summaryrefslogtreecommitdiff
path: root/yjit/src/cruby.rs
blob: 25149ab7304d45813b30091e43e5c64e64da0d11 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
//! This module deals with making relevant C functions available to Rust YJIT.
//! Some C functions we use we maintain, some are public C extension APIs,
//! some are internal CRuby APIs.
//!
//! ## General notes about linking
//!
//! The YJIT crate compiles to a native static library, which for our purposes
//! we can understand as a collection of object files. On ELF platforms at least,
//! object files can refer to "external symbols" which we could take some
//! liberty and understand as assembly labels that refer to code defined in other
//! object files resolved when linking. When we are linking, say to produce miniruby,
//! the linker resolves and put concrete addresses for each usage of C function in
//! the Rust static library.
//!
//! By declaring external functions and using them, we are asserting the symbols
//! we use have definition in one of the object files we pass to the linker. Declaring
//! a function here that has no definition anywhere causes a linking error.
//!
//! There are more things going on during linking and this section makes a lot of
//! simplifications but hopefully this gives a good enough working mental model.
//!
//! ## Difference from example in the Rustonomicon
//!
//! You might be wondering about why this is different from the [FFI example]
//! in the Nomicon, an official book about Unsafe Rust.
//!
//! There is no `#[link]` attribute because we are not linking against an external
//! library, but rather implicitly asserting that we'll supply a concrete definition
//! for all C functions we call, similar to how pure C projects put functions
//! across different compilation units and link them together.
//!
//! TODO(alan): is the model different enough on Windows that this setup is unworkable?
//!             Seems prudent to at least learn more about Windows binary tooling before
//!             committing to a design.
//!
//! Alan recommends reading the Nomicon cover to cover as he thinks the book is
//! not very long in general and especially for something that can save hours of
//! debugging Undefined Behavior (UB) down the road.
//!
//! UBs can cause Safe Rust to crash, at which point it's hard to tell which
//! usage of `unsafe` in the codebase invokes UB. Providing safe Rust interface
//! wrapping `unsafe` Rust is a good technique, but requires practice and knowledge
//! about what's well defined and what's undefined.
//!
//! For an extremely advanced example of building safe primitives using Unsafe Rust,
//! see the [GhostCell] paper. Some parts of the paper assume less background knowledge
//! than other parts, so there should be learning opportunities in it for all experience
//! levels.
//!
//! ## Binding generation
//!
//! For the moment declarations on the Rust side are hand written. The code is boilerplate
//! and could be generated automatically with a custom tooling that depend on
//! rust-lang/rust-bindgen. The output Rust code could be checked in to version control
//! and verified on CI like `make update-deps`.
//!
//! Upsides for this design:
//!  - the YJIT static lib that links with miniruby and friends will not need bindgen
//!    as a dependency at all. This is an important property so Ruby end users can
//!    build a YJIT enabled Ruby with no internet connection using a release tarball
//!  - Less hand-typed boilerplate
//!  - Helps reduce risk of C definitions and Rust declaration going out of sync since
//!    CI verifies synchronicity
//!
//! Downsides and known unknowns:
//!  - Using rust-bindgen this way seems unusual. We might be depending on parts
//!    that the project is not committed to maintaining
//!  - This setup assumes rust-bindgen gives deterministic output, which can't be taken
//!    for granted
//!  - YJIT contributors will need to install libclang on their system to get rust-bindgen
//!    to work if they want to run the generation tool locally
//!
//! The elephant in the room is that we'll still need to use Unsafe Rust to call C functions,
//! and the binding generation can't magically save us from learning Unsafe Rust.
//!
//!
//! [FFI example]: https://doc.rust-lang.org/nomicon/ffi.html
//! [GhostCell]: http://plv.mpi-sws.org/rustbelt/ghostcell/

// CRuby types use snake_case. Allow them so we use one name across languages.
#![allow(non_camel_case_types)]
// A lot of imported CRuby globals aren't all-caps
#![allow(non_upper_case_globals)]

use std::convert::From;
use std::ffi::CString;
use std::os::raw::{c_char, c_int, c_long, c_uint};
use std::panic::{catch_unwind, UnwindSafe};

// We check that we can do this with the configure script and a couple of
// static asserts. u64 and not usize to play nice with lowering to x86.
pub type size_t = u64;

/// A type alias for the redefinition flags coming from CRuby. These are just
/// shifted 1s but not explicitly an enum.
pub type RedefinitionFlag = u32;

#[allow(dead_code)]
mod autogened {
    use super::*;
    // Textually include output from rust-bindgen as suggested by its user guide.
    include!("cruby_bindings.inc.rs");
}
pub use autogened::*;

// TODO: For #defines that affect memory layout, we need to check for them
// on build and fail if they're wrong. e.g. USE_FLONUM *must* be true.

