18.726213 #96899]  INFO -- : step fixed[0] and items[4]
    I, [2025-07-29T12:41:18.726246 #96899]  INFO -- : 4 candidates
    I, [2025-07-29T12:41:18.797212 #96899]  INFO -- : 2 candidates
    Reduced JIT list:
    [email protected]:8
    plum%

We start with 4 compiled functions and shrink to just one.

Author: tekknolagi

tool/zjit_bisect.rb

@@ -0,0 +1,97 @@
+#!/usr/bin/env ruby
+require 'logger'
+require 'open3'
+require 'tempfile'
+require 'timeout'
+
+RUBY = ARGV[0] || raise("Usage: ruby jit_bisect.rb <path_to_ruby> <options>")
+OPTIONS = ARGV[1] || raise("Usage: ruby jit_bisect.rb <path_to_ruby> <options>")
+TIMEOUT_SEC = 5
+LOGGER = Logger.new($stdout)
+
+# From https://github.com/tekknolagi/omegastar
+# MIT License
+# Copyright (c) 2024 Maxwell Bernstein and Meta Platforms
+# Attempt to reduce the `items` argument as much as possible, returning the
+# shorter version. `fixed` will always be used as part of the items when
+# running `command`.
+# `command` should return True if the command succeeded (the failure did not
+# reproduce) and False if the command failed (the failure reproduced).
+def bisect_impl(command, fixed, items, indent="")
+  LOGGER.info("#{indent}step fixed[#{fixed.length}] and items[#{items.length}]")
+  while items.length > 1
+    LOGGER.info("#{indent}#{fixed.length + items.length} candidates")
+    # Return two halves of the given list. For odd-length lists, the second
+    # half will be larger.
+    half = items.length / 2
+    left = items[0...half]
+    right = items[half..]
+    if !command.call(fixed + left)
+      items = left
+      next
+    end
+    if !command.call(fixed + right)
+      items = right
+      next
+    end
+    # We need something from both halves to trigger the failure. Try
+    # holding each half fixed and bisecting the other half to reduce the
+    # candidates.
+    new_right = bisect_impl(command, fixed + left, right, indent + "< ")
+    new_left = bisect_impl(command, fixed + new_right, left, indent + "> ")
+    return new_left + new_right
+  end
+  items
+end
+
+# From https://github.com/tekknolagi/omegastar
+# MIT License
+# Copyright (c) 2024 Maxwell Bernstein and Meta Platforms
+def run_bisect(command, items)
+  LOGGER.info("Verifying items")
+  if command.call(items)
+    raise StandardError.new("Command succeeded with full items")
+  end
+  if !command.call([])
+    raise StandardError.new("Command failed with empty items")
+  end
+  bisect_impl(command, [], items)
+end
+
+def run_with_jit_list(ruby, options, jit_list)
+  # Make a new temporary file containing the JIT list
+  Tempfile.create("jit_list") do |temp_file|
+    temp_file.write(jit_list.join("\n"))
+    temp_file.flush
+    temp_file.close
+    # Run the JIT with the temporary file
+    Open3.capture3("#{ruby} --zjit-allowed-iseqs=#{temp_file.path} #{options}")
+  end
+end
+
+# Try running with no JIT list to get a stable baseline
+_, stderr, status = run_with_jit_list(RUBY, OPTIONS, [])
+if !status.success?
+  raise "Command failed with empty JIT list: #{stderr}"
+end
+# Collect the JIT list from the failing Ruby process
+jit_list = nil
+Tempfile.create "jit_list" do |temp_file|
+  Open3.capture3("#{RUBY} --zjit-log-compiled-iseqs=#{temp_file.path} #{OPTIONS}")
+  jit_list = File.readlines(temp_file.path).map(&:strip).reject(&:empty?)
+end
+LOGGER.info("Starting with JIT list of #{jit_list.length} items.")
+# Now narrow it down
+command = lambda do |items|
+  status = Timeout.timeout(TIMEOUT_SEC) do
+    _, _, status = run_with_jit_list(RUBY, OPTIONS, items)
+    status
+  end
+  status.success?
+end
+result = run_bisect(command, jit_list)
+File.open("jitlist.txt", "w") do |file|
+  file.puts(result)
+end
+puts "Reduced JIT list (available in jitlist.txt):"
+puts result

