Implement Options as part of ZJITState

Merged: https://github.com/ruby/ruby/pull/13131

internal/cmdlineopt.h

@@ -23,6 +23,9 @@ typedef struct ruby_cmdline_options {
     ruby_features_t warn;
     unsigned int dump;
     long backtrace_length_limit;
+#if USE_ZJIT
+    void *zjit;
+#endif
 
     const char *crash_report;
 
@@ -39,7 +42,6 @@ typedef struct ruby_cmdline_options {
 #if USE_YJIT
     unsigned int yjit: 1;
 #endif
-    unsigned int zjit: 1;
 } ruby_cmdline_options_t;
 
 struct ruby_opt_message {

ruby.c

@@ -1184,6 +1184,21 @@ setup_yjit_options(const char *s)
 }
 #endif
 
+#if USE_ZJIT
+static void
+setup_zjit_options(ruby_cmdline_options_t *opt, const char *s)
+{
+    // The option parsing is done in zjit/src/options.rs
+    extern void *rb_zjit_init_options(void);
+    extern bool rb_zjit_parse_option(void *options, const char *s);
+
+    if (!opt->zjit) opt->zjit = rb_zjit_init_options();
+    if (!rb_zjit_parse_option(opt->zjit, s)) {
+        rb_raise(rb_eRuntimeError, "invalid ZJIT option '%s' (--help will show valid zjit options)", s);
+    }
+}
+#endif
+
 /*
  * Following proc_*_option functions are tree kinds:
  *
@@ -1454,9 +1469,13 @@ proc_long_options(ruby_cmdline_options_t *opt, const char *s, long argc, char **
 #endif
     }
     else if (is_option_with_optarg("zjit", '-', true, false, false)) {
+#if USE_ZJIT
         FEATURE_SET(opt->features, FEATURE_BIT(zjit));
-        extern bool rb_zjit_parse_option();
+        setup_zjit_options(opt, s);
-        rb_zjit_parse_option();
+#else
+        rb_warn("Ruby was built without ZJIT support."
+                " You may need to install rustc to build Ruby with ZJIT.");
+#endif
     }
     else if (strcmp("yydebug", s) == 0) {
         if (envopt) goto noenvopt_long;
@@ -1799,9 +1818,8 @@ ruby_opt_init(ruby_cmdline_options_t *opt)
 #endif
 #if USE_ZJIT
     if (opt->zjit) {
-        fprintf(stderr, "test test\n");
+        extern void rb_zjit_init(void *options);
-        extern void rb_zjit_init();
+        rb_zjit_init(opt->zjit);
-        rb_zjit_init();
     }
 #endif
 
@@ -2339,8 +2357,9 @@ process_options(int argc, char **argv, ruby_cmdline_options_t *opt)
     }
 #endif
 #if USE_ZJIT
-    if (FEATURE_SET_P(opt->features, zjit)) {
+    if (FEATURE_SET_P(opt->features, zjit) && !opt->zjit) {
-        opt->zjit = true;
+        extern void *rb_zjit_init_options(void);
+        opt->zjit = rb_zjit_init_options();
     }
 #endif
 

zjit/src/codegen.rs

@@ -1,7 +1,9 @@
+use std::mem;
 use std::rc::Rc;
 use std::cell::RefCell;
 use crate::cruby::*;
 use crate::virtualmem::*;
+use crate::options::Options;
 
 /// Block of memory into which instructions can be assembled
 pub struct CodeBlock {
@@ -111,18 +113,20 @@ impl crate::virtualmem::CodePtrBase for CodeBlock {
 pub struct ZJITState {
     /// Inline code block (fast path)
     code_block: CodeBlock,
+
+    /// ZJIT command-line options
+    options: Options,
 }
 
 /// Private singleton instance of the codegen globals
 static mut ZJIT_STATE: Option<ZJITState> = None;
 
 impl ZJITState {
-    /// Initialize the ZJIT globals
+    /// Initialize the ZJIT globals, given options allocated by rb_zjit_init_options()
-    pub fn init() {
+    pub fn init(options: *const u8) {
-        let exec_mem_size: usize = 64 * 1024 * 1024; // TODO: support the option
-
         #[cfg(not(test))]
         let cb = {
+            let exec_mem_size: usize = 64 * 1024 * 1024; // TODO: implement the option
             let virt_block: *mut u8 = unsafe { rb_zjit_reserve_addr_space(64 * 1024 * 1024) };
 
             // Memory protection syscalls need page-aligned addresses, so check it here. Assuming
@@ -153,16 +157,18 @@ impl ZJITState {
             CodeBlock::new(mem_block.clone())
         };
 