// These are functions we expose from vm_insnhelper.c, not in any header.
// Parsing it would result in a lot of duplicate definitions.
// Use bindgen for functions that are defined in headers or in yjit.c.
#[cfg_attr(test, allow(unused))] // We don't link against C code when testing
extern "C" {
    pub fn rb_vm_splat_array(flag: VALUE, ary: VALUE) -> VALUE;
    pub fn rb_vm_concat_array(ary1: VALUE, ary2st: VALUE) -> VALUE;
    pub fn rb_vm_defined(
        ec: EcPtr,
        reg_cfp: CfpPtr,
        op_type: rb_num_t,
        obj: VALUE,
        v: VALUE,
    ) -> bool;
    pub fn rb_vm_set_ivar_idx(obj: VALUE, idx: u32, val: VALUE) -> VALUE;
    pub fn rb_vm_setinstancevariable(iseq: IseqPtr, obj: VALUE, id: ID, val: VALUE, ic: IVC);
    pub fn rb_aliased_callable_method_entry(
        me: *const rb_callable_method_entry_t,
    ) -> *const rb_callable_method_entry_t;
    pub fn rb_vm_getclassvariable(iseq: IseqPtr, cfp: CfpPtr, id: ID, ic: ICVARC) -> VALUE;
    pub fn rb_vm_setclassvariable(
        iseq: IseqPtr,
        cfp: CfpPtr,
        id: ID,
        val: VALUE,
        ic: ICVARC,
    ) -> VALUE;
    pub fn rb_vm_ic_hit_p(ic: IC, reg_ep: *const VALUE) -> bool;
    pub fn rb_str_bytesize(str: VALUE) -> VALUE;
}