zjit/src/codegen.rs

@@ -284,6 +284,10 @@ fn gen_function(cb: &mut CodeBlock, iseq: IseqPtr, function: &Function) -> Optio
             let iseq_name = iseq_get_location(iseq, 0);
             register_with_perf(iseq_name, start_usize, code_size);
         }
+        if ZJITState::should_log_compiled_iseqs() {
+            let iseq_name = iseq_get_location(iseq, 0);
+            ZJITState::log_compile(iseq_name);
+        }
     }
     result
 }

zjit/src/hir.rs

@@ -2480,6 +2480,7 @@ pub enum ParseError {
     UnknownParameterType(ParameterType),
     MalformedIseq(u32), // insn_idx into iseq_encoded
     Validation(ValidationError),
+    NotAllowed,
 }
 
 /// Return the number of locals in the current ISEQ (includes parameters)
@@ -2545,6 +2546,9 @@ fn filter_unknown_parameter_type(iseq: *const rb_iseq_t) -> Result<(), ParseErro
 
 /// Compile ISEQ into High-level IR
 pub fn iseq_to_hir(iseq: *const rb_iseq_t) -> Result<Function, ParseError> {
+    if !ZJITState::can_compile_iseq(iseq) {
+        return Err(ParseError::NotAllowed);
+    }
     filter_unknown_parameter_type(iseq)?;
     let payload = get_or_create_iseq_payload(iseq);
     let mut profiles = ProfileOracle::new(payload);

zjit/src/options.rs

@@ -1,6 +1,7 @@
 use std::{ffi::{CStr, CString}, ptr::null};
 use std::os::raw::{c_char, c_int, c_uint};
 use crate::cruby::*;
+use std::collections::HashSet;
 
 /// Number of calls to start profiling YARV instructions.
 /// They are profiled `rb_zjit_call_threshold - rb_zjit_profile_threshold` times,
@@ -19,7 +20,7 @@ pub static mut rb_zjit_call_threshold: u64 = 2;
 #[allow(non_upper_case_globals)]
 static mut zjit_stats_enabled_p: bool = false;
 
-#[derive(Clone, Copy, Debug)]
+#[derive(Clone, Debug)]
 pub struct Options {
     /// Number of times YARV instructions should be profiled.
     pub num_profiles: u8,
@@ -44,6 +45,12 @@ pub struct Options {
 
     /// Dump code map to /tmp for performance profilers.
     pub perf: bool,
+
+    /// List of ISEQs that can be compiled, identified by their iseq_get_location()
+    pub allowed_iseqs: Option<HashSet<String>>,
+
+    /// Path to a file where compiled ISEQs will be saved.
+    pub log_compiled_iseqs: Option<String>,
 }
 
 /// Return an Options with default values
@@ -57,6 +64,8 @@ pub fn init_options() -> Options {
         dump_lir: false,
         dump_disasm: false,
         perf: false,
+        allowed_iseqs: None,
+        log_compiled_iseqs: None,
     }
 }
 
@@ -67,6 +76,8 @@ pub const ZJIT_OPTIONS: &'static [(&str, &str)] = &[
     ("--zjit-num-profiles=num",   "Number of profiled calls before JIT (default: 1, max: 255)."),
     ("--zjit-stats",              "Enable collecting ZJIT statistics."),
     ("--zjit-perf",               "Dump ISEQ symbols into /tmp/perf-{}.map for Linux perf."),
+    ("--zjit-log-compiled-iseqs=path",
+                     "Log compiled ISEQs to the file. The file will be truncated."),
 ];
 