-        #[cfg(not(test))]
+        let options = unsafe { Box::from_raw(options as *mut Options) };
-        let zjit_state = ZJITState {
+        #[cfg(not(test))] // TODO: can we get rid of this #[cfg]?
-            code_block: cb,
+        {
-        };
+            let zjit_state = ZJITState {
+                code_block: cb,
+                options: *options,
+            };
 
-        // Initialize the codegen globals instance
+            // Initialize the codegen globals instance
-        #[cfg(not(test))]
+            unsafe { ZJIT_STATE = Some(zjit_state); }
-        unsafe {
-            ZJIT_STATE = Some(zjit_state);
         }
+        mem::drop(options);
     }
 
     /// Get a mutable reference to the codegen globals instance
@@ -174,4 +180,9 @@ impl ZJITState {
     pub fn get_code_block() -> &'static mut CodeBlock {
         &mut ZJITState::get_instance().code_block
     }
+
+    // Get a mutable reference to the options
+    pub fn get_options() -> &'static mut Options {
+        &mut ZJITState::get_instance().options
+    }
 }

zjit/src/lib.rs

@@ -10,6 +10,7 @@ mod virtualmem;
 mod asm;
 mod backend;
 mod disasm;
+mod options;
 
 use backend::x86_emit;
 use codegen::ZJITState;
@@ -19,12 +20,13 @@ use crate::cruby::*;
 #[no_mangle]
 pub static mut rb_zjit_enabled_p: bool = false;
 
+/// Initialize ZJIT, given options allocated by rb_zjit_init_options()
 #[no_mangle]
-pub extern "C" fn rb_zjit_init() {
+pub extern "C" fn rb_zjit_init(options: *const u8) {
     // Catch panics to avoid UB for unwinding into C frames.
     // See https://doc.rust-lang.org/nomicon/exception-safety.html
     let result = std::panic::catch_unwind(|| {
-        ZJITState::init();
+        ZJITState::init(options);
 
         rb_bug_panic_hook();
 
@@ -71,12 +73,6 @@ fn rb_bug_panic_hook() {
     }));
 }
 
-#[no_mangle]
-pub extern "C" fn rb_zjit_parse_option() -> bool {
-    println!("parsing zjit options");
-    false
-}
-
 #[no_mangle]
 pub extern "C" fn rb_zjit_iseq_gen_entry_point(iseq: IseqPtr, _ec: EcPtr) -> *const u8 {
     ir::iseq_to_ssa(iseq).unwrap();

zjit/src/options.rs

@@ -0,0 +1,56 @@
+use std::{ffi::CStr, os::raw::c_char};
+
+#[derive(Clone, Copy, Debug)]
+pub struct Options {
+    /// Dump all compiled instructions of target cb.
+    pub dump_disasm: bool,
+}
+
+/// Allocate Options on the heap, initialize it, and return the address of it.
+/// The return value will be modified by rb_zjit_parse_option() and then
+/// passed to rb_zjit_init() for initialization.
+#[no_mangle]
+pub extern "C" fn rb_zjit_init_options() -> *const u8 {
+    let options = Options {
+        dump_disasm: false,
+    };
+    Box::into_raw(Box::new(options)) as *const u8
+}
+
+/// Parse a --zjit* command-line flag
+#[no_mangle]
+pub extern "C" fn rb_zjit_parse_option(options: *const u8, str_ptr: *const c_char) -> bool {
+    let options = unsafe { &mut *(options as *mut Options) };
+    parse_option(options, str_ptr).is_some()
+}
+
+/// 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-".
+fn parse_option(options: &mut Options, str_ptr: *const std::os::raw::c_char) -> Option<()> {
+    let c_str: &CStr = unsafe { CStr::from_ptr(str_ptr) };
+    let opt_str: &str = c_str.to_str().ok()?;
+
+    // Split the option name and value strings
+    // Note that some options do not contain an assignment
+    let parts = opt_str.split_once('=');
+    let (opt_name, opt_val) = match parts {
+        Some((before_eq, after_eq)) => (before_eq, after_eq),
+        None => (opt_str, ""),
+    };
+
+    // Match on the option name and value strings
+    match (opt_name, opt_val) {
+        ("", "") => {}, // Simply --zjit
+
+        ("dump-disasm", "") => options.dump_disasm = true,
+
+        // Option name not recognized
+        _ => {
+            return None;
+        }
+    }
+
+    // Option successfully parsed
+    return Some(());
+}