// Renames
pub use rb_insn_name as raw_insn_name;
pub use rb_insn_len as raw_insn_len;
pub use rb_yarv_class_of as CLASS_OF;
pub use rb_get_ec_cfp as get_ec_cfp;
pub use rb_get_cfp_pc as get_cfp_pc;
pub use rb_get_cfp_sp as get_cfp_sp;
pub use rb_get_cfp_self as get_cfp_self;
pub use rb_get_cfp_ep as get_cfp_ep;
pub use rb_get_cfp_ep_level as get_cfp_ep_level;
pub use rb_get_cme_def_type as get_cme_def_type;
pub use rb_get_cme_def_body_attr_id as get_cme_def_body_attr_id;
pub use rb_get_cme_def_body_optimized_type as get_cme_def_body_optimized_type;
pub use rb_get_cme_def_body_optimized_index as get_cme_def_body_optimized_index;
pub use rb_get_cme_def_body_cfunc as get_cme_def_body_cfunc;
pub use rb_get_def_method_serial as get_def_method_serial;
pub use rb_get_def_original_id as get_def_original_id;
pub use rb_get_mct_argc as get_mct_argc;
pub use rb_get_mct_func as get_mct_func;
pub use rb_get_def_iseq_ptr as get_def_iseq_ptr;
pub use rb_iseq_encoded_size as get_iseq_encoded_size;
pub use rb_get_iseq_body_local_iseq as get_iseq_body_local_iseq;
pub use rb_get_iseq_body_iseq_encoded as get_iseq_body_iseq_encoded;
pub use rb_get_iseq_body_stack_max as get_iseq_body_stack_max;
pub use rb_get_iseq_flags_has_opt as get_iseq_flags_has_opt;
pub use rb_get_iseq_flags_has_kw as get_iseq_flags_has_kw;
pub use rb_get_iseq_flags_has_rest as get_iseq_flags_has_rest;
pub use rb_get_iseq_flags_ruby2_keywords as get_iseq_flags_ruby2_keywords;
pub use rb_get_iseq_flags_has_post as get_iseq_flags_has_post;
pub use rb_get_iseq_flags_has_kwrest as get_iseq_flags_has_kwrest;
pub use rb_get_iseq_flags_has_block as get_iseq_flags_has_block;
pub use rb_get_iseq_flags_has_accepts_no_kwarg as get_iseq_flags_has_accepts_no_kwarg;
pub use rb_get_iseq_body_local_table_size as get_iseq_body_local_table_size;
pub use rb_get_iseq_body_param_keyword as get_iseq_body_param_keyword;
pub use rb_get_iseq_body_param_size as get_iseq_body_param_size;
pub use rb_get_iseq_body_param_lead_num as get_iseq_body_param_lead_num;
pub use rb_get_iseq_body_param_opt_num as get_iseq_body_param_opt_num;
pub use rb_get_iseq_body_param_opt_table as get_iseq_body_param_opt_table;
pub use rb_get_cikw_keyword_len as get_cikw_keyword_len;
pub use rb_get_cikw_keywords_idx as get_cikw_keywords_idx;
pub use rb_get_call_data_ci as get_call_data_ci;
pub use rb_yarv_str_eql_internal as rb_str_eql_internal;
pub use rb_yarv_ary_entry_internal as rb_ary_entry_internal;
pub use rb_yarv_fix_mod_fix as rb_fix_mod_fix;
pub use rb_FL_TEST as FL_TEST;
pub use rb_FL_TEST_RAW as FL_TEST_RAW;
pub use rb_RB_TYPE_P as RB_TYPE_P;
pub use rb_BASIC_OP_UNREDEFINED_P as BASIC_OP_UNREDEFINED_P;
pub use rb_RSTRUCT_LEN as RSTRUCT_LEN;
pub use rb_RSTRUCT_SET as RSTRUCT_SET;
pub use rb_vm_ci_argc as vm_ci_argc;
pub use rb_vm_ci_mid as vm_ci_mid;
pub use rb_vm_ci_flag as vm_ci_flag;
pub use rb_vm_ci_kwarg as vm_ci_kwarg;
pub use rb_METHOD_ENTRY_VISI as METHOD_ENTRY_VISI;
pub use rb_RCLASS_ORIGIN as RCLASS_ORIGIN;

/// Helper so we can get a Rust string for insn_name()
pub fn insn_name(opcode: usize) -> String {
    use std::ffi::CStr;

    unsafe {
        // Look up Ruby's NULL-terminated insn name string
        let op_name = raw_insn_name(VALUE(opcode));

        // Convert the op name C string to a Rust string and concat
        let op_name = CStr::from_ptr(op_name).to_str().unwrap();

        // Convert into an owned string
        op_name.to_string()
    }
}

#[allow(unused_variables)]
pub fn insn_len(opcode: usize) -> u32 {
    #[cfg(test)]
    panic!("insn_len is a CRuby function, and we don't link against CRuby for Rust testing!");

    #[cfg(not(test))]
    unsafe {
        raw_insn_len(VALUE(opcode)).try_into().unwrap()
    }
}

/// Opaque iseq type for opaque iseq pointers from vm_core.h
/// See: <https://doc.rust-lang.org/nomicon/ffi.html#representing-opaque-structs>
#[repr(C)]
pub struct rb_iseq_t {
    _data: [u8; 0],
    _marker: core::marker::PhantomData<(*mut u8, core::marker::PhantomPinned)>,
}

/// An object handle similar to VALUE in the C code. Our methods assume
/// that this is a handle. Sometimes the C code briefly uses VALUE as
/// an unsigned integer type and don't necessarily store valid handles but
/// thankfully those cases are rare and don't cross the FFI boundary.
#[derive(Copy, Clone, PartialEq, Eq, Debug, Hash)]
#[repr(transparent)] // same size and alignment as simply `usize`
pub struct VALUE(pub usize);

/// Pointer to an ISEQ
pub type IseqPtr = *const rb_iseq_t;