 #[derive(Clone, Copy, Debug)]
@@ -108,6 +119,26 @@ pub extern "C" fn rb_zjit_parse_option(options: *const u8, str_ptr: *const c_cha
     parse_option(options, str_ptr).is_some()
 }
 
+fn parse_jit_list(path_like: &str) -> HashSet<String> {
+    // Read lines from the file
+    let mut result = HashSet::new();
+    if let Ok(lines) = std::fs::read_to_string(path_like) {
+        for line in lines.lines() {
+            let trimmed = line.trim();
+            if !trimmed.is_empty() {
+                result.insert(trimmed.to_string());
+            }
+        }
+    } else {
+        eprintln!("Failed to read JIT list from '{}'", path_like);
+    }
+    eprintln!("JIT list:");
+    for item in &result {
+        eprintln!("  {}", item);
+    }
+    result
+}
+
 /// Expected to receive what comes after the third dash in "--zjit-*".
 /// Empty string means user passed only "--zjit". C code rejects when
 /// they pass exact "--zjit-".
@@ -165,6 +196,19 @@ fn parse_option(options: &mut Options, str_ptr: *const std::os::raw::c_char) ->
 
         ("perf", "") => options.perf = true,
 
+        ("allowed-iseqs", _) if opt_val != "" => options.allowed_iseqs = Some(parse_jit_list(opt_val)),
+        ("log-compiled-iseqs", _) if opt_val != "" => {
+            // Truncate the file if it exists
+            std::fs::OpenOptions::new()
+                .create(true)
+                .write(true)
+                .truncate(true)
+                .open(opt_val)
+                .map_err(|e| eprintln!("Failed to open file '{}': {}", opt_val, e))
+                .ok();
+            options.log_compiled_iseqs = Some(opt_val.into());
+        }
+
         _ => return None, // Option name not recognized
     }
 

zjit/src/state.rs

@@ -136,6 +136,38 @@ impl ZJITState {
     pub fn get_counters() -> &'static mut Counters {
         &mut ZJITState::get_instance().counters
     }
+
+    /// Was --zjit-save-compiled-iseqs specified?
+    pub fn should_log_compiled_iseqs() -> bool {
+        ZJITState::get_instance().options.log_compiled_iseqs.is_some()
+    }
+
+    /// Log the name of a compiled ISEQ to the file specified in options.log_compiled_iseqs
+    pub fn log_compile(iseq_name: String) {
+        assert!(ZJITState::should_log_compiled_iseqs());
+        let filename = ZJITState::get_instance().options.log_compiled_iseqs.as_ref().unwrap();
+        use std::io::Write;
+        let mut file = match std::fs::OpenOptions::new().create(true).append(true).open(filename) {
+            Ok(f) => f,
+            Err(e) => {
+                eprintln!("ZJIT: Failed to create file '{}': {}", filename, e);
+                return;
+            }
+        };
+        if let Err(e) = writeln!(file, "{}", iseq_name) {
+            eprintln!("ZJIT: Failed to write to file '{}': {}", filename, e);
+        }
+    }
+
+    /// Check if we are allowed to compile a given ISEQ based on --zjit-allowed-iseqs
+    pub fn can_compile_iseq(iseq: cruby::IseqPtr) -> bool {
+        if let Some(ref allowed_iseqs) = ZJITState::get_instance().options.allowed_iseqs {
+            let name = cruby::iseq_get_location(iseq, 0);
+            allowed_iseqs.contains(&name)
+        } else {
+            true // If no restrictions, allow all ISEQs
+        }
+    }
 }
 
 /// Initialize ZJIT, given options allocated by rb_zjit_init_options()
@@ -148,7 +180,6 @@ pub extern "C" fn rb_zjit_init(options: *const u8) {
 
         let options = unsafe { Box::from_raw(options as *mut Options) };
         ZJITState::init(*options);
-        std::mem::drop(options);
 
         rb_bug_panic_hook();