/// Opaque execution-context type from vm_core.h
#[repr(C)]
pub struct rb_execution_context_struct {
    _data: [u8; 0],
    _marker: core::marker::PhantomData<(*mut u8, core::marker::PhantomPinned)>,
}
/// Alias for rb_execution_context_struct used by CRuby sometimes
pub type rb_execution_context_t = rb_execution_context_struct;

/// Pointer to an execution context (rb_execution_context_struct)
pub type EcPtr = *const rb_execution_context_struct;

// From method.h
#[repr(C)]
pub struct rb_method_definition_t {
    _data: [u8; 0],
    _marker: core::marker::PhantomData<(*mut u8, core::marker::PhantomPinned)>,
}
type rb_method_definition_struct = rb_method_definition_t;

/// Opaque cfunc type from method.h
#[repr(C)]
pub struct rb_method_cfunc_t {
    _data: [u8; 0],
    _marker: core::marker::PhantomData<(*mut u8, core::marker::PhantomPinned)>,
}

/// Opaque FILE type from the C standard library
#[repr(C)]
pub struct FILE {
    _data: [u8; 0],
    _marker: core::marker::PhantomData<(*mut u8, core::marker::PhantomPinned)>,
}

/// Opaque call-cache type from vm_callinfo.h
#[repr(C)]
pub struct rb_callcache {
    _data: [u8; 0],
    _marker: core::marker::PhantomData<(*mut u8, core::marker::PhantomPinned)>,
}

/// Opaque call-info type from vm_callinfo.h
#[repr(C)]
pub struct rb_callinfo_kwarg {
    _data: [u8; 0],
    _marker: core::marker::PhantomData<(*mut u8, core::marker::PhantomPinned)>,
}

/// Opaque control_frame (CFP) struct from vm_core.h
#[repr(C)]
pub struct rb_control_frame_struct {
    _data: [u8; 0],
    _marker: core::marker::PhantomData<(*mut u8, core::marker::PhantomPinned)>,
}

/// Pointer to a control frame pointer (CFP)
pub type CfpPtr = *mut rb_control_frame_struct;

/// Opaque struct from vm_core.h
#[repr(C)]
pub struct rb_cref_t {
    _data: [u8; 0],
    _marker: core::marker::PhantomData<(*mut u8, core::marker::PhantomPinned)>,
}

impl VALUE {
    /// Dump info about the value to the console similarly to rp(VALUE)
    pub fn dump_info(self) {
        unsafe { rb_obj_info_dump(self) }
    }

    /// Return whether the value is truthy or falsy in Ruby -- only nil and false are falsy.
    pub fn test(self) -> bool {
        let VALUE(cval) = self;
        let VALUE(qnilval) = Qnil;
        (cval & !qnilval) != 0
    }

    /// Return true if the number is an immediate integer, flonum or static symbol
    fn immediate_p(self) -> bool {
        let VALUE(cval) = self;
        (cval & 7) != 0
    }

    /// Return true if the value is a Ruby immediate integer, flonum, static symbol, nil or false
    pub fn special_const_p(self) -> bool {
        self.immediate_p() || !self.test()
    }

    /// Return true if the value is a Ruby Fixnum (immediate-size integer)
    pub fn fixnum_p(self) -> bool {
        let VALUE(cval) = self;
        (cval & 1) == 1
    }

    /// Return true if the value is an immediate Ruby floating-point number (flonum)
    pub fn flonum_p(self) -> bool {
        let VALUE(cval) = self;
        (cval & 3) == 2
    }

    /// Return true for a static (non-heap) Ruby symbol
    pub fn static_sym_p(self) -> bool {
        let VALUE(cval) = self;
        (cval & 0xff) == RUBY_SYMBOL_FLAG
    }

    /// Returns true or false depending on whether the value is nil
    pub fn nil_p(self) -> bool {
        self == Qnil
    }

    /// Read the flags bits from the RBasic object, then return a Ruby type enum (e.g. RUBY_T_ARRAY)
    pub fn builtin_type(self) -> ruby_value_type {
        assert!(!self.special_const_p());

        let VALUE(cval) = self;
        let rbasic_ptr = cval as *const RBasic;
        let flags_bits: usize = unsafe { (*rbasic_ptr).flags }.as_usize();
        (flags_bits & (RUBY_T_MASK as usize)) as ruby_value_type
    }

    pub fn class_of(self) -> VALUE {
        unsafe { CLASS_OF(self) }
    }

    pub fn as_isize(self) -> isize {
        let VALUE(is) = self;
        is as isize
    }

    pub fn as_i32(self) -> i32 {
        self.as_i64().try_into().unwrap()
    }

    pub fn as_u32(self) -> u32 {
        let VALUE(i) = self;
        i.try_into().unwrap()
    }

    pub fn as_i64(self) -> i64 {
        let VALUE(i) = self;
        i as i64
    }

    pub fn as_u64(self) -> u64 {
        let VALUE(i) = self;
        i.try_into().unwrap()
    }

    pub fn as_usize(self) -> usize {
        let VALUE(us) = self;
        us as usize
    }

    pub fn as_ptr<T>(self) -> *const T {
        let VALUE(us) = self;
        us as *const T
    }

    pub fn as_mut_ptr<T>(self) -> *mut T {
        let VALUE(us) = self;
        us as *mut T
    }

    /// For working with opague pointers and encoding null check.
    /// Similar to [std::ptr::NonNull], but for `*const T`. `NonNull<T>`
    /// is for `*mut T` while our C functions are setup to use `*const T`.
    /// Casting from `NonNull<T>` to `*const T` is too noisy.
    pub fn as_optional_ptr<T>(self) -> Option<*const T> {
        let ptr: *const T = self.as_ptr();

        if ptr.is_null() {
            None
        } else {
            Some(ptr)
        }
    }

    /// Assert that `self` is an iseq in debug builds
    pub fn as_iseq(self) -> IseqPtr {
        let ptr: IseqPtr = self.as_ptr();

        #[cfg(debug_assertions)]
        if !ptr.is_null() {
            unsafe { rb_assert_iseq_handle(self) }
        }

        ptr
    }

    /// Assert that `self` is a method entry in debug builds
    pub fn as_cme(self) -> *const rb_callable_method_entry_t {
        let ptr: *const rb_callable_method_entry_t = self.as_ptr();

        #[cfg(debug_assertions)]
        if !ptr.is_null() {
            unsafe { rb_assert_cme_handle(self) }
        }

        ptr
    }
}

impl VALUE {
    pub fn fixnum_from_usize(item: usize) -> Self {
        assert!(item <= (RUBY_FIXNUM_MAX as usize)); // An unsigned will always be greater than RUBY_FIXNUM_MIN
        let k: usize = item.wrapping_add(item.wrapping_add(1));
        VALUE(k)
    }
}

impl From<IseqPtr> for VALUE {
    /// For `.into()` convenience
    fn from(iseq: IseqPtr) -> Self {
        VALUE(iseq as usize)
    }
}

impl From<*const rb_callable_method_entry_t> for VALUE {
    /// For `.into()` convenience
    fn from(cme: *const rb_callable_method_entry_t) -> Self {
        VALUE(cme as usize)
    }
}

impl From<VALUE> for u64 {
    fn from(value: VALUE) -> Self {
        let VALUE(uimm) = value;
        uimm as u64
    }
}

impl From<VALUE> for i64 {
    fn from(value: VALUE) -> Self {
        let VALUE(uimm) = value;
        assert!(uimm <= (i64::MAX as usize));
        uimm as i64
    }
}

impl From<VALUE> for i32 {
    fn from(value: VALUE) -> Self {
        let VALUE(uimm) = value;
        assert!(uimm <= (i32::MAX as usize));
        uimm.try_into().unwrap()
    }
}

impl From<VALUE> for u16 {
    fn from(value: VALUE) -> Self {
        let VALUE(uimm) = value;
        uimm.try_into().unwrap()
    }
}

/// Produce a Ruby string from a Rust string slice
#[cfg(feature = "asm_comments")]
pub fn rust_str_to_ruby(str: &str) -> VALUE {
    unsafe { rb_utf8_str_new(str.as_ptr() as *const _, str.len() as i64) }
}

/// Produce a Ruby symbol from a Rust string slice
pub fn rust_str_to_sym(str: &str) -> VALUE {
    let c_str = CString::new(str).unwrap();
    let c_ptr: *const c_char = c_str.as_ptr();

    unsafe { rb_id2sym(rb_intern(c_ptr)) }
}

/// A location in Rust code for integrating with debugging facilities defined in C.
/// Use the [src_loc!] macro to crate an instance.
pub struct SourceLocation {
    pub file: CString,
    pub line: c_int,
}

/// Make a [SourceLocation] at the current spot.
macro_rules! src_loc {
    () => {
        // NOTE(alan): `CString::new` allocates so we might want to limit this to debug builds.
        $crate::cruby::SourceLocation {
            file: std::ffi::CString::new(file!()).unwrap(), // ASCII source file paths
            line: line!().try_into().unwrap(),              // not that many lines
        }
    };
}

pub(crate) use src_loc;

/// Run GC write barrier. Required after making a new edge in the object reference
/// graph from `old` to `young`.
macro_rules! obj_written {
    ($old: expr, $young: expr) => {
        let (old, young): (VALUE, VALUE) = ($old, $young);
        let src_loc = $crate::cruby::src_loc!();
        unsafe { rb_yjit_obj_written(old, young, src_loc.file.as_ptr(), src_loc.line) };
    };
}
pub(crate) use obj_written;

/// Acquire the VM lock, make sure all other Ruby threads are asleep then run
/// some code while holding the lock. Returns whatever `func` returns.
/// Use with [src_loc!].
///
/// Required for code patching in the presence of ractors.
pub fn with_vm_lock<F, R>(loc: SourceLocation, func: F) -> R
where
    F: FnOnce() -> R + UnwindSafe,
{
    let file = loc.file.as_ptr();
    let line = loc.line;
    let mut recursive_lock_level: c_uint = 0;

    unsafe { rb_yjit_vm_lock_then_barrier(&mut recursive_lock_level, file, line) };

    let ret = match catch_unwind(func) {
        Ok(result) => result,
        Err(_) => {
            // Theoretically we can recover from some of these panics,
            // but it's too late if the unwind reaches here.
            use std::{process, str};

            let _ = catch_unwind(|| {
                // IO functions can panic too.
                eprintln!(
                    "YJIT panicked while holding VM lock acquired at {}:{}. Aborting...",
                    str::from_utf8(loc.file.as_bytes()).unwrap_or("<not utf8>"),
                    line,
                );
            });
            process::abort();
        }
    };

    unsafe { rb_yjit_vm_unlock(&mut recursive_lock_level, file, line) };

    ret
}

// Non-idiomatic capitalization for consistency with CRuby code
#[allow(non_upper_case_globals)]
pub const Qfalse: VALUE = VALUE(0);
#[allow(non_upper_case_globals)]
pub const Qnil: VALUE = VALUE(8);
#[allow(non_upper_case_globals)]
pub const Qtrue: VALUE = VALUE(20);
#[allow(non_upper_case_globals)]
pub const Qundef: VALUE = VALUE(52);

#[allow(unused)]
mod manual_defs {
    use super::*;

    pub const SIZEOF_VALUE: usize = 8;
    pub const SIZEOF_VALUE_I32: i32 = SIZEOF_VALUE as i32;

    pub const RUBY_LONG_MIN: isize = std::os::raw::c_long::MIN as isize;
    pub const RUBY_LONG_MAX: isize = std::os::raw::c_long::MAX as isize;

    pub const RUBY_FIXNUM_MIN: isize = RUBY_LONG_MIN / 2;
    pub const RUBY_FIXNUM_MAX: isize = RUBY_LONG_MAX / 2;
    pub const RUBY_FIXNUM_FLAG: usize = 0x1;

    // All these are defined in include/ruby/internal/special_consts.h,
    // in the same enum as RUBY_Qfalse, etc.
    // Do we want to switch to using Ruby's definition of Qnil, Qfalse, etc?
    pub const RUBY_SYMBOL_FLAG: usize = 0x0c;
    pub const RUBY_FLONUM_FLAG: usize = 0x2;
    pub const RUBY_FLONUM_MASK: usize = 0x3;
    pub const RUBY_SPECIAL_SHIFT: usize = 8;
    pub const RUBY_IMMEDIATE_MASK: usize = 0x7;

    // From vm_callinfo.h - uses calculation that seems to confuse bindgen
    pub const VM_CALL_ARGS_SPLAT: u32 = 1 << VM_CALL_ARGS_SPLAT_bit;
    pub const VM_CALL_ARGS_BLOCKARG: u32 = 1 << VM_CALL_ARGS_BLOCKARG_bit;
    pub const VM_CALL_FCALL: u32 = 1 << VM_CALL_FCALL_bit;
    pub const VM_CALL_KWARG: u32 = 1 << VM_CALL_KWARG_bit;
    pub const VM_CALL_KW_SPLAT: u32 = 1 << VM_CALL_KW_SPLAT_bit;
    pub const VM_CALL_TAILCALL: u32 = 1 << VM_CALL_TAILCALL_bit;
    pub const VM_CALL_ZSUPER : u32 = 1 << VM_CALL_ZSUPER_bit;

    // From internal/struct.h - in anonymous enum, so we can't easily import it
    pub const RSTRUCT_EMBED_LEN_MASK: usize = (RUBY_FL_USER2 | RUBY_FL_USER1) as usize;

    // From iseq.h - via a different constant, which seems to confuse bindgen
    pub const ISEQ_TRANSLATED: usize = RUBY_FL_USER7 as usize;

    // We'll need to encode a lot of Ruby struct/field offsets as constants unless we want to
    // redeclare all the Ruby C structs and write our own offsetof macro. For now, we use constants.
    pub const RUBY_OFFSET_RBASIC_FLAGS: i32 = 0; // struct RBasic, field "flags"
    pub const RUBY_OFFSET_RBASIC_KLASS: i32 = 8; // struct RBasic, field "klass"
    pub const RUBY_OFFSET_RARRAY_AS_HEAP_LEN: i32 = 16; // struct RArray, subfield "as.heap.len"
    pub const RUBY_OFFSET_RARRAY_AS_HEAP_PTR: i32 = 32; // struct RArray, subfield "as.heap.ptr"
    pub const RUBY_OFFSET_RARRAY_AS_ARY: i32 = 16; // struct RArray, subfield "as.ary"

    pub const RUBY_OFFSET_RSTRUCT_AS_HEAP_PTR: i32 = 24; // struct RStruct, subfield "as.heap.ptr"
    pub const RUBY_OFFSET_RSTRUCT_AS_ARY: i32 = 16; // struct RStruct, subfield "as.ary"

    // Constants from rb_control_frame_t vm_core.h
    pub const RUBY_OFFSET_CFP_PC: i32 = 0;
    pub const RUBY_OFFSET_CFP_SP: i32 = 8;
    pub const RUBY_OFFSET_CFP_ISEQ: i32 = 16;
    pub const RUBY_OFFSET_CFP_SELF: i32 = 24;
    pub const RUBY_OFFSET_CFP_EP: i32 = 32;
    pub const RUBY_OFFSET_CFP_BLOCK_CODE: i32 = 40;
    pub const RUBY_OFFSET_CFP_BP: i32 = 48; // field __bp__
    pub const RUBY_OFFSET_CFP_JIT_RETURN: i32 = 56;
    pub const RUBY_SIZEOF_CONTROL_FRAME: usize = 64;

    // Constants from rb_execution_context_t vm_core.h
    pub const RUBY_OFFSET_EC_CFP: i32 = 16;
    pub const RUBY_OFFSET_EC_INTERRUPT_FLAG: i32 = 32; // rb_atomic_t (u32)
    pub const RUBY_OFFSET_EC_INTERRUPT_MASK: i32 = 36; // rb_atomic_t (u32)
    pub const RUBY_OFFSET_EC_THREAD_PTR: i32 = 48;

    // Constants from rb_thread_t in vm_core.h
    pub const RUBY_OFFSET_THREAD_SELF: i32 = 16;

    // Constants from iseq_inline_constant_cache (IC) and iseq_inline_constant_cache_entry (ICE) in vm_core.h
    pub const RUBY_OFFSET_IC_ENTRY: i32 = 0;
    pub const RUBY_OFFSET_ICE_VALUE: i32 = 8;
}
pub use manual_defs::*;